]> gcc.gnu.org Git - gcc.git/blob - gcc/ChangeLog
altivec.md (vec_unpacku_hi_v16qi): Adjust for little endian.
[gcc.git] / gcc / ChangeLog
1 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2
3 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
4 little endian.
5 (vec_unpacku_hi_v8hi): Likewise.
6 (vec_unpacku_lo_v16qi): Likewise.
7 (vec_unpacku_lo_v8hi): Likewise.
8
9 2013-10-20 Jan Hubicka <jh@suse.cz>
10
11 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
12 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
13 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
14
15 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
16
17 * config/mips/mips.h (ISA_HAS_WSBH): Define.
18 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
19 constants.
20 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
21
22 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
23
24 PR target/58603
25 * system.h: Undef m_slot.
26
27 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
28
29 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
30 all elements for both endian flavors.
31
32 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
33
34 PR target/58792
35 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
36 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
37 and SI_REG for 64bit SYSV ABI targets.
38
39 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
40
41 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
42 to multi_reg_return. Clarify that we are skipping USEs of multiple
43 return registers. Use bool type where appropriate.
44
45 2013-10-18 Jan Hubicka <jh@suse.cz>
46
47 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
48 for cold functions.
49 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
50 (X86_TUNE_PUSH_MEMORY): Likewise.
51 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
52 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
53 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
54 New.
55 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
56 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
57 Remove.
58 (ix86_option_override_internal): Update to use tune features instead
59 of variables.
60
61 2013-10-18 Cong Hou <congh@google.com>
62
63 PR tree-optimization/58508
64 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
65 statement that contains data refs with zero-step.
66
67 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
68
69 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
70 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
71 * gimple-low.c (gimple_check_call_arg,
72 gimple_check_call_matching_types): Move to cgraph.c.
73 * gimple-low.h: Remove prototype.
74 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
75 Relocate from gimple-low.c.
76 * cgraph.h: Add prototype. Don't include basic-block.h.
77 * gimplify.c: Add gimple-low to include list.
78 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
79 * tree-eh.c: Add gimple-low to include list.
80 * tree-nested.c: Likewise.
81 * cfgexpand.c: Add tree-ssa-address.h to include list.
82 * expr.c: Likewise.
83 * gimple-fold.c: Likewise.
84 * gimple-ssa-strength-reduction.c: Likewise.
85 * trans-mem.c: Likewise.
86 * tree-mudflap.c: Likewise.
87 * tree-ssa-loop-ivopts.c: Likewise.
88 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
89 (degenerate_phi_result): Move to tree-phinodes.c.
90 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
91 * tree-ssa-threadedge.c: Likewise.
92 * tree-vrp.c: Likewise.
93 * tree-phinodes.c (degenerate_phi_result): Relocate here.
94 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
95 * tree-phinodes.h (degenerate_phi_result): Add prototype.
96 * tree-ssa-copy.c: Include tree-ssa-dom.h.
97 * tree-ssa-forwprop.c: Likewise.
98 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
99 pass_data_cleanup_cfg_post_optimizing,
100 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
101 * tree-optimize.c: Delete File.
102 * graphite.c: Include tree-cfgcleanup.h.
103 * passes.c: Likewise.
104 * tree-cfg.c: Likewise.
105 * tree-profile.c: Likewise.
106 * tree-ssa-dse.c: Likewise.
107 * tree-ssa-loop-ivcanon.c: Likewise.
108 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
109 * tree-outof-ssa.c: Include sbitmap.h.
110 * tree-ssa-live.c: Likewise.
111 * tree-ssa-propagate.c: Likewise.
112 * tree-ssa-structalias.c: Likewise.
113 * tree-stdarg.c: Likewise.
114 * Makefile.in (OBJS): Delete tree-optimize.o.
115 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
116 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
117 * varasm.c: Include basic-block.h.
118 * cfgloop.h: Include function.h instead of basic-block.h
119 (bb_loop_depth): Move to cfgloop.c.
120 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
121
122 2013-10-18 Teresa Johnson <tejohnson@google.com>
123
124 * predict.c (probably_never_executed): Compare frequency-based
125 count to number of training runs.
126 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
127
128 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
129
130 * config/arm/arm.c (cortexa9_extra_costs): New table.
131 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
132
133 2013-10-18 Jeff Law <law@redhat.com>
134
135 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
136
137 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
138 (register_jump_thread): Add "registered_jump_thread" debug
139 counter support.
140 * dbgcnt.def (registered_jump_thread): New debug counter.
141
142 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
143
144 * config/rs6000/rs6000.c: Include cgraph.h.
145
146 2013-10-18 Teresa Johnson <tejohnson@google.com>
147
148 * tree-ssa-tail-merge.c (replace_block_by): Update edge
149 weights during merging.
150
151 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
152
153 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
154 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
155 * tree-ssa-operands.h: Remove prototype.
156 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
157 * gimple.c (virtual_operand_p): Relocate from gimple.c.
158 * gimple.h: Add prototype.
159 * gimple-ssa.h: Include tree-ssa-operands.h.
160 * tree-dump.c: Add tree-cfg.h to include list.
161 * tree-ssa-alias.c: Add ipa-reference.h to include list.
162 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
163 * config/i386/i386.c: Don't include tree-flow.h.
164 * config/rs6000/rs6000.c: Likewise.
165
166 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
167
168 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
169
170 2013-10-18 Richard Biener <rguenther@suse.de>
171
172 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
173 or TYPE_UNSIGNED of integral types.
174 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
175 NULL_TREE for zero-precision integral types.
176
177 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
178
179 * config/aarch64/arm_neon.h
180 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
181
182 2013-10-17 Sriraman Tallam <tmsriram@google.com>
183
184 * opth-gen.awk: Define target_flags_explicit.
185
186 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
187
188 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
189 fields to the reg_addr array that describes the valid addressing
190 mode for any register, general purpose registers, floating point
191 registers, and Altivec registers.
192 (FIRST_RELOAD_REG_CLASS): Likewise.
193 (LAST_RELOAD_REG_CLASS): Likewise.
194 (struct reload_reg_map_type): Likewise.
195 (reload_reg_map_type): Likewise.
196 (RELOAD_REG_VALID): Likewise.
197 (RELOAD_REG_MULTIPLE): Likewise.
198 (RELOAD_REG_INDEXED): Likewise.
199 (RELOAD_REG_OFFSET): Likewise.
200 (RELOAD_REG_PRE_INCDEC): Likewise.
201 (RELOAD_REG_PRE_MODIFY): Likewise.
202 (reg_addr): Likewise.
203 (mode_supports_pre_incdec_p): New helper functions to say whether
204 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
205 (mode_supports_pre_modify_p): Likewise.
206 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
207 print the valid address mode bits for each mode.
208 (rs6000_debug_print_mode): Likewise.
209 (rs6000_debug_reg_global): Likewise.
210 (rs6000_setup_reg_addr_masks): New function to set up the address
211 mask bits for each type.
212 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
213 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
214 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
215 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
216 PRE_MODIFY are supported.
217 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
218 registers, instead of {src,dest}_av_p.
219 (rs6000_print_options_internal): Tweak the debug output slightly.
220
221 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
222
223 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
224 isa attribute.
225
226 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
227
228 * tree-flow.h (struct omp_region): Move to omp-low.c.
229 Remove omp_ prototypes and variables.
230 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
231 (copy_var_decl): Relocate prototype from tree-flow.h.
232 * gimple.c (copy_var_decl): Relocate from omp-low.c.
233 * tree.h: Move prototype to omp-low.h.
234 * omp-low.h: New File. Relocate prototypes here.
235 * omp-low.c (struct omp_region): Make local here.
236 (root_omp_region): Make static.
237 (copy_var_decl) Move to gimple.c.
238 (new_omp_region): Make static.
239 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
240 * tree-cfg.c: Include omp-low.h.
241 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
242 * gimplify.c: Include omp-low.h.
243 * tree-parloops.c: Likewise.
244
245 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
246
247 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
248 operands are in memory, prefer to force non-matched operand 1 to
249 the register.
250
251 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
252
253 PR target/58673
254 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
255 restrict TImode addresses to single indirect registers if both
256 -mquad-memory and -mvsx-timode are used.
257 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
258 we should emit load/store quad. Remove using %y for quad memory
259 addresses.
260
261 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
262 constraints to allow load/store quad on machines where TImode is
263 not allowed in VSX registers. Use 'n' instead of 'F' constraint
264 for TImode to load integer constants.
265
266 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
267
268 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
269
270 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
271
272 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
273 handling of STACK_REG.
274
275 2013-10-17 Richard Biener <rguenther@suse.de>
276
277 PR tree-optimization/58143
278 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
279 New function.
280 (rewrite_to_defined_overflow): Likewise.
281 (move_computations_dom_walker::before_dom): Rewrite stmts
282 with undefined signed overflow that are not always executed
283 into unsigned arithmetic.
284
285 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
286
287 PR target/57756
288 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
289 explicit isa flag to be an options variable, instead of using
290 global_options_set. Remove define from rs6000.h.
291 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
292
293 * config/rs6000/rs6000.c (rs6000_option_override_internal):
294 Initialize rs6000_isa_flags_explicit.
295 (rs6000_function_specific_save): Add gcc_options* parameter, so
296 that the powerpc builds after the 2013-10-15 changes.
297 (rs6000_function_specific_restore): Likewise.
298
299 2013-10-16 DJ Delorie <dj@redhat.com>
300
301 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
302 op is a REG before checking REGNO.
303 (rl78_alloc_physical_registers): Verify pattern is a SET before
304 checking SET_SRC.
305
306 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
307
308 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
309 endianness.
310 (vec_unpacks_lo_v4sf): Likewise.
311 (vec_unpacks_float_hi_v4si): Likewise.
312 (vec_unpacks_float_lo_v4si): Likewise.
313 (vec_unpacku_float_hi_v4si): Likewise.
314 (vec_unpacku_float_lo_v4si): Likewise.
315
316 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
317
318 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
319 (vsx_concat_v2sf): Likewise.
320
321 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
322
323 * config/aarch64/aarch64.md
324 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
325
326 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
327
328 PR tree-optimization/58697
329 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
330 estimated_loop_iterations_int.
331 (max_stmt_executions_int): Call get_max_loop_iterations_int.
332 (get_max_loop_iterations_int): New. HWINT version of
333 get_max_loop_iterations.
334 * cfgloop.h: Add prototypes.
335 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
336 * loop-unroll.c (decide_peel_once_rolling): Call
337 get_estimated_loop_iterations_int.
338 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
339 * tree-ssa-loop-niter.h: Tweak prototypes.
340
341 2013-10-16 David Malcolm <dmalcolm@redhat.com>
342
343 * gengtype-parse.c (struct_field_seq): Ignore access-control
344 keywords ("public:" etc).
345
346 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
347
348 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
349 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
350
351 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
352
353 * config/arm/arm.opt (mlra): New option.
354 * config/arm/arm.c (arm_lra_p): New function.
355 (TARGET_LRA_P): Define.
356
357 2013-10-16 Paulo Matos <pmatos@broadcom.com>
358
359 * tree-core.h (tree_code_name): Remove.
360 * tree.h (get_tree_code_name): New prototype.
361 * tree.c (tree_code_name): Make static.
362 (get_tree_code_name): New function.
363 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
364 tree_class_check_failed, tree_range_check_failed,
365 tree_not_class_check_failed, omp_clause_check_failed,
366 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
367 wrapper get_tree_code_name instead of calling tree_code_name directly.
368 * tree-vrp.c (dump_asserts_for): Likewise.
369 * tree-dump.c (dequeue_and_dump): Likewise.
370 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
371 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
372 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
373 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
374 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
375 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
376 dump_gimple_omp_for): Likewise.
377 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
378 * tree-ssa-pre.c (print_pre_expr): Likewise.
379 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
380 * print-tree.c (print_node_brief, print_node): Likewise.
381 * gimple.c (gimple_check_failed): Likewise.
382 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
383 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
384 * config/mep/mep.c (mep_validate_vliw): Likewise.
385 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
386 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
387
388 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
389
390 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
391 for AMD bdver3.
392
393 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
394
395 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
396 (MULTILIB_MATCHES): Add multilib for -march=v8.
397
398 2013-10-15 Sriraman Tallam <tmsriram@google.com>
399
400 PR target/57756
401 * optc-save-gen.awk: Add extra parameter to the save and restore
402 target calls.
403 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
404 * tree.c (build_optimization_node): New parameter. Add extra parameter
405 to call to cl_optimization_save.
406 (build_target_option_node): New parameter. Add extra parameter
407 to call to cl_target_option_save.
408 * tree.h (build_optimization_node): New parameter.
409 (build_target_option_node): New parameter.
410 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
411 build_optimization_node and build_target_option_node.
412 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
413 (handle_pragma_push_options): Ditto.
414 * toplev.c (process_options): Ditto.
415 * opts.c (init_options_struct): Check for opts_set non-null.
416 * target.def (target_option.save): New parameter.
417 (target_option.restore): New parameter.
418 * tm.texi: Generate.
419 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
420 (ix86_pragma_target_parse): Ditto.
421 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
422 parameters.
423 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
424 to build_optimization_node and build_target_option_node.
425 (rs6000_valid_attribute_p): Ditto.
426 (rs6000_pragma_target_parse): Ditto.
427 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
428 data.
429 * config/i386/i386.h:
430 TARGET_64BIT_P: New Macro
431 TARGET_MMX_P: New Macro.
432 TARGET_3DNOW_P: New Macro.
433 TARGET_3DNOW_A_P: New Macro.
434 TARGET_SSE_P: New Macro.
435 TARGET_SSE2_P: New Macro.
436 TARGET_SSE3_P: New Macro.
437 TARGET_SSSE3_P: New Macro.
438 TARGET_SSE4_1_P: New Macro.
439 TARGET_SSE4_2_P: New Macro.
440 TARGET_AVX_P: New Macro.
441 TARGET_AVX2_P: New Macro.
442 TARGET_AVX512F_P: New Macro.
443 TARGET_AVX512PF_P: New Macro.
444 TARGET_AVX512ER_P: New Macro.
445 TARGET_AVX512CD_P: New Macro.
446 TARGET_FMA_P: New Macro.
447 TARGET_SSE4A_P: New Macro.
448 TARGET_FMA4_P: New Macro.
449 TARGET_XOP_P: New Macro.
450 TARGET_LWP_P: New Macro.
451 TARGET_ABM_P: New Macro.
452 TARGET_BMI_P: New Macro.
453 TARGET_BMI2_P: New Macro.
454 TARGET_LZCNT_P: New Macro.
455 TARGET_TBM_P: New Macro.
456 TARGET_POPCNT_P: New Macro.
457 TARGET_SAHF_P: New Macro.
458 TARGET_MOVBE_P: New Macro.
459 TARGET_CRC32_P: New Macro.
460 TARGET_AES_P: New Macro.
461 TARGET_PCLMUL_P: New Macro.
462 TARGET_CMPXCHG16B_P: New Macro.
463 TARGET_FSGSBASE_P: New Macro.
464 TARGET_RDRND_P: New Macro.
465 TARGET_F16C_P: New Macro.
466 TARGET_RTM_P: New Macro.
467 TARGET_HLE_P: New Macro.
468 TARGET_RDSEED_P: New Macro.
469 TARGET_PRFCHW_P: New Macro.
470 TARGET_ADX_P: New Macro.
471 TARGET_FXSR_P: New Macro.
472 TARGET_XSAVE_P: New Macro.
473 TARGET_XSAVEOPT_P: New Macro.
474 TARGET_LP64_P: New Macro.
475 TARGET_X32_P: New Macro.
476 TARGET_FPMATH_DEFAULT_P: New Macro.
477 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
478 * config/i386/i386.c (ix86_option_override_internal): New parameters.
479 opts and opts_set.
480 Change ix86_tune_string to access opts->x_ix86_tune_string.
481 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
482 Change ix86_arch_string to access opts->x_ix86_arch_string.
483 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
484 Change ix86_pmode to access opts->x_ix86_pmode.
485 Change ix86_abi to access opts->x_ix86_abi.
486 Change ix86_cmodel to access opts->x_ix86_cmodel.
487 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
488 Change ix86_isa_flags_explicit to access
489 opts->x_ix86_isa_flags_explicit.
490 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
491 Change ix86_regparm to access opts->x_ix86_regparm.
492 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
493 Change ix86_preferred_stack_boundary_arg to access
494 opts->x_ix86_preferred_stack_boundary_arg.
495 Change ix86_force_align_arg_pointer to access
496 opts->x_ix86_force_align_arg_pointer.
497 Change ix86_incoming_stack_boundar_arg to access
498 opts->x_ix86_incoming_stack_boundar_arg.
499 Change ix86_fpmath to access opts->x_ix86_fpmath.
500 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
501 Change ix86_recip_name to access opts->x_ix86_recip_name.
502 Change ix86_stack_protector_guard to access
503 opts->x_ix86_stack_protector_guard.
504 Change ix86_tune_memcpy_strategy to access
505 opts->x_ix86_tune_memcpy_strategy.
506 Change ix86_tune_memset_strategy to access
507 opts->x_ix86_tune_memset_strategy.
508 Change global_options to access opts.
509 Change global_options_set to access opts_set.
510 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
511 Change TARGET_MMX to TARGET_MMX_P (opts->...).
512 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
513 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
514 Change TARGET_SSE to TARGET_SSE_P (opts->...).
515 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
516 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
517 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
518 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
519 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
520 Change TARGET_AVX to TARGET_AVX_P (opts->...).
521 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
522 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
523 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
524 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
525 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
526 Change TARGET_FMA to TARGET_FMA_P (opts->...).
527 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
528 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
529 Change TARGET_XOP to TARGET_XOP_P (opts->...).
530 Change TARGET_LWP to TARGET_LWP_P (opts->...).
531 Change TARGET_ABM to TARGET_ABM_P (opts->...).
532 Change TARGET_BMI to TARGET_BMI_P (opts->...).
533 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
534 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
535 Change TARGET_TBM to TARGET_TBM_P (opts->...).
536 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
537 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
538 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
539 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
540 Change TARGET_AES to TARGET_AES_P (opts->...).
541 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
542 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
543 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
544 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
545 Change TARGET_F16C to TARGET_F16C_P (opts->...).
546 Change TARGET_RTM to TARGET_RTM_P (opts->...).
547 Change TARGET_HLE to TARGET_HLE_P (opts->...).
548 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
549 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
550 Change TARGET_ADX to TARGET_ADX_P (opts->...).
551 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
552 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
553 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
554 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
555 Change TARGET_X32 to TARGET_X32_P (opts->...).
556 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
557 Change TARGET_FLOAT_RETURNS_IN_80387 to
558 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
559 (ix86_function_specific_save): New parameter. Use opts-> fields
560 to replace global fields.
561 (ix86_function_specific_restore): Ditto.
562 (ix86_valid_target_attribute_inner_p): New parameters.
563 Fix recursive call.
564 Fix call to ix86_handle_option and set_option.
565 (ix86_valid_target_attribute_tree): New parameters.
566 Change global_options to access opts.
567 Change global_options_set to access opts_set.
568 Fix call to ix86_valid_target_attribute_inner_p.
569 Change ix86_tune_string to access opts->x_ix86_tune_string.
570 Change ix86_arch_string to access opts->x_ix86_arch_string.
571 Change ix86_fpmath to access opts->x_ix86_fpmath
572 Fix call to ix86_option_override_internal.
573 Fix call to ix86_add_new_builtins.
574 Fix calls to build_optimization_node and build_target_option_node.
575 (ix86_valid_target_attribute_p): Remove access to global_options.
576 Use new gcc_options structure func_options.
577 Fix call to ix86_valid_target_attribute_tree.
578 Fix call to build_optimization_node.
579 (get_builtin_code_for_version): Fix call to
580 ix86_valid_target_attribute_tree.
581
582 2013-10-15 David Malcolm <dmalcolm@redhat.com>
583
584 * Makefile.in (PICFLAG): New.
585 (enable_host_shared): New.
586 (INTERNAL_CFLAGS): Use PICFLAG.
587 (LIBIBERTY): Use pic build of libiberty.a if configured with
588 --enable-host-shared.
589 * configure.ac: Add --enable-host-shared, setting up new
590 PICFLAG variable.
591 * configure: Regenerate.
592 * doc/install.texi (--enable-shared): Add note contrasting it with ...
593 (--enable-host-shared): New option.
594
595 2013-10-15 Richard Biener <rguenther@suse.de>
596
597 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
598 for built-in functions.
599
600 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
601
602 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
603 (optimize_range_tests_1): New function,
604 extracted from optimize_range_tests.
605 (optimize_range_tests_xor): Similarly.
606 (optimize_range_tests_diff): New function.
607 (optimize_range_tests): Use optimize_range_tests_1.
608
609 2013-10-15 Cong Hou <congh@google.com>
610
611 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
612 requirement of the reduction pattern so that one operand of the
613 reduction operation can come from outside of the loop.
614
615 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
616
617 * config/arm/neon-schedgen.ml: Remove.
618 * config/arm/cortex-a9-neon.md: Remove comment regarding
619 neon-schedgen.ml.
620
621 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
622
623 * config/arm/types: Remove old neon types.
624
625 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
626
627 * config/arm/cortex-a7.md
628 (cortex_a7_neon_type): New.
629 (cortex_a7_neon_mul): Update for new types.
630 (cortex_a7_neon_mla): Likewise.
631 (cortex_a7_neon): Likewise.
632
633 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
634
635 * config/arm/cortex-a15-neon.md
636 (cortex_a15_neon_type): New,
637
638 (cortex_a15_neon_int_1): Remove.
639 (cortex_a15_neon_int_2): Likewise.
640 (cortex_a15_neon_int_3): Likewise.
641 (cortex_a15_neon_int_4): Likewise.
642 (cortex_a15_neon_int_5): Likewise.
643 (cortex_a15_neon_vqneg_vqabs): Likewise.
644 (cortex_a15_neon_vmov): Likewise.
645 (cortex_a15_neon_vaba): Likewise.
646 (cortex_a15_neon_vaba_qqq): Likewise.
647 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
648 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
649 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
650 Likewise.
651 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
652 (cortex_a15_neon_mla_qqq_8_16): Likewise.
653 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
654 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
655 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
656 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
657 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
658 (cortex_a15_neon_shift_1): Likewise.
659 (cortex_a15_neon_shift_2): Likewise.
660 (cortex_a15_neon_shift_3): Likewise.
661 (cortex_a15_neon_vshl_ddd): Likewise.
662 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
663 (cortex_a15_neon_vsra_vrsra): Likewise.
664 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
665 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
666 (cortex_a15_neon_bp_3cycle): Likewise.
667 (cortex_a15_neon_ldm_2): Likewise.
668 (cortex_a15_neon_stm_2): Likewise.
669 (cortex_a15_neon_mcr): Likewise.
670 (cortex_a15_neon_mrc): Likewise.
671 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
672 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
673 (cortex_a15_neon_fp_vmul_ddd): Likewise.
674 (cortex_a15_neon_fp_vmul_qqd): Likewise.
675 (cortex_a15_neon_fp_vmla_ddd): Likewise.
676 (cortex_a15_neon_fp_vmla_qqq): Likewise.
677 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
678 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
679 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
680 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
681 (cortex_a15_neon_bp_simple): Likewise.
682 (cortex_a15_neon_bp_2cycle): Likewise.
683 (cortex_a15_neon_bp_3cycle): Likewise.
684 (cortex_a15_neon_vld1_1_2_regs): Likewise.
685 (cortex_a15_neon_vld1_3_4_regs): Likewise.
686 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
687 (cortex_a15_neon_vld2_4_regs): Likewise.
688 (cortex_a15_neon_vld3_vld4): Likewise.
689 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
690 (cortex_a15_neon_vst1_3_4_regs): Likewise.
691 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
692 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
693 (cortex_a15_neon_vst3_vst4): Rename to...
694 (cortex_a15_neon_vst4): This, update for new attributes.
695 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
696 (cortex_a15_neon_vld3_vld4_lane): Likewise.
697 (cortex_a15_neon_vst1_vst2_lane): Likewise.
698 (cortex_a15_neon_vst3_vst4_lane): Likewise.
699 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
700 (cortex_a15_neon_ldm_2): Likewise.
701 (cortex_a15_neon_stm_2): Likewise.
702 (cortex_a15_neon_mcr): Likewise.
703 (cortex_a15_neon_mcr_2_mcrr): Likewise.
704 (cortex_a15_neon_mrc): Likewise.
705 (cortex_a15_neon_mrrc): Likewise.
706
707 (cortex_a15_neon_abd): New.
708 (cortex_a15_neon_abd_q): Likewise.
709 (cortex_a15_neon_aba): Likewise.
710 (cortex_a15_neon_aba_q): Likewise.
711 (cortex_a15_neon_acc): Likewise.
712 (cortex_a15_neon_acc_q): Likewise.
713 (cortex_a15_neon_arith_basic): Likewise.
714 (cortex_a15_neon_arith_complex): Likewise.
715 (cortex_a15_neon_multiply): Likewise.
716 (cortex_a15_neon_multiply_q): Likewise.
717 (cortex_a15_neon_mla): Likewise.
718 (cortex_a15_neon_mla_q): Likewise.
719 (cortex_a15_neon_sat_mla_long): Likewise.
720 (cortex_a15_neon_shift_acc): Likewise.
721 (cortex_a15_neon_shift_imm_basic): Likewise.
722 (cortex_a15_neon_shift_imm_complex): Likewise.
723 (cortex_a15_neon_shift_reg_basic): Likewise.
724 (cortex_a15_neon_shift_reg_basic_q): Likewise.
725 (cortex_a15_neon_shift_reg_complex): Likewise.
726 (cortex_a15_neon_shift_reg_complex_q): Likewise.
727 (cortex_a15_neon_fp_negabs): Likewise
728 (cortex_a15_neon_fp_arith): Likewise
729 (cortex_a15_neon_fp_arith_q): Likewise
730 (cortex_a15_neon_fp_cvt_int): Likewise
731 (cortex_a15_neon_fp_cvt_int_q): Likewise
732 (cortex_a15_neon_fp_cvt_16): Likewise
733 (cortex_a15_neon_fp_mul): Likewise
734 (cortex_a15_neon_fp_mul_q): Likewise
735 (cortex_a15_neon_fp_mla): Likewise
736 (cortex_a15_neon_fp_mla_q): Likewise
737 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
738 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
739 (cortex_a15_neon_bitops): Likewise.
740 (cortex_a15_neon_bitops_q): Likewise.
741 (cortex_a15_neon_from_gp): Likewise.
742 (cortex_a15_neon_from_gp_q): Likewise.
743 (cortex_a15_neon_tbl3_tbl4): Likewise.
744 (cortex_a15_neon_zip_q): Likewise.
745 (cortex_a15_neon_to_gp): Likewise.
746 (cortex_a15_neon_load_a): Likewise.
747 (cortex_a15_neon_load_b): Likewise.
748 (cortex_a15_neon_load_c): Likewise.
749 (cortex_a15_neon_load_d): Likewise.
750 (cortex_a15_neon_load_e): Likewise.
751 (cortex_a15_neon_load_f): Likewise.
752 (cortex_a15_neon_store_a): Likewise.
753 (cortex_a15_neon_store_b): Likewise.
754 (cortex_a15_neon_store_c): Likewise.
755 (cortex_a15_neon_store_d): Likewise.
756 (cortex_a15_neon_store_e): Likewise.
757 (cortex_a15_neon_store_f): Likewise.
758 (cortex_a15_neon_store_g): Likewise.
759 (cortex_a15_neon_store_h): Likewise.
760 (cortex_a15_vfp_to_from_gp): Likewise.
761
762 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
763
764 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
765
766 (cortex_a9_neon_vshl_ddd): Remove.
767 (cortex_a9_neon_vst3_vst4): Likewise.
768 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
769
770 (cortex_a9_neon_bit_ops_q): New.
771
772 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
773 (cortex_a9_neon_int_2): Likewise.
774 (cortex_a9_neon_int_3): Likewise.
775 (cortex_a9_neon_int_4): Likewise.
776 (cortex_a9_neon_int_5): Likewise.
777 (cortex_a9_neon_vqneg_vqabs): Likewise.
778 (cortex_a9_neon_vmov): Likewise.
779 (cortex_a9_neon_vaba): Likewise.
780 (cortex_a9_neon_vaba_qqq): Likewise.
781 (cortex_a9_neon_shift_1): Likewise.
782 (cortex_a9_neon_shift_2): Likewise.
783 (cortex_a9_neon_shift_3): Likewise.
784 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
785 (cortex_a9_neon_vsra_vrsra): Likewise.
786 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
787 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
788 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
789 Likewise.
790 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
791 (cortex_a9_neon_mla_qqq_8_16): Likewise.
792 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
793 Likewise.
794 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
795 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
796 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
797 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
798 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
799 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
800 (cortex_a9_neon_fp_vsum): Likewise.
801 (cortex_a9_neon_fp_vmul_ddd): Likewise.
802 (cortex_a9_neon_fp_vmul_qqd): Likewise.
803 (cortex_a9_neon_fp_vmla_ddd): Likewise.
804 (cortex_a9_neon_fp_vmla_qqq): Likewise.
805 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
806 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
807 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
808 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
809 (cortex_a9_neon_bp_simple): Likewise.
810 (cortex_a9_neon_bp_2cycle): Likewise.
811 (cortex_a9_neon_bp_3cycle): Likewise.
812 (cortex_a9_neon_ldr): Likewise.
813 (cortex_a9_neon_str): Likewise.
814 (cortex_a9_neon_vld1_1_2_regs): Likewise.
815 (cortex_a9_neon_vld1_3_4_regs): Likewise.
816 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
817 (cortex_a9_neon_vld2_4_regs): Likewise.
818 (cortex_a9_neon_vld3_vld4): Likewise.
819 (cortex_a9_neon_vld1_vld2_lane): Likewise.
820 (cortex_a9_neon_vld3_vld4_lane): Likewise.
821 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
822 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
823 (cortex_a9_neon_vst1_3_4_regs): Likewise.
824 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
825 (cortex_a9_neon_vst1_vst2_lane): Likewise.
826 (cortex_a9_neon_vst3_vst4_lane): Likewise.
827 (cortex_a9_neon_mcr): Likewise.
828 (cortex_a9_neon_mcr_2_mcrr): Likewise.
829 (cortex_a9_neon_mrc): Likewise.
830 (cortex_a9_neon_mrrc): Likewise.
831
832 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
833
834 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
835
836 (cortex_a8_neon_vshl_ddd): Remove.
837 (cortex_a8_neon_vst3_vst4): Likewise.
838 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
839
840 (cortex_a8_neon_bit_ops_q): New.
841
842 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
843 (cortex_a8_neon_int_2): Likewise..
844 (cortex_a8_neon_int_3): Likewise.
845 (cortex_a8_neon_int_5): Likewise.
846 (cortex_a8_neon_vqneg_vqabs): Likewise.
847 (cortex_a8_neon_int_4): Likewise.
848 (cortex_a8_neon_vaba): Likewise.
849 (cortex_a8_neon_vaba_qqq): Likewise.
850 (cortex_a8_neon_shift_1): Likewise.
851 (cortex_a8_neon_shift_2): Likewise.
852 (cortex_a8_neon_shift_3): Likewise.
853 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
854 (cortex_a8_neon_vsra_vrsra): Likewise.
855 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
856 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
857 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
858 Likewise.
859 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
860 (cortex_a8_neon_mla_qqq_8_16): Likewise.
861 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
862 Likewise.
863 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
864 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
865 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
866 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
867 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
868 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
869 (cortex_a8_neon_fp_vsum): Likewise.
870 (cortex_a8_neon_fp_vmul_ddd): Likewise.
871 (cortex_a8_neon_fp_vmul_qqd): Likewise.
872 (cortex_a8_neon_fp_vmla_ddd): Likewise.
873 (cortex_a8_neon_fp_vmla_qqq): Likewise.
874 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
875 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
876 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
877 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
878 (cortex_a8_neon_bp_simple): Likewise.
879 (cortex_a8_neon_bp_2cycle): Likewise.
880 (cortex_a8_neon_bp_3cycle): Likewise.
881 (cortex_a8_neon_ldr): Likewise.
882 (cortex_a8_neon_str): Likewise.
883 (cortex_a8_neon_vld1_1_2_regs): Likewise.
884 (cortex_a8_neon_vld1_3_4_regs): Likewise.
885 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
886 (cortex_a8_neon_vld2_4_regs): Likewise.
887 (cortex_a8_neon_vld3_vld4): Likewise.
888 (cortex_a8_neon_vld1_vld2_lane): Likewise.
889 (cortex_a8_neon_vld3_vld4_lane): Likewise.
890 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
891 (cortex_a8_neon_vst1_3_4_regs): Likewise.
892 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
893 (cortex_a8_neon_vst1_vst2_lane): Likewise.
894 (cortex_a8_neon_vst3_vst4_lane): Likewise.
895 (cortex_a8_neon_mcr): Likewise.
896 (cortex_a8_neon_mcr_2_mcrr): Likewise.
897 (cortex_a8_neon_mrc): Likewise.
898 (cortex_a8_neon_mrrc): Likewise.
899
900 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
901
902 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
903 (fp): New.
904 * config/aarch64/aarch64-simd.md (neon_type): Remove.
905 (aarch64_simd_dup<mode>): Add "type" attribute.
906 (aarch64_dup_lane<mode>): Likewise.
907 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
908 (*aarch64_simd_mov<mode>): Likewise.
909 (aarch64_simd_mov_from_<mode>low): Likewise.
910 (aarch64_simd_mov_from_<mode>high): Likewise.
911 (orn<mode>3): Likewise.
912 (bic<mode>3): Likewise.
913 (add<mode>3): Likewise.
914 (sub<mode>3): Likewise.
915 (mul<mode>3): Likewise.
916 (*aarch64_mul3_elt<mode>): Likewise.
917 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
918 (*aarch64_mul3_elt_to_128df): Likewise.
919 (*aarch64_mul3_elt_to_64v2df): Likewise.
920 (neg<mode>2): Likewise.
921 (abs<mode>2): Likewise.
922 (abd<mode>_3): Likewise.
923 (aba<mode>_3): Likewise.
924 (fabd<mode>_3): Likewise.
925 (*fabd_scalar<mode>3): Likewise.
926 (and<mode>3): Likewise.
927 (ior<mode>3): Likewise.
928 (xor<mode>3): Likewise.
929 (one_cmpl<mode>2): Likewise.
930 (aarch64_simd_vec_set<mode>): Likewise.
931 (aarch64_simd_lshr<mode>): Likewise.
932 (aarch64_simd_ashr<mode>): Likewise.
933 (aarch64_simd_imm_shl<mode>): Likewise.
934 (aarch64_simd_reg_sshl<mode): Likewise.
935 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
936 (aarch64_simd_reg_shl<mode>_signed): Likewise.
937 (aarch64_simd_vec_setv2di): Likewise.
938 (aarch64_simd_vec_set<mode>): Likewise.
939 (aarch64_mla<mode>): Likewise.
940 (*aarch64_mla_elt<mode>): Likewise.
941 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
942 (aarch64_mls<mode>): Likewise.
943 (*aarch64_mls_elt<mode>): Likewise.
944 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
945 (<su><maxmin><mode>3): Likewise.
946 (move_lo_quad_<mode>): Likewise.
947 (aarch64_simd_move_hi_quad_<mode>): Likewise.
948 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
949 (vec_pack_trunc_<mode>): Likewise.
950 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
951 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
952 (*aarch64_<su>mlal_lo<mode>): Likewise.
953 (*aarch64_<su>mlal_hi<mode>): Likewise.
954 (*aarch64_<su>mlsl_lo<mode>): Likewise.
955 (*aarch64_<su>mlsl_hi<mode>): Likewise.
956 (*aarch64_<su>mlal<mode>): Likewise.
957 (*aarch64_<su>mlsl<mode>): Likewise.
958 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
959 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
960 (add<mode>3): Likewise.
961 (sub<mode>3): Likewise.
962 (mul<mode>3): Likewise.
963 (div<mode>3): Likewise.
964 (neg<mode>2): Likewise.
965 (abs<mode>2): Likewise.
966 (fma<mode>4): Likewise.
967 (*aarch64_fma4_elt<mode>): Likewise.
968 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
969 (*aarch64_fma4_elt_to_128df): Likewise.
970 (*aarch64_fma4_elt_to_64v2df): Likewise.
971 (fnma<mode>4): Likewise.
972 (*aarch64_fnma4_elt<mode>): Likewise.
973 (*aarch64_fnma4_elt_<vswap_width_name><mode>
974 (*aarch64_fnma4_elt_to_128df): Likewise.
975 (*aarch64_fnma4_elt_to_64v2df): Likewise.
976 (<frint_pattern><mode>2): Likewise.
977 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
978 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
979 (vec_unpacks_lo_v4sf): Likewise.
980 (aarch64_float_extend_lo_v2df): Likewise.
981 (vec_unpacks_hi_v4sf): Likewise.
982 (aarch64_float_truncate_lo_v2sf): Likewise.
983 (aarch64_float_truncate_hi_v4sf): Likewise.
984 (aarch64_vmls<mode>): Likewise.
985 (<su><maxmin><mode>3): Likewise.
986 (<maxmin_uns><mode>3): Likewise.
987 (reduc_<sur>plus_<mode>): Likewise.
988 (reduc_<sur>plus_v2di): Likewise.
989 (reduc_<sur>plus_v2si): Likewise.
990 (reduc_<sur>plus_<mode>): Likewise.
991 (aarch64_addpv4sf): Likewise.
992 (clz<mode>2): Likewise.
993 (reduc_<maxmin_uns>_<mode>): Likewise.
994 (reduc_<maxmin_uns>_v2di): Likewise.
995 (reduc_<maxmin_uns>_v2si): Likewise.
996 (reduc_<maxmin_uns>_<mode>): Likewise.
997 (reduc_<maxmin_uns>_v4sf): Likewise.
998 (aarch64_simd_bsl<mode>_internal): Likewise.
999 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1000 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1001 (aarch64_get_lane<mode>): Likewise.
1002 (*aarch64_combinez<mode>): Likewise.
1003 (aarch64_combine<mode>): Likewise.
1004 (aarch64_simd_combine<mode>): Likewise.
1005 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1006 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1007 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1008 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1009 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1010 (aarch64_<sur>h<addsub><mode>): Likewise.
1011 (aarch64_<sur><addsub>hn<mode>): Likewise.
1012 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1013 (aarch64_pmul<mode>): Likewise.
1014 (aarch64_<su_optab><optab><mode>): Likewise.
1015 (aarch64_<sur>qadd<mode>): Likewise.
1016 (aarch64_sqmovun<mode>): Likewise.
1017 (aarch64_<sur>qmovn<mode>): Likewise.
1018 (aarch64_s<optab><mode>): Likewise.
1019 (aarch64_sq<r>dmulh<mode>): Likewise.
1020 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1021 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1022 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1023 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1024 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1025 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1026 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1027 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1028 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1029 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1030 (aarch64_sqdmull<mode>): Likewise.
1031 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1032 (aarch64_sqdmull_n<mode>): Likewise.
1033 (aarch64_sqdmull2<mode>_internal): Likewise.
1034 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1035 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1036 (aarch64_<sur>shl<mode>): Likewise.
1037 (aarch64_<sur>q<r>shl<mode>
1038 (aarch64_<sur>shll_n<mode>): Likewise.
1039 (aarch64_<sur>shll2_n<mode>): Likewise.
1040 (aarch64_<sur>shr_n<mode>): Likewise.
1041 (aarch64_<sur>sra_n<mode>): Likewise.
1042 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1043 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1044 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1045 (aarch64_cm<optab><mode>): Likewise.
1046 (aarch64_cm<optab>di): Likewise.
1047 (aarch64_cm<optab><mode>): Likewise.
1048 (aarch64_cm<optab>di): Likewise.
1049 (aarch64_cmtst<mode>): Likewise.
1050 (aarch64_cmtstdi): Likewise.
1051 (aarch64_cm<optab><mode>): Likewise.
1052 (*aarch64_fac<optab><mode>): Likewise.
1053 (aarch64_addp<mode>): Likewise.
1054 (aarch64_addpdi): Likewise.
1055 (sqrt<mode>2): Likewise.
1056 (vec_load_lanesoi<mode>): Likewise.
1057 (vec_store_lanesoi<mode>): Likewise.
1058 (vec_load_lanesci<mode>): Likewise.
1059 (vec_store_lanesci<mode>): Likewise.
1060 (vec_load_lanesxi<mode>): Likewise.
1061 (vec_store_lanesxi<mode>): Likewise.
1062 (*aarch64_mov<mode>): Likewise.
1063 (aarch64_ld2<mode>_dreg): Likewise.
1064 (aarch64_ld2<mode>_dreg): Likewise.
1065 (aarch64_ld3<mode>_dreg): Likewise.
1066 (aarch64_ld3<mode>_dreg): Likewise.
1067 (aarch64_ld4<mode>_dreg): Likewise.
1068 (aarch64_ld4<mode>_dreg): Likewise.
1069 (aarch64_tbl1<mode>): Likewise.
1070 (aarch64_tbl2v16qi): Likewise.
1071 (aarch64_combinev16qi): Likewise.
1072 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1073 (aarch64_st2<mode>_dreg): Likewise.
1074 (aarch64_st2<mode>_dreg): Likewise.
1075 (aarch64_st3<mode>_dreg): Likewise.
1076 (aarch64_st3<mode>_dreg): Likewise.
1077 (aarch64_st4<mode>_dreg): Likewise.
1078 (aarch64_st4<mode>_dreg): Likewise.
1079 (*aarch64_simd_ld1r<mode>): Likewise.
1080 (aarch64_frecpe<mode>): Likewise.
1081 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1082 (aarch64_frecps<mode>): Likewise.
1083
1084 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1085
1086 * config/arm/iterators.md (V_elem_ch): New.
1087 (q): Likewise.
1088 (VQH_type): Likewise.
1089 * config/arm/arm.md (is_neon_type): New.
1090 (conds): Use is_neon_type.
1091 (anddi3_insn): Update type attribute.
1092 (xordi3_insn): Likewise.
1093 (one_cmpldi2): Likewise.
1094 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1095 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1096 (*movmisalign<mode>_neon_store): Likewise.
1097 (*movmisalign<mode>_neon_load): Likewise.
1098 (vec_set<mode>_internal): Likewise.
1099 (vec_set<mode>_internal): Likewise.
1100 (vec_setv2di_internal): Likewise.
1101 (vec_extract<mode>): Likewise.
1102 (vec_extract<mode>): Likewise.
1103 (vec_extractv2di): Likewise.
1104 (*add<mode>3_neon): Likewise.
1105 (adddi3_neon): Likewise.
1106 (*sub<mode>3_neon): Likewise.
1107 (subdi3_neon): Likewise.
1108 (fma<VCVTF:mode>4): Likewise.
1109 (fma<VCVTF:mode>4_intrinsic): Likewise.
1110 (*fmsub<VCVTF:mode>4): Likewise.
1111 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1112 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1113 (ior<mode>3): Likewise.
1114 (and<mode>3): Likewise.
1115 (orn<mode>3_neon): Likewise.
1116 (orndi3_neon): Likewise.
1117 (bic<mode>3_neon): Likewise.
1118 (bicdi3_neon): Likewise.
1119 (xor<mode>3): Likewise.
1120 (one_cmpl<mode>2): Likewise.
1121 (abs<mode>2): Likewise.
1122 (neg<mode>2): Likewise.
1123 (negdi2_neon): Likewise.
1124 (*umin<mode>3_neon): Likewise.
1125 (*umax<mode>3_neon): Likewise.
1126 (*smin<mode>3_neon): Likewise.
1127 (*smax<mode>3_neon): Likewise.
1128 (vashl<mode>3): Likewise.
1129 (vashr<mode>3_imm): Likewise.
1130 (vlshr<mode>3_imm): Likewise.
1131 (ashl<mode>3_signed): Likewise.
1132 (ashl<mode>3_unsigned): Likewise.
1133 (neon_load_count): Likewise.
1134 (ashldi3_neon_noclobber): Likewise.
1135 (ashldi3_neon): Likewise.
1136 (signed_shift_di3_neon): Likewise.
1137 (unsigned_shift_di3_neon): Likewise.
1138 (ashrdi3_neon_imm_noclobber): Likewise.
1139 (lshrdi3_neon_imm_noclobber): Likewise.
1140 (<shift>di3_neon): Likewise.
1141 (widen_ssum<mode>3): Likewise.
1142 (widen_usum<mode>3): Likewise.
1143 (quad_halves_<code>v4si): Likewise.
1144 (quad_halves_<code>v4sf): Likewise.
1145 (quad_halves_<code>v8hi): Likewise.
1146 (quad_halves_<code>v16qi): Likewise.
1147 (reduc_splus_v2di): Likewise.
1148 (neon_vpadd_internal<mode>): Likewise.
1149 (neon_vpsmin<mode>): Likewise.
1150 (neon_vpsmax<mode>): Likewise.
1151 (neon_vpumin<mode>): Likewise.
1152 (neon_vpumax<mode>): Likewise.
1153 (*ss_add<mode>_neon): Likewise.
1154 (*us_add<mode>_neon): Likewise.
1155 (*ss_sub<mode>_neon): Likewise.
1156 (*us_sub<mode>_neon): Likewise.
1157 (neon_vadd<mode>_unspec): Likewise.
1158 (neon_vaddl<mode>): Likewise.
1159 (neon_vaddw<mode>): Likewise.
1160 (neon_vhadd<mode>): Likewise.
1161 (neon_vqadd<mode>): Likewise.
1162 (neon_vaddhn<mode>): Likewise.
1163 (neon_vmul<mode>): Likewise.
1164 (neon_vfms<VCVTF:mode>): Likewise.
1165 (neon_vmlal<mode>): Likewise.
1166 (neon_vmls<mode>): Likewise.
1167 (neon_vmlsl<mode>): Likewise.
1168 (neon_vqdmulh<mode>): Likewise.
1169 (neon_vqdmlal<mode>): Likewise.
1170 (neon_vqdmlsl<mode>): Likewise.
1171 (neon_vmull<mode>): Likewise.
1172 (neon_vqdmull<mode>): Likewise.
1173 (neon_vsub<mode>_unspec): Likewise.
1174 (neon_vsubl<mode>): Likewise.
1175 (neon_vsubw<mode>): Likewise.
1176 (neon_vqsub<mode>): Likewise.
1177 (neon_vhsub<mode>): Likewise.
1178 (neon_vsubhn<mode>): Likewise.
1179 (neon_vceq<mode>): Likewise.
1180 (neon_vcge<mode>): Likewise.
1181 (neon_vcgeu<mode>): Likewise.
1182 (neon_vcgt<mode>): Likewise.
1183 (neon_vcgtu<mode>): Likewise.
1184 (neon_vcle<mode>): Likewise.
1185 (neon_vclt<mode>): Likewise.
1186 (neon_vcage<mode>): Likewise.
1187 (neon_vcagt<mode>): Likewise.
1188 (neon_vtst<mode>): Likewise.
1189 (neon_vabd<mode>): Likewise.
1190 (neon_vabdl<mode>): Likewise.
1191 (neon_vaba<mode>): Likewise.
1192 (neon_vabal<mode>): Likewise.
1193 (neon_vmax<mode>): Likewise.
1194 (neon_vmin<mode>): Likewise.
1195 (neon_vpaddl<mode>): Likewise.
1196 (neon_vpadal<mode>): Likewise.
1197 (neon_vpmax<mode>): Likewise.
1198 (neon_vpmin<mode>): Likewise.
1199 (neon_vrecps<mode>): Likewise.
1200 (neon_vrsqrts<mode>): Likewise.
1201 (neon_vqabs<mode>): Likewise.
1202 (neon_vqneg<mode>): Likewise.
1203 (neon_vcls<mode>): Likewise.
1204 (clz<mode>2): Likewise.
1205 (popcount<mode>2): Likewise.
1206 (neon_vrecpe<mode>): Likewise.
1207 (neon_vrsqrte<mode>): Likewise.
1208 (neon_vget_lane<mode>_sext_internal): Likewise.
1209 (neon_vget_lane<mode>_zext_internal): Likewise.
1210 (neon_vdup_n<mode>): Likewise.
1211 (neon_vdup_n<mode>): Likewise.
1212 (neon_vdup_nv2di): Likewise.
1213 (neon_vdup_lane<mode>_interal): Likewise.
1214 (*neon_vswp<mode>): Likewise.
1215 (neon_vcombine<mode>): Likewise.
1216 (float<mode><V_cvtto>2): Likewise.
1217 (floatuns<mode><V_cvtto>2): Likewise.
1218 (fix_trunc<mode><V_cvtto>2): Likewise.
1219 (fixuns_trunc<mode><V_cvtto>2
1220 (neon_vcvt<mode>): Likewise.
1221 (neon_vcvt<mode>): Likewise.
1222 (neon_vcvtv4sfv4hf): Likewise.
1223 (neon_vcvtv4hfv4sf): Likewise.
1224 (neon_vcvt_n<mode>): Likewise.
1225 (neon_vcvt_n<mode>): Likewise.
1226 (neon_vmovn<mode>): Likewise.
1227 (neon_vqmovn<mode>): Likewise.
1228 (neon_vqmovun<mode>): Likewise.
1229 (neon_vmovl<mode>): Likewise.
1230 (neon_vmul_lane<mode>): Likewise.
1231 (neon_vmul_lane<mode>): Likewise.
1232 (neon_vmull_lane<mode>): Likewise.
1233 (neon_vqdmull_lane<mode>): Likewise.
1234 (neon_vqdmulh_lane<mode>): Likewise.
1235 (neon_vqdmulh_lane<mode>): Likewise.
1236 (neon_vmla_lane<mode>): Likewise.
1237 (neon_vmla_lane<mode>): Likewise.
1238 (neon_vmlal_lane<mode>): Likewise.
1239 (neon_vqdmlal_lane<mode>): Likewise.
1240 (neon_vmls_lane<mode>): Likewise.
1241 (neon_vmls_lane<mode>): Likewise.
1242 (neon_vmlsl_lane<mode>): Likewise.
1243 (neon_vqdmlsl_lane<mode>): Likewise.
1244 (neon_vext<mode>): Likewise.
1245 (neon_vrev64<mode>): Likewise.
1246 (neon_vrev32<mode>): Likewise.
1247 (neon_vrev16<mode>): Likewise.
1248 (neon_vbsl<mode>_internal): Likewise.
1249 (neon_vshl<mode>): Likewise.
1250 (neon_vqshl<mode>): Likewise.
1251 (neon_vshr_n<mode>): Likewise.
1252 (neon_vshrn_n<mode>): Likewise.
1253 (neon_vqshrn_n<mode>): Likewise.
1254 (neon_vqshrun_n<mode>): Likewise.
1255 (neon_vshl_n<mode>): Likewise.
1256 (neon_vqshl_n<mode>): Likewise.
1257 (neon_vqshlu_n<mode>): Likewise.
1258 (neon_vshll_n<mode>): Likewise.
1259 (neon_vsra_n<mode>): Likewise.
1260 (neon_vsri_n<mode>): Likewise.
1261 (neon_vsli_n<mode>): Likewise.
1262 (neon_vtbl1v8qi): Likewise.
1263 (neon_vtbl2v8qi): Likewise.
1264 (neon_vtbl3v8qi): Likewise.
1265 (neon_vtbl4v8qi): Likewise.
1266 (neon_vtbl1v16qi): Likewise.
1267 (neon_vtbl2v16qi): Likewise.
1268 (neon_vcombinev16qi): Likewise.
1269 (neon_vtbx1v8qi): Likewise.
1270 (neon_vtbx2v8qi): Likewise.
1271 (neon_vtbx3v8qi): Likewise.
1272 (neon_vtbx4v8qi): Likewise.
1273 (*neon_vtrn<mode>_insn): Likewise.
1274 (*neon_vzip<mode>_insn): Likewise.
1275 (*neon_vuzp<mode>_insn): Likewise.
1276 (neon_vld1<mode>): Likewise.
1277 (neon_vld1_lane<mode>): Likewise.
1278 (neon_vld1_lane<mode>): Likewise.
1279 (neon_vld1_dup<mode>): Likewise.
1280 (neon_vld1_dup<mode>): Likewise.
1281 (neon_vld1_dupv2di): Likewise.
1282 (neon_vst1<mode>): Likewise.
1283 (neon_vst1_lane<mode>): Likewise.
1284 (neon_vst1_lane<mode>): Likewise.
1285 (neon_vld2<mode>): Likewise.
1286 (neon_vld2<mode>): Likewise.
1287 (neon_vld2_lane<mode>): Likewise.
1288 (neon_vld2_lane<mode>): Likewise.
1289 (neon_vld2_dup<mode>): Likewise.
1290 (neon_vst2<mode>): Likewise.
1291 (neon_vst2<mode>): Likewise.
1292 (neon_vst2_lane<mode>): Likewise.
1293 (neon_vst2_lane<mode>): Likewise.
1294 (neon_vld3<mode>): Likewise.
1295 (neon_vld3qa<mode>): Likewise.
1296 (neon_vld3qb<mode>): Likewise.
1297 (neon_vld3_lane<mode>): Likewise.
1298 (neon_vld3_lane<mode>): Likewise.
1299 (neon_vld3_dup<mode>): Likewise.
1300 (neon_vst3<mode>): Likewise.
1301 (neon_vst3qa<mode>): Likewise.
1302 (neon_vst3qb<mode>): Likewise.
1303 (neon_vst3_lane<mode>): Likewise.
1304 (neon_vst3_lane<mode>): Likewise.
1305 (neon_vld4<mode>): Likewise.
1306 (neon_vld4qa<mode>): Likewise.
1307 (neon_vld4qb<mode>): Likewise.
1308 (neon_vld4_lane<mode>): Likewise.
1309 (neon_vld4_lane<mode>): Likewise.
1310 (neon_vld4_dup<mode>): Likewise.
1311 (neon_vst4<mode>): Likewise.
1312 (neon_vst4qa<mode>): Likewise.
1313 (neon_vst4qb<mode>): Likewise.
1314 (neon_vst4_lane<mode>): Likewise.
1315 (neon_vst4_lane<mode>): Likewise.
1316 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1317 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1318 (neon_vec_<US>mult_lo_<mode>): Likewise.
1319 (neon_vec_<US>mult_hi_<mode>): Likewise.
1320 (neon_vec_<US>shiftl_<mode>): Likewise.
1321 (neon_unpack<US>_<mode>): Likewise.
1322 (neon_vec_<US>mult_<mode>): Likewise.
1323 (vec_pack_trunc_<mode>): Likewise.
1324 (neon_vec_pack_trunc_<mode>): Likewise.
1325 (neon_vabd<mode>_2): Likewise.
1326 (neon_vabd<mode>_3): Likewise.
1327
1328 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1329
1330 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1331 (load_pair): Update type attribute.
1332 (store_pair): Update type attribute.
1333 * config/aarch64/iterators.md (q): New.
1334
1335 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1336
1337 * config/arm/types.md: Add new types for Neon insns.
1338
1339 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1340 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1341 Sergey Lega <sergey.s.lega@intel.com>
1342 Anna Tikhonova <anna.tikhonova@intel.com>
1343 Ilya Tocar <ilya.tocar@intel.com>
1344 Andrey Turetskiy <andrey.turetskiy@intel.com>
1345 Ilya Verbin <ilya.verbin@intel.com>
1346 Kirill Yukhin <kirill.yukhin@intel.com>
1347 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1348
1349 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1350 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1351 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1352 (rcp14<mode>): New.
1353 (srcp14<mode>): Ditto.
1354 (rsqrt14<mode>): Ditto.
1355 (rsqrt14<mode>): Ditto.
1356 (avx512f_vmscalef<mode>): Ditto.
1357 (avx512f_scalef<mode>): Ditto.
1358 (avx512f_getexp<mode>): Ditto.
1359 (avx512f_sgetexp<mode>): Ditto.
1360 (avx512f_fixupimm<mode>): Ditto.
1361 (avx512f_sfixupimm<mode>): Ditto.
1362 (avx512f_rndscale<mode>): Ditto.
1363 (*avx512er_exp2<mode>): Ditto.
1364 (*avx512er_rcp28<mode>): Ditto.
1365 (avx512er_rsqrt28<mode>): Ditto.
1366 (avx512f_getmant<mode>): Ditto.
1367 (avx512f_getmant<mode>): Ditto.
1368 (avx512f_rndscale<mode>): Fix formatting.
1369
1370 2013-10-15 Martin Jambor <mjambor@suse.cz>
1371
1372 * ipa-utils.h (ipa_edge_within_scc): Declare.
1373 * ipa-cp.c (edge_within_scc): Moved...
1374 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1375
1376 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1377 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1378 Sergey Lega <sergey.s.lega@intel.com>
1379 Anna Tikhonova <anna.tikhonova@intel.com>
1380 Ilya Tocar <ilya.tocar@intel.com>
1381 Andrey Turetskiy <andrey.turetskiy@intel.com>
1382 Ilya Verbin <ilya.verbin@intel.com>
1383 Kirill Yukhin <kirill.yukhin@intel.com>
1384 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1385
1386 * config/i386/predicates.md (const_8_to_15_operand): New.
1387 (const_16_to_31_operand): Ditto.
1388 * config/i386/sse.md (V8FI): New.
1389 (V16FI): Ditto.
1390 (reduc_splus_v8df): Ditto.
1391 (reduc_splus_v16sf): Ditto.
1392 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1393 (vec_extract_hi_<mode>): Ditto.
1394 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1395 (vec_set_lo_<mode>): Ditto.
1396 (vec_set_hi_<mode>): Ditto.
1397 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1398 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1399 (avx512f_pshufd_1): Ditto.
1400 (avx512f_broadcast<mode>): Ditto.
1401 (avx512f_broadcast<mode>): Ditto.
1402 (define_split): Split vec_extract_lo into move.
1403 (ssequartermode): Ditto.
1404 (ssedoublemode): Extened with wider modes.
1405 (vec_extract_lo_<mode>): Ditto.
1406
1407 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1408 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1409 Sergey Lega <sergey.s.lega@intel.com>
1410 Anna Tikhonova <anna.tikhonova@intel.com>
1411 Ilya Tocar <ilya.tocar@intel.com>
1412 Andrey Turetskiy <andrey.turetskiy@intel.com>
1413 Ilya Verbin <ilya.verbin@intel.com>
1414 Kirill Yukhin <kirill.yukhin@intel.com>
1415 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1416
1417 * config/i386/predicates.md (register_or_constm1_operand): New.
1418 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1419 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1420 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1421 UNSPEC_SCATTER_PREFETCH
1422 (VI48_512): New.
1423 (avx512f_ucmp<mode>3): Ditto.
1424 (avx512f_vternlog<mode>): Ditto.
1425 (avx512f_align<mode>): Ditto.
1426 (<shift_insn><mode>3): Ditto.
1427 (avx512f_<rotate>v<mode>): Ditto.
1428 (avx512f_<rotate><mode>): Ditto.
1429 (avx512f_eq<mode>3): Ditto.
1430 (avx512f_eq<mode>3_1): Ditto.
1431 (avx512f_gt<mode>3): Ditto.
1432 (avx512f_testm<mode>3): Ditto.
1433 (avx512f_testnm<mode>3): Ditto.
1434 (avx512pf_gatherpf<mode>): Ditto.
1435 (*avx512pf_gatherpf<mode>_mask): Ditto.
1436 (*avx512pf_gatherpf<mode>): Ditto.
1437 (avx512pf_scatterpf<mode>): Ditto.
1438 (*avx512pf_scatterpf<mode>_mask): Ditto.
1439 (*avx512pf_scatterpf<mode>): Ditto.
1440 (avx512f_vec_dup_gpr<mode>): Ditto.
1441 (clz<mode>2): Ditto.
1442 (conflict<mode>): Ditto.
1443 (REDUC_SMINMAX_MODE): Extened with wider modes.
1444 (reduc_<code>_<mode>): Ditto.
1445 (vlshr<mode>3): Ditto.
1446 (vashl<mode>3): Ditto.
1447
1448 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1449 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1450 Sergey Lega <sergey.s.lega@intel.com>
1451 Anna Tikhonova <anna.tikhonova@intel.com>
1452 Ilya Tocar <ilya.tocar@intel.com>
1453 Andrey Turetskiy <andrey.turetskiy@intel.com>
1454 Ilya Verbin <ilya.verbin@intel.com>
1455 Kirill Yukhin <kirill.yukhin@intel.com>
1456 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1457
1458 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1459 UNSPEC_SCATTER.
1460 (VI48F_512): New.
1461 (avx512fmaskmode): Ditto.
1462 (bcstscalarsuff): Ditto.
1463 (avx512f_blendm<mode>): Ditto.
1464 (cmp_imm_predicate): Ditto.
1465 (avx512f_cmp<mode>3): Ditto.
1466 (avx512f_vec_dup<mode>): Ditto.
1467 (avx512f_vec_dup_mem<mode>): Ditto.
1468 (avx512f_vpermi2var<mode>3): Ditto.
1469 (avx512f_vpermt2var<mode>3): Ditto.
1470 (vec_init<mode>): Ditto.
1471 (avx512f_gathersi<mode>): Ditto.
1472 (*avx512f_gathersi<mode>): Ditto.
1473 (*avx512f_gathersi<mode>_2): Ditto.
1474 (avx512f_gatherdi<mode>): Ditto.
1475 (*avx512f_gatherdi<mode>): Ditto.
1476 (*avx512f_gatherdi<mode>_2): Ditto.
1477 (avx512f_scattersi<mode>): Ditto.
1478 (*avx512f_scattersi<mode>): Ditto.
1479 (avx512f_scatterdi<mode>): Ditto.
1480 (*avx512f_scatterdi<mode>): Ditto.
1481 (sseintprefix): Extened with wider modes.
1482 (VEC_GATHER_IDXSI): Ditto.
1483 (VEC_GATHER_IDXDI): Ditto.
1484 (VEC_GATHER_SRCDI): Ditto.
1485
1486 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1487 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1488
1489 * config/arm/t-aprofile: New file.
1490 * config.gcc: Handle --with-multilib-list option.
1491
1492 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1493
1494 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1495 the input for r2.
1496
1497 2013-10-15 Richard Biener <rguenther@suse.de>
1498
1499 * tree-loop-distribution.c (build_empty_rdg): Inline into
1500 single user.
1501 (rdg_flag_vertex): Inline into single user.
1502 (rdg_flag_vertex_and_dependent): Likewise.
1503 (build_rdg_partition_for_vertex): Remove processed bitmap.
1504 (rdg_build_partitions): Simplify.
1505
1506 2013-10-15 Richard Biener <rguenther@suse.de>
1507
1508 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1509 Restructure forwarding through conversions and copies to
1510 avoid performing copy-propagation the wrong way. Adjust
1511 recursion invocations.
1512 (forward_propagate_addr_expr): Add argument stating if we
1513 are recursing from a single-use.
1514 (ssa_forward_propagate_and_combine): Adjust.
1515
1516 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1517
1518 * dumpfile.h (gcc::dump_manager): New class, to hold state
1519 relating to dumpfile management.
1520 (get_dump_file_name): Remove in favor of method of dump_manager.
1521 (dump_initialized_p): Likewise.
1522 (dump_start): Likewise.
1523 (dump_finish): Likewise.
1524 (dump_switch_p): Likewise.
1525 (dump_register): Likewise.
1526 (get_dump_file_info): Likewise.
1527 * context.c (gcc::context::context): Construct the dump_manager
1528 instance.
1529 * context.h (gcc::context::get_dumps): New.
1530 (gcc::context::m_dumps): New.
1531 * coverage.c (coverage_init): Port to dump_manager API.
1532 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1533 (extra_dump_files_in_use): Likewise.
1534 (extra_dump_files_alloced): Likewise.
1535 (gcc::dump_manager::dump_manager): New.
1536 (dump_register): Convert to...
1537 (gcc::dump_manager::dump_register): ...method, replacing
1538 function-static next_dump with m_next_dump field.
1539 (get_dump_file_info): Convert to...
1540 (gcc::dump_manager::get_dump_file_info): ...method.
1541 (get_dump_file_name): Convert to...
1542 (gcc::dump_manager::get_dump_file_name): ...method.
1543 (dump_start): Convert to...
1544 (gcc::dump_manager::dump_start): ...method.
1545 (dump_finish): Convert to...
1546 (gcc::dump_manager::dump_finish): ...method.
1547 (dump_begin): Replace body with...
1548 (gcc::dump_manager::dump_begin): ...new method.
1549 (dump_phase_enabled_p): Convert to...
1550 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1551 (dump_phase_enabled_p): Convert to...
1552 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1553 (dump_initialized_p): Convert to...
1554 (gcc::dump_manager::dump_initialized_p): ...method.
1555 (dump_flag_name): Replace body with...
1556 (gcc::dump_manager::dump_flag_name): ...new method.
1557 (dump_enable_all): Convert to...
1558 (gcc::dump_manager::dump_enable_all): ...new method.
1559 (opt_info_enable_passes): Convert to...
1560 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1561 (dump_switch_p_1): Convert to...
1562 (gcc::dump_manager::dump_switch_p_1): ...new method.
1563 (dump_switch_p): Convert to...
1564 (gcc::dump_manager::dump_switch_p): ...new method.
1565 (opt_info_switch_p): Port to dump_manager API.
1566 (enable_rtl_dump_file): Likewise.
1567 * opts-global.c (handle_common_deferred_options): Port to new
1568 dump_manager API.
1569 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1570 (pass_manager::register_one_dump_file): Likewise.
1571 (pass_manager::register_pass): Likewise.
1572 (pass_init_dump_file): Likewise.
1573 (pass_fini_dump_file): Likewise.
1574 * statistics.c (statistics_early_init): Likewise.
1575
1576 2013-10-14 Richard Biener <rguenther@suse.de>
1577
1578 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1579 iterative_hash_canonical_type, gimple_canonical_type_hash,
1580 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1581 gimple_register_canonical_type, print_gimple_types_stats,
1582 free_gimple_type_tables): Move to lto/lto.c
1583 (gt-gimple.h): Do not include.
1584 * gimple.h (gimple_register_canonical_type,
1585 print_gimple_types_stats, free_gimple_type_tables): Remove.
1586 * Makefile.in (GTFILES): Remove gimple.c.
1587
1588 2013-10-14 Travis Snoozy <quandary@remstate.com>
1589
1590 PR target/58716
1591 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1592 scanning for msp430x targets.
1593
1594 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1595
1596 PR bootstrap/58509
1597 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1598 (registers_ok_for_ldd_peep): Move around.
1599 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1600 * config/sparc/sparc.md (widening peepholes): Use it.
1601
1602 2013-10-14 Richard Biener <rguenther@suse.de>
1603
1604 PR middle-end/58712
1605 PR middle-end/55358
1606 * gimple.c (iterative_hash_canonical_type): Make sure to
1607 record the hash into the correct hashtable slot.
1608
1609 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1610
1611 PR rtl-optimization/58662
1612 * combine.c (try_combine): Take into account death nodes on I2 when
1613 splitting a PARALLEL of two independent SETs. Fix dump message.
1614
1615 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1616
1617 PR target/51244
1618 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1619 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1620 * config/sh/t-sh (sh_treg_combine.o): New entry.
1621 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1622 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1623 (register_sh_passes): New function. Register sh_treg_combine pass.
1624 (sh_option_override): Invoke it.
1625 (sh_canonicalize_comparison): Handle op0_preserve_value.
1626 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1627 opportunities. Canonicalize branch condition.
1628 (nott): Allow only if pseudos can be created for non-SH2A.
1629
1630 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1631
1632 PR target/58690
1633 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1634 (ix86_expand_movmem): Replace copy_addr_to_reg with
1635 ix86_copy_addr_to_reg.
1636 (ix86_expand_setmem): Likewise.
1637
1638 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1639
1640 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1641 provided by ix86_fp_compare_mode instead of CCFPUmode.
1642
1643 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1644
1645 * config/aarch64/arm_neon.h
1646 (vtbx<1,3>_<psu>8): Fix register constriants.
1647
1648 2013-10-11 Jeff Law <law@redhat.com>
1649
1650 PR tree-optimization/58640
1651 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1652 threading paths that cross over two loop entry points.
1653
1654 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1655
1656 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1657 handle vector float as well.
1658 (*vsx_le_perm_load_v4si): Likewise.
1659 (*vsx_le_perm_store_v2di): Likewise.
1660 (*vsx_le_perm_store_v4si): Likewise.
1661
1662 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1663
1664 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1665 directly to circumvent subtract from splat{31} workaround.
1666 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1667 prototype.
1668 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1669 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1670 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1671 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1672 separate big and little endian logic.
1673 (*altivec_vperm_<mode>_internal): New define_insn.
1674 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1675 separate big and little endian logic.
1676 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1677 (vec_permv16qi): Add little endian logic.
1678
1679 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1680
1681 * doc/extend.texi (returns_nonnull): Remove arguments.
1682
1683 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1684 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1685 Sergey Lega <sergey.s.lega@intel.com>
1686 Anna Tikhonova <anna.tikhonova@intel.com>
1687 Ilya Tocar <ilya.tocar@intel.com>
1688 Andrey Turetskiy <andrey.turetskiy@intel.com>
1689 Ilya Verbin <ilya.verbin@intel.com>
1690 Kirill Yukhin <kirill.yukhin@intel.com>
1691 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1692
1693 * config/i386/sse.md (VI48F_256_512): New.
1694 (avx2_permvar<mode>): Change to ...
1695 (<avx2_avx512f>_permvar<mode>): This.
1696
1697 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1698 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1699 Sergey Lega <sergey.s.lega@intel.com>
1700 Anna Tikhonova <anna.tikhonova@intel.com>
1701 Ilya Tocar <ilya.tocar@intel.com>
1702 Andrey Turetskiy <andrey.turetskiy@intel.com>
1703 Ilya Verbin <ilya.verbin@intel.com>
1704 Kirill Yukhin <kirill.yukhin@intel.com>
1705 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1706
1707 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1708 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1709 * config/i386/sse.md (VI4_AVX): New.
1710 (sf2simodelower): Ditto.
1711 (sse2_cvtps2dq): Change to ...
1712 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1713
1714 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1715 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1716 Sergey Lega <sergey.s.lega@intel.com>
1717 Anna Tikhonova <anna.tikhonova@intel.com>
1718 Ilya Tocar <ilya.tocar@intel.com>
1719 Andrey Turetskiy <andrey.turetskiy@intel.com>
1720 Ilya Verbin <ilya.verbin@intel.com>
1721 Kirill Yukhin <kirill.yukhin@intel.com>
1722 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1723
1724 * config/i386/sse.md (V_512): New.
1725 (VI_512): Ditto.
1726 (vcond<V_512:mode><VF_512:mode>): Ditto.
1727 (vcond<V_512:mode><VI_512:mode>): Ditto.
1728 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1729
1730 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1731 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1732 Sergey Lega <sergey.s.lega@intel.com>
1733 Anna Tikhonova <anna.tikhonova@intel.com>
1734 Ilya Tocar <ilya.tocar@intel.com>
1735 Andrey Turetskiy <andrey.turetskiy@intel.com>
1736 Ilya Verbin <ilya.verbin@intel.com>
1737 Kirill Yukhin <kirill.yukhin@intel.com>
1738 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1739
1740 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1741 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1742 (FMAMODE): Ditto.
1743 (fma<mode>4): Removed condition.
1744 (fms<mode>4): Ditto.
1745 (fnma<mode>4): Ditto.
1746 (fnms<mode>4): Ditto.
1747 (fma4i_fmadd_<mode>): Ditto.
1748 (*fma_fmadd_<mode>): Ditto.
1749 (*fma_fmsub_<mode>): Ditto.
1750 (*fma_fnmadd_<mode>): Ditto.
1751 (*fma_fnmsub_<mode>): Ditto.
1752 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1753 (*fma_fmaddsub_<mode>): Ditto.
1754 (*fma_fmsubadd_<mode>): Ditto.
1755 (*fmai_fmadd_<mode>): Ditto.
1756 (*fmai_fmsub_<mode>): Ditto.
1757 (*fmai_fnmadd_<mode>): Ditto.
1758 (*fmai_fnmsub_<mode>): Ditto.
1759
1760 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1761 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1762 Sergey Lega <sergey.s.lega@intel.com>
1763 Anna Tikhonova <anna.tikhonova@intel.com>
1764 Ilya Tocar <ilya.tocar@intel.com>
1765 Andrey Turetskiy <andrey.turetskiy@intel.com>
1766 Ilya Verbin <ilya.verbin@intel.com>
1767 Kirill Yukhin <kirill.yukhin@intel.com>
1768 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1769
1770 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1771 (VI124_256): Changed to ...
1772 (VI124_256_48_512): This.
1773 (ssepackmode): Extended with wider modes.
1774 (<code><mode>3): Changed iterator.
1775 (*avx2_<code><mode>3): Ditto.
1776 (vec_pack_trunc_<mode>): Ditto.
1777
1778 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1779 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1780 Sergey Lega <sergey.s.lega@intel.com>
1781 Anna Tikhonova <anna.tikhonova@intel.com>
1782 Ilya Tocar <ilya.tocar@intel.com>
1783 Andrey Turetskiy <andrey.turetskiy@intel.com>
1784 Ilya Verbin <ilya.verbin@intel.com>
1785 Kirill Yukhin <kirill.yukhin@intel.com>
1786 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1787
1788 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1789 (VI8F_256_512): Ditto.
1790 (abs<mode>2): Changed iterator.
1791 (avx2_perm<mode>): Changed to ...
1792 (<avx2_avx512f>_perm<mode>): This.
1793 (avx2_perm<mode>_1): Changed to ...
1794 (<avx2_avx512f>_perm<mode>_1): This.
1795
1796 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1797 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1798 Sergey Lega <sergey.s.lega@intel.com>
1799 Anna Tikhonova <anna.tikhonova@intel.com>
1800 Ilya Tocar <ilya.tocar@intel.com>
1801 Andrey Turetskiy <andrey.turetskiy@intel.com>
1802 Ilya Verbin <ilya.verbin@intel.com>
1803 Kirill Yukhin <kirill.yukhin@intel.com>
1804 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1805
1806 * config/i386/sse.md (VI48_AVX512F): New.
1807 (VI48_AVX2): Changed to ...
1808 (VI48_AVX2_48_AVX512F): This.
1809 (avx2_ashrv<mode>): Changed to ...
1810 (<avx2_avx512f>_ashrv<mode>): This.
1811 (avx2_<shift_insn>v<mode>): Changed to ...
1812 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1813
1814 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1815 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1816 Sergey Lega <sergey.s.lega@intel.com>
1817 Anna Tikhonova <anna.tikhonova@intel.com>
1818 Ilya Tocar <ilya.tocar@intel.com>
1819 Andrey Turetskiy <andrey.turetskiy@intel.com>
1820 Ilya Verbin <ilya.verbin@intel.com>
1821 Kirill Yukhin <kirill.yukhin@intel.com>
1822 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1823
1824 * config/i386/sse.md (VI4_AVX512F): New.
1825 (VI8_AVX2_AVX512F): Ditto.
1826 (mul<mode>3): Extended with wider modes.
1827 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1828 (mul<mode>3): Ditto.
1829 (vec_widen_<s>mult_odd_<mode>): Ditto.
1830
1831 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1832 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1833 Sergey Lega <sergey.s.lega@intel.com>
1834 Anna Tikhonova <anna.tikhonova@intel.com>
1835 Ilya Tocar <ilya.tocar@intel.com>
1836 Andrey Turetskiy <andrey.turetskiy@intel.com>
1837 Ilya Verbin <ilya.verbin@intel.com>
1838 Kirill Yukhin <kirill.yukhin@intel.com>
1839 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1840
1841 * config/i386/sse.md (VI2_AVX512F): New.
1842 (VI124_AVX512F): Ditto.
1843 (sseunpackmode): Extended with wider modes.
1844 (sseunpackfltmode): Ditto.
1845 (vec_unpacks_float_hi_<mode>): Ditto.
1846 (vec_unpacks_float_lo_<mode>): Ditto.
1847 (vec_unpacku_float_hi_<mode>): Ditto.
1848 (vec_unpacku_float_lo_<mode>): Ditto.
1849 (vec_unpacks_lo_<mode>): Ditto.
1850 (vec_unpacks_hi_<mode>): Ditto.
1851 (vec_unpacku_lo_<mode>): Ditto.
1852 (vec_unpacku_hi_<mode>): Ditto.
1853
1854 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1855 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1856 Sergey Lega <sergey.s.lega@intel.com>
1857 Anna Tikhonova <anna.tikhonova@intel.com>
1858 Ilya Tocar <ilya.tocar@intel.com>
1859 Andrey Turetskiy <andrey.turetskiy@intel.com>
1860 Ilya Verbin <ilya.verbin@intel.com>
1861 Kirill Yukhin <kirill.yukhin@intel.com>
1862 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1863
1864 * config/i386/i386.md (multdiv): New.
1865 (multdiv_mnemonic): Ditto.
1866 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1867 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1868 (<sse>_vmdiv<mode>3): Removed.
1869
1870 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1871 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1872 Sergey Lega <sergey.s.lega@intel.com>
1873 Anna Tikhonova <anna.tikhonova@intel.com>
1874 Ilya Tocar <ilya.tocar@intel.com>
1875 Andrey Turetskiy <andrey.turetskiy@intel.com>
1876 Ilya Verbin <ilya.verbin@intel.com>
1877 Kirill Yukhin <kirill.yukhin@intel.com>
1878 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1879
1880 * config/i386/sse.md (V): Extended with wider modes.
1881 (VF2): Ditto.
1882 (ssehalfvecmode): Ditto.
1883 (i128): Ditto.
1884 (ssepackfltmode): Ditto.
1885 (avx_vec_concat<mode>): Ditto.
1886 (V_256_512): New iterator.
1887 (VF2_512_256): Ditto.
1888 (si2dfmode): New attribute.
1889 (si2dfmodelower): Ditto.
1890 (sf2dfmode): Ditto.
1891 (concat_tg_mode): Ditto.
1892 (floatv4siv4df2): Changed to ...
1893 (float<si2dfmodelower><mode>2): This.
1894 (avx_cvtps2pd256): Changed to ...
1895 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1896 (vec_pack_trunc_v4df): Changed to ...
1897 (vec_pack_trunc_<mode>): This.
1898 (avx_vpermil<mode>): Changed to ...
1899 (<sse2_avx_avx512f>_vpermil<mode>): This.
1900 (<fixsuffix>fix_truncv8dfv8si2): New.
1901 (vec_pack_sfix_trunc_v8df): Ditto.
1902 (avx512f_rndscale<mode>): Ditto.
1903 (avx512f_roundpd512): Ditto.
1904 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1905
1906 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1907 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1908 Sergey Lega <sergey.s.lega@intel.com>
1909 Anna Tikhonova <anna.tikhonova@intel.com>
1910 Ilya Tocar <ilya.tocar@intel.com>
1911 Andrey Turetskiy <andrey.turetskiy@intel.com>
1912 Ilya Verbin <ilya.verbin@intel.com>
1913 Kirill Yukhin <kirill.yukhin@intel.com>
1914 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1915
1916 * config/i386/i386.md (any_fix): New iterator.
1917 (fixsuffix): New attribute.
1918 * config/i386/sse.md (VF1): Extened with wider modes.
1919 (VI): Ditto.
1920 (VI_AVX2): Ditto.
1921 (VI8): Ditto.
1922 (sseintvecmodelower): Ditto.
1923 (ssescalarmode): Ditto.
1924 (ssescalarnum): Ditto.
1925 (VF1_128_256): New.
1926 (ssexmmmode): Ditto.
1927 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1928 (<sse>_rcp<mode>2): Change iterator.
1929 (rsqrt<mode>2): Ditto.
1930 (<sse>_rsqrt<mode>2): Ditto.
1931 (avx2_vec_dup<mode>): Ditto.
1932 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1933 (round<mode>2_sfix): Ditto.
1934 (avx2_pbroadcast<mode>): Ditto.
1935 (*andnot<mode>3): Handle XI mode.
1936 (*<code><mode>3): Ditto.
1937 (AVXTOSSEMODE): Removed.
1938 (avx_vpermil<mode>): Changed to ...
1939 (<sse2_avx_avx512f>_vpermil<mode>): This.
1940
1941 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1942 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1943 Sergey Lega <sergey.s.lega@intel.com>
1944 Anna Tikhonova <anna.tikhonova@intel.com>
1945 Ilya Tocar <ilya.tocar@intel.com>
1946 Andrey Turetskiy <andrey.turetskiy@intel.com>
1947 Ilya Verbin <ilya.verbin@intel.com>
1948 Kirill Yukhin <kirill.yukhin@intel.com>
1949 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1950
1951 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1952 (<sse>_comi): Ditto.
1953 (<sse>_ucomi): Ditto.
1954 (sse_cvtss2siq_2): Ditto.
1955 (sse2_cvtsd2si): Ditto.
1956 (sse2_cvtsd2siq): Ditto.
1957 (sse2_cvttsd2si): Ditto.
1958 (sse2_cvttsd2siq): Ditto.
1959 (<shift_insn><mode>3): Ditto.
1960 (sse2_cvtsi2sdq): Update constraint and prefix.
1961 (sse_cvtsi2ss): Update prefix.
1962 (sse_cvtsi2ssq): Ditto.
1963
1964 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1965
1966 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1967 ignore internal calls.
1968
1969 2013-10-11 Richard Biener <rguenther@suse.de>
1970
1971 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
1972 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
1973 TREE_OVERFLOW is set.
1974
1975 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1976
1977 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1978
1979 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1980 * gimple.h: Likewise.
1981
1982 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1983
1984 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1985 macros.
1986
1987 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1988
1989 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1990 * configure: Regenerate.
1991
1992 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1993
1994 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1995 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1996 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1997 similarly pp_right_paren.
1998 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1999 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2000 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2001 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2002 clauses.
2003 (omp_declare_simd_clauses_equal,
2004 omp_remove_redundant_declare_simd_attrs): New functions.
2005 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2006 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2007 * tree.h (OMP_LOOP_CHECK): Define.
2008 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2009 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2010 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2011 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2012 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2013 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2014 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2015 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2016 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2017 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2018 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2019 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2020 (omp_remove_redundant_declare_simd_attrs): New prototype.
2021 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2022 GIMPLE_OMP_TEAMS): New codes.
2023 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2024 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2025 fields.
2026 (target_nesting_level): New variable.
2027 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2028 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2029 for collapse > 1 static schedule unless ordered.
2030 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2031 (determine_parallel_type): Adjust get_ws_args_for caller.
2032 (install_var_field): Handle mask & 4 for double indirection.
2033 (scan_sharing_clauses): Ignore shared clause on teams construct.
2034 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2035 (create_omp_child_function): If inside target or declare target
2036 constructs, set "omp declare target" attribute on the child function.
2037 (find_combined_for): New function.
2038 (scan_omp_parallel): Handle combined loops.
2039 (scan_omp_target, scan_omp_teams): New functions.
2040 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2041 restrictions and set ctx->cancellable for cancellable constructs.
2042 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2043 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2044 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2045 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2046 (omp_clause_aligned_alignment): New function.
2047 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2048 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2049 on teams constructs. Handle user defined reductions and new
2050 OpenMP 4.0 clauses.
2051 (lower_reduction_clauses): Don't set placeholder to address of ref
2052 if it has already the right type.
2053 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2054 (expand_parallel_call): Use the new non-_start suffixed builtins,
2055 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2056 and GOMP_parallel_end after the call.
2057 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2058 (expand_omp_for_init_counts): Handle combined loops.
2059 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2060 loops.
2061 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2062 end of cancellable loops.
2063 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2064 Likewise. Handle collapse > 1 loops.
2065 (expand_omp_simd): Handle combined loops.
2066 (expand_omp_for): Add inner_stmt argument, adjust callers of
2067 expand_omp_for* functions, use expand_omp_for_static*chunk even
2068 for collapse > 1 unless ordered.
2069 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2070 of cancellable sections.
2071 (expand_omp_single): Remove need_barrier variable, just rely on
2072 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2073 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2074 (expand_omp_atomic_load, expand_omp_atomic_store,
2075 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2076 (expand_omp_target): New function.
2077 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2078 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2079 (build_omp_regions_1): Immediately close region for
2080 GF_OMP_TARGET_KIND_UPDATE.
2081 (maybe_add_implicit_barrier_cancel): New function.
2082 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2083 cancellation.
2084 (lower_omp_single): Likewise. Add clobber after the barrier.
2085 (lower_omp_taskgroup): New function.
2086 (lower_omp_for): Handle combined loops. Adjust
2087 lower_rec_input_clauses caller. Handle cancellation.
2088 (lower_depend_clauses): New function.
2089 (lower_omp_taskreg): Lower depend clauses. Adjust
2090 lower_rec_input_clauses caller. Add clobber after the call. Handle
2091 cancellation.
2092 (lower_omp_target, lower_omp_teams): New functions.
2093 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2094 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2095 and GOMP_cancellation_point calls.
2096 (lower_omp): Fold stmts inside of target region.
2097 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2098 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2099 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2100 (BT_FN_VOID_OMPFN_PTR_UINT,
2101 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2102 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2103 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2104 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2105 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2106 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2107 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2108 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2109 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2110 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2111 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2112 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2113 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2114 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2115 BUILT_IN_GOMP_PARALLEL_END.
2116 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2117 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2118 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2119 GF_OMP_FOR_KIND_DISTRIBUTE.
2120 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2121 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2122 (dump_gimple_omp_return): Print lhs if it has any.
2123 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2124 gimple_omp_atomic_seq_cst_p.
2125 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2126 and GIMPLE_OMP_TEAMS.
2127 * langhooks.c (lhd_omp_mappable_type): New function.
2128 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2129 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2130 hook.
2131 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2132 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2133 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2134 (struct gimplify_omp_ctx): Add combined_loop field.
2135 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2136 on stmts inside of target region.
2137 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2138 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2139 ORT_TARGET and ORT_TARGET_DATA.
2140 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2141 Handle GOVD_MAP.
2142 (omp_notice_threadprivate_variable): Complain about threadprivate
2143 variables in target region.
2144 (omp_notice_variable): Complain about vars with non-mappable type
2145 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2146 (omp_check_private): Ignore ORT_TARGET* regions.
2147 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2148 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2149 (find_combined_omp_for): New function.
2150 (gimplify_omp_for): Handle gimplification of combined loops.
2151 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2152 OMP_TEAMS.
2153 (gimplify_omp_target_update): New function.
2154 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2155 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2156 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2157 (gimplify_body): If fndecl has "omp declare target" attribute, add
2158 implicit ORT_TARGET context around it.
2159 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2160 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2161 * tree-nested.c (convert_nonlocal_reference_stmt,
2162 convert_local_reference_stmt, convert_gimple_call): Handle
2163 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2164 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2165 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2166 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2167 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2168 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2169 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2170 BUILT_IN_GOMP_LOOP_END_CANCEL,
2171 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2172 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2173 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2174 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2175 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2176 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2177 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2178 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2179 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2180 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2181 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2182 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2183 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2184 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2185 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2186 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2187 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2188 gimple_build_omp_teams): New functions.
2189 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2190 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2191 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2192 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2193 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2194 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2195 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2196 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2197 GF_OMP_ATOMIC_SEQ_CST): New.
2198 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2199 gimple_build_omp_teams): New prototypes.
2200 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2201 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2202 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2203 GIMPLE_OMP_SINGLE as end of range.
2204 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2205 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2206 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2207 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2208 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2209 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2210 gimple_omp_target_kind, gimple_omp_target_set_kind,
2211 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2212 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2213 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2214 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2215 gimple_omp_teams_set_clauses): New inlines.
2216 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2217 and GIMPLE_OMP_TASKGROUP.
2218 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2219 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2220 enum omp_clause_proc_bind_kind): New.
2221 (union omp_clause_subcode): Add depend_kind, map_kind and
2222 proc_bind_kind fields.
2223 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2224 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2225 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2226 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2227 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2228
2229 2013-10-10 Teresa Johnson <tejohnson@google.com>
2230
2231 * predict.c (tree_estimate_probability): Add new parameter
2232 for estimate_bb_frequencies.
2233 (estimate_bb_frequencies): Add new parameter to force estimation.
2234 (rebuild_frequencies): When max frequency in function is small,
2235 recompute counts from frequencies.
2236 * predict.h (estimate_bb_frequencies): New parameter.
2237
2238 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2239
2240 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2241 optimizations are disabled.
2242
2243 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2244
2245 * coverage.c (coverage_finish): Fix leak of da_file_name.
2246
2247 2013-10-10 Jan Hubicka <jh@suse.cz>
2248
2249 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2250 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2251 Bobcat and generic.
2252
2253 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2254
2255 PR middle-end/58670
2256 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2257 if any labels are in FALLTHRU_BB, use a special label emitted
2258 immediately after the asm goto insn rather than label_rtx
2259 of the LABEL_DECL.
2260 (expand_asm_stmt): Adjust caller.
2261 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2262 edge if the last insn in predecessor is a jump with single successor,
2263 but it isn't simplejump_p.
2264
2265 2013-10-10 Richard Biener <rguenther@suse.de>
2266
2267 PR tree-optimization/58656
2268 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2269
2270 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2271
2272 * gimplify.c: Include expr.h and tm_p.h for targets with special
2273 va-arg padding requirements.
2274
2275 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2276
2277 * tree-flow.h: Move some prototypes to gimple.h.
2278 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2279 * gimple.h: Relocate some prototypes from tree-flow.h
2280 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2281 Move to gimplify.c.
2282 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2283 (build_va_arg_indirect_ref): Relocate and make static.
2284 (std_gimplify_va_arg_expr): Relocate here.
2285 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2286 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2287
2288 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2289
2290 * doc/md.texi: Document the mnemonic attribute.
2291
2292 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2293
2294 PR target/57377
2295 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2296 (set_attr_alternative x ...) when searching for user defined
2297 mnemonic attribute.
2298
2299 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2300
2301 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2302
2303 2013-10-09 Easwaran Raman <eraman@google.com>
2304
2305 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2306 * cfgexpand.c (defer_stack_allocation): ...use here
2307 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2308
2309 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2310
2311 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2312 (operand_equal_for_value_replacement): New function, extracted from
2313 value_replacement and enhanced to catch more cases.
2314 (value_replacement): Use operand_equal_for_value_replacement.
2315
2316 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2317
2318 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2319 get_max_loop_iterations.
2320
2321 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2322
2323 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2324 Place comment above function.
2325
2326 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2327
2328 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2329 are not related to tree-cfg.c.
2330 * tree-ssa-address.h: New file. Relocate prototypes.
2331 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2332 (addr_for_mem_ref): New. Combine call to get_address_description and
2333 return addr_for_mem_ref.
2334 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2335 * tree-ssa-live.h: Adjust prototypes.
2336 * passes.c: Include tree-ssa-live.h.
2337 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2338 * graphite.c (graphite_transform_loops): Make static.
2339 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2340 make_pass_graphite, pass_data_graphite_transforms,
2341 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2342 * ipa-pure-const.c (warn_function_noreturn): Make static.
2343 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2344 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2345 Relocate from tree-cfg.c
2346 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2347 static.
2348 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2349 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2350 Move to ipa-pure-const.c.
2351 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2352 Relocate from tree-optimize.c.
2353 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2354 make_pass_fixup_cfg): Move to tree-cfg.c.
2355 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2356 Relocate some prototypes.
2357 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2358 * tree-dump.c (dump_function_to_file): Remove prototype.
2359 Add tree-flow.h to the include file.
2360 * tree-dump.h: Remove prototype.
2361 * tree-parloops.h: New File. Add prototypes.
2362 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2363 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2364 from tree-ssa-loop.c.
2365 * tree-predcom.c (run_tree_predictive_commoning,
2366 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2367 Relocate here from tree-ssa-loop.c.
2368 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2369 ssa_name_values.release ().
2370 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2371 (ssa_name_values): Relocate from tree-flow.h.
2372 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2373 * tree-ssa-loop.c (run_tree_predictive_commoning,
2374 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2375 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2376 make_pass_graphite, pass_data_graphite_transforms,
2377 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2378 tree_parallelize_loops, pass_data_parallelize_loops,
2379 make_pass_parallelize_loops): Move to other files.
2380 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2381 moved here.
2382 * tree.h: Remove prototypes from tree-address.c.
2383
2384 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2385
2386 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2387 (struct int_tree_map): Move to tree-hasher.h
2388 (SCALE, LABEL, PERCENT): Move to gimple.h
2389 * tree-flow-inline.h: Delete. Move functions to other files.
2390 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2391 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2392 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2393 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2394 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2395 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2396 * tree-hasher.h: Don't include tree-flow.h.
2397 (struct int_tree_map): Relocate from tree-flow.h.
2398 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2399 tree-flow-inline.h and make static.
2400 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2401 tree-flow-inline.h.
2402 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2403 tree-flow-inline.h and make static.
2404 * tree.h (is_global_var, may_be_aliased): Relocate from
2405 tree-flow-inline.h.
2406 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2407 * value-prof.c: No longer include tree-flow-inline.h.
2408 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2409
2410 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2411
2412 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2413 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2414 (enum move_pos): Move to tree-ssa-loop-im.h
2415 * cfgloop.h: Move some prototypes.
2416 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2417 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2418 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2419 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2420 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2421 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2422 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2423 (enum move_pos): Relocate here.
2424 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2425 tree-ssa-loop.c.
2426 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2427 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2428 make_pass_lim): Relocate here from tree-ssa-loop.c.
2429 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2430 tree-ssa-loop.c.
2431 (loop_edge_to_cancel, unloop_loops): Make static.
2432 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2433 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2434 (tree_complete_unroll, gate_tree_complete_unroll,
2435 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2436 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2437 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2438 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2439 (stmt_invariant_in_loop_p): Remove unused function.
2440 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2441 * tree-ssa-loop-manip.h: New file. Add prototypes.
2442 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2443 (gcov_type_to_double_int): Move to cfgloop.h.
2444 (double_int_cmp, bound_index,
2445 estimate_numbers_of_iterations_loop): Make static.
2446 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2447 (max_loop_iterations): Factor out get_max_loop_iterations.
2448 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2449 cfgloop.c.
2450 * tree-ssa-loop-niter.h: New file. Add prototypes.
2451 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2452 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2453 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2454 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2455 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2456 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2457 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2458 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2459 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2460 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2461 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2462 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2463 pass_data_complete_unroll, make_pass_complete_unroll,
2464 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2465 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2466 tree-ssa-loop-ivcanon.c.
2467 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2468 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2469 tree-ssa-loop-prefetch.c.
2470 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2471 tree-ssa-loop-im.c.
2472 (get_lsm_tmp_name): Relocate and add suffix parameter.
2473 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2474 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2475 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2476 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2477 (get_estimated_loop_iterations): Factor out accessor from
2478 estimated_loop_iterations in tree-ssa-loop-niter.c.
2479 (get_max_loop_iterations): Factor out accessor from
2480 _max_loop_iterations in tree-ssa-niter.c.
2481 * loop-unroll.c (decide_unroll_constant_iterations,
2482 decide_unroll_runtime_iterations, decide_peel_simple,
2483 decide_unroll_stupid): Use new get_* accessors.
2484
2485 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2486
2487 PR tree-optimization/20318
2488 * doc/extend.texi (returns_nonnull): New function attribute.
2489 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2490 attribute.
2491 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2492 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2493
2494 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2495
2496 PR middle-end/58570
2497 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2498 false if both components are bitfields.
2499
2500 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2501
2502 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2503 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2504 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2505 * config/aarch64/aarch64.h
2506 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2507 * config/aarch64/aarch64-simd-builtins.def
2508 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2509
2510 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2511
2512 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2513
2514 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2515
2516 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2517
2518 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2519
2520 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2521 (vneg_f64): New intrinsic.
2522 (vneg_s8): Asm replaced with C.
2523 (vneg_s16): Likewise.
2524 (vneg_s32): Likewise.
2525 (vneg_s64): New intrinsic.
2526 (vnegq_f32): Asm replaced with C.
2527 (vnegq_f64): Likewise.
2528 (vnegq_s8): Likewise.
2529 (vnegq_s16): Likewise.
2530 (vnegq_s32): Likewise.
2531 (vnegq_s64): Likewise.
2532
2533 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2534
2535 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2536
2537 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2538
2539 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2540 packed stack special handling.
2541 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2542 back to fixed stack slots for FPRs saved due to stdarg.
2543
2544 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2545
2546 * config/s390/s390.c (s390_frame_info): Restructure function.
2547
2548 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2549
2550 * config/s390/s390.c (struct s390_frame_layout): New field
2551 gpr_save_slots.
2552 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2553 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2554 regs_ever_clobbered to char*.
2555 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2556 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2557 (s390_register_info): Enable FPR save slots. Move/Copy some
2558 functionality into ...
2559 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2560 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2561 function.
2562 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2563 not imply a stack frame.
2564 (s390_init_frame_layout): Remove variable clobbered_regs.
2565 (s390_update_register_info): Remove function.
2566 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2567 cannot be used for register renaming.
2568 (s390_hard_regno_scratch_ok): New function.
2569 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2570 (s390_initial_elimination_offset): Change offset calculation of
2571 the return address pointer.
2572 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2573 (restore_gprs): Set frame related flag.
2574 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2575 (s390_emit_prologue): Call s390_register_info instead of
2576 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2577 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2578 (s390_optimize_prologue): Call s390_optimize_register_info.
2579 Try to remove also FPR slot save/restore INSNs. Remove frame
2580 related flags from restore INSNs.
2581
2582 2013-10-08 DJ Delorie <dj@redhat.com>
2583
2584 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2585 (movhi): Likewise.
2586
2587 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2588 avoid conflict with the MI use of %c.
2589 * config/rl78/rl78-real.md: change %c to %C throughout.
2590 * config/rl78/rl78-virt.md: Likewise.
2591
2592 2013-10-08 Jan Hubicka <jh@suse.cz>
2593
2594 * config/i386/i386.c (ix86_option_override_internal): Switch
2595 to SSE math for -ffast-math when target ISA supports SSE2.
2596
2597 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
2598
2599 * tree-flow.h: Remove some prototypes.
2600 * tree.h: Remove some protypes, add a couple.
2601 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2602 using_eh_for_cleanups_p): Add interface routines for front ends.
2603 * tree-eh.h: New file. Add protoptyes.
2604 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2605 (add_stmt_to_eh_lp_fn): Make static.
2606 (lower_try_finally): Use new using_eh_for_cleanups_p.
2607 * emit-rtl.c: Include tree-eh.h.
2608 * gimple.h: Include tree-eh.h.
2609
2610 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2611
2612 PR tree-optimization/58480
2613 * tree-vrp.c (infer_nonnull_range): New function.
2614 (infer_value_range): Call infer_nonnull_range.
2615
2616 2013-10-08 Dehao Chen <dehao@google.com>
2617
2618 PR tree-optimization/58619
2619 * tree-inline.c (copy_phis_for_bb): Combine location data
2620 only if non-null.
2621
2622 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2623
2624 PR target/58423
2625 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2626 RTX_FRAME_RELATED_P on INSN.
2627
2628 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2629
2630 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2631 (altivec_expand_vec_perm_const): Call it.
2632
2633 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2634
2635 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2636 sequences for LE VSX loads and stores at expand time.
2637 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2638 prototype.
2639 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2640 (rs6000_gen_le_vsx_permute): New.
2641 (rs6000_gen_le_vsx_load): New.
2642 (rs6000_gen_le_vsx_store): New.
2643 (rs6000_gen_le_vsx_move): New.
2644 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2645 (*vsx_le_perm_load_v4si): New.
2646 (*vsx_le_perm_load_v8hi): New.
2647 (*vsx_le_perm_load_v16qi): New.
2648 (*vsx_le_perm_store_v2di): New.
2649 (*vsx_le_perm_store_v4si): New.
2650 (*vsx_le_perm_store_v8hi): New.
2651 (*vsx_le_perm_store_v16qi): New.
2652 (*vsx_xxpermdi2_le_<mode>): New.
2653 (*vsx_xxpermdi4_le_<mode>): New.
2654 (*vsx_xxpermdi8_le_V8HI): New.
2655 (*vsx_xxpermdi16_le_V16QI): New.
2656 (*vsx_lxvd2x2_le_<mode>): New.
2657 (*vsx_lxvd2x4_le_<mode>): New.
2658 (*vsx_lxvd2x8_le_V8HI): New.
2659 (*vsx_lxvd2x16_le_V16QI): New.
2660 (*vsx_stxvd2x2_le_<mode>): New.
2661 (*vsx_stxvd2x4_le_<mode>): New.
2662 (*vsx_stxvd2x8_le_V8HI): New.
2663 (*vsx_stxvd2x16_le_V16QI): New.
2664
2665 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2666
2667 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2668
2669 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2670
2671 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2672 loop to work also for 31bit ABI.
2673 Save the stack pointer for frame_size > 0.
2674
2675 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2676
2677 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2678 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2679 constraint letters from expanders.
2680 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2681 retry count to general_operand.
2682 ("tabort"): Give operand 0 a mode.
2683 ("tabort_1"): Add mode and constraint letter for operand 0.
2684 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2685
2686 2013-10-04 Jeff Law <law@redhat.com>
2687
2688 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2689
2690 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2691 out of ...
2692 (thread_across_edge): Here. Call it.
2693
2694 2013-10-04 Cary Coutant <ccoutant@google.com>
2695
2696 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2697 discarding a location list expression (or a piece of one).
2698
2699 2013-10-03 Jan Hubicka <jh@suse.cz>
2700
2701 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2702 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2703 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2704
2705 2013-10-03 Jan Hubicka <jh@suse.cz>
2706
2707 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2708 accumulate-outgoing-args when producing unwind info.
2709
2710 2013-10-03 Wei Mi <wmi@google.com>
2711
2712 * lra-constraints.c (insert_move_for_subreg): New function
2713 extracted from simplify_operand_subreg.
2714 (simplify_operand_subreg): Add reload for paradoxical subreg.
2715
2716 2013-10-03 Rong Xu <xur@google.com>
2717
2718 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2719 the candidate of builtin_expect such that we should fix the
2720 size/time estimation.
2721 (estimate_function_body_sizes): Do the acutally size/time fix-up
2722 for builtin_expect.
2723
2724 2013-10-03 Rong Xu <xur@google.com>
2725
2726 * predict.c (tree_predict_by_opcode): Get the probability
2727 for builtin_expect from param builtin_expect_probability.
2728 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2729 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2730 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2731
2732 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2733
2734 PR c++/19476
2735 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2736 * calls.c (alloca_call_p): Use get_callee_fndecl.
2737 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2738 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2739 Likewise.
2740 (vrp_visit_stmt): Remove duplicated code.
2741
2742 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2743
2744 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2745 ceildf2, btruncdf2, instead of vsx_* name.
2746
2747 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2748 iterators to only do V2DF and V4SF here. Move the DF code to
2749 rs6000.md where it is combined with SF mode. Replace <VSv> with
2750 just 'v' since only vector operations are handled with these insns
2751 after moving the DF support to rs6000.md.
2752 (vsx_sub<mode>3): Likewise.
2753 (vsx_mul<mode>3): Likewise.
2754 (vsx_div<mode>3): Likewise.
2755 (vsx_fre<mode>2): Likewise.
2756 (vsx_neg<mode>2): Likewise.
2757 (vsx_abs<mode>2): Likewise.
2758 (vsx_nabs<mode>2): Likewise.
2759 (vsx_smax<mode>3): Likewise.
2760 (vsx_smin<mode>3): Likewise.
2761 (vsx_sqrt<mode>2): Likewise.
2762 (vsx_rsqrte<mode>2): Likewise.
2763 (vsx_fms<mode>4): Likewise.
2764 (vsx_nfma<mode>4): Likewise.
2765 (vsx_copysign<mode>3): Likewise.
2766 (vsx_btrunc<mode>2): Likewise.
2767 (vsx_floor<mode>2): Likewise.
2768 (vsx_ceil<mode>2): Likewise.
2769 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2770 (vsx_sminsf3): Likewise.
2771 (vsx_fmadf4): Likewise.
2772 (vsx_fmsdf4): Likewise.
2773 (vsx_nfmadf4): Likewise.
2774 (vsx_nfmsdf4): Likewise.
2775 (vsx_cmpdf_internal1): Likewise.
2776
2777 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2778 simpler to select whether a target has SPE or traditional floating
2779 point support in iterators.
2780 (TARGET_DF_SPE): Likewise.
2781 (TARGET_SF_FPR): Likewise.
2782 (TARGET_DF_FPR): Likewise.
2783 (TARGET_SF_INSN): Macros to say whether floating point support
2784 exists for a given operation for expanders.
2785 (TARGET_DF_INSN): Likewise.
2786
2787 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2788 combining of SF/DF mode operations, using both traditional and VSX
2789 registers.
2790 (Fvsx): Likewise.
2791 (Ff): Likewise.
2792 (Fv): Likewise.
2793 (Fs): Likewise.
2794 (Ffre): Likewise.
2795 (FFRE): Likewise.
2796 (abs<mode>2): Combine SF/DF modes using traditional floating point
2797 instructions. Add support for using the upper DF registers with
2798 VSX support, and SF registers with power8-vector support. Update
2799 expanders for operations supported by both the SPE and traditional
2800 floating point units.
2801 (abs<mode>2_fpr): Likewise.
2802 (nabs<mode>2): Likewise.
2803 (nabs<mode>2_fpr): Likewise.
2804 (neg<mode>2): Likewise.
2805 (neg<mode>2_fpr): Likewise.
2806 (add<mode>3): Likewise.
2807 (add<mode>3_fpr): Likewise.
2808 (sub<mode>3): Likewise.
2809 (sub<mode>3_fpr): Likewise.
2810 (mul<mode>3): Likewise.
2811 (mul<mode>3_fpr): Likewise.
2812 (div<mode>3): Likewise.
2813 (div<mode>3_fpr): Likewise.
2814 (sqrt<mode>3): Likewise.
2815 (sqrt<mode>3_fpr): Likewise.
2816 (fre<Fs>): Likewise.
2817 (rsqrt<mode>2): Likewise.
2818 (cmp<mode>_fpr): Likewise.
2819 (smax<mode>3): Likewise.
2820 (smin<mode>3): Likewise.
2821 (smax<mode>3_vsx): Likewise.
2822 (smin<mode>3_vsx): Likewise.
2823 (negsf2): Delete SF operations that are merged with DF.
2824 (abssf2): Likewise.
2825 (addsf3): Likewise.
2826 (subsf3): Likewise.
2827 (mulsf3): Likewise.
2828 (divsf3): Likewise.
2829 (fres): Likewise.
2830 (fmasf4_fpr): Likewise.
2831 (fmssf4_fpr): Likewise.
2832 (nfmasf4_fpr): Likewise.
2833 (nfmssf4_fpr): Likewise.
2834 (sqrtsf2): Likewise.
2835 (rsqrtsf_internal1): Likewise.
2836 (smaxsf3): Likewise.
2837 (sminsf3): Likewise.
2838 (cmpsf_internal1): Likewise.
2839 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2840 (negdf2): Delete DF operations that are merged with SF.
2841 (absdf2): Likewise.
2842 (nabsdf2): Likewise.
2843 (adddf3): Likewise.
2844 (subdf3): Likewise.
2845 (muldf3): Likewise.
2846 (divdf3): Likewise.
2847 (fred): Likewise.
2848 (rsqrtdf_internal1): Likewise.
2849 (fmadf4_fpr): Likewise.
2850 (fmsdf4_fpr): Likewise.
2851 (nfmadf4_fpr): Likewise.
2852 (nfmsdf4_fpr): Likewise.
2853 (sqrtdf2): Likewise.
2854 (smaxdf3): Likewise.
2855 (smindf3): Likewise.
2856 (cmpdf_internal1): Likewise.
2857 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2858 (btrunc<mode>2): Delete separate expander, and combine with the
2859 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2860 (btrunc<mode>2_fpr): Likewise.
2861 (ceil<mode>2): Likewise.
2862 (ceil<mode>2_fpr): Likewise.
2863 (floor<mode>2): Likewise.
2864 (floor<mode>2_fpr): Likewise.
2865 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2866 Add support for using the upper registers with VSX and
2867 power8-vector. Move insns to be closer to the define_expands. On
2868 VSX systems, prefer the traditional form of FMA over the VSX
2869 version, since the traditional form allows the target not to
2870 overlap with the inputs.
2871 (fms<mode>4_fpr): Likewise.
2872 (nfma<mode>4_fpr): Likewise.
2873 (nfms<mode>4_fpr): Likewise.
2874
2875 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2876 Richard Earnshaw <richard.earnshaw@arm.com>
2877
2878 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2879 (struct mult_cost_table): Likewise.
2880 (struct mem_cost_table): Likewise.
2881 (struct fp_cost_table): Likewise.
2882 (struct vector_cost_table): Likewise.
2883 (cpu_cost_table): Likewise.
2884 * config/arm/arm.opt (mold-rts-costs): New option.
2885 (mnew-generic-costs): Likewise.
2886 * config/arm/arm.c (generic_extra_costs): New table.
2887 (cortexa15_extra_costs): Likewise.
2888 (arm_slowmul_tune): Use NULL as new costs.
2889 (arm_fastmul_tune): Likewise.
2890 (arm_strongarm_tune): Likewise.
2891 (arm_xscale_tune): Likewise.
2892 (arm_9e_tune): Likewise.
2893 (arm_v6t2_tune): Likewise.
2894 (arm_cortex_a5_tune): Likewise.
2895 (arm_cortex_a9_tune): Likewise.
2896 (arm_v6m_tune): Likewise.
2897 (arm_fa726te_tune): Likewise.
2898 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2899 (arm_cortex_tune): Use generict_extra_costs.
2900 (shifter_op_p): New function.
2901 (arm_unspec_cost): Likewise.
2902 (LIBCALL_COST): Define.
2903 (arm_new_rtx_costs): New function.
2904 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2905 table is available. Use old costs otherwise unless mnew-generic-costs
2906 is specified.
2907 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2908 (cpu_cost_table): Declare.
2909
2910 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2911
2912 PR target/58460
2913 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2914 (*subs_mul_imm_<mode>)
2915 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2916 (*sub_<shift>_<mode>)
2917 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2918 Remove k constraint.
2919
2920 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2921
2922 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2923 code.
2924 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2925
2926 2013-10-02 Teresa Johnson <tejohnson@google.com>
2927
2928 * predict.c (probably_never_executed): New function.
2929 (probably_never_executed_bb_p): Invoke probably_never_executed.
2930 (probably_never_executed_edge_p): Ditto.
2931 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2932 Treat profile insanities conservatively.
2933
2934 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2935
2936 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2937
2938 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2939
2940 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2941 use it. Use smaller increase for scratch. Don't increase reject
2942 for early clobber scratch.
2943 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2944 setting eliminated regs except setting fp from hfp.
2945 (lra_eliminate): Check lra_insn_recog_data on NULL.
2946
2947 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2948
2949 PR target/58587
2950 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2951 setting -mvsx-timode by default until the underlying problem is fixed.
2952 (RS6000_CPU, power7 defaults): Likewise.
2953
2954 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2955
2956 * config/x-linux (host-linux.o): Remove header dependencies.
2957 Use $(COMPILE) and $(POSTCOMPILE).
2958 * config/t-linux-android (linux-android.o): Ditto.
2959
2960 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2961
2962 * Makefile.in (expmed.o-warn): Remove.
2963
2964 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2965
2966 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2967 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2968
2969 2013-10-02 Teresa Johnson <tejohnson@google.com>
2970
2971 * dojump.c (do_jump_1): Divide probability between
2972 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2973
2974 2013-10-02 Tom Tromey <tromey@redhat.com>
2975
2976 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2977
2978 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2979
2980 * tree-flow.h: Remove some prototypes.
2981 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2982 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2983 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2984 mark_virtual_phi_result_for_renaming): Relocate here.
2985 * tree-into-ssa.h: Add prototypes.
2986 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2987 single_pred_before_succ_order.
2988 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2989 (nonfreeing_call_p) Move to gimple.c.
2990 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2991 tree-ssa-phiopt.c.
2992 * basic-block.h (single_pred_before_succ_order): Add prototype.
2993 * gimple.c (nonfreeing_call_p): Relocate here.
2994 * gimple.h: Add prototype.
2995 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2996 * tree-ssa-dom.h: New file. Relocate prototypes here.
2997 * tree-ssa.h: Include tree-ssa-dom.h.
2998
2999 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3000
3001 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3002 Use $(COMPILE) and $(POSTCOMPILE).
3003
3004 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3005
3006 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3007
3008 * tree-flow.h: Remove some prototypes.
3009 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3010 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3011 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3012 propagate_tree_value*): Move from here to...
3013 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3014 propagate_tree_value*): Relocate here.
3015 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3016 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3017 * gimple-fold.c: Remove gimple-fold.h from include list.
3018 * tree-vrp.c: Remove gimple-fold.h from include list.
3019 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3020 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3021 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3022 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3023 * sese.c: Add tree-ssa-propagate.h to include list.
3024
3025 2013-10-02 Richard Biener <rguenther@suse.de>
3026
3027 * tree-loop-distribution.c: Include tree-vectorizer.h for
3028 find_loop_location.
3029 (enum partition_kind): Remove PKIND_REDUCTION.
3030 (struct partition_s): Remove has_writes member, add reduction_p member.
3031 (partition_alloc): Adjust.
3032 (partition_builtin_p): Likewise.
3033 (partition_has_writes): Remove.
3034 (partition_reduction_p): New function.
3035 (partition_merge_into): Likewise.
3036 (generate_code_for_partition): Commonize builtin partition
3037 handling tail.
3038 (rdg_cannot_recompute_vertex_p): Remove.
3039 (already_processed_vertex_p): Likewise.
3040 (rdg_flag_vertex): Do not set has_writes.
3041 (classify_partition): Adjust.
3042 (rdg_build_partitions): Do not set has_writes, treat all
3043 partitions as useful.
3044 (distribute_loop): Record number of library calls generated. Adjust.
3045 (tree_loop_distribution): Report number of loops and library
3046 calls generated as opt-info.
3047
3048 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3049
3050 * tree-flow.h: Include new .h files. Move prototypes.
3051 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3052 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3053 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3054 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3055 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3056 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3057 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3058 * tree.h (get_ref_base_and_extent): Move prototype out.
3059 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3060 tree-dfa.h.
3061 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3062 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3063 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3064 * tree-scalar-evolution.c: Include tree.h.
3065 * sese.c: Include tree.h.
3066 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3067 * dwarf2out.c: Include tree-dfa.h.
3068 * tree-chrec.c: Include tree.h.
3069 * tree-data-ref.c: Include tree.h.
3070
3071 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3072
3073 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3074 Fix whitespace.
3075
3076 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3077
3078 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3079 Use $(COMPILE) and $(POSTCOMPILE).
3080 (sol2-cxx.o): Likewise.
3081 (sol2-stubs.o): Likewise.
3082 (sol2.o): Likewise.
3083 * config/x-solaris (host-solaris.o): Likewise.
3084
3085 * config/sparc/t-sparc (sparc.o): Remove.
3086 (sparc-c.o): Remove header dependencies.
3087 Use $(COMPILE) and $(POSTCOMPILE).
3088 * config/sparc/x-sparc: Likewise.
3089
3090 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3091
3092 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3093 * config/arc/arc700.md: Likewise.
3094 * config/arc/arc-modes.def: Likewise.
3095 * config/arc/arc-simd.h: Likewise.
3096 * config/arc/t-arc-uClibc: Likewise.
3097 * config/arc/t-arc-newlib: Likewise.
3098
3099 2013-10-02 Renlin Li <renlin.li@arm.com>
3100
3101 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3102 plus_constant.
3103 (aarch64_expand_epilogue): Likewise.
3104
3105 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3106 Yufeng Zhang <yufeng.zhang@arm.com>
3107
3108 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3109 declaration.
3110 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3111 'base_in' represent a conversion and legal_cast_p_1 holds; set
3112 'base_in' with the returned value from get_unwidened.
3113
3114 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3115
3116 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3117 plus_constant is not used.
3118
3119 2013-10-01 Wei Mi <wmi@google.com>
3120
3121 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3122 * config/i386/i386.md: Add define_peephole2 to
3123 break partial reg stall for cvtss2sd/cvtsd2ss.
3124
3125 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3126
3127 * config/arc/arc.c (pass_arc_ifcvt::clone):
3128 Update for ctxt_ -> m_ctxt change.
3129
3130 2013-10-01 Jeff Law <law@redhat.com>
3131
3132 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3133 outgoing_edge and intermediate_edge fields. Instead store the path.
3134 (redirection_data::hash): Hash on the last edge's destination index.
3135 (redirection_data::equal): Check the entire thread path.
3136 (lookup_redirectio_data): Corresponding changes.
3137 (create_edge_and_update_destination_phis): Likewise.
3138 (thread_single_edge): Likewise.
3139
3140 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3141 Diego Novillo <dnovillo@google.com>
3142
3143 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3144 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3145 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3146 versions of these patterns.
3147
3148 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3149 (__builtin_arc_aligned): Likewise.
3150
3151 * config/arc/arc.md: Expand adc_0 comment stating the intended
3152 purpose and why it isn't ready.
3153 Replace commented out call_value_via_label_mixed with a
3154 plain comment about bl_s.
3155
3156 * config/arc/arc.c (stdio.h): Don't include directly.
3157 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3158 Remove if (1) condition.
3159 (arc_encode_section_info): Fix comment.
3160
3161 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3162
3163 * config/arc/arc.c (arc_conditional_register_usage):
3164 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3165 Also set reg_alloc_order for DMA config regs.
3166
3167 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3168 Jeremy Bennett <jeremy.bennett@embecosm.com>
3169
3170 * doc/install.texi (--with-cpu): Mention ARC.
3171 (arc-*-elf32): New paragraph.
3172 (arc-linux-uclibc): Likewise.
3173 * doc/md.texi (Machine Constraints): Add ARC part.
3174 * doc/invoke.texi: (menu): Add ARC Options.
3175 (Machine Dependent Options) <ARC Options>: Add synopsis.
3176 (node ARC Options): Add.
3177 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3178 (ARC Built-in Functions): New section defining
3179 generic ARC built-in functions.
3180 (ARC SIMD Built-in Functions): New section defining SIMD specific
3181 built-in functions.
3182 (Declaring Attributes of Functions): Extended
3183 description of short_call and long_call attributes for ARC and
3184 added index entries.
3185
3186 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3187 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3188 Joern Rennecke <joern.rennecke@embecosm.com>
3189 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3190 Brendan Kehoe <brendan@zen.org>
3191 Michael Eager <eager@eagercon.com>
3192 Simon Cook <simon.cook@embecosm.com>
3193 Jeremy Bennett <jeremy.bennett@embecosm.com>
3194
3195 * config/arc, common/config/arc: New directories.
3196
3197 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3198 Brendan Kehoe <brendan@zen.org>
3199 Simon Cook <simon.cook@embecosm.com>
3200
3201 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3202
3203 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3204
3205 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3206 * tree-ssa-coalesce.h: New. Move prototype to here.
3207 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3208 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3209 (gimple_can_coalesce_p): Move to...
3210 * gimple.c (gimple_can_coalesce_p): Here.
3211
3212 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3213
3214 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3215 (dump_decl_set): Move to gimple.c.
3216 * gimple.h: Don't include tree-ssa-operands.h.
3217 (dump_decl_set): Add prototype.
3218 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3219 Move to gimple-ssa.h.
3220 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3221 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3222 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3223 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3224 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3225 than PHI_ARG_DEF.
3226 (dump_decl_set): Relocate here.
3227 * gimple-ssa.h: New file.
3228 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3229 Relocate from gimple.h.
3230 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3231 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3232 swap_tree_operands and remove non-ssa path.
3233 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3234 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3235 swap_ssa_operands.
3236 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3237 vect_is_simple_reduction_1): Use swap_ssa_operands.
3238 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3239 (enum need_phi_state): Move to tree-into-ssa.c.
3240 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3241 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3242 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3243 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3244 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3245 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3246 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3247 num_imm_uses): Move to ssa-iterators.h.
3248 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3249 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3250 tree-phinodes.h.
3251 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3252 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3253 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3254 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3255 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3256 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3257 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3258 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3259 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3260 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3261 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3262 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3263 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3264 (set_phi_nodes): Move to tree-phinodes.h.
3265 * tree-ssa-operands.h (enum ssa_op_iter_type,
3266 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3267 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3268 (dump_decl_set): Remove prototype.
3269 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3270 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3271 (set_phi_nodes): Relocate from tree-flow-inline.h.
3272 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3273 tree-flow-inline.h
3274 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3275 include list. Temporarily add gimple.h to include list.
3276 * ssa-iterators.h: New file.
3277 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3278 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3279 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3280 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3281 Relocate from tree-ssa-operands.h.
3282 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3283 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3284 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3285 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3286 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3287 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3288 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3289 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3290 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3291 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3292 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3293 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3294 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3295 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3296 Relocate from tree-flow-inline.h.
3297 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3298
3299 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3300
3301 * aarch64-simd.md
3302 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3303 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3304 Insert '\t' to output template.
3305 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3306 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3307 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3308 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3309
3310 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3311
3312 * doc/install.texi (Host/target specific installation notes for GCC):
3313 Put @anchor before @heading.
3314 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3315 Use @email for email addresses.
3316
3317 2013-10-01 Jeff Law <law@redhat.com>
3318
3319 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3320 a vec. Only delete the path if we create one without successfully
3321 registering a jump thread.
3322 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3323 as a pointer.
3324 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3325 (paths): New vector of jump threading paths.
3326 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3327 (THREAD_PATH): New accessor macro for the entire thread path.
3328 (lookup_redirection_data): Get intermediate and final outgoing edge
3329 from the thread path.
3330 (create_edge_and_update_destination_phis): Copy the threading path.
3331 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3332 jump threading path.
3333 (ssa_redirect_edges): Get edges and block types from the jump threading
3334 path. Free the path vector.
3335 (thread_block): Get edges from the jump threading path. Look at the
3336 entire path to see if we thread to a loop exit. If we cancel a jump
3337 thread request, then free the path vector.
3338 (thread_single_edge): Get edges and block types from the jump threading
3339 path. Free the path vector.
3340 (thread_through_loop_header): Get edges and block types from the jump
3341 threading path. Free the path vector.
3342 (mark_threaded_blocks): Iterate over the vector of paths and store
3343 the path on the appropriate edge. Get edges and block types from the
3344 jump threading path.
3345 (mark_threaded_blocks): Get edges and block types from the jump
3346 threading path. Free the path vector.
3347 (thread_through_all_blocks): Use the vector of paths rather than
3348 a vector of 3-edge sets.
3349 (register_jump_thread): Accept pointer to a path vector rather
3350 than the path vector itself. Store the path vector for later use.
3351 Simplify.
3352
3353 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3354 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3355
3356 PR target/58574
3357 * config/s390/s390.c (s390_split_branches): Modify check for table
3358 jump insns.
3359 (s390_chunkify_start): Rearrange table jump insn check in order to
3360 deal with compare and branch insns correctly.
3361
3362 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3363
3364 PR target/58578
3365 Revert
3366 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3367 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3368 define_insn_and_split.
3369 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3370 (shiftsi3_compare): New pattern.
3371 (rrx): New pattern.
3372 * config/arm/unspecs.md (UNSPEC_RRX): New.
3373
3374 2013-10-01 Alan Modra <amodra@gmail.com>
3375
3376 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3377 casing inout operands.
3378
3379 2013-10-01 Richard Biener <rguenther@suse.de>
3380
3381 PR tree-optimization/58553
3382 * tree-loop-distribution.c (struct partition_s): Add niter member.
3383 (classify_partition): Populate niter member for the partition
3384 and properly identify whether the relevant store happens before
3385 or after the loop exit.
3386 (generate_memset_builtin): Use niter member from the partition.
3387 (generate_memcpy_builtin): Likewise.
3388
3389 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3390
3391 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3392 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3393
3394 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3395
3396 * basic-block.h (edge_list): Prefix member names with "m_".
3397 * context.h (context): Likewise.
3398 * domwalk.h (dom_walker): Likewise.
3399 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3400 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3401 * hash-table.h (hash_table): Likewise.
3402 * machmode.h (bit_field_mode_iterator): Likewise.
3403 * pass_manager.h (pass_list): Likewise.
3404 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3405 * tree-pass.h (pass_data): Likewise.
3406 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3407 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3408 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3409 * asan.c (pass_data_asan): Update accordingly.
3410 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3411 (control_dependences::control_dependences): Likewise.
3412 (control_dependences::~control_dependences): Likewise.
3413 (control_dependences::~control_dependences): Likewise.
3414 (control_dependences::get_edges_dependent_on): Likewise.
3415 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3416 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3417 * context.c (gcc::context::context): Likewise.
3418 * cprop.c (pass_rtl_cprop::clone): Likewise.
3419 * domwalk.c (dom_walker::walk): Likewise.
3420 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3421 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3422 * mode-switching.c (pass_mode_switching::clone): Likewise.
3423 * passes.c (opt_pass::opt_pass): Likewise.
3424 (pass_manager::pass_manager): Likewise.
3425 * predict.c (pass_strip_predict_hints::clone): Likewise.
3426 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3427 (pass_split_all_insns::clone): Likewise.
3428 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3429 Likewise.
3430 (bit_field_mode_iterator::next_mode): Likewise.
3431 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3432 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3433 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3434 * tree-complex.c (pass_lower_complex::clone): Likewise.
3435 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3436 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3437 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3438 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3439 (pass_fold_builtins::clone): Likewise.
3440 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3441 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3442 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3443 (pass_cd_dce::clone): Likewise.
3444 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3445 (pass_phi_only_cprop::clone): Likewise.
3446 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3447 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3448 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3449 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3450 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3451 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3452 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3453 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3454 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3455 * tree-vrp.c (pass_vrp::clone): Likewise.
3456 * tsan.c (pass_tsan::clone): Likewise.
3457
3458 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3459
3460 PR middle-end/58564
3461 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3462 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3463
3464 PR middle-end/58564
3465 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3466 optimization, punt if sign_bit_p looked through any zero extension.
3467
3468 2013-09-30 Teresa Johnson <tejohnson@google.com>
3469
3470 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3471 Update redirected out edge count in joiner case.
3472 (ssa_redirect_edges): Common the joiner and non-joiner cases
3473 so that joiner case gets profile updates.
3474
3475 2013-09-30 Richard Biener <rguenther@suse.de>
3476
3477 PR tree-optimization/58554
3478 * tree-loop-distribution.c (classify_partition): Require
3479 unconditionally executed stores for memcpy and memset recognition.
3480 (tree_loop_distribution): Calculate dominance info.
3481
3482 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3483
3484 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3485 (NO_PROFILE_COUNTERS): Likewise.
3486 (PROFILE_HOOK): Likewise.
3487 (FUNCTION_PROFILER): Likewise.
3488 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3489
3490 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3491
3492 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3493 calls and defines in TARGET_MACHO conditional.
3494 (load_macho_picbase_di): Likewise.
3495 (reload_macho_picbase): Likewise.
3496 (reload_macho_picbase_si): Likewise.
3497 (reload_macho_picbase_di): Likewise.
3498 (nonlocal_goto_receiver): Likewise.
3499
3500 2013-09-30 Nick Clifton <nickc@redhat.com>
3501
3502 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3503 that use the MSP430X ISA.
3504 (msp430_option_override): Scan -mmcu command line option for any
3505 MCU name that supports the MSP430X ISA.
3506 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3507 -mmcu options which enable the MSP430X ISA.
3508
3509 2013-09-30 Richard Biener <rguenther@suse.de>
3510
3511 PR middle-end/58532
3512 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3513 before looking for setjmp-like calls.
3514
3515 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3516
3517 PR target/10901
3518 * config/darwin-protos.h (machopic_get_function_picbase): New.
3519 * config/darwin.c (machopic_get_function_picbase): New.
3520 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3521 label for a new func. (load_macho_picbase_di): Likewise.
3522 (reload_macho_picbase): New expand.
3523 (reload_macho_picbase_si): New insn.
3524 (reload_macho_picbase_di): New insn.
3525 (nonlocal_goto_receiver): New define and split.
3526 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3527 (unspecv enum): Add UNSPECV_NLGR.
3528
3529 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3530
3531 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3532 that altivec registers are correctly sized on Darwin.
3533
3534 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3535
3536 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3537 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3538 * config/x-darwin (host-darwin.o): Likewise.
3539 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3540 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3541 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3542
3543 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3544
3545 * doc/invoke.texi: Fix usage of @tie{} command.
3546
3547 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3548
3549 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3550
3551 2013-09-28 Jan Hubicka <jh@suse.cz>
3552
3553 * config/i386/x86-tune.def: Add documentation for each of the options;
3554 add whitespace.
3555
3556 2013-09-28 Jan Hubicka <jh@suse.cz>
3557
3558 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3559 generic.
3560 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3561 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3562 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3563 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3564 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3565
3566 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3567
3568 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3569 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3570 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3571 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3572 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3573 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3574 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3575 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3576 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3577 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3578 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3579 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3580 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3581 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3582 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3583 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3584 gimple.h, godump.c, graphite-clast-to-gimple.c,
3585 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3586 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3587 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3588 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3589 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3590 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3591 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3592 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3593 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3594 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3595 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3596 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3597 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3598 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3599 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3600 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3601 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3602 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3603 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3604 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3605 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3606 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3607 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3608 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3609 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3610 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3611 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3612 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3613 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3614 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3615 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3616 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3617 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3618 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3619 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3620 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3621 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3622 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3623 whitespace before "(".
3624
3625 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3626
3627 * expr.h (extract_bit_field): Remove packedp parameter.
3628 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3629 from forward declaration.
3630 (store_split_bit_field): Remove packedp arg from calls to
3631 extract_fixed_bit_field.
3632 (extract_bit_field_1): Remove packedp parameter and packedp
3633 argument from recursive calls and calls to extract_fixed_bit_field.
3634 (extract_bit_field): Remove packedp parameter and corresponding
3635 arg to extract_bit_field_1.
3636 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3637 to issue warnings.
3638 (extract_split_bit_field): Remove packedp arg from call to
3639 extract_fixed_bit_field.
3640 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3641 (copy_blkmode_from_reg): Likewise.
3642 (copy_blkmode_to_reg): Likewise.
3643 (read_complex_part): Likewise.
3644 (store_field): Likewise.
3645 (expand_expr_real_1): Likewise.
3646 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3647 to extract_bit_field.
3648 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3649 call to extract_bit_field.
3650 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3651 call to extract_bit_field.
3652 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3653 and special packedp behavior from -fstrict-volatile-bitfields
3654 documentation.
3655
3656 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3657
3658 * lra-eliminations.c (init_elim_table): Guard value_p.
3659
3660 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3661
3662 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3663 DFmode, DImode, and SFmode in the upper VSX registers based on the
3664 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3665 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3666 of the VSX register handling.
3667
3668 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3669 (f32_sv): Likewise.
3670 (zero_extendsidi2_lfiwzx): Add support for loading into the
3671 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3672 only do VSX memory options on Altivec registers.
3673 (extendsidi2_lfiwax): Likewise.
3674 (extendsfdf2_fpr): Likewise.
3675 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3676 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3677 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3678 (movdi_internal64): Likewise.
3679
3680 2013-09-27 Xinliang David Li <davidxl@google.com>
3681
3682 * opts.c (finish_options): Adjust parameters
3683 according to vect cost model.
3684 (common_handle_option): Set dynamic vect cost
3685 model for FDO.
3686 targhooks.c (default_add_stmt_cost): Compute stmt cost
3687 unconditionally.
3688 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3689 Use helper function.
3690 * tree-vectorizer.h (unlimited_cost_model): New function.
3691 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3692 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3693 function.
3694 (vect_enhance_data_refs_alignment): Ditto.
3695 * flag-types.h: New enum.
3696 * common/config/i386/i386-common.c (ix86_option_init_struct):
3697 No need to initialize vect_cost_model flag.
3698 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3699 unconditionally.
3700
3701 2013-09-27 Diego Novillo <dnovillo@google.com>
3702
3703 * gimple.h (enum ssa_mode): Remove.
3704
3705 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3706
3707 * cfgloop.h (number_of_loops): Fix typo in check for null.
3708
3709 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3710
3711 PR middle-end/58551
3712 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3713 are children of outermost saved_cfun's loop, and set it up to
3714 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3715 if loop != loop0 and SESE region contains bbs that belong to loop0.
3716
3717 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3718
3719 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3720 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3721 (set_address_segment, set_address_base, set_address_index)
3722 (set_address_disp): Accept the argument unconditionally.
3723 (baseness): Remove must_be_base_p and must_be_index_p checks.
3724 (decompose_normal_address): Classify as much as possible in the
3725 main loop.
3726
3727 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3728
3729 * cse.c (count_reg_usage): Handle INT_LIST.
3730 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3731 * reginfo.c (reg_scan_mark_refs): Likewise.
3732 * reload1.c (eliminate_regs_1): Likewise.
3733
3734 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3735
3736 PR middle-end/58547
3737 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3738 signed.
3739
3740 2013-09-27 Richard Biener <rguenther@suse.de>
3741
3742 PR tree-optimization/58459
3743 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3744 restriction not propagating into loops.
3745
3746 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3747
3748 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3749 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3750 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3751
3752 2013-09-26 Richard Biener <rguenther@suse.de>
3753
3754 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3755
3756 2013-09-26 Richard Biener <rguenther@suse.de>
3757
3758 * alias.h (component_uses_parent_alias_set): Rename to ...
3759 (component_uses_parent_alias_set_from): ... this.
3760 * alias.c (component_uses_parent_alias_set): Rename to ...
3761 (component_uses_parent_alias_set_from): ... this and return
3762 the desired parent.
3763 (reference_alias_ptr_type_1): Use the result from
3764 component_uses_parent_alias_set_from instead of stripping
3765 components one at a time.
3766 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3767
3768 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3769
3770 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3771 Move prototypes to...
3772 * tree-ssa-ter.h: New File. Move prototypes here.
3773 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3774 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3775 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3776 from is_replaceable_p.
3777 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3778 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3779 (process_replaceable): Use ter_is_replaceable_p.
3780 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3781 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3782 newly refactored ssa_is_replaceable_p.
3783 * cfgexpand.c: Include tree-outof-ssa.h.
3784 * ssaexpand.h: Delete.
3785
3786 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3787
3788 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3789 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3790 tree-ssa.c
3791 (create_gimple_tmp): Delete.
3792 (get_expr_type, build_assign, build_type_cast): Move to...
3793 * gimple-builder.c: New File.
3794 (get_expr_type): Relocate from gimple.c.
3795 (build_assign, build_type_cast): Change to only create ssanames.
3796 * gimple.h: Move prototypes to...
3797 * gimple-builder.h: New File. Here.
3798 * tree-ssa.h: And here.
3799 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3800 count_uses_and_derefs): Relocate from gimple.c.
3801 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3802 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3803 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3804 gimple_replace_ssa_lhs.
3805 * asan.c: Include gimple-builder.h.
3806 * Makefile.in: Add gimple-builder.o.
3807
3808 2013-09-26 Richard Biener <rguenther@suse.de>
3809
3810 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3811 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3812 (loe_visit_block): Use gcc_checking_assert.
3813 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3814 gimple_assign_ssa_name_copy_p.
3815 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3816 change.
3817
3818 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3819
3820 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3821 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3822
3823 2013-09-26 Richard Biener <rguenther@suse.de>
3824
3825 PR tree-optimization/58539
3826 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3827 the fact that debug statements are not taking part in loop-closed
3828 SSA construction.
3829
3830 2013-09-26 Nick Clifton <nickc@redhat.com>
3831
3832 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3833 time warning message.
3834 (msp430_print_operand_raw): Delete unused letter parameter.
3835 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3836 (msp430_print_operand_address): New function.
3837 (msp430_print_operand): Move address printing code from here to
3838 new function.
3839 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3840 assembler.
3841 (zero_extendpsisi2): Likewise.
3842 (extendpsisi2): New pattern.
3843 (andneghi3): New pattern.
3844
3845 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3846
3847 * config/aarch64/aarch64.opt (mlra): New option.
3848 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3849 (TARGET_LRA_P): Define.
3850
3851 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3852
3853 * expr.c (expand_assignment): Remove obsolete comment.
3854
3855 2013-09-25 Jeff Law <law@redhat.com>
3856
3857 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3858 tree-ssa-threadupdate.h.
3859 (register_jump_thread): Similarly.
3860 * tree-ssa-threadupdate.h: New header file.
3861 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3862 * tree-vrp.c: Likewise.
3863 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3864 (thread_around_empty_blocks): Change type of path vector argument to
3865 an edge,type pair from just an edge. Initialize both elements when
3866 appending to a jump threading path. Tweak references to elements
3867 appropriately.
3868 (thread_across_edge): Similarly. Release memory for the elements
3869 as needed.
3870 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3871 (dump_jump_thread_path): New function broken out from
3872 register_jump_thread.
3873 (register_jump_thread): Use dump_jump_thread_path. Change type of
3874 path vector entries. Search the path for NULL edges and dump
3875 the path if one is found. Tweak the conversion of path to 3-edge
3876 form to use the block copy type information embedded in the path.
3877
3878 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3879
3880 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3881
3882 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3883 Vladimir Makarov <vmakarov@redhat.com>
3884
3885 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3886 from the least significant bit.
3887 (strip_address_mutations): Add bitfield operations handling.
3888 (must_be_index_p): Add shifting and rotate operations handling.
3889 (set_address_base): Use must_be_base_p predicate.
3890 (set_address_index): Use must_be_index_p predicate.
3891
3892 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3893 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3894 Sergey Lega <sergey.s.lega@intel.com>
3895 Anna Tikhonova <anna.tikhonova@intel.com>
3896 Ilya Tocar <ilya.tocar@intel.com>
3897 Andrey Turetskiy <andrey.turetskiy@intel.com>
3898 Ilya Verbin <ilya.verbin@intel.com>
3899 Kirill Yukhin <kirill.yukhin@intel.com>
3900 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3901
3902 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3903 Use new names.
3904 (ix86_expand_vector_move_misalign): Support new unaligned load and
3905 stores and use new names.
3906 (CODE_FOR_sse2_storedqu): Rename to ...
3907 (CODE_FOR_sse2_storedquv16qi): ... this.
3908 (CODE_FOR_sse2_loaddqu): Rename to ...
3909 (CODE_FOR_sse2_loaddquv16qi): ... this.
3910 (CODE_FOR_avx_loaddqu256): Rename to ...
3911 (CODE_FOR_avx_loaddquv32qi): ... this.
3912 (CODE_FOR_avx_storedqu256): Rename to ...
3913 (CODE_FOR_avx_storedquv32qi): ... this.
3914 * config/i386/i386.md (fpint_logic): New.
3915 * config/i386/sse.md (VMOVE): Extend for AVX512.
3916 (VF): Ditto.
3917 (VF_128_256): New.
3918 (VF_512): Ditto.
3919 (VI_UNALIGNED_LOADSTORE): Ditto.
3920 (sse2_avx_avx512f): Ditto.
3921 (sse2_avx2): Extend for AVX512.
3922 (sse4_1_avx2): Ditto.
3923 (avx2_avx512f): New.
3924 (sse): Extend for AVX512.
3925 (sse2): Ditto.
3926 (sse4_1): Ditto.
3927 (avxsizesuffix): Ditto.
3928 (sseintvecmode): Ditto.
3929 (ssePSmode): Ditto.
3930 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3931 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3932 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3933 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3934 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3935 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3936 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3937 (STORENT_MODE): Extend for AVX512.
3938 (*absneg<mode>2): Replace constraint "x" with "v".
3939 (*mul<mode>3): Ditto.
3940 (*ieee_smin<mode>3): Ditto.
3941 (*ieee_smax<mode>3): Ditto.
3942 (avx_cmp<mode>3): Replace VF with VF_128_256.
3943 (*<sse>_maskcmp<mode>3_comm): Ditto.
3944 (<sse>_maskcmp<mode>3): Ditto.
3945 (<sse>_andnot<mode>3): Extend for AVX512.
3946 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3947 (<code><mode>3, fpint_logic): New.
3948 (*<code><mode>3): Extend for AVX512.
3949 (avx512flogicsuff): New.
3950 (avx512f_<logic><mode>): Ditto.
3951 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3952 VF_128_256.
3953 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3954 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3955 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3956 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3957 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3958 (xop_vpermil2<mode>3): Ditto.
3959 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3960 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3961 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3962 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3963
3964 2013-09-25 Tom Tromey <tromey@redhat.com>
3965
3966 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3967 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3968 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3969 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3970 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3971 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3972 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3973 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3974 (GRAPHITE_HTAB_H): Remove.
3975
3976 2013-09-25 Tom Tromey <tromey@redhat.com>
3977
3978 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3979
3980 2013-09-25 Tom Tromey <tromey@redhat.com>
3981
3982 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3983
3984 2013-09-25 Tom Tromey <tromey@redhat.com>
3985
3986 * config/i386/t-i386 (i386.o): Remove.
3987 (i386-c.o): Use COMPILE and POSTCOMPILE.
3988
3989 2013-09-25 Tom Tromey <tromey@redhat.com>
3990
3991 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3992
3993 2013-09-25 Tom Tromey <tromey@redhat.com>
3994
3995 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3996 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3997 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3998 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3999 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4000 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4001 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4002 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4003 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4004 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4005 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4006 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4007 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4008 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4009 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4010 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4011 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4012 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4013 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4014 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4015 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4016 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4017 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4018 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4019 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4020 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4021 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4022 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4023 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4024 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4025 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4026 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4027 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4028 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4029 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4030 (graphite-clast-to-gimple.o, graphite-dependences.o)
4031 (graphite-interchange.o, graphite-poly.o)
4032 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4033 (graphite-optimize-isl.o, tree-vect-loop.o)
4034 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4035 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4036 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4037 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4038 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4039 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4040 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4041 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4042 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4043 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4044 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4045 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4046 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4047 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4048 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4049 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4050 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4051 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4052 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4053 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4054 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4055 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4056 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4057 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4058 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4059 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4060 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4061 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4062 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4063 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4064 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4065 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4066 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4067 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4068 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4069 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4070 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4071 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4072 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4073 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4074 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4075 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4076 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4077 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4078 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4079 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4080 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4081 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4082 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4083 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4084 (gcov-dump.o): Remove.
4085 (default-c.o): Use COMPILE and POSTCOMPILE.
4086 (CFLAGS-gcc.o): New variable.
4087 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4088
4089 2013-09-25 Tom Tromey <tromey@redhat.com>
4090
4091 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4092 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4093 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4094 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4095 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4096 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4097 (c-family/array-notation-common.o, c-family/stub-objc.o)
4098 (c-family/c-ubsan.o): Remove.
4099
4100 2013-09-25 Tom Tromey <tromey@redhat.com>
4101
4102 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4103
4104 2013-09-25 Tom Tromey <tromey@redhat.com>
4105
4106 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4107 to add -DENABLE_SHARED_LIBGCC.
4108 (gcc.o): Don't use subshell.
4109
4110 2013-09-25 Tom Tromey <tromey@redhat.com>
4111
4112 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4113 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4114 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4115 * configure, config.in: Rebuild.
4116
4117 2013-09-25 Tom Tromey <tromey@redhat.com>
4118
4119 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4120 (COMPILE, POSTCOMPILE): New variables.
4121 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4122 (DEPFILES): New variable.
4123 Include ".Po" files.
4124 * configure.ac: Add checks for dependency checking.
4125 * configure, aclocal.m4: Regenerate.
4126
4127 2013-09-25 Tom Tromey <tromey@redhat.com>
4128
4129 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4130 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4131
4132 2013-09-25 Tom Tromey <tromey@redhat.com>
4133
4134 * Makefile.in (generated_files): Add options.h,
4135 target-hooks-def.h, insn-opinit.h,
4136 common/common-target-hooks-def.h, pass-instances.def,
4137 c-family/c-target-hooks-def.h.
4138
4139 2013-09-25 Jeff Law <law@redhat.com>
4140
4141 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4142 than foo[foo.length () - 1] to access last member in a vec.
4143 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4144
4145 2013-09-25 Richard Biener <rguenther@suse.de>
4146
4147 PR middle-end/58521
4148 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4149
4150 2013-09-25 Jan Hubicka <jh@suse.cz>
4151
4152 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4153 test.
4154
4155 2013-09-25 Marek Polacek <polacek@redhat.com>
4156
4157 PR sanitizer/58420
4158 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4159 when determining the type name.
4160
4161 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4162
4163 * config/sh/sh.md: Fix formatting.
4164
4165 2013-09-24 Xinliang David Li <davidxl@google.com>
4166
4167 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4168 max peel iterations parameter.
4169 * param.def: New parameter.
4170 * doc/invoke.texi: Document New parameter.
4171
4172 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4173
4174 * gimple-pretty-print.c: Various whitespace tweaks.
4175 * tree-core.h: Likewise.
4176 * tree-pretty-print.c: Likewise.
4177 * tree-ssa-alias.c: Likewise.
4178 * tree-ssa-copy.c: Likewise.
4179 * tree-ssanames.c: Likewise.
4180 * tree-ssanames.h: Likewise.
4181 * tree-vrp.c: Likewise.
4182
4183 2013-09-24 Alan Modra <amodra@gmail.com>
4184
4185 PR middle-end/57134
4186 PR middle-end/57586
4187 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4188 for output operands that disallow regs. Don't use EXPAND_WRITE on
4189 inout operands.
4190
4191 2013-09-24 Richard Biener <rguenther@suse.de>
4192
4193 PR middle-end/58513
4194 * tree.c (reference_alias_ptr_type): Move ...
4195 * alias.c (reference_alias_ptr_type): ... here and implement
4196 in terms of the new reference_alias_ptr_type_1.
4197 (ref_all_alias_ptr_type_p): New helper.
4198 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4199 use ref_all_alias_ptr_type_p.
4200 (get_deref_alias_set): Add flag_strict_aliasing check here.
4201 (reference_alias_ptr_type_1): New function, split out from ...
4202 (get_alias_set): ... here.
4203 (alias_ptr_types_compatible_p): New function.
4204 * alias.h (reference_alias_ptr_type): Declare.
4205 (alias_ptr_types_compatible_p): Likewise.
4206 * tree.h (reference_alias_ptr_type): Remove.
4207 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4208 to compare MEM_REF alias types.
4209
4210 2013-09-24 Richard Biener <rguenther@suse.de>
4211
4212 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4213
4214 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4215
4216 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4217 reload helper function arrays into a single array reg_addr.
4218 (reload_fpr_gpr): Likewise.
4219 (reload_gpr_vsx): Likewise.
4220 (reload_vsx_gpr): Likewise.
4221 (struct rs6000_reg_addr): Likewise.
4222 (reg_addr): Likewise.
4223 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4224 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4225 (rs6000_init_hard_regno_mode_ok): Likewise.
4226 (rs6000_secondary_reload_direct_move): Likewise.
4227 (rs6000_secondary_reload): Likewise.
4228
4229 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4230 constraints: wu, ww, and wy. Repurpose wv constraint added during
4231 power8 changes. Put wg constraint in alphabetical order.
4232
4233 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4234 for future work to add ISA 2.07 VSX single precision support.
4235 (-mvsx-scalar-double): Change default from -1 to 1, update
4236 documentation comment.
4237 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4238 (-mupper-regs-df): New debug switch to control whether DF values
4239 can go in the traditional Altivec registers.
4240 (-mupper-regs-sf): New debug switch to control whether SF values
4241 can go in the traditional Altivec registers.
4242
4243 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4244 and wy constraints.
4245 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4246 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4247 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4248 (rs6000_debug_legitimate_address_p): Print if we are running
4249 before, during, or after reload.
4250 (rs6000_secondary_reload): Add a comment.
4251 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4252
4253 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4254 constraints. Update documentation string.
4255 (wd constraint): Likewise.
4256 (wf constraint): Likewise.
4257 (wg constraint): Likewise.
4258 (wn constraint): Likewise.
4259 (ws constraint): Likewise.
4260 (wt constraint): Likewise.
4261 (wx constraint): Likewise.
4262 (wz constraint): Likewise.
4263 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4264 instructions.
4265 (ww constraint): Likewise.
4266 (wy constraint): Likewise.
4267 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4268 in the previous submissions.
4269 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4270
4271 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4272
4273 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4274 (REG_BR_PROB): Say that the probability is stored in an int_list.
4275 * reg-notes.def: Update commentary to mention INT_LIST.
4276 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4277 (INT_LIST): New rtx.
4278 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4279 * rtlanal.c (int_reg_note_p): New function.
4280 (alloc_reg_note): Assert that the note does not have an int argument.
4281 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4282 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4283 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4284 rather than an INSN_LIST. Handle INT_LIST.
4285 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4286 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4287 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4288 Manipulate them as ints rather than rtxes.
4289 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4290 * regmove.c (copy_src_to_dest): Likewise.
4291 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4292
4293 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4294 into the cases that need it.
4295 * config/arm/arm.c (arm_unwind_emit): Likewise.
4296
4297 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4298 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4299 * loop-doloop.c (add_test, doloop_modify): Likewise.
4300 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4301 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4302 * predict.c (combine_predictions_for_insn): Likewise.
4303 * print-rtl.c (print_rtx): Handle INT_LIST.
4304 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4305 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4306 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4307 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4308 (ix86_split_fp_branch, predict_jump): Likewise.
4309 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4310 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4311 * config/spu/spu.c (ea_load_store_inline): Likewise.
4312
4313 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4314 value of a REG_BR_PROB note.
4315 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4316 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4317 * emit-rtl.c (try_split): Likewise.
4318 * predict.c (br_prob_note_reliable_p): Likewise.
4319 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4320 * reorg.c (mostly_true_jump): Likewise.
4321 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4322 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4323 * config/i386/i386.c (ix86_print_operand): Likewise.
4324 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4325 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4326 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4327 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4328 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4329 * config/sparc/sparc.c (output_cbranch): Likewise.
4330 * config/spu/spu.c (get_branch_target): Likewise.
4331 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4332 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4333
4334 2013-09-23 Jan Hubicka <jh@suse.cz>
4335
4336 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4337 for ipa-devirt.
4338 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4339 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4340 of external calls
4341 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4342 (gimple_fold_call): Dump inconsistent devirtualizations; add
4343 sanity check for type based devirtualizations.
4344 * ipa-prop.c: Include ipa-utils.h
4345 (ipa_intraprocedural_devirtualization): Add sanity check.
4346 (try_make_edge_direct_virtual_call): Likewise.
4347
4348 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4349
4350 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4351 assignment statements.
4352
4353 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4354
4355 * gimple-pretty-print.c (dump_ssaname_info): New function.
4356 (dump_gimple_phi): Call it.
4357 (pp_gimple_stmt_1): Likewise.
4358 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4359 (range_info_def): Declare.
4360 * tree-pretty-print.c (pp_double_int): New function.
4361 (dump_generic_node): Call it.
4362 * tree-pretty-print.h (pp_double_int): Declare.
4363 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4364 * tree-ssanames.h (range_info_def): New structure.
4365 (value_range_type): Move definition here.
4366 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4367 Declare.
4368 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4369 initialization.
4370 (set_range_info): New function.
4371 (get_range_info): Likewise.
4372 (duplicate_ssa_name_range_info): Likewise.
4373 (duplicate_ssa_name_fn): Check pointer type and call
4374 duplicate_ssa_name_range_info.
4375 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4376 * tree-vrp.c (value_range_type): Remove definition, now in
4377 tree-ssanames.h.
4378 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4379 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4380 (SSA_NAME_RANGE_INFO): New macro.
4381
4382 2013-09-23 Richard Biener <rguenther@suse.de>
4383
4384 PR tree-optimization/58464
4385 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4386 (phi_trans_add): Change to add conditionally on being not present.
4387 (phi_translate_1): Remove recursion detection here.
4388 (phi_translate): Pre-seed the cache with NULL to catch
4389 recursion here in a more generic way.
4390 (bitmap_find_leader): Adjust comment.
4391 (get_representative_for): Dump value-numbers.
4392 (create_expression_by_pieces): Likewise.
4393 (insert_into_preds_of_block): Likewise.
4394
4395 2013-09-23 Christian Bruel <christian.bruel@st.com>
4396
4397 PR target/58475
4398 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4399 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4400
4401 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4402
4403 Revert r202780:
4404 2013-09-20 Renlin Li <renlin.li@arm.com>
4405
4406 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4407 plus_constant.
4408 (aarch64_expand_epilogue): Likewise.
4409 (aarch64_legitimize_reload_address): Likewise.
4410
4411 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4412
4413 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4414 NULL_TREE before pushing them onto the vector. Likewise for labels.
4415
4416 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4417
4418 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4419 comparison operators when -fno-trapping-math is in effect.
4420 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4421 comparison operators in TFmode and assert that unsupported operators
4422 cannot reach here.
4423 (ia64_print_operand): Likewise.
4424
4425 2013-09-21 Jan Hubicka <jh@suse.cz>
4426
4427 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4428 (sse_typeless_stores): Enable for core
4429 (sse_load0_by_pxor): Likewise.
4430 (four_jump_limit): Disable for core.
4431 (pad_returns): Likewise.
4432 (avoid_vector_decode): Likewise.
4433 (fuse_cmp_and_branch): Enable for cores.
4434 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4435
4436 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4437
4438 PR middle-end/56791
4439 * config/pa/pa.c (pa_option_override): Disable auto increment and
4440 decrement instructions until reload is completed.
4441
4442 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4443 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4444 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4445
4446 2013-09-20 DJ Delorie <dj@redhat.com>
4447 Nick Clifton <nickc@redhat.com>
4448
4449 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4450 (need_to_save): Save bank 0 on interrupts.
4451 (characterize_address): Strip far address wrappers.
4452 (rl78_as_legitimate_address): Likewise.
4453 (transcode_memory_rtx): Likewise.
4454 (rl78_peep_movhi_p): Disable this peephole after devirt.
4455 (rl78_propogate_register_origins): Forget all origins when a
4456 CLOBBER is seen.
4457 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4458 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4459 peephole2's.
4460 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4461 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4462 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4463 relocating.
4464 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4465
4466 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4467
4468 * config/pa/pa.md: In "scc" insn patterns, change output template to
4469 handle const0_rtx in reg_or_0_operand operands.
4470
4471 2013-09-20 Martin Husemann <martin@NetBSD.org>
4472
4473 PR target/56875
4474 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4475 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4476
4477 2013-09-20 Richard Biener <rguenther@suse.de>
4478
4479 PR middle-end/58484
4480 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4481 remembering SSA name version and block index.
4482 (new_scev_info_str): Adjust.
4483 (hash_scev_info): Likewise. Also hash the block index.
4484 (eq_scev_info): Adjust.
4485 (find_var_scev_info): Likewise.
4486 (struct instantiate_cache_entry): Remove.
4487 (struct instantiate_cache_type): Use a htab to map name, block
4488 to chrec.
4489 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4490 (get_instantiated_value_entry): Likewise.
4491 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4492 (instantiate_scev_name): Adjust.
4493
4494 2013-09-20 Jeff Law <law@redhat.com>
4495
4496 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4497
4498 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4499
4500 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4501 Call aarch64_simd_expand_args to update op[argc].
4502
4503 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4504
4505 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4506 plugin argument.
4507
4508 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4509
4510 * gengtype.c (file_rules): Added rule for *.cc files.
4511 (get_output_file_with_visibility): Give fatal message when no
4512 rules found.
4513
4514 2013-09-20 Renlin Li <renlin.li@arm.com>
4515
4516 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4517 (aarch64_expand_epilogue): Likewise.
4518 (aarch64_legitimize_reload_address): Likewise.
4519
4520 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4521
4522 PR middle-end/57748
4523 * expr.c (expand_assignment): Remove misalignp code path.
4524
4525 2013-09-20 Marek Polacek <polacek@redhat.com>
4526
4527 PR sanitizer/58413
4528 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4529 TYPE_PRECISION. Add asserts.
4530
4531 2013-09-20 Richard Biener <rguenther@suse.de>
4532
4533 PR tree-optimization/58453
4534 * tree-loop-distribution.c (distribute_loop): Apply the cost
4535 model for -ftree-loop-distribute-patterns, too.
4536
4537 2013-09-20 Richard Biener <rguenther@suse.de>
4538
4539 PR middle-end/58473
4540 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4541 make type comparison less strict.
4542
4543 2013-09-20 Alan Modra <amodra@gmail.com>
4544
4545 * configure: Regenerate.
4546 * aclocal.m4: Regenerate.
4547
4548 2013-09-20 Marek Polacek <polacek@redhat.com>
4549
4550 PR other/58467
4551 * doc/extend.texi: Document that attribute used is meant to be used
4552 on variables with static storage duration.
4553
4554 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4555
4556 PR tree-optimization/58472
4557 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4558 simd_lane_access set inv_p = false.
4559 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4560 the simduid magic VAR_DECL.
4561
4562 2013-09-19 Jan Hubicka <jh@suse.cz>
4563
4564 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4565
4566 2013-09-17 Jeff Law <law@redhat.com>
4567
4568 * tree-ssa-dom.c (record_temporary_equivalences): New function
4569 split out of dom_opt_dom_walker::after_dom_children.
4570 (dom_opt_dom_walker::thread_across_edge): Move common code
4571 in here from dom_opt_dom_walker::after_dom_children.
4572 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4573
4574 2013-09-19 Jan Hubicka <jh@suse.cz>
4575
4576 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4577 (TARGET_GENERIC): Use PROCESOR_GENERIC
4578 (enum processor_type): Unify generic32 and 64.
4579 * i386.md (cpu): Likewise.
4580 * x86-tune.def (use_leave): Enable for generic32.
4581 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4582 * athlon.md: Change generic64 to generic in all occurences.
4583 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4584 (ix86_target_macros_internal): Likewise.
4585 * driver-i386.c (host_detect_local_cpu): Likewise.
4586 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4587 to ..
4588 (generic_memcpy, generic_memset, generic_cost): This one.
4589 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4590 (m_GENERIC32, m_GENERIC64): Remove.
4591 (m_GENERIC): Turn into one flag.
4592 (processor_target): Unify generic tunnings.
4593 (ix86_option_override_internal): Replace generic32/64 by generic.
4594 (ix86_issue_rate): Likewise.
4595 (ix86_adjust_cost): Likewise.
4596
4597 2013-09-19 Jan Hubicka <jh@suse.cz>
4598
4599 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4600 of speculative flag.
4601
4602 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4603
4604 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4605 GOMP_sections_start, even if !exit_reachable.
4606
4607 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4608
4609 * lra-constraints.c (need_for_all_save_p): Use macro
4610 HARD_REGNO_CALL_PART_CLOBBERED.
4611 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4612 set up pseudo conflict hard regs.
4613
4614 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4615
4616 PR target/58452
4617 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4618 operands.
4619
4620 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4621
4622 PR rtl-optimization/58438
4623 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4624 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4625 from previous subpasses.
4626
4627 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4628
4629 * arm.c (arm_get_frame_offsets): Validate architecture supports
4630 LDRD/STRD before accepting the tuning preference.
4631 (arm_expand_prologue): Likewise.
4632 (arm_expand_epilogue): Likewise.
4633
4634 2013-09-18 Richard Biener <rguenther@suse.de>
4635
4636 PR tree-optimization/58417
4637 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4638 have chrecs with symbols defined in the loop as operands.
4639 (chrec_fold_multiply): Likewise.
4640 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4641 parameters before folding binary operations.
4642 (struct instantiate_cache_entry_hasher): Remove.
4643 (struct instantiate_cache_type): Use a pointer-map.
4644 (instantiate_cache_type::instantiate_cache_type): New function.
4645 (instantiate_cache_type::get): Likewise.
4646 (instantiate_cache_type::set): Likewise.
4647 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4648 (get_instantiated_value_entry): Likewise.
4649 (global_cache): New global.
4650 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4651 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4652 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4653 (instantiate_scev_name): Adjust.
4654 (instantiate_scev): Construct global instead of local cache.
4655 (resolve_mixers): Likewise.
4656
4657 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4658 Paolo Carlini <paolo.carlini@oracle.com>
4659
4660 PR c++/58458
4661 * doc/implement-cxx.texi: Fix references to the C++ standards.
4662
4663 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4664
4665 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4666 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4667 simd_lane_access drs, update also DR_ALIGNED_TO.
4668
4669 2013-09-18 Marek Polacek <polacek@redhat.com>
4670
4671 PR sanitizer/58411
4672 * doc/extend.texi: Document no_sanitize_undefined attribute.
4673 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4674 no_sanitize_undefined attribute.
4675
4676 2013-09-18 Nick Clifton <nickc@redhat.com>
4677
4678 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4679 (ASM_DECLARE_FUNCTION_NAME): Define.
4680
4681 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4682
4683 * compare-elim.c (find_comparison_dom_walker): New class
4684 (find_comparisons_in_bb): Rename to
4685 find_comparison_dom_walker::before_dom_children
4686 (find_comparisons): Adjust
4687 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4688 adjust.
4689 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4690 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4691 (init_walk_dominator_tree): Remove declaration.
4692 (fini_walk_dominator_tree): Remove declaration.
4693 * fwprop.c (single_def_use_dom_walker): New class
4694 (single_def_use_enter_block): Convert to
4695 single_def_use_dom_walker::before_dom_children.
4696 (single_def_use_leave_block): Convert to
4697 single_def_use_dom_walker::after_dom_children.
4698 (build_single_def_use_links): Adjust.
4699 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4700 class.
4701 (find_candidates_in_block): Convert to
4702 find_candidates_dom_walker::before_dom_children.
4703 (execute_strength_reduction): Adjust.
4704 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4705 (sese_dom_walker): New class.
4706 (sese_dom_walker::sese_dom_walker): New constructor.
4707 (sese_dom_walker::~sese_dom_walker): New destructor.
4708 (build_sese_conditions_before): Convert to
4709 sese_dom_walker::before_dom_children.
4710 (build_sese_conditions_after): Convert to
4711 sese_dom_walker::after_dom_children.
4712 (build_poly_scop): Adjust
4713 * tree-into-ssa.c (rewrite_dom_walker): New class
4714 (rewrite_enter_block): Convert to
4715 rewrite_dom_walker::before_dom_children.
4716 (rewrite_leave_block): Convert to
4717 rewrite_dom_walker::after_dom_children.
4718 (rewrite_update_dom_walker): New class.
4719 (rewrite_update_enter_block): Convert to
4720 rewrite_update_dom_walker::before_dom_children.
4721 (rewrite_update_leave_block): Convert to
4722 rewrite_update_dom_walker::after_dom_children.
4723 (rewrite_blocks, rewrite_into_ssa): Adjust.
4724 (mark_def_dom_walker): New class.
4725 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4726 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4727 (mark_def_sites_blocks): Convert to
4728 mark_def_dom_walker::before_dom_children.
4729 (mark_def_site_blocks): Remove.
4730 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4731 (tree_ssa_dominator_optimize): Adjust.
4732 (dom_thread_across_edge): Convert to method
4733 dom_opt_dom_walker::thread_across_edge.
4734 (dom_opt_enter_block): Convert to member function
4735 dom_opt_dom_walker::before_dom_children.
4736 (dom_opt_leave_block): Convert to member function
4737 dom_opt_dom_walker::after_dom_children.
4738 * tree-ssa-dse.c (dse_dom_walker): New class.
4739 (dse_enter_block): Convert to member function
4740 dse_dom_walker::before_dom_children.
4741 (tree_ssa_dse): Adjust.
4742 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4743 (determine_invariantness_stmt): Convert to method
4744 invariantness_dom_walker::before_dom_children.
4745 (determine_invariantness): Remove
4746 (move_computations_dom_walker): New class.
4747 (move_computations_stmt): Convert to method
4748 move_computations_dom_walker::before_dom_children.
4749 (move_computations, tree_ssa_lim): Adjust.
4750 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4751 (nt_init_block): Make method
4752 notrappping_dom_walker::before_dom_children.
4753 (nt_fini_block): Make
4754 method nontrapping_dom_walker::after_dom_children.
4755 (get_non_trapping): Adjust.
4756 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4757 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4758 (eliminate_leave_block): Make method.
4759 eliminate_dom_walker::after_dom_children.
4760 (eliminate): Adjust
4761 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4762 (strlen_enter_block): Make method
4763 strlen_dom_walker::before_dom_children.
4764 (strlen_leave_block): Make
4765 method strlen_dom_walker::after_dom_children.
4766 (tree_ssa_strlen): Adjust.
4767 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4768 (tree_ssa_uncprop): Adjust.
4769 (uncprop_leave_block): Make method
4770 uncprop_dom_walker::after_dom_children.
4771 (uncprop_leave_block): Make method
4772 uncprop_dom_walker::before_dom_children.
4773
4774 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4775
4776 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4777 branch in same basic block. Check both src and dest of the move insn.
4778
4779 2013-09-17 Nick Clifton <nickc@redhat.com>
4780
4781 * config/rl78/rl78-real.md (bf): New pattern.
4782 (bt): New pattern.
4783 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4784 (rl78_print_operand): Do not put a # before a %B.
4785 * config/rl78/rl78.opt: Tweak doc strings.
4786
4787 2013-09-17 DJ Delorie <dj@redhat.com>
4788
4789 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4790 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4791 (rl78_option_override): Likewise, if -mallregs.
4792 (is_virtual_register): Likewise.
4793 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4794 (REGNO_OK_FOR_BASE_P): Likewise.
4795 * config/rl78/rl78.opt (-mallregs): New.
4796
4797 2013-09-17 Nick Clifton <nickc@redhat.com>
4798
4799 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4800 For interrupt functions: save all call clobbered registers if the
4801 interrupt handler is not a leaf function.
4802 (rl78_expand_prologue): Always recompute the frame information.
4803 For interrupt functions: only select bank 0 if one of the bank 0
4804 registers is going to be psuhed.
4805
4806 2013-09-17 DJ Delorie <dj@redhat.com>
4807
4808 * config/rl78/constraints.md: For each W* constraint, rename to C*
4809 and create a W* constraint that checks for an optional ES: prefix
4810 pattern also.
4811 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4812 (es_addr): New. Used to wrap far addresses.
4813 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4814 (rl78_es_base): New.
4815 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4816 wrapped far addresses.
4817 (rl78_print_operand_1): Unwrap far addresses before processing.
4818 (rl78_lo16): Wrap far addresses in unspecs.
4819 (rl78_es_addr): New.
4820 (rl78_es_base): New.
4821 (insn_ok_now): Check for not-yet-wrapped far addresses.
4822 (transcode_memory_rtx): Properly re-wrap far addresses.
4823
4824 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4825
4826 * config/sparc/t-rtems: Add leon3 multilibs.
4827
4828 2013-09-17 Cong Hou <congh@google.com>
4829
4830 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4831 when checking the dot production pattern. The type of rhs operand
4832 of multiply is now checked correctly.
4833
4834 2013-09-17 Jeff Law <law@redhat.com>
4835
4836 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4837 edge implied equivalences into successor phis.
4838 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4839 here from tree-ssa-threadedge.c.
4840 (mark_threaded_blocks): When threading through a joiner, if both
4841 successors of the joiner's clone reach the same block, verify the
4842 PHI arguments are equal. If not, cancel the jump threading request.
4843 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4844 tree-ssa-threadupdate.c
4845 (thread_across_edge): Don't check PHI argument equality when
4846 threading through joiner block here.
4847
4848 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4849
4850 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4851 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4852 (warn_uninit, warn_uninitialized_vars,
4853 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4854 Move to tree-ssa-uninit.c.
4855 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4856 (has_undefined_value_p): New. Pass dependant parts of
4857 ssa_undefined_value_p.
4858 (uninit_undefined_value_p): Use has_undefined_value_p.
4859 (warn_uninit, warn_uninitialized_vars,
4860 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4861 Move from tree-ssa.c.
4862 * tree-ssa.h: Adjust prototypes.
4863
4864 2013-09-17 Jan Hubicka <jh@suse.cz>
4865
4866 PR middle-end/58332
4867 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4868 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4869 FUNCTION_NOT_OPTIMIZED.
4870 * ipa-inline-analysis.c (compute_inline_parameters): Function
4871 not optimized is not inlinable unless it is alwaysinline.
4872 (inline_analyze_function): Force calls in not optimized
4873 function not inlinable.
4874
4875 2013-09-17 Jan Hubicka <jh@suse.cz>
4876
4877 PR middle-end/58329
4878 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4879 to return NULL.
4880 * ipa.c (function_and_variable_visibility): Likewise.
4881 * ipa-profile.c (ipa_profile): Likewise.
4882
4883 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4884
4885 PR ipa/58398
4886 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4887 attribute, and don't inline the resolver in this case.
4888
4889 2013-09-17 Teresa Johnson <tejohnson@google.com>
4890
4891 * coverage.c (get_coverage_counts): Add missing newline.
4892
4893 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4894
4895 PR tree-optimization/58088
4896 * fold-const.c (mask_with_trailing_zeros): New function.
4897 (fold_binary_loc): Make sure we don't recurse infinitely
4898 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4899 Use mask_with_trailing_zeros where appropriate.
4900
4901 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4902
4903 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4904 of 'prev' var to get better distance estimation.
4905
4906 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4907
4908 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4909 * tree-inline.c (is_parameter_of): New predicate.
4910 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4911 a parameter has been remapped.
4912 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4913 (optimize_inline_calls): Initialize transform_parameter.
4914 (copy_gimple_seq_and_replace_locals): Likewise.
4915 (tree_function_versioning): Likewise.
4916 (maybe_inline_call_in_expr): Likewise.
4917
4918 2013-09-17 Nick Clifton <nickc@redhat.com>
4919
4920 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4921 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4922 interrupt handlers.
4923 (is_attr_func): New function.
4924 (msp430_is_interrupt_func): New function.
4925 (is_naked_func): New function.
4926 (is_reentrant_func): New function.
4927 (is_critical_func): New function.
4928 (msp430_start_function): Add annotations for function attributes.
4929 (msp430_attr): New function.
4930 (msp430_attribute_table): New.
4931 (msp430_function_section): New function.
4932 (TARGET_ASM_FUNCTION_SECTION): Define.
4933 (msp430_builtin): New enum.
4934 (msp430_init_builtins): New function.
4935 (msp430_builtin_devl): New function.
4936 (msp430_expand_builtin): New function.
4937 (TARGET_INIT_BUILTINS): Define.
4938 (TARGET_EXPAND_BUILTINS): Define.
4939 (TARGET_BUILTIN_DECL): Define.
4940 (msp430_expand_prologue): Add support for naked, interrupt,
4941 critical and reentrant functions.
4942 (msp430_expand_epilogue): Likewise.
4943 (msp430_print_operand): Handle 'O' character.
4944 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4945 NO_TRAMPOLINES.
4946 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4947 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4948 (pushm): Use a 'n' rather than an 'i' constraint.
4949 (msp_return): Add generation of the interrupt return instruction.
4950 (disable_interrupts): New pattern.
4951 (enable_interrupts): New pattern.
4952 (push_intr_state): New pattern.
4953 (pop_intr_state): New pattern.
4954 (bic_SR): New pattern.
4955 (bis_SR): New pattern.
4956 * doc/extend.texi: Document MSP430 function attributes and builtin
4957 functions.
4958
4959 2013-09-17 Richard Biener <rguenther@suse.de>
4960
4961 PR tree-optimization/58432
4962 * tree-loop-distribution.c (tree_loop_distribution): Also
4963 scan PHIs for outside loop uses and seed a partition from them.
4964
4965 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4966
4967 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4968 (restructure_reference): Call backtrace_base_for_ref.
4969
4970 2013-09-17 Alan Modra <amodra@gmail.com>
4971
4972 PR target/57589
4973 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4974 patch.
4975
4976 2013-09-16 DJ Delorie <dj@redhat.com>
4977
4978 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4979 vregs location for RL78/G10.
4980 (rl78_expand_prologue): Avoid SEL on G10.
4981 (rl78_expand_epilogue): Likewise.
4982 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4983 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4984 __RL78_G10__ when appropriate.
4985 (ASM_SPEC): Pass -mg10 along to the assembler.
4986 * config/rl78/rl78.md (sel_rb): Disable for G10.
4987 * config/rl78/rl78.opt: Add -mg10 option.
4988 * config/rl78/t-rl78: Add -mg10 multilib.
4989
4990 2013-09-16 Xinliang David Li <davidxl@google.com>
4991
4992 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4993 * omp-low.c (omp_max_vf): Ditto.
4994 (expand_omp_simd): Ditto.
4995 * tree-vectorizer.c (vectorize_loops): Ditto.
4996 (gate_vect_slp): Ditto.
4997 (gate_increase_alignment): Ditto.
4998 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4999 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5000 (gate_tree_vectorize): Name change.
5001 (tree_vectorize): Ditto.
5002 (pass_vectorize::gate): Call new function.
5003 (pass_vectorize::execute): Ditto.
5004 * opts.c: O3 default setting change.
5005 (finish_options): Check new flag.
5006 * doc/invoke.texi: Document new flags.
5007 * common.opt: New flags.
5008
5009 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5010
5011 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5012 (Condition Code): Don't reference it.
5013 * doc/tm.texi: Regenerate.
5014
5015 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5016
5017 PR middle-end/58418
5018 * lra-constraints.c (undo_optional_reloads): Consider all optional
5019 reload even if it did not get a hard reg.
5020
5021 2013-09-16 Teresa Johnson <tejohnson@google.com>
5022
5023 * dumpfile.c (dump_loc): Remove newline emission.
5024 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5025 emission to dump_printf_loc calls where missing.
5026 (vect_mark_for_runtime_alias_test): Ditto.
5027 (vect_analyze_data_ref_dependence): Ditto.
5028 (vect_analyze_data_ref_dependences): Ditto.
5029 (vect_slp_analyze_data_ref_dependence): Ditto.
5030 (vect_slp_analyze_data_ref_dependences): Ditto.
5031 (vect_compute_data_ref_alignment): Ditto.
5032 (vect_update_misalignment_for_peel): Ditto.
5033 (vect_verify_datarefs_alignment): Ditto.
5034 (vector_alignment_reachable_p): Ditto.
5035 (vect_get_data_access_cost): Ditto.
5036 (vect_enhance_data_refs_alignment): Ditto.
5037 (vect_find_same_alignment_drs): Ditto.
5038 (vect_analyze_data_refs_alignment): Ditto.
5039 (vect_analyze_group_access): Ditto.
5040 (vect_analyze_data_ref_access): Ditto.
5041 (vect_analyze_data_ref_accesses): Ditto.
5042 (vect_prune_runtime_alias_test_list): Ditto.
5043 (vect_analyze_data_refs): Ditto.
5044 (vect_create_addr_base_for_vector_ref): Ditto.
5045 (vect_create_data_ref_ptr): Ditto.
5046 (vect_grouped_store_supported): Ditto.
5047 (vect_grouped_load_supported): Ditto.
5048 * value-prof.c (check_counter): Ditto.
5049 (check_ic_target): Ditto.
5050 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5051 (vect_recog_widen_mult_pattern): Ditto.
5052 (vect_recog_widen_sum_pattern): Ditto.
5053 (vect_recog_over_widening_pattern): Ditto.
5054 (vect_recog_widen_shift_pattern): Ditto.
5055 (vect_recog_rotate_pattern): Ditto.
5056 (vect_recog_vector_vector_shift_pattern): Ditto.
5057 (vect_recog_divmod_pattern): Ditto.
5058 (vect_recog_mixed_size_cond_pattern): Ditto.
5059 (vect_recog_bool_pattern): Ditto.
5060 (vect_pattern_recog_1): Ditto.
5061 (vect_pattern_recog): Ditto.
5062 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5063 (vect_is_simple_iv_evolution): Ditto.
5064 (vect_analyze_scalar_cycles_1): Ditto.
5065 (vect_get_loop_niters): Ditto.
5066 (vect_analyze_loop_1): Ditto.
5067 (vect_analyze_loop_form): Ditto.
5068 (vect_analyze_loop_operations): Ditto.
5069 (vect_analyze_loop_2): Ditto.
5070 (vect_analyze_loop): Ditto.
5071 (report_vect_op): Ditto.
5072 (vect_is_slp_reduction): Ditto.
5073 (vect_is_simple_reduction_1): Ditto.
5074 (vect_get_known_peeling_cost): Ditto.
5075 (vect_estimate_min_profitable_iters): Ditto.
5076 (vect_model_reduction_cost): Ditto.
5077 (vect_model_induction_cost): Ditto.
5078 (get_initial_def_for_induction): Ditto.
5079 (vect_create_epilog_for_reduction): Ditto.
5080 (vectorizable_reduction): Ditto.
5081 (vectorizable_induction): Ditto.
5082 (vectorizable_live_operation): Ditto.
5083 (vect_loop_kill_debug_uses): Ditto.
5084 (vect_transform_loop): Ditto.
5085 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5086 (vect_stmt_relevant_p): Ditto.
5087 (process_use): Ditto.
5088 (vect_mark_stmts_to_be_vectorized): Ditto.
5089 (vect_model_simple_cost): Ditto.
5090 (vect_model_promotion_demotion_cost): Ditto.
5091 (vect_model_store_cost): Ditto.
5092 (vect_get_store_cost): Ditto.
5093 (vect_model_load_cost): Ditto.
5094 (vect_get_load_cost): Ditto.
5095 (vect_init_vector_1): Ditto.
5096 (vect_get_vec_def_for_operand): Ditto.
5097 (vect_finish_stmt_generation): Ditto.
5098 (vectorizable_call): Ditto.
5099 (vectorizable_conversion): Ditto.
5100 (vectorizable_assignment): Ditto.
5101 (vectorizable_shift): Ditto.
5102 (vectorizable_operation): Ditto.
5103 (vectorizable_store): Ditto.
5104 (vectorizable_load): Ditto.
5105 (vectorizable_condition): Ditto.
5106 (vect_analyze_stmt): Ditto.
5107 (vect_transform_stmt): Ditto.
5108 (vect_is_simple_use): Ditto.
5109 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5110 (vect_can_advance_ivs_p): Ditto.
5111 (vect_update_ivs_after_vectorizer): Ditto.
5112 (vect_do_peeling_for_loop_bound): Ditto.
5113 (vect_gen_niters_for_prolog_loop): Ditto.
5114 (vect_update_inits_of_drs): Ditto.
5115 (vect_create_cond_for_alias_checks): Ditto.
5116 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5117 (vect_build_slp_tree_1): Ditto.
5118 (vect_supported_load_permutation_p): Ditto.
5119 (vect_analyze_slp_instance): Ditto.
5120 (vect_analyze_slp): Ditto.
5121 (vect_make_slp_decision): Ditto.
5122 (vect_detect_hybrid_slp): Ditto.
5123 (vect_bb_vectorization_profitable_p): Ditto.
5124 (vect_slp_analyze_bb_1): Ditto.
5125 (vect_update_slp_costs_according_to_vf): Ditto.
5126 (vect_get_mask_element): Ditto.
5127 (vect_transform_slp_perm_load): Ditto.
5128 (vect_schedule_slp_instance): Ditto.
5129 (vect_schedule_slp): Ditto.
5130 (vect_slp_transform_bb): Ditto.
5131 * profile.c (read_profile_edge_counts): Ditto.
5132 (compute_branch_probabilities): Ditto.
5133 * coverage.c (get_coverage_counts): Ditto.
5134
5135 2013-09-16 Diego Novillo <dnovillo@google.com>
5136
5137 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5138
5139 2013-09-16 Jan Hubicka <jh@suse.cz>
5140
5141 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5142 abstract functions; for static functions check the presence of body.
5143
5144 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5145
5146 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5147 * config/aarch64/aarch64-simd.md
5148 (aarch64_mla_elt<mode>): New.
5149 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5150 (aarch64_mls_elt<mode>): Likewise.
5151 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5152 (aarch64_fma4_elt<mode>): Likewise.
5153 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5154 (aarch64_fma4_elt_to_128v2df): Likewise.
5155 (aarch64_fma4_elt_to_64df): Likewise.
5156 (fnma<mode>4): Likewise.
5157 (aarch64_fnma4_elt<mode>): Likewise.
5158 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5159 (aarch64_fnma4_elt_to_128v2df): Likewise.
5160 (aarch64_fnma4_elt_to_64df): Likewise.
5161 * config/aarch64/iterators.md (VDQSF): New.
5162 * config/aarch64/arm_neon.h
5163 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5164 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5165
5166 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5167
5168 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5169 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5170 (aarch64_mul3_elt_to_128df): Likewise.
5171 (aarch64_mul3_elt_to_64v2df): Likewise.
5172 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5173 (VMUL): New.
5174 (VMUL_CHANGE_NLANES) Likewise.
5175 (h_con): Likewise.
5176 (f): Likewise.
5177 * config/aarch64/arm_neon.h
5178 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5179
5180 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5181
5182 * config/aarch64/arm_neon.h
5183 (vcvtx_high_f32_f64): Fix parameters.
5184
5185 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5186 Uros Bizjak <ubizjak@gmail.com>
5187
5188 * config/alpha.c: Include tree-ssanames.h.
5189
5190 2013-09-16 Richard Biener <rguenther@suse.de>
5191
5192 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5193 (dot_rdg_1): Handle control_dd.
5194 (create_edge_for_control_dependence): New function.
5195 (create_rdg_edges): Add control dependences if asked for.
5196 (build_rdg): Likewise.
5197 (generate_loops_for_partition): If there are not necessary
5198 control stmts remove all their dependencies.
5199 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5200 (distribute_loop): Pass on control dependences.
5201 (tree_loop_distribution): Compute control dependences and remove
5202 restriction on number of loop nodes.
5203
5204 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5205
5206 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5207 for internal calls.
5208
5209 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5210
5211 * cse.c (try_const_anchors): Punt on CC modes.
5212
5213 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5214
5215 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5216
5217 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5218
5219 PR target/58382
5220 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5221 calls to word_mode.
5222
5223 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5224
5225 PR target/48094
5226 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5227 seen.
5228 (darwin_objc1_section): Likewise.
5229 (darwin_file_end): Emit Image Info section when required.
5230
5231 2013-09-14 Jan Hubicka <jh@suse.cz>
5232
5233 * tree-into-ssa.c (gate_into_ssa): New.
5234 (pass_data_build_ssa): Use it.
5235 * cgraph.h (expand_thunk): Update prototype.
5236 * cgraphunit.c (analyze_function): Expand thunks early.
5237 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5238 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5239 set lowered flag; do not add new function.
5240 (assemble_thunks_and_aliases): Update.
5241 * tree-ssa.c (gate_init_datastructures): New gate.
5242 (pass_data_init_datastructures): Use it.
5243
5244 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5245
5246 PR target/58269
5247 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5248 xmm register set described in the psABI.
5249
5250 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5251
5252 * dwarf2out.c (should_emit_struct_debug): Add check
5253 for type_decl variable is not NULL.
5254
5255 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5256
5257 * config.gcc: Use new winnt-c.c target hooks
5258 * config/t-winnt: New file
5259 * config/winnt-c.c: New file
5260 * doc/tm.texi.in: Document new hook
5261 * doc/tm.texi: Regenerated
5262
5263 2013-09-13 Jan Hubicka <jh@suse.cz>
5264
5265 PR middle-end/58094
5266 * ipa-inline.c (check_callers): New function.
5267 (check_caller_edge): Remove.
5268 (want_inline_function_to_all_callers_p): Also permit alises that are
5269 called dirrectly.
5270 (inline_to_all_callers): Terminate the walk when devirtualization
5271 introduce new calls.
5272
5273 2013-09-13 Jan Hubicka <jh@suse.cz>
5274
5275 * ipa-inline-analysis.c (struct growth_data): Add node.
5276 (do_estimate_growth_1): Fix detection of recursion.
5277
5278 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5279
5280 PR tree-optimization/58392
5281 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5282 to avoid shadowing of outer loop variable. If
5283 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5284 replace_by_duplicate_decl simduid of loops that have it set and
5285 set dest_cfun->has_simduid_loops and/or
5286 dest_cfun->has_force_vect_loops.
5287 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5288 instead of maybe_lookup_decl.
5289 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5290 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5291 call. Copy over force_vect and copy and remap simduid. Set
5292 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5293 (copy_cfg_body): Remove blocks_to_copy argument. Use
5294 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5295 caller. Don't set cfun->has_simduid_loops and/or
5296 cfun->has_force_vect_loops here.
5297 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5298 caller.
5299 (expand_call_inline, tree_function_versioning): Adjust copy_body
5300 callers.
5301
5302 2013-09-13 Martin Jambor <mjambor@suse.cz>
5303
5304 PR bootstrap/58388
5305 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5306 the assert if the edge was a speculative one.
5307
5308 2013-09-13 Richard Biener <rguenther@suse.de>
5309
5310 * tree-data-ref.h (known_dependences_p): Move here ...
5311 * tree-loop-distribution.c (known_dependences_p): ... from here.
5312 (dump_rdg_component, debug_rdg_component): Remove.
5313 (dump_rdg): Adjust.
5314 (generate_loops_for_partition): Use gimple_uid instead of
5315 relying on matching stmt visit order.
5316 (rdg_build_partitions): Take starting stmt vector.
5317 (ldist_gen): Merge into ...
5318 (distribute_loop): ... this function. Do not compute starting
5319 vertices vector.
5320 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5321
5322 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5323
5324 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5325 Set type attribute correctly. Set predicable_short_it attribute.
5326 (cmpsi_shiftsi): Remove %? from output template.
5327
5328 2013-09-13 Richard Biener <rguenther@suse.de>
5329
5330 * tree-loop-distribution.c (struct rdg_component,
5331 rdg_defs_used_in_other_loops_p, free_rdg_components,
5332 rdg_build_components): Remove.
5333 (stmts_from_loop): Do not record virtual PHIs.
5334 (generate_loops_for_partition): Skip virtual PHIs.
5335 (build_rdg_partition_for_component): Rename to ...
5336 (build_rdg_partition_for_vertex): ... this and adjust.
5337 (rdg_build_partitions): Take a vector of starting vertices
5338 instead of components. Remove unnecessary leftover handling.
5339 (ldist_gen): Do not build components or record other stores.
5340 (distribute_loop): Do not distribute loops containing stmts
5341 with side-effects.
5342
5343 2013-09-13 Christian Bruel <christian.bruel@st.com>
5344
5345 PR target/58314
5346 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5347
5348 2013-09-13 Kai Tietz <ktietz@redhat.com>
5349
5350 * config.gcc: Separate cases for mingw and cygwin targets,
5351 and add 64-bit cygwin target case.
5352
5353 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5354 dll-export inline-functions.
5355 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5356
5357
5358 2013-09-13 Jeff Law <law@redhat.com>
5359
5360 PR middle-end/58387
5361 Revert:
5362 2013-09-06 Jeff Law <law@redhat.com>
5363
5364 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5365 edge implied equivalences into successor phis.
5366
5367 2013-09-12 DJ Delorie <dj@redhat.com>
5368
5369 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5370 separators.
5371
5372 2013-09-12 Brooks Moses <bmoses@google.com>
5373
5374 PR driver/42955
5375 * Makefile.in: Do not install driver binaries in $(target)/bin.
5376
5377 2013-09-12 DJ Delorie <dj@redhat.com>
5378
5379 * config/rl78/rl78.opt (mrelax): New.
5380 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5381 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5382
5383 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5384 between SP and FP.
5385 (rl78_expand_epilogue): Likewise.
5386
5387 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5388
5389 PR middle-end/58335
5390 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5391 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5392 pointer elimination with using remove_reg_equal_offset_note.
5393
5394 2013-09-12 DJ Delorie <dj@redhat.com>
5395
5396 * config/msp430/: New port.
5397 * config.gcc (msp430): Added.
5398 * doc/invoke.texi: Document MSP430 options.
5399 * doc/install.texi: Document msp430-elf
5400 * doc/md.texi: Document msp430-elf
5401 * doc/contrib.texi: Document msp430-elf
5402
5403 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5404 PSImode.
5405
5406 2013-09-12 Martin Jambor <mjambor@suse.cz>
5407
5408 PR ipa/58389
5409 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5410 reference descriptor is NULL.
5411 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5412 edge to NULL.
5413
5414 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5415
5416 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5417 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5418 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5419 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5420 make_temp_ssa_name): move to tree-ssanames.h
5421 * tree-ssa-alias.h: Move prototype.
5422 * tree-ssa.h: Include tree-ssanames.h.
5423 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5424 * tree-ssanames.h: New. Move items from tree-flow*.h
5425 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5426
5427 2013-09-12 Richard Biener <rguenther@suse.de>
5428
5429 PR tree-optimization/58404
5430 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5431 propagate non-invariant addresses into dereferences wrapped
5432 in component references.
5433
5434 2013-09-12 Richard Biener <rguenther@suse.de>
5435
5436 PR tree-optimization/58402
5437 * passes.def: Move pass_late_warn_uninitialized later.
5438
5439 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5440
5441 * tree-ssa.h: New. Move content from tree-flow.h and
5442 tree-flow-inline.h.
5443 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5444 Move prototypes belonging to tree-ssa.c.
5445 * tree-flow-inline.h (redirect_edge_var_map_def,
5446 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5447 tree-ssa.h.
5448 * gimple.h: Adjust prototypes.
5449 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5450 to...
5451 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5452 * tree.h: Move prototype to tree-ssa.h.
5453 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5454 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5455 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5456 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5457 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5458 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5459 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5460 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5461 graphite-blocking.c, graphite-clast-to-gimple.c,
5462 graphite-dependences.c, graphite-interchange.c,
5463 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5464 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5465 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5466 ipa-reference.c, ipa-split.c, ipa-utils.c,
5467 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5468 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5469 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5470 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5471 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5472 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5473 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5474 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5475 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5476 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5477 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5478 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5479 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5480 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5481 value-prof.c, var-tracking.c,
5482 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5483
5484 2013-09-12 Richard Biener <rguenther@suse.de>
5485
5486 PR tree-optimization/58396
5487 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5488 (build_rdg): Take a loop-nest parameter, fix memleaks.
5489 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5490
5491 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5492
5493 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5494 for SLM.
5495
5496 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5497
5498 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5499 The FMA instruction names should have a 'v' prefix.
5500
5501 2013-09-12 Richard Biener <rguenther@suse.de>
5502
5503 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5504 (dot_rdg): Use popen instead of system in optional code.
5505 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5506 (already_processed_vertex_p): Adjust.
5507 (has_anti_or_output_dependence, predecessor_has_mem_write,
5508 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5509 rdg_flag_uses): Remove.
5510 (rdg_flag_vertex): Simplify.
5511 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5512 remove recursion.
5513 (build_rdg_partition_for_component): Process the first vertex
5514 of a component only.
5515 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5516
5517 2013-09-12 Alan Modra <amodra@gmail.com>
5518
5519 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5520
5521 2013-09-11 DJ Delorie <dj@redhat.com>
5522 Nick Clifton <nickc@redhat.com>
5523
5524 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5525 (rl78_stack_based_mem): New.
5526 * config/rl78/constraints.md (Iv08): New.
5527 (Iv16): New.
5528 (Iv24): New.
5529 (Is09): New.
5530 (Is17): New.
5531 (Is25): New.
5532 (ISsi): New.
5533 (IShi): New.
5534 (ISqi): New.
5535 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5536 (movhi): Likewise.
5537 (movsi): Change from expand to insn-and-split.
5538 (ashrsi3): Clobber AX.
5539 (lshrsi3): New.
5540 (ashlsi3): New.
5541 (cbranchsi4): New.
5542 * config/rl78/rl78.md (CC_REG): Fix.
5543 (addsi3): Allow memory and immediate operands.
5544 (addsi3_internal): Split into...
5545 (addsi3_internal_virt): ...new, and ...
5546 (addsi3_internal_real): ...new.
5547 (subsi): New.
5548 (subsi3_internal_virt): New.
5549 (subsi3_internal_real): New.
5550 (mulsi3): Add memory operand.
5551 (mulsi3_rl78): Likewise.
5552 (mulsi3_g13): Likewise.
5553 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5554 (cbranchqi4_real): Add more constraint options.
5555 (cbranchhi4_real): Expand pattern.
5556 (cbranchhi4_real_signed): New.
5557 (cbranchhi4_real_inverted): New.
5558 (cbranchsi4_real_lt): New.
5559 (cbranchsi4_real_ge): New.
5560 (cbranchsi4_real_signed): New.
5561 (cbranchsi4_real): New.
5562 (peephole2): New.
5563 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5564 constant shifts.
5565 (lshrsi3_virt): Likewise.
5566 (ashlsi3_virt): Likewise.
5567 (cbranchqi4_virt_signed): New.
5568 (cbranchhi4_virt_signed): New.
5569 (cbranchsi4_virt): New.
5570 * config/rl78/rl78.c: Whitespace fixes throughout.
5571 (move_elim_pass): New.
5572 (pass_data_rl78_move_elim): New.
5573 (pass_rl78_move_elim): New.
5574 (make_pass_rl78_move_elim): New.
5575 (rl78_devirt_info): Run devirt earlier.
5576 (rl78_move_elim_info): New.
5577 (rl78_asm_file_start): Register it.
5578 (rl78_split_movsi): New.
5579 (rl78_as_legitimate_address): Allow virtual base registers when
5580 appropriate.
5581 (rl78_addr_space_convert): Remove spurious debug stuff.
5582 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5583 (rl78_print_operand): More cases for not printing '#'.
5584 (rl78_expand_compare): Remove most of the logic.
5585 (content_memory): New.
5586 (clear_content_memory): New.
5587 (get_content_index): New.
5588 (get_content_name): New.
5589 (display_content_memory): New.
5590 (update_content): New.
5591 (record_content): New.
5592 (already_contains): New.
5593 (insn_ok_now): Re-recog insns with virtual registers.
5594 (add_postponed_content_update): New.
5595 (process_postponed_content_update): New.
5596 (gen_and_emit_move): New.
5597 (transcode_memory_rtx): Record new location content.
5598 Use gen_and_emit_move.
5599 (force_into_acc): New.
5600 (move_to_acc): Use gen_and_emit_move.
5601 (move_from_acc): Likewise.
5602 (move_acc_to_reg): Likewise.
5603 (move_to_x): Likewise.
5604 (move_to_hl): Likewise.
5605 (move_to_de): Likewise.
5606 (rl78_alloc_physical_registers_op1): Record location content.
5607 (has_constraint): New.
5608 (rl78_alloc_physical_registers_op2): Record location content.
5609 Optimize use of HL.
5610 (rl78_alloc_physical_registers_ro1): Likewise.
5611 (rl78_alloc_physical_registers_cmp): Likewise.
5612 (rl78_alloc_physical_registers_umul): Likewise.
5613 (rl78_alloc_address_registers_macax): New.
5614 (rl78_alloc_physical_registers): Initialize and set location
5615 content memory as needed.
5616 (rl78_reorg): Make sure split2 is called.
5617 (rl78_rtx_costs): New.
5618
5619 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5620
5621 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5622 for (not (neg ...)) and (neg (not ...)) cases.
5623
5624 2013-09-11 Richard Biener <rguenther@suse.de>
5625
5626 PR middle-end/58377
5627 * passes.def: Split critical edges before late uninit warning passes.
5628 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5629
5630 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5631
5632 PR tree-optimization/58385
5633 * fold-const.c (build_range_check): If both low and high are NULL,
5634 use omit_one_operand_loc to preserve exp side-effects.
5635
5636 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5637
5638 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5639
5640 2013-09-11 Richard Biener <rguenther@suse.de>
5641
5642 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5643 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5644 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5645 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5646 stmts_from_loop, known_dependences_p, build_empty_rdg,
5647 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5648 * tree-loop-distribution.c: ... here.
5649 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5650 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5651 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5652 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5653 * tree-loop-distribution.c: ... here.
5654 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5655 (struct partition_s): Add loops member.
5656 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5657 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5658 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5659
5660 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5661 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5662 Sergey Lega <sergey.s.lega@intel.com>
5663 Anna Tikhonova <anna.tikhonova@intel.com>
5664 Ilya Tocar <ilya.tocar@intel.com>
5665 Andrey Turetskiy <andrey.turetskiy@intel.com>
5666 Ilya Verbin <ilya.verbin@intel.com>
5667 Kirill Yukhin <kirill.yukhin@intel.com>
5668 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5669
5670 * config/i386/constraints.md (k): New.
5671 (Yk): Ditto.
5672 * config/i386/i386.c (const regclass_map): Add new mask registers.
5673 (dbx_register_map): Ditto.
5674 (dbx64_register_map): Ditto.
5675 (svr4_dbx_register_map): Ditto.
5676 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5677 disabled.
5678 (ix86_preferred_reload_class): Disable constants for mask registers.
5679 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5680 (ix86_hard_regno_mode_ok): Support new mask registers.
5681 (x86_order_regs_for_local_alloc): Ditto.
5682 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5683 (FIXED_REGISTERS): Add new mask registers.
5684 (CALL_USED_REGISTERS): Ditto.
5685 (REG_ALLOC_ORDER): Ditto.
5686 (VALID_MASK_REG_MODE): New.
5687 (FIRST_MASK_REG): Ditto.
5688 (LAST_MASK_REG): Ditto.
5689 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5690 (MAYBE_MASK_CLASS_P): New.
5691 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5692 (REG_CLASS_CONTENTS): Ditto.
5693 (MASK_REGNO_P): New.
5694 (ANY_MASK_REG_P): Ditto.
5695 (HI_REGISTER_NAMES): Add new mask registers.
5696 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5697 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5698 mask registers.
5699 (attribute "type"): Add mskmov, msklog.
5700 (attribute "length_immediate"): Support them.
5701 (attribute "memory"): Ditto.
5702 (attribute "prefix_0f"): Ditto.
5703 (*movhi_internal): Support new mask registers.
5704 (*movqi_internal): Ditto.
5705 (define_split): Split out clobber pattern is a logic
5706 insn on mask registers.
5707 (*k<logic><mode>): New.
5708 (*andhi_1): Extend to support mask regs.
5709 (*andqi_1): Extend to support mask regs.
5710 (kandn<mode>): New.
5711 (define_split): Split and-not to and and not if operands
5712 are not mask regs.
5713 (*<code><mode>_1): Separate HI mode to new pattern...
5714 (*<code>hi_1): This.
5715 (*<code>qi_1): Extend to support mask regs.
5716 (kxnor<mode>): New.
5717 (kortestzhi): Ditto.
5718 (kortestchi): Ditto.
5719 (kunpckhi): Ditto.
5720 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5721 (*one_cmplhi2_1): ...Here, now with mask registers support.
5722 (*one_cmplqi2_1): Support new mask registers.
5723 (HI/QImode arithmetics splitter): Don't split if mask registers
5724 are used.
5725 (HI/QImode not splitter): Ditto.
5726 * config/i386/predicated.md (mask_reg_operand): New.
5727 (general_reg_operand): Ditto.
5728
5729 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5730
5731 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5732 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5733
5734 2013-09-10 Jeff Law <law@redhat.com>
5735
5736 PR tree-optimization/58380
5737 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5738 of threading through a buried loop header.
5739
5740 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5741 return value for single successor case.
5742
5743 2013-09-10 Jan Hubicka <jh@suse.cz>
5744
5745 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5746
5747 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5748
5749 PR target/58361
5750 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5751 support conditional execution.
5752 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5753
5754 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5755
5756 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5757 constraint pass.
5758 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5759 Check destination too to check move insn.
5760 (undo_optional_reloads): Add check that the original peudo did not
5761 changed its allocation and the optional reload was inherited on last
5762 inheritance pass. Break loop after deciding to keep optional reload.
5763 (lra_undo_inheritance): Add check that inherited pseudo still in
5764 memory.
5765
5766 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5767
5768 * config/aarch64/aarch64.md (generic_sched): New.
5769 * config/aarch64/aarch64-generic.md (load): Make conditional
5770 on generic_sched attribute.
5771 (nonload): Likewise.
5772
5773 2013-09-10 Jan Hubicka <jh@suse.cz>
5774
5775 * lto-cgraph.c: Include ipa-utils.h.
5776 (compute_ltrans_boundary): Also add possible targets into the boundary.
5777
5778 2013-09-10 Jan Hubicka <jh@suse.cz>
5779
5780 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5781 VAR_DECL of vtable rather than full expression.
5782
5783 2013-09-10 Jan Hubicka <jh@suse.cz>
5784 Paolo Carlini <paolo.carlini@oracle.com>
5785
5786 * cgraphunit.c (analyze_functions): Save input_location, set it
5787 to UNKNOWN_LOCATION and restore it at the end.
5788
5789 2013-09-10 Martin Jambor <mjambor@suse.cz>
5790
5791 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5792 represented by a thunk.
5793
5794 2013-09-10 Jeff Law <law@redhat.com>
5795
5796 PR tree-optimization/58343
5797 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5798 jump threading requests through loop headers buried in the middle
5799 of a jump threading path.
5800
5801 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5802 in return value/type.
5803
5804 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5805
5806 PR rtl-optimization/58365
5807 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5808 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5809 it differs.
5810
5811 2013-09-10 Richard Biener <rguenther@suse.de>
5812
5813 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5814 * tree-data-ref.c (create_rdg_vertices): Collect all data
5815 references, signal failure to the caller, use data-ref API.
5816 (build_rdg): Compute data references only once. Maintain lifetime
5817 of data references and data dependences from within RDG.
5818 (free_rdg): Free dependence relations.
5819 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5820 inventing extra dependences.
5821 (distribute_loop): Update for RDG API changes.
5822
5823 2013-09-10 Kai Tietz <ktietz@redhat.com>
5824
5825 * doc/invoke.texi (fms-extensions): Document changed
5826 behavior for ms-abi targets.
5827 * config/i386/i386.c (ix86_option_override_internal):
5828 Set default value of option -fms-extension for ms-abi targets.
5829
5830 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5831
5832 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5833
5834 2013-09-10 Alan Modra <amodra@gmail.com>
5835
5836 PR target/58330
5837 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5838
5839 2013-09-10 Alan Modra <amodra@gmail.com>
5840
5841 * config/rs6000/predicates.md (add_cint_operand): New.
5842 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5843 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5844 using add_cint_operand.
5845 (largetoc_high_plus_aix): Likewise.
5846
5847 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5848
5849 PR tree-optimization/58364
5850 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5851 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5852 the current range can't be an unconditional true or false.
5853
5854 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5855
5856 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5857
5858 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5859
5860 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5861
5862 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5863
5864 PR c++/43452
5865 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5866
5867 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5868
5869 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5870 NO_REGS for immediate that can't be moved directly into FP_REGS.
5871
5872 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5873
5874 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5875 comparison with negated operand.
5876 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5877 RTL form.
5878
5879 2013-09-09 Richard Biener <rguenther@suse.de>
5880
5881 PR middle-end/58326
5882 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5883 of a subloop record all its block as affecting loop-closed SSA form.
5884
5885 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5886
5887 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5888 of an rtx/bitpos pair.
5889 (store_fixed_bit_field): Update accordingly.
5890
5891 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5892
5893 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5894 GEN_INT.
5895 * builtins.c (expand_errno_check): Likewise.
5896 * dwarf2cfi.c (init_return_column_size): Likewise.
5897 * except.c (sjlj_mark_call_sites): Likewise.
5898 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5899 * lra-constraints.c (emit_inc): Likewise.
5900 * ree.c (combine_set_extension): Likewise.
5901 * regmove.c (fixup_match_2): Likewise.
5902 * reload1.c (inc_for_reload): Likewise.
5903
5904 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5905
5906 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5907 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5908 Use gen_int_mode with the mode of the associated simplify_* call.
5909 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5910 * expmed.c (expand_shift_1): Likewise.
5911 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5912 * loop-iv.c (iv_number_of_iterations): Likewise.
5913 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5914 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5915
5916 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5917
5918 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5919 of the associated expand_* call.
5920 (asan_emit_stack_protection): Likewise.
5921 * builtins.c (round_trampoline_addr): Likewise.
5922 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5923 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5924 (emit_store_flag): Likewise.
5925 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5926 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5927 Likewise.
5928 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5929 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5930 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5931 * modulo-sched.c (generate_prolog_epilog): Likewise.
5932 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5933 (expand_ctz, expand_ffs, expand_unop): Likewise.
5934
5935 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5936
5937 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5938 of the associated gen_rtx_* call.
5939 * caller-save.c (init_caller_save): Likewise.
5940 * combine.c (find_split_point, make_extraction): Likewise.
5941 (make_compound_operation): Likewise.
5942 * dwarf2out.c (mem_loc_descriptor): Likewise.
5943 * explow.c (plus_constant, probe_stack_range): Likewise.
5944 * expmed.c (expand_mult_const): Likewise.
5945 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5946 * reload1.c (init_reload): Likewise.
5947 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5948 * var-tracking.c (adjust_mems): Likewise.
5949 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5950 rather than gen_rtx_fmt_ee.
5951
5952 2013-09-09 Jan Hubicka <jh@suse.cz>
5953
5954 PR middle-end/58294
5955 * value-prof.c (gimple_ic): Copy also abnormal edges.
5956
5957 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5958
5959 * asan.c (asan_shadow_cst): Use gen_int_mode.
5960
5961 2013-09-08 Jan Hubicka <jh@suse.cz>
5962
5963 * ipa-profile.c: Add toplevel comment.
5964 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5965 (contains_hot_call_p): New function.
5966 (ipa_propagate_frequency): Set frequencies based on counts when
5967 profile is read.
5968 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5969 profile; do not tamper with profile after inlining if it is read.
5970
5971 2013-09-08 Jan Hubicka <jh@suse.cz>
5972
5973 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5974 speculative edges.
5975
5976 2013-09-08 Jan Hubicka <jh@suse.cz>
5977
5978 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5979 summary generation.
5980
5981 2013-09-08 Jeff Law <law@redhat.com>
5982
5983 PR bootstrap/58340
5984 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5985 of 'found'.
5986
5987 2013-09-08 Andi Kleen <ak@linux.intel.com>
5988
5989 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5990
5991 2013-09-08 Jan Hubicka <jh@suse.cz>
5992
5993 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5994
5995 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5996
5997 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5998 for non-debug insns.
5999 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6000 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6001 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6002 accordingly.
6003
6004 2013-09-08 Jan Hubicka <jh@suse.cz>
6005
6006 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6007 targets and devirtualize to BUILT_IN_UNREACHABLE.
6008 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6009 * ipa.c (walk_polymorphic_call_targets): New function.
6010 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6011 functions; use the new timevar.
6012 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6013 was removed from the program.
6014 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6015 not consider it in the walk when its vtable is dead.
6016 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6017 record_binfo.
6018 (devirt_variable_node_removal_hook): New function.
6019 (possible_polymorphic_call_targets): Also register
6020 devirt_variable_node_removal_hook.
6021 (ipa_devirt): Do not do non-speculative devirtualization.
6022 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6023
6024 2013-09-08 Jan Hubicka <jh@suse.cz>
6025
6026 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6027 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6028 varpool_remove_variable_insertion_hook): Declare.
6029 * varpool.c (varpool_node_hook_list): New structure.
6030 (first_varpool_node_removal_hook,
6031 first_varpool_variable_insertion_hook): New variables.
6032 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6033 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6034 varpool_remove_variable_insertion_hook,
6035 varpool_call_variable_insertion_hooks): New functions.
6036 (varpool_remove_node): Use it.
6037
6038 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6039
6040 PR c++/54941
6041 * diagnostic.c (diagnostic_build_prefix): When s.file is
6042 "<built-in>" don't output line and column numbers.
6043
6044 2013-09-06 Jan Hubicka <jh@suse.cz>
6045
6046 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6047 * lto-streamer-out.c: Stream thunks, too.
6048 * lto-streamer-in.c (input_function): Pop cfun here
6049 (lto_read_body): Instead of here.
6050
6051 2013-09-06 Caroline Tice <cmtice@google.com>
6052
6053 * doc/install.texi: Add documentation for the --enable-vtable-verify
6054 and the --disable-libvtv configure options.
6055
6056 2013-09-06 Jeff Law <law@redhat.com>
6057
6058 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6059 edge implied equivalences into successor phis.
6060
6061 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6062
6063 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6064
6065 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6066
6067 * resource.c (mark_target_live_regs): Compute resources taking
6068 into account if a call is predicated or not.
6069
6070 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6071
6072 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6073 the compiler in the function names.
6074
6075 2013-09-06 Jan Hubicka <jh@suse.cz>
6076
6077 PR middle-end/58094
6078 * ipa-inline.c (has_caller_p): New function.
6079 (want_inline_function_to_all_callers_p): Use it.
6080 (sum_callers, inline_to_all_callers): Break out from ...
6081 (ipa_inline): ... here.
6082
6083 2013-09-06 Jan Hubicka <jh@suse.cz>
6084
6085 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6086 only when AVX is enabled.
6087
6088 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6089
6090 * config/aarch64/aarch64.md
6091 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6092 is fpsimd_<load/store>2.
6093 (load_pair<mode>): Likewise.
6094 (store_pair<mode>): Likewise.
6095
6096 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6097
6098 * config/arm/types.md (type): Add "mrs" type.
6099 * config/aarch64/aarch64.md
6100 (aarch64_load_tp_hard): Make type "mrs".
6101 * config/arm/arm.md
6102 (load_tp_hard): Make type "mrs".
6103 * config/arm/cortex-a15.md: Update with new attributes.
6104 * config/arm/cortex-a5.md: Update with new attributes.
6105 * config/arm/cortex-a53.md: Update with new attributes.
6106 * config/arm/cortex-a7.md: Update with new attributes.
6107 * config/arm/cortex-a8.md: Update with new attributes.
6108 * config/arm/cortex-a9.md: Update with new attributes.
6109 * config/arm/cortex-m4.md: Update with new attributes.
6110 * config/arm/cortex-r4.md: Update with new attributes.
6111 * config/arm/fa526.md: Update with new attributes.
6112 * config/arm/fa606te.md: Update with new attributes.
6113 * config/arm/fa626te.md: Update with new attributes.
6114 * config/arm/fa726te.md: Update with new attributes.
6115
6116 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6117
6118 * config/aarch64/aarch64.md
6119 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6120 (*movtf_aarch64): Likewise.
6121 * config/arm/arm.md
6122 (thumb1_movdi_insn): Use "multiple" for type where more than one
6123 instruction is used for a move.
6124 (*arm32_movhf): Likewise.
6125 (*thumb_movdf_insn): Likewise.
6126
6127 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6128
6129 * config/arm/types.md (type): Rename fcpys to fmov.
6130 * config/arm/vfp.md
6131 (*arm_movsi_vfp): Rename type fcpys as fmov.
6132 (*thumb2_movsi_vfp): Likewise
6133 (*movhf_vfp_neon): Likewise
6134 (*movhf_vfp): Likewise
6135 (*movsf_vfp): Likewise
6136 (*thumb2_movsf_vfp): Likewise
6137 (*movsfcc_vfp): Likewise
6138 (*thumb2_movsfcc_vfp): Likewise
6139 * config/aarch64/aarch64-simd.md
6140 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6141 * config/aarch64/aarch64.md
6142 (*movsi_aarch64): Replace type mov_reg with fmovs.
6143 (*movdi_aarch64): Likewise
6144 (*movsf_aarch64): Likewise
6145 (*movdf_aarch64): Likewise
6146 * config/arm/arm.c
6147 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6148 * config/arm/iwmmxt.md
6149 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6150 * config/arm/arm1020e.md: Update with new attributes.
6151 * config/arm/cortex-a15-neon.md: Update with new attributes.
6152 * config/arm/cortex-a5.md: Update with new attributes.
6153 * config/arm/cortex-a53.md: Update with new attributes.
6154 * config/arm/cortex-a7.md: Update with new attributes.
6155 * config/arm/cortex-a8-neon.md: Update with new attributes.
6156 * config/arm/cortex-a9.md: Update with new attributes.
6157 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6158 * config/arm/cortex-r4f.md: Update with new attributes.
6159 * config/arm/marvell-pj4.md: Update with new attributes.
6160 * config/arm/vfp11.md: Update with new attributes.
6161
6162 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6163
6164 * config/aarch64/aarch64.md
6165 (*madd<mode>): Fix type attribute.
6166 (*maddsi_uxtw): Likewise.
6167 (*msub<mode>): Likewise.
6168 (*msubsi_uxtw): Likewise.
6169 (<su_optab>maddsidi4): Likewise.
6170 (<su_optab>msubsidi4): Likewise.
6171
6172 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6173
6174 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6175 * config/arm/arm.md (core_cycles): Remove fdiv.
6176 * config/arm/vfp.md:
6177 (*sqrtsf2_vfp): Update for attribute changes.
6178 (*sqrtdf2_vfp): Likewise.
6179 * config/aarch64/aarch64.md:
6180 (sqrt<mode>2): Update for attribute changes.
6181 * config/arm/arm1020e.md: Update with new attributes.
6182 * config/arm/cortex-a15-neon.md: Update with new attributes.
6183 * config/arm/cortex-a5.md: Update with new attributes.
6184 * config/arm/cortex-a53.md: Update with new attributes.
6185 * config/arm/cortex-a7.md: Update with new attributes.
6186 * config/arm/cortex-a8-neon.md: Update with new attributes.
6187 * config/arm/cortex-a9.md: Update with new attributes.
6188 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6189 * config/arm/cortex-r4f.md: Update with new attributes.
6190 * config/arm/marvell-pj4.md: Update with new attributes.
6191 * config/arm/vfp11.md: Update with new attributes.
6192
6193 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6194
6195 * config/arm/types.md
6196 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6197 * config/aarch64/aarch64.md
6198 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6199 new attributes.
6200 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6201 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6202 (float<GPI:mode><GPF:mode>2): Likewise.
6203 * config/arm/vfp.md
6204 (*truncsisf2_vfp): Update with new attributes.
6205 (*truncsidf2_vfp): Likewise.
6206 (fixuns_truncsfsi2): Likewise.
6207 (fixuns_truncdfsi2): Likewise.
6208 (*floatsisf2_vfp): Likewise.
6209 (*floatsidf2_vfp): Likewise.
6210 (floatunssisf2): Likewise.
6211 (floatunssidf2): Likewise.
6212 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6213 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6214 * config/arm/arm1020e.md: Update with new attributes.
6215 * config/arm/cortex-a15-neon.md: Update with new attributes.
6216 * config/arm/cortex-a5.md: Update with new attributes.
6217 * config/arm/cortex-a53.md: Update with new attributes.
6218 * config/arm/cortex-a7.md: Update with new attributes.
6219 * config/arm/cortex-a8-neon.md: Update with new attributes.
6220 * config/arm/cortex-a9.md: Update with new attributes.
6221 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6222 * config/arm/cortex-r4f.md: Update with new attributes.
6223 * config/arm/marvell-pj4.md: Update with new attributes.
6224 * config/arm/vfp11.md: Update with new attributes.
6225
6226 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6227
6228 * config/aarch64/arm_neon.h
6229 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6230 (vqtbx<1,2,3,4><q>_s8): Likewise.
6231
6232 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6233
6234 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6235 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6236 (add<mode>3): Add type attribute.
6237 (add<mode>3): Likewise.
6238 (usadd<mode>3): Likewise.
6239 (ssadd<mode>3): Likewise.
6240 (sub<mode>3): Likewise.
6241 (sub<mode>3): Likewise.
6242 (ussub<mode>3): Likewise.
6243 (sssub<mode>3): Likewise.
6244 (ssmulsa3): Likewise.
6245 (usmulusa3): Likewise.
6246 (arm_usatsihi): Likewise.
6247 * config/arm/vfp.md
6248 (*movdi_vfp): Add types for all instructions.
6249 (*movdi_vfp_cortexa8): Likewise.
6250 (*movhf_vfp_neon): Likewise.
6251 (*movhf_vfp): Likewise.
6252 (*movdf_vfp): Likewise.
6253 (*thumb2_movdf_vfp): Likewise.
6254 (*thumb2_movdfcc_vfp): Likewise.
6255 * config/arm/arm.md: Add type attribute to all insn patterns.
6256 (*thumb1_adddi3): Add type attribute.
6257 (*arm_adddi3): Likewise.
6258 (*adddi_sesidi_di): Likewise.
6259 (*adddi_zesidi_di): Likewise.
6260 (*thumb1_addsi3): Likewise.
6261 (addsi3_compare0): Likewise.
6262 (*addsi3_compare0_scratch): Likewise.
6263 (*compare_negsi_si): Likewise.
6264 (cmpsi2_addneg): Likewise.
6265 (*addsi3_carryin_<optab>): Likewise.
6266 (*addsi3_carryin_alt2_<optab>): Likewise.
6267 (*addsi3_carryin_clobercc_<optab>): Likewise.
6268 (*subsi3_carryin): Likewise.
6269 (*subsi3_carryin_const): Likewise.
6270 (*subsi3_carryin_compare): Likewise.
6271 (*subsi3_carryin_compare_const): Likewise.
6272 (*arm_subdi3): Likewise.
6273 (*thumb_subdi3): Likewise.
6274 (*subdi_di_zesidi): Likewise.
6275 (*subdi_di_sesidi): Likewise.
6276 (*subdi_zesidi_di): Likewise.
6277 (*subdi_sesidi_di): Likewise.
6278 (*subdi_zesidi_ze): Likewise.
6279 (thumb1_subsi3_insn): Likewise.
6280 (*arm_subsi3_insn): Likewise.
6281 (*anddi3_insn): Likewise.
6282 (*anddi_zesidi_di): Likewise.
6283 (*anddi_sesdi_di): Likewise.
6284 (*ne_zeroextracts): Likewise.
6285 (*ne_zeroextracts): Likewise.
6286 (*ite_ne_zeroextr): Likewise.
6287 (*ite_ne_zeroextr): Likewise.
6288 (*anddi_notdi_di): Likewise.
6289 (*anddi_notzesidi): Likewise.
6290 (*anddi_notsesidi): Likewise.
6291 (andsi_notsi_si): Likewise.
6292 (thumb1_bicsi3): Likewise.
6293 (*iordi3_insn): Likewise.
6294 (*iordi_zesidi_di): Likewise.
6295 (*iordi_sesidi_di): Likewise.
6296 (*thumb1_iorsi3_insn): Likewise.
6297 (*xordi3_insn): Likewise.
6298 (*xordi_zesidi_di): Likewise.
6299 (*xordi_sesidi_di): Likewise.
6300 (*arm_xorsi3): Likewise.
6301 (*andsi_iorsi3_no): Likewise.
6302 (*smax_0): Likewise.
6303 (*smax_m1): Likewise.
6304 (*arm_smax_insn): Likewise.
6305 (*smin_0): Likewise.
6306 (*arm_smin_insn): Likewise.
6307 (*arm_umaxsi3): Likewise.
6308 (*arm_uminsi3): Likewise.
6309 (*minmax_arithsi): Likewise.
6310 (*minmax_arithsi_): Likewise.
6311 (*satsi_<SAT:code>): Likewise.
6312 (arm_ashldi3_1bit): Likewise.
6313 (arm_ashrdi3_1bit): Likewise.
6314 (arm_lshrdi3_1bit): Likewise.
6315 (*arm_negdi2): Likewise.
6316 (*thumb1_negdi2): Likewise.
6317 (*arm_negsi2): Likewise.
6318 (*thumb1_negsi2): Likewise.
6319 (*negdi_extendsid): Likewise.
6320 (*negdi_zero_extend): Likewise.
6321 (*arm_abssi2): Likewise.
6322 (*thumb1_abssi2): Likewise.
6323 (*arm_neg_abssi2): Likewise.
6324 (*thumb1_neg_abss): Likewise.
6325 (one_cmpldi2): Likewise.
6326 (extend<mode>di2): Likewise.
6327 (*compareqi_eq0): Likewise.
6328 (*arm_extendhisi2addsi): Likewise.
6329 (*arm_movdi): Likewise.
6330 (*thumb1_movdi_insn): Likewise.
6331 (*arm_movt): Likewise.
6332 (*thumb1_movsi_insn): Likewise.
6333 (pic_add_dot_plus_four): Likewise.
6334 (pic_add_dot_plus_eight): Likewise.
6335 (tls_load_dot_plus_eight): Likewise.
6336 (*thumb1_movhi_insn): Likewise.
6337 (*thumb1_movsf_insn): Likewise.
6338 (*movdf_soft_insn): Likewise.
6339 (*thumb_movdf_insn): Likewise.
6340 (cbranchsi4_insn): Likewise.
6341 (cbranchsi4_scratch): Likewise.
6342 (*negated_cbranchsi4): Likewise.
6343 (*tbit_cbranch): Likewise.
6344 (*tlobits_cbranch): Likewise.
6345 (*tstsi3_cbranch): Likewise.
6346 (*cbranchne_decr1): Likewise.
6347 (*addsi3_cbranch): Likewise.
6348 (*addsi3_cbranch_scratch): Likewise.
6349 (*arm_cmpdi_insn): Likewise.
6350 (*arm_cmpdi_unsig): Likewise.
6351 (*arm_cmpdi_zero): Likewise.
6352 (*thumb_cmpdi_zero): Likewise.
6353 (*deleted_compare): Likewise.
6354 (*mov_scc): Likewise.
6355 (*mov_negscc): Likewise.
6356 (*mov_notscc): Likewise.
6357 (*cstoresi_eq0_thumb1_insn): Likewise.
6358 (cstoresi_nltu_thumb1): Likewise.
6359 (cstoresi_ltu_thu): Likewise.
6360 (thumb1_addsi3_addgeu): Likewise.
6361 (*arm_jump): Likewise.
6362 (*thumb_jump): Likewise.
6363 (*check_arch2): Likewise.
6364 (arm_casesi_internal): Likewise.
6365 (thumb1_casesi_dispatch): Likewise.
6366 (*arm_indirect_jump): Likewise.
6367 (*thumb1_indirect_jump): Likewise.
6368 (nop): Likewise.
6369 (*and_scc): Likewise.
6370 (*ior_scc): Likewise.
6371 (*compare_scc): Likewise.
6372 (*cond_move): Likewise.
6373 (*cond_arith): Likewise.
6374 (*cond_sub): Likewise.
6375 (*cmp_ite0): Likewise.
6376 (*cmp_ite1): Likewise.
6377 (*cmp_and): Likewise.
6378 (*cmp_ior): Likewise.
6379 (*ior_scc_scc): Likewise.
6380 (*ior_scc_scc_cmp): Likewise.
6381 (*and_scc_scc): Likewise.
6382 (*and_scc_scc_cmp): Likewise.
6383 (*and_scc_scc_nod): Likewise.
6384 (*negscc): Likewise.
6385 (movcond_addsi): Likewise.
6386 (movcond): Likewise.
6387 (*ifcompare_plus_move): Likewise.
6388 (*if_plus_move): Likewise.
6389 (*ifcompare_move_plus): Likewise.
6390 (*if_move_plus): Likewise.
6391 (*ifcompare_arith_arith): Likewise.
6392 (*if_arith_arith): Likewise.
6393 (*ifcompare_arith_move): Likewise.
6394 (*if_arith_move): Likewise.
6395 (*ifcompare_move_arith): Likewise.
6396 (*if_move_arith): Likewise.
6397 (*ifcompare_move_not): Likewise.
6398 (*if_move_not): Likewise.
6399 (*ifcompare_not_move): Likewise.
6400 (*if_not_move): Likewise.
6401 (*ifcompare_shift_move): Likewise.
6402 (*if_shift_move): Likewise.
6403 (*ifcompare_move_shift): Likewise.
6404 (*if_move_shift): Likewise.
6405 (*ifcompare_shift_shift): Likewise.
6406 (*ifcompare_not_arith): Likewise.
6407 (*ifcompare_arith_not): Likewise.
6408 (*if_arith_not): Likewise.
6409 (*ifcompare_neg_move): Likewise.
6410 (*if_neg_move): Likewise.
6411 (*ifcompare_move_neg): Likewise.
6412 (*if_move_neg): Likewise.
6413 (prologue_thumb1_interwork): Likewise.
6414 (*cond_move_not): Likewise.
6415 (*sign_extract_onebit): Likewise.
6416 (*not_signextract_onebit): Likewise.
6417 (stack_tie): Likewise.
6418 (align_4): Likewise.
6419 (align_8): Likewise.
6420 (consttable_end): Likewise.
6421 (consttable_1): Likewise.
6422 (consttable_2): Likewise.
6423 (consttable_4): Likewise.
6424 (consttable_8): Likewise.
6425 (consttable_16): Likewise.
6426 (*thumb1_tablejump): Likewise.
6427 (prefetch): Likewise.
6428 (force_register_use): Likewise.
6429 (thumb_eh_return): Likewise.
6430 (load_tp_hard): Likewise.
6431 (load_tp_soft): Likewise.
6432 (tlscall): Likewise.
6433 (*arm_movtas_ze): Likewise.
6434 (*arm_rev): Likewise.
6435 (*arm_revsh): Likewise.
6436 (*arm_rev16): Likewise.
6437 * config/arm/thumb2.md
6438 (*thumb2_smaxsi3): Likewise.
6439 (*thumb2_sminsi3): Likewise.
6440 (*thumb32_umaxsi3): Likewise.
6441 (*thumb2_uminsi3): Likewise.
6442 (*thumb2_negdi2): Likewise.
6443 (*thumb2_abssi2): Likewise.
6444 (*thumb2_neg_abss): Likewise.
6445 (*thumb2_movsi_insn): Likewise.
6446 (tls_load_dot_plus_four): Likewise.
6447 (*thumb2_movhi_insn): Likewise.
6448 (*thumb2_mov_scc): Likewise.
6449 (*thumb2_mov_negs): Likewise.
6450 (*thumb2_mov_negs): Likewise.
6451 (*thumb2_mov_nots): Likewise.
6452 (*thumb2_mov_nots): Likewise.
6453 (*thumb2_movsicc_): Likewise.
6454 (*thumb2_movsfcc_soft_insn): Likewise.
6455 (*thumb2_indirect_jump): Likewise.
6456 (*thumb2_and_scc): Likewise.
6457 (*thumb2_ior_scc): Likewise.
6458 (*thumb2_ior_scc_strict_it): Likewise.
6459 (*thumb2_cond_move): Likewise.
6460 (*thumb2_cond_arith): Likewise.
6461 (*thumb2_cond_ari): Likewise.
6462 (*thumb2_cond_sub): Likewise.
6463 (*thumb2_negscc): Likewise.
6464 (*thumb2_movcond): Likewise.
6465 (thumb2_casesi_internal): Likewise.
6466 (thumb2_casesi_internal_pic): Likewise.
6467 (*thumb2_alusi3_short): Likewise.
6468 (*thumb2_mov<mode>_shortim): Likewise.
6469 (*thumb2_addsi_short): Likewise.
6470 (*thumb2_subsi_short): Likewise.
6471 (thumb2_addsi3_compare0): Likewise.
6472 (*thumb2_cbz): Likewise.
6473 (*thumb2_cbnz): Likewise.
6474 (*thumb2_one_cmplsi2_short): Likewise.
6475 (*thumb2_negsi2_short): Likewise.
6476 (*orsi_notsi_si): Likewise.
6477 * config/arm/arm1020e.md: Update with new attributes.
6478 * config/arm/arm1026ejs.md: Update with new attributes.
6479 * config/arm/arm1136jfs.md: Update with new attributes.
6480 * config/arm/arm926ejs.md: Update with new attributes.
6481 * config/arm/cortex-a15.md: Update with new attributes.
6482 * config/arm/cortex-a5.md: Update with new attributes.
6483 * config/arm/cortex-a53.md: Update with new attributes.
6484 * config/arm/cortex-a7.md: Update with new attributes.
6485 * config/arm/cortex-a8.md: Update with new attributes.
6486 * config/arm/cortex-a9.md: Update with new attributes.
6487 * config/arm/cortex-m4.md: Update with new attributes.
6488 * config/arm/cortex-r4.md: Update with new attributes.
6489 * config/arm/fa526.md: Update with new attributes.
6490 * config/arm/fa606te.md: Update with new attributes.
6491 * config/arm/fa626te.md: Update with new attributes.
6492 * config/arm/fa726te.md: Update with new attributes.
6493
6494 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6495
6496 * config/aarch64/aarch64-simd.md
6497 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6498 <vwx> iterator to ensure correct register choice.
6499 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6500 (aarch64_sqdmull_n<mode>): Likewise.
6501 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6502 * config/aarch64/arm_neon.h
6503 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6504 (vml<as><q>_n_<su>16): Likewise.
6505 (vml<as>l_high_lane<q>_<su>16): Likewise.
6506 (vml<as>l_high_n_<su>16): Likewise.
6507 (vml<as>l_lane<q>_<su>16): Likewise.
6508 (vml<as>l_n_<su>16): Likewise.
6509 (vmul<q>_lane<q>_<su>16): Likewise.
6510 (vmul<q>_n_<su>16): Likewise.
6511 (vmull_lane<q>_<su>16): Likewise.
6512 (vmull_n_<su>16): Likewise.
6513 (vmull_high_lane<q>_<su>16): Likewise.
6514 (vmull_high_n_<su>16): Likewise.
6515 (vqrdmulh<q>_n_s16): Likewise.
6516
6517 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6518
6519 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6520 have the correct lane parameter.
6521
6522 2013-09-06 Richard Biener <rguenther@suse.de>
6523
6524 * cfganal.c (control_dependences::~control_dependences):
6525 Properly free all of the vector.
6526
6527 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6528
6529 PR target/58269
6530 * config/i386/i386.c (ix86_conditional_register_usage):
6531 Proper initialize extended SSE registers.
6532
6533 2013-09-06 Jan Hubicka <jh@suse.cz>
6534
6535 PR tree-optimization/58311
6536 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6537
6538 2013-09-06 Jan Hubicka <jh@suse.cz>
6539
6540 * Makefile.in (tree-sra.o): Update dependencies.
6541 * tree-sra.c: Include ipa-utils.h
6542 (scan_function): Use recursive_call_p.
6543 (has_caller_p): New function.
6544 (cgraph_for_node_and_aliases): Count also callers of aliases.
6545
6546 2013-09-06 Jan Hubicka <jh@suse.cz>
6547
6548 PR middle-end/58094
6549 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6550 * tree-tailcall.c: Include ipa-utils.h.
6551 (find_tail_calls): Use it.
6552 * ipa-pure-const.c (check_call): Likewise.
6553 * ipa-utils.c (recursive_call_p): New function.
6554 * ipa-utils.h (recursive_call_p): Dclare.
6555 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6556 (symtab_semantically_equivalent_p): New function.
6557 * Makefile.in (tree-tailcall.o): Update dependencies.
6558
6559 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6560
6561 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6562 non-inlinable part.
6563
6564 2013-09-06 Richard Biener <rguenther@suse.de>
6565
6566 * lto-streamer.h (lto_global_var_decls): Remove.
6567 * Makefile.in (OBJS): Remove lto-symtab.o.
6568 (lto-symtab.o): Remove.
6569 (GTFILES): Remove lto-symtab.c
6570 * lto-symtab.c: Move to lto/
6571
6572 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6573
6574 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6575 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6576 (UNSPEC_FPINT_RINT): New constant definitions.
6577 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6578 definition with 2 attributes.
6579 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6580 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6581 definitions.
6582
6583 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6584
6585 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6586 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6587 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6588 attribute to "z196_alone".
6589 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6590 "zEC12_simple".
6591
6592 2013-09-06 Richard Biener <rguenther@suse.de>
6593
6594 * basic-block.h (class control_dependences): New.
6595 * tree-ssa-dce.c (control_dependence_map): Remove.
6596 (cd): New global.
6597 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6598 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6599 find_pdom, find_control_dependence, find_all_control_dependences):
6600 Move to cfganal.c.
6601 (mark_control_dependent_edges_necessary,
6602 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6603 tree_dce_done, perform_tree_ssa_dce): Adjust.
6604 * cfganal.c (set_control_dependence_map_bit,
6605 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6606 find_all_control_dependences): Move from tree-ssa-dce.c and
6607 implement as methods of control_dependences class.
6608 (control_dependences::control_dependences): New.
6609 (control_dependences::~control_dependences): Likewise.
6610 (control_dependences::get_edges_dependent_on): Likewise.
6611 (control_dependences::get_edge): Likewise.
6612
6613 2013-09-04 Jan Hubicka <jh@suse.cz>
6614
6615 * tree.c (types_same_for_odr): Drop overactive check.
6616 * ipa-devirt.c (hash_type_name): Likewise.
6617
6618 2013-09-04 Jan Hubicka <jh@suse.cz>
6619
6620 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6621 (analyze_functions): ... here.
6622
6623 2013-09-04 Jan Hubicka <jh@suse.cz>
6624
6625 PR middle-end/58201
6626 * cgraphunit.c (analyze_functions): Clear AUX fields
6627 after processing; initialize assembler name has.
6628
6629 2013-09-05 Jeff Law <law@redhat.com>
6630
6631 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6632 from thread_around_empty_block. Record threading path into PATH.
6633 Recurse if threading through the initial block is successful.
6634 (thread_across_edge): Corresponding changes to slightly simplify.
6635
6636 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6637
6638 * config/aarch64/aarch64.md
6639 (type): Remove frecpe, frecps, frecpx.
6640 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6641 fix to be a TARGET_SIMD instruction.
6642 (aarch64_frecps): Remove.
6643 * config/aarch64/aarch64-simd.md
6644 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6645 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6646
6647 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6648 Sofiane Naci <sofiane.naci@arm.com>
6649
6650 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6651 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6652 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6653 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6654 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6655 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6656 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6657 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6658 into "clz, "rbit". Rename "shift" to "shift_imm".
6659 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6660 changes. Update for attribute changes all occurrences of arlo_* and
6661 shift* types.
6662 * config/arm/arm-fixed.md: Update for attribute changes
6663 all occurrences of arlo_* types.
6664 * config/arm/thumb2.md: Update for attribute changes all occurrences
6665 of arlo_* types.
6666 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6667 (cortexa7_older_only): Likewise.
6668 (cortexa7_younger): Likewise.
6669 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6670 (1020alu_shift_op): Likewise.
6671 (1020alu_shift_reg_op): Likewise.
6672 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6673 (alu_shift_op): Likewise.
6674 (alu_shift_reg_op): Likewise.
6675 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6676 (11_alu_shift_op): Likewise.
6677 (11_alu_shift_reg_op): Likewise.
6678 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6679 (9_alu_shift_reg_op): Likewise.
6680 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6681 attribute changes.
6682 (cortex_a15_alu_shift): Likewise.
6683 (cortex_a15_alu_shift_reg): Likewise.
6684 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6685 attribute changes.
6686 (cortex_a5_alu_shift): Likewise.
6687 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6688 attribute changes.
6689 (cortex_a53_alu_shift): Likewise.
6690 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6691 attribute changes.
6692 (cortex_a7_alu_reg): Likewise.
6693 (cortex_a7_alu_shift): Likewise.
6694 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6695 attribute changes.
6696 (cortex_a8_alu_shift): Likewise.
6697 (cortex_a8_alu_shift_reg): Likewise.
6698 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6699 (cortex_a9_dp_shift): Likewise.
6700 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6701 attribute changes.
6702 * config/arm/cortex-r4.md
6703 (cortex_r4_alu): Update for attribute changes.
6704 (cortex_r4_mov): Likewise.
6705 (cortex_r4_alu_shift_reg): Likewise.
6706 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6707 (526_alu_shift_op): Likewise.
6708 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6709 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6710 (626te_alu_shift_op): Likewise.
6711 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6712 (726te_alu_shift_op): Likewise.
6713 (726te_alu_shift_reg_op): Likewise.
6714 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6715 (mp626_alu_shift_op): Likewise.
6716 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6717 (pj4_alu_conds): Likewise.
6718 (pj4_shift): Likewise.
6719 (pj4_shift_conds): Likewise.
6720 (pj4_alu_shift): Likewise.
6721 (pj4_alu_shift_conds): Likewise.
6722 * config/aarch64/aarch64.md: Update for attribute change
6723 all occurrences of arlo_* and shift* types.
6724
6725 2013-09-05 Mike Stump <mikestump@comcast.net>
6726
6727 * tree.h: Move documentation for tree_function_decl to tree-core.h
6728 with the declaration.
6729
6730 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6731
6732 PR target/58139
6733 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6734 looking for widest mode.
6735
6736 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6737
6738 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6739
6740 2013-09-05 Richard Biener <rguenther@suse.de>
6741
6742 PR tree-optimization/58137
6743 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6744 Do not create vectors of pointers.
6745 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6746 types for the components of the vector initializer.
6747 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6748 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6749
6750 2013-09-05 Martin Jambor <mjambor@suse.cz>
6751
6752 * ipa-prop.c (remove_described_reference): Accept missing references,
6753 return false if that hppens, otherwise return true.
6754 (cgraph_node_for_jfunc): New function.
6755 (try_decrement_rdesc_refcount): Likewise.
6756 (try_make_edge_direct_simple_call): Use them.
6757 (ipa_edge_removal_hook): Remove references from rdescs.
6758 (ipa_edge_duplication_hook): Clone rdescs and their references
6759 when the new edge has the same caller as the old one.
6760 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6761 reference before removing any edges.
6762
6763 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6764
6765 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6766 initial store.
6767 * thumb2.md (thumb2_storewb_parisi): New pattern.
6768
6769 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6770
6771 * config/aarch64/aarch64-option-extensions.def: Add
6772 AARCH64_OPT_EXTENSION of 'crc'.
6773 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6774 (AARCH64_ISA_CRC): Ditto.
6775 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6776 description of the CRC extension.
6777
6778 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6779
6780 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6781 * config/rs6000/linux.h: Ditto.
6782 * alpha/linux.h: Ditto.
6783 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6784 no_c99_libc_has_function.
6785 * config/c6x/uclinux-elf.h: Ditto.
6786 * config/lm32/uclinux-elf.h: Ditto.
6787 * config/m68k/uclinux.h: Ditto.
6788 * config/moxie/uclinux.h: Ditto.
6789 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6790 (crisv32-*-linux*, cris-*-linux*): Ditto.
6791 * config/bfin/bfin.c: Include "tm_p.h".
6792
6793 2013-09-05 Richard Biener <rguenther@suse.de>
6794
6795 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6796 check for a definition without a basic-block.
6797
6798 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6799 Sofiane Naci <sofiane.naci@arm.com>
6800
6801 * config/aarch64/aarch64.md
6802 (*movti_aarch64): Rename r_2_f and f_2_r.
6803 (*movsf_aarch64): Likewise.
6804 (*movdf_aarch64): Likewise.
6805 (*movtf_aarch64): Likewise.
6806 (aarch64_movdi_<mode>low): Likewise.
6807 (aarch64_movdi_<mode>high): Likewise.
6808 (aarch64_mov<mode>high_di): Likewise.
6809 (aarch64_mov<mode>low_di): Likewise.
6810 (aarch64_movtilow_tilow): Likewise.
6811 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6812 values to config/arm/types.md
6813 (attribute "conds"): Update for attribute change.
6814 (anddi3_insn): Likewise.
6815 (iordi3_insn): Likewise.
6816 (xordi3_insn): Likewise.
6817 (one_cmpldi2): Likewise.
6818 * config/arm/types.md (type): Add Neon types.
6819 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6820 use "type" attribute.
6821 (movmisalign<mode>_neon_store): Likewise.
6822 (movmisalign<mode>_neon_load): Likewise.
6823 (vec_set<mode>_internal): Likewise.
6824 (vec_setv2di_internal): Likewise.
6825 (vec_extract<mode>): Likewise.
6826 (vec_extractv2di): Likewise.
6827 (add<mode>3_neon): Likewise.
6828 (adddi3_neon): Likewise.
6829 (sub<mode>3_neon): Likewise.
6830 (subdi3_neon): Likewise.
6831 (mul<mode>3_neon): Likewise.
6832 (mul<mode>3add<mode>_neon): Likewise.
6833 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6834 (fma<VCVTF:mode>4)): Likewise.
6835 (fma<VCVTF:mode>4_intrinsic): Likewise.
6836 (fmsub<VCVTF:mode>4)): Likewise.
6837 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6838 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6839 (ior<mode>3): Likewise.
6840 (and<mode>3): Likewise.
6841 (anddi3_neon): Likewise.
6842 (orn<mode>3_neon): Likewise.
6843 (orndi3_neon): Likewise.
6844 (bic<mode>3_neon): Likewise.
6845 (bicdi3_neon): Likewise.
6846 (xor<mode>3): Likewise.
6847 (one_cmpl<mode>2): Likewise.
6848 (abs<mode>2): Likewise.
6849 (neg<mode>2): Likewise.
6850 (umin<mode>3_neon): Likewise.
6851 (umax<mode>3_neon): Likewise.
6852 (smin<mode>3_neon): Likewise.
6853 (smax<mode>3_neon): Likewise.
6854 (vashl<mode>3): Likewise.
6855 (vashr<mode>3_imm): Likewise.
6856 (vlshr<mode>3_imm): Likewise.
6857 (ashl<mode>3_signed): Likewise.
6858 (ashl<mode>3_unsigned): Likewise.
6859 (neon_load_count): Likewise.
6860 (ashldi3_neon_noclobber): Likewise.
6861 (signed_shift_di3_neon): Likewise.
6862 (unsigned_shift_di3_neon): Likewise.
6863 (ashrdi3_neon_imm_noclobber): Likewise.
6864 (lshrdi3_neon_imm_noclobber): Likewise.
6865 (widen_ssum<mode>3): Likewise.
6866 (widen_usum<mode>3): Likewise.
6867 (quad_halves_<code>v4si): Likewise.
6868 (quad_halves_<code>v4sf): Likewise.
6869 (quad_halves_<code>v8hi): Likewise.
6870 (quad_halves_<code>v16qi): Likewise.
6871 (reduc_splus_v2di): Likewise.
6872 (neon_vpadd_internal<mode>): Likewise.
6873 (neon_vpsmin<mode>): Likewise.
6874 (neon_vpsmax<mode>): Likewise.
6875 (neon_vpumin<mode>): Likewise.
6876 (neon_vpumax<mode>): Likewise.
6877 (ss_add<mode>_neon): Likewise.
6878 (us_add<mode>_neon): Likewise.
6879 (ss_sub<mode>_neon): Likewise.
6880 (us_sub<mode>_neon): Likewise.
6881 (neon_vadd<mode>_unspec): Likewise.
6882 (neon_vaddl<mode>): Likewise.
6883 (neon_vaddw<mode>): Likewise.
6884 (neon_vhadd<mode>): Likewise.
6885 (neon_vqadd<mode>): Likewise.
6886 (neon_vaddhn<mode>): Likewise.
6887 (neon_vmul<mode>): Likewise.
6888 (neon_vmla<mode>): Likewise.
6889 (neon_vmlal<mode>): Likewise.
6890 (neon_vmls<mode>): Likewise.
6891 (neon_vmlsl<mode>): Likewise.
6892 (neon_vqdmulh<mode>): Likewise.
6893 (neon_vqdmlal<mode>): Likewise.
6894 (neon_vqdmlsl<mode>): Likewise.
6895 (neon_vmull<mode>): Likewise.
6896 (neon_vqdmull<mode>): Likewise.
6897 (neon_vsub<mode>_unspec): Likewise.
6898 (neon_vsubl<mode>): Likewise.
6899 (neon_vsubw<mode>): Likewise.
6900 (neon_vqsub<mode>): Likewise.
6901 (neon_vhsub<mode>): Likewise.
6902 (neon_vsubhn<mode>): Likewise.
6903 (neon_vceq<mode>): Likewise.
6904 (neon_vcge<mode>): Likewise.
6905 (neon_vcgeu<mode>): Likewise.
6906 (neon_vcgt<mode>): Likewise.
6907 (neon_vcgtu<mode>): Likewise.
6908 (neon_vcle<mode>): Likewise.
6909 (neon_vclt<mode>): Likewise.
6910 (neon_vcage<mode>): Likewise.
6911 (neon_vcagt<mode>): Likewise.
6912 (neon_vtst<mode>): Likewise.
6913 (neon_vabd<mode>): Likewise.
6914 (neon_vabdl<mode>): Likewise.
6915 (neon_vaba<mode>): Likewise.
6916 (neon_vabal<mode>): Likewise.
6917 (neon_vmax<mode>): Likewise.
6918 (neon_vmin<mode>): Likewise.
6919 (neon_vpaddl<mode>): Likewise.
6920 (neon_vpadal<mode>): Likewise.
6921 (neon_vpmax<mode>): Likewise.
6922 (neon_vpmin<mode>): Likewise.
6923 (neon_vrecps<mode>): Likewise.
6924 (neon_vrsqrts<mode>): Likewise.
6925 (neon_vqabs<mode>): Likewise.
6926 (neon_vqneg<mode>): Likewise.
6927 (neon_vcls<mode>): Likewise.
6928 (clz<mode>2): Likewise.
6929 (popcount<mode>2): Likewise.
6930 (neon_vrecpe): Likewise.
6931 (neon_vrsqrte): Likewise.
6932 (neon_vget_lane<mode>_sext_internal): Likewise.
6933 (neon_vget_lane<mode>_zext_internal): Likewise.
6934 (neon_vdup_n<mode>): Likewise.
6935 (neon_vdup_nv2di): Likewise.
6936 (neon_vdpu_lane<mode>_internal): Likewise.
6937 (neon_vswp<mode>): Likewise.
6938 (float<mode><V_cvtto>2): Likewise.
6939 (floatuns<mode><V_cvtto>2): Likewise.
6940 (fix_trunc<mode><V_cvtto>)2): Likewise
6941 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6942 (neon_vcvt<mode>): Likewise.
6943 (neon_vcvtv4sfv4hf): Likewise.
6944 (neon_vcvtv4hfv4sf): Likewise.
6945 (neon_vcvt_n<mode>): Likewise.
6946 (neon_vmovn<mode>): Likewise.
6947 (neon_vqmovn<mode>): Likewise.
6948 (neon_vqmovun<mode>): Likewise.
6949 (neon_vmovl<mode>): Likewise.
6950 (neon_vmul_lane<mode>): Likewise.
6951 (neon_vmull_lane<mode>): Likewise.
6952 (neon_vqdmull_lane<mode>): Likewise.
6953 (neon_vqdmulh_lane<mode>): Likewise.
6954 (neon_vmla_lane<mode>): Likewise.
6955 (neon_vmlal_lane<mode>): Likewise.
6956 (neon_vqdmlal_lane<mode>): Likewise.
6957 (neon_vmls_lane<mode>): Likewise.
6958 (neon_vmlsl_lane<mode>): Likewise.
6959 (neon_vqdmlsl_lane<mode>): Likewise.
6960 (neon_vext<mode>): Likewise.
6961 (neon_vrev64<mode>): Likewise.
6962 (neon_vrev32<mode>): Likewise.
6963 (neon_vrev16<mode>): Likewise.
6964 (neon_vbsl<mode>_internal): Likewise.
6965 (neon_vshl<mode>): Likewise.
6966 (neon_vqshl<mode>): Likewise.
6967 (neon_vshr_n<mode>): Likewise.
6968 (neon_vshrn_n<mode>): Likewise.
6969 (neon_vqshrn_n<mode>): Likewise.
6970 (neon_vqshrun_n<mode>): Likewise.
6971 (neon_vshl_n<mode>): Likewise.
6972 (neon_vqshl_n<mode>): Likewise.
6973 (neon_vqshlu_n<mode>): Likewise.
6974 (neon_vshll_n<mode>): Likewise.
6975 (neon_vsra_n<mode>): Likewise.
6976 (neon_vsri_n<mode>): Likewise.
6977 (neon_vsli_n<mode>): Likewise.
6978 (neon_vtbl1v8qi): Likewise.
6979 (neon_vtbl2v8qi): Likewise.
6980 (neon_vtbl3v8qi): Likewise.
6981 (neon_vtbl4v8qi): Likewise.
6982 (neon_vtbx1v8qi): Likewise.
6983 (neon_vtbx2v8qi): Likewise.
6984 (neon_vtbx3v8qi): Likewise.
6985 (neon_vtbx4v8qi): Likewise.
6986 (neon_vtrn<mode>_internal): Likewise.
6987 (neon_vzip<mode>_internal): Likewise.
6988 (neon_vuzp<mode>_internal): Likewise.
6989 (neon_vld1<mode>): Likewise.
6990 (neon_vld1_lane<mode>): Likewise.
6991 (neon_vld1_dup<mode>): Likewise.
6992 (neon_vld1_dupv2di): Likewise.
6993 (neon_vst1<mode>): Likewise.
6994 (neon_vst1_lane<mode>): Likewise.
6995 (neon_vld2<mode>): Likewise.
6996 (neon_vld2_lane<mode>): Likewise.
6997 (neon_vld2_dup<mode>): Likewise.
6998 (neon_vst2<mode>): Likewise.
6999 (neon_vst2_lane<mode>): Likewise.
7000 (neon_vld3<mode>): Likewise.
7001 (neon_vld3qa<mode>): Likewise.
7002 (neon_vld3qb<mode>): Likewise.
7003 (neon_vld3_lane<mode>): Likewise.
7004 (neon_vld3_dup<mode>): Likewise.
7005 (neon_vst3<mode>): Likewise.
7006 (neon_vst3qa<mode>): Likewise.
7007 (neon_vst3qb<mode>): Likewise.
7008 (neon_vst3_lane<mode>): Likewise.
7009 (neon_vld4<mode>): Likewise.
7010 (neon_vld4qa<mode>): Likewise.
7011 (neon_vld4qb<mode>): Likewise.
7012 (neon_vld4_lane<mode>): Likewise.
7013 (neon_vld4_dup<mode>): Likewise.
7014 (neon_vst4<mode>): Likewise.
7015 (neon_vst4qa<mode>): Likewise.
7016 (neon_vst4qb<mode>): Likewise.
7017 (neon_vst4_lane<mode>): Likewise.
7018 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7019 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7020 (neon_vec_<US>mult_lo_<mode>): Likewise.
7021 (neon_vec_<US>mult_hi_<mode>): Likewise.
7022 (neon_vec_<US>shiftl_<mode>): Likewise.
7023 (neon_unpack<US>_<mode>): Likewise.
7024 (neon_vec_<US>mult_<mode>): Likewise.
7025 (vec_pack_trunc_<mode>): Likewise.
7026 (neon_vec_pack_trunk_<mode>): Likewise.
7027 (neon_vabd<mode>_2): Likewise.
7028 (neon_vabd<mode>_3): Likewise.
7029 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7030 (thumb2_movsi_vfp): Likewise.
7031 (movdi_vfp): Likewise.
7032 (movdi_vfp_cortexa8): Likewise.
7033 (movhf_vfp_neon): Likewise.
7034 (movhf_vfp): Likewiwse.
7035 (movsf_vfp): Likewiwse.
7036 (thumb2_movsf_vfp): Likewiwse.
7037 (movdf_vfp): Likewise.
7038 (thumb2_movdf_vfp): Likewise.
7039 (movsfcc_vfp): Likewise.
7040 (thumb2_movsfcc_vfp): Likewise.
7041 (movdfcc_vfp): Likewise.
7042 (thumb2_movdfcc_vfp): Likewise.
7043 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7044 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7045 (v10_v2c): Likewise.
7046 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7047 attribute change.
7048 (cortex_a15_neon_int_2): Likewise.
7049 (cortex_a15_neon_int_3): Likewise.
7050 (cortex_a15_neon_int_4): Likewise.
7051 (cortex_a15_neon_int_5): Likewise.
7052 (cortex_a15_neon_vqneg_vqabs): Likewise.
7053 (cortex_a15_neon_vmov): Likewise.
7054 (cortex_a15_neon_vaba): Likewise.
7055 (cortex_a15_neon_vaba_qqq): Likewise.
7056 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7057 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7058 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7059 scalar_64_32_long_scalar): Likewise.
7060 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7061 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7062 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7063 lotype_qdd_64_32_long): Likewise.
7064 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7065 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7066 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7067 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7068 (cortex_a15_neon_shift_1): Likewise.
7069 (cortex_a15_neon_shift_2): Likewise.
7070 (cortex_a15_neon_shift_3): Likewise.
7071 (cortex_a15_neon_vshl_ddd): Likewise.
7072 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7073 (cortex_a15_neon_vsra_vrsra): Likewise.
7074 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7075 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7076 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7077 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7078 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7079 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7080 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7081 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7082 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7083 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7084 (cortex_a15_neon_bp_simple): Likewise.
7085 (cortex_a15_neon_bp_2cycle): Likewise.
7086 (cortex_a15_neon_bp_3cycle): Likewise.
7087 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7088 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7089 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7090 (cortex_a15_neon_vld2_4_regs): Likewise.
7091 (cortex_a15_neon_vld3_vld4): Likewise.
7092 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7093 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7094 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7095 (cortex_a15_neon_vst3_vst4): Likewise.
7096 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7097 (cortex_a15_neon_vld3_vld4_lane" 10
7098 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7099 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7100 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7101 (cortex_a15_neon_ldm_2): Likewise.0
7102 (cortex_a15_neon_stm_2): Likewise.
7103 (cortex_a15_neon_mcr): Likewise.
7104 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7105 (cortex_a15_neon_mrc): Likewise.
7106 (cortex_a15_neon_mrrc): Likewise.
7107 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7108 change.
7109 (cortex_a15_alu_shift): Likewise.
7110 (cortex_a15_alu_shift_reg): Likewise.
7111 (cortex_a15_mult32): Likewise.
7112 (cortex_a15_mult64): Likewise.
7113 (cortex_a15_block): Likewise.
7114 (cortex_a15_branch): Likewise.
7115 (cortex_a15_load1): Likewise.
7116 (cortex_a15_load3): Likewise.
7117 (cortex_a15_store1): Likewise.
7118 (cortex_a15_store3): Likewise.
7119 (cortex_a15_call): Likewise.
7120 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7121 (cortex_a5_f2r): Likewise.
7122 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7123 change.
7124 (cortex_a53_f2r): Likewise.
7125 * config/arm/cortex-a7.md
7126 (cortex_a7_branch): Update for attribute change.
7127 (cortex_a7_call): Likewise.
7128 (cortex_a7_alu_imm): Likewise.
7129 (cortex_a7_alu_reg): Likewise.
7130 (cortex_a7_alu_shift): Likewise.
7131 (cortex_a7_mul): Likewise.
7132 (cortex_a7_load1): Likewise.
7133 (cortex_a7_store1): Likewise.
7134 (cortex_a7_load2): Likewise.
7135 (cortex_a7_store2): Likewise.
7136 (cortex_a7_load3): Likewise.
7137 (cortex_a7_store3): Likewise.
7138 (cortex_a7_load4): Likewise.
7139 (cortex_a7_store4): Likewise.
7140 (cortex_a7_fpalu): Likewise.
7141 (cortex_a7_fconst): Likewise.
7142 (cortex_a7_fpmuls): Likewise.
7143 (cortex_a7_neon_mul): Likewise.
7144 (cortex_a7_fpmacs): Likewise.
7145 (cortex_a7_neon_mla: Likewise.
7146 (cortex_a7_fpmuld: Likewise.
7147 (cortex_a7_fpmacd: Likewise.
7148 (cortex_a7_fpfmad: Likewise.
7149 (cortex_a7_fdivs: Likewise.
7150 (cortex_a7_fdivd: Likewise.
7151 (cortex_a7_r2f: Likewise.
7152 (cortex_a7_f2r: Likewise.
7153 (cortex_a7_f_flags: Likewise.
7154 (cortex_a7_f_loads: Likewise.
7155 (cortex_a7_f_loadd: Likewise.
7156 (cortex_a7_f_stores: Likewise.
7157 (cortex_a7_f_stored: Likewise.
7158 (cortex_a7_neon): Likewise.
7159 * config/arm/cortex-a8-neon.md
7160 (cortex_a8_neon_mrc): Update for attribute change.
7161 (cortex_a8_neon_mrrc): Likewise.
7162 (cortex_a8_neon_int_1): Likewise.
7163 (cortex_a8_neon_int_2): Likewise.
7164 (cortex_a8_neon_int_3): Likewise.
7165 (cortex_a8_neon_int_4): Likewise.
7166 (cortex_a8_neon_int_5): Likewise.
7167 (cortex_a8_neon_vqneg_vqabs): Likewise.
7168 (cortex_a8_neon_vmov): Likewise.
7169 (cortex_a8_neon_vaba): Likewise.
7170 (cortex_a8_neon_vaba_qqq): Likewise.
7171 (cortex_a8_neon_vsma): Likewise.
7172 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7173 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7174 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7175 long_scalar): Likewise.
7176 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7177 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7178 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7179 long_scalar_qdd_64_32_long): Likewise.
7180 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7181 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7182 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7183 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7184 (cortex_a8_neon_shift_1): Likewise.
7185 (cortex_a8_neon_shift_2): Likewise.
7186 (cortex_a8_neon_shift_3): Likewise.
7187 (cortex_a8_neon_vshl_ddd): Likewise.
7188 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7189 (cortex_a8_neon_vsra_vrsra): Likewise.
7190 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7191 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7192 (cortex_a8_neon_fp_vsum): Likewise.
7193 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7194 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7195 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7196 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7197 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7198 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7199 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7200 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7201 (cortex_a8_neon_bp_simple): Likewise.
7202 (cortex_a8_neon_bp_2cycle): Likewise.
7203 (cortex_a8_neon_bp_3cycle): Likewise.
7204 (cortex_a8_neon_ldr): Likewise.
7205 (cortex_a8_neon_str): Likewise.
7206 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7207 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7208 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7209 (cortex_a8_neon_vld2_4_regs): Likewise.
7210 (cortex_a8_neon_vld3_vld4): Likewise.
7211 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7212 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7213 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7214 (cortex_a8_neon_vst3_vst4): Likewise.
7215 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7216 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7217 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7218 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7219 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7220 (cortex_a8_neon_mcr): Likewise.
7221 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7222 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7223 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7224 change.
7225 (ca9_neon_mrrc): Likewise.
7226 (cortex_a9_neon_int_1): Likewise.
7227 (cortex_a9_neon_int_2): Likewise.
7228 (cortex_a9_neon_int_3): Likewise.
7229 (cortex_a9_neon_int_4): Likewise.
7230 (cortex_a9_neon_int_5): Likewise.
7231 (cortex_a9_neon_vqneg_vqabs): Likewise.
7232 (cortex_a9_neon_vmov): Likewise.
7233 (cortex_a9_neon_vaba): Likewise.
7234 (cortex_a9_neon_vaba_qqq): Likewise.
7235 (cortex_a9_neon_vsma): Likewise.
7236 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7237 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7238 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7239 long_scalar): Likewise.
7240 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7241 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7242 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7243 long_scalar_qdd_64_32_long): Likewise.
7244 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7245 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7246 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7247 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7248 (cortex_a9_neon_shift_1): Likewise.
7249 (cortex_a9_neon_shift_2): Likewise.
7250 (cortex_a9_neon_shift_3): Likewise.
7251 (cortex_a9_neon_vshl_ddd): Likewise.
7252 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7253 (cortex_a9_neon_vsra_vrsra): Likewise.
7254 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7255 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7256 (cortex_a9_neon_fp_vsum): Likewise.
7257 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7258 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7259 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7260 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7261 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7262 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7263 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7264 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7265 (cortex_a9_neon_bp_simple): Likewise.
7266 (cortex_a9_neon_bp_2cycle): Likewise.
7267 (cortex_a9_neon_bp_3cycle): Likewise.
7268 (cortex_a9_neon_ldr): Likewise.
7269 (cortex_a9_neon_str): Likewise.
7270 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7271 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7272 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7273 (cortex_a9_neon_vld2_4_regs): Likewise.
7274 (cortex_a9_neon_vld3_vld4): Likewise.
7275 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7276 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7277 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7278 (cortex_a9_neon_vst3_vst4): Likewise.
7279 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7280 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7281 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7282 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7283 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7284 (cortex_a9_neon_mcr): Likewise.
7285 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7286 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7287 (cortex_a9_fps): Likewise.
7288 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7289 change.
7290 (cortex_m4_fmuls): Likewise.
7291 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7292 change.
7293 (cortex_r4_mrc): Likewise.
7294 * config/arm/iterators.md: Update comment referring to neon_type.
7295 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7296 (iwmmxt_movsi_insn): Likewise.
7297 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7298 attribute change.
7299 (pj4_core_to_vfp): Likewise.
7300 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7301 attribute change.
7302 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7303 (vfp_fstore): Likewise.
7304 * doc/md.texi: Change references to neon_type to refer to type.
7305
7306 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7307
7308 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7309
7310 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7311
7312 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7313 lglobl if not weak.
7314
7315 2013-09-04 Easwaran Raman <eraman@google.com>
7316
7317 PR middle-end/57370
7318 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7319 (build_and_add_sum): Use it.
7320 (appears_later_in_bb): Simplify code.
7321
7322 2013-09-04 Teresa Johnson <tejohnson@google.com>
7323
7324 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7325
7326 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7327
7328 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7329
7330 2013-09-04 Jan Hubicka <jh@suse.cz>
7331
7332 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7333 * ipa-devirt.c: Include diganostic.h
7334 (odr_type_d): Add types and types_set.
7335 (hash_type_name): Work for types with vtables during LTO.
7336 (odr_hasher::remove): Fix comment; destroy types_set.
7337 (add_type_duplicate): New function,
7338 (get_odr_type): Use it.
7339 (dump_type_inheritance_graph): Dump type duplicates.
7340 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7341 graph.
7342 * tree.c (types_same_for_odr): Give exact answers on types with
7343 virtual tables.
7344
7345 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7346
7347 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7348 explaining their differences.
7349
7350 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7351
7352 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7353
7354 2013-09-03 Jeff Law <law@redhat.com>
7355
7356 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7357 when not threading through a joiner block. Pass joiner/no joiner
7358 state to register_jump_thread.
7359 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7360 state from argument rather than implying on path length.
7361 Dump the entire jump thread path into debugging dump.
7362 * tree-flow.h (register_jump_thread): Update prototype.
7363
7364 2013-08-29 Xinliang David Li <davidxl@google.com>
7365
7366 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7367 Remove a trivial gcc_assert.
7368
7369 2013-08-29 Xinliang David Li <davidxl@google.com>
7370
7371 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7372 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7373 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7374 Delay base decl alignment adjustment.
7375 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7376 * tree-vectorizer.h: New data structure.
7377 (set_dr_misalignment): New function.
7378 (dr_misalignment): Ditto.
7379 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7380 (vectorizable_load): Ditto.
7381 (ensure_base_align): New function.
7382 (vectorize_loops): Add dbg_cnt support.
7383 (execute_vect_slp): Ditto.
7384 * dbgcnt.def: New debug counter.
7385 * Makefile: New dependency.
7386
7387 2013-09-03 Meador Inge <meadori@codesourcery.com>
7388
7389 Revert:
7390
7391 2013-08-30 Meador Inge <meadori@codesourcery.com>
7392
7393 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7394
7395 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7396
7397 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7398 function descriptor.
7399
7400 2013-09-03 Richard Biener <rguenther@suse.de>
7401
7402 * tree-affine.c (add_elt_to_tree): Fix association issue,
7403 avoid useless converts and make sure to always return a
7404 properly typed result.
7405
7406 2013-09-03 Richard Biener <rguenther@suse.de>
7407
7408 PR middle-end/57656
7409 * fold-const.c (negate_expr_p): Fix division case.
7410 (negate_expr): Likewise.
7411
7412 2013-09-03 Richard Biener <rguenther@suse.de>
7413
7414 PR lto/58285
7415 * tree-streamer-out.c: Include tm.h.
7416 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7417
7418 2013-09-03 Jan Hubicka <jh@suse.cz>
7419
7420 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7421
7422 2013-09-03 Alan Modra <amodra@gmail.com>
7423
7424 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7425 multilibs to big-endian target and vice versa.
7426 * config/rs6000/t-linux64: Use := assignment on all vars.
7427 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7428 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7429 * config/rs6000/t-linux64le: New file.
7430 * config/rs6000/t-linux64bele: New file.
7431 * config/rs6000/t-linux64lebe: New file.
7432
7433 2013-09-02 Jan Hubicka <jh@suse.cz>
7434
7435 * ipa-inline-transform.c (inline_transform): Do not
7436 optimize_inline_calls when not optimizing.
7437
7438 2013-09-02 Jan Hubicka <jh@suse.cz>
7439
7440 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7441 duplicated nodes for assembler names.
7442 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7443 hard registers from assembler name hash.
7444
7445 2013-09-02 Jan Hubicka <jh@suse.cz>
7446
7447 * ipa-split.c (execute_split_functions): Split externally visible
7448 functions called once.
7449
7450 2013-09-02 Martin Jambor <mjambor@suse.cz>
7451
7452 PR ipa/58106
7453 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7454 linked list. When finding the correct duplicate, also consider also
7455 the caller in additon to its inlined_to node.
7456
7457 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7458
7459 * config/aarch64/aarch64-simd-builtins.def
7460 (dup_lane_scalar): Remove.
7461 * config/aarch64/aarch64-simd.md
7462 (aarch64_simd_dup): Add 'w->w' alternative.
7463 (aarch64_dup_lane<mode>): Allow for VALL.
7464 (aarch64_dup_lane_scalar<mode>): Remove.
7465 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7466 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7467 (aarch64_get_lane_unsigned<mode>): Likewise.
7468 (aarch64_get_lane<mode>): Likewise.
7469 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7470 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7471 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7472 (VCON): Change container of V2SF.
7473 (vswap_width_name): Likewise.
7474 * config/aarch64/arm_neon.h
7475 (__aarch64_vdup_lane_any): New.
7476 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7477 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7478 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7479
7480 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7481
7482 PR middle-end/56382
7483 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7484 the source or the destination is a single hard register.
7485
7486 2013-09-02 Richard Biener <rguenther@suse.de>
7487
7488 PR middle-end/57511
7489 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7490 non-linear SCEVs.
7491
7492 2013-09-02 Richard Biener <rguenther@suse.de>
7493
7494 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7495 arithmetic to sizetype.
7496
7497 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7498
7499 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7500 Find auto-increment use both before and after candidate.
7501
7502 2013-09-02 Marek Polacek <polacek@redhat.com>
7503
7504 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7505
7506 2013-09-01 Jan Hubicka <jh@suse.cz>
7507
7508 * Makefile.in: Add ipa-profile.o
7509 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7510 * cgraph.c (struct cgraph_propagate_frequency_data,
7511 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7512 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7513 * cgraph.h (cgraph_propagate_frequency): Remove.
7514 * ipa-inline-analysis.c: Include ipa-utils.h;
7515 drop duplicated cfgloop.h.
7516 (inline_update_callee_summaries): Update.
7517 * ipa-profile.c: New file.
7518 * ipa-utils.h (ipa_propagate_frequency): Declare.
7519 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7520 data-streamer.h, value-prof.h.
7521 (symtab_remove_unreachable_nodes): Update profile.
7522 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7523 account_time_size, cmp_counts, dump_histogram,
7524 ipa_profile_generate_summary, ipa_profile_write_summary,
7525 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7526 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7527 Move to ipa-profile.c.
7528
7529 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7530
7531 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7532
7533 2013-09-01 Jan Hubicka <jh@suse.cz>
7534
7535 * common.opt (fdevirtualize-speculatively): New function.
7536 * invoke.texi (fdevirtualize-speculatively): Document.
7537 * ipa-devirt.c: Include ipa-inline.h
7538 (likely_target_p): New function.
7539 (ipa_devirt): New function.
7540 (gate_ipa_devirt): New function.
7541 (pass_data_ipa_devirt): New static var.
7542 (pass_ipa_devirt): Likewise.
7543 (make_pass_ipa_devirt): New function.
7544 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7545 (common_handle_option): Disable devirtualization when
7546 value range profiling is available.
7547 * passes.def (pass_ipa_devirt): Add.
7548 * timever.def (TV_IPA_DEVIRT): New timevar.
7549 * tree-pass.h (make_pass_ipa_devirt):
7550
7551 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7552
7553 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7554 include sanitize(undefined).
7555
7556 2013-08-31 Diego Novillo <dnovillo@google.com>
7557
7558 * Makefile.in (TREE_CORE_H): Define.
7559 (TREE_H): Use.
7560 (GTFILES): Add tree-core.h.
7561 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7562 size the array.
7563 * tree-core.h: New file.
7564 Move all data structures, enum, typedefs, global
7565 declarations and constants from ...
7566 * tree.h: ... here.
7567
7568 2013-08-31 Jan Hubicka <jh@suse.cz>
7569
7570 * bulitins.c (expand_builtin): Do not early exit for gcov
7571 instrumented functions.
7572
7573 2013-08-31 Marek Polacek <polacek@redhat.com>
7574
7575 * ubsan.c: Include tm_p.h.
7576
7577 2013-08-31 Jan Hubicka <jh@suse.cz>
7578
7579 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7580 warning.
7581
7582 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7583 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7584 * tree-cfg.c (verify_gimple_label): ... here.
7585 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7586 (ipa_merge_profiles): New function.
7587 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7588 (lto_input_function_body): Likewise.
7589 * ipa-utils.h (ipa_merge_profiles): Declare.
7590 * lto-streamer.h (lto_input_function_body): Update prototype.
7591 (emit_label_in_global_context_p): Remove.
7592 * lto-symtab.c: Include ipa-utils.h
7593 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7594
7595 2013-08-31 Jan Hubicka <jh@suse.cz>
7596
7597 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7598
7599 2013-08-31 Jan Hubicka <jh@suse.cz>
7600
7601 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7602
7603 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
7604
7605 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7606 "cmp" RTX before signed_comparison_operator check to account
7607 for "code" changes.
7608
7609 2013-08-30 Jan Hubicka <jh@suse.cz>
7610
7611 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7612 (detect_type_change_1): Rename to ...
7613 (detect_type_change): ... this one; early return on non-polymorphic
7614 types.
7615 (detect_type_change_ssa): Add comp_type parameter; update
7616 use of detect_type_change.
7617 (compute_complex_assign_jump_func): Add param_type parameter;
7618 update use of detect_type_change_ssa.
7619 (compute_complex_ancestor_jump_func): Likewise.
7620 (ipa_get_callee_param_type): New function.
7621 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7622 update calls to the jump function computation functions.
7623
7624 2013-08-30 Teresa Johnson <tejohnson@google.com>
7625 Steven Bosscher <steven@gcc.gnu.org>
7626
7627 * cfgrtl.c (fixup_new_cold_bb): New routine.
7628 (commit_edge_insertions): Invoke fixup_partitions.
7629 (find_partition_fixes): New routine.
7630 (fixup_partitions): Ditto.
7631 (verify_hot_cold_block_grouping): Update comments.
7632 (rtl_verify_edges): Invoke find_partition_fixes.
7633 (rtl_verify_bb_pointers): Update comments.
7634 (rtl_verify_bb_layout): Ditto.
7635 * basic-block.h (probably_never_executed_edge_p): Declare.
7636 (fixup_partitions): Ditto.
7637 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7638 * bb-reorder.c (sanitize_hot_paths): New function.
7639 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7640 sanitize_hot_paths.
7641 * predict.c (probably_never_executed_edge_p): New routine.
7642 * cfg.c (check_bb_profile): Add partition insanity warnings.
7643
7644 2013-08-30 Meador Inge <meadori@codesourcery.com>
7645
7646 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7647
7648 2013-08-30 Marek Polacek <polacek@redhat.com>
7649
7650 * Makefile.in (ubsan.o): Add.
7651 (c-family/c-ubsan.o): Add.
7652 (builtins.o): Add ubsan.h dependency.
7653 * ubsan.h: New file.
7654 * ubsan.c: New file.
7655 * common.opt: Add -fsanitize=undefined option.
7656 (flag_sanitize): Add variable.
7657 (fsanitize=): Add option. Add Driver.
7658 (fsanitize=thread): Remove option.
7659 (fsanitize=address): Likewise.
7660 (static-libubsan): New option.
7661 * doc/invoke.texi: Document the new flag and -static-libubsan.
7662 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7663 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7664 * builtin-attrs.def (ATTR_COLD): Define.
7665 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7666 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7667 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7668 * flag-types.h (sanitize_code): New enum.
7669 * opts.c (common_handle_option): Parse command line arguments
7670 of -fsanitize=. Add -fsanitize=unreachable option.
7671 * varasm.c (get_variable_section): Adjust.
7672 (assemble_noswitch_variable): Likewise.
7673 (assemble_variable): Likewise.
7674 (output_constant_def_contents): Likewise.
7675 (categorize_decl_for_section): Likewise.
7676 (place_block_symbol): Likewise.
7677 (output_object_block): Likewise.
7678 * builtins.def: Likewise.
7679 * toplev.c (compile_file): Likewise.
7680 (process_options): Likewise.
7681 * cppbuiltin.c: Likewise.
7682 * tsan.c (tsan_pass): Likewise.
7683 (tsan_gate): Likewise.
7684 (tsan_gate_O0): Likewise.
7685 * cfgexpand.c (partition_stack_vars): Likewise.
7686 (expand_stack_vars): Likewise.
7687 (defer_stack_allocation): Likewise.
7688 (expand_used_vars): Likewise.
7689 * cfgcleanup.c (old_insns_match_p): Likewise.
7690 * asan.c (asan_finish_file): Likewise.
7691 (asan_instrument): Likewise.
7692 (gate_asan): Likewise.
7693 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7694 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7695 (asan_global_struct): Use pointer_sized_int_node instead
7696 calling build_nonstandard_integer_type.
7697 (initialize_sanitizer_builtins): Likewise.
7698 (asan_finish_file): Likewise.
7699 * gcc.c: Document %{%:function(args):X}.
7700 (static_spec_functions): Add sanitize.
7701 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7702 store funcval != NULL there.
7703 (do_spec_1): Adjust handle_spec_function caller.
7704 (handle_braces): Allow %:function(args) as condition.
7705 (sanitize_spec_function): New function.
7706 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7707 (LIBUBSAN_SPEC): Likewise.
7708 (LIBUBSAN_EARLY_SPEC): Likewise.
7709 (SANITIZER_SPEC): Handle libubsan.
7710 (SANITIZER_EARLY_SPEC): Likewise.
7711 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7712 instead of fsanitize=address.
7713 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7714 instead of fsanitize=address*.
7715 * builtins.c: Include ubsan.h.
7716 (fold_builtin_0): Instrument __builtin_unreachable.
7717 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7718 instead of flag_asan.
7719 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7720 (pointer_sized_int_node): Define.
7721 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7722
7723 2013-08-30 Mike Stump <mikestump@comcast.net>
7724
7725 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7726 with RE patterns.
7727
7728 2013-08-29 Jan Hubicka <jh@suse.cz>
7729
7730 * cgraph.c (cgraph_function_body_availability): Handle weakref
7731 correctly.
7732 * passes.def: Remove pass_fixup_cfg.
7733 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7734 track when we need to remove functions.
7735 (gate_ipa_inline): Execute inlining always; add comment why.
7736 (pass_data_ipa_inline): Remove TODO_remove_functions.
7737 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7738 do not produce summaries.
7739 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7740 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7741 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7742 constant pool and vtable.
7743
7744 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7745
7746 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7747 New arm_neon.h's internal macros to specify 64-bit constants.
7748 Avoid using stdint.h's macros.
7749
7750 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7751
7752 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7753 before applying REG_P.
7754
7755 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7756
7757 PR tree-optimization/58277
7758 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7759 after seeing too many stmts with vdef in between dombb and current
7760 bb, invalidate everything.
7761
7762 2013-08-30 Richard Biener <rguenther@suse.de>
7763
7764 * fold-const.c (fold_single_bit_test): Fix overflow test.
7765
7766 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7767
7768 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7769 and which can live in a register, always retrieve the value on entry.
7770 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7771 passed by invisible reference specially.
7772 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7773 (vt_add_function_parameter): Correctly deal with a parameter passed by
7774 invisible reference.
7775
7776 2013-08-30 Jan Hubicka <jh@suse.cz>
7777
7778 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7779
7780 2013-08-30 Richard Biener <rguenther@suse.de>
7781
7782 PR tree-optimization/58228
7783 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7784 allow invariant loads in nested loop vectorization.
7785
7786 2013-08-30 Richard Biener <rguenther@suse.de>
7787
7788 PR tree-optimization/58223
7789 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7790 (has_anti_or_output_dependence): ... this and adjust to also
7791 look for output dependences.
7792 (mark_nodes_having_upstream_mem_writes): Adjust.
7793 (rdg_flag_uses): Likewise.
7794
7795 2013-08-30 Richard Biener <rguenther@suse.de>
7796
7797 PR tree-optimization/58010
7798 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7799 assert that we have a loop-closed PHI.
7800
7801 2013-08-29 Jan Hubicka <jh@suse.cz>
7802
7803 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7804 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7805 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7806 (lto_free_function_in_decl_state_for_node): New function.
7807
7808 2013-08-29 Xinliang David Li <davidxl@google.com>
7809
7810 * loop-unroll.c (report_unroll_peel): Minor message change.
7811 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7812 Emit alignment peeling message with default -fopt-info.
7813 (vect_loop_versioning): Emit loop version info message.
7814 * tree-vectorizer.c (vectorize_loops): Minor message change.
7815 (execute_vect_slp): Ditto.
7816
7817 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7818
7819 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7820 of the clone from the DECL_NAME of the original function.
7821
7822 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7823
7824 * passes.c (register_pass): Add overload.
7825 * tree-pass.h (register_pass): Forward declare it. Add comment.
7826
7827 2013-08-29 Jan Hubicka <jh@suse.cz>
7828
7829 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7830 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7831 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7832 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7833 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7834 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7835 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7836 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7837
7838 2013-08-29 Teresa Johnson <tejohnson@google.com>
7839
7840 * dumpfile.c (dump_loc): Output column number.
7841 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7842 * doc/invoke.texi: Document optall -fopt-info flag.
7843 * profile.c (read_profile_edge_counts): Use new dump framework.
7844 (compute_branch_probabilities): Ditto.
7845 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7846 when pass not in any opt group.
7847 * pass_manager.h (pass_manager::get_pass_profile): New method.
7848 * value-prof.c (check_counter): Use new dump framework.
7849 (check_ic_target): Ditto.
7850 * coverage.c (get_coverage_counts): Ditto.
7851 (coverage_init): Setup new dump framework.
7852
7853 2013-08-29 Richard Biener <rguenther@suse.de>
7854
7855 PR tree-optimization/58246
7856 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7857 handle the dominance check inside a basic-block.
7858
7859 2013-08-29 Richard Biener <rguenther@suse.de>
7860
7861 PR middle-end/57287
7862 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7863 of default defs that appear in abnormal PHI nodes.
7864
7865 2013-08-29 Richard Biener <rguenther@suse.de>
7866
7867 PR tree-optimization/57685
7868 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7869 single-use operands to avoid exponential complexity.
7870
7871 2013-08-28 Dehao Chen <dehao@google.com>
7872
7873 * ipa-inline.c (edge_badness): Fix integer underflow.
7874
7875 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7876
7877 * gtm-builtins.def (_ITM_free): Declare leaf.
7878
7879 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7880
7881 PR target/58067
7882 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7883 (*tls_local_dynamic_base_64_largepic): Likewise.
7884 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7885 Remove predicate from call operand.
7886 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7887 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7888
7889 2013-08-28 Jeff Law <law@redhat.com>
7890
7891 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7892 checks for the number of predecessors and successors allowed.
7893 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7894 which require copying a joiner block if there is a request which
7895 is a subpath that requires no joiner block copying.
7896
7897 2013-08-28 Jan Hubicka <jh@suse.cz>
7898
7899 * lto-streamer-out.c (DFS_write_tree_body): Drop
7900 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7901 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7902 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7903 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7904 Do not read DECL_ERROR_ISSUED.
7905 (unpack_ts_decl_with_vis_value_fields): Do not read
7906 DECL_DEFER_OUTPUT.
7907 (lto_input_ts_binfo_tree_pointers): Do not read
7908 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7909 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7910 write DECL_ERROR_ISSUED..
7911 (pack_ts_decl_with_vis_value_fields): Do not write
7912 DECL_DEFER_OUTPUT.
7913 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7914 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7915 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7916 * tree.h (tree_decl_common): Update comment.
7917 (DECL_ERROR_ISSUED): Remove.
7918
7919 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7920
7921 PR middle-end/58257
7922 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7923
7924 2013-08-28 Jan Hubicka <jh@suse.cz>
7925
7926 * builtins.def (free): Declare leaf.
7927
7928 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7929
7930 * gdbhooks.py: New.
7931 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7932 * configure: Regenerate.
7933
7934 2013-08-27 Martin Jambor <mjambor@suse.cz>
7935
7936 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7937 (ipa_ancestor_jf_data): Likewise.
7938 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7939 (ipa_get_jf_pass_through_type_preserved): New function.
7940 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7941 (ipa_get_jf_ancestor_type_preserved): New function.
7942 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7943 (ipa_get_jf_ancestor_result): Likewise.
7944 (propagate_vals_accross_pass_through): Use
7945 ipa_get_jf_pass_through_result to do all the value mappings.
7946 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7947 type_preserved flag.
7948 (ipa_set_jf_cst_copy): New function.
7949 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7950 (ipa_set_jf_arith_pass_through): Likewise.
7951 (ipa_set_ancestor_jf): Likewise.
7952 (compute_complex_assign_jump_func): Set type_preserved instead of
7953 punting.
7954 (ipa_compute_jump_functions_for_edge): Likewise.
7955 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7956 (update_jump_functions_after_inlining): Update type_preserved.
7957 Explicitely create jump functions when combining one with pass_through.
7958 (ipa_write_jump_function): Stream the type_preserved flags.
7959 (ipa_read_jump_function): Likewise.
7960
7961 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7962 Aldy Hernandez <aldyh@redhat.com>
7963
7964 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7965 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7966 * function.h (struct function): Add has_force_vect_loops and
7967 has_simduid_loops.
7968 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7969 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7970 handle it.
7971 * gimple.def: Update CLAUSES comments.
7972 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7973 (gimple_build_omp_for): Add argument to prototype.
7974 (gimple_omp_for_kind): New.
7975 (gimple_omp_for_set_kind): New.
7976 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7977 GOVD_DATA_SHARE_CLASS.
7978 (enum omp_region_type): Add ORT_SIMD.
7979 (gimple_add_tmp_var): Handle ORT_SIMD.
7980 (gimplify_var_or_parm_decl): Same.
7981 (is_gimple_stmt): Same.
7982 (omp_firstprivatize_variable): Same.
7983 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7984 (omp_notice_variable): Handle ORT_SIMD.
7985 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7986 (omp_check_private): Handle ORT_SIMD.
7987 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7988 OMP_CLAUSE_SAFELEN.
7989 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7990 Handle OMP_CLAUSE_LASTPRIVATE.
7991 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7992 OMP_CLAUSE_SAFELEN.
7993 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7994 (gimplify_expr): Handle OMP_SIMD.
7995 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7996 (expand_GOMP_SIMD_VF): New.
7997 (expand_GOMP_SIMD_LAST_LANE): New.
7998 * internal-fn.def (GOMP_SIMD_LANE): New.
7999 (GOMP_SIMD_VF): New.
8000 (GOMP_SIMD_LAST_LANE): New.
8001 * omp-low.c: Include target.h.
8002 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8003 OMP_CLAUSE_SAFELEN.
8004 (check_omp_nesting_restrictions): Same.
8005 (omp_max_vf): New.
8006 (lower_rec_simd_input_clauses): New.
8007 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8008 OMP_CLAUSE_LINEAR.
8009 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8010 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8011 (expand_omp_build_assign): New.
8012 (expand_omp_for_init_counts): New.
8013 (expand_omp_for_init_vars): New.
8014 (extract_omp_for_update_vars): New.
8015 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8016 and rewrite accordingly.
8017 (expand_omp_simd): New.
8018 (expand_omp_for): Use expand_omp_simd.
8019 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8020 (lower_omp_for): Do not lower the body.
8021 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8022 in their own loops.
8023 * tree-flow.h (find_omp_clause): Remove prototype.
8024 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8025 forcing vectorization of the loop, or if flag_tree_vectorize.
8026 (gate_tree_if_conversion): Similarly.
8027 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8028 gimple_build_omp_for.
8029 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8030 * tree-parloops (create_parallel_loop): Pass kind argument to
8031 gimple_build_omp_for.
8032 * tree-pretty-print.c (dump_omp_clause): Add cases for
8033 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8034 OMP_CLAUSE__SIMDUID_.
8035 (dump_generic_node): Handle OMP_SIMD.
8036 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8037 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8038 unroll OMP_SIMD loops here.
8039 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8040 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8041 loop->safelen.
8042 (vect_analyze_data_refs): Handle simd loops.
8043 * tree-vect-loop.c (vectorizable_live_operation): Handle
8044 IFN_GOMP_SIMD*.
8045 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8046 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8047 (vectorizable_load): Same.
8048 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8049 (struct simduid_to_vf): New.
8050 (simduid_to_vf::hash): New.
8051 (simduid_to-vf::equal): New.
8052 (struct simd_array_to_simduid): New.
8053 (simd_array_to_simduid::hash): New.
8054 (simd_array_to_simduid::equal): New.
8055 (adjust_simduid_builtins): New.
8056 (struct note_simd_array_uses_struct): New.
8057 (note_simd_array_uses_cb): New.
8058 (note_simd_array_uses): New.
8059 (vectorize_loops): Handle simd hints and adjust simd builtins
8060 accordingly.
8061 * tree-vectorizer.h (struct _stmt_vec_info): Add
8062 simd_lane_access_p field.
8063 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8064 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8065 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8066 (omp_clause_code_name): Same.
8067 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8068 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8069 * tree.def (OMP_SIMD): New entry.
8070 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8071 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8072 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8073 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8074 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8075 (OMP_CLAUSE_LINEAR_STEP): New.
8076 (OMP_CLAUSE_SAFELEN_EXPR): New.
8077 (OMP_CLAUSE__SIMDUID__DECL): New.
8078 (find_omp_clause): New prototype.
8079
8080 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8081
8082 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8083 Haswell processor detection.
8084
8085 2013-08-27 Christian Widmer <shadow@umbrox.de>
8086
8087 PR target/57927
8088 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8089 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8090 AVX2 capable processors.
8091
8092 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8093
8094 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8095 of vget_low_* with implementations in terms of other intrinsics.
8096
8097 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8098
8099 PR middle-end/57219
8100 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8101 values to -1, 0 and 1.
8102
8103 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8104
8105 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8106 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8107 (<optab><mode>3_insn): Remove.
8108 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8109 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8110 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8111 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8112 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8113 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8114 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8115 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8116 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8117 (ror<mode>3_insn): Likewise.
8118 * config/aarch64/predicates.md (aarch64_simd_register): New.
8119
8120 2013-08-27 Richard Biener <rguenther@suse.de>
8121
8122 PR tree-optimization/57521
8123 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8124 one edge is non-critical.
8125 (find_phi_replacement_condition): Make sure to use a non-critical
8126 edge. Cleanup and remove old bug workarounds.
8127 (bb_postdominates_preds): Remove.
8128 (if_convertible_loop_p_1): Do not compute post-dominators.
8129 (combine_blocks): Do not free post-dominators.
8130 (main_tree_if_conversion): Likewise.
8131 (pass_data_if_conversion): Add TODO_verify_ssa.
8132
8133 2013-08-27 DJ Delorie <dj@redhat.com>
8134
8135 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8136
8137 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8138
8139 * function.c (assign_parm_find_data_types): Set passed_mode and
8140 nominal_mode to the TYPE_MODE of nominal_type for the built
8141 pointer type in case of the struct-pass-by-reference.
8142
8143 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8144
8145 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8146 (UINT16_TYPE): Change default to "unsigned int".
8147
8148 * config/avr/avr.opt (mfract-convert-truncate): New option.
8149 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8150 is set, round negative fractional integers according to n1169
8151 when converting to integer types.
8152
8153 2013-08-26 Jan Hubicka <jh@suse.cz>
8154
8155 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8156 methods can not be called indirectly when their address is not taken.
8157
8158 2013-08-26 Jan Hubicka <jh@suse.cz>
8159
8160 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8161 ctor_for_folding.
8162
8163 2013-08-26 Jan Hubicka <jh@suse.cz>
8164
8165 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8166 can be unshared.
8167
8168 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8169
8170 * reload.c (find_valid_class): Allow classes that do not include
8171 FIRST_PSEUDO_REGISTER - 1.
8172
8173 2013-08-26 Jan Hubicka <jh@suse.cz>
8174
8175 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8176 fix edge count/frequency when speculation failed; fix type check
8177 for the direct call.
8178
8179 2013-08-26 Jan Hubicka <jh@suse.cz>
8180
8181 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8182
8183 2013-08-26 Jan Hubicka <jh@suse.cz>
8184
8185 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8186 to be changed by edge redirection.
8187
8188 2013-08-26 Jan Hubicka <jh@suse.cz>
8189
8190 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8191 formating; add sanity check.
8192 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8193 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8194 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8195 (resolve_noninline_speculation): Update callee keys, too.
8196
8197 2013-08-26 Jan Hubicka <jh@suse.cz>
8198
8199 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8200 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8201
8202 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8203
8204 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8205 into proper place.
8206
8207 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8208
8209 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8210 (x86_64_elf_select_section): Ditto.
8211 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8212 arguments.
8213 (ix86_pass_by_reference): Ditto.
8214 (output_set_got): Ditto.
8215 (ix86_unary_operator_ok): Ditto.
8216 (ix86_expand_builtin): Ditto.
8217
8218 2013-08-23 Jan Hubicka <jh@suse.cz>
8219
8220 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8221
8222 2013-08-23 Jan Hubicka <jh@suse.cz>
8223
8224 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8225 (tree_decl_with_vis): Add FINAL field.
8226
8227 2013-08-23 Jeff Law <law@redhat.com>
8228
8229 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8230 the debugging dump when the expression is fully redundant.
8231
8232 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8233
8234 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8235 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8236 * pretty-print.c (pp_formatted_text_data): Likewise.
8237 (pp_write_text_to_stream): Likewise.
8238 (pp_write_text_as_dot_label_to_stream): Likewise.
8239 (pp_append_r): Likewise.
8240 (pp_format): Likewise.
8241 (pp_flush): Likewise.
8242 (pp_clear_output_area): Likewise.
8243 (pp_append_text): Likewise.
8244 (pp_formatted_text): Likewise.
8245 (pp_remaining_character_count_for_line): Likewise.
8246 (pp_newline): Likewise.
8247 (pp_character): Likewise.
8248 (output_buffer::~output_buffer): Define.
8249 (pretty_printer::~pretty_printer): Destruct output buffer.
8250 * pretty-print.h (output_buffer::~output_buffer): Declare.
8251 (pretty_printer::~pretty_printer): Declare virtual.
8252
8253 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8254
8255 PR other/57324
8256 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8257 HOST_WIDE_INT_M1U): New macros.
8258 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8259 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8260 unsigned -1 for lshift.
8261 * cse.c (cse_insn): Likewise.
8262 * double-int.c (rshift_double, lshift_double): Likewise.
8263 * builtins.c (fold_builtin_bitop): Likewise.
8264 * combine.c (force_to_mode): Likewise.
8265 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8266 * simplify-rtx.c (simplify_const_unary_operation,
8267 simplify_const_binary_operation): Likewise.
8268 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8269 check_va_list_escapes): Likewise.
8270 * rtlanal.c (nonzero_bits1): Likewise.
8271 * expmed.c (expand_smod_pow2): Likewise.
8272 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8273
8274 2013-08-23 Jan Hubicka <jh@suse.cz>
8275
8276 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8277 as having address taken.
8278
8279 2013-08-23 Jan Hubicka <jh@suse.cz>
8280
8281 * ipa-utils.h (method_class_type): Declare.
8282 * ipa-devirt.c (method_class_type): Export.
8283
8284 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8285 do not walk base classes of anonymous types.
8286
8287 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8288
8289 PR rtl-optimization/58220
8290 PR regression/58221
8291 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8292 handle SEQUENCE insns properly.
8293
8294 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8295
8296 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8297 definition.
8298 (pp_newline_and_indent): Likewise.
8299 (pp_separate_with): Likewise.
8300 * pretty-print.c (pp_newline_and_flush): Define.
8301 (pp_newline_and_indent): Likewise.
8302 (pp_separate_with): Likewise.
8303
8304 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8305
8306 PR target/58218
8307 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8308 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8309
8310 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8311
8312 * config/i386/predicates.md (ext_sse_reg_operand): New.
8313 * config/i386/i386.md (*movti_internal): Use
8314 predicate to determine if EVEX is needed.
8315 (*movsi_internal): Ditto.
8316 (*movdf_internal): Ditto.
8317 (*movsf_internal): Ditto.
8318 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8319
8320 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8321
8322 PR tree-optimization/58209
8323 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8324 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8325 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8326 emit POINTER_PLUS_EXPR.
8327 (create_tailcall_accumulator): For pointer result type accumulate in
8328 sizetype type.
8329
8330 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8331
8332 * configure.ac: Add backslashes missing from the last change.
8333 * configure: Regenerate.
8334
8335 2013-08-22 Jan Hubicka <jh@susue.cz>
8336
8337 * ipa.c (function_and_variable_visibility): First remember function
8338 was global and then make it local.
8339
8340 2013-08-22 Julian Brown <julian@codesourcery.com>
8341
8342 * configure.ac: Add aarch64 to list of arches which use "nop" in
8343 debug_line test.
8344 * configure: Regenerate.
8345
8346 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8347
8348 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8349 gnu_libc_has_function.
8350 * config/s390/tpf.h: Likewise.
8351
8352 2013-08-22 Jan Hubicka <jh@susue.cz>
8353
8354 * timevar.c (validate_phases): Add cast.
8355
8356 2013-08-22 Jan Hubicka <jh@susue.cz>
8357
8358 * timevar.c (validate_phases): Use size_t for memory.
8359 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8360
8361 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8362
8363 * pretty-print.h (output_buffer::output_buffer): Declare.
8364 (pretty_printer::pretty_printer): Likewise.
8365 (pp_construct): Remove.
8366 * pretty-print.c (output_buffer::output_buffer): Define.
8367 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8368 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8369 (print_gimple_expr): Likewise.
8370 (print_gimple_seq): Likewise.
8371 (gimple_dump_bb): Likewise.
8372 * sched-vis.c (dump_value_slim): Likewise.
8373 (dump_insn_slim): Likewise.
8374 (dump_rtl_slim): Likewise.
8375 (str_pattern_slim): Likewise.
8376 * tree-mudflap.c (mf_varname_tree): Likewise.
8377 * graph.c (print_graph_cfg): Likewise.
8378 (start_graph_dump): Likewise.
8379 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8380 placement-new.
8381 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8382 pretty printer initialization.
8383 * coretypes.h (diagnostic_context): Remove superflous type alias
8384 declaration.
8385 (pretty_printer): Likewise. Declare directly as a class.
8386 (pretty_print_info): Remove declaration as class.
8387 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8388 and pp_clear_output_area.
8389 (asan_add_global): Likewise.
8390
8391 2013-08-22 Jan Hubicka <jh@suse.cz>
8392
8393 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8394 * ipa-utils.h (update_type_inheritance_graph): Declare.
8395 (possible_polymorphic_call_target_p): Declare.
8396 (possible_polymorphic_call_target_p): New.
8397 * ipa-devirt.c: Update toplevel comments.
8398 (cached_polymorphic_call_targets): Move up.
8399 (odr_type_d): Move ID down.
8400 (polymorphic_type_binfo_p): Update comment.
8401 (odr_hasher::remove): Likewise;
8402 (get_odr_type): Set anonymous_namespace.
8403 (dump_odr_type): Dump it.
8404 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8405 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8406 (record_binfo): Add comment.
8407 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8408 built.
8409 (devirt_node_removal_hook): Do not iCE when cache is freed.
8410 (possible_polymorphic_call_target_p): New predicate.
8411 (update_type_inheritance_graph): New function.
8412
8413 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8414 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8415 Sergey Lega <sergey.s.lega@intel.com>
8416 Anna Tikhonova <anna.tikhonova@intel.com>
8417 Ilya Tocar <ilya.tocar@intel.com>
8418 Andrey Turetskiy <andrey.turetskiy@intel.com>
8419 Ilya Verbin <ilya.verbin@intel.com>
8420 Kirill Yukhin <kirill.yukhin@intel.com>
8421 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8422
8423 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8424 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8425 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8426 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8427 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8428 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8429 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8430 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8431 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8432 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8433 OPT_mavx512pf, OPT_mavx512er cases.
8434 * config/i386/constraints.md (v): New constraint.
8435 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8436 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8437 (bit_AVX512CD): New.
8438 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8439 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8440 * config/i386/i386-c.c (ix86_target_macros_internal):
8441 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8442 __AVX512PF__.
8443 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8444 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8445 * config/i386/i386.c (regclass_map, dbx_register_map)
8446 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8447 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8448 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8449 -mavx512pf options.
8450 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8451 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8452 -mavx512pf options. Fix formatting.
8453 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8454 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8455 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8456 -mavx512cd, -mavx512pf options.
8457 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8458 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8459 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8460 ALL_SSE_REGS.
8461 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8462 (ix86_set_reg_reg_cost): Ditto.
8463 (x86_order_regs_for_local_alloc): Ditto.
8464 (MAX_VECT_LEN): Extend to 64-byte.
8465 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8466 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8467 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8468 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8469 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8470 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8471 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8472 (SSE_REG_MODE_P): Support new modes.
8473 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8474 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8475 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8476 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8477 (REG_CLASS_CONTENTS): Add new registers.
8478 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8479 (EXT_REX_SSE_REGNO_P): New.
8480 (HI_REGISTER_NAMES): Add new registers.
8481 * config/i386/i386.md: Define constants for new registers.
8482 (mode): Add new 512-bit modes.
8483 (prefix): Support evex prefix.
8484 (isa): Support avx512f, noavx512f, fma_avx512f.
8485 (ssemodesuffix): Add new 512-bit modes.
8486 (movxi): New.
8487 (*movxi_internal_avx512f): Ditto.
8488 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8489 Support MODE_XI.
8490 (*movsi_internal): Likewise.
8491 (*movdf_internal): Likewise.
8492 (*movsf_internal): Likewise.
8493 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8494 (<code><mode>3): Likewise.
8495 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8496 New.
8497 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8498 with the new constraint "v".
8499 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8500 modes.
8501 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8502 with the new constraint "v".
8503 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8504 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8505 (*<plusminus_insn><mode>3): Likewise.
8506 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8507 (*mul<mode>3): Likewise.
8508 (<sse>_vmmul<mode>3): Likewise.
8509 (<sse>_div<mode>3): Likewise.
8510 (<sse>_vmdiv<mode>3): Likewise.
8511 (<sse>_sqrt<mode>2): Likewise.
8512 (<sse>_vmsqrt<mode>2): Likewise.
8513 (*<code><mode>3_finite): Likewise.
8514 (*<code><mode>3) <smaxmin>: Likewise.
8515 (<sse>_vm<code><mode>3): Likewise.
8516 (*<code><mode>3) <any_logic>: Likewise.
8517 (*fma_fmadd_<mode>): Likewise.
8518 (*fma_fmsub_<mode>): Likewise.
8519 (*fma_fnmadd_<mode>): Likewise.
8520 (*fma_fnmsub_<mode>): Likewise.
8521 (*fma_fmaddsub_<mode>): Likewise.
8522 (*fma_fmsubadd_<mode>): Likewise.
8523 (*fmai_fmadd_<mode>): Likewise.
8524 (*fmai_fmsub_<mode>): Likewise.
8525 (*fmai_fnmadd_<mode>): Likewise.
8526 (*fmai_fnmsub_<mode>): Likewise.
8527 (sse_cvtsi2ss): Likewise.
8528 (sse_cvtsi2ssq): Likewise.
8529 (sse_cvtss2si): Likewise.
8530 (sse_cvtss2si_2): Likewise.
8531 (sse_cvtss2siq): Likewise.
8532 (sse_cvtss2siq_2): Likewise.
8533 (sse_cvttss2si): Likewise.
8534 (sse_cvtss2siq_2): Likewise.
8535 (float<sseintvecmodelower><mode>2): Likewise.
8536 (sse2_cvtsd2si_2): Likewise.
8537 (sse2_cvtsd2siq_2): Likewise.
8538 (*<plusminus_insn><mode>3): Likewise.
8539 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8540 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8541 (ashr<mode>3): Likewise.
8542 (<shift_insn><mode>3): Likewise.
8543 (avx2_<code><mode>3): Likewise.
8544 (*avx2_<code><mode>3): Likewise.
8545 (*andnot<mode>3): Likewise.
8546 (*<code><mode>3) <any_logic>: Likewise.
8547 (abs<mode>2): Likewise.
8548 (avx2_permvar<mode>): Likewise.
8549 (avx2_perm<mode>_1): Likewise.
8550 (*avx_vpermilp<mode>): Likewise.
8551 (avx_vpermilvar<mode>3): Likewise.
8552 (avx2_ashrv<mode>): Likewise.
8553 (avx2_<shift_insn>v<mode>): Likewise.
8554 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8555 -mavx512cd.
8556 * doc/rtl.texi: Document XImode.
8557
8558 2013-08-21 Jeff Law <law@redhat.com>
8559
8560 * tree-flow.h (register_jump_thread): Pass vector of edges
8561 instead of each important edge.
8562 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8563 thread path into a vector and pass that to register_jump_thread.
8564 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8565 passed in edge vector to the current 3-edge form.
8566
8567 Revert:
8568 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8569
8570 * dce.c (fini_dce): Call df_analyze again just in case
8571 delete_unmarked_insns removed anything.
8572
8573 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8574
8575 * reload.h (struct reg_equivs): Rename to ..
8576 (struct reg_equivs_s): .. this.
8577
8578 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8579
8580 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8581
8582 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8583
8584 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8585
8586 2013-08-21 Jeff Law <law@redhat.com>
8587
8588 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8589 simplify assignments too. If the RHS collapses to a singleton
8590 range, then return the value for the range.
8591
8592 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8593
8594 * config/i386/sse.md (V16): Rename to...
8595 (VMOVE): this.
8596 (mov<mode>): Update iterator name.
8597 (*mov<mode>_internal): Ditto.
8598 (push<mode>1): Ditto.
8599 (movmisalign<mode>): Ditto.
8600
8601 2013-08-20 Jan Hubicka <jh@suse.cz>
8602
8603 PR bootstrap/58186
8604 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8605 entry for direct edges.
8606 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8607
8608 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8609
8610 Revert my last two changes, r201865 and r201864:
8611
8612 Revert r201865:
8613 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8614
8615 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8616 instances can own GC refs.
8617
8618 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8619 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8620 (gcc::context::gt_pch_nx): Likewise.
8621 (gcc::context::gt_pch_nx): Likewise.
8622 * ggc.h (gt_ggc_mx <T>): New.
8623 (gt_pch_nx_with_op <T>): New.
8624 (gt_pch_nx <T>): New.
8625 * passes.c (opt_pass::gt_ggc_mx): New.
8626 (opt_pass::gt_pch_nx): New.
8627 (opt_pass::gt_pch_nx_with_op): New.
8628 (pass_manager::gt_ggc_mx): New.
8629 (pass_manager::gt_pch_nx): New.
8630 (pass_manager::gt_pch_nx_with_op): New.
8631 (pass_manager::operator new): Use
8632 ggc_internal_cleared_alloc_stat rather than xcalloc.
8633 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8634 (pass_manager::gt_ggc_mx): New.
8635 (pass_manager::gt_pch_nx): New.
8636 (pass_manager::gt_pch_nx_with_op): New.
8637 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8638 (opt_pass::operator new): New.
8639 (opt_pass::gt_ggc_mx): New.
8640 (opt_pass::gt_pch_nx): New.
8641 (opt_pass::gt_pch_nx_with_op): New.
8642
8643 Revert r201864:
8644 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8645
8646 * Makefile.in (GTFILES): Add context.h.
8647 * context.c (gcc::context::operator new): New.
8648 (gcc::context::gt_ggc_mx): New.
8649 (gcc::context::gt_pch_nx): New.
8650 (gcc::context::gt_pch_nx): New.
8651 * context.h (gcc::context): Add GTY((user)) marking.
8652 (gcc::context::operator new): New.
8653 (gcc::context::gt_ggc_mx): New.
8654 (gcc::context::gt_pch_nx): New.
8655 (gcc::context::gt_pch_nx): New.
8656 (g): Add GTY marking.
8657 (gt_ggc_mx (gcc::context *)): New.
8658 (gt_pch_nx (gcc::context *)): New.
8659 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8660 void *cookie)): New.
8661 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8662
8663 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8664
8665 * dce.c (fini_dce): Call df_analyze again just in case
8666 delete_unmarked_insns removed anything.
8667
8668 2013-08-20 Teresa Johnson <tejohnson@google.com>
8669
8670 PR rtl-optimizations/57451
8671 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8672 from crossing split section boundaries.
8673
8674 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8675
8676 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8677 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8678 with MULTLIB_DEFAULTS.
8679
8680 2013-08-20 Nick Clifton <nickc@redhat.com>
8681
8682 * target.def (narrow_volatile_bitfield): Note that the default
8683 value is false, not !TARGET_STRICT_ALIGN.
8684 * doc/tm.texi: Regenerate.
8685
8686 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8687
8688 Fix LIB_SPEC for systems without libpthread.
8689
8690 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8691 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8692 for Android.
8693 * config/i386/linux-common.h: Likewise.
8694 * config/mips/linux-common.h: Likewise.
8695
8696 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8697
8698 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8699 checks.
8700
8701 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8702
8703 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8704 instances can own GC refs.
8705
8706 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8707 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8708 (gcc::context::gt_pch_nx): Likewise.
8709 (gcc::context::gt_pch_nx): Likewise.
8710 * ggc.h (gt_ggc_mx <T>): New.
8711 (gt_pch_nx_with_op <T>): New.
8712 (gt_pch_nx <T>): New.
8713 * passes.c (opt_pass::gt_ggc_mx): New.
8714 (opt_pass::gt_pch_nx): New.
8715 (opt_pass::gt_pch_nx_with_op): New.
8716 (pass_manager::gt_ggc_mx): New.
8717 (pass_manager::gt_pch_nx): New.
8718 (pass_manager::gt_pch_nx_with_op): New.
8719 (pass_manager::operator new): Use
8720 ggc_internal_cleared_alloc_stat rather than xcalloc.
8721 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8722 (pass_manager::gt_ggc_mx): New.
8723 (pass_manager::gt_pch_nx): New.
8724 (pass_manager::gt_pch_nx_with_op): New.
8725 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8726 (opt_pass::operator new): New.
8727 (opt_pass::gt_ggc_mx): New.
8728 (opt_pass::gt_pch_nx): New.
8729 (opt_pass::gt_pch_nx_with_op): New.
8730
8731 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8732
8733 * Makefile.in (GTFILES): Add context.h.
8734 * context.c (gcc::context::operator new): New.
8735 (gcc::context::gt_ggc_mx): New.
8736 (gcc::context::gt_pch_nx): New.
8737 (gcc::context::gt_pch_nx): New.
8738 * context.h (gcc::context): Add GTY((user)) marking.
8739 (gcc::context::operator new): New.
8740 (gcc::context::gt_ggc_mx): New.
8741 (gcc::context::gt_pch_nx): New.
8742 (gcc::context::gt_pch_nx): New.
8743 (g): Add GTY marking.
8744 (gt_ggc_mx (gcc::context *)): New.
8745 (gt_pch_nx (gcc::context *)): New.
8746 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8747 void *cookie)): New.
8748 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8749
8750 2013-08-20 Alan Modra <amodra@gmail.com>
8751
8752 PR target/57865
8753 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8754 (rs6000_emit_epilogue): Likewise.
8755
8756 2013-08-19 Dehao Chen <dehao@google.com>
8757
8758 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8759
8760 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8761 Jakub Jelinek <jakub@redhat.com>
8762
8763 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8764 (BUILT_IN_FABSD64): Likewise.
8765 (BUILT_IN_FABSD128): Likewise.
8766 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8767 (fold_builtin_1): Likewise.
8768 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8769 destination and source operands.
8770 (*abstd2_fpr): Likewise.
8771 (*nabstd2_fpr): Likewise.
8772
8773 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8774
8775 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8776 JUMP_P and INSN_P.
8777
8778 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8779
8780 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8781 incomplete.
8782
8783 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8784
8785 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8786 * builtins.c (default_libc_has_function): New.
8787 (gnu_libc_has_function): Ditto.
8788 (no_c99_libc_has_function): Ditto.
8789 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8790 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8791 (fold_builtin_sincos): Likewise.
8792 (fold_builtin_cexp): Likewise.
8793 * builtins.def (DEF_C94_BUILTIN): Likewise.
8794 (DEF_C99_BUILTIN): Likewise.
8795 (DEF_C99_C90RES_BUILTIN): Likewise.
8796 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8797 definitions to using this define.
8798 * config/darwin-protos.h (darwin_libc_has_function): New.
8799 * config/darwin.c (darwin_libc_has_function): Ditto.
8800 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8801 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8802 * config/darwin.h: Ditto.
8803 * config/elfos.h: Ditto.
8804 * config/freebsd.h: Ditto.
8805 * config/i386/cygming.h: Ditto.
8806 * config/i386/djgpp.h: Ditto.
8807 * config/i386/i386-interix.h: Ditto.
8808 * config/microblaze/microblaze.h: Ditto.
8809 * config/mmix/mmix.h: Ditto.
8810 * config/gnu-user.h: Ditto.
8811 * config/ia64/hpux.h: Ditto.
8812 * config/pa/pa-hpux.h: Ditto.
8813 * config/pdp11/pdp11.h: Ditto.
8814 * config/picochip/picochip.h: Ditto.
8815 * config/linux.h: Ditto.
8816 * config/netbsd.h: Ditto.
8817 * config/openbsd.h: Ditto.
8818 * config/rs6000/aix43.h: Ditto.
8819 * config/rs6000/aix51.h: Ditto.
8820 * config/rs6000/aix52.h: Ditto.
8821 * config/rs6000/aix53.h: Ditto.
8822 * config/rs6000/aix61.h: Ditto.
8823 * config/rs6000/darwin.h: Ditto.
8824 * config/rs6000/linux.h: Ditto.
8825 * config/rs6000/linux64.h: Ditto.
8826 * config/s390/tpf.h: Ditto.
8827 * config/sol2-10.h: Ditto.
8828 * config/sol2.h: Ditto.
8829 * config/vms/vms.h: Ditto.
8830 * config/vxworks.h: Ditto.
8831 * config/linux-android.c (linux_android_libc_has_function):
8832 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8833 * config/linux-protos.h (linux_android_libc_has_function):
8834 New declaration.
8835 * config/i386/i386.c (ix86_libc_has_function): New.
8836 * config/i386/i386-protos.h
8837 (ix86_libc_has_function): New declaration.
8838 * config/i386/i386.md
8839 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8840 ("isinf<mode>2): Likewise.
8841 * convert.c (convert_to_integer): Using new target hook
8842 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8843 TARGET_C99_FUNCTIONS.
8844 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8845 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8846 * coretypes.h (function_class): New enum for different
8847 classes of functions.
8848 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8849 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8850 (TARGET_HAS_SINCOS): Likewise.
8851 (TARGET_LIBC_HAS_FUNCTION): New.
8852 * doc/tm.texi: Regenerated.
8853 * targhooks.h (default_libc_has_function): New declaration.
8854 (no_c99_libc_has_function): Ditto.
8855 (gnu_libc_has_function): Ditto.
8856 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8857 and TARGET_HAS_SINCOS.
8858
8859 2013-08-18 Jan Hubicka <jh@suse.cz>
8860
8861 * Makeifle-in (ipa-devirt.o): New.
8862 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8863 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8864 (analyze_functions): Look into possible targets of polymorphic call.
8865 * dumpfile.c (dump_files): Add type-inheritance dump.
8866 * dumpfile.h (TDI_inheritance): New.
8867 * ipa-devirt.c: New file.
8868 * ipa-utils.h (odr_type_d): Forward declare.
8869 (odr_type): New type.
8870 (build_type_inheritance_graph): Declare.
8871 (possible_polymorphic_call_targets): Declare and introduce inline
8872 variant when only edge is pased.
8873 (dump_possible_polymorphic_call_targets): Likewise.
8874 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8875 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8876 (types_same_for_odr): ... here.
8877 * tree.h (type_in_anonymous_namespace_p): Declare.
8878
8879 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8880
8881 PR tree-optimization/58006
8882 * tree-parloops.c (take_address_of): Don't ICE if get_name
8883 returns NULL.
8884 (eliminate_local_variables_stmt): Remove clobber stmts.
8885
8886 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8887
8888 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8889 error message is issued for an alias to undefined symbol.
8890
8891 2013-08-18 Jan Hubicka <jh@suse.cz>
8892
8893 * cgraph.c (cgraph_create_indirect_edge): Discover
8894 polymorphic calls and record basic info into indirect_info.
8895 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8896 devirtualization, ignore objc function calls.
8897 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8898 call with no parm index info.
8899 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8900 * tree.c (virtual_method_call_p): New function.
8901 * tree.h (virtual_method_call_p): Declare.
8902
8903 2013-08-16 Jan Hubicka <jh@suse.cz>
8904
8905 PR middle-end/58179
8906 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8907
8908 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8909
8910 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8911 attribute.
8912
8913 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8914
8915 * gengtype.c (type_for_name): Add special-case support for
8916 locating types within the "gcc::" namespace.
8917 (open_base_files): Emit a "using namespace gcc" directive.
8918
8919 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8920
8921 PR target/58160
8922 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8923 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8924
8925 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8926 array instead of each individual operand as a separate argument.
8927 (emit_fusion_gpr_load): Likewise.
8928 (expand_fusion_gpr_load): Add new function declaration.
8929
8930 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8931 signature to have the operands passed as an array, instead of as
8932 separate arguments. Allow ZERO_EXTEND to be in the memory
8933 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8934 depend on the register live/dead flags when peepholes are run.
8935 (expand_fusion_gpr_load): New function to be called from the
8936 peephole2 pass, to change the register that addis sets to be the
8937 target register.
8938 (emit_fusion_gpr_load): Change the calling signature to have the
8939 operands passed as an array, instead of as separate arguments.
8940 Allow ZERO_EXTEND to be in the memory address, and also
8941 SIGN_EXTEND if -mpower8-fusion-sign.
8942
8943 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8944 unspec enumeration.
8945 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8946 adjust the register addis loads up in the peephole2 pass. Do not
8947 depend on the register live/dead state when the peephole pass is done.
8948
8949 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8950
8951 * gengtype.c (create_user_defined_type): Ensure that the kind
8952 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8953 declaration is seen before the GTY((user)) marking.
8954
8955 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8956
8957 PR target/58105
8958 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8959
8960 2013-08-16 Jan Hubicka <jh@suse.cz>
8961
8962 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8963 arugment expected_type.
8964 (gimple_fold_call): Use it.
8965 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8966 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8967 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8968 (try_make_edge_direct_virtual_call): Likewise.
8969 * tree.c (obj_type_ref_class): New.
8970 * tree.h (obj_type_ref_class): Use it.
8971
8972 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8973
8974 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8975 (rtl_slim_pp): Likewise.
8976 (init_rtl_slim_pretty_print): Likewise.
8977 (dump_value_slim): Don't call it. Use local pretty printer.
8978 (dump_insn_slim): Likewise.
8979 (dump_rtl_slim): Likewise.
8980 (str_pattern_slim): Likewise.
8981 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8982 Simplify.
8983
8984 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8985
8986 PR tree-optimization/58164
8987 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8988 walk gimple_goto_dest of GIMPLE_GOTO.
8989
8990 PR tree-optimization/58165
8991 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8992 bi_call must be the last stmt in a bb, don't split_block, instead
8993 use fallthru edge from it and give up if there is none.
8994 Release conds vector when returning early.
8995
8996 2013-08-14 Xinliang David Li <davidxl@google.com>
8997
8998 * config/i386/i386.c (ix86_option_override_internal):
8999 Remove unused variable and field.
9000
9001 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9002
9003 PR target/57949
9004 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9005 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9006 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9007 and Linux, correct BLKmode alignment when 128-bit alignment is
9008 required and compatibility flag is not set.
9009 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9010 for zero-size arguments when compatibility flag is not set.
9011
9012 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9013
9014 PR tree-optimization/58145
9015 * tree-sra.c (build_ref_for_offset): If prev_base has
9016 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9017
9018 2013-08-14 Xinliang David Li <davidxl@google.com>
9019
9020 * config/i386/i386.c (ix86_option_override_internal):
9021 Fix uninitialized variable error.
9022
9023 2013-08-14 Xinliang David Li <davidxl@google.com>
9024
9025 * config/i386/i386.opt: Define two new options.
9026 * config/i386/x86-tune.def: Add arch selector field in macros.
9027 * config/i386/i386.h: Adjust macro definition.
9028 * config/i386/i386.c (ix86_option_override_internal):
9029 Refactor the code.
9030 (parse_mtune_ctrl_str): New function.
9031 (set_ix86_tune_features): New function.
9032 (ix86_function_specific_restore): Call the new helper function.
9033
9034 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9035
9036 PR rtl-optimization/57662
9037 * sel-sched.c (code_motion_process_successors): When the current insn
9038 is removed after the recursive traversal, break from the loop.
9039 Add comments and debug printouts.
9040
9041 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9042 Alexandre Oliva <aoliva@redhat.com>
9043
9044 PR target/58067
9045 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9046 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9047 there also UNSPEC_PLTOFF.
9048
9049 2013-08-14 Marek Polacek <polacek@redhat.com>
9050
9051 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9052 NUM_CONDITIONS bit positions.
9053
9054 2013-08-13 Cary Coutant <ccoutant@google.com>
9055
9056 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9057 (attr_checksum): Hash vector contents instead of pointer.
9058 (attr_checksum_ordered): Likewise.
9059
9060 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9061
9062 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9063 when Pmode != word_mode. Add length_address attribute.
9064 (sse3_monitor_<mode>): Merge from sse3_monitor and
9065 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9066 Pmode != word_mode. Update insn length attribute.
9067 * config/i386/i386.c (ix86_option_override_internal): Update
9068 ix86_gen_monitor selection for merged sse3_monitor insn.
9069
9070 2013-08-13 Julian Brown <julian@codesourcery.com>
9071
9072 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9073 perform invalid legitimization on greater-than-word-size modes for
9074 TARGET_E500_DOUBLE.
9075
9076 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9077
9078 * ira.c (setup_class_translate_array): Use aclass instead of cl
9079 for classes not fully covered by allocno classes.
9080
9081 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9082
9083 PR tree-optimization/57661
9084 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9085 * tree-inline.c (tree_function_versioning): Initialize it.
9086 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9087 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9088 that is not being copied.
9089
9090 PR sanitizer/56417
9091 * asan.c (instrument_strlen_call): Fix typo in comment.
9092 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9093
9094 2013-08-13 Steve Ellcey <sellcey@mips.com>
9095
9096 * config/mips/mips.md (prefetch): Use lw instead of ld on
9097 loongson in 32bit mode.
9098
9099 2013-08-13 Nick Clifton <nickc@redhat.com>
9100
9101 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9102
9103 2013-08-13 Jan Hubicka <jh@suse.cz>
9104
9105 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9106 introduced edge; fix typo in sanity check.
9107 (cgraph_resolve_speculation): Export; improve diagnostic.
9108 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9109 speculation at type mismatch.
9110 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9111 (cgraph_resolve_speculation): Declare.
9112 (symtab_can_be_discarded): New function.
9113 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9114 * ipa-inline-transform.c (speculation_removed): New global var.
9115 (clone_inlined_nodes): See if speculation can be removed.
9116 (inline_call): If speculations was removed, we growths may not match.
9117 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9118 (speculation_useful_p): New function.
9119 (resolve_noninline_speculation): New function.
9120 (inline_small_functions): Resolve useless speculations.
9121 * ipa-inline.h (speculation_useful_p): Declare
9122 * ipa.c (can_replace_by_local_alias): Simplify.
9123 (ipa_profile): Produce speculative calls in non-lto, too;
9124 add simple cost model; produce local aliases.
9125
9126 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9127
9128 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9129 PASS_MANAGER_H.
9130
9131 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9132
9133 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9134
9135 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9136
9137 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9138 iterator instead of X87MODEF.
9139
9140 2013-08-12 Perez Read <netfirewall@gmail.com>
9141
9142 PR target/58132
9143 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9144 operand 0 for intel asm alternative.
9145 (*movabs<mode>_2): Ditto for operand 1.
9146
9147 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9148
9149 * config/aarch64/arm_none.h
9150 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9151
9152 2013-08-12 Nick Clifton <nickc@redhat.com>
9153
9154 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9155
9156 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9157
9158 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9159 expand for QI/HImode operand to produce more effictive code for
9160 unsigned char(short) --> float(double) conversion.
9161
9162 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9163
9164 * doc/invoke.texi: Mention that -ftls-model does not force the final
9165 model.
9166
9167 2013-08-12 Marek Polacek <polacek@redhat.com>
9168 Marc Glisse <marc.glisse@inria.fr>
9169
9170 PR tree-optimization/57980
9171 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9172 when creating -1 constant.
9173
9174 2013-08-10 Jan Hubicka <jh@suse.cz>
9175
9176 Workaround binutils PR14342.
9177 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9178 (gimple_init_edge_profiler): Likewise.
9179 (gimple_gen_ic_func_profiler): Likewise.
9180
9181 2013-08-09 Jan Hubicka <jh@suse.cz>
9182
9183 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9184
9185 2013-08-09 Xinliang David Li <davidxl@google.com>
9186
9187 * config/i386/stringop.def: New file.
9188 * config/i386/stringop.opt: New file.
9189 * config/i386/i386-opts.h: Include stringopt.def.
9190 * config/i386/i386.opt: Include stringopt.opt.
9191 * config/i386/i386.c (ix86_option_override_internal):
9192 Override default size based stringop inline strategies with options.
9193 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9194 New function.
9195
9196 2013-08-09 Jan Hubicka <jh@suse.cz>
9197
9198 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9199
9200 2013-08-09 Jan Hubicka <jh@suse.cz>
9201
9202 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9203 CGRAPH_FREQ_MAX.
9204 (dump_cgraph_node): Dump profile-id.
9205 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9206 and common_target_probability.
9207 * lto-cgraph.c (lto_output_edge): Stream common targets.
9208 (lto_output_node): Stream profile ids.
9209 (input_node): Stream profile ids.
9210 (input_edge): Stream common targets.
9211 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9212 * ipa.c: Include value-prof.h
9213 (ipa_profile_generate_summary): Turn indirect call statement histograms
9214 into common targets.
9215 (ipa_profile): Turn common targets into speculative edges.
9216
9217 2013-08-09 Jan Hubicka <jh@suse.cz>
9218
9219 * cgraph.h (cgraph_node): Add profile_id.
9220 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9221 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9222 (del_node_map): Update.
9223 (find_func_by_funcdef_no): Replace by ...
9224 (find_func_by_profile_id): ... this one.
9225 (gimple_ic_transform): Do not remove useful histograms when
9226 speculation is not done; dump info when indirect call removal
9227 can happen at LTO.
9228 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9229 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9230 (__gcov_indirect_call_profiler_v2): .. this one.
9231 * profile.h (init_node_map): Update.
9232 * coverage.c (coverage_compute_profile_id): New function.
9233 * coverage.h (coverage_compute_profile_id): Declare.
9234 * tree-profile.c (init_ic_make_global_vars): Make
9235 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9236 (gimple_init_edge_profiler): Update prototype of
9237 __gcov_indirect_call_profiler.
9238 (gimple_gen_ic_func_profiler): Simplify.
9239 (tree_profiling): Use init_node_map
9240
9241 2013-08-09 Jan Hubicka <jh@suse.cz>
9242
9243 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9244 non-speculative refs.
9245 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9246 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9247 (cgraph_set_call_stmt): Likewise.
9248 (cgraph_create_edge_1): Fix release checking compilatoin;
9249 clear lto_stmt_uid.
9250 (cgraph_free_edge): Free indirect info.
9251 (cgraph_turn_edge_to_speculative): New function.
9252 (cgraph_speculative_call_info): New function.
9253 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9254 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9255 (dump_cgraph_node): Dump speculation.
9256 (verify_edge_count_and_frequency): Accept speculative edges.
9257 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9258 (verify_cgraph_node): Handle speculation.
9259 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9260 (cgraph_set_call_stmt): Update prototype.
9261 (cgraph_make_edge_direct): Update prototype.
9262 (cgraph_speculative_call_info): Declare.
9263 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9264 to change; update call of ipa_find_references.
9265 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9266 lto_stmt_uid and speculative flags.
9267 (ipa_dump_references): Dump speculation.
9268 (ipa_clone_references): Clone speculative flag.
9269 (ipa_clone_referring): Likewise.
9270 (ipa_clone_ref): New function.
9271 (ipa_find_reference): Look into lto_stmt_uids
9272 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9273 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9274 (ipa_find_reference): Update declaration.
9275 (ipa_clone_ref): Declare.
9276 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9277 stream speculative flag.
9278 (lto_output_ref): Stream statements uids and speculation.
9279 (input_ref): Likewise.
9280 (input_edge): Stream speuclation.
9281 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9282 (cgraph_set_call_stmt_including_clones): Handle speculation.
9283 * ipa-inline.c (heap_edge_removal_hook): New function.
9284 (inline_small_functions): Register it.
9285 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9286 also initialize refs.
9287 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9288 edge to change.
9289 (try_make_edge_direct_simple_call): Likewise.
9290 (try_make_edge_direct_simple_call): Likewise.
9291 (update_indirect_edges_after_inlining): Likewise.
9292 (remove_described_reference): Look proper lto_stmt_uid.
9293 (propagate_controlled_uses): Likewise.
9294 (propagate_controlled_uses): Liekwise.
9295 * tree-inline.c (copy_bb): Copy speculative edges.
9296 (redirect_all_calls): New function.
9297 (copy_cfg_body): Do redirection after loop info is updated.
9298 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9299
9300 2013-08-09 Jan Hubicka <jh@suse.cz>
9301
9302 * lto-streamer-out.c (output_function): Renumber PHIs.
9303 * lto-streamer-in.c (input_function): Likewise.
9304
9305 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9306
9307 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9308 (get_lane_unsigned): Likewise.
9309 (dup_lane_scalar): Likewise.
9310 (get_lane): enable for VALL.
9311 * config/aarch64/aarch64-simd.md
9312 (aarch64_dup_lane_scalar<mode>): Remove.
9313 (aarch64_get_lane_signed<mode>): Likewise.
9314 (aarch64_get_lane_unsigned<mode>): Likewise.
9315 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9316 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9317 (aarch64_get_lane<mode>): Enable for all vector modes.
9318 (aarch64_get_lanedi): Remove misleading constraints.
9319 * config/aarch64/arm_neon.h
9320 (__aarch64_vget_lane_any): Define.
9321 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9322 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9323 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9324 * config/aarch64/iterators.md (VDQQH): New.
9325 (VDQQHS): Likewise.
9326 (vwcore): Likewise.
9327
9328 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9329
9330 * configure.ac: Add GAS check for LEON instructions on SPARC.
9331 * configure: Regenerate.
9332 * config.in: Likewise.
9333 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9334 sparc*-*-* block.
9335 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9336 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9337 for LEON or LEON3.
9338 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9339 (AS_LEON_FLAG): New macro.
9340 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9341 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9342 Deal with LEON and LEON3 for the memory model.
9343 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9344 (atomic_compare_and_swap<mode>_1): Likewise.
9345 (*atomic_compare_and_swap<mode>_1): Likewise.
9346
9347 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9348
9349 * config/arm/neon.md (vcond): Fix floating-point vector
9350 comparisons against 0.
9351
9352 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9353
9354 * lra-constraints.c (emit_spill_move): Remove assert.
9355 (process_alt_operands): Add more debugging
9356 output. Increase reject for spilling into memory. Decrease
9357 reject for reloading scratch.
9358 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9359
9360 2013-08-08 Steve Ellcey <sellcey@mips.com>
9361
9362 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9363 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9364 micromips incompatible. Add nan2008.
9365 (MULTILIB_DIRNAMES): Add nan2008.
9366 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9367 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9368 and micromips incompatible. Add nan2008.
9369 (MULTILIB_DIRNAMES): Add nan2008.
9370 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9371
9372 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9373
9374 PR rtl-optimization/58079
9375 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9376 simplify_comparison has widened a comparison with an integer.
9377
9378 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9379
9380 * config/arm/neon.md (movmisalign<mode>): Disable when we
9381 don't allow unaligned accesses.
9382 (*movmisalign<mode>_neon_store): Likewise.
9383 (*movmisalign<mode>_neon_load): Likewise.
9384 (*movmisalign<mode>_neon_store): Likewise.
9385 (*movmisalign<mode>_neon_load): Likewise.
9386
9387 2013-08-08 Jan Hubicka <jh@suse.cz>
9388
9389 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9390 (make_pass_rebuild_cgraph_edges): Also clear references.
9391 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9392 * ipa-inline-transform.c (inline_transform): Remove all references
9393 after inlining.
9394 * cgraphunit.c (expand_function): Remove all references after
9395 expansion.
9396 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9397 (ipa_find_reference): Rewrite to iterator.
9398 (remove_stmt_references): Likewise.
9399 (ipa_clear_stmts_in_references): New function.
9400 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9401 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9402 clear references.
9403 * ipa-split.c (split_function): Remove references in split function.
9404
9405 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9406
9407 PR target/57431
9408 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9409 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9410
9411 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9412
9413 PR target/56979
9414 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9415 suggested mode for the assignment isn't compatible with the
9416 registers required.
9417
9418 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9419
9420 PR target/58065
9421 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9422
9423 2013-08-07 Xinliang David Li <davidxl@google.com>
9424
9425 * config/i386/i386.opt: New option -mtune-ctrl=.
9426 * config/i386/x86-tune.def: New file.
9427 * config/i386/i386.h: include x86-tune.def.
9428 * config/i386/i386.c (ix86_option_override_internal):
9429 Parsing -mtune-ctrl= option and set tune features.
9430
9431 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9432
9433 PR other/12081
9434 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9435 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9436 to gen_2arg_fn_t.
9437
9438 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9439
9440 * rtl.h (update_alignments): Declare.
9441 * final.c (grow_label_align): New function extracted from...
9442 (shorten_branches): ...here. Call it.
9443 (update_alignments): New function.
9444 * reorg.c (sibling_labels): New variable.
9445 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9446 the new label along with it onto the sibling_labels vector.
9447 (fill_simple_delay_slots): Adjust call to get_label_before.
9448 (fill_slots_from_thread): Likewise.
9449 (relax_delay_slots): Likewise.
9450 (make_return_insns): Likewise.
9451 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9452
9453 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9454
9455 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9456 document its semantics.
9457 (diagnostic_report_diagnostic): Adjust accordingly.
9458
9459 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9460
9461 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9462 (sparc_option_override): ...and port to new C++ pass API.
9463 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9464
9465 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9466
9467 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9468
9469 2013-08-06 Caroline Tice <cmtice@google.com>
9470
9471 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9472 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9473 * tree-pass.h: Add pass_vtable_verify.
9474 * varasm.c (assemble_variable): Add code to properly set the comdat
9475 section and name for the .vtable_map_vars section.
9476 (assemble_vtyv_preinit_initializer): New function.
9477 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9478 LINK_ONCE flag.
9479 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9480 * vtable-verify.c: New file.
9481 * vtable-verify.h: New file.
9482 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9483 initialiation levels.
9484 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9485 * passes.def: Insert pass_vtable_verify.
9486 * aclocal.m4: Reorder includes.
9487 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9488 -fvtv-counts options.
9489 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9490 as appropriate, if -fvtable-verify=... is used.
9491 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9492 -fvtable-verify=... is used.
9493 * Makefile.in (OBJS): Add vtable-verify.o to list.
9494 (vtable-verify.o): Add new build rule.
9495 (GTFILES): Add vtable-verify.c to list.
9496 * common.opt (fvtable-verify=): New flag.
9497 (vtv_priority): Values for fvtable-verify= flag.
9498 (fvtv-counts): New flag.
9499 (fvtv-debug): New flag.
9500 * tree.h (save_vtable_map_decl): New extern function decl.
9501
9502 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9503
9504 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9505 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9506 (pass_data_rl78_devirt): ...new pass_data instance and...
9507 (make_pass_rl78_devirt): ...new function.
9508 (rl78_asm_file_start): Port pass registration to new C++ API.
9509
9510 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9511
9512 * coretypes.h (rtl_opt_pass): Add.
9513 (gcc::context): Add.
9514 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9515 (epiphany_init): Port to new C++ pass API.
9516 (epiphany_optimize_mode_switching): Likewise.
9517 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9518 (pass_manager::get_pass_mode_switching): New.
9519 (pass_manager::get_pass_peephole2): New.
9520 * mode-switching.c (pass_mode_switching): Add clone method.
9521 * recog.c (pass_peephole2): Add clone method.
9522 (pass_split_all_insns): Add clone method.
9523
9524 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9525
9526 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9527 (mips_option_override): ...here, porting to new C++ API for passes.
9528
9529 2013-08-06 Jan Hubicka <jh@suse.cz>
9530
9531 * cgraph.c (cgraph_get_body): New function based on lto.c
9532 implementation.
9533 * cgraph.h (cgraph_get_body): Declare.
9534 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9535 LTO paths.
9536 * cgraphunit.c (expand_function): Get body prior expanding.
9537 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9538 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9539 really need.
9540 * passes.c (do_per_function_toporder): Get body.
9541 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9542 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9543
9544 2013-08-06 Martin Jambor <mjambor@suse.cz>
9545
9546 PR fortran/57987
9547 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9548 is not re-finalized. Rename second parameter to no_collect.
9549
9550 2013-08-06 Martin Jambor <mjambor@suse.cz>
9551
9552 PR middle-end/58041
9553 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9554 MEM_REF has proper alignment information.
9555
9556 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9557
9558 PR other/12081
9559 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9560 class insn_gen_fn.
9561 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9562 rtx (*) (rtx, ...) with insn_gen_fn.
9563 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9564 insn_gen_fn::stored_funcptr. Add initializer braces.
9565
9566 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9567
9568 Rewrite how instances of passes are cloned to remove assumptions
9569 about their sizes (thus allowing pass subclasses to have
9570 additional data fields, albeit non-GC-managed ones at this point).
9571
9572 * passes.c (make_pass_instance): Now that passes have clone
9573 methods, rewrite this function to eliminate XNEW and memcpy
9574 calls that used hardcoded sizes. Since this function no longer
9575 creates pass instances, rename it to...
9576 (add_pass_instance): ...this. Document the old way that passes were
9577 numbered and flagged, and rework this function to continue using it.
9578 (next_pass_1): Add an initial_pass argument for use by
9579 add_pass_instance.
9580 (position_pass): When adding multiple instances of a pass, use
9581 the pass's clone method, rather than relying on the XNEW/memcpy
9582 within the former make_pass_instance (now add_pass_instance).
9583 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9584 the initial instance of the current pass within the pass manager.
9585
9586 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9587
9588 This is the automated part of the conversion of passes from C
9589 structs to C++ classes.
9590
9591 Patch autogenerated by refactor_passes.py from
9592 https://github.com/davidmalcolm/gcc-refactoring-scripts
9593 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9594
9595 * asan.c (pass_asan): Convert from a global struct to a subclass of
9596 gimple_opt_pass along with...
9597 (pass_data_asan): ...new pass_data instance and...
9598 (make_pass_asan): ...new function.
9599 (pass_asan_O0): Convert from a global struct to a subclass of
9600 gimple_opt_pass along with...
9601 (pass_data_asan_O0): ...new pass_data instance and...
9602 (make_pass_asan_O0): ...new function.
9603 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9604 subclass of rtl_opt_pass along with...
9605 (pass_data_inc_dec): ...new pass_data instance and...
9606 (make_pass_inc_dec): ...new function.
9607 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9608 a subclass of rtl_opt_pass along with...
9609 (pass_data_reorder_blocks): ...new pass_data instance and...
9610 (make_pass_reorder_blocks): ...new function.
9611 (pass_duplicate_computed_gotos): Convert from a global struct to a
9612 subclass of rtl_opt_pass along with...
9613 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9614 (make_pass_duplicate_computed_gotos): ...new function.
9615 (pass_partition_blocks): Convert from a global struct to a subclass of
9616 rtl_opt_pass along with...
9617 (pass_data_partition_blocks): ...new pass_data instance and...
9618 (make_pass_partition_blocks): ...new function.
9619 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9620 struct to a subclass of rtl_opt_pass along with...
9621 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9622 and...
9623 (make_pass_branch_target_load_optimize1): ...new function.
9624 (pass_branch_target_load_optimize2): Convert from a global struct to a
9625 subclass of rtl_opt_pass along with...
9626 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9627 and...
9628 (make_pass_branch_target_load_optimize2): ...new function.
9629 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9630 of rtl_opt_pass along with...
9631 (pass_data_jump): ...new pass_data instance and...
9632 (make_pass_jump): ...new function.
9633 (pass_jump2): Convert from a global struct to a subclass of
9634 rtl_opt_pass along with...
9635 (pass_data_jump2): ...new pass_data instance and...
9636 (make_pass_jump2): ...new function.
9637 * cfgexpand.c (pass_expand): Convert from a global struct to a
9638 subclass of rtl_opt_pass along with...
9639 (pass_data_expand): ...new pass_data instance and...
9640 (make_pass_expand): ...new function.
9641 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9642 of rtl_opt_pass along with...
9643 (pass_data_free_cfg): ...new pass_data instance and...
9644 (make_pass_free_cfg): ...new function.
9645 (pass_into_cfg_layout_mode): Convert from a global struct to a
9646 subclass of rtl_opt_pass along with...
9647 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9648 (make_pass_into_cfg_layout_mode): ...new function.
9649 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9650 subclass of rtl_opt_pass along with...
9651 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9652 (make_pass_outof_cfg_layout_mode): ...new function.
9653 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9654 struct to a subclass of gimple_opt_pass along with...
9655 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9656 (make_pass_build_cgraph_edges): ...new function.
9657 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9658 subclass of gimple_opt_pass along with...
9659 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9660 (make_pass_rebuild_cgraph_edges): ...new function.
9661 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9662 subclass of gimple_opt_pass along with...
9663 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9664 and...
9665 (make_pass_remove_cgraph_callee_edges): ...new function.
9666 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9667 struct to a subclass of rtl_opt_pass along with...
9668 (pass_data_stack_adjustments): ...new pass_data instance and...
9669 (make_pass_stack_adjustments): ...new function.
9670 * combine.c (pass_combine): Convert from a global struct to a subclass
9671 of rtl_opt_pass along with...
9672 (pass_data_combine): ...new pass_data instance and...
9673 (make_pass_combine): ...new function.
9674 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9675 global struct to a subclass of rtl_opt_pass along with...
9676 (pass_data_compare_elim_after_reload): ...new pass_data instance
9677 and...
9678 (make_pass_compare_elim_after_reload): ...new function.
9679 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9680 of rtl_opt_pass along with...
9681 (pass_data_rtl_cprop): ...new pass_data instance and...
9682 (make_pass_rtl_cprop): ...new function.
9683 * cse.c (pass_cse): Convert from a global struct to a subclass of
9684 rtl_opt_pass along with...
9685 (pass_data_cse): ...new pass_data instance and...
9686 (make_pass_cse): ...new function.
9687 (pass_cse2): Convert from a global struct to a subclass of
9688 rtl_opt_pass along with...
9689 (pass_data_cse2): ...new pass_data instance and...
9690 (make_pass_cse2): ...new function.
9691 (pass_cse_after_global_opts): Convert from a global struct to a
9692 subclass of rtl_opt_pass along with...
9693 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9694 (make_pass_cse_after_global_opts): ...new function.
9695 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9696 of rtl_opt_pass along with...
9697 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9698 (make_pass_ud_rtl_dce): ...new function.
9699 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9700 rtl_opt_pass along with...
9701 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9702 (make_pass_fast_rtl_dce): ...new function.
9703 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9704 a subclass of rtl_opt_pass along with...
9705 (pass_data_df_initialize_opt): ...new pass_data instance and...
9706 (make_pass_df_initialize_opt): ...new function.
9707 (pass_df_initialize_no_opt): Convert from a global struct to a
9708 subclass of rtl_opt_pass along with...
9709 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9710 (make_pass_df_initialize_no_opt): ...new function.
9711 (pass_df_finish): Convert from a global struct to a subclass of
9712 rtl_opt_pass along with...
9713 (pass_data_df_finish): ...new pass_data instance and...
9714 (make_pass_df_finish): ...new function.
9715 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9716 rtl_opt_pass along with...
9717 (pass_data_rtl_dse1): ...new pass_data instance and...
9718 (make_pass_rtl_dse1): ...new function.
9719 (pass_rtl_dse2): Convert from a global struct to a subclass of
9720 rtl_opt_pass along with...
9721 (pass_data_rtl_dse2): ...new pass_data instance and...
9722 (make_pass_rtl_dse2): ...new function.
9723 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9724 subclass of rtl_opt_pass along with...
9725 (pass_data_dwarf2_frame): ...new pass_data instance and...
9726 (make_pass_dwarf2_frame): ...new function.
9727 * except.c (pass_set_nothrow_function_flags): Convert from a global
9728 struct to a subclass of rtl_opt_pass along with...
9729 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9730 and...
9731 (make_pass_set_nothrow_function_flags): ...new function.
9732 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9733 subclass of rtl_opt_pass along with...
9734 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9735 and...
9736 (make_pass_convert_to_eh_region_ranges): ...new function.
9737 * final.c (pass_compute_alignments): Convert from a global struct to a
9738 subclass of rtl_opt_pass along with...
9739 (pass_data_compute_alignments): ...new pass_data instance and...
9740 (make_pass_compute_alignments): ...new function.
9741 (pass_final): Convert from a global struct to a subclass of
9742 rtl_opt_pass along with...
9743 (pass_data_final): ...new pass_data instance and...
9744 (make_pass_final): ...new function.
9745 (pass_shorten_branches): Convert from a global struct to a subclass of
9746 rtl_opt_pass along with...
9747 (pass_data_shorten_branches): ...new pass_data instance and...
9748 (make_pass_shorten_branches): ...new function.
9749 (pass_clean_state): Convert from a global struct to a subclass of
9750 rtl_opt_pass along with...
9751 (pass_data_clean_state): ...new pass_data instance and...
9752 (make_pass_clean_state): ...new function.
9753 * function.c (pass_instantiate_virtual_regs): Convert from a global
9754 struct to a subclass of rtl_opt_pass along with...
9755 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9756 (make_pass_instantiate_virtual_regs): ...new function.
9757 (pass_leaf_regs): Convert from a global struct to a subclass of
9758 rtl_opt_pass along with...
9759 (pass_data_leaf_regs): ...new pass_data instance and...
9760 (make_pass_leaf_regs): ...new function.
9761 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9762 subclass of rtl_opt_pass along with...
9763 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9764 and...
9765 (make_pass_thread_prologue_and_epilogue): ...new function.
9766 (pass_match_asm_constraints): Convert from a global struct to a
9767 subclass of rtl_opt_pass along with...
9768 (pass_data_match_asm_constraints): ...new pass_data instance and...
9769 (make_pass_match_asm_constraints): ...new function.
9770 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9771 subclass of rtl_opt_pass along with...
9772 (pass_data_rtl_fwprop): ...new pass_data instance and...
9773 (make_pass_rtl_fwprop): ...new function.
9774 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9775 rtl_opt_pass along with...
9776 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9777 (make_pass_rtl_fwprop_addr): ...new function.
9778 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9779 rtl_opt_pass along with...
9780 (pass_data_rtl_pre): ...new pass_data instance and...
9781 (make_pass_rtl_pre): ...new function.
9782 (pass_rtl_hoist): Convert from a global struct to a subclass of
9783 rtl_opt_pass along with...
9784 (pass_data_rtl_hoist): ...new pass_data instance and...
9785 (make_pass_rtl_hoist): ...new function.
9786 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9787 subclass of gimple_opt_pass along with...
9788 (pass_data_lower_cf): ...new pass_data instance and...
9789 (make_pass_lower_cf): ...new function.
9790 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9791 from a global struct to a subclass of gimple_opt_pass along with...
9792 (pass_data_strength_reduction): ...new pass_data instance and...
9793 (make_pass_strength_reduction): ...new function.
9794 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9795 of rtl_opt_pass along with...
9796 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9797 (make_pass_rtl_ifcvt): ...new function.
9798 (pass_if_after_combine): Convert from a global struct to a subclass of
9799 rtl_opt_pass along with...
9800 (pass_data_if_after_combine): ...new pass_data instance and...
9801 (make_pass_if_after_combine): ...new function.
9802 (pass_if_after_reload): Convert from a global struct to a subclass of
9803 rtl_opt_pass along with...
9804 (pass_data_if_after_reload): ...new pass_data instance and...
9805 (make_pass_if_after_reload): ...new function.
9806 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9807 a subclass of rtl_opt_pass along with...
9808 (pass_data_initialize_regs): ...new pass_data instance and...
9809 (make_pass_initialize_regs): ...new function.
9810 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9811 of ipa_opt_pass_d along with...
9812 (pass_data_ipa_cp): ...new pass_data instance and...
9813 (make_pass_ipa_cp): ...new function.
9814 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9815 global struct to a subclass of gimple_opt_pass along with...
9816 (pass_data_inline_parameters): ...new pass_data instance and...
9817 (make_pass_inline_parameters): ...new function.
9818 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9819 subclass of gimple_opt_pass along with...
9820 (pass_data_early_inline): ...new pass_data instance and...
9821 (make_pass_early_inline): ...new function.
9822 (pass_ipa_inline): Convert from a global struct to a subclass of
9823 ipa_opt_pass_d along with...
9824 (pass_data_ipa_inline): ...new pass_data instance and...
9825 (make_pass_ipa_inline): ...new function.
9826 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9827 struct to a subclass of gimple_opt_pass along with...
9828 (pass_data_local_pure_const): ...new pass_data instance and...
9829 (make_pass_local_pure_const): ...new function.
9830 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9831 ipa_opt_pass_d along with...
9832 (pass_data_ipa_pure_const): ...new pass_data instance and...
9833 (make_pass_ipa_pure_const): ...new function.
9834 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9835 to a subclass of ipa_opt_pass_d along with...
9836 (pass_data_ipa_reference): ...new pass_data instance and...
9837 (make_pass_ipa_reference): ...new function.
9838 * ipa-split.c (pass_split_functions): Convert from a global struct to
9839 a subclass of gimple_opt_pass along with...
9840 (pass_data_split_functions): ...new pass_data instance and...
9841 (make_pass_split_functions): ...new function.
9842 (pass_feedback_split_functions): Convert from a global struct to a
9843 subclass of gimple_opt_pass along with...
9844 (pass_data_feedback_split_functions): ...new pass_data instance and...
9845 (make_pass_feedback_split_functions): ...new function.
9846 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9847 global struct to a subclass of simple_ipa_opt_pass along with...
9848 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9849 instance and...
9850 (make_pass_ipa_function_and_variable_visibility): ...new function.
9851 (pass_ipa_free_inline_summary): Convert from a global struct to a
9852 subclass of simple_ipa_opt_pass along with...
9853 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9854 (make_pass_ipa_free_inline_summary): ...new function.
9855 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9856 subclass of ipa_opt_pass_d along with...
9857 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9858 and...
9859 (make_pass_ipa_whole_program_visibility): ...new function.
9860 (pass_ipa_profile): Convert from a global struct to a subclass of
9861 ipa_opt_pass_d along with...
9862 (pass_data_ipa_profile): ...new pass_data instance and...
9863 (make_pass_ipa_profile): ...new function.
9864 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9865 ipa_opt_pass_d along with...
9866 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9867 (make_pass_ipa_cdtor_merge): ...new function.
9868 * ira.c (pass_ira): Convert from a global struct to a subclass of
9869 rtl_opt_pass along with...
9870 (pass_data_ira): ...new pass_data instance and...
9871 (make_pass_ira): ...new function.
9872 (pass_reload): Convert from a global struct to a subclass of
9873 rtl_opt_pass along with...
9874 (pass_data_reload): ...new pass_data instance and...
9875 (make_pass_reload): ...new function.
9876 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9877 subclass of rtl_opt_pass along with...
9878 (pass_data_cleanup_barriers): ...new pass_data instance and...
9879 (make_pass_cleanup_barriers): ...new function.
9880 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9881 of rtl_opt_pass along with...
9882 (pass_data_loop2): ...new pass_data instance and...
9883 (make_pass_loop2): ...new function.
9884 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9885 rtl_opt_pass along with...
9886 (pass_data_rtl_loop_init): ...new pass_data instance and...
9887 (make_pass_rtl_loop_init): ...new function.
9888 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9889 rtl_opt_pass along with...
9890 (pass_data_rtl_loop_done): ...new pass_data instance and...
9891 (make_pass_rtl_loop_done): ...new function.
9892 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9893 subclass of rtl_opt_pass along with...
9894 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9895 (make_pass_rtl_move_loop_invariants): ...new function.
9896 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9897 rtl_opt_pass along with...
9898 (pass_data_rtl_unswitch): ...new pass_data instance and...
9899 (make_pass_rtl_unswitch): ...new function.
9900 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9901 subclass of rtl_opt_pass along with...
9902 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9903 and...
9904 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9905 (pass_rtl_doloop): Convert from a global struct to a subclass of
9906 rtl_opt_pass along with...
9907 (pass_data_rtl_doloop): ...new pass_data instance and...
9908 (make_pass_rtl_doloop): ...new function.
9909 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9910 a subclass of rtl_opt_pass along with...
9911 (pass_data_lower_subreg): ...new pass_data instance and...
9912 (make_pass_lower_subreg): ...new function.
9913 (pass_lower_subreg2): Convert from a global struct to a subclass of
9914 rtl_opt_pass along with...
9915 (pass_data_lower_subreg2): ...new pass_data instance and...
9916 (make_pass_lower_subreg2): ...new function.
9917 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9918 struct to a subclass of ipa_opt_pass_d along with...
9919 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9920 (make_pass_ipa_lto_gimple_out): ...new function.
9921 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9922 of ipa_opt_pass_d along with...
9923 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9924 (make_pass_ipa_lto_finish_out): ...new function.
9925 * mode-switching.c (pass_mode_switching): Convert from a global struct
9926 to a subclass of rtl_opt_pass along with...
9927 (pass_data_mode_switching): ...new pass_data instance and...
9928 (make_pass_mode_switching): ...new function.
9929 * modulo-sched.c (pass_sms): Convert from a global struct to a
9930 subclass of rtl_opt_pass along with...
9931 (pass_data_sms): ...new pass_data instance and...
9932 (make_pass_sms): ...new function.
9933 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9934 subclass of gimple_opt_pass along with...
9935 (pass_data_expand_omp): ...new pass_data instance and...
9936 (make_pass_expand_omp): ...new function.
9937 (pass_lower_omp): Convert from a global struct to a subclass of
9938 gimple_opt_pass along with...
9939 (pass_data_lower_omp): ...new pass_data instance and...
9940 (make_pass_lower_omp): ...new function.
9941 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9942 of gimple_opt_pass along with...
9943 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9944 (make_pass_diagnose_omp_blocks): ...new function.
9945 * passes.c (pass_early_local_passes): Convert from a global struct to
9946 a subclass of simple_ipa_opt_pass along with...
9947 (pass_data_early_local_passes): ...new pass_data instance and...
9948 (make_pass_early_local_passes): ...new function.
9949 (pass_all_early_optimizations): Convert from a global struct to a
9950 subclass of gimple_opt_pass along with...
9951 (pass_data_all_early_optimizations): ...new pass_data instance and...
9952 (make_pass_all_early_optimizations): ...new function.
9953 (pass_all_optimizations): Convert from a global struct to a subclass
9954 of gimple_opt_pass along with...
9955 (pass_data_all_optimizations): ...new pass_data instance and...
9956 (make_pass_all_optimizations): ...new function.
9957 (pass_all_optimizations_g): Convert from a global struct to a subclass
9958 of gimple_opt_pass along with...
9959 (pass_data_all_optimizations_g): ...new pass_data instance and...
9960 (make_pass_all_optimizations_g): ...new function.
9961 (pass_rest_of_compilation): Convert from a global struct to a subclass
9962 of rtl_opt_pass along with...
9963 (pass_data_rest_of_compilation): ...new pass_data instance and...
9964 (make_pass_rest_of_compilation): ...new function.
9965 (pass_postreload): Convert from a global struct to a subclass of
9966 rtl_opt_pass along with...
9967 (pass_data_postreload): ...new pass_data instance and...
9968 (make_pass_postreload): ...new function.
9969 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9970 subclass of rtl_opt_pass along with...
9971 (pass_data_gcse2): ...new pass_data instance and...
9972 (make_pass_gcse2): ...new function.
9973 * postreload.c (pass_postreload_cse): Convert from a global struct to
9974 a subclass of rtl_opt_pass along with...
9975 (pass_data_postreload_cse): ...new pass_data instance and...
9976 (make_pass_postreload_cse): ...new function.
9977 * predict.c (pass_profile): Convert from a global struct to a subclass
9978 of gimple_opt_pass along with...
9979 (pass_data_profile): ...new pass_data instance and...
9980 (make_pass_profile): ...new function.
9981 (pass_strip_predict_hints): Convert from a global struct to a subclass
9982 of gimple_opt_pass along with...
9983 (pass_data_strip_predict_hints): ...new pass_data instance and...
9984 (make_pass_strip_predict_hints): ...new function.
9985 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9986 of rtl_opt_pass along with...
9987 (pass_data_peephole2): ...new pass_data instance and...
9988 (make_pass_peephole2): ...new function.
9989 (pass_split_all_insns): Convert from a global struct to a subclass of
9990 rtl_opt_pass along with...
9991 (pass_data_split_all_insns): ...new pass_data instance and...
9992 (make_pass_split_all_insns): ...new function.
9993 (pass_split_after_reload): Convert from a global struct to a subclass
9994 of rtl_opt_pass along with...
9995 (pass_data_split_after_reload): ...new pass_data instance and...
9996 (make_pass_split_after_reload): ...new function.
9997 (pass_split_before_regstack): Convert from a global struct to a
9998 subclass of rtl_opt_pass along with...
9999 (pass_data_split_before_regstack): ...new pass_data instance and...
10000 (make_pass_split_before_regstack): ...new function.
10001 (pass_split_before_sched2): Convert from a global struct to a subclass
10002 of rtl_opt_pass along with...
10003 (pass_data_split_before_sched2): ...new pass_data instance and...
10004 (make_pass_split_before_sched2): ...new function.
10005 (pass_split_for_shorten_branches): Convert from a global struct to a
10006 subclass of rtl_opt_pass along with...
10007 (pass_data_split_for_shorten_branches): ...new pass_data instance
10008 and...
10009 (make_pass_split_for_shorten_branches): ...new function.
10010 * ree.c (pass_ree): Convert from a global struct to a subclass of
10011 rtl_opt_pass along with...
10012 (pass_data_ree): ...new pass_data instance and...
10013 (make_pass_ree): ...new function.
10014 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10015 subclass of rtl_opt_pass along with...
10016 (pass_data_stack_regs): ...new pass_data instance and...
10017 (make_pass_stack_regs): ...new function.
10018 (pass_stack_regs_run): Convert from a global struct to a subclass of
10019 rtl_opt_pass along with...
10020 (pass_data_stack_regs_run): ...new pass_data instance and...
10021 (make_pass_stack_regs_run): ...new function.
10022 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10023 subclass of rtl_opt_pass along with...
10024 (pass_data_cprop_hardreg): ...new pass_data instance and...
10025 (make_pass_cprop_hardreg): ...new function.
10026 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10027 subclass of rtl_opt_pass along with...
10028 (pass_data_reginfo_init): ...new pass_data instance and...
10029 (make_pass_reginfo_init): ...new function.
10030 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10031 of rtl_opt_pass along with...
10032 (pass_data_regmove): ...new pass_data instance and...
10033 (make_pass_regmove): ...new function.
10034 * regrename.c (pass_regrename): Convert from a global struct to a
10035 subclass of rtl_opt_pass along with...
10036 (pass_data_regrename): ...new pass_data instance and...
10037 (make_pass_regrename): ...new function.
10038 * reorg.c (pass_delay_slots): Convert from a global struct to a
10039 subclass of rtl_opt_pass along with...
10040 (pass_data_delay_slots): ...new pass_data instance and...
10041 (make_pass_delay_slots): ...new function.
10042 (pass_machine_reorg): Convert from a global struct to a subclass of
10043 rtl_opt_pass along with...
10044 (pass_data_machine_reorg): ...new pass_data instance and...
10045 (make_pass_machine_reorg): ...new function.
10046 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10047 of rtl_opt_pass along with...
10048 (pass_data_sched): ...new pass_data instance and...
10049 (make_pass_sched): ...new function.
10050 (pass_sched2): Convert from a global struct to a subclass of
10051 rtl_opt_pass along with...
10052 (pass_data_sched2): ...new pass_data instance and...
10053 (make_pass_sched2): ...new function.
10054 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10055 to a subclass of rtl_opt_pass along with...
10056 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10057 (make_pass_stack_ptr_mod): ...new function.
10058 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10059 to a subclass of rtl_opt_pass along with...
10060 (pass_data_rtl_store_motion): ...new pass_data instance and...
10061 (make_pass_rtl_store_motion): ...new function.
10062 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10063 of gimple_opt_pass along with...
10064 (pass_data_tracer): ...new pass_data instance and...
10065 (make_pass_tracer): ...new function.
10066 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10067 to a subclass of gimple_opt_pass along with...
10068 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10069 (make_pass_diagnose_tm_blocks): ...new function.
10070 (pass_lower_tm): Convert from a global struct to a subclass of
10071 gimple_opt_pass along with...
10072 (pass_data_lower_tm): ...new pass_data instance and...
10073 (make_pass_lower_tm): ...new function.
10074 (pass_tm_init): Convert from a global struct to a subclass of
10075 gimple_opt_pass along with...
10076 (pass_data_tm_init): ...new pass_data instance and...
10077 (make_pass_tm_init): ...new function.
10078 (pass_tm_mark): Convert from a global struct to a subclass of
10079 gimple_opt_pass along with...
10080 (pass_data_tm_mark): ...new pass_data instance and...
10081 (make_pass_tm_mark): ...new function.
10082 (pass_tm_edges): Convert from a global struct to a subclass of
10083 gimple_opt_pass along with...
10084 (pass_data_tm_edges): ...new pass_data instance and...
10085 (make_pass_tm_edges): ...new function.
10086 (pass_tm_memopt): Convert from a global struct to a subclass of
10087 gimple_opt_pass along with...
10088 (pass_data_tm_memopt): ...new pass_data instance and...
10089 (make_pass_tm_memopt): ...new function.
10090 (pass_ipa_tm): Convert from a global struct to a subclass of
10091 simple_ipa_opt_pass along with...
10092 (pass_data_ipa_tm): ...new pass_data instance and...
10093 (make_pass_ipa_tm): ...new function.
10094 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10095 subclass of gimple_opt_pass along with...
10096 (pass_data_call_cdce): ...new pass_data instance and...
10097 (make_pass_call_cdce): ...new function.
10098 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10099 subclass of gimple_opt_pass along with...
10100 (pass_data_build_cfg): ...new pass_data instance and...
10101 (make_pass_build_cfg): ...new function.
10102 (pass_split_crit_edges): Convert from a global struct to a subclass of
10103 gimple_opt_pass along with...
10104 (pass_data_split_crit_edges): ...new pass_data instance and...
10105 (make_pass_split_crit_edges): ...new function.
10106 (pass_warn_function_return): Convert from a global struct to a
10107 subclass of gimple_opt_pass along with...
10108 (pass_data_warn_function_return): ...new pass_data instance and...
10109 (make_pass_warn_function_return): ...new function.
10110 (pass_warn_function_noreturn): Convert from a global struct to a
10111 subclass of gimple_opt_pass along with...
10112 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10113 (make_pass_warn_function_noreturn): ...new function.
10114 (pass_warn_unused_result): Convert from a global struct to a subclass
10115 of gimple_opt_pass along with...
10116 (pass_data_warn_unused_result): ...new pass_data instance and...
10117 (make_pass_warn_unused_result): ...new function.
10118 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10119 a subclass of gimple_opt_pass along with...
10120 (pass_data_merge_phi): ...new pass_data instance and...
10121 (make_pass_merge_phi): ...new function.
10122 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10123 a subclass of gimple_opt_pass along with...
10124 (pass_data_lower_complex): ...new pass_data instance and...
10125 (make_pass_lower_complex): ...new function.
10126 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10127 gimple_opt_pass along with...
10128 (pass_data_lower_complex_O0): ...new pass_data instance and...
10129 (make_pass_lower_complex_O0): ...new function.
10130 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10131 subclass of gimple_opt_pass along with...
10132 (pass_data_lower_eh): ...new pass_data instance and...
10133 (make_pass_lower_eh): ...new function.
10134 (pass_refactor_eh): Convert from a global struct to a subclass of
10135 gimple_opt_pass along with...
10136 (pass_data_refactor_eh): ...new pass_data instance and...
10137 (make_pass_refactor_eh): ...new function.
10138 (pass_lower_resx): Convert from a global struct to a subclass of
10139 gimple_opt_pass along with...
10140 (pass_data_lower_resx): ...new pass_data instance and...
10141 (make_pass_lower_resx): ...new function.
10142 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10143 of gimple_opt_pass along with...
10144 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10145 (make_pass_lower_eh_dispatch): ...new function.
10146 (pass_cleanup_eh): Convert from a global struct to a subclass of
10147 gimple_opt_pass along with...
10148 (pass_data_cleanup_eh): ...new pass_data instance and...
10149 (make_pass_cleanup_eh): ...new function.
10150 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10151 to a subclass of simple_ipa_opt_pass along with...
10152 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10153 (make_pass_ipa_lower_emutls): ...new function.
10154 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10155 a subclass of gimple_opt_pass along with...
10156 (pass_data_if_conversion): ...new pass_data instance and...
10157 (make_pass_if_conversion): ...new function.
10158 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10159 subclass of gimple_opt_pass along with...
10160 (pass_data_build_ssa): ...new pass_data instance and...
10161 (make_pass_build_ssa): ...new function.
10162 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10163 global struct to a subclass of gimple_opt_pass along with...
10164 (pass_data_loop_distribution): ...new pass_data instance and...
10165 (make_pass_loop_distribution): ...new function.
10166 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10167 subclass of gimple_opt_pass along with...
10168 (pass_data_mudflap_1): ...new pass_data instance and...
10169 (make_pass_mudflap_1): ...new function.
10170 (pass_mudflap_2): Convert from a global struct to a subclass of
10171 gimple_opt_pass along with...
10172 (pass_data_mudflap_2): ...new pass_data instance and...
10173 (make_pass_mudflap_2): ...new function.
10174 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10175 subclass of gimple_opt_pass along with...
10176 (pass_data_mudflap_1): ...new pass_data instance and...
10177 (make_pass_mudflap_1): ...new function.
10178 (pass_mudflap_2): Convert from a global struct to a subclass of
10179 gimple_opt_pass along with...
10180 (pass_data_mudflap_2): ...new pass_data instance and...
10181 (make_pass_mudflap_2): ...new function.
10182 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10183 gimple_opt_pass along with...
10184 (pass_data_nrv): ...new pass_data instance and...
10185 (make_pass_nrv): ...new function.
10186 (pass_return_slot): Convert from a global struct to a subclass of
10187 gimple_opt_pass along with...
10188 (pass_data_return_slot): ...new pass_data instance and...
10189 (make_pass_return_slot): ...new function.
10190 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10191 to a subclass of gimple_opt_pass along with...
10192 (pass_data_object_sizes): ...new pass_data instance and...
10193 (make_pass_object_sizes): ...new function.
10194 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10195 global struct to a subclass of gimple_opt_pass along with...
10196 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10197 and...
10198 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10199 (pass_fixup_cfg): Convert from a global struct to a subclass of
10200 gimple_opt_pass along with...
10201 (pass_data_fixup_cfg): ...new pass_data instance and...
10202 (make_pass_fixup_cfg): ...new function.
10203 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10204 (make_pass_mudflap_1): ...new function.
10205 (pass_mudflap_2): Replace declaration with that of...
10206 (make_pass_mudflap_2): ...new function.
10207 (pass_asan): Replace declaration with that of...
10208 (make_pass_asan): ...new function.
10209 (pass_asan_O0): Replace declaration with that of...
10210 (make_pass_asan_O0): ...new function.
10211 (pass_tsan): Replace declaration with that of...
10212 (make_pass_tsan): ...new function.
10213 (pass_tsan_O0): Replace declaration with that of...
10214 (make_pass_tsan_O0): ...new function.
10215 (pass_lower_cf): Replace declaration with that of...
10216 (make_pass_lower_cf): ...new function.
10217 (pass_refactor_eh): Replace declaration with that of...
10218 (make_pass_refactor_eh): ...new function.
10219 (pass_lower_eh): Replace declaration with that of...
10220 (make_pass_lower_eh): ...new function.
10221 (pass_lower_eh_dispatch): Replace declaration with that of...
10222 (make_pass_lower_eh_dispatch): ...new function.
10223 (pass_lower_resx): Replace declaration with that of...
10224 (make_pass_lower_resx): ...new function.
10225 (pass_build_cfg): Replace declaration with that of...
10226 (make_pass_build_cfg): ...new function.
10227 (pass_early_tree_profile): Replace declaration with that of...
10228 (make_pass_early_tree_profile): ...new function.
10229 (pass_cleanup_eh): Replace declaration with that of...
10230 (make_pass_cleanup_eh): ...new function.
10231 (pass_sra): Replace declaration with that of...
10232 (make_pass_sra): ...new function.
10233 (pass_sra_early): Replace declaration with that of...
10234 (make_pass_sra_early): ...new function.
10235 (pass_early_ipa_sra): Replace declaration with that of...
10236 (make_pass_early_ipa_sra): ...new function.
10237 (pass_tail_recursion): Replace declaration with that of...
10238 (make_pass_tail_recursion): ...new function.
10239 (pass_tail_calls): Replace declaration with that of...
10240 (make_pass_tail_calls): ...new function.
10241 (pass_tree_loop): Replace declaration with that of...
10242 (make_pass_tree_loop): ...new function.
10243 (pass_tree_loop_init): Replace declaration with that of...
10244 (make_pass_tree_loop_init): ...new function.
10245 (pass_lim): Replace declaration with that of...
10246 (make_pass_lim): ...new function.
10247 (pass_tree_unswitch): Replace declaration with that of...
10248 (make_pass_tree_unswitch): ...new function.
10249 (pass_predcom): Replace declaration with that of...
10250 (make_pass_predcom): ...new function.
10251 (pass_iv_canon): Replace declaration with that of...
10252 (make_pass_iv_canon): ...new function.
10253 (pass_scev_cprop): Replace declaration with that of...
10254 (make_pass_scev_cprop): ...new function.
10255 (pass_empty_loop): Replace declaration with that of...
10256 (make_pass_empty_loop): ...new function.
10257 (pass_record_bounds): Replace declaration with that of...
10258 (make_pass_record_bounds): ...new function.
10259 (pass_graphite): Replace declaration with that of...
10260 (make_pass_graphite): ...new function.
10261 (pass_graphite_transforms): Replace declaration with that of...
10262 (make_pass_graphite_transforms): ...new function.
10263 (pass_if_conversion): Replace declaration with that of...
10264 (make_pass_if_conversion): ...new function.
10265 (pass_loop_distribution): Replace declaration with that of...
10266 (make_pass_loop_distribution): ...new function.
10267 (pass_vectorize): Replace declaration with that of...
10268 (make_pass_vectorize): ...new function.
10269 (pass_slp_vectorize): Replace declaration with that of...
10270 (make_pass_slp_vectorize): ...new function.
10271 (pass_complete_unroll): Replace declaration with that of...
10272 (make_pass_complete_unroll): ...new function.
10273 (pass_complete_unrolli): Replace declaration with that of...
10274 (make_pass_complete_unrolli): ...new function.
10275 (pass_parallelize_loops): Replace declaration with that of...
10276 (make_pass_parallelize_loops): ...new function.
10277 (pass_loop_prefetch): Replace declaration with that of...
10278 (make_pass_loop_prefetch): ...new function.
10279 (pass_iv_optimize): Replace declaration with that of...
10280 (make_pass_iv_optimize): ...new function.
10281 (pass_tree_loop_done): Replace declaration with that of...
10282 (make_pass_tree_loop_done): ...new function.
10283 (pass_ch): Replace declaration with that of...
10284 (make_pass_ch): ...new function.
10285 (pass_ccp): Replace declaration with that of...
10286 (make_pass_ccp): ...new function.
10287 (pass_phi_only_cprop): Replace declaration with that of...
10288 (make_pass_phi_only_cprop): ...new function.
10289 (pass_build_ssa): Replace declaration with that of...
10290 (make_pass_build_ssa): ...new function.
10291 (pass_build_alias): Replace declaration with that of...
10292 (make_pass_build_alias): ...new function.
10293 (pass_build_ealias): Replace declaration with that of...
10294 (make_pass_build_ealias): ...new function.
10295 (pass_dominator): Replace declaration with that of...
10296 (make_pass_dominator): ...new function.
10297 (pass_dce): Replace declaration with that of...
10298 (make_pass_dce): ...new function.
10299 (pass_dce_loop): Replace declaration with that of...
10300 (make_pass_dce_loop): ...new function.
10301 (pass_cd_dce): Replace declaration with that of...
10302 (make_pass_cd_dce): ...new function.
10303 (pass_call_cdce): Replace declaration with that of...
10304 (make_pass_call_cdce): ...new function.
10305 (pass_merge_phi): Replace declaration with that of...
10306 (make_pass_merge_phi): ...new function.
10307 (pass_split_crit_edges): Replace declaration with that of...
10308 (make_pass_split_crit_edges): ...new function.
10309 (pass_pre): Replace declaration with that of...
10310 (make_pass_pre): ...new function.
10311 (pass_profile): Replace declaration with that of...
10312 (make_pass_profile): ...new function.
10313 (pass_strip_predict_hints): Replace declaration with that of...
10314 (make_pass_strip_predict_hints): ...new function.
10315 (pass_lower_complex_O0): Replace declaration with that of...
10316 (make_pass_lower_complex_O0): ...new function.
10317 (pass_lower_complex): Replace declaration with that of...
10318 (make_pass_lower_complex): ...new function.
10319 (pass_lower_vector): Replace declaration with that of...
10320 (make_pass_lower_vector): ...new function.
10321 (pass_lower_vector_ssa): Replace declaration with that of...
10322 (make_pass_lower_vector_ssa): ...new function.
10323 (pass_lower_omp): Replace declaration with that of...
10324 (make_pass_lower_omp): ...new function.
10325 (pass_diagnose_omp_blocks): Replace declaration with that of...
10326 (make_pass_diagnose_omp_blocks): ...new function.
10327 (pass_expand_omp): Replace declaration with that of...
10328 (make_pass_expand_omp): ...new function.
10329 (pass_expand_omp_ssa): Replace declaration with that of...
10330 (make_pass_expand_omp_ssa): ...new function.
10331 (pass_object_sizes): Replace declaration with that of...
10332 (make_pass_object_sizes): ...new function.
10333 (pass_strlen): Replace declaration with that of...
10334 (make_pass_strlen): ...new function.
10335 (pass_fold_builtins): Replace declaration with that of...
10336 (make_pass_fold_builtins): ...new function.
10337 (pass_stdarg): Replace declaration with that of...
10338 (make_pass_stdarg): ...new function.
10339 (pass_early_warn_uninitialized): Replace declaration with that of...
10340 (make_pass_early_warn_uninitialized): ...new function.
10341 (pass_late_warn_uninitialized): Replace declaration with that of...
10342 (make_pass_late_warn_uninitialized): ...new function.
10343 (pass_cse_reciprocals): Replace declaration with that of...
10344 (make_pass_cse_reciprocals): ...new function.
10345 (pass_cse_sincos): Replace declaration with that of...
10346 (make_pass_cse_sincos): ...new function.
10347 (pass_optimize_bswap): Replace declaration with that of...
10348 (make_pass_optimize_bswap): ...new function.
10349 (pass_optimize_widening_mul): Replace declaration with that of...
10350 (make_pass_optimize_widening_mul): ...new function.
10351 (pass_warn_function_return): Replace declaration with that of...
10352 (make_pass_warn_function_return): ...new function.
10353 (pass_warn_function_noreturn): Replace declaration with that of...
10354 (make_pass_warn_function_noreturn): ...new function.
10355 (pass_cselim): Replace declaration with that of...
10356 (make_pass_cselim): ...new function.
10357 (pass_phiopt): Replace declaration with that of...
10358 (make_pass_phiopt): ...new function.
10359 (pass_forwprop): Replace declaration with that of...
10360 (make_pass_forwprop): ...new function.
10361 (pass_phiprop): Replace declaration with that of...
10362 (make_pass_phiprop): ...new function.
10363 (pass_tree_ifcombine): Replace declaration with that of...
10364 (make_pass_tree_ifcombine): ...new function.
10365 (pass_dse): Replace declaration with that of...
10366 (make_pass_dse): ...new function.
10367 (pass_nrv): Replace declaration with that of...
10368 (make_pass_nrv): ...new function.
10369 (pass_rename_ssa_copies): Replace declaration with that of...
10370 (make_pass_rename_ssa_copies): ...new function.
10371 (pass_sink_code): Replace declaration with that of...
10372 (make_pass_sink_code): ...new function.
10373 (pass_fre): Replace declaration with that of...
10374 (make_pass_fre): ...new function.
10375 (pass_check_data_deps): Replace declaration with that of...
10376 (make_pass_check_data_deps): ...new function.
10377 (pass_copy_prop): Replace declaration with that of...
10378 (make_pass_copy_prop): ...new function.
10379 (pass_vrp): Replace declaration with that of...
10380 (make_pass_vrp): ...new function.
10381 (pass_uncprop): Replace declaration with that of...
10382 (make_pass_uncprop): ...new function.
10383 (pass_return_slot): Replace declaration with that of...
10384 (make_pass_return_slot): ...new function.
10385 (pass_reassoc): Replace declaration with that of...
10386 (make_pass_reassoc): ...new function.
10387 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10388 (make_pass_rebuild_cgraph_edges): ...new function.
10389 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10390 (make_pass_remove_cgraph_callee_edges): ...new function.
10391 (pass_build_cgraph_edges): Replace declaration with that of...
10392 (make_pass_build_cgraph_edges): ...new function.
10393 (pass_local_pure_const): Replace declaration with that of...
10394 (make_pass_local_pure_const): ...new function.
10395 (pass_tracer): Replace declaration with that of...
10396 (make_pass_tracer): ...new function.
10397 (pass_warn_unused_result): Replace declaration with that of...
10398 (make_pass_warn_unused_result): ...new function.
10399 (pass_diagnose_tm_blocks): Replace declaration with that of...
10400 (make_pass_diagnose_tm_blocks): ...new function.
10401 (pass_lower_tm): Replace declaration with that of...
10402 (make_pass_lower_tm): ...new function.
10403 (pass_tm_init): Replace declaration with that of...
10404 (make_pass_tm_init): ...new function.
10405 (pass_tm_mark): Replace declaration with that of...
10406 (make_pass_tm_mark): ...new function.
10407 (pass_tm_memopt): Replace declaration with that of...
10408 (make_pass_tm_memopt): ...new function.
10409 (pass_tm_edges): Replace declaration with that of...
10410 (make_pass_tm_edges): ...new function.
10411 (pass_split_functions): Replace declaration with that of...
10412 (make_pass_split_functions): ...new function.
10413 (pass_feedback_split_functions): Replace declaration with that of...
10414 (make_pass_feedback_split_functions): ...new function.
10415 (pass_strength_reduction): Replace declaration with that of...
10416 (make_pass_strength_reduction): ...new function.
10417 (pass_ipa_lower_emutls): Replace declaration with that of...
10418 (make_pass_ipa_lower_emutls): ...new function.
10419 (pass_ipa_function_and_variable_visibility): Replace declaration with
10420 that of...
10421 (make_pass_ipa_function_and_variable_visibility): ...new function.
10422 (pass_ipa_tree_profile): Replace declaration with that of...
10423 (make_pass_ipa_tree_profile): ...new function.
10424 (pass_early_local_passes): Replace declaration with that of...
10425 (make_pass_early_local_passes): ...new function.
10426 (pass_ipa_whole_program_visibility): Replace declaration with that
10427 of...
10428 (make_pass_ipa_whole_program_visibility): ...new function.
10429 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10430 (make_pass_ipa_lto_gimple_out): ...new function.
10431 (pass_ipa_increase_alignment): Replace declaration with that of...
10432 (make_pass_ipa_increase_alignment): ...new function.
10433 (pass_ipa_inline): Replace declaration with that of...
10434 (make_pass_ipa_inline): ...new function.
10435 (pass_ipa_free_lang_data): Replace declaration with that of...
10436 (make_pass_ipa_free_lang_data): ...new function.
10437 (pass_ipa_free_inline_summary): Replace declaration with that of...
10438 (make_pass_ipa_free_inline_summary): ...new function.
10439 (pass_ipa_cp): Replace declaration with that of...
10440 (make_pass_ipa_cp): ...new function.
10441 (pass_ipa_reference): Replace declaration with that of...
10442 (make_pass_ipa_reference): ...new function.
10443 (pass_ipa_pure_const): Replace declaration with that of...
10444 (make_pass_ipa_pure_const): ...new function.
10445 (pass_ipa_pta): Replace declaration with that of...
10446 (make_pass_ipa_pta): ...new function.
10447 (pass_ipa_lto_finish_out): Replace declaration with that of...
10448 (make_pass_ipa_lto_finish_out): ...new function.
10449 (pass_ipa_tm): Replace declaration with that of...
10450 (make_pass_ipa_tm): ...new function.
10451 (pass_ipa_profile): Replace declaration with that of...
10452 (make_pass_ipa_profile): ...new function.
10453 (pass_ipa_cdtor_merge): Replace declaration with that of...
10454 (make_pass_ipa_cdtor_merge): ...new function.
10455 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10456 of...
10457 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10458 (pass_init_datastructures): Replace declaration with that of...
10459 (make_pass_init_datastructures): ...new function.
10460 (pass_fixup_cfg): Replace declaration with that of...
10461 (make_pass_fixup_cfg): ...new function.
10462 (pass_expand): Replace declaration with that of...
10463 (make_pass_expand): ...new function.
10464 (pass_instantiate_virtual_regs): Replace declaration with that of...
10465 (make_pass_instantiate_virtual_regs): ...new function.
10466 (pass_rtl_fwprop): Replace declaration with that of...
10467 (make_pass_rtl_fwprop): ...new function.
10468 (pass_rtl_fwprop_addr): Replace declaration with that of...
10469 (make_pass_rtl_fwprop_addr): ...new function.
10470 (pass_jump): Replace declaration with that of...
10471 (make_pass_jump): ...new function.
10472 (pass_jump2): Replace declaration with that of...
10473 (make_pass_jump2): ...new function.
10474 (pass_lower_subreg): Replace declaration with that of...
10475 (make_pass_lower_subreg): ...new function.
10476 (pass_cse): Replace declaration with that of...
10477 (make_pass_cse): ...new function.
10478 (pass_fast_rtl_dce): Replace declaration with that of...
10479 (make_pass_fast_rtl_dce): ...new function.
10480 (pass_ud_rtl_dce): Replace declaration with that of...
10481 (make_pass_ud_rtl_dce): ...new function.
10482 (pass_rtl_dce): Replace declaration with that of...
10483 (make_pass_rtl_dce): ...new function.
10484 (pass_rtl_dse1): Replace declaration with that of...
10485 (make_pass_rtl_dse1): ...new function.
10486 (pass_rtl_dse2): Replace declaration with that of...
10487 (make_pass_rtl_dse2): ...new function.
10488 (pass_rtl_dse3): Replace declaration with that of...
10489 (make_pass_rtl_dse3): ...new function.
10490 (pass_rtl_cprop): Replace declaration with that of...
10491 (make_pass_rtl_cprop): ...new function.
10492 (pass_rtl_pre): Replace declaration with that of...
10493 (make_pass_rtl_pre): ...new function.
10494 (pass_rtl_hoist): Replace declaration with that of...
10495 (make_pass_rtl_hoist): ...new function.
10496 (pass_rtl_store_motion): Replace declaration with that of...
10497 (make_pass_rtl_store_motion): ...new function.
10498 (pass_cse_after_global_opts): Replace declaration with that of...
10499 (make_pass_cse_after_global_opts): ...new function.
10500 (pass_rtl_ifcvt): Replace declaration with that of...
10501 (make_pass_rtl_ifcvt): ...new function.
10502 (pass_into_cfg_layout_mode): Replace declaration with that of...
10503 (make_pass_into_cfg_layout_mode): ...new function.
10504 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10505 (make_pass_outof_cfg_layout_mode): ...new function.
10506 (pass_loop2): Replace declaration with that of...
10507 (make_pass_loop2): ...new function.
10508 (pass_rtl_loop_init): Replace declaration with that of...
10509 (make_pass_rtl_loop_init): ...new function.
10510 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10511 (make_pass_rtl_move_loop_invariants): ...new function.
10512 (pass_rtl_unswitch): Replace declaration with that of...
10513 (make_pass_rtl_unswitch): ...new function.
10514 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10515 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10516 (pass_rtl_doloop): Replace declaration with that of...
10517 (make_pass_rtl_doloop): ...new function.
10518 (pass_rtl_loop_done): Replace declaration with that of...
10519 (make_pass_rtl_loop_done): ...new function.
10520 (pass_web): Replace declaration with that of...
10521 (make_pass_web): ...new function.
10522 (pass_cse2): Replace declaration with that of...
10523 (make_pass_cse2): ...new function.
10524 (pass_df_initialize_opt): Replace declaration with that of...
10525 (make_pass_df_initialize_opt): ...new function.
10526 (pass_df_initialize_no_opt): Replace declaration with that of...
10527 (make_pass_df_initialize_no_opt): ...new function.
10528 (pass_reginfo_init): Replace declaration with that of...
10529 (make_pass_reginfo_init): ...new function.
10530 (pass_inc_dec): Replace declaration with that of...
10531 (make_pass_inc_dec): ...new function.
10532 (pass_stack_ptr_mod): Replace declaration with that of...
10533 (make_pass_stack_ptr_mod): ...new function.
10534 (pass_initialize_regs): Replace declaration with that of...
10535 (make_pass_initialize_regs): ...new function.
10536 (pass_combine): Replace declaration with that of...
10537 (make_pass_combine): ...new function.
10538 (pass_if_after_combine): Replace declaration with that of...
10539 (make_pass_if_after_combine): ...new function.
10540 (pass_ree): Replace declaration with that of...
10541 (make_pass_ree): ...new function.
10542 (pass_partition_blocks): Replace declaration with that of...
10543 (make_pass_partition_blocks): ...new function.
10544 (pass_match_asm_constraints): Replace declaration with that of...
10545 (make_pass_match_asm_constraints): ...new function.
10546 (pass_regmove): Replace declaration with that of...
10547 (make_pass_regmove): ...new function.
10548 (pass_split_all_insns): Replace declaration with that of...
10549 (make_pass_split_all_insns): ...new function.
10550 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10551 (make_pass_fast_rtl_byte_dce): ...new function.
10552 (pass_lower_subreg2): Replace declaration with that of...
10553 (make_pass_lower_subreg2): ...new function.
10554 (pass_mode_switching): Replace declaration with that of...
10555 (make_pass_mode_switching): ...new function.
10556 (pass_sms): Replace declaration with that of...
10557 (make_pass_sms): ...new function.
10558 (pass_sched): Replace declaration with that of...
10559 (make_pass_sched): ...new function.
10560 (pass_ira): Replace declaration with that of...
10561 (make_pass_ira): ...new function.
10562 (pass_reload): Replace declaration with that of...
10563 (make_pass_reload): ...new function.
10564 (pass_clean_state): Replace declaration with that of...
10565 (make_pass_clean_state): ...new function.
10566 (pass_branch_prob): Replace declaration with that of...
10567 (make_pass_branch_prob): ...new function.
10568 (pass_value_profile_transformations): Replace declaration with that
10569 of...
10570 (make_pass_value_profile_transformations): ...new function.
10571 (pass_postreload_cse): Replace declaration with that of...
10572 (make_pass_postreload_cse): ...new function.
10573 (pass_gcse2): Replace declaration with that of...
10574 (make_pass_gcse2): ...new function.
10575 (pass_split_after_reload): Replace declaration with that of...
10576 (make_pass_split_after_reload): ...new function.
10577 (pass_branch_target_load_optimize1): Replace declaration with that
10578 of...
10579 (make_pass_branch_target_load_optimize1): ...new function.
10580 (pass_thread_prologue_and_epilogue): Replace declaration with that
10581 of...
10582 (make_pass_thread_prologue_and_epilogue): ...new function.
10583 (pass_stack_adjustments): Replace declaration with that of...
10584 (make_pass_stack_adjustments): ...new function.
10585 (pass_peephole2): Replace declaration with that of...
10586 (make_pass_peephole2): ...new function.
10587 (pass_if_after_reload): Replace declaration with that of...
10588 (make_pass_if_after_reload): ...new function.
10589 (pass_regrename): Replace declaration with that of...
10590 (make_pass_regrename): ...new function.
10591 (pass_cprop_hardreg): Replace declaration with that of...
10592 (make_pass_cprop_hardreg): ...new function.
10593 (pass_reorder_blocks): Replace declaration with that of...
10594 (make_pass_reorder_blocks): ...new function.
10595 (pass_branch_target_load_optimize2): Replace declaration with that
10596 of...
10597 (make_pass_branch_target_load_optimize2): ...new function.
10598 (pass_leaf_regs): Replace declaration with that of...
10599 (make_pass_leaf_regs): ...new function.
10600 (pass_split_before_sched2): Replace declaration with that of...
10601 (make_pass_split_before_sched2): ...new function.
10602 (pass_compare_elim_after_reload): Replace declaration with that of...
10603 (make_pass_compare_elim_after_reload): ...new function.
10604 (pass_sched2): Replace declaration with that of...
10605 (make_pass_sched2): ...new function.
10606 (pass_stack_regs): Replace declaration with that of...
10607 (make_pass_stack_regs): ...new function.
10608 (pass_stack_regs_run): Replace declaration with that of...
10609 (make_pass_stack_regs_run): ...new function.
10610 (pass_df_finish): Replace declaration with that of...
10611 (make_pass_df_finish): ...new function.
10612 (pass_compute_alignments): Replace declaration with that of...
10613 (make_pass_compute_alignments): ...new function.
10614 (pass_duplicate_computed_gotos): Replace declaration with that of...
10615 (make_pass_duplicate_computed_gotos): ...new function.
10616 (pass_variable_tracking): Replace declaration with that of...
10617 (make_pass_variable_tracking): ...new function.
10618 (pass_free_cfg): Replace declaration with that of...
10619 (make_pass_free_cfg): ...new function.
10620 (pass_machine_reorg): Replace declaration with that of...
10621 (make_pass_machine_reorg): ...new function.
10622 (pass_cleanup_barriers): Replace declaration with that of...
10623 (make_pass_cleanup_barriers): ...new function.
10624 (pass_delay_slots): Replace declaration with that of...
10625 (make_pass_delay_slots): ...new function.
10626 (pass_split_for_shorten_branches): Replace declaration with that of...
10627 (make_pass_split_for_shorten_branches): ...new function.
10628 (pass_split_before_regstack): Replace declaration with that of...
10629 (make_pass_split_before_regstack): ...new function.
10630 (pass_convert_to_eh_region_ranges): Replace declaration with that
10631 of...
10632 (make_pass_convert_to_eh_region_ranges): ...new function.
10633 (pass_shorten_branches): Replace declaration with that of...
10634 (make_pass_shorten_branches): ...new function.
10635 (pass_set_nothrow_function_flags): Replace declaration with that of...
10636 (make_pass_set_nothrow_function_flags): ...new function.
10637 (pass_dwarf2_frame): Replace declaration with that of...
10638 (make_pass_dwarf2_frame): ...new function.
10639 (pass_final): Replace declaration with that of...
10640 (make_pass_final): ...new function.
10641 (pass_rtl_seqabstr): Replace declaration with that of...
10642 (make_pass_rtl_seqabstr): ...new function.
10643 (pass_release_ssa_names): Replace declaration with that of...
10644 (make_pass_release_ssa_names): ...new function.
10645 (pass_early_inline): Replace declaration with that of...
10646 (make_pass_early_inline): ...new function.
10647 (pass_inline_parameters): Replace declaration with that of...
10648 (make_pass_inline_parameters): ...new function.
10649 (pass_update_address_taken): Replace declaration with that of...
10650 (make_pass_update_address_taken): ...new function.
10651 (pass_convert_switch): Replace declaration with that of...
10652 (make_pass_convert_switch): ...new function.
10653 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10654 to a subclass of simple_ipa_opt_pass along with...
10655 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10656 (make_pass_ipa_tree_profile): ...new function.
10657 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10658 subclass of gimple_opt_pass along with...
10659 (pass_data_sra_early): ...new pass_data instance and...
10660 (make_pass_sra_early): ...new function.
10661 (pass_sra): Convert from a global struct to a subclass of
10662 gimple_opt_pass along with...
10663 (pass_data_sra): ...new pass_data instance and...
10664 (make_pass_sra): ...new function.
10665 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10666 gimple_opt_pass along with...
10667 (pass_data_early_ipa_sra): ...new pass_data instance and...
10668 (make_pass_early_ipa_sra): ...new function.
10669 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10670 subclass of gimple_opt_pass along with...
10671 (pass_data_ccp): ...new pass_data instance and...
10672 (make_pass_ccp): ...new function.
10673 (pass_fold_builtins): Convert from a global struct to a subclass of
10674 gimple_opt_pass along with...
10675 (pass_data_fold_builtins): ...new pass_data instance and...
10676 (make_pass_fold_builtins): ...new function.
10677 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10678 subclass of gimple_opt_pass along with...
10679 (pass_data_copy_prop): ...new pass_data instance and...
10680 (make_pass_copy_prop): ...new function.
10681 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10682 global struct to a subclass of gimple_opt_pass along with...
10683 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10684 (make_pass_rename_ssa_copies): ...new function.
10685 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10686 subclass of gimple_opt_pass along with...
10687 (pass_data_dce): ...new pass_data instance and...
10688 (make_pass_dce): ...new function.
10689 (pass_dce_loop): Convert from a global struct to a subclass of
10690 gimple_opt_pass along with...
10691 (pass_data_dce_loop): ...new pass_data instance and...
10692 (make_pass_dce_loop): ...new function.
10693 (pass_cd_dce): Convert from a global struct to a subclass of
10694 gimple_opt_pass along with...
10695 (pass_data_cd_dce): ...new pass_data instance and...
10696 (make_pass_cd_dce): ...new function.
10697 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10698 subclass of gimple_opt_pass along with...
10699 (pass_data_dominator): ...new pass_data instance and...
10700 (make_pass_dominator): ...new function.
10701 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10702 gimple_opt_pass along with...
10703 (pass_data_phi_only_cprop): ...new pass_data instance and...
10704 (make_pass_phi_only_cprop): ...new function.
10705 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10706 subclass of gimple_opt_pass along with...
10707 (pass_data_dse): ...new pass_data instance and...
10708 (make_pass_dse): ...new function.
10709 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10710 a subclass of gimple_opt_pass along with...
10711 (pass_data_forwprop): ...new pass_data instance and...
10712 (make_pass_forwprop): ...new function.
10713 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10714 struct to a subclass of gimple_opt_pass along with...
10715 (pass_data_tree_ifcombine): ...new pass_data instance and...
10716 (make_pass_tree_ifcombine): ...new function.
10717 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10718 subclass of gimple_opt_pass along with...
10719 (pass_data_ch): ...new pass_data instance and...
10720 (make_pass_ch): ...new function.
10721 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10722 subclass of gimple_opt_pass along with...
10723 (pass_data_tree_loop): ...new pass_data instance and...
10724 (make_pass_tree_loop): ...new function.
10725 (pass_tree_loop_init): Convert from a global struct to a subclass of
10726 gimple_opt_pass along with...
10727 (pass_data_tree_loop_init): ...new pass_data instance and...
10728 (make_pass_tree_loop_init): ...new function.
10729 (pass_lim): Convert from a global struct to a subclass of
10730 gimple_opt_pass along with...
10731 (pass_data_lim): ...new pass_data instance and...
10732 (make_pass_lim): ...new function.
10733 (pass_tree_unswitch): Convert from a global struct to a subclass of
10734 gimple_opt_pass along with...
10735 (pass_data_tree_unswitch): ...new pass_data instance and...
10736 (make_pass_tree_unswitch): ...new function.
10737 (pass_predcom): Convert from a global struct to a subclass of
10738 gimple_opt_pass along with...
10739 (pass_data_predcom): ...new pass_data instance and...
10740 (make_pass_predcom): ...new function.
10741 (pass_vectorize): Convert from a global struct to a subclass of
10742 gimple_opt_pass along with...
10743 (pass_data_vectorize): ...new pass_data instance and...
10744 (make_pass_vectorize): ...new function.
10745 (pass_graphite): Convert from a global struct to a subclass of
10746 gimple_opt_pass along with...
10747 (pass_data_graphite): ...new pass_data instance and...
10748 (make_pass_graphite): ...new function.
10749 (pass_graphite_transforms): Convert from a global struct to a subclass
10750 of gimple_opt_pass along with...
10751 (pass_data_graphite_transforms): ...new pass_data instance and...
10752 (make_pass_graphite_transforms): ...new function.
10753 (pass_check_data_deps): Convert from a global struct to a subclass of
10754 gimple_opt_pass along with...
10755 (pass_data_check_data_deps): ...new pass_data instance and...
10756 (make_pass_check_data_deps): ...new function.
10757 (pass_iv_canon): Convert from a global struct to a subclass of
10758 gimple_opt_pass along with...
10759 (pass_data_iv_canon): ...new pass_data instance and...
10760 (make_pass_iv_canon): ...new function.
10761 (pass_scev_cprop): Convert from a global struct to a subclass of
10762 gimple_opt_pass along with...
10763 (pass_data_scev_cprop): ...new pass_data instance and...
10764 (make_pass_scev_cprop): ...new function.
10765 (pass_record_bounds): Convert from a global struct to a subclass of
10766 gimple_opt_pass along with...
10767 (pass_data_record_bounds): ...new pass_data instance and...
10768 (make_pass_record_bounds): ...new function.
10769 (pass_complete_unroll): Convert from a global struct to a subclass of
10770 gimple_opt_pass along with...
10771 (pass_data_complete_unroll): ...new pass_data instance and...
10772 (make_pass_complete_unroll): ...new function.
10773 (pass_complete_unrolli): Convert from a global struct to a subclass of
10774 gimple_opt_pass along with...
10775 (pass_data_complete_unrolli): ...new pass_data instance and...
10776 (make_pass_complete_unrolli): ...new function.
10777 (pass_parallelize_loops): Convert from a global struct to a subclass
10778 of gimple_opt_pass along with...
10779 (pass_data_parallelize_loops): ...new pass_data instance and...
10780 (make_pass_parallelize_loops): ...new function.
10781 (pass_loop_prefetch): Convert from a global struct to a subclass of
10782 gimple_opt_pass along with...
10783 (pass_data_loop_prefetch): ...new pass_data instance and...
10784 (make_pass_loop_prefetch): ...new function.
10785 (pass_iv_optimize): Convert from a global struct to a subclass of
10786 gimple_opt_pass along with...
10787 (pass_data_iv_optimize): ...new pass_data instance and...
10788 (make_pass_iv_optimize): ...new function.
10789 (pass_tree_loop_done): Convert from a global struct to a subclass of
10790 gimple_opt_pass along with...
10791 (pass_data_tree_loop_done): ...new pass_data instance and...
10792 (make_pass_tree_loop_done): ...new function.
10793 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10794 struct to a subclass of gimple_opt_pass along with...
10795 (pass_data_cse_reciprocals): ...new pass_data instance and...
10796 (make_pass_cse_reciprocals): ...new function.
10797 (pass_cse_sincos): Convert from a global struct to a subclass of
10798 gimple_opt_pass along with...
10799 (pass_data_cse_sincos): ...new pass_data instance and...
10800 (make_pass_cse_sincos): ...new function.
10801 (pass_optimize_bswap): Convert from a global struct to a subclass of
10802 gimple_opt_pass along with...
10803 (pass_data_optimize_bswap): ...new pass_data instance and...
10804 (make_pass_optimize_bswap): ...new function.
10805 (pass_optimize_widening_mul): Convert from a global struct to a
10806 subclass of gimple_opt_pass along with...
10807 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10808 (make_pass_optimize_widening_mul): ...new function.
10809 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10810 subclass of gimple_opt_pass along with...
10811 (pass_data_phiopt): ...new pass_data instance and...
10812 (make_pass_phiopt): ...new function.
10813 (pass_cselim): Convert from a global struct to a subclass of
10814 gimple_opt_pass along with...
10815 (pass_data_cselim): ...new pass_data instance and...
10816 (make_pass_cselim): ...new function.
10817 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10818 subclass of gimple_opt_pass along with...
10819 (pass_data_phiprop): ...new pass_data instance and...
10820 (make_pass_phiprop): ...new function.
10821 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10822 subclass of gimple_opt_pass along with...
10823 (pass_data_pre): ...new pass_data instance and...
10824 (make_pass_pre): ...new function.
10825 (pass_fre): Convert from a global struct to a subclass of
10826 gimple_opt_pass along with...
10827 (pass_data_fre): ...new pass_data instance and...
10828 (make_pass_fre): ...new function.
10829 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10830 subclass of gimple_opt_pass along with...
10831 (pass_data_reassoc): ...new pass_data instance and...
10832 (make_pass_reassoc): ...new function.
10833 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10834 subclass of gimple_opt_pass along with...
10835 (pass_data_sink_code): ...new pass_data instance and...
10836 (make_pass_sink_code): ...new function.
10837 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10838 subclass of gimple_opt_pass along with...
10839 (pass_data_strlen): ...new pass_data instance and...
10840 (make_pass_strlen): ...new function.
10841 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10842 struct to a subclass of gimple_opt_pass along with...
10843 (pass_data_build_alias): ...new pass_data instance and...
10844 (make_pass_build_alias): ...new function.
10845 (pass_build_ealias): Convert from a global struct to a subclass of
10846 gimple_opt_pass along with...
10847 (pass_data_build_ealias): ...new pass_data instance and...
10848 (make_pass_build_ealias): ...new function.
10849 (pass_ipa_pta): Convert from a global struct to a subclass of
10850 simple_ipa_opt_pass along with...
10851 (pass_data_ipa_pta): ...new pass_data instance and...
10852 (make_pass_ipa_pta): ...new function.
10853 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10854 subclass of gimple_opt_pass along with...
10855 (pass_data_uncprop): ...new pass_data instance and...
10856 (make_pass_uncprop): ...new function.
10857 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10858 global struct to a subclass of gimple_opt_pass along with...
10859 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10860 (make_pass_late_warn_uninitialized): ...new function.
10861 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10862 to a subclass of gimple_opt_pass along with...
10863 (pass_data_init_datastructures): ...new pass_data instance and...
10864 (make_pass_init_datastructures): ...new function.
10865 (pass_early_warn_uninitialized): Convert from a global struct to a
10866 subclass of gimple_opt_pass along with...
10867 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10868 (make_pass_early_warn_uninitialized): ...new function.
10869 (pass_update_address_taken): Convert from a global struct to a
10870 subclass of gimple_opt_pass along with...
10871 (pass_data_update_address_taken): ...new pass_data instance and...
10872 (make_pass_update_address_taken): ...new function.
10873 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10874 struct to a subclass of gimple_opt_pass along with...
10875 (pass_data_release_ssa_names): ...new pass_data instance and...
10876 (make_pass_release_ssa_names): ...new function.
10877 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10878 subclass of gimple_opt_pass along with...
10879 (pass_data_stdarg): ...new pass_data instance and...
10880 (make_pass_stdarg): ...new function.
10881 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10882 global struct to a subclass of gimple_opt_pass along with...
10883 (pass_data_convert_switch): ...new pass_data instance and...
10884 (make_pass_convert_switch): ...new function.
10885 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10886 to a subclass of gimple_opt_pass along with...
10887 (pass_data_tail_recursion): ...new pass_data instance and...
10888 (make_pass_tail_recursion): ...new function.
10889 (pass_tail_calls): Convert from a global struct to a subclass of
10890 gimple_opt_pass along with...
10891 (pass_data_tail_calls): ...new pass_data instance and...
10892 (make_pass_tail_calls): ...new function.
10893 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10894 struct to a subclass of gimple_opt_pass along with...
10895 (pass_data_lower_vector): ...new pass_data instance and...
10896 (make_pass_lower_vector): ...new function.
10897 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10898 gimple_opt_pass along with...
10899 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10900 (make_pass_lower_vector_ssa): ...new function.
10901 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10902 to a subclass of gimple_opt_pass along with...
10903 (pass_data_slp_vectorize): ...new pass_data instance and...
10904 (make_pass_slp_vectorize): ...new function.
10905 (pass_ipa_increase_alignment): Convert from a global struct to a
10906 subclass of simple_ipa_opt_pass along with...
10907 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10908 (make_pass_ipa_increase_alignment): ...new function.
10909 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10910 gimple_opt_pass along with...
10911 (pass_data_vrp): ...new pass_data instance and...
10912 (make_pass_vrp): ...new function.
10913 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10914 subclass of simple_ipa_opt_pass along with...
10915 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10916 (make_pass_ipa_free_lang_data): ...new function.
10917 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10918 gimple_opt_pass along with...
10919 (pass_data_tsan): ...new pass_data instance and...
10920 (make_pass_tsan): ...new function.
10921 (pass_tsan_O0): Convert from a global struct to a subclass of
10922 gimple_opt_pass along with...
10923 (pass_data_tsan_O0): ...new pass_data instance and...
10924 (make_pass_tsan_O0): ...new function.
10925 * var-tracking.c (pass_variable_tracking): Convert from a global
10926 struct to a subclass of rtl_opt_pass along with...
10927 (pass_data_variable_tracking): ...new pass_data instance and...
10928 (make_pass_variable_tracking): ...new function.
10929 * web.c (pass_web): Convert from a global struct to a subclass of
10930 rtl_opt_pass along with...
10931 (pass_data_web): ...new pass_data instance and...
10932 (make_pass_web): ...new function.
10933 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10934 declaration with that of...
10935 (make_pass_mode_switch_use): ...new function.
10936 (pass_resolve_sw_modes): Replace declaration with that of...
10937 (make_pass_resolve_sw_modes): ...new function.
10938 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10939 from a global struct to a subclass of rtl_opt_pass along with...
10940 (pass_data_mode_switch_use): ...new pass_data instance and...
10941 (make_pass_mode_switch_use): ...new function.
10942 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10943 from a global struct to a subclass of rtl_opt_pass along with...
10944 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10945 (make_pass_resolve_sw_modes): ...new function.
10946 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10947 struct to a subclass of rtl_opt_pass along with...
10948 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10949 (make_pass_insert_vzeroupper): ...new function.
10950 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10951 global struct to a subclass of rtl_opt_pass along with...
10952 (pass_data_work_around_errata): ...new pass_data instance and...
10953 (make_pass_work_around_errata): ...new function.
10954 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10955 struct to a subclass of rtl_opt_pass along with...
10956 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10957 (make_pass_mips_machine_reorg2): ...new function.
10958
10959 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10960
10961 * passes.c (pass_manager::operator new): New.
10962
10963 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10964
10965 Handwritten part of conversion of passes to C++ classes.
10966
10967 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10968 (toplev.o): Add dep on PASS_MANAGER_H.
10969 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10970 of early local pases to reflect this moving from a global to a
10971 member of gcc::pass_manager.
10972 (cgraph_add_new_function): Likewise.
10973 * lto-cgraph.c (lto_output_node): Update for conversion of
10974 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10975 * passes.c (opt_pass::clone): New.
10976 (opt_pass::gate): New.
10977 (opt_pass::execute): New.
10978 (opt_pass::opt_pass): New.
10979 (pass_manager::execute_early_local_passes): New.
10980 (pass_manager::execute_pass_mode_switching): new.
10981 (finish_optimization_passes): Convert to...
10982 (pass_manager::finish_optimization_passes): ...this.
10983 (finish_optimization_passes): Update for conversion of passes to
10984 C++ classes.
10985 (register_dump_files_1): Use has_gate since we cannot portably
10986 check a vtable entry against NULL.
10987 (dump_one_pass): Likewise.
10988 (ipa_write_summaries_2): Likewise.
10989 (ipa_write_optimization_summaries_1): Likewise.
10990 (ipa_read_summaries_1): Likewise.
10991 (ipa_read_optimization_summaries_1): Likewise.
10992 (execute_ipa_stmt_fixups): Likewise.
10993 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10994 pass-creation functions rather than wiring up globals, and
10995 storing the results in fields of pass_manager generated using
10996 pass-instances.def.
10997 (pass_manager::dump_profile_report): Update for conversion of
10998 passes to C++ classes.
10999 (pass_manager::execute_ipa_summary_passes): Likewise.
11000 (execute_one_ipa_transform_pass): Likewise.
11001 (execute_one_pass): Use has_gate and has_execute since we cannot
11002 portably check a vtable entry against NULL.
11003 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11004 (pass_manager): Use pass-instances.def to add fields for the
11005 various pass instances.
11006 * toplev.c (finalize): Update for move of
11007 finish_optimization_passes to a method of gcc::pass_manager.
11008 * toplev.h (finish_optimization_passes): Move to method of class
11009 pass_manager.
11010 * tree-pass.h (struct pass_data): New.
11011 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11012 (opt_pass::gate): Convert to virtual function.
11013 (opt_pass::~opt_pass): New.
11014 (opt_pass::clone): New.
11015 (opt_pass::execute): Convert to virtual function.
11016 (opt_pass::opt_pass): New.
11017 (opt_pass::ctxt_): new.
11018 (gimple_opt_pass): Convert to subclass of opt_pass.
11019 (gimple_opt_pass::gimple_opt_pass): New.
11020 (rtl_opt_pass): Convert to subclass of opt_pass.
11021 (rtl_opt_pass::rtl_opt_pass): New.
11022 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11023 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11024 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11025 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11026 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11027 invocation of pass_mode_switching to reflect this moving from a
11028 global to a member of gcc::pass_manager.
11029 (ix86_option_override): Rework how pass_insert_vzeroupper is
11030 added to the pass_manager to reflect autogenerated changes.
11031 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11032
11033 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11034
11035 PR rtl-optimization/57708
11036 * recog.c (peep2_find_free_register): Validate all regs in a
11037 multi-reg mode.
11038
11039 2013-08-05 Jan Hubicka <jh@suse.cz>
11040
11041 PR lto/57602
11042 * cgraph.c (verify_cgraph_node): Accept local flags from other
11043 partitions.
11044 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11045 (function_and_variable_visibility): Likewise.
11046 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11047
11048 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11049
11050 * graph.c (init_graph_slim_pretty_print): Remove.
11051 (print_graph_cfg): Do not call it. Use local pretty printer.
11052 (start_graph_dump): Likewise.
11053
11054 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11055
11056 * gimple-pretty-print.c (buffer): Remove.
11057 (initialized): Likewise.
11058 (maybe_init_pretty_print): Likewise.
11059 (print_gimple_stmt): Do not call it. Use non-static local
11060 pretty_printer variable.
11061 (print_gimple_expr): Likewise.
11062 (print_gimple_seq): Likewise.
11063 (gimple_dump_bb): Likewise.
11064
11065 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11066
11067 * asan.c (asan_pp): Remove.
11068 (asan_pp_initialized): Likewise.
11069 (asan_pp_initialize): Likewise.
11070 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11071 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11072 (asan_add_global): Likewise.
11073
11074 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11075
11076 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11077 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11078 * pretty-print.c (pp_formatted_text_data): Likewise.
11079 (pp_indent): Rename from pp_base_indent.
11080 (pp_format): Rename from pp_base_format.
11081 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11082 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11083 (pp_flush): Rename from pp_base_flush.
11084 (pp_set_line_maximum_length): Rename from
11085 pp_base_set_line_maximum_length.
11086 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11087 (pp_set_prefix): Rename from pp_base_set_prefix.
11088 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11089 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11090 (pp_append_text): Rename from pp_base_append_text.
11091 (pp_formatted_text): Rename from pp_base_formatted_text.
11092 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11093 (pp_remaining_character_count_for_line): Rename from
11094 pp_base_remaining_character_count_for_line.
11095 (pp_newline): Rename from pp_base_newline.
11096 (pp_character): Rename from pp_base_character.
11097 (pp_string): Rename from pp_base_string.
11098 (pp_maybe_space): Rename from pp_base_maybe_space.
11099 * asan.c (asan_pp_string): Adjust.
11100 (asan_emit_stack_protection): Likewise.
11101 (asan_add_global): Likewise.
11102 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11103 * tree-mudflap.c (mf_varname_tree): Likewise.
11104 * tree-pretty-print.c (pp_tree_identifier): Rename from
11105 pp_base_tree_identifier.
11106 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11107 Declare as function.
11108
11109 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11110
11111 * pretty-print.h (pp_bar_bar): New.
11112 (pp_ampersand_ampersand): Likewise.
11113 (pp_less_equal): Likewise.
11114 (pp_greater_equal): Likewise.
11115 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11116 printer functions instead of pp_string or operators and punctuators.
11117 (dump_gimple_call): Likewise.
11118 (dump_gimple_omp_for): Likewise.
11119 (dump_gimple_transaction): Likewise.
11120 (dump_gimple_phi): Likewise.
11121 (pp_gimple_stmt_1): Likewise.
11122 * sched-vis.c (print_insn): Likewise.
11123 * tree-mudflap.c (mf_varname_tree): Likewise.
11124 * tree-pretty-print.c (dump_block_node): Likewise.
11125 (dump_generic_node): Likewise.
11126
11127 2013-08-02 Jan Hubicka <jh@suse.cz>
11128
11129 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11130 boundaries.
11131 * lto-streamer-out.c (tree_is_indexable): Results decls and
11132 parm decls are not indexable.
11133 (DFS_write_tree_body): Do not follow args and results.
11134 (hash_tree): Likewise.
11135 (output_functions): Rearrange so struct function is needed
11136 only when real body is output; be able to also ouptut abstract
11137 functions; output DECL_ARGUMENTS and DECL_RESULT.
11138 (lto_output): When not in WPA, ale store abstract functions.
11139 (write_symbol): Do not care about RESULT_DECL.
11140 (output_symbol_p): Handle correctly sbtract decls.
11141 * lto-streamer-in.c (input_function): Rearrange so struct
11142 function can be NULL at entry; allow streaming of
11143 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11144 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11145 sanity check during LTO.
11146 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11147 RESULT_DECl and DECL_ARGUMENTS.
11148 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11149 Likewise.
11150
11151 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11152
11153 * pretty-print.h (pp_underscore): New.
11154 (pp_comma): Tidy.
11155 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11156 printer functions instead of pp_character.
11157 (dump_binary_rhs): Likewise.
11158 (dump_ternary_rhs): Likewise.
11159 (dump_gimple_call_args): Likewise.
11160 (pp_points_to_solution): Likewise.
11161 (dump_gimple_call): Likewise.
11162 (dump_gimple_switch): Likewise.
11163 (dump_gimple_cond): Likewise.
11164 (dump_gimple_bind): Likewise.
11165 (dump_gimple_try): Likewise.
11166 (dump_gimple_omp_for): Likewise.
11167 (dump_gimple_omp_continue): Likewise.
11168 (dump_gimple_omp_single): Likewise.
11169 (dump_gimple_omp_sections): Likewise.
11170 (dump_gimple_omp_block): Likewise.
11171 (dump_gimple_omp_critical): Likewise.
11172 (dump_gimple_transaction): Likewise.
11173 (dump_gimple_asm): Likewise.
11174 (dump_gimple_phi): Likewise.
11175 (dump_gimple_omp_parallel): Likewise.
11176 (dump_gimple_omp_task): Likewise.
11177 (dump_gimple_omp_atomic_load): Likewise.
11178 (dump_gimple_omp_atomic_store): Likewise.
11179 (dump_gimple_mem_ops): Likewise.
11180 (pp_gimple_stmt_1): Likewise.
11181 (pp_cfg_jump): Likewise.
11182 (dump_implicit_edges): Likewise.
11183 (gimple_dump_bb_for_graph): Likewise.
11184 * graph.c (draw_cfg_node): Likewise.
11185 * langhooks.c (lhd_print_error_function): Likewise.
11186 * sched-vis.c (print_exp): Likewise.
11187 (print_value): Likewise.
11188 (print_pattern): Likewise.
11189 (print_insn): Likewise.
11190 (rtl_dump_bb_for_graph): Likewise.
11191 * tree-pretty-print.c (dump_function_declaration): Likewise.
11192 (dump_array_domain): Likewise.
11193 (dump_omp_clause): Likewise.
11194 (dump_location): Likewise.
11195 (dump_generic_node): Likewise.
11196 (print_struct_decl): Likewise.
11197 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11198
11199 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11200
11201 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11202 candidate table when replacing a candidate statement.
11203 (replace_rhs_if_not_dup): Likewise.
11204 (replace_one_candidate): Likewise.
11205
11206 2013-08-02 Jan Hubicka <jh@suse.cz>
11207 Martin Liska <marxin.liska@gmail.com>
11208
11209 * cgraphunit.c (add_new_function): Fix logic when adding from
11210 late IPA pass.
11211 (assemble_thunk): Rename to ...
11212 (expand_thunk); .. this one; export; get it working with
11213 general functions; make produced gimple valid.
11214 * cgraph.h (expand_thunk): Declare.
11215
11216 2013-08-02 Jan Hubicka <jh@suse.cz>
11217
11218 * ipa-cp.c (gather_context_independent_values): Use
11219 ipa_get_param_move_cost.
11220 (get_replacement_map): Remove PARAM; move parameter folding
11221 into tree-inline.c
11222 (create_specialized_node): Update.
11223 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11224 assert that we have gimple body; update move_cost.
11225 (count_formal_params): Assert that we have gimple body.
11226 (ipa_dump_param): New function.
11227 (ipa_alloc_node_params): Break out from ...
11228 (ipa_initialize_node_params): ... here.
11229 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11230 (ipa_write_node_info): Stream move costs.
11231 (ipa_read_node_info): Read move costs.
11232 (ipa_update_after_lto_read): Do not recompute node params.
11233 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11234 (ipa_get_param): Check we are not in WPA.
11235 (ipa_get_param_move_cost): New.
11236 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11237 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11238 parm numbers to be present.
11239
11240 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11241
11242 PR rtl-optimization/58048
11243 * lra-constraints.c (process_alt_operands): Don't check asm
11244 operand on register.
11245
11246 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11247
11248 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11249 the implied StoreLoad barrier for atomic operations if before.
11250
11251 2013-08-02 Jan Hubicka <jh@suse.cz>
11252 Martin Liska <marxin.liska@gmail.com>
11253
11254 * cgraph.c (cgraph_function_body_availability): Do not check
11255 cgraph flags.
11256 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11257 symtab_node_availability): Declare.
11258 * ipa.c (can_replace_by_local_alias): New.
11259 (function_and_variable_visibility): Use it.
11260 * symtab.c (symtab_for_node_and_aliases,
11261 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11262
11263 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11264
11265 PR rtl-optimization/57963
11266 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11267 (lra_constraints): Use them.
11268
11269 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11270
11271 * config/arm/types.md (define_attr "type"): Add "load_acq"
11272 and "store_rel".
11273 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11274 changes.
11275 (cortex_a53_store1): Likewise.
11276
11277 2013-08-01 Jan Hubicka <jh@suse.cz>
11278
11279 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11280 partitions are not needed.
11281
11282 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11283
11284 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11285 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11286 MAYBE_NON_Q_CLASS_P where appropriate.
11287
11288 2013-08-01 Jan Hubicka <jh@suse.cz>
11289
11290 * cgraph.h (release_function_body): Declare.
11291 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11292 of unused delcarations.
11293
11294 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11295
11296 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11297 RTL form when subtracting a constant.
11298
11299 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11300
11301 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11302 Generate canonical plus rtx with negated immediate instead of minus
11303 where appropriate.
11304 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11305
11306 2013-08-01 Jan Hubicka <jh@suse.cz>
11307
11308 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11309 (cgraph_release_function_body): Likewise.
11310 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11311 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11312 to used_as_abstract_origin.
11313 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11314 symbols used as abstract origins.
11315 * cgraphunit.c (analyze_functions): Update.
11316 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11317 used_as_abstract_origin.
11318 * tree-inline.c (tree_function_versioning): Update
11319 used_as_abstract_origin; be ready for DECL_RESULT and
11320 DECL_ARGUMENTS to be NULL.
11321
11322 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11323 for abstract functions.
11324 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11325 real symbols.
11326
11327 2013-08-01 Jan Hubicka <jh@suse.cz>
11328
11329 * profile.c (compute_value_histograms): Fix thinko.
11330
11331 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11332
11333 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11334 aarch-common-protos.h to extra_headers.
11335 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11336 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11337 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11338
11339 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11340
11341 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11342 Include "../arm/types.md". Define "type" attribute for all patterns.
11343 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11344 attribute changes.
11345
11346 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11347
11348 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11349 to support power8 load fusion.
11350 (fusion_gpr_mem_load): Likewise.
11351
11352 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11353
11354 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11355 declarations for power8 load fusion.
11356 (emit_fusion_gpr_load): Likewise.
11357
11358 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11359 tuning for power8, turn on fusion mode by default. Turn on sign
11360 extending fusion mode if normal fusion mode is on, and we are at
11361 -O2 or -O3.
11362 (fusion_gpr_load_p): New function, return true if we can fuse an
11363 addis instruction with a dependent load to a GPR.
11364 (emit_fusion_gpr_load): Emit the instructions for power8 load
11365 fusion to GPRs.
11366
11367 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11368 (VSX load fusion peepholes): New peepholes to fuse together an
11369 addi instruction with a VSX load instruction.
11370
11371 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11372 peepholes to fuse an addis instruction with a load to a GPR base
11373 register. If we are supporting sign extending fusions, convert
11374 sign extending loads to zero extending loads and add an explicit
11375 sign extension.
11376
11377 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11378
11379 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11380 aarch-common-protos.h to extra_headers.
11381 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11382 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11383 (arm_early_store_addr_dep): Likewise.
11384 (arm_no_early_alu_shift_dep): Likewise.
11385 (arm_no_early_alu_shift_value_dep): Likewise.
11386 (arm_no_early_mul_dep): Likewise.
11387 (arm_no_early_store_addr_dep): Likewise.
11388 (arm_mac_accumulator_is_mul_result): Likewise.
11389 (arm_mac_accumulator_is_result): Likewise.
11390 * config/arm/aarch-common.c: ... here. New file.
11391 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11392 here to ...
11393 (arm_early_store_addr_dep): Likewise.
11394 (arm_no_early_alu_shift_dep): Likewise.
11395 (arm_no_early_alu_shift_value_dep): Likewise.
11396 (arm_no_early_mul_dep): Likewise.
11397 (arm_no_early_store_addr_dep): Likewise.
11398 (arm_mac_accumulator_is_mul_result): Likewise.
11399 (arm_mac_accumulator_is_result): Likewise.
11400 * config/arm/aarch-common-protos.h: ... here. New file.
11401 * config/arm/t-arm (aarch-common.o): Define.
11402
11403 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11404
11405 * config/arm/arm.md: Include new file "types.md".
11406 (define_attr "type"): Move from here to ...
11407 (define_attr "mul32"): Likewise.
11408 (define_attr "mul64"): Likewise.
11409 * config/arm/types.md: ... here. New file.
11410
11411 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11412
11413 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11414 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11415
11416 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11417
11418 * gen-pass-instances.awk: Fix offset of substr().
11419
11420 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11421
11422 * Makefile.in (pass-instances.def): New.
11423 (passes.o): Replace dependency on passes.def with one on
11424 pass-instances.def
11425
11426 * gen-pass-instances.awk: New.
11427
11428 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11429 rather than passes.def, updating local definition of NEXT_PASS
11430 macro to add an extra NUM parameter (currently unused).
11431
11432 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11433
11434 * Makefile.in (PASS_MANAGER_H): New.
11435 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11436 (passes.o): Likewise.
11437 (statistics.o): Likewise.
11438 (cgraphunit.o): Likewise.
11439 (context.o): Depend on PASS_MANAGER_H.
11440
11441 * pass_manager.h: New.
11442
11443 * cgraphunit.c (cgraph_add_new_function): Update for moves
11444 of globals to fields of pass_manager.
11445 (analyze_function): Likewise.
11446 (expand_function): Likewise.
11447 (ipa_passes): Likewise.
11448 (compile): Likewise.
11449
11450 * context.c (context::context): New.
11451 * context.h (context::context): New.
11452 (context::get_passes): New.
11453 (context::passes_): New.
11454
11455 * lto-cgraph.c (input_node): Update for moves of globals to
11456 fields of pass_manager.
11457
11458 * passes.c (all_passes): Remove, in favor of a field of the
11459 same name within the new class pass_manager.
11460 (all_small_ipa_passes): Likewise.
11461 (all_lowering_passes): Likewise.
11462 (all_regular_ipa_passes): Likewise.
11463 (all_late_ipa_passes): Likewise.
11464 (all_lto_gen_passes): Likewise.
11465 (passes_by_id): Likewise.
11466 (passes_by_id_size): Likewise.
11467 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11468 the new class pass_manager.
11469 (set_pass_for_id): Convert to...
11470 (pass_manager::set_pass_for_id): ...method.
11471 (get_pass_for_id): Convert to...
11472 (pass_manager::get_pass_for_id): ...method.
11473 (register_one_dump_file): Move body of implementation into...
11474 (pass_manager::register_one_dump_file): ...here.
11475 (register_dump_files_1): Convert to...
11476 (pass_manager::register_dump_files_1): ...method.
11477 (register_dump_files): Convert to...
11478 (pass_manager::register_dump_files): ...method.
11479 (create_pass_tab): Update for moves of globals to fields of
11480 pass_manager.
11481 (dump_passes): Move body of implementation into...
11482 (pass_manager::dump_passes): ...here.
11483 (register_pass): Move body of implementation into...
11484 (pass_manager::register_pass): ...here.
11485 (init_optimization_passes): Convert into...
11486 (pass_manager::pass_manager): ...constructor for new
11487 pass_manager class, and initialize the pass_lists array.
11488 (check_profile_consistency): Update for moves of globals to
11489 fields of pass_manager.
11490 (dump_profile_report): Move body of implementation into...
11491 (pass_manager::dump_profile_report): ...here.
11492 (ipa_write_summaries_1): Update for moves of pass lists from
11493 being globals to fields of pass_manager.
11494 (ipa_write_optimization_summaries): Likewise.
11495 (ipa_read_summaries): Likewise.
11496 (ipa_read_optimization_summaries): Likewise.
11497 (execute_all_ipa_stmt_fixups): Likewise.
11498
11499 * statistics.c (statistics_fini): Update for moves of globals to
11500 fields of pass_manager.
11501
11502 * toplev.c (general_init): Replace call to
11503 init_optimization_passes with construction of the pass_manager
11504 instance.
11505
11506 * tree-pass.h (all_passes): Remove, in favor of a field of the
11507 same name within the new class pass_manager.
11508 (all_small_ipa_passes): Likewise.
11509 (all_lowering_passes): Likewise.
11510 (all_regular_ipa_passes): Likewise.
11511 (all_lto_gen_passes): Likewise.
11512 (all_late_ipa_passes): Likewise.
11513 (passes_by_id): Likewise.
11514 (passes_by_id_size): Likewise.
11515 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11516 the new class pass_manager.
11517 (get_pass_for_id): Remove.
11518
11519 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11520
11521 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11522 configs.
11523
11524 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11525
11526 * arm.md (mulhi3): New expand pattern.
11527
11528 2013-07-30 Jan Hubicka <jh@suse.cz>
11529 Martin Liska <marxin.liska@gmail.com>
11530
11531 * profile.c (compute_value_histograms): Do not ICE when
11532 there is mismatch only on some counters.
11533
11534 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11535
11536 PR rtl-optimization/57637
11537 * function.c (move_insn_for_shrink_wrap): Also check the
11538 GEN set of the LIVE problem for the liveness analysis
11539 if it exists, otherwise give up.
11540
11541 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11542
11543 PR tree-optimization/57993
11544 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11545 replaced statement in the candidate table.
11546 (phi_add_costs): Return infinite cost when the hidden basis does
11547 not dominate all phis on which the candidate is dependent.
11548 (replace_one_candidate): Record replaced statement in the
11549 candidate table.
11550
11551 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11552
11553 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11554 (ashlv2si3): New expander.
11555 (*ashlv2si3_i): New define_insn_and_split.
11556 * predicates.md (float_operation): Allow patterns with three
11557 basic sub-patterns.
11558
11559 PR rtl-optimization/58021
11560 * mode-switching.c (create_pre_exit): Always split off preceding
11561 insns if we are not at the basic block head.
11562
11563 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11564
11565 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11566 (UCLIBC_DYNAMIC_LINKER): New macro.
11567 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11568 `-mnan=2008'.
11569 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11570 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11571 `-mnan=2008'.
11572 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11573 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11574 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11575 for SF and DF modes. Use ieee_quad_format for TF mode.
11576 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11577 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11578 (mips_option_override): Handle `-mnan=legacy'.
11579 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11580 `-mabs=2008' and `-mnan=2008'.
11581 (OPTION_DEFAULT_SPECS): Add "nan" default.
11582 (ASM_SPEC): Handle `-mnan='.
11583 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11584 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11585 comment accordingly.
11586 (neg<mode>2): Likewise.
11587 * config/mips/mips.opt (mabs, mnan): New options.
11588 * doc/install.texi (Configuration): Document `--with-nan=' option.
11589 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11590 `-mnan=' options.
11591 (MIPS Options): Document them.
11592 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11593 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11594 * configure: Regenerate.
11595 * config.in: Regenerate.
11596
11597 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11598
11599 * config/i386/i386.md (float post-reload splitters): Do not check
11600 for subregs of SSE registers.
11601
11602 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
11603 H.J. Lu <hongjiu.lu@intel.com>
11604
11605 PR target/57954
11606 PR target/57988
11607 * config/i386/i386.md (post-reload splitter
11608 to avoid partial SSE reg dependency stalls): New pattern.
11609
11610 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11611
11612 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11613
11614 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11615
11616 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11617 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11618
11619 2013-07-26 Cary Coutant <ccoutant@google.com>
11620
11621 * dwarf2out.c (die_checksum_ordered): Don't include template
11622 instantiations in signature.
11623 (is_template_parameter): New function.
11624 (is_template_instantiation): New function.
11625 (generate_skeleton_bottom_up): Don't include template instantiations
11626 in type unit DIE.
11627 (generate_skeleton): Likewise.
11628 (break_out_comdat_types): Move recursive call to break out nested
11629 types earlier.
11630 (prune_unused_types_mark_generic_parms_dies): Call
11631 is_template_parameter.
11632
11633 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11634
11635 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11636 uses vector registers.
11637 * config/aarch64/iterators.md: Add attributes rtn and vas.
11638
11639 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11640 Richard Earnshaw <richard.earnshaw@arm.com>
11641
11642 * combine.c (simplify_comparison): Re-canonicalize operands
11643 where appropriate.
11644 * config/arm/arm.md (movcond_addsi): New splitter.
11645
11646 2013-07-25 Sterling Augustine <saugustine@google.com>
11647
11648 * dwarf2out.c (size_of_pubnames): Move code to...
11649 (include_pubname_in_output): ...here. New.
11650 (want_pubnames): Rearrange.
11651 (output_pubnames): Call include_pubname_in_output. Move assertion.
11652
11653 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11654
11655 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11656
11657 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11658
11659 PR target/38836
11660 * doc/extend.texi: Remove obsolete builtins. Fix
11661 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11662
11663 2013-07-25 Jan Hubicka <jh@suse.cz>
11664
11665 * cgraph.c (release_function_body): Break out from ...
11666 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11667 and DECL_ARGUMENTS.
11668 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11669 old_tree in the map.
11670 (create_specialized_node): Update.
11671 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11672 into index.
11673 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11674 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11675 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11676 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11677 DECL_RESULT.
11678
11679 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11680
11681 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11682 addsi3_carryin_alt2_<optab>): Correct output template.
11683
11684 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11685
11686 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11687 Adjust for arm_restrict_it.
11688 Remove trailing whitespace.
11689
11690 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11691
11692 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11693 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11694
11695 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11696
11697 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11698
11699 PR rtl-optimization/57960
11700 * lra-constraints.c (process_alt_operands): Use the right mode
11701 when checking strict_low.
11702
11703 2013-07-25 Jan Hubicka <jh@suse.cz>
11704
11705 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11706 * cgraph.c (cgraph_remove_node): Do not release function body
11707 when in cgraph streaming.
11708 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11709 in other partitions are not considered reachable; fix handling of
11710 clones.
11711
11712 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11713
11714 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11715
11716 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11717
11718 PR target/19599
11719 PR target/57731
11720 PR target/57837
11721 * config/arm/arm.md ("*sibcall_insn): Replace use of
11722 Ss with US. Adjust output for v5 and v4t.
11723 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11724
11725 * config/arm/constraints.md ("Ss"): Rename to US.
11726
11727 2013-07-25 Terry Guo <terry.guo@arm.com>
11728
11729 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11730 shift_add/shift_sub0/shift_sub1 RTXs.
11731
11732 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11733 Anton Blanchard <anton@au1.ibm.com>
11734
11735 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11736 (altivec_vpks<VI_char>ss): Likewise.
11737 (altivec_vpks<VI_char>us): Likewise.
11738 (altivec_vpku<VI_char>us): Likewise.
11739 (altivec_vpku<VI_char>um): Likewise.
11740
11741 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11742
11743 Introduce context class.
11744
11745 * Makefile.in (CONTEXT_H): New.
11746 (OBJS): Add context.o.
11747 (toplev.o): Add CONTEXT_H to dependencies.
11748 (context.o): New.
11749
11750 * toplev.c (general_init): Create the singleton gcc::context instance.
11751
11752 * context.c: New.
11753
11754 * context.h: New.
11755
11756 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11757
11758 PR rtl-optimization/57968
11759 * mode-switching.c (create_pre_exit): Allow instructions that
11760 don't set a return register to need a non-exit mode.
11761
11762 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11763 Anton Blanchard <anton@au1.ibm.com>
11764
11765 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11766 operands to vperm for little endian.
11767 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11768 of lvsl to create the control mask for a vperm for little endian.
11769
11770 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11771 Anton Blanchard <anton@au1.ibm.com>
11772
11773 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11774 two operands for little-endian.
11775
11776 2013-07-23 Steve Ellcey <sellcey@mips.com>
11777
11778 * config/mips/mips.c (mips_case_values_threshold): New.
11779 (TARGET_CASE_VALUES_THRESHOLD): Define.
11780
11781 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11782 Anton Blanchard <anton@au1.ibm.com>
11783
11784 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11785 selection of field for vector splat in little endian mode.
11786
11787 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11788
11789 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11790 expanders to rs6000.md.
11791 (ior<mode>3): Likewise.
11792 (and<mode>3): Likewise.
11793 (one_cmpl<mode>2): Likewise.
11794 (nor<mode>3): Likewise.
11795 (andc<mode>3): Likewise.
11796 (eqv<mode>3): Likewise.
11797 (nand<mode>3): Likewise.
11798 (orc<mode>3): Likewise.
11799
11800 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11801 declaration.
11802
11803 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11804 to split multi-word logical operations.
11805 (rs6000_split_logical_di): Likewise.
11806 (rs6000_split_logical): Likewise.
11807
11808 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11809 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11810 and allow TImode operations in 32-bit.
11811 (vsx_and<mode>3_64bit): Likewise.
11812 (vsx_ior<mode>3_32bit): Likewise.
11813 (vsx_ior<mode>3_64bit): Likewise.
11814 (vsx_xor<mode>3_32bit): Likewise.
11815 (vsx_xor<mode>3_64bit): Likewise.
11816 (vsx_one_cmpl<mode>2_32bit): Likewise.
11817 (vsx_one_cmpl<mode>2_64bit): Likewise.
11818 (vsx_nor<mode>3_32bit): Likewise.
11819 (vsx_nor<mode>3_64bit): Likewise.
11820 (vsx_andc<mode>3_32bit): Likewise.
11821 (vsx_andc<mode>3_64bit): Likewise.
11822 (vsx_eqv<mode>3_32bit): Likewise.
11823 (vsx_eqv<mode>3_64bit): Likewise.
11824 (vsx_nand<mode>3_32bit): Likewise.
11825 (vsx_nand<mode>3_64bit): Likewise.
11826 (vsx_orc<mode>3_32bit): Likewise.
11827 (vsx_orc<mode>3_64bit): Likewise.
11828
11829 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11830 logical types in GPRs.
11831
11832 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11833 logical insns to rs6000.md, and allow TImode operations in
11834 32-bit.
11835 (altivec_ior<mode>3): Likewise.
11836 (altivec_xor<mode>3): Likewise.
11837 (altivec_one_cmpl<mode>2): Likewise.
11838 (altivec_nor<mode>3): Likewise.
11839 (altivec_andc<mode>3): Likewise.
11840
11841 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11842 attributes for moving the 128-bit logical operations into
11843 rs6000.md.
11844 (BOOL_REGS_OUTPUT): Likewise.
11845 (BOOL_REGS_OP1): Likewise.
11846 (BOOL_REGS_OP2): Likewise.
11847 (BOOL_REGS_UNARY): Likewise.
11848 (BOOL_REGS_AND_CR0): Likewise.
11849 (one_cmpl<mode>2): Add support for DI logical operations on
11850 32-bit, splitting the operations to 32-bit.
11851 (anddi3): Likewise.
11852 (iordi3): Likewise.
11853 (xordi3): Likewise.
11854 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11855 changes to combine the 32/64-bit code, allow logical operations on
11856 TI mode in 32-bit, and to use similar match_operator patterns like
11857 scalar mode uses. Combine the Altivec and VSX code for logical
11858 operations, and move it here.
11859 (ior<mode>3, 128-bit types): Likewise.
11860 (xor<mode>3, 128-bit types): Likewise.
11861 (one_cmpl<mode>3, 128-bit types): Likewise.
11862 (nor<mode>3, 128-bit types): Likewise.
11863 (andc<mode>3, 128-bit types): Likewise.
11864 (eqv<mode>3, 128-bit types): Likewise.
11865 (nand<mode>3, 128-bit types): Likewise.
11866 (orc<mode>3, 128-bit types): Likewise.
11867 (and<mode>3_internal): Likewise.
11868 (bool<mode>3_internal): Likewise.
11869 (boolc<mode>3_internal1): Likewise.
11870 (boolc<mode>3_internal2): Likewise.
11871 (boolcc<mode>3_internal1): Likewise.
11872 (boolcc<mode>3_internal2): Likewise.
11873 (eqv<mode>3_internal1): Likewise.
11874 (eqv<mode>3_internal2): Likewise.
11875 (one_cmpl1<mode>3_internal): Likewise.
11876
11877 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11878
11879 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11880 Rename flag_stack_usage to flag_stack_usage_info.
11881
11882 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11883
11884 * config/microblaze/sync.md: New file.
11885 * config/microblaze/microblaze.md: Include sync.md
11886 * config/microblaze/microblaze.c: Add print_operand 'y'.
11887 * config/microblaze/constraints.md: Add memory_contraint
11888 'Q' which is a single register.
11889
11890 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11891
11892 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11893
11894 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11895
11896 * reload.c (find_reloads): Exit loop once we find this operand
11897 cannot be reloaded somehow for this alternative.
11898
11899 * reload.c (find_reloads): Exit loop once we find a hard register.
11900
11901 * rtlanal.c (computed_jump_p): Exit loop once we find label
11902 reference is used.
11903
11904 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11905
11906 * cfgloopmanip.c (remove_path): Exit loop after setting
11907 irred_invalidated.
11908
11909 * gensupport.c (subst_dup): Avoid loop if code is not
11910 MATCH_DUP nor MATCH_OP_DUP.
11911
11912 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11913
11914 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11915
11916 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11917
11918 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11919 true for SP_REGNUM if mode == ptr_mode.
11920 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11921 with value R0_REGNUM + 31.
11922
11923 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11924
11925 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11926 pad pointer-typed argument downward.
11927
11928 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11929
11930 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11931 and __ILP32__ when the ILP32 model is in use.
11932
11933 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11934
11935 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11936 (aarch64_load_symref_appropriately): In the case of
11937 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11938 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11939 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11940 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11941 if 'mode' doesn't equal to 'ptr_mode'.
11942 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11943 'vcall_offset'; change to call aarch64_emit_move differently depending
11944 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11945 to calculate the upper bound of 'vcall_offset'.
11946 (aarch64_cannot_force_const_mem): Change to also return true if
11947 mode != ptr_mode.
11948 (aarch64_legitimize_reload_address): In the case of large
11949 displacements, add new local variable 'xmode' and an assertion
11950 based on it; change to use 'xmode' to generate the new rtx and
11951 reload.
11952 (aarch64_asm_trampoline_template): Change to generate the template
11953 differently depending on TARGET_ILP32 or not; change to use
11954 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11955 (aarch64_trampoline_size): Removed.
11956 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11957 and replace immediate literals with it. Change to use 'ptr_mode'
11958 instead of 'DImode' and call convert_memory_address if the mode
11959 of 'fnaddr' doesn't equal to 'ptr_mode'.
11960 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11961 to output symbol.
11962 (aarch64_elf_asm_destructor): Likewise.
11963 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11964 on TARGET_ILP32 instead of aarch64_trampoline_size.
11965 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11966 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11967 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11968 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11969 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11970 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11971 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11972 depending on the value of 'mode'.
11973 (add_losym_<mode>): New.
11974 (ldr_got_small_<mode>): New, based on ldr_got_small.
11975 (ldr_got_small): Remove.
11976 (ldr_got_small_sidi): New.
11977 * config/aarch64/iterators.md (P): New.
11978 (PTR): Change to 'ptr_mode' in the condition.
11979
11980 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11981
11982 * config.gcc (aarch64*-*-*): Support --with-abi.
11983 (aarch64*-*-elf): Support --with-multilib-list.
11984 (aarch64*-*-linux*): Likewise.
11985 (supported_defaults): Add abi to aarch64*-*-*.
11986 * configure.ac: Mention AArch64 for --with-multilib-list.
11987 * configure: Re-generated.
11988 * config/aarch64/biarchilp32.h: New file.
11989 * config/aarch64/biarchlp64.h: New file.
11990 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11991 (ABI_SPEC): Ditto.
11992 (MULTILIB_DEFAULTS): Ditto.
11993 (DRIVER_SELF_SPECS): Ditto.
11994 (ASM_SPEC): Update to also substitute -mabi.
11995 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11996 file whose name depends on -mabi= and -mbig-endian.
11997 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11998 TARGET_ILP32.
11999 (POINTER_SIZE): New define.
12000 (POINTERS_EXTEND_UNSIGNED): Ditto.
12001 (enum aarch64_abi_type): New enumeration tag.
12002 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12003 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12004 (TARGET_ILP32): New define.
12005 * config/aarch64/aarch64.opt (mabi): New.
12006 (aarch64_abi): New.
12007 (ilp32, lp64): New values for -mabi.
12008 * config/aarch64/t-aarch64 (comma): New define.
12009 (MULTILIB_OPTIONS): Ditto.
12010 (MULTILIB_DIRNAMES): Ditto.
12011 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12012 * doc/invoke.texi: Document -mabi for AArch64.
12013
12014 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12015
12016 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12017
12018 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12019 Anton Blanchard <anton@au1.ibm.com>
12020
12021 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12022 endianness when selecting field to splat.
12023
12024 2013-07-22 Eric Christopher <echristo@gmail.com>
12025
12026 * dwarf2out.c (die_odr_checksum): New function to use
12027 CHECKSUM_ macros and ULEB128 for DIE tag.
12028 (generate_type_signature): Use.
12029
12030 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12031
12032 * config.gcc (sparc*-*-*): Accept leon3 processor.
12033 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12034 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12035 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12036 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12037 (mfix-ut699): Adjust comment.
12038 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12039 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12040 (CPP_CPU_SPEC): Likewise.
12041 (ASM_CPU_SPEC): Likewise.
12042 * config/sparc/sparc.c (leon3_cost): New constant.
12043 (sparc_option_override): Add leon3 support.
12044 (mem_ref): New function.
12045 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12046 (sparc_do_work_around_errata): Look into the instruction in the delay
12047 slot and adjust accordingly. Add fix for the data cache nullify issues
12048 of the UT699. Change insertion position for the NOP.
12049 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12050 (leon3_load): New reservation.
12051 (leon_store): Bump latency to 2.
12052 (grfpu): New automaton.
12053 (grfpu_alu): New unit.
12054 (grfpu_ds): Likewise.
12055 (leon_fp_alu): Adjust.
12056 (leon_fp_mult): Delete.
12057 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12058 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12059 * config/sparc/sparc.md (cpu): Add leon3.
12060 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12061 (swapsi): Likewise.
12062 (atomic_test_and_set): Likewise.
12063 (ldstub): Likewise.
12064
12065 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12066
12067 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12068 default for R5900 targets.
12069 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12070 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12071 * config/mips/mips.c (mips_option_override): Report an error for
12072 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12073 for -march=r5900 -mhard-float.
12074
12075 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12076
12077 * df-problems.c (can_move_insns_across): Exit loop once we
12078 find a non-fixed, non-global register.
12079
12080 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12081 setting can_throw.
12082
12083 * omega.c (omega_eliminate_red): Break after setting red_found.
12084 (omega_problem_has_red_equations): Similarly after setting found.
12085 (omega_query_variable): Similarly after setting coupled.
12086
12087 2013-07-22 Marek Polacek <polacek@redhat.com>
12088
12089 * gimplify.c: Don't include gimple.h twice.
12090
12091 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12092
12093 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12094 instead of TARGET_THUMB1.
12095 (Pz): New constraint.
12096 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12097 encodings.
12098 (compare_negsi_si): Likewise.
12099 (compare_addsi2_op0): Likewise.
12100 (compare_addsi2_op1): Likewise.
12101 (addsi3_carryin_<optab>): Likewise.
12102 (addsi3_carryin_alt2_<optab>): Likewise.
12103 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12104 for arm_restrict_it.
12105 (subsi3_carryin): Likewise.
12106 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12107 (minmax_arithsi): Disable for arm_restrict_it.
12108 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12109 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12110 (satsi_<SAT:code>_shift): Likewise.
12111 (arm_shiftsi3): Add alternative for 16-bit encoding.
12112 (arm32_movhf): Disable for arm_restrict_it.
12113 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12114 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12115
12116 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12117
12118 * config/arm/arm.md (attribute "insn"): Delete.
12119 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12120 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12121 (not_shiftsi): Update for attribute change.
12122 (not_shiftsi_compare0): Likewise.
12123 (not_shiftsi_compare0_scratch): Likewise.
12124 (arm_one_cmplsi2): Likewise.
12125 (thumb1_one_cmplsi2): Likewise.
12126 (notsi_compare0): Likewise.
12127 (notsi_compare0_scratch): Likewise.
12128 (thumb1_movdi_insn): Likewise.
12129 (arm_movsi_insn): Likewise.
12130 (movhi_insn_arch4): Likewise.
12131 (movhi_bytes): Likewise.
12132 (arm_movqi_insn): Likewise.
12133 (thumb1_movqi_insn): Likewise.
12134 (arm32_movhf): Likewise.
12135 (thumb1_movhf): Likewise.
12136 (arm_movsf_soft_insn): Likewise.
12137 (thumb1_movsf_insn): Likewise.
12138 (thumb_movdf_insn): Likewise.
12139 (movsicc_insn): Likewise.
12140 (movsfcc_soft_insn): Likewise.
12141 (and_scc): Likewise.
12142 (cond_move): Likewise.
12143 (if_move_not): Likewise.
12144 (if_not_move): Likewise.
12145 (if_shift_move): Likewise.
12146 (if_move_shift): Likewise.
12147 (if_shift_shift): Likewise.
12148 (if_not_arith): Likewise.
12149 (if_arith_not): Likewise.
12150 (cond_move_not): Likewise.
12151 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12152 (neon_mov<mode>): Likewise.
12153 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12154 (thumb2_movsi_vfp): Likewise.
12155 (movsf_vfp): Likewise.
12156 (thumb2_movsf_vfp): Likewise.
12157 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12158 change.
12159 (cortexa7_older_only): Likewise.
12160 (cortexa7_younger): Likewise.
12161 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12162 (1020alu_shift_op): Likewise.
12163 (1020alu_shift_reg_op): Likewise.
12164 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12165 (alu_shift_op): Likewise.
12166 (alu_shift_reg_op): Likewise.
12167 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12168 (11_alu_shift_op): Likewise.
12169 (11_alu_shift_reg_op): Likewise.
12170 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12171 (9_alu_shift_reg_op): Likewise.
12172 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12173 change.
12174 (cortex_a15_alu_shift): Likewise.
12175 (cortex_a15_alu_shift_reg): Likewise.
12176 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12177 (cortex_a5_alu_shift): Likewise.
12178 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12179 change.
12180 (cortex_a53_alu_shift): Likewise.
12181 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12182 change.
12183 (cortex_a7_alu_reg): Likewise.
12184 (cortex_a7_alu_shift): Likewise.
12185 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12186 (cortex_a8_alu_shift): Likewise.
12187 (cortex_a8_alu_shift_reg): Likewise.
12188 (cortex_a8_mov): Likewise.
12189 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12190 (cortex_a9_dp_shift): Likewise.
12191 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12192 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12193 (cortex_r4_mov): Likewise.
12194 (cortex_r4_alu_shift): Likewise.
12195 (cortex_r4_alu_shift_reg): Likewise.
12196 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12197 (526_alu_shift_op): Likewise.
12198 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12199 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12200 (626te_alu_shift_op): Likewise.
12201 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12202 (726te_alu_op): Likewise.
12203 (726te_alu_shift_op): Likewise.
12204 (726te_alu_shift_reg_op): Likewise.
12205 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12206 (mp626_alu_shift_op): Likewise.
12207 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12208 (pj4_alu_e1_conds): Likewise.
12209 (pj4_alu): Likewise.
12210 (pj4_alu_conds): Likewise.
12211 (pj4_shift): Likewise.
12212 (pj4_shift_conds): Likewise.
12213 (pj4_alu_shift): Likewise.
12214 (pj4_alu_shift_conds): Likewise.
12215
12216 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12217
12218 * config/arm/predicates.md (shiftable_operator_strict_it):
12219 New predicate.
12220 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12221 Disable cond_exec version for arm_restrict_it.
12222 (thumb2_smaxsi3): Convert to generate cond_exec.
12223 (thumb2_sminsi3): Likewise.
12224 (thumb32_umaxsi3): Likewise.
12225 (thumb2_uminsi3): Likewise.
12226 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12227 (thumb2_neg_abssi2): Likewise.
12228 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12229 (thumb2_movsicc_insn): Adjust alternatives.
12230 (thumb2_mov_negscc): Disable for arm_restrict_it.
12231 (thumb2_mov_negscc_strict_it): New pattern.
12232 (thumb2_mov_notscc_strict_it): New pattern.
12233 (thumb2_mov_notscc): Disable for arm_restrict_it.
12234 (thumb2_ior_scc): Likewise.
12235 (thumb2_ior_scc_strict_it): New pattern.
12236 (thumb2_cond_move): Adjust for arm_restrict_it.
12237 (thumb2_cond_arith): Disable for arm_restrict_it.
12238 (thumb2_cond_arith_strict_it): New pattern.
12239 (thumb2_cond_sub): Adjust for arm_restrict_it.
12240 (thumb2_movcond): Likewise.
12241 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12242 (thumb2_zero_extendhisi2_v6): Likewise.
12243 (thumb2_zero_extendqisi2_v6): Likewise.
12244 (orsi_notsi_si): Likewise.
12245 (orsi_not_shiftsi_si): Likewise.
12246
12247 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12248
12249 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12250 instruction sequence is 1 byte shorter.
12251
12252 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12253
12254 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12255 it is not needed after split.
12256
12257 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12258
12259 PR target/51784
12260 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12261 second label for nonlocal goto receivers. Don't output pic base labels
12262 unless we're producing PIC; mark that action unreachable().
12263 (ix86_save_reg): If the function contains a nonlocal label, save the
12264 PIC base reg.
12265 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12266 * config/darwin.c (emitted_pic_label_num): New GTY.
12267 (update_pic_label_number_if_needed): New.
12268 (machopic_output_function_base_name): Adjust for nonlocal receiver
12269 case.
12270 (machopic_should_output_picbase_label): New.
12271 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12272 (nonlocal_goto_receiver): New insn and split.
12273
12274 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12275
12276 * config/aarch64/aarch64-builtins.c
12277 (aarch64_fold_builtin): Fold abs in all modes.
12278 * config/aarch64/aarch64-simd-builtins.def
12279 (abs): Enable for all modes.
12280 * config/aarch64/arm_neon.h
12281 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12282 (vabs_f64): Add missing intrinsic.
12283
12284 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12285
12286 * config/aarch64/arm_neon.h (vabs_s64): New function
12287
12288 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12289
12290 PR target/57516
12291 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12292 * config/avr/avr.md (adjust_len): Add `round'.
12293 * config/avr/avr-protos.h (avr_out_round): New prototype.
12294 (avr_out_plus): Add `out_label' argument.
12295 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12296 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12297 Handle the case where `insn' is just a pattern.
12298 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12299 (avr_out_round): New function.
12300 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12301
12302 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12303
12304 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12305 Add check for flag_stack_usage to handle -fstack-usage support
12306
12307 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12308
12309 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12310 interaction for new Power8 flags and VSX.
12311
12312 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12313
12314 PR middle-end/57698
12315 * tree-inline.c (expand_call_inline): Emit errors during
12316 early_inlining only if optimization is not turned on.
12317
12318 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12319
12320 * passes.def: New.
12321
12322 * passes.c (init_optimization_passes): Move the construction of
12323 the pass hierarchy into a new passes.def file.
12324
12325 * Makefile.in (passes.o): Add dependency on passes.def.
12326
12327 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12328
12329 * passes.c (init_optimization_passes): Introduce macros for
12330 constructing the tree of passes (INSERT_PASSES_AFTER,
12331 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12332 TERMINATE_PASS_LIST).
12333
12334 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12335 Wei Mi <wmi@google.com>
12336
12337 PR rtl-optimization/57878
12338 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12339 top.
12340 (reload_pseudo_compare_func): Check nregs first for reload
12341 pseudos.
12342
12343 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12344
12345 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12346
12347 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12348
12349 * read-rtl.c (validate_const_int): Once an invalid character is
12350 seen, quit the loop.
12351
12352 * gengtype.c (write_roots): Similarly once we find the "deletable"
12353 or "if_marked" option.
12354
12355 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12356
12357 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12358 "xtab" and "sat". Move value "clz" from here to ...
12359 (attriubte "type"): ... here.
12360 (satsi_<SAT:code>): Delete "insn" attribute.
12361 (satsi_<SAT:code>_shift): Likewise.
12362 (arm_zero_extendqisi2addsi): Likewise.
12363 (arm_extendqisi2addsi): Likewise.
12364 (clzsi2): Update for attribute changes.
12365 (rbitsi2): Likewise.
12366 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12367 attribute.
12368 (arm_usatsihi): Likewise.
12369 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12370
12371 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12372
12373 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12374 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12375 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12376 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12377 in alphabetical order.
12378 (attribute "core_cycles"): Update for attribute changes.
12379 (arm_addsi3): Likewise.
12380 (addsi3_compare0): Likewise.
12381 (addsi3_compare0_scratch): Likewise.
12382 (addsi3_compare_op1): Likewise.
12383 (addsi3_compare_op2): Likewise.
12384 (compare_addsi2_op0): Likewise.
12385 (compare_addsi2_op1): Likewise.
12386 (addsi3_carryin_shift_<optab>): Likewise.
12387 (subsi3_carryin_shift): Likewise.
12388 (rsbsi3_carryin_shift): Likewise.
12389 (arm_subsi3_insn): Likewise.
12390 (subsi3_compare0): Likewise.
12391 (subsi3_compare): Likewise.
12392 (arm_andsi3_insn): Likewise.
12393 (thumb1_andsi3_insn): Likewise.
12394 (andsi3_compare0): Likewise.
12395 (andsi3_compare0_scratch): Likewise.
12396 (zeroextractsi_compare0_scratch
12397 (andsi_not_shiftsi_si): Likewise.
12398 (iorsi3_insn): Likewise.
12399 (iorsi3_compare0): Likewise.
12400 (iorsi3_compare0_scratch): Likewise.
12401 (arm_xorsi3): Likewise.
12402 (thumb1_xorsi3_insn): Likewise.
12403 (xorsi3_compare0): Likewise.
12404 (xorsi3_compare0_scratch): Likewise.
12405 (satsi_<SAT:code>_shift): Likewise.
12406 (rrx): Likewise.
12407 (arm_shiftsi3): Likewise.
12408 (shiftsi3_compare0): Likewise.
12409 (not_shiftsi): Likewise.
12410 (not_shiftsi_compare0): Likewise.
12411 (not_shiftsi_compare0_scratch): Likewise.
12412 (arm_one_cmplsi2): Likewise.
12413 (thumb_one_complsi2): Likewise.
12414 (notsi_compare0): Likewise.
12415 (notsi_compare0_scratch): Likewise.
12416 (thumb1_zero_extendhisi2): Likewise.
12417 (arm_zero_extendhisi2): Likewise.
12418 (arm_zero_extendhisi2_v6): Likewise.
12419 (arm_zero_extendhisi2addsi): Likewise.
12420 (thumb1_zero_extendqisi2): Likewise.
12421 (thumb1_zero_extendqisi2_v6): Likewise.
12422 (arm_zero_extendqisi2): Likewise.
12423 (arm_zero_extendqisi2_v6): Likewise.
12424 (arm_zero_extendqisi2addsi): Likewise.
12425 (thumb1_extendhisi2): Likewise.
12426 (arm_extendhisi2): Likewise.
12427 (arm_extendhisi2_v6): Likewise.
12428 (arm_extendqisi): Likewise.
12429 (arm_extendqisi_v6): Likewise.
12430 (arm_extendqisi2addsi): Likewise.
12431 (thumb1_extendqisi2): Likewise.
12432 (thumb1_movdi_insn): Likewise.
12433 (arm_movsi_insn): Likewise.
12434 (movsi_compare0): Likewise.
12435 (movhi_insn_arch4): Likewise.
12436 (movhi_bytes): Likewise.
12437 (arm_movqi_insn): Likewise.
12438 (thumb1_movqi_insn): Likewise.
12439 (arm32_movhf): Likewise.
12440 (thumb1_movhf): Likewise.
12441 (arm_movsf_soft_insn): Likewise.
12442 (thumb1_movsf_insn): Likewise.
12443 (movdf_soft_insn): Likewise.
12444 (thumb_movdf_insn): Likewise.
12445 (arm_cmpsi_insn): Likewise.
12446 (cmpsi_shiftsi): Likewise.
12447 (cmpsi_shiftsi_swp): Likewise.
12448 (arm_cmpsi_negshiftsi_si): Likewise.
12449 (movsicc_insn): Likewise.
12450 (movsfcc_soft_insn): Likewise.
12451 (arith_shiftsi): Likewise.
12452 (arith_shiftsi_compare0
12453 (arith_shiftsi_compare0_scratch
12454 (sub_shiftsi): Likewise.
12455 (sub_shiftsi_compare0
12456 (sub_shiftsi_compare0_scratch
12457 (and_scc): Likewise.
12458 (cond_move): Likewise.
12459 (if_plus_move): Likewise.
12460 (if_move_plus): Likewise.
12461 (if_move_not): Likewise.
12462 (if_not_move): Likewise.
12463 (if_shift_move): Likewise.
12464 (if_move_shift): Likewise.
12465 (if_shift_shift): Likewise.
12466 (if_not_arith): Likewise.
12467 (if_arith_not): Likewise.
12468 (cond_move_not): Likewise.
12469 (thumb1_ashlsi3): Set type attribute.
12470 (thumb1_ashrsi3): Likewise.
12471 (thumb1_lshrsi3): Likewise.
12472 (thumb1_rotrsi3): Likewise.
12473 (shiftsi3_compare0_scratch): Likewise.
12474 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12475 (neon_mov<mode>): Likewise.
12476 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12477 attribute changes.
12478 (thumb2_movsi_insn): Likewise.
12479 (thumb2_cmpsi_neg_shiftsi): Likewise.
12480 (thumb2_extendqisi_v6): Likewise.
12481 (thumb2_zero_extendhisi2_v6): Likewise.
12482 (thumb2_zero_extendqisi2_v6): Likewise.
12483 (thumb2_shiftsi3_short): Likewise.
12484 (thumb2_addsi3_compare0_scratch): Likewise.
12485 (orsi_not_shiftsi_si): Likewise.
12486 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12487 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12488 changes.
12489 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12490 (1020alu_shift_op): Likewise.
12491 (1020alu_shift_reg_op): Likewise.
12492 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12493 (alu_shift_op): Likewise.
12494 (alu_shift_reg_op): Likewise.
12495 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12496 (11_alu_shift_op): Likewise.
12497 (11_alu_shift_reg_op): Likewise.
12498 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12499 (9_alu_shift_reg_op): Likewise.
12500 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12501 changes.
12502 (cortex_a15_alu_shift): Likewise.
12503 (cortex_a15_alu_shift_reg): Likewise.
12504 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12505 changes.
12506 (cortex_a5_alu_shift): Likewise.
12507 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12508 changes.
12509 (cortex_a53_alu_shift): Likewise.
12510 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12511 changes.
12512 (cortex_a7_alu_reg): Likewise.
12513 (cortex_a7_alu_shift): Likewise.
12514 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12515 changes.
12516 (cortex_a8_alu_shift): Likewise.
12517 (cortex_a8_alu_shift_reg): Likewise.
12518 (cortex_a8_mov): Likewise.
12519 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12520 (cortex_a9_dp_shift): Likewise.
12521 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12522 changes.
12523 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12524 changes.
12525 (cortex_r4_mov): Likewise.
12526 (cortex_r4_alu_shift): Likewise.
12527 (cortex_r4_alu_shift_reg): Likewise.
12528 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12529 (526_alu_shift_op): Likewise.
12530 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12531 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12532 (626te_alu_shift_op): Likewise.
12533 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12534 (726te_alu_op): Likewise.
12535 (726te_alu_shift_op): Likewise.
12536 (726te_alu_shift_reg_op): Likewise.
12537 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12538 (mp626_alu_shift_op): Likewise.
12539 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12540 (pj4_alu_e1_conds): Likewise.
12541 (pj4_alu): Likewise.
12542 (pj4_alu_conds): Likewise.
12543 (pj4_shift): Likewise.
12544 (pj4_shift_conds): Likewise.
12545 (pj4_alu_shift): Likewise.
12546 (pj4_alu_shift_conds): Likewise.
12547 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12548 changes.
12549 (cortexa7_older_only): Likewise.
12550 (cortexa7_younger): Likewise.
12551
12552 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12553
12554 * ipa-pure-const.c (generate_summary): Rename to...
12555 (pure_const_generate_summary): ... this.
12556
12557 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12558
12559 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12560
12561 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12562
12563 PR target/57909
12564 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12565 usage in HI mode.
12566
12567 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12568
12569 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12570 enabled without -march=zEC12.
12571 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12572 flags to be set.
12573
12574 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12575
12576 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12577 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12578 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12579 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12580 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12581 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12582 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12583 ISA_HAS_FP4.
12584 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12585 and ISA_HAS_NMADD3_NMSUB3.
12586 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12587 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12588 (nmsub4<mode>, nmsub3<mode>): Likewise.
12589 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12590
12591 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12592
12593 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12594 TARGET_MIPS5400 checking.
12595
12596 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12597 Peter Bergner <bergner@vnet.ibm.com>
12598
12599 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12600 registers in the comment.
12601 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12602 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12603 rather than FIRST_PSEUDO_REGISTERS.
12604
12605 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12606
12607 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12608 enable extra ISA flags with TARGET_HTM.
12609
12610 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12611
12612 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12613 Fix comment typos.
12614
12615 2013-07-15 Cong Hou <congh@google.com>
12616
12617 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12618 in compare function for sorting.
12619
12620 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12621
12622 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12623 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12624 * config/rs6000/rs6000.opt: Add -mhtm option.
12625 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12626 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12627 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12628 __HTM__ if the HTM instructions are available.
12629 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12630 htm_spr_reg_operand): New define_predicates.
12631 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12632 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12633 Include htm.md.
12634 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12635 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12636 HTM builtin functions.
12637 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12638 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12639 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12640 (rs6000_builtin_mask_calculate): Likewise.
12641 (rs6000_option_override_internal): Likewise.
12642 (bdesc_htm): Add new HTM builtin support.
12643 (htm_spr_num): New function.
12644 (htm_spr_regno): Likewise.
12645 (rs6000_htm_spr_icode): Likewise.
12646 (htm_expand_builtin): Likewise.
12647 (htm_init_builtins): Likewise.
12648 (rs6000_expand_builtin): Add support for HTM builtin functions.
12649 (rs6000_init_builtins): Likewise.
12650 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12651 option.
12652 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12653 (TARGET_HTM, MASK_HTM): Define macros.
12654 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12655 (FIXED_REGISTERS): Likewise.
12656 (CALL_USED_REGISTERS): Likewise.
12657 (CALL_REALLY_USED_REGISTERS): Likewise.
12658 (REG_ALLOC_ORDER): Likewise.
12659 (enum reg_class): Likewise.
12660 (REG_CLASS_NAMES): Likewise.
12661 (REG_CLASS_CONTENTS): Likewise.
12662 (REGISTER_NAMES): Likewise.
12663 (ADDITIONAL_REGISTER_NAMES): Likewise.
12664 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12665 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12666 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12667 * config/rs6000/htm.md: New file.
12668 * config/rs6000/htmintrin.h: New file.
12669 * config/rs6000/htmxlintrin.h: New file.
12670
12671 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12672
12673 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12674 Define SYMBOL_TINY_GOT, update comment.
12675 * config/aarch64/aarch64.c
12676 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12677 (aarch64_expand_mov_immediate): Likewise.
12678 (aarch64_print_operand): Likewise.
12679 (aarch64_classify_symbol): Likewise.
12680 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12681 (ldr_got_tiny): Define.
12682
12683 2013-07-13 Tobias Grosser <tobias@grosser.es>
12684
12685 PR tree-optimization/54094
12686 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12687 scheduling dimension for the parallelism check from the polyhedral
12688 information in the AST.
12689 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12690 in 2D + 1 form.
12691
12692 2013-07-13 Jason Merrill <jason@redhat.com>
12693
12694 * print-tree.c (debug_vec_tree): Use debug_raw.
12695 (debug_raw (vec<tree, va_gc> &)): New.
12696 (debug_raw (vec<tree, va_gc> *)): New.
12697 * tree.h: Declare them.
12698
12699 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12700
12701 * ifcvt.c (ifcvt_after_combine): New static variable.
12702 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12703 for size.
12704 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12705 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12706 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12707
12708 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12709
12710 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12711
12712 2013-07-12 Michael Matz <matz@suse.de>
12713
12714 PR middle-end/55771
12715 * convert.c (convert_to_real): Reject non-float inner types.
12716
12717 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12718
12719 * config/aarch64/aarch64-protos.h
12720 (aarch64_simd_immediate_valid_for_move): Remove.
12721 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12722 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12723 cases.
12724 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12725
12726 2013-07-11 Steve Ellcey <sellcey@mips.com>
12727
12728 * config/mips/mips.c (mips_conditional_register_usage): Do not
12729 use t[0-7] registers in MIPS16 mode when optimizing for size.
12730
12731 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12732
12733 * config/i386/i386.c (dispatch_function_versions): Fix array
12734 indexing of function_version_info to match actual_versions.
12735
12736 2013-07-11 Teresa Johnson <tejohnson@google.com>
12737
12738 * vec.h (struct va_gc): Move release out-of-line.
12739 (va_gc::release): Call ggc_free on released vec.
12740
12741 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12742
12743 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12744 Require GOT register as additional operand in UNSPEC.
12745 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12746 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12747 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12748 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12749 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12750 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12751 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12752
12753 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12754
12755 PR target/57631
12756 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12757 name seen by assembler/linker rather if available.
12758
12759 2013-07-11 Andreas Schwab <schwab@suse.de>
12760
12761 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12762
12763 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12764
12765 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12766
12767 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12768
12769 * doc/tm.texi.in: Move hook documentation to ....
12770 * target.def: ... here.
12771
12772 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12773 text on @hook line.
12774 * doc/tm.texi: Regenerate.
12775
12776 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12777
12778 PR c++/57869
12779 * doc/invoke.texi: Document -Wconditionally-supported.
12780
12781 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12782
12783 PR target/57844
12784 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12785 of my_fp.
12786
12787 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12788
12789 PR target/57506
12790 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12791 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12792 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12793 Remove duplicate devices.
12794 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12795 * config/avr/t-multilib: Regenerate.
12796 * config/avr/avr-tables.opt: Regenerate.
12797 * doc/avr-mmcu.texi: Regenerate.
12798
12799 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12800
12801 PR target/56987
12802 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12803
12804 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12805
12806 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12807 the cost of MULT when optimizing for size.
12808
12809 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12810
12811 * config/cr16/cr16-protos.h: Don't include target.h.
12812
12813 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12814
12815 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12816 adjust register size for TDmode and TFmode for VSX registers.
12817
12818 2013-07-08 Kai Tietz <ktietz@redhat.com>
12819
12820 PR target/56892
12821 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12822 hook_bool_const_tree_true.
12823
12824 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12825
12826 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12827 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12828 * config/s390/s390.md: Define FPR*_REGNUM constants.
12829 Fix FPR2_REGNUM constant (18 -> 17).
12830 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12831 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12832 ("trunc<BFP:mode><DFP_ALL:mode>2")
12833 ("trunc<DFP_ALL:mode><BFP:mode>2")
12834 ("*extend<BFP:mode><DFP_ALL:mode>2")
12835 ("*extend<DFP_ALL:mode><BFP:mode>2")
12836 ("extend<BFP:mode><DFP_ALL:mode>2")
12837 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12838 FPR4_REGNUM.
12839
12840 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12841
12842 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12843
12844 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12845
12846 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12847 and cfun_fpr_bit_p to cfun_fpr_save_p.
12848 (s390_frame_area, s390_register_info, s390_frame_info)
12849 (s390_emit_prologue, s390_emit_epilogue)
12850 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12851 register numbers.
12852 * config/s390/s390.h: Define *_REGNUM macros for floating point
12853 register numbers.
12854
12855 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12856
12857 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12858
12859 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12860
12861 PR rtl-optimization/57786
12862 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12863 and break out of the loop when it is set to false.
12864
12865 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12866
12867 PR target/57819
12868 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12869 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12870 (const_int 63)) 0)).
12871 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12872 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12873 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12874
12875 PR rtl-optimization/57829
12876 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12877 mask bits outside of mode are just sign-extension from mode to HWI.
12878
12879 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12880
12881 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12882 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12883 adjust_address instead of change_address to keep info about alignment.
12884 (emit_strmov): Remove.
12885 (emit_memmov): New function.
12886 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12887 (expand_movmem_epilogue): Likewise and return updated rtx for
12888 destination.
12889 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12890 destination and source.
12891 (decide_alignment): Refactor, handle vector_loop.
12892 (ix86_expand_movmem): Likewise.
12893 (ix86_expand_setmem): Likewise.
12894 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12895
12896 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12897
12898 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12899 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12900
12901 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12902
12903 * config/i386/sse.md (sse_movlhps): Change alternative 3
12904 of operand 2 to "m".
12905
12906 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12907
12908 PR target/57807
12909 * config/i386/sse.md (iptr): New mode attribute.
12910 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12911 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12912 (<sse>_vmmul<mode>3): Ditto.
12913 (<sse>_vmdiv<mode>3): Ditto.
12914 (sse_vmrcpv4sf2): Ditto.
12915 (<sse>_vmsqrt<mode>2): Ditto.
12916 (sse_vmrsqrtv4sf2): Ditto.
12917 (<sse>_vm<code><mode>3): Ditto.
12918 (avx_vmcmp<mode>3): Ditto.
12919 (<sse>_vmmaskcmp<mode>3): Ditto.
12920 (<sse>_comi): Ditto.
12921 (<sse>_ucomi): Ditto.
12922 (*xop_vmfrcz_<mode>): Ditto.
12923 (*fmai_fmadd_<mode>): Ditto.
12924 (*fmai_fmsub_<mode>): Ditto.
12925 (*fmai_fnmadd_<mode>): Ditto.
12926 (*fmai_fnmsub_<mode>): Ditto.
12927 (*fma4i_vmfmadd_<mode>): Ditto.
12928 (*fma4i_vmfmsub_<mode>): Ditto.
12929 (*fma4i_vmfnmadd_<mode>): Ditto.
12930 (*fma4i_vmfnmsub_<mode>): Ditto.
12931 (*xop_vmfrcz_<mode>): Ditto.
12932 (sse_cvtps2pi): Ditto.
12933 (sse_cvttps2pi): Ditto.
12934 (sse_cvtss2si): Ditto.
12935 (sse_cvtss2si_2): Ditto.
12936 (sse_cvtss2siq_2): Ditto.
12937 (sse_cvttss2si): Ditto.
12938 (sse_cvttss2siq): Ditto.
12939 (sse_cvtsd2si): Ditto.
12940 (sse_cvtsd2si_2): Ditto.
12941 (sse_cvtsd2siq_2): Ditto.
12942 (sse_cvttsd2si): Ditto.
12943 (sse_cvttsd2siq): Ditto.
12944 (sse_cvtsd2ss): Ditto.
12945 (sse_cvtss2sd): Ditto.
12946 (avx2_pbroadcast<mode>): Ditto.
12947 (avx2_pbroadcast<mode>_1): Ditto.
12948 (*avx_vperm_broadcast_v4sf): Ditto.
12949
12950 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12951 (sse_movlhps): Ditto.
12952 (sse_storehps): Ditto.
12953 (sse_loadhps): Ditto.
12954 (sse_storelps): Ditto.
12955 (sse_loadlps): Ditto.
12956 (*vec_concatv4sf): Ditto.
12957 (*vec_interleave_highv2df): Ditto.
12958 (*vec_interleave_lowv2df): Ditto.
12959 (*vec_extractv2df_1_sse): Ditto.
12960 (*vec_extractv2df_0_sse): Ditto.
12961 (sse2_storelpd): Ditto.
12962 (sse2_loadlpd): Ditto.
12963 (sse2_movsd): Ditto.
12964 (*vec_concatv4si): Ditto.
12965 (vec_concatv2di): Ditto.
12966
12967 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12968 for Intel asm dialect.
12969 (mmx_punpcklwd): Ditto.
12970 (mmx_punpckldq): Ditto.
12971
12972 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12973 for intel assembler dialect.
12974
12975 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12976
12977 PR target/29776
12978 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12979 for BUILT_IN_C{LZ,LRSB}*.
12980 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12981 * tree-vrp.c (extract_range_basic): Handle
12982 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12983 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12984 fall thru to code calling set_value*.
12985 * builtins.c (expand_builtin): Remove *IMAX cases.
12986 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12987 if width is bigger than 2*HWI.
12988
12989 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12990
12991 PR rtl-optimization/55342
12992 * lra-int.h (lra_subreg_reload_pseudos): New.
12993 * lra.c: Add undoing optional reloads to the block diagram.
12994 (lra_subreg_reload_pseudos): New.
12995 (lra_optional_reload_pseudos): Change comments.
12996 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12997 lra_optional_reload_pseudos after undo transformations.
12998 * lra-assigns.c (pseudo_prefix_title): New.
12999 (lra_setup_reg_renumber): Use it.
13000 (spill_for): Ditto. Check subreg reload pseudos too.
13001 (assign_by_spills): Consider subreg reload pseudos too.
13002 * lra-constraints.c (simplify_operand_subreg): Use
13003 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13004 (curr_insn_transform): Recognize and do optional reloads.
13005 (undo_optional_reloads): New.
13006 (lra_undo_inheritance): Call undo_optional_reloads.
13007
13008 2013-07-05 Thomas Quinot <quinot@adacore.com>
13009
13010 * tree-complex.c (expand_complex_operations_1): Fix typo.
13011
13012 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13013
13014 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13015 (tune_params): New member 'const vec_costs'.
13016 * config/aarch64/aarch64.c (generic_vector_cost): New.
13017 (generic_tunings): New member 'generic_vector_cost'.
13018 (aarch64_builtin_vectorization_cost): New.
13019 (aarch64_add_stmt_cost): New.
13020 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13021 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13022
13023 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13024
13025 PR target/57777
13026 * config/i386/predicates.md (vsib_address_operand): Disallow
13027 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13028
13029 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13030
13031 PR middle-end/55030
13032 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13033 expand_builtin_setjmp_receiver.
13034 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13035 with NULL for the label parameter.
13036 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13037 the frame-pointer. Adjust comments.
13038 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13039 only if LABEL is non-NULL.
13040
13041 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13042
13043 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13044 (ARM_ABI_AAPCS64): Ditto.
13045 (arm_abi): Ditto.
13046 (ARM_DEFAULT_ABI): Ditto.
13047
13048 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13049
13050 * config/aarch64/aarch64-builtins.c
13051 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13052 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13053 (st1): Likewise.
13054 * config/aarch64/aarch64-simd.md
13055 (aarch64_ld1<VALL:mode>): New.
13056 (aarch64_st1<VALL:mode>): Likewise.
13057 * config/aarch64/arm_neon.h
13058 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13059
13060 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13061
13062 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13063 target ISA is AVX.
13064 (ix86_option_override_internal):Turn on all -mavx target flags by
13065 default as they are dependent on AVX anyway.
13066
13067 2013-07-02 Cary Coutant <ccoutant@google.com>
13068
13069 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13070 deterministic hash.
13071 (loc_checksum_ordered): Likewise.
13072 (hash_loc_operands): Remove inline keyword.
13073
13074 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13075
13076 PR tree-optimization/57741
13077 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13078 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13079 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13080 Allow REAL_CST step_exprs if flag_associative_math.
13081 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13082
13083 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13084
13085 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13086
13087 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13088
13089 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13090
13091 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13092
13093 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13094 encoding.
13095 (iorsi3_insn): Likewise.
13096 (arm_xorsi3): Likewise.
13097
13098 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13099
13100 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13101 to ...
13102 (attribute "type"): ... here, and prefix with "wmmx_".
13103 (attribute "core_cycles"): Update for attribute changes.
13104 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13105 (tbcstv4hi): Likewise.
13106 (tbcstv2si): Likewise.
13107 (iwmmxt_iordi3): Likewise.
13108 (iwmmxt_xordi3): Likewise.
13109 (iwmmxt_anddi3): Likewise.
13110 (iwmmxt_nanddi3): Likewise.
13111 (iwmmxt_arm_movdi): Likewise.
13112 (iwmmxt_movsi_insn): Likewise.
13113 (mov<mode>_internal): Likewise.
13114 (and<mode>3_iwmmxt): Likewise.
13115 (ior<mode>3_iwmmxt): Likewise.
13116 (xor<mode>3_iwmmxt): Likewise.
13117 (add<mode>3_iwmmxt): Likewise.
13118 (ssaddv8qi3): Likewise.
13119 (ssaddv4hi3): Likewise.
13120 (ssaddv2si3): Likewise.
13121 (usaddv8qi3): Likewise.
13122 (usaddv4hi3): Likewise.
13123 (usaddv2si3): Likewise.
13124 (sub<mode>3_iwmmxt): Likewise.
13125 (sssubv8qi3): Likewise.
13126 (sssubv4hi3): Likewise.
13127 (sssubv2si3): Likewise.
13128 (ussubv8qi3): Likewise.
13129 (ussubv4hi3): Likewise.
13130 (ussubv2si3): Likewise.
13131 (mulv4hi3_iwmmxt): Likewise.
13132 (smulv4hi3_highpart): Likewise.
13133 (umulv4hi3_highpart): Likewise.
13134 (iwmmxt_wmacs): Likewise.
13135 (iwmmxt_wmacsz): Likewise.
13136 (iwmmxt_wmacu): Likewise.
13137 (iwmmxt_wmacuz): Likewise.
13138 (iwmmxt_clrdi): Likewise.
13139 (iwmmxt_clrv8qi): Likewise.
13140 (iwmmxt_clr4hi): Likewise.
13141 (iwmmxt_clr2si): Likewise.
13142 (iwmmxt_uavgrndv8qi3): Likewise.
13143 (iwmmxt_uavgrndv4hi3): Likewise.
13144 (iwmmxt_uavgv8qi3): Likewise.
13145 (iwmmxt_uavgv4hi3): Likewise.
13146 (iwmmxt_tinsrb): Likewise.
13147 (iwmmxt_tinsrh): Likewise.
13148 (iwmmxt_tinsrw): Likewise.
13149 (iwmmxt_textrmub): Likewise.
13150 (iwmmxt_textrmsb): Likewise.
13151 (iwmmxt_textrmuh): Likewise.
13152 (iwmmxt_textrmsh): Likewise.
13153 (iwmmxt_textrmw): Likewise.
13154 (iwmxxt_wshufh): Likewise.
13155 (eqv8qi3): Likewise.
13156 (eqv4hi3): Likewise.
13157 (eqv2si3): Likewise.
13158 (gtuv8qi3): Likewise.
13159 (gtuv4hi3): Likewise.
13160 (gtuv2si3): Likewise.
13161 (gtv8qi3): Likewise.
13162 (gtv4hi3): Likewise.
13163 (gtv2si3): Likewise.
13164 (smax<mode>3_iwmmxt): Likewise.
13165 (umax<mode>3_iwmmxt): Likewise.
13166 (smin<mode>3_iwmmxt): Likewise.
13167 (umin<mode>3_iwmmxt): Likewise.
13168 (iwmmxt_wpackhss): Likewise.
13169 (iwmmxt_wpackwss): Likewise.
13170 (iwmmxt_wpackdss): Likewise.
13171 (iwmmxt_wpackhus): Likewise.
13172 (iwmmxt_wpackwus): Likewise.
13173 (iwmmxt_wpackdus): Likewise.
13174 (iwmmxt_wunpckihb): Likewise.
13175 (iwmmxt_wunpckihh): Likewise.
13176 (iwmmxt_wunpckihw): Likewise.
13177 (iwmmxt_wunpckilb): Likewise.
13178 (iwmmxt_wunpckilh): Likewise.
13179 (iwmmxt_wunpckilw): Likewise.
13180 (iwmmxt_wunpckehub): Likewise.
13181 (iwmmxt_wunpckehuh): Likewise.
13182 (iwmmxt_wunpckehuw): Likewise.
13183 (iwmmxt_wunpckehsb): Likewise.
13184 (iwmmxt_wunpckehsh): Likewise.
13185 (iwmmxt_wunpckehsw): Likewise.
13186 (iwmmxt_wunpckelub): Likewise.
13187 (iwmmxt_wunpckeluh): Likewise.
13188 (iwmmxt_wunpckeluw): Likewise.
13189 (iwmmxt_wunpckelsb): Likewise.
13190 (iwmmxt_wunpckelsh): Likewise.
13191 (iwmmxt_wunpckelsw): Likewise.
13192 (ror<mode>3): Likewise.
13193 (ashr<mode>3_iwmmxt): Likewise.
13194 (lshr<mode>3_iwmmxt): Likewise.
13195 (ashl<mode>3_iwmmxt): Likewise.
13196 (ror<mode>3_di): Likewise.
13197 (ashr<mode>3_di): Likewise.
13198 (lshr<mode>3_di): Likewise.
13199 (ashl<mode>3_di): Likewise.
13200 (iwmmxt_wmadds): Likewise.
13201 (iwmmxt_wmaddu): Likewise.
13202 (iwmmxt_tmia): Likewise.
13203 (iwmmxt_tmiaph): Likewise.
13204 (iwmmxt_tmiabb): Likewise.
13205 (iwmmxt_tmiatb): Likewise.
13206 (iwmmxt_tmiabt): Likewise.
13207 (iwmmxt_tmiatt): Likewise.
13208 (iwmmxt_tmovmskb): Likewise.
13209 (iwmmxt_tmovmskh): Likewise.
13210 (iwmmxt_tmovmskw): Likewise.
13211 (iwmmxt_waccb): Likewise.
13212 (iwmmxt_wacch): Likewise.
13213 (iwmmxt_waccw): Likewise.
13214 (iwmmxt_waligni): Likewise.
13215 (iwmmxt_walignr): Likewise.
13216 (iwmmxt_walignr0): Likewise.
13217 (iwmmxt_walignr1): Likewise.
13218 (iwmmxt_walignr2): Likewise.
13219 (iwmmxt_walignr3): Likewise.
13220 (iwmmxt_wsadb): Likewise.
13221 (iwmmxt_wsadh): Likewise.
13222 (iwmmxt_wsadbz): Likewise.
13223 (iwmmxt_wsadhz): Likewise.
13224 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13225 (iwmmxt_wabsdiffb): Likewise.
13226 (iwmmxt_wabsdiffh): Likewise.
13227 (iwmmxt_wabsdiffw): Likewise.
13228 (iwmmxt_waddsubhx): Likewise
13229 (iwmmxt_wsubaddhx): Likewise.
13230 (addc<mode>3): Likewise.
13231 (iwmmxt_avg4): Likewise.
13232 (iwmmxt_avg4r): Likewise.
13233 (iwmmxt_wmaddsx): Likewise.
13234 (iwmmxt_wmaddux): Likewise.
13235 (iwmmxt_wmaddsn): Likewise.
13236 (iwmmxt_wmaddun): Likewise.
13237 (iwmmxt_wmulwsm): Likewise.
13238 (iwmmxt_wmulwum): Likewise.
13239 (iwmmxt_wmulsmr): Likewise.
13240 (iwmmxt_wmulumr): Likewise.
13241 (iwmmxt_wmulwsmr): Likewise.
13242 (iwmmxt_wmulwumr): Likewise.
13243 (iwmmxt_wmulwl): Likewise.
13244 (iwmmxt_wqmulm): Likewise.
13245 (iwmmxt_wqmulwm): Likewise.
13246 (iwmmxt_wqmulmr): Likewise.
13247 (iwmmxt_wqmulwmr): Likewise.
13248 (iwmmxt_waddbhusm): Likewise.
13249 (iwmmxt_waddbhusl): Likewise.
13250 (iwmmxt_wqmiabb): Likewise.
13251 (iwmmxt_wqmiabt): Likewise.
13252 (iwmmxt_wqmiatb): Likewise.
13253 (iwmmxt_wqmiatt): Likewise.
13254 (iwmmxt_wqmiabbn): Likewise.
13255 (iwmmxt_wqmiabtn): Likewise.
13256 (iwmmxt_wqmiatbn): Likewise.
13257 (iwmmxt_wqmiattn): Likewise.
13258 (iwmmxt_wmiabb): Likewise.
13259 (iwmmxt_wmiabt): Likewise.
13260 (iwmmxt_wmiatb): Likewise.
13261 (iwmmxt_wmiatt): Likewise.
13262 (iwmmxt_wmiabbn): Likewise.
13263 (iwmmxt_wmiabtn): Likewise.
13264 (iwmmxt_wmiatbn): Likewise.
13265 (iwmmxt_wmiattn): Likewise.
13266 (iwmmxt_wmiawbb): Likewise.
13267 (iwmmxt_wmiawbt): Likewise.
13268 (iwmmxt_wmiawtb): Likewise.
13269 (iwmmxt_wmiawtt): Likewise.
13270 (iwmmxt_wmiawbbn): Likewise.
13271 (iwmmxt_wmiawbtn): Likewise.
13272 (iwmmxt_wmiawtbn): Likewise.
13273 (iwmmxt_wmiawttn): Likewise.
13274 (iwmmxt_wmerge): Likewise.
13275 (iwmmxt_tandc<mode>3): Likewise.
13276 (iwmmxt_torc<mode>3): Likewise.
13277 (iwmmxt_torvsc<mode>3): Likewise.
13278 (iwmmxt_textrc<mode>3): Likewise.
13279 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13280 (wmmxt_pack): Likewise.
13281 (wmmxt_mult_c1): Likewise.
13282 (wmmxt_mult_c2): Likewise.
13283 (wmmxt_alu_c1): Likewise.
13284 (wmmxt_alu_c2): Likewise.
13285 (wmmxt_alu_c3): Likewise.
13286 (wmmxt_transfer_c1): Likewise.
13287 (wmmxt_transfer_c2): Likewise.
13288 (wmmxt_transfer_c3): Likewise.
13289 (marvell_f_iwmmxt_wstr): Likewise.
13290 (marvell_f_iwmmxt_wldr): Likewise.
13291
13292 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13293
13294 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13295
13296 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13297
13298 Revert:
13299 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13300 * lra-constraints.c (need_for_split_p): Check call used hard regs
13301 living through calls.
13302
13303 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13304 call used regs for call insn.
13305
13306 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13307
13308 PR target/57736
13309 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13310 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13311
13312 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13313
13314 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13315
13316 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13317
13318 * lra-constraints.c (need_for_split_p): Check call used hard regs
13319 living through calls.
13320
13321 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13322
13323 PR target/57744
13324 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13325 to tie with any other modes. Eliminate Altivec vector mode tests,
13326 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13327 code, to return 0 if testing MODE2 for a condition, if we've
13328 already tested MODE1 for the same condition.
13329
13330 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13331
13332 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13333 layout.
13334
13335 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13336
13337 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13338 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13339
13340 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13341
13342 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13343 Define.
13344 (aarch64_symbolic_constant_p): Remove.
13345 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13346 static. Fix line length and white space.
13347 (aarch64_symbolic_constant_p): Remove.
13348 * config/aarch64/predicates.md (aarch64_valid_symref):
13349 Use aarch64_classify_symbol_expression.
13350
13351 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13352
13353 * config/arm/constraints.md (Ts): New constraint.
13354 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13355 16-bit encodings.
13356 (compare_scc): Use "Ts" constraint for operand 0.
13357 (ior_scc_scc): Likewise.
13358 (and_scc_scc): Likewise.
13359 (and_scc_scc_nodom): Likewise.
13360 (ior_scc_scc_cmp): Likewise for operand 7.
13361 (and_scc_scc_cmp): Likewise.
13362 * config/arm/thumb2.md (thumb2_movsi_insn):
13363 Add alternatives for 16-bit encodings.
13364 (thumb2_movhi_insn): Likewise.
13365 (thumb2_movsicc_insn): Likewise.
13366 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13367 (thumb2_negscc): Use "Ts" constraint.
13368 Move mvn instruction outside cond_exec block.
13369 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13370 for 16-bit encodings.
13371
13372 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13373
13374 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13375 encoding.
13376 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13377 (mulsi3subsi): Likewise.
13378 (mulsidi3adddi): Likewise.
13379 (mulsidi3_v6): Likewise.
13380 (umulsidi3_v6): Likewise.
13381 (umulsidi3adddi_v6): Likewise.
13382 (smulsi3_highpart_v6): Likewise.
13383 (umulsi3_highpart_v6): Likewise.
13384 (mulhisi3tb): Likewise.
13385 (mulhisi3bt): Likewise.
13386 (mulhisi3tt): Likewise.
13387 (maddhisi4): Likewise.
13388 (maddhisi4tb): Likewise.
13389 (maddhisi4tt): Likewise.
13390 (maddhidi4): Likewise.
13391 (maddhidi4tb): Likewise.
13392 (maddhidi4tt): Likewise.
13393 (zeroextractsi_compare0_scratch): Likewise.
13394 (insv_zero): Likewise.
13395 (insv_t2): Likewise.
13396 (anddi_notzesidi_di): Likewise.
13397 (anddi_notsesidi_di): Likewise.
13398 (andsi_notsi_si): Likewise.
13399 (iordi_zesidi_di): Likewise.
13400 (xordi_zesidi_di): Likewise.
13401 (andsi_iorsi3_notsi): Likewise.
13402 (smax_0): Likewise.
13403 (smax_m1): Likewise.
13404 (smin_0): Likewise.
13405 (not_shiftsi): Likewise.
13406 (unaligned_loadsi): Likewise.
13407 (unaligned_loadhis): Likewise.
13408 (unaligned_loadhiu): Likewise.
13409 (unaligned_storesi): Likewise.
13410 (unaligned_storehi): Likewise.
13411 (extv_reg): Likewise.
13412 (extzv_t2): Likewise.
13413 (divsi3): Likewise.
13414 (udivsi3): Likewise.
13415 (arm_zero_extendhisi2addsi): Likewise.
13416 (arm_zero_extendqisi2addsi): Likewise.
13417 (compareqi_eq0): Likewise.
13418 (arm_extendhisi2_v6): Likewise.
13419 (arm_extendqisi2addsi): Likewise.
13420 (arm_movt): Likewise.
13421 (thumb2_ldrd): Likewise.
13422 (thumb2_ldrd_base): Likewise.
13423 (thumb2_ldrd_base_neg): Likewise.
13424 (thumb2_strd): Likewise.
13425 (thumb2_strd_base): Likewise.
13426 (thumb2_strd_base_neg): Likewise.
13427 (arm_negsi2): Add alternative for 16-bit encoding.
13428 (arm_one_cmplsi2): Likewise.
13429
13430 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13431
13432 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13433 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13434 (movdfcc): Likewise.
13435 * config/arm/vfp.md (*thumb2_movsf_vfp):
13436 Disable predication for arm_restrict_it.
13437 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13438 (*thumb2_movdfcc_vfp): Likewise.
13439 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13440 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13441 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13442 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13443 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13444 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13445 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13446 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13447 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13448 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13449 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13450 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13451 Disable predication for arm_restrict_it.
13452
13453 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13454
13455 * config/i386/bmiintrin.h (_bextr_u32): New.
13456 (_bextr_u64): Ditto.
13457
13458 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13459
13460 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13461 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13462 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13463 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13464 * config/mips/n32-elf.h: ...this new file.
13465
13466 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13467
13468 PR target/57224
13469 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13470 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13471
13472 2013-06-27 Catherine Moore <clm@codesourcery.com>
13473
13474 * config/mips/mips-tables.opt: Regenerate.
13475 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13476 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13477 * doc/invoke.texi: Add -m14kc.
13478
13479 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13480
13481 PR target/57623
13482 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13483 constraints of operand 1 and 2.
13484
13485 PR target/57623
13486 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13487 to match RTL canonicalization. Swap predicates and
13488 constraints of operand 1 and 2.
13489
13490 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13491
13492 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13493 Process OP_INOUT regs for splitting too.
13494
13495 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13496
13497 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13498 decl before the loop, initialize to NULL.
13499 (vectorizable_load): Initialize ptr_incr to NULL.
13500
13501 2013-06-27 Martin Jambor <mjambor@suse.cz>
13502
13503 PR lto/57208
13504 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13505 * ipa-ref.c (ipa_maybe_record_reference): New function.
13506 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13507 * ipa-cp.c (create_specialized_node): Record potential references from
13508 aggvals.
13509 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13510
13511 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13512
13513 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13514 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13515 'mode' to force_reg.
13516 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13517 (aarch64_expand_mov_immediate): Likewise.
13518
13519 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13520
13521 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13522 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13523
13524 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13525
13526 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13527 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13528 (struct machine_function): Add tbegin_p.
13529 (s390_canonicalize_comparison): Fold CC mode compares to
13530 conditional jump if possible.
13531 (s390_emit_jump): Return the emitted jump.
13532 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13533 Handle CCRAWmode compares.
13534 (s390_option_override): Default to -mhtm if available.
13535 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13536 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13537 FPRs instead of df_regs_ever_live_p.
13538 (s390_optimize_nonescaping_tx): New function.
13539 (s390_init_frame_layout): Extend clobbered_regs array to cover
13540 FPRs as well.
13541 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13542 (s390_expand_tbegin): New function.
13543 (enum s390_builtin): New enum definition.
13544 (code_for_builtin): New array definition.
13545 (s390_init_builtins): New function.
13546 (s390_expand_builtin): New function.
13547 (TARGET_INIT_BUILTINS): Define.
13548 (TARGET_EXPAND_BUILTIN): Define.
13549 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13550 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13551 (s390_alc_comparison): Likewise.
13552 * config/s390/s390-modes.def: Add CCRAWmode.
13553 * config/s390/s390.h (processor_flags): Add PF_TX.
13554 (TARGET_CPU_HTM): Define macro.
13555 (TARGET_HTM): Define macro.
13556 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13557 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13558 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13559 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13560 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13561 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13562 ("*cc_to_int"): Move up.
13563 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13564 constants other than 0.
13565 ("*ccraw_to_int"): New insn and splitter definition.
13566 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13567 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13568 ("tx_assist"): New expander.
13569 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13570 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13571 * config/s390/s390.opt: Add -mhtm option.
13572 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13573 * config/s390/htmxlintrin.h: New file.
13574 * config/s390/htmintrin.h: New file.
13575 * config/s390/s390intrin.h: New file.
13576 * doc/extend.texi: Document htm builtins.
13577 * config.gcc: Add the new header files to extra_headers.
13578
13579 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13580
13581 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13582 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13583
13584 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13585 Pat Haugen <pthaugen@us.ibm.com>
13586 Peter Bergner <bergner@vnet.ibm.com>
13587
13588 * config/rs6000/power8.md: New.
13589 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13590 setting for power8 entry.
13591 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13592 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13593 test for Power4/Power5 only.
13594 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13595 support.
13596 (force_new_group): Adjust comment.
13597 * config/rs6000/rs6000.md: Include power8.md.
13598
13599 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13600
13601 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13602 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13603 declaration.
13604 (tune_params): Update comment.
13605 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13606 (arm_max_conditional_execute): New function.
13607 (thumb2_final_prescan_insn): Use max_insn_skipped and
13608 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13609
13610 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13611
13612 PR tree-optimization/57705
13613 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13614 SSA_NAME step, provided that it is not defined inside the loop.
13615 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13616 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13617
13618 2013-06-25 Martin Jambor <mjambor@suse.cz>
13619
13620 PR middle-end/57670
13621 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13622 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13623 calls in the dump.
13624 (ipa_note_param_call): Initialize member_ptr flag.
13625 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13626 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13627 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13628 (ipa_read_indirect_edge_info): Likewise.
13629
13630 2013-06-25 Richard Biener <rguenther@suse.de>
13631
13632 PR middle-end/56977
13633 * passes.c (init_optimization_passes): Move pass_fold_builtins
13634 and pass_dce earlier with -Og.
13635
13636 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13637
13638 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13639 <BIT_FIELD_REF>: Remove trailing TAB.
13640 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13641 remove blank line.
13642
13643 2013-06-24 Martin Jambor <mjambor@suse.cz>
13644
13645 PR tree-optimization/57358
13646 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13647 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13648 (ipa_analyze_params_uses): Generate pessimistic info when true.
13649
13650 2013-06-24 Martin Jambor <mjambor@suse.cz>
13651
13652 PR tree-optimization/57539
13653 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13654 global.inlined_to of the new node to it. All callers changed.
13655 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13656 inlining_into, pass it to cgraph_clone_node.
13657 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13658 ipa_free_edge_args_substructures.
13659 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13660 rdesc linked list. Do not assert rdesc edges have inlined caller.
13661 Assert we have found an rdesc in the rdesc list.
13662
13663 2013-06-24 Richard Biener <rguenther@suse.de>
13664
13665 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13666 (pointer_set_lookup): Declare.
13667 (class pointer_map): New template class implementing a
13668 generic pointer to T map.
13669 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13670 pointer_map<T>::contains, pointer_map<T>::insert,
13671 pointer_map<T>::traverse): New functions.
13672 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13673 (pointer_set_lookup): New function.
13674 (pointer_set_contains): Use pointer_set_lookup.
13675 (pointer_set_insert): Likewise.
13676 (insert_aux): Remove.
13677 (struct pointer_map_t): Embed a pointer_set_t.
13678 (pointer_map_create): Adjust.
13679 (pointer_map_destroy): Likewise.
13680 (pointer_map_contains): Likewise.
13681 (pointer_map_insert): Likewise.
13682 (pointer_map_traverse): Likewise.
13683 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13684 pointer_map<unsigned> instead of a pointer_map_t.
13685 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13686 (streamer_tree_cache_lookup): Likewise.
13687 (streamer_tree_cache_create): Likewise.
13688 (streamer_tree_cache_delete): Likewise.
13689 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13690 pointer_map<unsigned> instead of a pointer_map_t.
13691 (lto_init_tree_ref_encoder): Adjust.
13692 (lto_destroy_tree_ref_encoder): Likewise.
13693 * lto-section-out.c (lto_output_decl_index): Likewise.
13694 (lto_record_function_out_decl_state): Likewise.
13695 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13696
13697 2013-06-24 Richard Biener <rguenther@suse.de>
13698
13699 PR tree-optimization/57488
13700 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13701
13702 2013-06-24 Alan Modra <amodra@gmail.com>
13703
13704 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13705 (gen_easy_altivec_constant): Likewise.
13706 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13707 easy_vector_constant_msb): Likewise.
13708
13709 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13710
13711 PR target/57688
13712 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13713 add missing return true.
13714
13715 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13716
13717 PR target/52483
13718 * config/sh/predicates.md (general_extend_operand): Invoke
13719 general_movsrc_operand for memory operands.
13720 (general_movsrc_operand): Allow reg+reg addressing, do not use
13721 general_operand for memory operands.
13722
13723 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13724
13725 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13726 when current target options does not apply.
13727 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13728 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13729 * config/i386/bmiintrin.h: Pass appropriate target
13730 attributes to header.
13731 * config/i386/mmintrin.h: Ditto.
13732 * config/i386/nmmintrin.h: Ditto.
13733 * config/i386/avx2intrin.h: Ditto.
13734 * config/i386/fxsrintrin.h: Ditto.
13735 * config/i386/tbmintrin.h: Ditto.
13736 * config/i386/xsaveintrin.h: Ditto.
13737 * config/i386/f16cintrin.h: Ditto.
13738 * config/i386/xtestintrin.h: Ditto.
13739 * config/i386/xsaveoptintrin.h: Ditto.
13740 * config/i386/bmi2intrin.h: Ditto.
13741 * config/i386/lzcntintrin.h: Ditto.
13742 * config/i386/smmintrin.h: Ditto.
13743 * config/i386/wmmintrin.h: Ditto.
13744 * config/i386/x86intrin.h: Remove all header include guards.
13745 * config/i386/prfchwintrin.h: Ditto.
13746 * config/i386/pmmintrin.h: Ditto.
13747 * config/i386/tmmintrin.h: Ditto.
13748 * config/i386/xmmintrin.h: Ditto.
13749 * config/i386/popcntintrin.h: Ditto.
13750 * config/i386/rdseedintrin.h: Ditto.
13751 * config/i386/ammintrin.h: Ditto.
13752 * config/i386/emmintrin.h: Ditto.
13753 * config/i386/immintrin.h: Remove all header include guards.
13754 * config/i386/fma4intrin.h: Ditto.
13755 * config/i386/lwpintrin.h: Ditto.
13756 * config/i386/xopintrin.h: Ditto.
13757 * config/i386/ia32intrin.h: Ditto.
13758 * config/i386/avxintrin.h: Ditto.
13759 * config/i386/rtmintrin.h: Ditto.
13760 * config/i386/fmaintrin.h: Ditto.
13761 * config/i386/mm3dnow.h: Ditto.
13762
13763 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13764
13765 * common/config/i386/i386-common.c: Handle LZCNT.
13766
13767 2013-06-22 Andi Kleen <ak@linux.intel.com>
13768
13769 * doc/extend.texi: Use __atomic_store_n instead of
13770 __atomic_store in HLE example.
13771
13772 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13773
13774 * config/sh/sh.c: Remove <cstdlib> workaround.
13775
13776 2013-06-21 Andi Kleen <ak@linux.intel.com>
13777
13778 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13779
13780 2013-06-21 Andi Kleen <ak@linux.intel.com>
13781
13782 * doc/extend.texi: Document that __atomic_clear and
13783 __atomic_test_and_set should only be used with bool.
13784
13785 2013-06-20 Jan Hubicka <jh@suse.cz>
13786
13787 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13788 types_same_for_odr.
13789 * tree.c (decls_same_for_odr): New function.
13790 (same_for_edr): New function.
13791 (types_same_for_odr): New function.
13792 (get_binfo_at_offset): Use it.
13793 * tree.h (types_same_for_odr): Declare.
13794
13795 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13796 Jason Merrill <jason@redhat.com>
13797
13798 * system.h: Include <cstdlib> as well as <stdlib.h>.
13799
13800 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13801
13802 PR target/57655
13803 * config/i386/i386.c (construct_container): Report error if
13804 long double is used with disabled x87 float returns.
13805
13806 2013-06-20 Jan Hubicka <jh@suse.cz>
13807
13808 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13809
13810 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13811
13812 PR rtl-optimization/57425
13813 PR rtl-optimization/57569
13814 * alias.c (write_dependence_p): Remove parameters mem_mode and
13815 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13816 Changed all callers.
13817 (canon_anti_dependence): Get comments and semantics in sync.
13818 Add parameter mem_canonicalized. Changed all callers.
13819 * rtl.h (canon_anti_dependence): Update prototype.
13820
13821 2013-06-20 Richard Biener <rguenther@suse.de>
13822
13823 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13824 byte case, inline streamer_read_uchar and defer section
13825 overrun check.
13826
13827 2013-06-20 Richard Biener <rguenther@suse.de>
13828
13829 PR tree-optimization/57584
13830 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13831 SSA names into the expanded expression that take part in
13832 abnormal coalescing.
13833
13834 2013-06-19 Sharad Singhai <singhai@google.com>
13835
13836 * gcov.c (print_usage): Handle new option.
13837 (process_args): Ditto.
13838 (get_gcov_intermediate_filename): New function.
13839 (output_intermediate_file): New function.
13840 (output_gcov_file): New function
13841 (generate_results): Handle new option.
13842 (release_function): Relase demangled name.
13843 (read_graph_file): Handle demangled name.
13844 (output_lines): Ditto.
13845 * doc/gcov.texi: Document gcov intermediate format.
13846
13847 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13848
13849 PR bootstrap/57604
13850 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13851 (lra_emit_add): Use the functions. Add comment about Y as an
13852 address segment.
13853
13854 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13855
13856 PR driver/57652
13857 * collect2.c (collect_atexit): New.
13858 (collect_exit): Delete.
13859 (main): Register collect_atexit with atexit.
13860 (collect_wait): Change collect_exit to exit.
13861 (do_wait): Same.
13862 * collect2.h (collect_exit): Delete.
13863 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13864
13865 2013-06-19 Wei Mi <wmi@google.com>
13866
13867 PR rtl-optimization/57518
13868 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13869 if regno is used in paradoxical subreg.
13870 (update_equiv_regs): Check pdx_subregs[regno] before
13871 set a reg to be equivalent with a mem.
13872
13873 2013-06-19 Matthias Klose <doko@ubuntu.com>
13874
13875 PR driver/57651
13876 * file-find.h (find_a_file): Add a mode parameter.
13877 * file-find.c (find_a_file): Likewise.
13878 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13879 with X_OK for the executables.
13880 * collect2.c (main): Call find_a_file with X_OK.
13881
13882 2013-06-19 Steve Ellcey <sellcey@mips.com>
13883
13884 PR target/56942
13885 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13886 Use NEXT_INSN instead of next_real_insn.
13887
13888 2013-06-19 Jan Hubicka <jh@suse.cz>
13889
13890 * cgraph.h (const_value_known_p): Replace by ...
13891 (ctor_for_folding): .. this one.
13892 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13893 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13894 * expr.c (expand_expr_real_1): Likewise.
13895 (string_constant): Likewise.
13896 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13897 * ipa.c (process_references): Likewise.
13898 (symtab_remove_unreachable_nodes): Likewise.
13899 * ipa-inline-analysis.c (param_change_prob): Likewise.
13900 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13901 (get_base_constructor): Likwise.
13902 * varpool.c (varpool_remove_node): Likewise.
13903 (varpool_remove_initializer): LIkewise.
13904 (dump_varpool_node): LIkwise.
13905 (const_value_known_p): Rewrite to ...
13906 (ctor_for_folding): ... this one.
13907
13908 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13909
13910 PR driver/57651
13911 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13912 PERSONALITY in $PATH derived prefixes.
13913
13914 2013-06-19 Jeff Law <law@redhat.com>
13915
13916 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13917 in comment.
13918
13919 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13920 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13921 booleans.
13922
13923 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13924
13925 * config/arm/vfp.md: Move VFP instruction classification documentation
13926 to ...
13927 * config/arm/arm.md: ... here. Update instruction classification
13928 documentation.
13929
13930 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13931
13932 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13933 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13934 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13935
13936 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13937
13938 PR target/57609
13939 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13940 with NEXT_INSN. Use tablejump_p to check for jump table data
13941 insns.
13942
13943 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13944
13945 PR c++/56544
13946 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13947 that now in C++ the value is correct per the C++ standards.
13948
13949 2013-06-19 Richard Biener <rguenther@suse.de>
13950
13951 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13952 for global context.
13953
13954 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13955
13956 Revert:
13957 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13958
13959 PR target/57609
13960 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13961 with next_active_insn.
13962
13963 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13964
13965 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13966 functions are inlined during failures to flag an error.
13967 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13968 in early inline pass.
13969
13970 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13971
13972 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13973 in comments.
13974
13975 2013-06-18 Julian Brown <julian@codesourcery.com>
13976
13977 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13978 Permit virtual register pre-reload if !strict.
13979 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13980 change.
13981 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13982 prototype.
13983 * config/arm/neon.md (movmisalign<mode>): Use
13984 neon_perm_struct_or_reg_operand instead of
13985 neon_struct_or_register_operand.
13986 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13987 neon_permissive_struct_operand instead of neon_struct_operand.
13988 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13989 neon_vector_mem_operand.
13990 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13991 neon_vector_mem_operand.
13992 (neon_permissive_struct_operand): New.
13993 (neon_struct_or_register_operand): Rename to...
13994 (neon_perm_struct_or_reg_operand): This. Adjust call to
13995 neon_vector_mem_operand.
13996
13997 2013-06-18 Richard Biener <rguenther@suse.de>
13998
13999 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14000 * lto-streamer.h: Include pointer-set.h.
14001 (struct lto_decl_slot): Remove.
14002 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14003 Remove next_index entry.
14004 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14005 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14006 (lto_init_tree_ref_encoder): Adjust.
14007 (lto_destroy_tree_ref_encoder): Likewise.
14008 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14009 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14010 (lto_output_decl_index): Adjust.
14011 (lto_new_out_decl_state): Likewise.
14012 (lto_record_function_out_decl_state): Likewise.
14013 * lto-streamer-out.c (copy_function): Likewise.
14014
14015 2013-06-18 Richard Biener <rguenther@suse.de>
14016
14017 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14018 * cgraphunit.c: Include cfgloop.h.
14019 (init_lowered_empty_function): Initialize the loop tree.
14020 (assemble_thunk): Insert new BBs into loops.
14021
14022 2013-06-18 Richard Biener <rguenther@suse.de>
14023
14024 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14025 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14026 the map from cache entry to cache index optional.
14027 (streamer_tree_cache_replace_tree): Adjust accordingly.
14028 (streamer_tree_cache_append): Likewise.
14029 (streamer_tree_cache_delete): Likewise.
14030 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14031 streamer cache map from cache entry to cache index.
14032 * lto-streamer-out.c (create_output_block): Adjust.
14033
14034 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14035
14036 * config/arm/arm.md (attribute "insn"): Move multiplication and
14037 division attributes to...
14038 (attribute "type"): ... here. Remove mult.
14039 (attribute "mul32"): New attribute.
14040 (attribute "mul64"): Add umaal.
14041 (*arm_mulsi3): Update attributes.
14042 (*arm_mulsi3_v6): Likewise.
14043 (*thumb_mulsi3): Likewise.
14044 (*thumb_mulsi3_v6): Likewise.
14045 (*mulsi3_compare0): Likewise.
14046 (*mulsi3_compare0_v6): Likewise.
14047 (*mulsi_compare0_scratch): Likewise.
14048 (*mulsi_compare0_scratch_v6): Likewise.
14049 (*mulsi3addsi): Likewise.
14050 (*mulsi3addsi_v6): Likewise.
14051 (*mulsi3addsi_compare0): Likewise.
14052 (*mulsi3addsi_compare0_v6): Likewise.
14053 (*mulsi3addsi_compare0_scratch): Likewise.
14054 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14055 (*mulsi3subsi): Likewise.
14056 (*mulsidi3adddi): Likewise.
14057 (*mulsi3addsi_v6): Likewise.
14058 (*mulsidi3adddi_v6): Likewise.
14059 (*mulsidi3_nov6): Likewise.
14060 (*mulsidi3_v6): Likewise.
14061 (*umulsidi3_nov6): Likewise.
14062 (*umulsidi3_v6): Likewise.
14063 (*umulsidi3adddi): Likewise.
14064 (*umulsidi3adddi_v6): Likewise.
14065 (*smulsi3_highpart_nov6): Likewise.
14066 (*smulsi3_highpart_v6): Likewise.
14067 (*umulsi3_highpart_nov6): Likewise.
14068 (*umulsi3_highpart_v6): Likewise.
14069 (mulhisi3): Likewise.
14070 (*mulhisi3tb): Likewise.
14071 (*mulhisi3bt): Likewise.
14072 (*mulhisi3tt): Likewise.
14073 (maddhisi4): Likewise.
14074 (*maddhisi4tb): Likewise.
14075 (*maddhisi4tt): Likewise.
14076 (maddhidi4): Likewise.
14077 (*maddhidi4tb): Likewise.
14078 (*maddhidi4tt): Likewise.
14079 (divsi3): Likewise.
14080 (udivsi3): Likewise.
14081 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14082 (thumb2_mulsi_short_compare0): Likewise.
14083 (thumb2_mulsi_short_compare0_scratch): Likewise.
14084 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14085 (1020mult2): Likewise.
14086 (1020mult3): Likewise.
14087 (1020mult4): Likewise.
14088 (1020mult5): Likewise.
14089 (1020mult6): Likewise.
14090 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14091 change.
14092 (cortex_a15_mult64): Likewise.
14093 (cortex_a15_sdiv): Likewise.
14094 (cortex_a15_udiv): Likewise.
14095 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14096 (mult2): Likewise.
14097 (mult3): Likewise.
14098 (mult4): Likewise.
14099 (mult5): Likewise.
14100 (mult6): Likewise.
14101 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14102 (pj4_ir_div): Likewise.
14103 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14104 (11_mult2): Likewise.
14105 (11_mult3): Likewise.
14106 (11_mult4): Likewise.
14107 (11_mult5): Likewise.
14108 (11_mult6): Likewise.
14109 (11_mult7): Likewise.
14110 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14111 (cortex_a8_mla): Likewise.
14112 (cortex_a8_mull): Likewise.
14113 (cortex_a8_smulwy): Likewise.
14114 (cortex_a8_smlald): Likewise.
14115 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14116 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14117 (cortex_r4_mul_3): Likewise.
14118 (cortex_r4_mla_4): Likewise.
14119 (cortex_r4_mla_3): Likewise.
14120 (cortex_r4_smlald): Likewise.
14121 (cortex_r4_mull): Likewise.
14122 (cortex_r4_sdiv): Likewise.
14123 (cortex_r4_udiv): Likewise.
14124 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14125 (cortex_a7_idiv): Likewise.
14126 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14127 (9_mult2): Likewise.
14128 (9_mult3): Likewise.
14129 (9_mult4): Likewise.
14130 (9_mult5): Likewise.
14131 (9_mult6): Likewise.
14132 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14133 (cortex_a53_sdiv): Likewise.
14134 (cortex_a53_udiv): Likewise.
14135 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14136 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14137 (mp626_mult2): Likewise.
14138 (mp626_mult3): Likewise.
14139 (mp626_mult4): Likewise.
14140 * config/arm/fa526.md (526_mult1): Update attribute change.
14141 (526_mult2): Likewise.
14142 * config/arm/arm-generic.md (mult): Update attribute change.
14143 (mult_ldsched_strongarm): Likewise.
14144 (mult_ldsched): Likewise.
14145 (multi_cycle): Likewise.
14146 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14147 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14148 (606te_mult2): Likewise.
14149 (606te_mult3): Likewise.
14150 (606te_mult4): Likewise.
14151 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14152 (cortex_a9_mac16): Likewise.
14153 (cortex_a9_multiply): Likewise.
14154 (cortex_a9_mac): Likewise.
14155 (cortex_a9_multiply_long): Likewise.
14156 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14157 (626te_mult2): Likewise.
14158 (626te_mult3): Likewise.
14159 (626te_mult4): Likewise.
14160
14161 2013-06-18 Richard Biener <rguenther@suse.de>
14162
14163 PR lto/57334
14164 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14165
14166 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14167
14168 PR target/57609
14169 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14170 with next_active_insn.
14171
14172 2013-06-18 Alan Modra <amodra@gmail.com>
14173
14174 * config/rs6000/rs6000.h (enum data_align): New.
14175 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14176 (DATA_ABI_ALIGNMENT): Define.
14177 (CONSTANT_ALIGNMENT): Correct comment.
14178 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14179 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14180
14181 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14182
14183 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14184 ATTRIBUTE_UNUSED marking.
14185
14186 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14187
14188 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14189 alternative and update.
14190 (aarch64_dup_lanedi): Delete.
14191 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14192 * config/aarch64/aarch64-simd-builtins.def: Update.
14193
14194 2013-06-17 Richard Biener <rguenther@suse.de>
14195
14196 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14197 (lto_input_scc): Declare.
14198 (lto_input_tree_1): Likewise.
14199 (struct lto_stats_d): Add num_tree_bodies_output and
14200 num_pickle_refs_output.
14201 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14202 (lto_read_tree_1): Split out from ...
14203 (lto_read_tree): ... this.
14204 (lto_input_scc): New function.
14205 (lto_input_tree_1): Split out from ...
14206 (lto_input_tree): ... this. Handle LTO_tree_scc.
14207 (lto_data_in_create): Create the streamer cache without hashes.
14208 * lto-streamer-out.c (create_output_block): Create the streamer
14209 cache with hashes when not doing WPA.
14210 (lto_write_tree_1): Split out from ...
14211 (lto_write_tree): ... this.
14212 (get_symbol_initial_value): New function.
14213 (lto_output_tree_1): Split out from ...
14214 (lto_output_tree): ... this. Write trees as series of SCCs
14215 using a DFS walk via DFS_write_tree.
14216 (struct sccs, struct scc_entry): New types.
14217 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14218 (DFS_write_tree_body): New function.
14219 (DFS_write_tree): Likewise.
14220 (hash_tree): Likewise.
14221 (scc_entry_compare): Likewise.
14222 (hash_scc): Likewise.
14223 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14224 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14225 TREE_CHAIN as regular reference.
14226 (streamer_read_integer_cst): Remove.
14227 (streamer_get_pickled_tree): Adjust.
14228 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14229 of DECL_EXTERNALs in BLOCK_VARS for now.
14230 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14231 reference.
14232 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14233 Add hash value argument and record that if hashes are recorded
14234 in the cache.
14235 (streamer_tree_cache_insert_1): Adjust.
14236 (streamer_tree_cache_insert): Likewise.
14237 (streamer_tree_cache_insert_at): Rename to ...
14238 (streamer_tree_cache_replace_tree): ... this and adjust.
14239 (streamer_tree_cache_append): Adjust.
14240 (record_common_node): Likewise.
14241 (streamer_tree_cache_create): Add argument whether to
14242 record hash values together with trees.
14243 (streamer_tree_cache_delete): Adjust.
14244 * tree-streamer.h (struct streamer_tree_cache_d): Add
14245 vector of hashes.
14246 (streamer_read_integer_cst): Remove.
14247 (streamer_tree_cache_insert): Adjust.
14248 (streamer_tree_cache_append): Likewise.
14249 (streamer_tree_cache_insert_at): Rename to ...
14250 (streamer_tree_cache_replace_tree): ... this and adjust.
14251 (streamer_tree_cache_create): Add argument whether to record hashes.
14252 (streamer_tree_cache_get): Rename to ...
14253 (streamer_tree_cache_get_tree): ... this.
14254 (streamer_tree_cache_get_hash): New function.
14255 * tree.c (cache_integer_cst): New function.
14256 * tree.h (cache_integer_cst): Declare.
14257 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14258 * lto-symtab.c (lto_varpool_replace_node): Only release
14259 DECL_INITIAL of non-prevailing decls.
14260 * varpool.c (varpool_remove_initializer): Do not release
14261 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14262
14263 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14264
14265 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14266 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14267 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14268 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14269 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14270 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14271 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14272 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14273 instead of TARGET_64BIT.
14274 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14275 Require ISA_HAS_<D>DIV.
14276
14277 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14278
14279 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14280 (mips*-*-linux*): Move default with_llsc setting to where other
14281 defaults are set.
14282 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14283 with_arch block.
14284 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14285 Likewise. Remove default with_tune setting. Move default float
14286 setting to its own block. Handle with_llsc in the same block as above.
14287
14288 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14289
14290 PR rtl-optimization/57425
14291 PR rtl-optimization/57569
14292 * alias.c (write_dependence_p): Add new parameters mem_mode,
14293 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14294 Changed all callers.
14295 (canon_anti_dependence): New function.
14296 * cse.c (check_dependence): Use canon_anti_dependence.
14297 * cselib.c (cselib_invalidate_mem): Likewise.
14298 * rtl.h (canon_anti_dependence): Declare.
14299
14300 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14301
14302 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14303 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14304 ".set mips3" for 64-bit targets.
14305
14306 2013-06-15 Dehao Chen <dehao@google.com>
14307
14308 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14309 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14310 (gimple_check_call_args): Likewise.
14311 * value-prof.c (check_ic_target): Likewise.
14312 * ipa-inline.c (early_inliner): Likewise.
14313 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14314 * cgraph.c (cgraph_create_edge_1): Likewise.
14315 (cgraph_make_edge_direct): Likewise.
14316
14317 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14318
14319 PR target/57615
14320 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14321 rs6000_output_move_128bit to handle emitting quad memory
14322 operations. Set attribute length to 8 bytes.
14323
14324 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14325
14326 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14327 New pattern.
14328 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14329 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14330 (aarch64_<su>mlsl<mode>): Likewise.
14331
14332 2013-06-14 Mike Stump <mikestump@comcast.net>
14333
14334 * Makefile.in (TARGET_H): Add insn-codes.h.
14335
14336 2013-06-14 Alan Modra <amodra@gmail.com>
14337
14338 PR middle-end/57134
14339 PR middle-end/57586
14340 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14341 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14342 bitfield expansion when EXPAND_MEMORY.
14343 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14344
14345 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14346
14347 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14348 test for clearing quad memory on 32-bit later.
14349
14350 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14351
14352 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14353 (fold_negate_expr): Likewise.
14354 (fold_real_zero_addition_p): Handle vectors.
14355 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14356
14357 2013-06-14 Alan Modra <amodra@gmail.com>
14358
14359 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14360
14361 2013-06-13 Jan Hubicka <jh@suse.cz>
14362
14363 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14364 Local comdats are not externally visible.
14365 * symtab.c (dump_symtab_base): Dump externally visible.
14366 (verify_symtab_base): Verify back links in the symtab hash.
14367
14368 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14369
14370 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14371 CONVERT_EXPR as equal nodes.
14372
14373 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14374
14375 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14376
14377 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14378
14379 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14380 Generalize to complex and vector.
14381 * tree.c (build_all_ones_cst): New function.
14382 * tree.h (build_all_ones_cst): Declare it.
14383
14384 2013-06-13 Alan Modra <amodra@gmail.com>
14385
14386 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14387 * config/rs6000/rs6000.md (signbittf2): New insn.
14388 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14389 (abstf2_internal, cmptf_internal2): Likewise.
14390 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14391
14392 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14393 Pat Haugen <pthaugen@us.ibm.com>
14394 Peter Bergner <bergner@vnet.ibm.com>
14395
14396 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14397 power8 byte, half-word, and quad-word atomic instructions.
14398 (emit_store_conditional): Likewise.
14399 (rs6000_expand_atomic_compare_and_swap): Likewise.
14400 (rs6000_expand_atomic_op): Likewise.
14401
14402 * config/rs6000/sync.md (larx): Add new modes for power8.
14403 (stcx): Likewise.
14404 (AINT): New mode iterator to include TImode as well as normal
14405 integer modes on power8.
14406 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14407 that VSX registers are not considered. Use AINT mode iterator
14408 instead of INT1 to allow inclusion of quad word atomic operations
14409 on power8.
14410 (load_locked<mode>): Likewise.
14411 (store_conditional<mode>): Likewise.
14412 (atomic_compare_and_swap<mode>): Likewise.
14413 (atomic_exchange<mode>): Likewise.
14414 (atomic_nand<mode>): Likewise.
14415 (atomic_fetch_<fetchop_name><mode>): Likewise.
14416 (atomic_nand_fetch<mode>): Likewise.
14417 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14418 each type.
14419 (ATOMIC): On power8, add QImode, HImode modes.
14420 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14421 modes that promote to SImode.
14422 (load_lockedti): Convert TImode arguments to PTImode, so that we
14423 get a guaranteed even/odd register pair.
14424 (load_lockedpti): Likewise.
14425 (store_conditionalti): Likewise.
14426 (store_conditionalpti): Likewise.
14427
14428 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14429 atomic load/store instructions.
14430 (HSI): Likewise.
14431
14432 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14433
14434 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14435 loads.
14436 (insn_count): New attribute, with most cases extracted from...
14437 (length): ...here. Redefine most cases in terms of insn_count.
14438 (single_insn): Delete.
14439 (can_delay): Use insn_count to check for single instructions.
14440 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14441 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14442 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14443 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14444 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14445 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14446 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14447 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14448 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14449 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14450 rather than "length".
14451 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14452 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14453 Use "insn_count" rather than "length".
14454 * config/mips/mips-dsp.md
14455 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14456 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14457 length attributes.
14458
14459 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14460
14461 PR tree-optimization/57361
14462 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14463
14464 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14465
14466 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14467 to split.
14468 (aarch64_simd_combine<mode>): New instruction expansion.
14469 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14470 function prototype.
14471 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14472 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14473
14474 2013-06-12 Jan Hubicka <jh@suse.cz>
14475
14476 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14477 decl has when in streaming stage.
14478 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14479 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14480
14481 2013-06-12 Roland Stigge <stigge@antcom.de>
14482
14483 PR target/57578
14484 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14485
14486 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14487
14488 PR tree-optimization/57537
14489 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14490 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14491
14492 2013-06-12 Richard Biener <rguenther@suse.de>
14493
14494 * data-streamer.h (streamer_write_char_stream): CSE
14495 obs->current_pointer.
14496 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14497 streamer_write_char_stream manually and optimize the resulting loop.
14498 (streamer_write_hwi_stream): Likewise.
14499
14500 2013-06-12 Jan Hubicka <jh@suse.cz>
14501
14502 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14503 * cgraph.h (varpool_create_empty_node): Declare.
14504 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14505 duplicated nodes.
14506 * symtab.c (symtab_unregister_node): Be lax about missin entries
14507 in node hash.
14508 (symtab_get_node): Update comment.
14509 * varpool.c (varpool_create_empty_node): Break out from ...
14510 (varpool_node_for_decl): ... here.
14511 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14512
14513 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14514
14515 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14516 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14517 part. Use straight-line flow at the end.
14518 <COMPONENT_REF>: Remove superfluous else.
14519 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14520
14521 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14522
14523 PR target/56564
14524 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14525 target hook even for !TREE_PUBLIC decls. If no resolution info
14526 is available, return false for common and external decls.
14527
14528 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14529
14530 * config/rl78/constraints.md (U): New constraint.
14531 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14532 valloc attribute.
14533
14534 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14535
14536 PR target/57589
14537 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14538 to allow returning address to AT_PLATFORM name.
14539
14540 2013-06-11 Jan Hubicka <jh@suse.cz>
14541
14542 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14543 * cgraph.h (symtab_node_base): Add weakref flag.
14544 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14545 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14546 (output_weakrefs): Use weakref flag.
14547 * fold-const.c (simple_operand_p): Handle WEAK.
14548 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14549 * ipa.c (varpool_externally_visible_p): Drop weakref.
14550 (function_and_variable_visibility): Update comment; fix weakref
14551 sanity checks; do not clear DECL_WEAK on them.
14552 * lto-cgraph.c (lto_output_node): update.
14553 (lto_output_varpool_node): Update.
14554 (input_overwrite_node): Update.
14555 (input_node): Update.
14556 (input_varpool_node): Update.
14557 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14558 (lto_symtab_merge_symbols): Add sanity check.
14559 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14560 * passes.c (rest_of_decl_compilation): Set static flag, too.
14561 * symtab.c (dump_symtab_base): Dump weakref.
14562 (verify_symtab_base): Sanity check weakrefs.
14563 (symtab_make_decl_local): Remove duplicated code.
14564 (symtab_alias_ultimate_target): Simplify.
14565 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14566
14567 2013-06-11 Tom de Vries <tom@codesourcery.com>
14568
14569 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14570 sequence_vect == NULL.
14571
14572 2013-06-11 DJ Delorie <dj@redhat.com>
14573
14574 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14575 (rl78_unwind_word_mode): New.
14576
14577 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14578
14579 * final.c (debug_prefix_maps): Make static.
14580
14581 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14582
14583 * function.c (initial_trampoline): Remove stray copy.
14584
14585 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14586
14587 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14588
14589 2013-06-11 Martin Jambor <mjambor@suse.cz>
14590
14591 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14592 within bounds at the beginning of the function.
14593
14594 2013-06-11 Alan Modra <amodra@gmail.com>
14595
14596 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14597 reporting.
14598 (get_named_section): Don't NULL !DECL_P decl.
14599
14600 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14601
14602 * doc/invoke.texi (core-avx2): Document.
14603 (slm): Likewise.
14604 (atom): Updated with MOVBE.
14605
14606 2013-06-11 Richard Biener <rguenther@suse.de>
14607
14608 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14609
14610 2013-06-11 Anton Blanchard <anton@samba.org>
14611
14612 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14613 correct shift value in little-endian mode.
14614
14615 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14616
14617 PR target/56564
14618 * varasm.c (get_variable_align): Move #endif to the right place.
14619
14620 2013-06-10 Cary Coutant <ccoutant@google.com>
14621
14622 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14623 for hash so that hash table traversal order is deterministic.
14624
14625 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14626 Pat Haugen <pthaugen@us.ibm.com>
14627 Peter Bergner <bergner@vnet.ibm.com>
14628
14629 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14630 of vectors in GPRS if they are direct moves or quad word load or
14631 store moves.
14632
14633 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14634 declaration.
14635 (direct_move_p): Likewise.
14636 (quad_load_store_p): Likewise.
14637
14638 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14639 classes into bins based on the physical register type.
14640 (reg_class_to_reg_type): Likewise.
14641 (IS_STD_REG_TYPE): Likewise.
14642 (IS_FP_VECT_REG_TYPE): Likewise.
14643 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14644 use direct move instructions.
14645 (reload_gpr_vsx): Likewise.
14646 (reload_vsx_gpr): Likewise.
14647 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14648 information that is a simplification of register classes. Also
14649 precalculate direct move reload helpers.
14650 (direct_move_p): New function to return true if the operation can
14651 be done as a direct move instruciton.
14652 (quad_load_store_p): New function to return true if the operation
14653 is a quad memory operation.
14654 (rs6000_legitimize_address): If quad memory, only allow register
14655 indirect for TImode addresses.
14656 (rs6000_legitimate_address_p): Likewise.
14657 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14658 (rs6000_reload_register_type): Likewise.
14659 (register_to_reg_type): Return register type.
14660 (rs6000_secondary_reload_simple_move): New helper function for
14661 secondary reload and secondary memory needed to identify anything
14662 that is a simple move, and does not need reloading.
14663 (rs6000_secondary_reload_direct_move): New helper function for
14664 secondary reload to identify cases that can be done with several
14665 instructions via the direct move instructions.
14666 (rs6000_secondary_reload_move): New helper function for secondary
14667 reload to identify moves between register types that can be done.
14668 (rs6000_secondary_reload): Add support for quad memory operations
14669 and for direct move.
14670 (rs6000_secondary_memory_needed): Likewise.
14671 (rs6000_debug_secondary_memory_needed): Change argument names.
14672 (rs6000_output_move_128bit): New function to return the move to
14673 use for 128-bit moves, including knowing about the various
14674 limitations of quad memory operations.
14675
14676 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14677 memory operations. call rs6000_output_move_128bit for the actual
14678 instruciton(s) to generate.
14679 (vsx_movti_64bit): Likewise.
14680
14681 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14682 (UNSPEC_P8V_MTVSRWZ): Likewise.
14683 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14684 (UNSPEC_P8V_MTVSRD): Likewise.
14685 (UNSPEC_P8V_XXPERMDI): Likewise.
14686 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14687 (UNSPEC_FUSION_GPR): Likewise.
14688 (FMOVE128_GPR): New iterator for direct move.
14689 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14690 (f32_sv): Likewise.
14691 (f32_dm): Likewise.
14692 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14693 loads and direct move instructions.
14694 (zero_extendsidi2_lfiwzx): Likewise.
14695 (extendsidi2_lfiwax): Likewise.
14696 (extendsidi2_nocell): Likewise.
14697 (floatsi<mode>2_lfiwax): Likewise.
14698 (lfiwax): Likewise.
14699 (floatunssi<mode>2_lfiwzx): Likewise.
14700 (lfiwzx): Likewise.
14701 (fix_trunc<mode>_stfiwx): Likewise.
14702 (fixuns_trunc<mode>_stfiwx): Likewise.
14703 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14704 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14705 (parity<mode>2_cmpb): Set length/type attr.
14706 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14707 for 'mr.' to fast_compare.
14708 (bpermd_<mode>): Change type attr to popcnt.
14709 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14710 (p8_mtvsrwz_1): Likewise.
14711 (p8_mtvsrwz_2): Likewise.
14712 (reload_fpr_from_gpr<mode>): Likewise.
14713 (p8_mtvsrd_1): Likewise.
14714 (p8_mtvsrd_2): Likewise.
14715 (p8_xxpermdi_<mode>): Likewise.
14716 (reload_vsx_from_gpr<mode>): Likewise.
14717 (reload_vsx_from_gprsf): Likewise.
14718 (p8_mfvsrd_3_<mode>): LIkewise.
14719 (reload_gpr_from_vsx<mode>): Likewise.
14720 (reload_gpr_from_vsxsf): Likewise.
14721 (p8_mfvsrd_4_disf): Likewise.
14722 (multi-word GPR splits): Do not split direct moves or quad memory
14723 operations.
14724
14725 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14726
14727 * tree-into-ssa.c (interesting_blocks): Make static.
14728
14729 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14730
14731 PR target/56564
14732 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14733 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14734 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14735 (get_variable_align): New function.
14736 (get_variable_section, emit_bss, emit_common,
14737 assemble_variable_contents, place_block_symbol): Use
14738 get_variable_align instead of DECL_ALIGN.
14739 (assemble_noswitch_variable): Add align argument, use it
14740 instead of DECL_ALIGN.
14741 (assemble_variable): Adjust caller. Use get_variable_align
14742 instead of DECL_ALIGN.
14743 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14744 caller.
14745 (DATA_ABI_ALIGNMENT): Define.
14746 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14747 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14748 opt is false, only return the psABI mandated alignment increase.
14749 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14750 (DATA_ABI_ALIGNMENT): ... this.
14751 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14752 (DATA_ABI_ALIGNMENT): ... this.
14753 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14754 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14755 (DATA_ABI_ALIGNMENT): ... this.
14756 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14757 * doc/tm.texi: Regenerated.
14758
14759 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14760
14761 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14762 cmp_code to construct REG_EQUAL note.
14763
14764 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14765
14766 PR target/57568
14767 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14768 that operands[2] doesn't overlap with operands[0].
14769
14770 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14771 Jan Hubicka <jh@suse.cz>
14772
14773 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14774 hack to mark symbols as used.
14775
14776 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14777
14778 PR rtl-optimization/57559
14779 * lra-constraints.c (process_alt_operands): Don't discourage
14780 memory with known offset for offsetable memory constraint.
14781 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14782
14783 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14784
14785 * varasm.c (struct oc_local_state): Reorder fields.
14786 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14787 and adjust accordingly.
14788 (output_constructor): Reorder initialization code and adjust call to
14789 output_constructor_bitfield.
14790
14791 2013-06-07 Jan Hubicka <jh@suse.cz>
14792
14793 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14794
14795 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14796
14797 * tree-object-size.c (unknown): Make const.
14798
14799 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14800
14801 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14802 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14803 for last alternative in the cpu_facility attribute.
14804
14805 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14806
14807 PR target/56315
14808 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14809 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14810 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14811 * config/arm/constraints.md (Dg): New constraint.
14812 * config/arm/neon.md (xordi3_neon): Remove.
14813 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14814 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14815
14816 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14817
14818 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14819 Clean up alternatives.
14820
14821 2013-06-07 Alan Modra <amodra@gmail.com>
14822
14823 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14824 va_list_gpr_size.
14825
14826 2013-06-07 Alan Modra <amodra@gmail.com>
14827
14828 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14829
14830 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14831
14832 * config/arm/constraints.md (Df): New constraint.
14833 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14834 Correct length attribute for last two alternatives.
14835
14836 2013-06-07 Alan Modra <amodra@gmail.com>
14837
14838 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14839 override user -mfp-in-toc.
14840 (offsettable_ok_by_alignment): Consider just the current access
14841 rather than the whole object, unless BLKmode. Handle
14842 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14843 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14844 for -mcmodel=medium.
14845 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14846 override user -mfp-in-toc or -msum-in-toc. Default to
14847 -mno-fp-in-toc for -mcmodel=medium.
14848
14849 2013-06-06 DJ Delorie <dj@redhat.com>
14850
14851 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14852 TARGET_VALID_POINTER_MODE.
14853
14854 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14855 Pat Haugen <pthaugen@us.ibm.com>
14856 Peter Bergner <bergner@vnet.ibm.com>
14857
14858 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14859 Document new power8 builtins.
14860
14861 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14862 condition code register, to allow 128-bit logical operations to be
14863 done in the VSX or GPR registers.
14864 (nor<mode>3): Use the canonical form for nor.
14865 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14866 vclz*, and vpopcnt* vector instructions.
14867 (nand<mode>3): Likewise.
14868 (orc<mode>3): Likewise.
14869 (clz<mode>2): LIkewise.
14870 (popcount<mode>2): Likewise.
14871
14872 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14873 that only the GPRs are recognized.
14874
14875 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14876 support for new power8 builtins.
14877
14878 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14879 builtin functions.
14880 (xscvdpspn): Likewise.
14881 (vclz): Likewise.
14882 (vclzb): Likewise.
14883 (vclzh): Likewise.
14884 (vclzw): Likewise.
14885 (vclzd): Likewise.
14886 (vpopcnt): Likewise.
14887 (vpopcntb): Likewise.
14888 (vpopcnth): Likewise.
14889 (vpopcntw): Likewise.
14890 (vpopcntd): Likewise.
14891 (vgbbd): Likewise.
14892 (vmrgew): Likewise.
14893 (vmrgow): Likewise.
14894 (eqv): Likewise.
14895 (eqv_v16qi3): Likewise.
14896 (eqv_v8hi3): Likewise.
14897 (eqv_v4si3): Likewise.
14898 (eqv_v2di3): Likewise.
14899 (eqv_v4sf3): Likewise.
14900 (eqv_v2df3): Likewise.
14901 (nand): Likewise.
14902 (nand_v16qi3): Likewise.
14903 (nand_v8hi3): Likewise.
14904 (nand_v4si3): Likewise.
14905 (nand_v2di3): Likewise.
14906 (nand_v4sf3): Likewise.
14907 (nand_v2df3): Likewise.
14908 (orc): Likewise.
14909 (orc_v16qi3): Likewise.
14910 (orc_v8hi3): Likewise.
14911 (orc_v4si3): Likewise.
14912 (orc_v2di3): Likewise.
14913 (orc_v4sf3): Likewise.
14914 (orc_v2df3): Likewise.
14915
14916 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14917 allow power8 quad mode in 64-bit.
14918 (rs6000_builtin_vectorized_function): Add support to vectorize
14919 ISA 2.07 count leading zeros, population count builtins.
14920 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14921 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14922 (builtin_function_type): Add vgbbd builtin function which takes an
14923 unsigned argument.
14924 (altivec_expand_vec_perm_const): Add support for new power8 merge
14925 instructions.
14926
14927 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14928 that does not include TImdoe for use with 32-bit.
14929 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14930 instructions.
14931 (UNSPEC_VSX_CVDPSPN): Likewise.
14932 (vsx_xscvdpspn): Likewise.
14933 (vsx_xscvspdpn): Likewise.
14934 (vsx_xscvdpspn_scalar): Likewise.
14935 (vsx_xscvspdpn_directmove): Likewise.
14936 (vsx_and<mode>3): Split logical operations into 32-bit and
14937 64-bit. Add support to do logical operations on TImode as well as
14938 VSX vector types. Allow logical operations to be done in either
14939 VSX registers or in general purpose registers in 64-bit mode. Add
14940 splitters if GPRs were used. For AND, add clobber of CCmode to
14941 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14942 encoding.
14943 (vsx_and<mode>3_32bit): Likewise.
14944 (vsx_and<mode>3_64bit): Likewise.
14945 (vsx_ior<mode>3): Likewise.
14946 (vsx_ior<mode>3_32bit): Likewise.
14947 (vsx_ior<mode>3_64bit): Likewise.
14948 (vsx_xor<mode>3): Likewise.
14949 (vsx_xor<mode>3_32bit): Likewise.
14950 (vsx_xor<mode>3_64bit): Likewise.
14951 (vsx_one_cmpl<mode>2): Likewise.
14952 (vsx_one_cmpl<mode>2_32bit): Likewise.
14953 (vsx_one_cmpl<mode>2_64bit): Likewise.
14954 (vsx_nor<mode>3): Likewise.
14955 (vsx_nor<mode>3_32bit): Likewise.
14956 (vsx_nor<mode>3_64bit): Likewise.
14957 (vsx_andc<mode>3): Likewise.
14958 (vsx_andc<mode>3_32bit): Likewise.
14959 (vsx_andc<mode>3_64bit): Likewise.
14960 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14961 and xxlorc instructions.
14962 (vsx_eqv<mode>3_64bit): Likewise.
14963 (vsx_nand<mode>3_32bit): Likewise.
14964 (vsx_nand<mode>3_64bit): Likewise.
14965 (vsx_orc<mode>3_32bit): Likewise.
14966 (vsx_orc<mode>3_64bit): Likewise.
14967
14968 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14969
14970 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14971 instruction.
14972 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14973 (p8_vmrgow): Likewise.
14974 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14975 GPRs to be split under VSX.
14976 (p8v_clz<mode>2): Add power8 count leading zero support.
14977 (p8v_popcount<mode>2): Add power8 population count support.
14978 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14979 support.
14980
14981 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14982 instruction.
14983
14984 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14985 builtin functions.
14986 (vec_nand): Likewise.
14987 (vec_vclz): Likewise.
14988 (vec_vclzb): Likewise.
14989 (vec_vclzd): Likewise.
14990 (vec_vclzh): Likewise.
14991 (vec_vclzw): Likewise.
14992 (vec_vgbbd): Likewise.
14993 (vec_vmrgew): Likewise.
14994 (vec_vmrgow): Likewise.
14995 (vec_vpopcnt): Likewise.
14996 (vec_vpopcntb): Likewise.
14997 (vec_vpopcntd): Likewise.
14998 (vec_vpopcnth): Likewise.
14999 (vec_vpopcntw): Likewise.
15000
15001 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15002
15003 PR rtl-optimization/57468
15004 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15005 spilled pseudos.
15006
15007 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15008
15009 PR rtl-optimization/57459
15010 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15011 type when setting live regs.
15012
15013 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15014
15015 * config/s390/s390.opt (mlra): New option.
15016 * config/s390/s390.c (s390_decompose_address): Check displacement
15017 for all registers for LRA.
15018 (s390_secondary_reload): Don't used secondary reloads for LRA.
15019 (s390_lra_p): New function.
15020 (TARGET_LRA_P): Define.
15021 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15022 of attribute cpu_facility to zarch for the last alternative.
15023 (*cmpmem_short): Ditto.
15024
15025 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15026
15027 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15028 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15029 (arm_expand_prologue): Likewise.
15030
15031 2013-06-06 Teresa Johnson <tejohnson@google.com>
15032
15033 PR c++/53743
15034 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15035 as this is now done by redirect_edge_and_branch_force.
15036 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15037 barriers, and fix interaction with splitting.
15038 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15039 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15040 reflect changes made in the routine.
15041 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15042 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15043 since this is called in cfglayout mode, and replace partition fixup
15044 with assert as that is now done by force_nonfallthru_and_redirect.
15045 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15046 already be marked with region crossing note.
15047 (insert_section_boundary_note): Make non-static, gate on flag
15048 has_bb_partition, rewrite to also check for multiple partitions.
15049 (rest_of_handle_reorder_blocks): Remove call to
15050 insert_section_boundary_note, now done later during free_cfg.
15051 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15052 * bb-reorder.h (insert_section_boundary_note): Declare.
15053 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15054 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15055 invoke insert_section_boundary_note.
15056 (try_redirect_by_replacing_jump): Remove unnecessary
15057 check for region crossing note.
15058 (fixup_partition_crossing): New function.
15059 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15060 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15061 in non-cfglayout mode.
15062 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15063 remove old code that tried to do this. Emit barrier correctly
15064 when we are in cfglayout mode.
15065 (last_bb_in_partition): New function.
15066 (rtl_split_edge): Correctly fixup partition boundaries.
15067 (commit_one_edge_insertion): Remove old code that tried to
15068 fixup region crossing edge since this is now handled in
15069 split_block, and set up insertion point correctly since
15070 block may now end in a jump.
15071 (verify_hot_cold_block_grouping): Guard against checking when not in
15072 linearized RTL mode.
15073 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15074 notes.
15075 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15076 rtl_verify_flow_info, so not called in cfglayout mode.
15077 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15078 (fixup_reorder_chain): Remove old code that attempted to fixup region
15079 crossing note as this is now handled in force_nonfallthru_and_redirect.
15080 (duplicate_insn_chain): Don't duplicate switch section notes.
15081 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15082 note.
15083 * basic-block.h (emit_barrier_after_bb): Declare.
15084
15085 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15086
15087 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15088 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15089 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15090
15091 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15092
15093 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15094 where appropriate.
15095 * config/arm/ldmstm.md: Regenerate.
15096
15097 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15098
15099 * config/arm/sync.md (atomic_loaddi_1):
15100 Disable predication for arm_restrict_it.
15101 (arm_load_exclusive<mode>): Likewise.
15102 (arm_load_exclusivesi): Likewise.
15103 (arm_load_exclusivedi): Likewise.
15104 (arm_load_acquire_exclusive<mode>): Likewise.
15105 (arm_load_acquire_exclusivesi): Likewise.
15106 (arm_load_acquire_exclusivedi): Likewise.
15107 (arm_store_exclusive<mode>): Likewise.
15108 (arm_store_exclusive<mode>): Likewise.
15109 (arm_store_release_exclusivedi): Likewise.
15110 (arm_store_release_exclusive<mode>): Likewise.
15111
15112 2013-06-06 Richard Biener <rguenther@suse.de>
15113
15114 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15115 after LTO_null.
15116 (lto_tag_is_tree_code_p): Adjust.
15117 (lto_tag_is_gimple_code_p): Likewise.
15118 (lto_gimple_code_to_tag): Likewise.
15119 (lto_tag_to_gimple_code): Likewise.
15120 (lto_tree_code_to_tag): Likewise.
15121 (lto_tag_to_tree_code): Likewise.
15122 * data-streamer.h (streamer_write_hwi_in_range): Use
15123 uhwi streaming to stream the normalized range.
15124 (streamer_read_hwi_in_range): Likewise.
15125
15126 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15127
15128 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15129 (predicable_short_it): Likewise.
15130 (predicated): Likewise.
15131 (enabled): Handle above.
15132 (define_cond_exec): Set predicated attribute to yes.
15133
15134 2013-06-05 Mike Stump <mikestump@comcast.net>
15135
15136 * gdbinit.in (__FUNCTION__): Add.
15137
15138 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15139
15140 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15141 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15142
15143 2013-06-05 Jan Hubicka <jh@suse.cz>
15144
15145 * varasm.c (mark_decl_referenced): Revert the removal until targets
15146 are fixed.
15147
15148 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15149
15150 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15151 instead of mark_decl_referenced.
15152
15153 2013-06-05 Jan Hubicka <jh@suse.cz>
15154
15155 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15156 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15157 and symtab_used_from_object_file_p.
15158 (cgraph_make_node_local_1): Clear forced_by_abi.
15159 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15160 * cgraph.h (symtab_node_base): Add forced_by_abi.
15161 (decide_is_variable_needed): Remove.
15162 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15163 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15164 (decide_is_symbol_needed): ... this one; handle symbols in general;
15165 always analyze virtuals; honnor forced_by_abi.
15166 (cgraph_finalize_function): Update.
15167 (varpool_finalize_decl): Update.
15168 (symbol_defined_and_needed): Remove.
15169 (analyze_functions): Update.
15170 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15171 output_refs, input_overwrite_node): Handle forced_by_abi.
15172 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15173 (address_taken_from_non_vtable_p): ... this one.
15174 (comdat_can_be_unshared_p_1): New function.
15175 (cgraph_comdat_can_be_unshared_p): Rename to ...
15176 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15177 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15178 (function_and_variable_visibility): Clear forced_by_abi as needed.
15179 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15180 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15181 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15182 * varpool.c (decide_is_variable_needed): Remove.
15183
15184 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15185
15186 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15187 (arm_option_override): Override arm_restrict_it where appropriate.
15188 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15189 * config/arm/arm.opt (mrestrict-it): New command-line option.
15190 * doc/invoke.texi: Document -mrestrict-it.
15191
15192 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15193
15194 * tsan.c (tsan_atomic_table): Make const.
15195
15196 2013-06-05 Richard Biener <rguenther@suse.de>
15197
15198 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15199 index associated with the tree we are supposed to replace.
15200 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15201 TREE_ASM_WRITTEN as zero for everything but SSA names.
15202
15203 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15204
15205 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15206
15207 2013-06-04 Jan Hubicka <jh@suse.cz>
15208
15209 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15210 (input_node, input_varpool_node): Handle correctly external same
15211 body aliases.
15212 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15213 nodes at ltrans stage.
15214
15215 2013-06-04 Jan Hubicka <jh@suse.cz>
15216
15217 * ipa-inline.c (update_caller_keys): Fix availability test.
15218 (update_callee_keys): Likewise.
15219 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15220 to follow ELF standard.
15221
15222 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15223
15224 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15225 (mips64r5900el-*-elf*): New configurations.
15226 * config/mips/mips-cpus.def (r5900): New processor.
15227 * config/mips/mips-tables.opt: Regenerate.
15228 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15229 (mips_issue_rate): Handle PROCESSOR_R5900.
15230 (mips_reorg_process_insns): Force reorder mode for the R5900.
15231 * config/mips/mips.h (TARGET_MIPS5900): Define.
15232 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15233 TARGET_MIPS5900.
15234 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15235 TARGET_MIPS5900.
15236 * config/mips/mips.md (processor): Add r5900.
15237 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15238
15239 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15240
15241 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15242 into function to generate MOVI instruction.
15243 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15244 (aarch64_preferred_simd_mode): Turn into wrapper.
15245 (aarch64_output_scalar_simd_mov_immediate): New function.
15246 * config/aarch64/aarch64-protos.h: Add prototype for above.
15247
15248 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15249
15250 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15251 element_char member.
15252 (sizetochar): Return signed char.
15253 (aarch64_simd_valid_immediate): Remove elchar and other
15254 unnecessary variables.
15255 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15256 Calculate element_char as required.
15257 * config/aarch64/aarch64-protos.h: Update and move prototype
15258 for aarch64_output_simd_mov_immediate.
15259 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15260 Update arguments.
15261
15262 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15263
15264 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15265 information completed by aarch64_simd_valid_immediate.
15266 (aarch64_legitimate_constant_p): Update arguments.
15267 (aarch64_simd_valid_immediate): Work with struct rather than many
15268 pointers.
15269 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15270 (aarch64_simd_make_constant): Update arguments.
15271 (aarch64_output_simd_mov_immediate): Work with struct rather than
15272 many pointers. Output immediate directly rather than as operand.
15273 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15274 Update prototype.
15275 * config/aarch64/constraints.md (Dn): Update arguments.
15276
15277 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15278
15279 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15280 longer static.
15281 (aarch64_simd_immediate_valid_for_move): Remove.
15282 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15283 (aarch64_simd_make_constant): Update call.
15284 (aarch64_output_simd_mov_immediate): Update call.
15285 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15286 Add prototype.
15287 * config/aarch64/constraints.md (Dn): Update call.
15288
15289 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15290
15291 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15292 return type to bool for prototype.
15293 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15294 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15295 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15296
15297 2013-06-04 Catherine Moore <clm@codesourcery.com>
15298
15299 * config/mips/mips.opt (meva): New.
15300 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15301 (ASM_SPEC): Handle -meva.
15302 * doc/invoke.texi (meva): Document.
15303
15304 2013-06-04 Alan Modra <amodra@gmail.com>
15305
15306 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15307 constant output.
15308
15309 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15310
15311 * rtl.def: Add extra fourth optional field to define_cond_exec.
15312 * gensupport.c (process_one_cond_exec): Process attributes from
15313 define_cond_exec.
15314 * doc/md.texi: Document fourth field in define_cond_exec.
15315
15316 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15317
15318 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15319 out the processing order as in store_bit_field_1.
15320
15321 2013-06-04 Jan Hubicka <jh@suse.cz>
15322
15323 PR middle-end/57500
15324 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15325 non-VAR_DECL node if it does not exist yet.
15326
15327 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15328
15329 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15330 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15331 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15332 target_cpu_default setting.
15333
15334 2013-06-03 Teresa Johnson <tejohnson@google.com>
15335
15336 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15337 default to -fopt-info=optimized instead of all.
15338 * doc/invoke.texi: Ditto.
15339 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15340 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15341 (execute_vect_slp): Emit BB vectorization success under
15342 MSG_OPTIMIZED_LOCATIONS.
15343 * tree-vect-slp.c (vect_slp_transform_bb): Change
15344 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15345 * tree-vect-loop.c (vect_transform_loop): Ditto.
15346
15347 2013-06-03 Jason Merrill <jason@redhat.com>
15348
15349 PR c++/57415
15350 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15351 Use TARGET_EXPR for C++.
15352
15353 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15354
15355 PR rtl-optimization/57268
15356 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15357 if DEBUG_INSN_P (insn).
15358
15359 Reapply
15360 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15361
15362 PR rtl-optimization/57268
15363 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15364 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15365
15366 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15367
15368 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15369 (ix86_avoid_lea_for_addr): Likewise.
15370 (exact_dependency_1): Likewise.
15371 (ix86_adjust_cost): Likewise.
15372 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15373 removed.
15374 (do_reorder_for_imul): Fix typo, formatting and
15375 !reload_completed check removed.
15376 (ix86_sched_reorder): Fix typo and formatting.
15377 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15378 list.
15379
15380 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15381
15382 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15383
15384 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15385
15386 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15387 <STRING_CST>: Likewise.
15388 <VECTOR_CST>: Likewise.
15389
15390 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15391 Mikael Morin <mikael@gcc.gnu.org>
15392
15393 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15394 * config.in: Regenerated.
15395 * configure: Regenerated.
15396
15397 2013-06-01 Jan Hubicka <jh@suse.cz>
15398
15399 PR middle-end/57366
15400 * cgraphunit.c (compile): When weakref is not supported,
15401 set up transparent aliases before final output pass.
15402 * varasm.c (assemble_alias): Do not try to do it here.
15403
15404 2013-06-01 Jan Hubicka <jh@suse.cz>
15405
15406 PR middle-end/57467
15407 * passes.c (for_per_function): Skip unanalyzed functions.
15408
15409 2013-06-01 Jan Hubicka <jh@suse.cz>
15410
15411 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15412 (lto_symtab_merge_symbols_1): ... this one.
15413 (lto_symtab_merge_cgraph_nodes): Rename to ...
15414 (lto_symtab_merge_symbols): ... this one; simplify.
15415 * cgraph.c (same_body_aliases_done): Rename to ...
15416 (cpp_implicit_aliases_done): ... this one.
15417 (cgraph_create_function_alias): Update.
15418 (cgraph_same_body_alias): Update.
15419 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15420 (verify_edge_corresponds_to_fndecl): Simplify.
15421 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15422 (cgraph_node): Remove same_body_alias.
15423 (varpool_node): Remove alias_of and extra_name_alias.
15424 (same_body_aliases_done): Rename to ..
15425 (cpp_implicit_aliases_done): ... this one.
15426 (symtab_alias_ultimate_target): Add default parameter.
15427 (symtab_resolve_alias): New function.
15428 (fixup_same_cpp_alias_visibility): Declare.
15429 (cgraph_function_node): Add default parameter.
15430 (cgraph_node_asm_name): Likewise.
15431 (cgraph_function_or_thunk_node): Add default parameter; do
15432 not ICE when it is NULL.
15433 (varpool_variable_node): Likewise.
15434 * tree-emutls.c (create_emultls_var): Update.
15435 (ipa_lower_emutls): Update.
15436 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15437 (cgraph_reset_node): Reset alias info.
15438 (cgraph_finalize_function): Update.
15439 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15440 (analyze_function): Simplify.
15441 (cgraph_process_same_body_aliases): Simplify.
15442 (analyze_functions): Fixup same body aliases.
15443 (handle_alias_pairs): Simplify.
15444 (assemble_thunk): Update.
15445 (assemble_thunks_and_aliases): Update.
15446 (output_weakrefs): Rewrite.
15447 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15448 (lto_output_varpool_node): Likewise.
15449 (compute_ltrans_boundary): Remve assert.
15450 (get_alias_symbol): New functoin.
15451 (input_node): Rewrite alias handling.
15452 (input_varpool_node): Likewise.
15453 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15454 * ipa.c (process_references): Handle weakrefs correctly.
15455 (symtab_remove_unreachable_nodes): Likewise.
15456 * trans-mem.c (get_cg_data): Update.
15457 (ipa_tm_create_version_alias): Update.
15458 (ipa_tm_execute): Update.
15459 * symtab.c (dump_symtab_base): Dump aliases.
15460 (verify_symtab_base): Verify aliases.
15461 (symtab_node_availability): New function.
15462 (symtab_alias_ultimate_target): Simplify.
15463 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15464 handle all the fixup cases.
15465 (symtab_resolve_alias): New function.
15466 * passes.c (ipa_write_summaries): Handle weakrefs.
15467 * varpool.c (varpool_analyze_node): Simplify.
15468 (assemble_aliases): Update.
15469 (varpool_create_variable_alias): Simplify.
15470 (varpool_extra_name_alias): Simplify.
15471 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15472 (lto_symtab_merge_symbols): ... this one.
15473
15474 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15475
15476 Revert
15477 PR rtl-optimization/57268
15478 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15479 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15480
15481 2013-06-01 Tobias Burnus <burnus@net-b.de>
15482
15483 Partially reverted:
15484 2013-05-31 Tobias Burnus <burnus@net-b.de>
15485
15486 PR middle-end/57073
15487 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15488 further up.
15489
15490 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15491
15492 PR rtl-optimization/57268
15493 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15494 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15495
15496 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15497
15498 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15499 unordered comparison operators when -fno-trapping-math is in effect
15500 on the e500.
15501 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15502 and implement unordered comparison operators properly on the e500.
15503
15504 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15505
15506 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15507 for constant scalar integers.
15508 (simplify_relational_operation_1): Likewise.
15509
15510 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15511
15512 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15513 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15514 Fix comment.
15515
15516 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15517 Igor Zamyatin <igor.zamyatin@intel.com>
15518
15519 Silvermont (SLM) architecture performance tuning.
15520 * config/i386/i386.h (enum ix86_tune_indices): Add
15521 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15522 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15523
15524 * config/i386/i386.c (initial_ix86_tune_features)
15525 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15526 (ix86_lea_outperforms): Handle Silvermont tuning.
15527 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15528 call.
15529 (ix86_use_lea_for_mov): Likewise.
15530 (ix86_avoid_lea_for_addr): Likewise.
15531 (ix86_lea_for_add_ok): Likewise.
15532 (exact_dependency_1): New function.
15533 (exact_store_load_dependency): Likewise.
15534 (ix86_adjust_cost): Handle Silvermont tuning.
15535 (do_reoder_for_imul): Likewise.
15536 (swap_top_of_ready_list): New function.
15537 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15538
15539 * config/i386/i386.md (peepholes that split memory operand in fp
15540 converts): New.
15541
15542 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15543
15544 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15545 Remove un-necessary braces.
15546
15547 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15548
15549 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15550 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15551
15552 2013-05-31 Tobias Burnus <burnus@net-b.de>
15553
15554 PR middle-end/57073
15555 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15556
15557 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15558
15559 PR target/56315
15560 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15561 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15562 * config/arm/neon.md (iordi3_neon): Remove.
15563 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15564 * config/arm/predicates.md (imm_for_neon_logic_operand):
15565 Move to earlier in the file.
15566 (neon_logic_op2): Likewise.
15567 (arm_iordi_operand_neon): New predicate.
15568
15569 2013-05-31 Richard Biener <rguenther@suse.de>
15570
15571 PR tree-optimization/57478
15572 PR tree-optimization/57453
15573 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15574 are life as well.
15575
15576 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15577
15578 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15579 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15580
15581 2013-05-30 Tobias Burnus <burnus@net-b.de>
15582 Thomas Koenig <tkoenig@gcc.gnu.org>
15583
15584 PR middle-end/57073
15585 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15586 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15587
15588 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15589
15590 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15591
15592 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15593
15594 * target.def (register_usage_leveling_p): New hook.
15595 * targhooks.c (default_register_usage_leveling_p): New.
15596 * targhooks.h (default_register_usage_leveling_p): New prototype.
15597 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15598 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15599 * doc/tm.texi: Update.
15600 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15601
15602 2013-05-30 Ian Bolton <ian.bolton@arm.com>
15603
15604 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15605 (*insv_reg<mode>): New define_insn.
15606
15607 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15608
15609 PR rtl-optimization/57439
15610 * postreload.c (move2add_valid_value_p): Check that we have
15611 a zero subreg_regno_offset when accessing the register in
15612 the requested mode.
15613
15614 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15615 Igor Zamyatin <igor.zamyatin@intel.com>
15616
15617 Silvermont (SLM) architecture pipeline model, tuning and
15618 insn selection.
15619 * config.gcc: Add slm config options and target.
15620
15621 * config/i386/slm.md: New.
15622
15623 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15624
15625 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15626 PROCESSOR_SLM.
15627 (ix86_target_macros_internal): Likewise.
15628
15629 * config/i386/i386.c (slm_cost): New cost.
15630 (m_SLM): New macro flag.
15631 (initial_ix86_tune_features): Set m_SLM.
15632 (x86_accumulate_outgoing_args): Likewise.
15633 (x86_arch_always_fancy_math_387): Likewise.
15634 (processor_target_table): Add slm cost.
15635 (cpu_names): Add slm cpu name.
15636 (x86_option_override_internal): Set SLM ISA.
15637 (ix86_issue_rate): New case PROCESSOR_SLM.
15638 (ia32_multipass_dfa_lookahead): Likewise.
15639 (fold_builtin_cpu): Add slm.
15640
15641 * config/i386/i386.h (TARGET_SLM): New target macro.
15642 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15643 (processor_type): Add PROCESSOR_SLM.
15644
15645 * config/i386/i386.md (cpu): Add new value "slm".
15646 (slm.md): Include slm.md.
15647
15648 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15649 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15650
15651 * config/arm/arm-protos.h: Add and update function protos.
15652 * config/arm/arm.c (use_simple_return_p): New added.
15653 (thumb2_expand_return): Check simple_return flag.
15654 * config/arm/arm.md: Add simple_return and conditional simple_return.
15655 * config/arm/iterators.md: Add iterator for return and simple_return.
15656
15657 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15658
15659 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15660 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15661 (arm_emit_vfp_multi_reg_pop): Likewise.
15662 (thumb2_emit_ldrd_pop): Likewise.
15663 (arm_expand_epilogue): Add misc REG_CFA notes.
15664 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15665
15666 2013-05-29 Lawrence Crowl <crowl@google.com>
15667
15668 * config/arm/t-arm: Update for below.
15669
15670 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15671 Change type to hash_table. Update dependent calls and types.
15672
15673 * config/i386/t-cygming: Update for below.
15674
15675 * config/i386/t-interix: Update for below.
15676
15677 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15678 Change type to hash_table. Update dependent calls and types.
15679 (i386_find_on_wrapper_list::wrappers): Likewise.
15680
15681 * config/ia64/t-ia64: Update for below.
15682
15683 * config/ia64/ia64.c (bundle_state_table):
15684 Change type to hash_table. Update dependent calls and types.
15685
15686 * config/mips/mips.c (mips_reorg_process_insns::htab):
15687 Change type to hash_table. Update dependent calls and types.
15688
15689 * config/sol2.c (solaris_comdat_htab):
15690 Change type to hash_table. Update dependent calls and types.
15691
15692 * config/t-sol2: Update for above.
15693
15694 2013-05-29 Teresa Johnson <tejohnson@google.com>
15695
15696 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15697 functions are not yet marked as defined.
15698
15699 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15700 Pat Haugen <pthaugen@us.ibm.com>
15701 Peter Bergner <bergner@vnet.ibm.com>
15702
15703 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15704 instructions.
15705 (VEC_A): Likewise.
15706 (VEC_C): Likewise.
15707 (vrotl<mode>3): Likewise.
15708 (vashl<mode>3): Likewise.
15709 (vlshr<mode>3): Likewise.
15710 (vashr<mode>3): Likewise.
15711
15712 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15713 support for power8 V2DI builtins.
15714
15715 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15716 power8 V2DI builtins.
15717 (vupkhsw): Likewise.
15718 (vupklsw): Likewise.
15719 (vaddudm): Likewise.
15720 (vminsd): Likewise.
15721 (vmaxsd): Likewise.
15722 (vminud): Likewise.
15723 (vmaxud): Likewise.
15724 (vpkudum): Likewise.
15725 (vpksdss): Likewise.
15726 (vpkudus): Likewise.
15727 (vpksdus): Likewise.
15728 (vrld): Likewise.
15729 (vsld): Likewise.
15730 (vsrd): Likewise.
15731 (vsrad): Likewise.
15732 (vsubudm): Likewise.
15733 (vcmpequd): Likewise.
15734 (vcmpgtsd): Likewise.
15735 (vcmpgtud): Likewise.
15736 (vcmpequd_p): Likewise.
15737 (vcmpgtsd_p): Likewise.
15738 (vcmpgtud_p): Likewise.
15739 (vupkhsw): Likewise.
15740 (vupklsw): Likewise.
15741 (vaddudm): Likewise.
15742 (vmaxsd): Likewise.
15743 (vmaxud): Likewise.
15744 (vminsd): Likewise.
15745 (vminud): Likewise.
15746 (vpksdss): Likewise.
15747 (vpksdus): Likewise.
15748 (vpkudum): Likewise.
15749 (vpkudus): Likewise.
15750 (vrld): Likewise.
15751 (vsld): Likewise.
15752 (vsrad): Likewise.
15753 (vsrd): Likewise.
15754 (vsubudm): Likewise.
15755
15756 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15757 support for power8 V2DI instructions.
15758
15759 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15760 power8 V2DI instructions. Combine pack and unpack insns to use an
15761 iterator for each mode. Check whether a particular mode supports
15762 Altivec instructions instead of just checking TARGET_ALTIVEC.
15763 (UNSPEC_VPKUWUM): Likewise.
15764 (UNSPEC_VPKSHSS): Likewise.
15765 (UNSPEC_VPKSWSS): Likewise.
15766 (UNSPEC_VPKUHUS): Likewise.
15767 (UNSPEC_VPKSHUS): Likewise.
15768 (UNSPEC_VPKUWUS): Likewise.
15769 (UNSPEC_VPKSWUS): Likewise.
15770 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15771 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15772 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15773 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15774 (UNSPEC_VUPKHSB): Likewise.
15775 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15776 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15777 (UNSPEC_VUPKHSH): Likewise.
15778 (UNSPEC_VUPKLSB): Likewise.
15779 (UNSPEC_VUPKLSH): Likewise.
15780 (VI2): Likewise.
15781 (VI_char): Likewise.
15782 (VI_scalar): Likewise.
15783 (VI_unit): Likewise.
15784 (VP): Likewise.
15785 (VP_small): Likewise.
15786 (VP_small_lc): Likewise.
15787 (VU_char): Likewise.
15788 (add<mode>3): Likewise.
15789 (altivec_vaddcuw): Likewise.
15790 (altivec_vaddu<VI_char>s): Likewise.
15791 (altivec_vadds<VI_char>s): Likewise.
15792 (sub<mode>3): Likewise.
15793 (altivec_vsubcuw): Likewise.
15794 (altivec_vsubu<VI_char>s): Likewise.
15795 (altivec_vsubs<VI_char>s): Likewise.
15796 (altivec_vavgs<VI_char>): Likewise.
15797 (altivec_vcmpbfp): Likewise.
15798 (altivec_eq<mode>): Likewise.
15799 (altivec_gt<mode>): Likewise.
15800 (altivec_gtu<mode>): Likewise.
15801 (umax<mode>3): Likewise.
15802 (smax<mode>3): Likewise.
15803 (umin<mode>3): Likewise.
15804 (smin<mode>3): Likewise.
15805 (altivec_vpkuhum): Likewise.
15806 (altivec_vpkuwum): Likewise.
15807 (altivec_vpkshss): Likewise.
15808 (altivec_vpkswss): Likewise.
15809 (altivec_vpkuhus): Likewise.
15810 (altivec_vpkshus): Likewise.
15811 (altivec_vpkuwus): Likewise.
15812 (altivec_vpkswus): Likewise.
15813 (altivec_vpks<VI_char>ss): Likewise.
15814 (altivec_vpks<VI_char>us): Likewise.
15815 (altivec_vpku<VI_char>us): Likewise.
15816 (altivec_vpku<VI_char>um): Likewise.
15817 (altivec_vrl<VI_char>): Likewise.
15818 (altivec_vsl<VI_char>): Likewise.
15819 (altivec_vsr<VI_char>): Likewise.
15820 (altivec_vsra<VI_char>): Likewise.
15821 (altivec_vsldoi_<mode>): Likewise.
15822 (altivec_vupkhsb): Likewise.
15823 (altivec_vupkhs<VU_char>): Likewise.
15824 (altivec_vupkls<VU_char>): Likewise.
15825 (altivec_vupkhsh): Likewise.
15826 (altivec_vupklsb): Likewise.
15827 (altivec_vupklsh): Likewise.
15828 (altivec_vcmpequ<VI_char>_p): Likewise.
15829 (altivec_vcmpgts<VI_char>_p): Likewise.
15830 (altivec_vcmpgtu<VI_char>_p): Likewise.
15831 (abs<mode>2): Likewise.
15832 (vec_unpacks_hi_v16qi): Likewise.
15833 (vec_unpacks_hi_v8hi): Likewise.
15834 (vec_unpacks_lo_v16qi): Likewise.
15835 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15836 (vec_unpacks_lo_v8hi): Likewise.
15837 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15838 (vec_pack_trunc_v8h): Likewise.
15839 (vec_pack_trunc_v4si): Likewise.
15840 (vec_pack_trunc_<mode>): Likewise.
15841
15842 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15843 V2DI builtins.
15844 (vec_vmaxsd): Likewise.
15845 (vec_vmaxud): Likewise.
15846 (vec_vminsd): Likewise.
15847 (vec_vminud): Likewise.
15848 (vec_vpksdss): Likewise.
15849 (vec_vpksdus): Likewise.
15850 (vec_vpkudum): Likewise.
15851 (vec_vpkudus): Likewise.
15852 (vec_vrld): Likewise.
15853 (vec_vsld): Likewise.
15854 (vec_vsrad): Likewise.
15855 (vec_vsrd): Likewise.
15856 (vec_vsubudm): Likewise.
15857 (vec_vupkhsw): Likewise.
15858 (vec_vupklsw): Likewise.
15859
15860 2013-05-29 Jan Hubicka <jh@suse.cz>
15861
15862 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15863 flags; reorder rest of fields in more consistent way.
15864 (varpool_node): Remove analyzed, finalized and alias.
15865 (cgraph_ndoe): Likewise.
15866 (symtab_alias_ultimate_target): New function.
15867 (cgraph_function_node): Move offline.
15868 (cgraph_reset_node): Declare.
15869 (cgraph_comdat_can_be_unshared_p): Remove.
15870 (varpool_remove_initializer): Declare.
15871 (varpool_first_defined_variable, varpool_next_defined_variable
15872 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15873 (cgraph_function_with_gimple_body_p): Update.
15874 (varpool_all_refs_explicit_p): Update.
15875 (symtab_alias_target): New function.
15876 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15877 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15878 (cgraph_function_or_thunk_node): Simplify using
15879 symtab_alias_ultimate_target.
15880 (varpool_variable_node): Likewise.
15881 * cgraph.c (cgraph_create_function_alias): Update.
15882 (cgraph_add_thunk): Update.
15883 (cgraph_remove_node): Update.
15884 (dump_cgraph_node): Do not dump removed flags.
15885 (cgraph_function_body_availability): Update.
15886 (cgraph_propagate_frequency): Update.
15887 (verify_cgraph_node): Check sanity of local flag.
15888 (cgraph_function_node): Move here from cgraph.h; revamp for
15889 cgraph_function_or_thunk_node.
15890 * lto-symtab.c (lto_varpool_replace_node): Update.
15891 (lto_symtab_resolve_can_prevail_p): Update.
15892 (lto_symtab_merge_cgraph_nodes): Update.
15893 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15894 propagate_constants_accross_call, devirtualization_time_bonus,
15895 ipcp_propagate_stage): Update.
15896 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15897 * ipa-inline-transform.c (clone_inlined_nodes,
15898 preserve_function_body_p): Update.
15899 * ipa-reference.c (propagate): Update.
15900 (write_node_summary_p): Update.
15901 * toplev.c (wrapup_global_declaration_2): Update.
15902 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15903 (analyze_function) ... this one.
15904 (cgraph_process_new_functions): Update.
15905 (cgraph_reset_node): Export.
15906 (cgraph_finalize_function): Update.
15907 (cgraph_add_new_function): Update.
15908 (process_function_and_variable_attributes): Update.
15909 (varpool_finalize_decl): Update.
15910 (symbol_finalized): Remove.
15911 (symbol_finalized_and_needed): Rename to ...
15912 (symbol_defined_and_needed): ... update.
15913 (cgraph_analyze_functions): Update.
15914 (handle_alias_pairs): Update.
15915 (mark_functions_to_output): Update.
15916 (assemble_thunk): Update.
15917 (output_in_order): Update.
15918 (output_weakrefs): Update.
15919 (finalize_compilation_unit): Update.
15920 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15921 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15922 input_node, input_varpool_node): Update.
15923 * dbxout.c (dbxout_expand_expr): Update.
15924 * cgraphclones.c (cgraph_clone_node): Update.
15925 (cgraph_copy_node_for_versioning): Update.
15926 (cgraph_materialize_clone): Update.
15927 (cgraph_materialize_all_clones): Update.
15928 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15929 propagate_pure_const, propagate_nothrow): Update.
15930 * lto-streamer-out.c (lto_output, write_symbol): Update.
15931 * ipa-utils.c (ipa_reverse_postorder): Update.
15932 * ipa-inline.c (can_inline_edge_p): Update.
15933 (update_caller_keys, ipa_inline): Update.
15934 * dwarf2out.c (reference_to_unused,
15935 premark_types_used_by_global_vars_helper): Update.
15936 * tree-eh.c (tree_could_trap_p): Update.
15937 * ipa-split.c (consider_split, execute_split_functions): Update.
15938 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15939 has_addr_references_p): Update; move ahead in file for better
15940 readability.
15941 (process_references): Simplify.
15942 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15943 bodies are removed.
15944 (cgraph_comdat_can_be_unshared_p): Make static.
15945 (cgraph_externally_visible_p): Update.
15946 (varpool_externally_visible_p): Update.
15947 (function_and_variable_visibility): Update.
15948 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15949 ipa_tm_mark_force_output_node): Update.
15950 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15951 estimate_edge_devirt_benefit, inline_generate_summary,
15952 inline_write_summary): Update.
15953 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15954 * ipa-prop.c (ipa_compute_jump_functions): Update.
15955 (ipa_print_node_params, ipa_prop_read_section,
15956 ipa_update_after_lto_read, read_replacements_section): Update.
15957 * varasm.c (mark_decl_referenced): Update.
15958 (assemble_alias, dump_tm_clone_pairs): Update.
15959 * tree-inline.c (copy_bb): Update.
15960 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15961 Update.
15962 * symtab.c (dump_symtab_base): Print new flags.
15963 (verify_symtab_base): Verify new flags.
15964 (symtab_alias_ultimate_target): New function.
15965 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15966 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15967 Update.
15968 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15969 Update.
15970 * i386.c (ix86_get_function_versions_dispatcher,
15971 ix86_generate_version_dispatcher_body): Update.
15972 (fold_builtin_cpu): Use varpool_add_new_variable.
15973 * varpool.c (varpool_remove_initializer): Break out from ...
15974 (varpool_remove_node): ... this one.
15975 (dump_varpool_node, varpool_node_for_asm,
15976 cgraph_variable_initializer_availability, varpool_analyze_node,
15977 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15978 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15979 Update.
15980
15981 2013-05-29 Jan Hubicka <jh@suse.cz>
15982
15983 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15984
15985 2013-05-29 Easwaran Raman <eraman@google.com>
15986
15987 PR tree-optimization/57442
15988 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15989 when control exits the main loop.
15990
15991 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15992
15993 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15994 and RX600.
15995 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15996 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15997 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15998 option.
15999
16000 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16001
16002 PR tree-optimization/57441
16003 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16004 Don't limit size of incr_vec to number of candidates.
16005
16006 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16007
16008 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16009 and mips16 directories.
16010 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16011 (MULTILIB_DIRNAMES): Ditto.
16012 (MULTILIB_EXCEPTIONS): Add new exceptions.
16013 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16014 (MULTILIB_DIRNAMES): Ditto.
16015 (MULTILIB_EXCEPTIONS): Add new exceptions.
16016
16017 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16018 Marcus Shawcroft <marcus.shawcroft@arm.com>
16019
16020 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16021 SYMBOL_TINY_ABSOLUTE.
16022 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16023 SYMBOL_TINY_ABSOLUTE.
16024 (aarch64_expand_mov_immediate): Likewise.
16025 (aarch64_classify_symbol): Likewise.
16026 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16027 Permit SYMBOL_TINY_ABSOLUTE.
16028 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16029
16030 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16031 Marcus Shawcroft <marcus.shawcroft@arm.com>
16032
16033 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16034 Refactor if/switch. Replace gcc_assert with if.
16035
16036 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16037
16038 * config/i386/i386.c (initial_ix86_tune_features): Enable
16039 FP Reassociation for AMD bdver1 and bdver2.
16040
16041 2013-05-29 Martin Jambor <mjambor@suse.cz>
16042
16043 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16044 and IMAGPART_EXPR do not occur within other handled_components.
16045
16046 2013-05-29 Richard Biener <rguenther@suse.de>
16047
16048 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16049 access on whether the use is in the BB we currently try to
16050 vectorize.
16051 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16052 vectorize to vect_bb_slp_scalar_cost.
16053
16054 2013-05-29 Richard Biener <rguenther@suse.de>
16055
16056 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16057 computing scalar cost offsetted by stmts that are kept live
16058 by scalar uses.
16059 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16060 for computation of scalar cost.
16061
16062 2013-05-28 Steve Ellcey <sellcey@mips.com>
16063
16064 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16065
16066 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16067
16068 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16069 array notation built-in reduction functions.
16070 * doc/passes.texi (Passes): Added documentation about changes done
16071 for Cilk Plus.
16072 * doc/invoke.texi (C Dialect Options): Added documentation about
16073 the -fcilkplus flag.
16074 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16075 (BUILTINS_DEF): Depend on cilkplus.def.
16076 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16077 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16078 * cilkplus.def: New file.
16079
16080 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16081
16082 PR rtl-optimization/57439
16083 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16084
16085 2013-05-28 Easwaran Raman <eraman@google.com>
16086
16087 PR tree-optimization/57337
16088 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16089 (find_insert_point): Correctly identify the insertion point
16090 when two statements with the same UID is compared.
16091
16092 2013-05-28 Richard Biener <rguenther@suse.de>
16093
16094 PR tree-optimization/56787
16095 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16096 from the list of data references.
16097 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16098 clobbers.
16099 (vect_analyze_loop_operations): Likewise.
16100 (vect_transform_loop): Remove clobbers.
16101
16102 2013-05-28 Martin Jambor <mjambor@suse.cz>
16103
16104 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16105 and REALPART_EXPRs have scalar type.
16106
16107 2013-05-28 Richard Biener <rguenther@suse.de>
16108
16109 PR tree-optimization/57411
16110 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16111 virtual operands.
16112 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16113 virtual operand propagation.
16114
16115 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16116
16117 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16118 destination register for bmasksi_vis.
16119 (vector_init_bshuffle): Likewise.
16120 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16121
16122 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16123
16124 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16125 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16126 mode if the instruction isn't available in the original mode.
16127 * config/sparc/sparc.opt (mfix-ut699): New option.
16128 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16129 (divdf3): Turn into expander.
16130 (divdf3_nofix): New insn.
16131 (divdf3_fix): Likewise.
16132 (divsf3): Disable if -mfix-ut699.
16133 (sqrtdf2): Turn into expander.
16134 (sqrtdf2_nofix): New insn.
16135 (sqrtdf2_fix): Likewise.
16136 (sqrtsf2): Disable if -mfix-ut699.
16137
16138 2013-05-27 Richard Biener <rguenther@suse.de>
16139
16140 PR middle-end/57412
16141 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16142 block for the new loop.
16143
16144 2013-05-27 Richard Biener <rguenther@suse.de>
16145
16146 PR tree-optimization/57343
16147 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16148 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16149 (number_of_iterations_cond): Do not build the folded tree.
16150
16151 2013-05-27 Richard Biener <rguenther@suse.de>
16152
16153 Revert
16154 PR middle-end/57381
16155 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16156 OEP_CONSTANT_ADDRESS_OF retained.
16157
16158 PR tree-optimization/57417
16159 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16160 for unchanged base.
16161 (set_ssa_val_to): Compare addresses using
16162 get_addr_base_and_unit_offset.
16163
16164 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16165
16166 PR rtl-optimization/56833
16167 * postreload.c (move2add_record_mode): New function.
16168 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16169 (move2add_use_add2_insn): Use move2add_record_sym_value.
16170 (move2add_use_add3_insn): Likewise.
16171 (reload_cse_move2add): Use move2add_valid_value_p and
16172 move2add_record_mode. Invalidate call-clobbered and REG_INC
16173 affected regs by setting reg_mode to VOIDmode.
16174 (move2add_note_store): Don't pretend the inside of a SUBREG is
16175 the actual destination. Invalidate single/leading registers by
16176 setting reg_mode to VOIDmode.
16177 Use move2add_record_sym_value, move2add_valid_value_p and
16178 move2add_record_mode.
16179
16180 2013-05-27 Richard Biener <rguenther@suse.de>
16181
16182 PR tree-optimization/57396
16183 * tree-affine.c (double_int_constant_multiple_p): Properly
16184 return false for val == 0 and div != 0.
16185
16186 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16187
16188 * config/mips/mips.h: Use #elif in preprocessor conditions.
16189
16190 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16191
16192 PR target/53916
16193 * config/mips/constraints.md (kl): New constraint.
16194 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16195 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16196 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16197 from LO for MIPS16.
16198 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16199
16200 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16201
16202 PR target/55777
16203 * config/mips/mips.c (mips_can_inline_p): New function.
16204 (TARGET_CAN_INLINE_P): Define.
16205
16206 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16207
16208 * sched-int.h (ds_t, dw_t): Make unsigned int.
16209 Fix documentation that describes how all the ds_t bits are used.
16210 Reserve the last bit for delayed-branch scheduling.
16211 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16212 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16213 (gen_dep_weak_1): Remove prototype.
16214 * sched-deps.c (get_dep_weak_1): Make static.
16215 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16216 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16217 * doc/tm.texi: Regenerate.
16218 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16219
16220 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16221
16222 PR debug/56950
16223 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16224
16225 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16226 Sandra Loosemore <sandra@codesourcery.com>
16227
16228 * config.gcc (powerpc-*): Allow native for with-cpu.
16229
16230 2013-05-24 Jeff Law <law@redhat.com>
16231
16232 PR tree-optimization/57124
16233 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16234 conversion feeding a condition if the range has an overflow
16235 if -fstrict-overflow. Add warnings for when we do make the
16236 transformation.
16237
16238 2013-05-24 Dehao Chen <dehao@google.com>
16239
16240 * tree-cfg.c (locus_discrim_map): Fix the typo.
16241 (locus_discrim_hasher): Likewise.
16242 (locus_discrim_hasher::hash): Likewise.
16243 (locus_discrim_hasher::equal): Likewise.
16244
16245 2013-05-24 Martin Jambor <mjambor@suse.cz>
16246
16247 PR tree-optimization/57294
16248 * cgraph.h (ipa_record_stmt_references): Declare.
16249 * cgraphbuild.c (ipa_record_stmt_references): New function.
16250 (build_cgraph_edges): Use ipa_record_stmt_references.
16251 (rebuild_cgraph_edges): Likewise.
16252 (cgraph_rebuild_references): Likewise.
16253 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16254 associated with the old statement and build references from the
16255 newly built statements.
16256 * ipa-ref.c (ipa_remove_stmt_references): New function.
16257 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16258
16259 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16260
16261 * lra-constraints.c (emit_spill_move): Use smaller mode for
16262 mem-mem moves.
16263 (check_and_process_move): Consider mem-reg moves for secondary
16264 too.
16265 (curr_insn_transform): Don't lose insns emitted before for
16266 secondary memory moves.
16267 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16268 reg set up in the current insn.
16269
16270 2013-05-24 Dehao Chen <dehao@google.com>
16271
16272 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16273 hash function.
16274 (locus_descrim_hasher::equal): Likewise.
16275 (build_gimple_cfg): New discriminator assignment algorithm.
16276 (make_edges): Likewise.
16277 (next_discriminator_for_locus): Likewise.
16278 (same_line_p): Likewise.
16279 (assign_discriminators): Likewise.
16280 (make_cond_expr_edges): Likewise.
16281 (make_gimple_switch_edges): Likewise.
16282 (make_goto_expr_edges): Likewise.
16283 (make_gimple_asm_edges): Likewise.
16284
16285 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16286
16287 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16288 X format specifier to only display bottom 16 bits.
16289 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16290 immediate to match for operand 2, since it will be masked.
16291
16292 2013-05-24 Richard Biener <rguenther@suse.de>
16293
16294 PR tree-optimization/57287
16295 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16296 all SSA names that occur in abnormal PHIs.
16297
16298 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16299
16300 PR tree-ssa/57385
16301 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16302 that index is not negative.
16303
16304 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16305
16306 PR rtl-optimization/55177
16307 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16308 (simplify_byte_swapping_operation): New.
16309 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16310 (simplify_relational_operation_1): Deal with BSWAP.
16311
16312 2013-05-23 Richard Henderson <rth@redhat.com>
16313
16314 PR target/56742
16315 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16316 (ix86_reorg): Call it.
16317
16318 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16319
16320 PR target/57379
16321 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16322 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16323 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16324
16325 2013-05-23 Christian Bruel <christian.bruel@st.com>
16326
16327 PR debug/57351
16328 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16329
16330 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16331 Marcus Shawcroft <marcus.shawcroft@arm.com>
16332
16333 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16334 * config/aarch64/constraints.md (Usa): Remove.
16335 * doc/md.texi (AArch64 Usa): Remove.
16336
16337 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16338 Marcus Shawcroft <marcus.shawcroft@arm.com>
16339
16340 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16341 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16342 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16343 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16344
16345 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16346
16347 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16348 instruction (AdvSIMD).
16349 * config/aarch64/aarch64-builtins.c
16350 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16351 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16352
16353 2013-05-23 Martin Jambor <mjambor@suse.cz>
16354
16355 PR middle-end/57347
16356 * tree.h (contains_bitfld_component_ref_p): Declare.
16357 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16358 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16359 caller.
16360 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16361 not access a bit-field. Assert all final offsets are byte-aligned.
16362
16363 2013-05-23 Richard Biener <rguenther@suse.de>
16364
16365 PR tree-optimization/57380
16366 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16367 least one invariant or re-used load.
16368 * passes.c (init_optimization_passes): Move pass_phiprop before
16369 pass_forwprop.
16370
16371 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16372
16373 * config/aarch64/aarch64-simd.md
16374 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16375
16376 2013-05-23 Richard Biener <rguenther@suse.de>
16377
16378 PR middle-end/57381
16379 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16380 OEP_CONSTANT_ADDRESS_OF retained.
16381
16382 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16383
16384 PR middle-end/57344
16385 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16386 don't lower unit. Handle unit not being always BITS_PER_WORD.
16387
16388 2013-05-23 Richard Biener <rguenther@suse.de>
16389
16390 PR rtl-optimization/57341
16391 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16392 instead of true_dependence.
16393
16394 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16395
16396 * bb-reorder.c (branch_threshold): Make const.
16397 (exec_threshold): Ditto.
16398
16399 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16400 Pat Haugen <pthaugen@us.ibm.com>
16401 Peter Bergner <bergner@vnet.ibm.com>
16402
16403 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16404 documentation for the power8 crypto builtins.
16405
16406 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16407
16408 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16409 macros for defining power8 builtin functions.
16410 (BU_P8V_AV_2): Likewise.
16411 (BU_P8V_AV_P): Likewise.
16412 (BU_P8V_VSX_1): Likewise.
16413 (BU_P8V_OVERLOAD_1): Likewise.
16414 (BU_P8V_OVERLOAD_2): Likewise.
16415 (BU_CRYPTO_1): Likewise.
16416 (BU_CRYPTO_2): Likewise.
16417 (BU_CRYPTO_3): Likewise.
16418 (BU_CRYPTO_OVERLOAD_1): Likewise.
16419 (BU_CRYPTO_OVERLOAD_2): Likewise.
16420 (XSCVSPDP): Fix typo, point to the correct instruction.
16421 (VCIPHER): Add power8 crypto builtins.
16422 (VCIPHERLAST): Likewise.
16423 (VNCIPHER): Likewise.
16424 (VNCIPHERLAST): Likewise.
16425 (VPMSUMB): Likewise.
16426 (VPMSUMH): Likewise.
16427 (VPMSUMW): Likewise.
16428 (VPERMXOR_V2DI): Likewise.
16429 (VPERMXOR_V4SI: Likewise.
16430 (VPERMXOR_V8HI: Likewise.
16431 (VPERMXOR_V16QI: Likewise.
16432 (VSHASIGMAW): Likewise.
16433 (VSHASIGMAD): Likewise.
16434 (VPMSUM): Likewise.
16435 (VPERMXOR): Likewise.
16436 (VSHASIGMA): Likewise.
16437
16438 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16439 __CRYPTO__ if the crypto instructions are available.
16440 (altivec_overloaded_builtins): Add support for overloaded power8
16441 builtins.
16442
16443 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16444 support for power8 crypto builtins.
16445 (builtin_function_type): Likewise.
16446 (altivec_init_builtins): Add support for builtins that take vector
16447 long long (V2DI) arguments.
16448
16449 * config/rs6000/crypto.md: New file, define power8 crypto
16450 instructions.
16451
16452 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16453 Pat Haugen <pthaugen@us.ibm.com>
16454 Peter Bergner <bergner@vnet.ibm.com>
16455
16456 * doc/invoke.texi (Option Summary): Add power8 options.
16457 (RS/6000 and PowerPC Options): Likewise.
16458
16459 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16460 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16461 wm, wn, wr documentation.
16462
16463 * config/rs6000/constraints.md (wm): New constraint for VSX
16464 registers if direct move instructions are enabled.
16465 (wn): New constraint for no registers.
16466 (wq): New constraint for quad word even GPR registers.
16467 (wr): New constraint if 64-bit instructions are enabled.
16468 (wv): New constraint if power8 vector instructions are enabled.
16469 (wQ): New constraint for quad word memory locations.
16470
16471 * config/rs6000/predicates.md (const_0_to_15_operand): New
16472 constraint for 0..15 for crypto instructions.
16473 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16474 as GPR and floating point registers.
16475 (int_reg_operand): New predicate to match only GPR registers.
16476 (base_reg_operand): New predicate to match base registers.
16477 (quad_int_reg_operand): New predicate to match even GPR registers
16478 for quad memory operations.
16479 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16480 operations in both GPR and VSX registers.
16481 (quad_memory_operand): New predicate for quad memory operations.
16482 (reg_or_indexed_operand): New predicate for direct move support.
16483
16484 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16485 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16486 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16487 (POWERPC_MASKS): Add power8 options.
16488 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16489 various options.
16490
16491 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16492 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16493
16494 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16495 (-mpower8-fusion): New power8 options.
16496 (-mpower8-fusion-sign): Likewise.
16497 (-mpower8-vector): Likewise.
16498 (-mcrypto): Likewise.
16499 (-mdirect-move): Likewise.
16500 (-mquad-memory): Likewise.
16501
16502 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16503 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16504 registers.
16505 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16506 (rs6000_debug_vector_unit): Add p8_vector.
16507 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16508 definitions. Also print fusion state.
16509 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16510 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16511 (rs6000_option_override_internal): Add support for power8.
16512 (rs6000_common_init_builtins): Add debugging for skipped builtins
16513 if -mdebug=builtin.
16514 (rs6000_adjust_cost): Add power8 support.
16515 (rs6000_issue_rate): Likewise.
16516 (insn_must_be_first_in_group): Likewise.
16517 (insn_must_be_last_in_group): Likewise.
16518 (force_new_group): Likewise.
16519 (rs6000_register_move_cost): Likewise.
16520 (rs6000_opt_masks): Likewise.
16521
16522 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16523 power8 capable assembler, default to power7 options.
16524 (TARGET_DIRECT_MOVE): Likewise.
16525 (TARGET_CRYPTO): Likewise.
16526 (TARGET_P8_VECTOR): Likewise.
16527 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16528 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16529 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16530 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16531 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16532 (TARGET_XSCVDPSPN): Likewise.
16533 (TARGET_XSCVSPDPN): Likewsie.
16534 (TARGET_SYNC_HI_QI): Likewise.
16535 (TARGET_SYNC_TI): Likewise.
16536 (MASK_CRYPTO): Likewise.
16537 (MASK_DIRECT_MOVE): Likewise.
16538 (MASK_P8_FUSION): Likewise.
16539 (MASK_P8_VECTOR): Likewise.
16540 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16541 temporary used by some of the direct move instructions to get two FP
16542 temporary registers does not force creation of a stack frame.
16543 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16544 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16545 that any VSX registers are tieable, even if they are also an
16546 Altivec vector mode.
16547 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16548 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16549 (RS6000_BTM_CRYPTO): Likewise.
16550 (RS6000_BTM_COMMON): Likewise.
16551
16552 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16553 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16554 (enum rs6000_vector): Add power8 vector support.
16555
16556 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16557
16558 PR target/19599
16559 PR target/57340
16560 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16561 (any_sibcall_could_use_r3): this and handle indirect calls.
16562 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16563
16564 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16565
16566 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16567
16568 2013-05-22 Richard Biener <rguenther@suse.de>
16569
16570 PR middle-end/57349
16571 * profile.c (branch_prob): Do not split blocks that are
16572 abnormally receiving from ECF_RETURNS_TWICE functions.
16573
16574 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16575
16576 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16577 address mode. Move pointer mode initialization to the same place.
16578
16579 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16580
16581 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16582 while it has any effect.
16583
16584 2013-05-21 Easwaran Raman <eraman@google.com>
16585
16586 PR tree-optimization/57322
16587 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16588 UID of the statement added to the BB to be 1.
16589
16590 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16591
16592 PR tree-optimization/57331
16593 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16594 of conversion from pointer type to integral type with integer.
16595
16596 2013-05-21 Martin Jambor <mjambor@suse.cz>
16597
16598 PR lto/57289
16599 * ipa-prop.c (ipa_read_node_info): Process param_used and
16600 controlled_uses in the same order as when writing.
16601
16602 2013-05-21 Magnus Granberg <baldrick@free.fr>
16603
16604 PR plugins/56754
16605 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16606
16607 2013-05-21 Richard Biener <rguenther@suse.de>
16608
16609 PR tree-optimization/57318
16610 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16611 estimate stmts with side-effects as likely eliminated.
16612
16613 2013-05-21 Richard Biener <rguenther@suse.de>
16614
16615 PR tree-optimization/57330
16616 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16617 preserve the call stmts fntype.
16618
16619 2013-05-21 Richard Biener <rguenther@suse.de>
16620
16621 PR tree-optimization/57303
16622 * tree-ssa-sink.c (statement_sink_location): Improve killing
16623 stmt detection and properly handle self-assignments.
16624
16625 2013-05-21 Christian Bruel <christian.bruel@st.com>
16626
16627 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16628 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16629 registers. Set register size out of the PARALLEL loop.
16630
16631 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16632
16633 PR target/56547
16634 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16635 (*fmasf4, *fmasf4_media): New insns.
16636
16637 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16638
16639 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16640 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16641 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16642 (mips_idiv_insns): Update the comments to say that the returned
16643 instruction counts are in units of BASE_INSN_LENGTH.
16644 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16645 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16646 using 2 rather than 4 as the length of indirect MIPS16 and
16647 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16648 length of a NOP. Don't divide MIPS16 lengths by 2.
16649 (mips16_split_long_branches): Assume a branch is long if the
16650 length is greater than 4 rather than 8.
16651 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16652 rather than multiplying them by 2. Multiply instruction counts
16653 by BASE_INSN_LENGTH rather than 4.
16654 (*jump_mips16, tls_get_tp_mips16_<mode>)
16655 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16656
16657 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16658
16659 * config/mips/mips.md (extended_mips16): Remove branch case.
16660 (length): Remove duplicated extended_mips16 test.
16661
16662 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16663
16664 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16665
16666 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16667
16668 * recog.h (Recog_data): Rename to...
16669 (recog_data_d): ...this.
16670 (recog_data): Update accordingly.
16671 * recog.c (recog_data): Likewise.
16672 * reload.c (save_recog_data): Likewise.
16673 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16674 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16675
16676 2013-05-17 Julian Brown <julian@codesourcery.com>
16677
16678 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16679 found in a REG_EQUAL note, invalidate it.
16680
16681 2013-05-17 Easwaran Raman <eraman@google.com>
16682
16683 * tree-ssa-reassoc.c (find_insert_point): New function.
16684 (insert_stmt_after): Likewise.
16685 (get_def_stmt): Likewise.
16686 (ensure_ops_are_available): Likewise.
16687 (not_dominated_by): Likewise.
16688 (rewrite_expr_tree): Do not move statements beyond what is
16689 necessary. Remove call to swap_ops_for_binary_stmt...
16690 (reassociate_bb): ... and move it here.
16691 (build_and_add_sum): Assign UIDs for new statements.
16692 (linearize_expr): Likewise.
16693 (do_reassoc): Renumber gimple statement UIDs.
16694
16695 2013-05-17 Jan Hubicka <jh@suse.cz>
16696
16697 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16698 weakrefs.
16699 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16700 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16701 weakrefs.
16702 (output_weakrefs): Update.
16703
16704 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16705 Martin Jambor <mjambor@suse.cz>
16706
16707 PR middle-end/57276
16708 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16709 value that corresponds to the given aggval is found in values vector.
16710
16711 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16712
16713 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16714 sse, sse2, sse3, ssse3 and sse4a flags to options.
16715
16716 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16717
16718 * gengtype-state.c: (s_expr_writer): New class, to handle
16719 prettifying of output layout of s-expressions.
16720 (state_writer): New class, to write out gtype.state.
16721 (state_written_type_count): Move this variable into member data of
16722 state_writer.
16723 (s_expr_writer::s_expr_writer): New code: constructor for new class
16724 (state_writer::state_writer(): ditto
16725 (s_expr_writer::write_new_line): New function
16726 (s_expr_writer::write_any_indent): ditto
16727 (s_expr_writer::begin_s_expr): ditto
16728 (s_expr_writer::end_s_expr): ditto
16729 (write_state_fileloc): convert to method of state_writer...
16730 (state_writer:: write_state_fileloc): ...and use methods of
16731 s_expr_writer to write indentation into the gtype.state output file
16732 to visually represent the hierarchical structure of the list
16733 structures
16734 (write_state_fields): ditto, renaming to...
16735 (state_writer::write_state_fields)
16736 (write_state_a_string): ditto, renaming to...
16737 (state_writer::write_state_a_string)
16738 (write_state_string_option): ditto, renaming to...
16739 (state_writer::write_state_string_option)
16740 (write_state_type_option): ditto, renaming to...
16741 (state_writer::write_state_type_option)
16742 (write_state_nested_option): ditto, renaming to...
16743 (state_writer::write_state_nested_option)
16744 (write_state_option): ditto, renaming to...
16745 (state_writer::write_state_option)
16746 (write_state_options): ditto, renaming to...
16747 (state_writer::write_state_options)
16748 (write_state_lang_bitmap): ditto, renaming to...
16749 (state_writer::write_state_lang_bitmap)
16750 (write_state_version): ditto, renaming to...
16751 (state_writer::write_state_version)
16752 (write_state_scalar_type): ditto, renaming to...
16753 (state_writer::write_state_scalar_type)
16754 (write_state_string_type): ditto, renaming to...
16755 (state_writer::write_state_string_type)
16756 (write_state_undefined_type): ditto, renaming to...
16757 (state_writer::write_state_undefined_type)
16758 (write_state_struct_union_type): ditto, renaming to...
16759 (state_writer::write_state_struct_union_type)
16760 (write_state_struct_type): ditto, renaming to...
16761 (state_writer::write_state_struct_type)
16762 (write_state_user_struct_type): ditto, renaming to...
16763 (state_writer::write_state_user_struct_type)
16764 (write_state_lang_struct_type): ditto, renaming to...
16765 (state_writer::write_state_lang_struct_type)
16766 (write_state_param_struct_type): ditto, renaming to...
16767 (state_writer::write_state_param_struct_type)
16768 (write_state_pointer_type): ditto, renaming to...
16769 (state_writer::write_state_pointer_type)
16770 (write_state_array_type): ditto, renaming to...
16771 (state_writer::write_state_array_type)
16772 (write_state_gc_used): ditto, renaming to...
16773 (state_writer::write_state_gc_used)
16774 (write_state_common_type_content): ditto, renaming to...
16775 (state_writer::write_state_common_type_content)
16776 (write_state_type): ditto, renaming to...
16777 (state_writer::write_state_type)
16778 (write_state_pair_list): ditto, renaming to...
16779 (state_writer::write_state_pair_list)
16780 (write_state_pair): ditto, renaming to...
16781 (state_writer::write_state_pair)
16782 (write_state_typedefs): ditto, renaming to...
16783 (state_writer::write_state_typedefs)
16784 (write_state_structures): ditto, renaming to...
16785 (state_writer::write_state_structures)
16786 (write_state_param_structs): ditto, renaming to...
16787 (state_writer::write_state_param_structs)
16788 (write_state_variables): ditto, renaming to...
16789 (state_writer::write_state_variables)
16790 (write_state_srcdir): ditto, renaming to...
16791 (state_writer::write_state_srcdir)
16792 (write_state_files_list): ditto, renaming to...
16793 (state_writer::write_state_files_list)
16794 (write_state_languages): ditto, renaming to...
16795 (state_writer::write_state_languages)
16796 (write_state): create a state_writer instance and use it when
16797 writing out the state file
16798
16799 2013-05-17 Mike Stump <mikestump@comcast.net>
16800
16801 PR rtl-optimization/57304
16802 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16803 accessing DF_REF_REAL_LOC.
16804
16805 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16806
16807 PR rtl-optimization/57281
16808 PR rtl-optimization/57300
16809 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16810 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16811 what the other splitter did if the registers are dead.
16812
16813 2013-05-17 Richard Biener <rguenther@suse.de>
16814
16815 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16816 MEM_REF offsets.
16817
16818 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16819
16820 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16821 linking.
16822
16823 2013-05-17 Marek Polacek <polacek@redhat.com>
16824
16825 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16826 length when doing non-zero store of storing '\0' to '\0'.
16827
16828 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16829
16830 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16831 vect_external_def oprnd1 with loop_vinfo, try to emit
16832 optional cast, negation and and stmts on the loop preheader
16833 edge instead of into the pattern def seq.
16834
16835 PR tree-optimization/57051
16836 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16837 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16838
16839 2013-05-16 Nick Clifton <nickc@redhat.com>
16840
16841 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16842 (rl78_is_naked_func): New function.
16843 (rl78_expand_prologue): Skip prologue generation for naked functions.
16844 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16845 * doc/extend.texi (naked): Add RL78 to the list of processors
16846 that supports this attribute.
16847
16848 2013-05-16 Jeff Law <law@redhat.com>
16849
16850 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16851
16852 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16853
16854 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16855 cache parameters using detect_caches_amd also for CYRIX,
16856 NSC and TM2 signatures.
16857
16858 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16859 Dzianis Kahanovich <mahatma@eu.by>
16860
16861 PR target/45359
16862 PR target/46396
16863 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16864 VIA/Centaur processors and determine their cache parameters
16865 using detect_caches_amd.
16866
16867 2013-05-16 Teresa Johnson <tejohnson@google.com>
16868
16869 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16870 (rtl_verify_edges): New function.
16871 (rtl_verify_bb_insns): Ditto.
16872 (rtl_verify_bb_pointers): Ditto.
16873 (rtl_verify_bb_insn_chain): Ditto.
16874 (rtl_verify_fallthru): Ditto.
16875 (rtl_verify_bb_layout): Ditto.
16876 (rtl_verify_flow_info_1): Outline checks into new functions.
16877 (rtl_verify_flow_info): Ditto.
16878
16879 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16880
16881 * cfghooks.c (copy_bbs): Add update_dominance argument.
16882 * cfghooks.h (copy_bbs): Update prototype.
16883 * tree-cfg.c (gimple_duplicate_sese_region):
16884 Add update_dominance argument.
16885 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16886 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16887 gimple_duplicate_sese_region call.
16888 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16889 Update copy_bbs call.
16890 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16891 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16892
16893 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16894
16895 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16896 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16897 vect_recog_rotate_pattern.
16898 (vect_recog_rotate_pattern): New function.
16899
16900 2013-05-16 Jason Merrill <jason@redhat.com>
16901
16902 * Makefile.in (LLINKER): New variable.
16903 (mostlyclean): Remove link mutex.
16904 * configure.ac: Handle --enable-link-mutex.
16905 * lock-and-run.sh: New script.
16906
16907 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16908
16909 PR target/19599
16910 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16911 for NULL decl.
16912
16913 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16914
16915 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16916
16917 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16918
16919 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16920 * config/arm/arm.c (next_consecutive_mem): New function.
16921 (gen_movmem_ldrd_strd): Likewise.
16922 * config/arm/arm.md (movmemqi): Update condition and code.
16923 (unaligned_loaddi, unaligned_storedi): New patterns.
16924
16925 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16926
16927 * config.gcc: Obsolete *-*-solaris2.9*.
16928 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16929
16930 2013-05-16 Richard Biener <rguenther@suse.de>
16931
16932 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16933 earlier, after GRAPHITE transforms and IV canonicalization.
16934
16935 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16936
16937 * omp-low.c (extract_omp_for_data): For collapsed loops,
16938 if at least one of the loops is known at compile time to
16939 iterate zero times, set count to 0.
16940 (expand_omp_regimplify_p): New function.
16941 (expand_omp_for_generic): For collapsed loops, if at least
16942 one of the loops isn't known to iterate at least once,
16943 add runtime check with setting count to 0.
16944 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16945 For unsigned types if it isn't known at compile time that
16946 the loop will iterate at least once, add runtime check to bypass
16947 the whole loop if initial condition isn't true.
16948
16949 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16950
16951 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16952
16953 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16954
16955 PR middle-end/57286
16956 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16957 transformations to avoid an infinite loop.
16958
16959 2013-05-16 Marek Polacek <polacek@redhat.com>
16960
16961 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16962
16963 2013-05-15 Leif Ekblad <leif@rdos.net>
16964
16965 * config/i386/i386.c (ix86_decompose_address): Use
16966 DEFAULT_TLS_SEG_REG to access TLS segment register.
16967 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16968 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16969 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16970
16971 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16972
16973 PR target/57260
16974 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16975 sibling calls to functions that would normally be lazily bound,
16976 unless $gp is call-clobbered.
16977
16978 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16979
16980 * config/i386/i386.c (ix86_option_override_internal): Update
16981 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16982 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16983 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16984 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16985 of TARGET_3DNOW.
16986 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16987
16988 2013-05-15 Andreas Schwab <schwab@suse.de>
16989
16990 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16991 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16992 third operand.
16993
16994 2013-05-15 Teresa Johnson <tejohnson@google.com>
16995
16996 * loop-unroll.c (report_unroll_peel): Check decision before
16997 emitting unroll/peel message.
16998
16999 2013-05-15 Teresa Johnson <tejohnson@google.com>
17000
17001 * function.h (has_bb_partition): New rtl_data flag.
17002 (bb_reorder_complete): Ditto.
17003 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17004 instead of flag_reorder_blocks_and_partition.
17005 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17006 with some enhancements.
17007 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17008 * bb-reorder.c (connect_traces): Check for has_bb_partition
17009 instead of flag_reorder_blocks_and_partition.
17010 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17011 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17012 verify_hot_cold_block_grouping.
17013 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17014
17015 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17016
17017 PR target/19599
17018 * config/arm/predicates.md (call_insn_operand): New predicate.
17019 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17020 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17021 if insn is not a tail call.
17022 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17023 registers.
17024 * config/arm/arm.h (enum reg_class): New caller save register class.
17025 (REG_CLASS_NAMES): Likewise.
17026 (REG_CLASS_CONTENTS): Likewise.
17027 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17028 without decls.
17029
17030 2013-05-15 Richard Biener <rguenther@suse.de>
17031
17032 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17033 of MSG_OPTIMIZED_LOCATIONS.
17034 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17035 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17036 message.
17037 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17038 of MSG_OPTIMIZED_LOCATIONS.
17039 (execute_vect_slp): Likewise.
17040 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17041 (vect_create_cond_for_alias_checks): Likewise.
17042 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17043 (vect_recog_widen_mult_pattern): Likewise.
17044 (vect_recog_widen_sum_pattern): Likewise.
17045 (vect_recog_over_widening_pattern): Likewise.
17046 (vect_recog_widen_shift_pattern): Likewise.
17047 (vect_recog_vector_vector_shift_pattern): Likewise.
17048 (vect_recog_divmod_pattern): Likewise.
17049 (vect_recog_mixed_size_cond_pattern): Likewise.
17050 (vect_recog_bool_pattern): Likewise.
17051 (vect_pattern_recog_1): Likewise.
17052
17053 2013-05-15 Martin Jambor <mjambor@suse.cz>
17054
17055 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17056 non-functions to builtin_unreachable.
17057 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17058 correct when new direct edges were discovered.
17059
17060 2013-05-15 Martin Jambor <mjambor@suse.cz>
17061
17062 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17063 header, print symbol order instead of node uid, print more information
17064 about indirect edge targets.
17065 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17066 uids.
17067 (ipa_make_edge_direct_to_target): Likewise.
17068 (remove_described_reference): Likewise.
17069 (propagate_controlled_uses): Likewise.
17070 (ipa_print_node_params): Also print symbol order.
17071 (ipcp_transform_function): Print symbol order instead of node uids.
17072 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17073 (cgraph_get_create_real_symbol_node): Likewise.
17074 * ipa-cp.c (print_lattice): Likewise.
17075 (print_all_lattices): Likewise.
17076 (determine_versionability): Likewise.
17077 (initialize_node_lattices): Likewise.
17078 (estimate_local_effects): Likewise.
17079 (update_profiling_info): Likewise.
17080 (create_specialized_node): Likewise.
17081 (perhaps_add_new_callers): Likewise.
17082 (decide_about_value): Likewise.
17083 (decide_whether_version_node): Likewise.
17084 (identify_dead_nodes): Likewise.
17085 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17086 (dump_inline_summary): Likewise.
17087 (estimate_node_size_and_time): Likewise.
17088 (inline_analyze_function): Likewise.
17089 * ipa-inline.c (report_inline_failed_reason): Likewise.
17090 (want_early_inline_function_p): Likewise.
17091 (edge_badness): Likewise.
17092 (update_edge_key): Likewise.
17093 (inline_small_functions): Likewise. Add dumping of order to two other
17094 dumps.
17095 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17096 instead of node uids.
17097 (propagate_pure_const): Likewise.
17098 (propagate_pure_const): Likewise.
17099 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17100 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17101 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17102 of node uids.
17103 * tree-pretty-print.c (dump_function_header): Likewise.
17104 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17105 Print symbol order instead of node uids.
17106
17107 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17108
17109 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17110 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17111
17112 2013-05-15 Richard Biener <rguenther@suse.de>
17113
17114 PR tree-optimization/57275
17115 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17116 return value for fail to do runtime alias checks for gather loads.
17117
17118 2013-05-15 Jan Hubicka <jh@suse.cz>
17119
17120 PR lto/57038
17121 PR lto/47375
17122 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17123 weakrefs are not external.
17124 (lto_symtab_merge_decls): Fix thinko when dealing with
17125 non-lto_symtab decls.
17126 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17127 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17128 * varpool.c (dump_varpool_node): Dump more flags.
17129
17130 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17131
17132 * config/i386/i386.c (processor_alias_table): Add instruction
17133 FSGSBASE for AMD bdver3 architecture.
17134
17135 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17136
17137 * tree.c (warn_deprecated_use): Print file:line using locus color.
17138 * diagnostic.c (diagnostic_report_current_module): Print file:line
17139 and file:line:column using locus color.
17140
17141 2013-05-14 Mike Stump <mikestump@comcast.net>
17142
17143 * gdbinit.in: Add __null.
17144
17145 2013-05-14 Mike Stump <mikestump@comcast.net>
17146
17147 * recog.h: Rename struct recog_data to Recog_data.
17148 * recog.c: Likewise.
17149 * reload.c (can_reload_into): Likewise.
17150 * config/picochip/picochip.c: Likewise.
17151
17152 2013-05-14 Mike Stump <mikestump@comcast.net>
17153
17154 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17155
17156 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17157
17158 * resource.h (struct resources): Remove unch_memory member.
17159 (CLEAR_RESOURCE): Don't clear unch_memory.
17160 * resource.c (mark_referenced_resources): Don't set it.
17161 (mark_set_resources): Likewise.
17162 (mark_target_live_regs): Don't clear it.
17163 (init_resource_info): Likewise.
17164 * reorg.c (resource_conflicts_p): Don't compare it.
17165 (redundant_insn): Don't set it.
17166
17167 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17168 Remove prototypes.
17169 * emit-rtl.c (next_label): Remove unused function.
17170 (skip_consecutive_labels, link_cc0_insns): Move to ...
17171 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17172 only place where these functions are used, and make them static.
17173
17174 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17175
17176 * fold-const.c (fold_negate_expr): Handle vectors.
17177 (fold_truth_not_expr): Make it static.
17178 (fold_invert_truthvalue): New static function.
17179 (invert_truthvalue_loc): Handle vectors. Do not call
17180 fold_truth_not_expr directly.
17181 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17182 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17183 (fold_comparison): Handle vector constants.
17184 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17185 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17186 * tree.h (fold_truth_not_expr): Remove declaration.
17187
17188 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17189
17190 * config/aarch64/aarch64-simd.md
17191 (aarch64_vcond_internal<mode>): Rename to...
17192 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17193 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17194 float modes. Clarify all iterator modes.
17195 (vcond<mode><mode>): Use new name for vcond expanders.
17196 (vcond<v_cmp_result><mode>): Likewise.
17197 (vcondu<mode><mode>: Likewise.
17198 * config/aarch64/iterators.md (VDQF_COND): New.
17199
17200 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17201
17202 PR bootstrap/57266
17203 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17204 variable for the shift amount. Check that we shift by non-negative
17205 amounts.
17206
17207 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17208
17209 PR target/42017
17210 * config/arm/arm.h (EPILOGUE_USES): Only return true
17211 for LR_REGNUM after epilogue_completed.
17212
17213 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17214
17215 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17216 is error_mark_node.
17217
17218 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17219
17220 PR target/57261
17221 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17222 and Solaris 11+/x86 with gld.
17223 * configure: Regenerate.
17224
17225 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17226
17227 * expmed.c (expand_shift_1): Canonicalize rotates by
17228 constant bitsize / 2 to bitsize - 1.
17229 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17230 case ROTATERT>: Likewise.
17231
17232 Revert:
17233 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17234
17235 * config/i386/i386.md (rotateinv): New code attr.
17236 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17237 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17238 roll $31, %eax, etc.
17239
17240 2013-05-14 Richard Biener <rguenther@suse.de>
17241
17242 PR middle-end/57235
17243 * tree-eh.c (sink_clobbers): Give up for successors with
17244 multiple predecessors and no virtual uses.
17245
17246 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17247
17248 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17249 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17250
17251 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17252
17253 PR middle-end/57251
17254 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17255 the case when both op0 and op1 have VOIDmode.
17256
17257 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17258
17259 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17260 in multiply-accumulate mode.
17261
17262 2013-05-13 Guozhi Wei <carrot@google.com>
17263
17264 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17265
17266 2013-05-13 Kai Tietz <ktietz@redhat.com>
17267
17268 PR target/56975
17269 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17270 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17271 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17272 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17273 (ix86_expand_prologue): Likewise.
17274 (ix86_expand_split_stack_prologue): Likewise.
17275 (legitimate_pic_address_disp_p): Likewise.
17276 (legitimize_pic_address): Likewise.
17277 (legitimize_tls_address): Likewise.
17278 (legitimize_pe_coff_symbol): Likewise.
17279 (output_pic_addr_const): Likewise.
17280 (construct_plt_address): Likewise.
17281 (ix86_expand_call): Likewise.
17282 (x86_output_mi_thunk): Likewise.
17283 (x86_function_profiler): Likewise.
17284
17285 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17286
17287 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17288 similar switch cases.
17289 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17290 (aarch64_simd_mov_to_<mode>low): Delete.
17291 (aarch64_simd_mov_to_<mode>high): Delete.
17292 (move_lo_quad_<mode>): Add w<-r alternative.
17293 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17294 (aarch64_simd_mov_from_*): Update type attribute.
17295 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17296 statement.
17297
17298 2013-05-13 Jan Hubicka <jh@suse.cz>
17299
17300 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17301 * config/i386/i386.c (ix86_compute_frame_layout,
17302 ix86_expand_epilogue, emit_i387_cw_initialization,
17303 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17304 ix86_local_alignment): Fix use of size/speed predicates.
17305
17306 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17307
17308 PR tree-optimization/45216
17309 PR tree-optimization/57157
17310 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17311 the (-Y) & (B - 1) variant if OP is |.
17312 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17313 return shifted. Use (-op1) & (prec - 1) as other_amount
17314 instead of prec - op1.
17315
17316 2013-05-13 Martin Jambor <mjambor@suse.cz>
17317
17318 PR middle-end/42371
17319 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17320 (ipa_constant_data): New type.
17321 (ipa_jump_func): Use ipa_constant_data to hold information about
17322 constant jump functions.
17323 (ipa_get_jf_constant): Adjust to jump function type changes.
17324 (ipa_get_jf_constant_rdesc): New function.
17325 (ipa_param_descriptor): New field controlled_uses.
17326 (ipa_get_controlled_uses): New function.
17327 (ipa_set_controlled_uses): Likewise.
17328 * ipa-ref.h (ipa_find_reference): Declare.
17329 * ipa-prop.c (ipa_cst_ref_desc): New type.
17330 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17331 changes.
17332 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17333 New parameter cs. Adjust all callers.
17334 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17335 (remove_described_reference): New function.
17336 (jfunc_rdesc_usable): Likewise.
17337 (try_make_edge_direct_simple_call): Decrement controlled use count,
17338 attempt to remove reference if it hits zero.
17339 (combine_controlled_uses_counters): New function.
17340 (propagate_controlled_uses): Likewise.
17341 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17342 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17343 (ipa_print_node_params): Print described use counter.
17344 (ipa_write_jump_function): Adjust to jump function type changes.
17345 (ipa_read_jump_function): New parameter CS, pass it to
17346 ipa_set_jf_constant. Adjust caller.
17347 (ipa_write_node_info): Stream controlled use count
17348 (ipa_read_node_info): Likewise.
17349 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17350 asserting.
17351 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17352 count. Remove cloning-added reference if it reaches zero.
17353 * ipa-ref.c (ipa_find_reference): New function.
17354
17355 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17356
17357 * config/i386/i386.c (processor_target_table): Modified default
17358 alignment values for AMD BD and BT architectures.
17359
17360 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17361
17362 * tree-vect-generic.c (uniform_vector_p): Move ...
17363 * tree.c (uniform_vector_p): ... here.
17364 * tree.h (uniform_vector_p): Declare it.
17365 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17366 into a scalar.
17367
17368 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17369
17370 PR tree-optimization/57230
17371 * tree-ssa-strlen.c (handle_char_store): Record length for
17372 array store from STRING_CST.
17373
17374 PR tree-optimization/57230
17375 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17376 check.
17377
17378 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17379
17380 * config/epiphany/epiphany.c (epiphany_init): Check size of
17381 NUM_MODES_FOR_MODE_SWITCHING.
17382 (epiphany_expand_prologue):
17383 Remove CONFIG_REGNUM initial value handling code.
17384 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17385 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17386 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17387 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17388 Don't return 1 for FP_MODE_NONE.
17389 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17390 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17391 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17392 * config/epiphany/epiphany.md (save_config): New pattern.
17393
17394 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17395
17396 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17397
17398 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17399
17400 * config/i386/i386.md (memory): Handle sseishft1.
17401 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17402 (*vec_extractv2di_1): Ditto.
17403
17404 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17405
17406 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17407 saved registers.
17408
17409 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17410
17411 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17412 Add mthumb/march=armv7-a multilib.
17413 Add mthumb/march=armv7-r multilib.
17414 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17415
17416 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17417
17418 * config/v850/t-rtems: Add more multilibs.
17419
17420 2013-05-10 Richard Biener <rguenther@suse.de>
17421
17422 PR tree-optimization/57214
17423 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17424 not propagate from SSA names that occur in abnormal PHI nodes.
17425
17426 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17427
17428 * stor-layout.c (element_precision): New function.
17429 * machmode.h (element_precision): Declare it.
17430 * tree.c (build_minus_one_cst): New function.
17431 (element_precision): Likewise.
17432 * tree.h (build_minus_one_cst): Declare new function.
17433 (element_precision): Likewise.
17434 * fold-const.c (operand_equal_p): Use element_precision.
17435 (fold_binary_loc): Handle vector types.
17436 * convert.c (convert_to_integer): Use element_precision.
17437 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17438 separately.
17439
17440 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17441
17442 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17443 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17444 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17445 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17446 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17447 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17448 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17449 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17450 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17451 (Uuw8): New constraints.
17452 (Usb4): Move into alphabetical order.
17453 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17454 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17455 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17456 previously unnamed patterns.
17457 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17458 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17459 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17460 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17461 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17462 of specific lengths.
17463
17464 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17465
17466 * config/i386/i386.md (rotateinv): New code attr.
17467 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17468 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17469 roll $31, %eax, etc.
17470
17471 PR tree-optimization/45216
17472 PR tree-optimization/57157
17473 * tree-ssa-forwprop.c (simplify_rotate): New function.
17474 (ssa_forward_propagate_and_combine): Call it.
17475
17476 2013-05-10 Richard Biener <rguenther@suse.de>
17477
17478 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17479 disable peeling when we version for aliasing.
17480 (vector_alignment_reachable_p): Honor explicit user alignment.
17481 (vect_supportable_dr_alignment): Likewise.
17482 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17483 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17484 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17485 then peeling to arrange for the cost-model check to come first.
17486
17487 2013-05-10 Alan Modra <amodra@gmail.com>
17488
17489 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17490 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17491 * configure: Regenerate.
17492
17493 2013-05-10 Alan Modra <amodra@gmail.com>
17494
17495 PR target/55033
17496 * varasm.c (default_elf_select_section): Move !DECL_P check..
17497 (get_named_section): ..to here before calling get_section_name.
17498 Adjust assertion.
17499 (default_section_type_flags): Add DECL_P check.
17500 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17501 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17502
17503 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17504
17505 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17506 When using gen_stack_adjust_str with a register offset, add a
17507 REG_FRAME_RELATED_EXPR note.
17508
17509 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17510
17511 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17512 (*vec_extractv4si_zext_mem): Ditto.
17513 (*vec_extractv2di): Add 0->x and x->x alternatives.
17514 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17515 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17516
17517 2013-05-09 Jason Merrill <jason@redhat.com>
17518
17519 N3639 C++1y VLA support
17520 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17521 DECL_VALUE_EXPR.
17522
17523 * tree.c (build_constructor_va): New.
17524 * tree.h: Declare it.
17525
17526 2013-05-09 Martin Jambor <mjambor@suse.cz>
17527
17528 PR lto/57084
17529 * gimple-fold.c (canonicalize_constructor_val): Call
17530 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17531
17532 2013-05-09 Jan Hubicka <jh@suse.cz>
17533 Richard Biener <rguenther@suse.de>
17534
17535 PR lto/54095
17536 * symtab.c (symtab_make_decl_local): Do not add private names.
17537
17538 2013-05-09 Jan Hubicka <jh@suse.cz>
17539
17540 PR lto/54095
17541 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17542 (unlink_from_assembler_name_hash): Likewise.
17543 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17544 symtab_unregister_node, symtab_initialize_asm_name_hash,
17545 change_decl_assembler_name): Update.
17546
17547 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17548
17549 * config/aarch64/aarch64.md: New movtf split.
17550 (*movtf_aarch64): Update.
17551 (aarch64_movdi_tilow): Handle TF modes and rename to
17552 aarch64_movdi_<mode>low.
17553 (aarch64_movdi_tihigh): Handle TF modes and rename to
17554 aarch64_movdi_<mode>high
17555 (aarch64_movtihigh_di): Handle TF modes and rename to
17556 aarch64_mov<mode>high_di
17557 (aarch64_movtilow_di): Handle TF modes and rename to
17558 aarch64_mov<mode>low_di
17559 (aarch64_movtilow_tilow): Remove spurious whitespace.
17560 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17561 splits.
17562 (aarch64_print_operand): Update.
17563
17564 2013-05-09 Alan Modra <amodra@gmail.com>
17565
17566 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17567 powerpc64le.
17568 * configure: Regenerate.
17569
17570 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17571
17572 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17573 splitter preparation statements.
17574 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17575 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17576 adjust_address_nv.
17577
17578 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17579
17580 * gimple-ssa-strength-reduction.c (count_candidates): Change
17581 return value to int.
17582 (analyze_candidates_and_replace): Change type of length to int.
17583
17584 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17585
17586 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17587 (*vec_extract<mode>): Use VI12_128 mode iterator.
17588 (*vec_extract<mode>_mem): Ditto.
17589 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17590 attribute.
17591
17592 2013-05-08 Diego Novillo <dnovillo@google.com>
17593
17594 PR bootstrap/54659
17595
17596 Revert:
17597 2012-08-17 Diego Novillo <dnovillo@google.com>
17598
17599 PR bootstrap/54281
17600 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17601 * config.in: Regenerate.
17602 * configure: Regenerate.
17603 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
17604
17605 2013-05-08 Jan Hubicka <jh@suse.cz>
17606
17607 PR lto/54095
17608 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17609 * cgraph.h (symtab_node_base): Add unique_name.
17610 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17611 input_overwrite_node, input_varpool_node): Stream unique_name.
17612 * cgraphclones.c (cgraph_create_virtual_clone,
17613 cgraph_function_versioning): Set unique_name.
17614 * ipa.c (function_and_variable_visibility): Set unique_name.
17615
17616 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17617
17618 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17619 (alloc_cand_and_find_basis): Restrict conditional candidate
17620 processing to CAND_MULTs.
17621
17622 2013-05-08 Jan Hubicka <jh@suse.cz>
17623
17624 PR lto/54095
17625 lto-symtab.c (lto_symtab_symbol_p): New function.
17626 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17627 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17628 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17629 Skip static symbols.
17630
17631 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17632
17633 PR tree-optimization/57200
17634 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17635 Only call inform if the preceding warning_at returns true.
17636
17637 2013-05-07 Han Shen <shenhan@google.com>
17638
17639 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17640 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17641 * common.opt (fstack-protector-strong): New option.
17642 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17643 * doc/invoke.texi (Optimization Options): Document
17644 "-fstack-protector-strong".
17645 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17646
17647 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17648
17649 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17650
17651 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17652
17653 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17654 it is not lhs of the insn.
17655 (setup_reg_equiv): Remove insn having equiv memory which it is not
17656 lhs of the insn.
17657 * lra-constraints.c (process_address): Try to improve generation
17658 code for address base + disp.
17659 (lra_constraints): Make correct the code for checking insn setting
17660 up backward equivalence. Remove insn only if it is in the init
17661 insn list.
17662 * lra-eliminations.c (update_reg_eliminate): Change return value.
17663 (lra_eliminate): Use the result.
17664
17665 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17666
17667 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17668 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17669 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17670 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17671 register target operands.
17672 (*vec_extractv8hi_sse2): New pattern.
17673 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17674 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17675 (*vec_extract<mode>_mem): New insn and split pattern.
17676
17677 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17678
17679 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17680 (TARGET_ASAN_SHADOW_OFFSET): Define.
17681 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17682 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17683
17684 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17685
17686 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17687 (incr_vec_index): Return -1 if increment not found.
17688 (create_add_on_incoming_edge): Assert if increment not found.
17689 (record_increment): Limit number of increments recorded.
17690 (all_phi_incrs_profitable): Return false if an increment not found.
17691 (replace_profitable_candidates): Don't process increments that were
17692 not recorded.
17693 (analyze_candidates_and_replace): Limit size of incr_vec.
17694
17695 2013-05-07 Richard Biener <rguenther@suse.de>
17696
17697 * calls.c (special_function_p): setjmp-like functions are leaf.
17698 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17699 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17700
17701 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17702
17703 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17704 (aarch64_simd_mov<mode>): New expander.
17705 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17706 (aarch64_simd_mov_to_<mode>high): Likewise.
17707 (aarch64_simd_mov_from_<mode>low): Likewise.
17708 (aarch64_simd_mov_from_<mode>high): Likewise.
17709 (aarch64_dup_lane<mode>): Update.
17710 (aarch64_dup_lanedi): New instruction pattern.
17711 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17712 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17713
17714 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17715
17716 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17717 (replace_mult_candidate): Remove unnecessary argument; remove
17718 unnecessary parameter from call to introduce_cast_before_cand.
17719 (replace_unconditional_candidate): Remove unnecessary parameter
17720 from call to replace_mult_candidate.
17721 (replace_conditional_candidate): Likewise.
17722 (insert_initializers): Use make_temp_ssa_name.
17723 (introduce_cast_before_cand): Remove unnecessary argument; use
17724 make_temp_ssa_name.
17725 (replace_one_candidate): Remove unnecessary argument; remove
17726 unnecessary parameter from calls to introduce_cast_before_cand.
17727 (replace_profitable_candidates): Remove unnecessary parameters
17728 from calls to replace_one_candidate.
17729
17730 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17731
17732 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17733 phi def as possibly hiding a basis for a CAND_ADD whose operands
17734 have been commuted in the analysis.
17735 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17736
17737 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17738
17739 * config/aarch64/aarch64.md
17740 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17741 shift value between 0-4.
17742
17743 2013-05-07 Richard Biener <rguenther@suse.de>
17744
17745 * double-int.h (rshift): New overload.
17746 * double-int.c (rshift): New function.
17747 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17748 (create_reference_ops_from_ref): Remove.
17749 (vn_reference_insert): Use shared ops for constructing the
17750 reference and copy it.
17751
17752 2013-05-07 Richard Biener <rguenther@suse.de>
17753
17754 PR middle-end/57190
17755 * tree-eh.c (sink_clobbers): Properly propagate
17756 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17757
17758 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17759
17760 PR tree-optimization/57149
17761 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17762 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17763 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17764 uninit_undefined_value_p instead of ssa_undefined_value_p.
17765
17766 PR debug/57184
17767 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17768 for modifier == EXPAND_INITIALIZER.
17769
17770 2013-05-07 Anton Blanchard <anton@samba.org>
17771
17772 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17773 for powerpc64 little endian.
17774 * configure: Regenerate.
17775
17776 2013-05-06 Graham Stott <grahams@btinternet.com>
17777
17778 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17779 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17780 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17781 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17782
17783 2013-05-06 Graham Stott <grahams@btinternet.com>
17784
17785 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17786 codes which allow non-lvalues.
17787
17788 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17789
17790 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17791 components are all 1s.
17792 (integer_minus_onep): New function.
17793 * tree.h (integer_minus_onep): Declare it.
17794 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17795 integer_minus_onep instead of integer_all_onesp.
17796
17797 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17798
17799 PR target/52933
17800 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17801 variations of these patterns.
17802
17803 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17804
17805 * config/i386/i386.md (isa): Add x64_sse4 member.
17806 (enabled): Handle x64_sse4.
17807 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17808 instruction for 64bit SSE4_1 targets. Update insn attributes.
17809 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17810 instruction for SSE4_1 targets. Update insn attributes.
17811 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17812 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17813 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17814 const_1 selector.
17815 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17816 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17817 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17818
17819 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17820
17821 PR target/57108
17822 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17823
17824 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17825
17826 * final.c (do_assembler_dialects): Don't handle curly braces and
17827 vertical bar escaped by % as dialect delimiters.
17828 (output_asm_insn): Print curly braces and vertical bar if escaped
17829 by % and ASSEMBLER_DIALECT defined.
17830 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17831 * doc/tm.texi: Regenerated.
17832
17833 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17834
17835 * config/mips/mips.c: Include tree-pass.h.
17836 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17837 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17838 (pass_mips_machine_reorg2): New machine specific pass.
17839 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17840 (mips_option_override): Register the new pass.
17841 * rtl.h (cleanup_barriers): Remove prototype.
17842 (dbr_schedule): Likewise.
17843 * jump.c (cleanup_barriers): Make static.
17844 * reorg.c (dbr_schedule): Likewise.
17845
17846 2013-05-06 Richard Biener <rguenther@suse.de>
17847
17848 PR tree-optimization/57185
17849 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17850 SSA names properly.
17851
17852 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17853
17854 PR target/57106
17855 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17856 between "first_arg" and "insn", not "last" and "insn".
17857
17858 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17859
17860 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17861 (find_candidates_in_block): Re-enable slsr_process_phi.
17862 (create_phi_basis): Fix double counting of candidate adjustment.
17863
17864 2013-05-06 Richard Biener <rguenther@suse.de>
17865
17866 PR middle-end/57147
17867 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17868 the edge is also fallthru, preserve it and just clear the
17869 abnormal flag.
17870 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17871 also complex, preserve that and just clear the fallthru flag.
17872 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17873 update virtual operands.
17874
17875 2013-05-06 Alan Modra <amodra@gmail.com>
17876
17877 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17878 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17879 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17880 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17881 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17882 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17883
17884 2013-05-06 Alan Modra <amodra@gmail.com>
17885
17886 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17887 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17888 (DEFAULT_ASM_ENDIAN): Define.
17889 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17890 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17891 Update -K PIC clause from sysv4.h.
17892 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17893 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17894
17895 2013-05-06 Alan Modra <amodra@gmail.com>
17896
17897 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17898 twice for little-endian.
17899 (ashrdi3_no_power, ashrdi3): Support little-endian.
17900
17901 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17902
17903 PR target/55303
17904 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17905 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17906 related expanders.
17907 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17908 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17909 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17910 New predicates.
17911
17912 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17913 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17914
17915 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17916 * config/pa/pa.opt: Make mbig-switch a no-op.
17917 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17918 (CASE_VECTOR_MODE): Always return SImode.
17919 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17920 for the !TARGET_BIG_SWITCH case.
17921 * config/pa/pa-linux.h: Likewise.
17922 * config/pa/pa-openbsd.h: Likewise.
17923 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17924 * config/pa/pa.md (short_jump): Remove define_insn.
17925 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17926 (casesi0): Remove define_insn.
17927 (type): Remove btable_branch.
17928 (pa_combine_type): Likewise.
17929 (in_nullified_branch_delay): Likewise.
17930 (in_call_delay): Likewise.
17931 (define_delay): Likewise.
17932 (define_insn_reservation "Z3"): Likewise.
17933 (define_insn_reservation "Z4"): Likewise.
17934 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17935 (pa_adjust_insn_length): Remove adjustment for btable branches.
17936 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17937 and mno-big-switch
17938
17939 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17940
17941 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17942 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17943 Add m->r,x alternatives.
17944 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17945 splitters using SWI48x mode iterator.
17946 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17947 TARGET_64BIT. Add m->x alternative.
17948 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17949 Add o->x alternative. Enable for TARGET_SSE.
17950 (sse_storeq): Remove expander.
17951 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17952 with memory input operand.
17953 (*vec_extractv2di_1 splitter): New.
17954 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17955 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17956
17957 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17958
17959 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17960 (INT_LOWPART): Delete.
17961 (extract_MB): Adjust.
17962 (extract_ME): Adjust.
17963 (print_operand): Adjust.
17964
17965 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17966
17967 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17968 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17969 (reg_or_logical_cint_operand, easy_fp_constant,
17970 logical_const_operand): Delete "CONST_DOUBLE" case.
17971 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17972 "HOST_BITS_PER_WIDE_INT == 64" test.
17973 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17974 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17975 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17976 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17977 test.
17978 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17979 CONST_DOUBLE DImode/VOIDmode case.
17980 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17981 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17982 CONST_DOUBLE VOIDmode case.
17983 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17984 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17985 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17986 Delete CONST_DOUBLE case.
17987 (splitters for mov FMOVE64 const_double): Delete
17988 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17989 "HOST_BITS_PER_WIDE_INT >= 64" test.
17990 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17991 case.
17992 (mov DI const_double): Delete.
17993
17994 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17995
17996 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17997 on op shows all bits zero in mode of a lowpart subreg, return zero.
17998
17999 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18000
18001 PR target/57150
18002 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18003 to save TFmode registers and DImode to save TImode registers for
18004 caller save operations.
18005 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18006 mark being partially clobbered since they only use the first
18007 double word.
18008
18009 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18010 and TDmode only use the upper 64-bits of each VSX register.
18011
18012 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18013
18014 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18015 (find_candidates_in_block): Disable slsr_process_phi.
18016
18017 2013-05-03 Guozhi Wei <carrot@google.com>
18018
18019 * coverage.c (coverage_obj_init): Move the construction of gcov
18020 constructor to ...
18021 (build_init_ctor): ... here.
18022
18023 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18024
18025 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18026 (slsr_cand_d): Redefine def_phi.
18027 (stride_status, phi_adjust_status, count_phis_status): New enums.
18028 (find_phi_def): New.
18029 (find_basis_for_base_expr): New.
18030 (find_basis_for_candidate): Handle hidden bases.
18031 (alloc_cand_and_find_basis): Handle phi candidates.
18032 (slsr_process_phi): New.
18033 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18034 (create_mul_imm_cand): Likewise.
18035 (create_add_ssa_cand): Exclude phi base candidates.
18036 (create_add_imm_cand): Likewise.
18037 (slsr_process_cast): Likewise.
18038 (slsr_process_copy): Likewise.
18039 (find_candidates_in_block): Handle phi candidates.
18040 (dump_candidate): Likewise.
18041 (unconditional_cands): Delete.
18042 (unconditional_cands_with_known_stride_p): Delete.
18043 (phi_dependent_cand_p): New.
18044 (cand_increment): Handle phi-dependent candidates.
18045 (replace_dependent): Delete.
18046 (replace_mult_candidate): New.
18047 (replace_unconditional_candidate): New.
18048 (incr_vec_index): Move to avoid forward reference.
18049 (create_add_on_incoming_edge): New.
18050 (create_phi_basis): New.
18051 (replace_dependents): Delete.
18052 (replace_conditional_candidate): New.
18053 (phi_add_costs): New.
18054 (replace_uncond_cands_and_profitable_phis): New.
18055 (record_increment): Handle phi adjustments.
18056 (record_phi_increments): New.
18057 (record_increments): Handle phi adjustments.
18058 (phi_incr_cost): New.
18059 (lowest_cost_path): Handle phis.
18060 (total_savings): Likewise.
18061 (analyze_increments): Likewise.
18062 (ncd_with_phi): New.
18063 (ncd_of_cand_and_phis): New.
18064 (nearest_common_dominator_for_cands): Handle phi increments.
18065 (all_phi_incrs_profitable): New.
18066 (replace_profitable_candidates): Handle phi-dependent candidates.
18067 (analyze_candidates_and_replace): Likewise.
18068
18069 2013-05-03 Teresa Johnson <tejohnson@google.com>
18070
18071 PR bootstrap/57154
18072 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18073 do not exceed REG_BR_PROB_BASE.
18074
18075 2013-05-03 Jeff Law <law@redhat.com>
18076
18077 PR tree-optimization/57144
18078 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18079 operand of the condition will bit into the new type when eliminating
18080 a cast feeding a condition.
18081
18082 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18083
18084 PR rtl-optimization/57130
18085 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18086 of COMPARE as in_code to the recursive call if needed.
18087
18088 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18089
18090 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18091 (enabled): Handle new members.
18092 * config/i386/sse.md (*vec_concatv2si): Merge from
18093 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18094 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18095
18096 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18097
18098 PR tree-optimization/57027
18099 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18100 for fnms opportunity, check we got the prerequisite kind
18101 of tree / gimple before using accessor functions.
18102
18103 2013-05-03 Richard Biener <rguenther@suse.de>
18104
18105 * double-int.h (lshift): New overload without precision
18106 and arith argument.
18107 (operator *=, operator +=, operator -=): Move ...
18108 * double-int.c (operator *=, operator +=, operator -=): ... here
18109 and implement more efficiently.
18110 (mul_double_with_sign): Remove.
18111 (lshift_double): Adjust to take unsinged shift argument, push
18112 dispatching code to callers.
18113 (mul_double_wide_with_sign): Add early out for callers that
18114 are not interested in high parts or overflow.
18115 (lshift): New function.
18116 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18117 dispatch code here.
18118 (lrotate, rrotate): Use logical shifts.
18119 * expr.c (get_inner_reference): Use lshift.
18120 * fixed-value.c (do_fixed_divide): Likewise.
18121 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18122 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18123 (indirect_refs_may_alias_p): Likewise.
18124 (stmt_kills_ref_p_1): Likewise.
18125
18126 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18127
18128 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18129
18130 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18131
18132 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18133 scalar form of FABD instruction.
18134
18135 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18136
18137 * lra-constraints.c (process_alt_operands): Add checking alt
18138 number to choose the best alternative.
18139
18140 2013-05-02 Richard Biener <rguenther@suse.de>
18141
18142 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18143 bitmap and its handling.
18144 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18145
18146 2013-05-02 Richard Biener <rguenther@suse.de>
18147
18148 PR middle-end/57140
18149 * tree-inline.c (copy_loops): Properly handle removed loops.
18150 (copy_cfg_body): Mark destination loops for fixup if source
18151 loops needed fixup.
18152
18153 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18154
18155 PR target/56732
18156 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18157 generating simple_return for naked functions.
18158
18159 2013-05-02 Martin Jambor <mjambor@suse.cz>
18160
18161 PR middle-end/56988
18162 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18163 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18164 flags match.
18165 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18166 ipa_agg_replacement_value structures.
18167 (known_aggs_to_agg_replacement_list): Likewise.
18168 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18169 (read_agg_replacement_chain): Likewise.
18170 (ipcp_transform_function): Also check that by_ref flags match.
18171
18172 2013-05-02 Richard Biener <rguenther@suse.de>
18173
18174 * graphds.h (struct graph): Add obstack member.
18175 * graphds.c (new_graph): Initialize obstack and allocate
18176 vertices from it.
18177 (add_edge): Allocate edge from the obstack.
18178 (free_graph): Free the obstack instead of all edges and vertices.
18179
18180 2013-05-02 Teresa Johnson <tejohnson@google.com>
18181
18182 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18183 divides.
18184 * cfg.c (update_bb_profile_for_threading): Ditto.
18185 * tree-inline.c (copy_bb): Ditto.
18186 (copy_edges_for_bb): Ditto.
18187 (initialize_cfun): Ditto.
18188 (copy_cfg_body): Ditto.
18189 (expand_call_inline): Ditto.
18190 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18191 (estimate_node_size_and_time): Ditto.
18192 (inline_merge_summary): Ditto.
18193 * cgraphclones.c (cgraph_clone_edge): Ditto.
18194 (cgraph_clone_node): Ditto.
18195 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18196 (compute_trg_info): Ditto.
18197
18198 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18199
18200 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18201 S reg when fp attribute set.
18202 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18203
18204 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18205
18206 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18207 New pattern.
18208 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18209 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18210 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18211
18212 2013-05-02 Richard Biener <rguenther@suse.de>
18213
18214 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18215 (struct instantiate_cache_entry): New type.
18216 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18217 (struct instantiate_cache_type): New type.
18218 (set_instantiated_value, get_instantiated_value): Remove.
18219 (get_instantiated_value_entry): New function.
18220 (instantiate_scev_name): Use the new cache and adjust.
18221 (instantiate_scev_poly): Adjust.
18222 (instantiate_scev_binary): Likewise.
18223 (instantiate_array_ref): Likewise.
18224 (instantiate_scev_convert): Likewise.
18225 (instantiate_scev_not): Likewise.
18226 (instantiate_scev_3): Likewise.
18227 (instantiate_scev_2): Likewise.
18228 (instantiate_scev_r): Likewise.
18229 (instantiate_scev): Likewise.
18230 (resolve_mixers): Likewise.
18231
18232 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18233
18234 PR target/57091
18235 * lra-constraints.c (best_small_class_operands_num): Remove.
18236 (process_alt_operands): Remove small_class_operands_num. Take
18237 small classes operands into losers and only if the operand is not
18238 matched. Modify debugging output.
18239 (curr_insn_transform): Remove best_small_class_operands_num.
18240 Print insn name.
18241
18242 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18243
18244 * config/aarch64/aarch64-builtins.c
18245 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18246 * config/aarch64/aarch64-simd-builtins.def
18247 (reduc_splus_): Add new modes.
18248 (reduc_uplus_): New.
18249 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18250 (reduc_uplus_v4sf): Likewise.
18251 (reduc_splus_v4sf): Likewise.
18252 (aarch64_addv<mode>): Likewise.
18253 (reduc_uplus_<mode>): Likewise.
18254 (reduc_splus_<mode>): Likewise.
18255 (aarch64_addvv2di): Likewise.
18256 (reduc_uplus_v2di): Likewise.
18257 (reduc_splus_v2di): Likewise.
18258 (aarch64_addvv2si): Likewise.
18259 (reduc_uplus_v2si): Likewise.
18260 (reduc_splus_v2si): Likewise.
18261 (reduc_<sur>plus_<mode>): New.
18262 (reduc_<sur>plus_v2di): Likewise.
18263 (reduc_<sur>plus_v2si): Likewise.
18264 (reduc_<sur>plus_v4sf): Likewise.
18265 (aarch64_addpv4sf): Likewise.
18266 * config/aarch64/arm_neon.h
18267 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18268 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18269 add UNSPEC_SADDV, UNSPEC_UADDV.
18270 (SUADDV): New.
18271 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18272
18273 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18274
18275 * config/aarch64/arm_neon.h
18276 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18277
18278 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18279
18280 * config/aarch64/aarch64-builtins
18281 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18282
18283 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18284
18285 * config/aarch64/aarch64-simd-builtins.def
18286 (reduc_smax_): New.
18287 (reduc_smin_): Likewise.
18288 (reduc_umax_): Likewise.
18289 (reduc_umin_): Likewise.
18290 (reduc_smax_nan_): Likewise.
18291 (reduc_smin_nan_): Likewise.
18292 (fmax): Remove.
18293 (fmin): Likewise.
18294 (smax): Update for V2SF, V4SF and V2DF modes.
18295 (smin): Likewise.
18296 (smax_nan): New.
18297 (smin_nan): Likewise.
18298 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18299 (<su><maxmin><mode>3): ...This, refactor.
18300 (s<maxmin><mode>3): New.
18301 (<maxmin_uns><mode>3): Likewise.
18302 (reduc_<maxmin_uns>_<mode>): Refactor.
18303 (reduc_<maxmin_uns>_v4sf): Likewise.
18304 (reduc_<maxmin_uns>_v2si): Likewise.
18305 (aarch64_<fmaxmin><mode>: Remove.
18306 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18307 new builtin names.
18308 (vmin<q>_f<32,64>): Likewise.
18309 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18310 (FMAXMIN): New.
18311 (su): Add mappings for smax, smin, umax, umin.
18312 (maxmin): New.
18313 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18314 (FMAXMIN): Rename as...
18315 (FMAXMIN_UNS): ...This.
18316 (maxminv): Remove.
18317 (fmaxminv): Likewise.
18318 (fmaxmin): Likewise.
18319 (maxmin_uns): New.
18320 (maxmin_uns_op): Likewise.
18321
18322 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18323
18324 * config/aarch64/arm_neon.h
18325 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18326 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18327 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18328
18329 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18330
18331 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18332 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18333
18334 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18335
18336 * config/aarch64/aarch64-simd.md
18337 (vcond<mode>_internal): Handle special cases for constant masks.
18338 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18339 (vcondu<mode><mode>): Likewise.
18340 (vcond<v_cmp_result><mode>): New.
18341
18342 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18343
18344 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18345 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18346 * config/aarch64/aarch64-simd-builtins.def
18347 (cmeq): Update to BUILTIN_VALLDI.
18348 (cmgt): Likewise.
18349 (cmge): Likewise.
18350 (cmle): Likewise.
18351 (cmlt): Likewise.
18352 * config/aarch64/arm_neon.h
18353 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18354 to builtins or C as appropriate.
18355
18356 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18357
18358 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18359 (cmgeu): ...This.
18360 (cmhi): Rename to...
18361 (cmgtu): ...This.
18362 * config/aarch64/aarch64-simd.md
18363 (simd_mode): Add SF.
18364 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18365 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18366 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18367 (cstore<mode>_neg): ...This.
18368 * config/aarch64/iterators.md
18369 (VALLF): new.
18370 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18371 (COMPARISONS): New.
18372 (UCOMPARISONS): Likewise.
18373 (optab): Add missing comparisons.
18374 (n_optab): New.
18375 (cmp_1): Likewise.
18376 (cmp_2): Likewise.
18377 (CMP): Likewise.
18378 (cmp): Remove.
18379 (VCMP_S): Likewise.
18380 (VCMP_U): Likewise.
18381 (V_cmp_result): Add DF, SF modes.
18382 (v_cmp_result): Likewise.
18383 (v): Likewise.
18384 (vmtype): Likewise.
18385 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18386
18387 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18388
18389 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18390 define_insn to define_insn_and_split.
18391 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18392 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18393 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18394 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18395 (thumb2_negscc): Likewise.
18396
18397 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18398
18399 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18400
18401 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18402
18403 * config/arm/thumb2.md: Remove trailing whitespaces.
18404
18405 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18406
18407 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18408 Use gen_int_mode rather than GEN_INT.
18409
18410 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18411
18412 * value-prof.c (stream_in_histogram_value): Remove the strayed
18413 debug_gimple_stmt.
18414
18415 2013-04-30 Richard Biener <rguenther@suse.de>
18416
18417 PR middle-end/57122
18418 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18419
18420 2013-04-30 Richard Biener <rguenther@suse.de>
18421
18422 PR middle-end/57107
18423 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18424
18425 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18426
18427 PR rtl-optimization/56957
18428 PR rtl-optimization/57105
18429 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18430 variable. Use just INSN_UID for determining whether an insn
18431 should be only disconnected from the insn stream.
18432 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18433
18434 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18435
18436 PR tree-optimization/57104
18437 * tsan.c (instrument_expr): Don't instrument accesses to
18438 DECL_HARD_REGISTER VAR_DECLs.
18439
18440 2013-04-30 Richard Biener <rguenther@suse.de>
18441
18442 * function.h (loops_for_fn): New inline function.
18443 (set_loops_for_fn): Likewise.
18444 * cfgloop.h (place_new_loop): Add struct function parameter.
18445 (get_loop): Likewise.
18446 (get_loops): Likewise.
18447 (number_of_loops): Likewise.
18448 (fel_next): Adjust.
18449 (fel_init): Likewise.
18450 * cfg.c (get_loop_copy): Adjust.
18451 * cfgloop.c (flow_loops_dump): Likewise.
18452 (record_loop_exits): Likewise.
18453 (verify_loop_structure): Likewise.
18454 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18455 (estimate_reg_pressure_cost): Likewise.
18456 (mark_loop_exit_edges): Likewise.
18457 * cfgloopmanip.c (place_new_loop): Likewise.
18458 (add_loop): Likewise.
18459 (duplicate_loop): Likewise.
18460 * graph.c (draw_cfg_nodes): Likewise.
18461 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18462 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18463 (extract_affine_chrec): Likewise.
18464 (build_scop_iteration_domain): Likewise.
18465 * graphite.c (graphite_initialize): Likewise.
18466 * ira-build.c (create_loop_tree_nodes): Likewise.
18467 (more_one_region_p): Likewise.
18468 (rebuild_regno_allocno_maps): Likewise.
18469 (mark_loops_for_removal): Likewise.
18470 (mark_all_loops_for_removal): Likewise.
18471 (remove_unnecessary_regions): Likewise.
18472 (ira_build): Likewise.
18473 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18474 * loop-init.c (fix_loop_structure): Likewise.
18475 (gate_rtl_move_loop_invariants): Likewise.
18476 (gate_rtl_unswitch): Likewise.
18477 (gate_rtl_unroll_and_peel_loops): Likewise.
18478 (rtl_doloop): Likewise.
18479 * lto-streamer-in.c (input_cfg): Likewise.
18480 * lto-streamer-out.c (output_cfg): Likewise.
18481 * modulo-sched.c (sms_schedule): Likewise.
18482 * predict.c (tree_estimate_probability): Likewise.
18483 (tree_estimate_probability_driver): Likewise.
18484 (estimate_loops): Likewise.
18485 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18486 (move_sese_region_to_fn): Likewise.
18487 (debug_loop_num): Likewise.
18488 * tree-chrec.c (chrec_evaluate): Likewise.
18489 (hide_evolution_in_other_loops_than_loop): Likewise.
18490 (chrec_component_in_loop_num): Likewise.
18491 (reset_evolution_in_loop): Likewise.
18492 (evolution_function_is_invariant_rec_p): Likewise.
18493 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18494 * tree-inline.c (copy_loops): Likewise.
18495 (copy_cfg_body): Likewise.
18496 (tree_function_versioning): Likewise.
18497 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18498 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18499 Likewise.
18500 (add_to_evolution_1): Likewise.
18501 (scev_const_prop): Likewise.
18502 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18503 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18504 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18505 (tree_ssa_lim_initialize): Likewise.
18506 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18507 (verify_loop_closed_ssa): Likewise.
18508 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18509 (tree_ssa_loop_im): Likewise.
18510 (tree_ssa_loop_unswitch): Likewise.
18511 (tree_vectorize): Likewise.
18512 (check_data_deps): Likewise.
18513 (tree_ssa_loop_ivcanon): Likewise.
18514 (tree_ssa_loop_bounds): Likewise.
18515 (tree_complete_unroll): Likewise.
18516 (tree_complete_unroll_inner): Likewise.
18517 (tree_parallelize_loops): Likewise.
18518 (tree_ssa_loop_prefetch): Likewise.
18519 (tree_ssa_loop_ivopts): Likewise.
18520 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18521 * tree-vectorizer.c (vectorize_loops): Likewise.
18522
18523 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18524
18525 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18526 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18527 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18528 with EABI_LINK_SPEC.
18529
18530 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18531
18532 PR target/44578
18533 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18534 alternative.
18535
18536 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18537
18538 PR target/57097
18539 * lra-constraints.c (process_alt_operands): Discourage a bit more
18540 using memory for pseudos. Print cost dump for alternatives.
18541 Modify cost values for conflicts with early clobbers.
18542 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18543
18544 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18545
18546 PR target/57098
18547 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18548
18549 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18550
18551 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18552 from/to S register.
18553 (movdi_aarch64): Support LDR/STR from/to D register.
18554
18555 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18556
18557 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18558 or higher by default.
18559
18560 2013-04-29 Richard Biener <rguenther@suse.de>
18561
18562 PR middle-end/57075
18563 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18564 even if not adding abnormal edges for calls that can make
18565 abnormal gotos.
18566
18567 2013-04-29 Richard Biener <rguenther@suse.de>
18568
18569 PR middle-end/57103
18570 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18571 TREE_BLOCK.
18572 (move_stmt_r): Remove redundant checking.
18573
18574 2013-04-29 Teresa Johnson <tejohnson@google.com>
18575
18576 PR bootstrap/57077
18577 * basic-block.h (apply_scale): New function.
18578 (apply_probability): Use apply_scale.
18579 * gimple-streamer-in.c (input_bb): Ditto.
18580 * lto-streamer-in.c (input_cfg): Ditto.
18581 * lto-cgraph.c (merge_profile_summaries): Ditto.
18582 * tree-optimize.c (execute_fixup_cfg): Ditto.
18583 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18584 (copy_edges_for_bb): Ditto.
18585 (copy_cfg_body): Ditto.
18586
18587 2013-04-29 Tom de Vries <tom@codesourcery.com>
18588
18589 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18590 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18591 (tail_merge_optimize): Handle current_loops == NULL.
18592
18593 2013-04-26 Jeff Law <law@redhat.com>
18594
18595 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18596 (simplify_cond_using_ranges): Generalize code to simplify
18597 COND_EXPRs where one argument is a constant and the other
18598 is an SSA_NAME created by an integral type conversion.
18599
18600 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18601
18602 * config/arm/arm.md (store_minmaxsi): Use only when
18603 optimize_insn_for_size_p.
18604
18605 2013-04-29 Christian Bruel <christian.bruel@st.com>
18606
18607 PR target/57108
18608 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18609
18610 2013-04-29 Richard Biener <rguenther@suse.de>
18611
18612 PR middle-end/57089
18613 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18614 loop tree make sure to schedule a fixup for the child as well.
18615 (expand_omp_for_generic): Properly add loops.
18616 (expand_omp_for_static_nochunk): Likewise.
18617 (expand_omp_for_static_chunk): Likewise.
18618 (expand_omp_for): For the degenerate case fixup loops.
18619 (expand_omp_sections): Fix default bb placement in loops.
18620 (expand_omp_atomic_pipeline): Properly add loops.
18621
18622 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18623
18624 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18625
18626 2013-04-29 Tom de Vries <tom@codesourcery.com>
18627
18628 * tree-ssa-tail-merge.c: Update header comment.
18629
18630 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18631
18632 * config/aarch64/arm_neon.h
18633 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18634 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18635 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18636 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18637 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18638 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18639 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18640 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18641
18642 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18643
18644 * config/aarch64/aarch64-simd.md
18645 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18646 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18647 fix_trunc, fixuns_trunc.
18648 (ftrunc<VDQF:mode>2): New.
18649 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18650 (fix_trunc_optab): New.
18651
18652 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18653
18654 * config/aarch64/aarch64-builtins.c
18655 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18656 iceilf, lround, iroundf.
18657
18658 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18659
18660 PR target/54349
18661 * config/i386/i386.h (enum ix86_tune_indices)
18662 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18663 New, split from X86_TUNE_INTER_UNIT_MOVES.
18664 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18665 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18666 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18667 (TARGET_INTER_UNIT_MOVES): Remove.
18668 * config/i386/i386.c (initial_ix86_tune_features): Update.
18669 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18670 (ix86_expand_convert_uns_didf_sse): Use
18671 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18672 (ix86_expand_vector_init_one_nonzero): Ditto.
18673 (ix86_expand_vector_init_interleave): Ditto.
18674 (inline_secondary_memory_needed): Return true for moves from SSE class
18675 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18676 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18677 * config/i386/constraints.md (Yi, Ym): Depend on
18678 TARGET_INTER_UNIT_MOVES_TO_VEC.
18679 (Yj, Yn): New constraints.
18680 * config/i386/i386.md (*movdi_internal): Change constraints of
18681 operand 1 from Yi to Yj and from Ym to Yn.
18682 (*movsi_internal): Ditto.
18683 (*movdf_internal): Ditto.
18684 (*movsf_internal): Ditto.
18685 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18686 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18687 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18688 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18689 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18690 * config/i386/sse.md (movdi_to_sse): Ditto.
18691 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18692 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18693 TARGET_INTER_UNIT_MOVES.
18694 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18695 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18696 instead of TARGET_INTER_UNIT_MOVES.
18697 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18698 operand 1 from Yi to Yj and from Ym to Yn.
18699
18700 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18701
18702 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18703 (float_truncate_hi_): Likewise.
18704 (float_extend_lo_): Likewise.
18705 (float_truncate_lo_): Likewise.
18706 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18707 (aarch64_float_extend_lo_v2df): Likewise.
18708 (vec_unpacks_hi_v4sf): Likewise.
18709 (aarch64_float_truncate_lo_v2sf): Likewise.
18710 (aarch64_float_truncate_hi_v4sf): Likewise.
18711 (vec_pack_trunc_v2df): Likewise.
18712 (vec_pack_trunc_df): Likewise.
18713
18714 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18715
18716 * config/aarch64/aarch64-builtins.c
18717 (aarch64_fold_builtin): Fold float conversions.
18718 * config/aarch64/aarch64-simd-builtins.def
18719 (floatv2si, floatv4si, floatv2di): New.
18720 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18721 * config/aarch64/aarch64-simd.md
18722 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18723 * config/aarch64/iterators.md (FLOATUORS): New.
18724 (optab): Add float, floatuns.
18725 (su_optab): Likewise.
18726
18727 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18728
18729 * config/aarch64/aarch64-builtins.c
18730 (aarch64_builtin_vectorized_function): Use new names for
18731 fcvt builtins.
18732 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18733 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18734 (fcvtzu): Split as...
18735 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18736 (fcvtas): Split as...
18737 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18738 (fcvtau): Split as...
18739 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18740 (fcvtps): Split as...
18741 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18742 (fcvtpu): Split as...
18743 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18744 (fcvtms): Split as...
18745 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18746 (fcvtmu): Split as...
18747 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18748 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18749 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18750 (lfrintnusf, lfrintnudf): Likewise.
18751 * config/aarch64/aarch64-simd.md
18752 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18753 define_insn.
18754 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18755 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18756 (fcvt_pattern): Likewise.
18757
18758 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18759
18760 * config/aarch64/aarch64-simd.md
18761 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18762 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18763
18764 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18765
18766 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18767 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18768 (vrnd<a,m,n,p>_f32): Implement using builtins.
18769 (vrnd<i,x><q>_f<32, 64>): New.
18770
18771 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18772
18773 * config/aarch64/aarch64-builtins.c
18774 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18775 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18776 (frintz): Rename to...
18777 (btrunc): ...this.
18778 (frintp): Rename to...
18779 (ceil): ...this.
18780 (frintm): Rename to...
18781 (floor): ...this.
18782 (frinti): Rename to...
18783 (nearbyint): ...this.
18784 (frintx): Rename to...
18785 (rint): ...this.
18786 (frinta): Rename to...
18787 (round): ...this.
18788 * config/aarch64/aarch64-simd.md
18789 (aarch64_frint<frint_suffix><mode>): Delete.
18790 (<frint_pattern><mode>2): Convert to insn.
18791 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18792 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18793 (frint_pattern): Likewise.
18794 (frint_suffix): Likewise.
18795
18796 2013-04-29 Richard Biener <rguenther@suse.de>
18797
18798 PR tree-optimization/57081
18799 * loop-init.c: Include tree-flow.h.
18800 (loop_optimizer_finalize): Free number of iteration estimates.
18801 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18802
18803 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18804
18805 PR tree-optimization/57083
18806 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18807 non-singleton shift count range, zero extend low_bound for uns case.
18808
18809 * config/i386/predicates.md (general_vector_operand): New predicate.
18810 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18811 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18812 if they aren't nonimmediate operands. If their original values
18813 satisfy const_vector_equal_evenodd_p, don't shift them.
18814 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18815 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18816 if not nonimmediate_operand.
18817 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18818 instead of register_operand.
18819 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18820
18821 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18822
18823 * stor-layout.c (finalize_size_functions): Allocate a structure and
18824 reset cfun before dumping the functions.
18825
18826 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18827
18828 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18829
18830 PR target/56866
18831 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18832 use xop_pmacsdqh if uns_p.
18833 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18834 the immediate rotate count.
18835
18836 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18837
18838 * rtl.h (struct rtx_def): Add comment for field jump.
18839 (LRA_SUBREG_P): New macro.
18840 * recog.c (register_operand): Check LRA_SUBREG_P.
18841 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18842 stack frame.
18843 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18844 (lra_final_code_change): Skip subreg change for operators.
18845 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18846 if there are no operand changes.
18847 * lra-constraints.c (curr_insn_set): New.
18848 (match_reload): Set LRA_SUBREG_P.
18849 (emit_spill_move): Ditto.
18850 (check_and_process_move): Use curr_insn_set. Process only single
18851 set insns. Don't initialize sec_mem_p and change_p.
18852 (simplify_operand_subreg): Use LRA_SUBREG_P.
18853 (reg_in_class_p): New function.
18854 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18855 of #ifdef. Add code to remove cycling.
18856 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18857 non-null disp. Reload inner instead of disp when base and index
18858 are null. Try to put lo_sum into register.
18859 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18860 (check_and_process_move): Move code for move cost check to
18861 simple_move_p. Remove equiv_substitution.
18862 (simple_move_p): New function.
18863 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18864 curr_insn_set. Call check_and_process_move only for single set
18865 insns. Use the new function. Move call of check_and_process_move
18866 after operand equiv substitution and address process.
18867
18868 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18869
18870 PR go/57045
18871 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18872 with nonlocal goto receivers or returns twice calls, ignore
18873 unininitialized values from abnormal edges to nl goto receiver
18874 or returns twice call.
18875
18876 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18877
18878 PR tree-optimization/57051
18879 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18880 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18881 bitsize.
18882
18883 2013-04-26 Richard Biener <rguenther@suse.de>
18884
18885 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18886 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18887 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18888 (fixup_loop_arrays_after_move): New function.
18889 (move_sese_region_to_fn): Properly outline the loop tree parts
18890 of the SESE region.
18891
18892 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18893
18894 * config/i386/i386.md (type, unit): Fix long lines.
18895
18896 2013-04-26 Richard Biener <rguenther@suse.de>
18897
18898 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18899 (lto-streamer-out.o): Likewise.
18900 * cfgloop.c (init_loops_structure): Export, add struct function
18901 argument and adjust.
18902 (flow_loops_find): Adjust.
18903 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18904 (init_loops_structure): Declare.
18905 * lto-streamer-in.c: Include cfgloop.h.
18906 (input_cfg): Input the loop tree.
18907 * lto-streamer-out.c: Include cfgloop.h.
18908 (output_cfg): Output the loop tree.
18909 (output_struct_function_base): Do not drop PROP_loops.
18910
18911 2013-03-26 Richard Biener <rguenther@suse.de>
18912
18913 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18914 (pass_build_cfg): Provide PROP_loops.
18915 (move_sese_region_to_fn): Remove loops that are outlined into fn
18916 for now.
18917 * tree-inline.c: Include cfgloop.h.
18918 (initialize_cfun): Do not drop PROP_loops.
18919 (copy_loops): New function.
18920 (copy_cfg_body): Copy loop structure.
18921 (tree_function_versioning): Initialize destination loop tree.
18922 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18923 (pass_parallelize_loops): Do IL verification.
18924 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18925 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18926 the CFG make sure we fixup loops as well.
18927 * tree-ssa-tail-merge.c: Include cfgloop.h.
18928 (replace_block_by): When merging loop latches mark loops for fixup.
18929 * lto-streamer-out.c (output_struct_function_base): Drop
18930 PROP_loops for now.
18931 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18932 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18933 * ipa-split.c: Include cfgloop.h.
18934 (split_function): Add the new return block to the loop tree root.
18935 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18936 whether we have removed the forwarder block.
18937 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18938 * cfgloop.h (place_new_loop): Declare.
18939 * cfgloopmanip.c (place_new_loop): Export.
18940 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18941 (tree-switch-conversion.o): Likewise.
18942 (tree-complex.o): Likewise.
18943 (tree-inline.o): Likewise.
18944 (tree-ssa-tailmerge.o): Likewise.
18945 (ipa-split.o): Likewise.
18946 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18947 (tree-ssa-copy.o): Likewise.
18948 * tree-switch-conversion.c: Include cfgloop.h
18949 (process_switch): If we emit a bit-test cascade, schedule loops
18950 for fixup.
18951 * tree-complex.c: Include cfgloop.h.
18952 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18953 * asan.c: Include cfgloop.h.
18954 (create_cond_insert_point): Properly add new basic-blocks to
18955 loops, schedule loop fixup.
18956 * cfgloop.c (verify_loop_structure): Check that looks are not
18957 marked for fixup.
18958 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18959 to loops.
18960 (expand_omp_for_generic): Likewise.
18961 (expand_omp_sections): Likewise.
18962 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18963 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18964 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18965 is initialized, not when loops are present.
18966 * tree-parloops.c (parallelize_loops): Remove checking here.
18967 * passes.c (init_optimization_passes): Schedule a copy-propagation
18968 pass before complete unrolling of inner loops.
18969
18970 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18971
18972 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18973 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18974 (colorize_init): Add argument to _WIN32 version.
18975 * toplev.c: Include diagnostic-color.h.
18976 (process_options): Default to -fdiagnostics-color=auto if
18977 GCC_COLORS env var is in the environment.
18978 * common.opt (fdiagnostics-color=): Add Var and Init.
18979 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18980 env var is in the environment, the default is auto rather than never.
18981
18982 * diagnostic.h (file_name_as_prefix): Add context argument.
18983 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18984 the string as locus.
18985 * langhooks.c (lhd_print_error_function): Adjust caller.
18986
18987 2013-04-25 Lawrence Crowl <crowl@google.com>
18988
18989 * var-tracking.c (shared_hash_def::htab):
18990 Change type to hash_table. Update dependent calls and types.
18991
18992 2013-04-25 Lawrence Crowl <crowl@google.com>
18993
18994 * Makefile.in: Update as needed below.
18995
18996 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18997 Move declaration to after the type's method definitons.
18998
18999 * attribs.c (htab_t scoped_attributes::attribute_hash):
19000 Change type to hash_table. Update dependent calls and types.
19001
19002 * bitmap.c (htab_t bitmap_desc_hash):
19003 Change type to hash_table. Update dependent calls and types.
19004
19005 * cselib.c (htab_t cselib_hash_table):
19006 Change type to hash_table. Update dependent calls and types.
19007
19008 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19009 (hash_string_slot_node): Move implementation into lto-streamer.h
19010 struct string_slot_hasher.
19011 (eq_string_slot_node): Likewise.
19012
19013 * data-streamer-out.c: Update output_block::string_hash_table
19014 dependent calls and types.
19015
19016 * dwarf2cfi.c (htab_t trace_index):
19017 Change type to hash_table. Update dependent calls and types.
19018
19019 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19020 Change type to hash_table. Update dependent calls and types.
19021 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19022 (htab_t optimize_external_refs::map): Likewise.
19023 (htab_t output_comp_unit::extern_map): Likewise.
19024 (htab_t output_comdat_type_unit::extern_map): Likewise.
19025 (htab_t output_macinfo::macinfo_htab): Likewise.
19026 (htab_t optimize_location_lists::htab): Likewise.
19027 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19028
19029 * except.c (htab_t ehspec_hash_type):
19030 Change type to hash_table. Update dependent calls and types.
19031 (assign_filter_values::ttypes): Likewise.
19032 (assign_filter_values::ehspec): Likewise.
19033 (sjlj_assign_call_site_values::ar_hash): Likewise.
19034 (convert_to_eh_region_ranges::ar_hash): Likewise.
19035
19036 * gcse.c (htab_t pre_ldst_table):
19037 Change type to hash_table. Update dependent calls and types.
19038
19039 * ggc-common.c (htab_t saving_htab):
19040 Change type to hash_table. Update dependent calls and types.
19041 (htab_t loc_hash): Likewise.
19042 (htab_t ptr_hash): Likewise.
19043 (call_count): Rename ggc_call_count.
19044 (call_alloc): Rename ggc_call_alloc.
19045 (loc_descriptor): Rename make_loc_descriptor.
19046 (add_statistics): Rename ggc_add_statistics.
19047
19048 * ggc-common.c (saving_htab):
19049 Change type to hash_table. Update dependent calls and types.
19050
19051 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19052 (push_gimplify_context): Likewise.
19053 (pop_gimplify_context): Likewise.
19054 (struct gimple_temp_hash_elt): Added.
19055 (struct gimplify_hasher): Likewise.
19056 (struct gimplify_ctx.temp_htab):
19057 Change type to hash_table. Update dependent calls and types.
19058
19059 * gimple-fold.c: Include gimplify-ctx.h.
19060
19061 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19062 Change type to hash_table. Update dependent calls and types.
19063 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19064 avoid potential global name collision.
19065
19066 * gimplify.c: Include gimplify-ctx.h.
19067 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19068 (htab_t gimplify_ctx::temp_htab):
19069 Update dependent calls and types for new type hash_table.
19070 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19071 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19072
19073 * gimplify-ctx.h: New.
19074 (struct gimple_temp_hash_elt): Move from gimplify.c.
19075 (class gimplify_hasher): New.
19076 (struct gimplify_ctx): Move from gimple.h.
19077 (htab_t gimplify_ctx::temp_htab):
19078 Change type to hash_table. Update dependent calls and types.
19079
19080 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19081 (htab_t ivs_params::newivs_index):
19082 Change type to hash_table. Update dependent calls and types.
19083 (htab_t ivs_params::params_index): Likewise.
19084 (htab_t print_generated_program::params_index): Likewise.
19085 (htab_t gloog::newivs_index): Likewise.
19086 (htab_t gloog::params_index): Likewise.
19087
19088 * graphite.c: Include graphite-htab.h.
19089 4htab_t graphite_transform_loops::bb_pbb_mapping):
19090 Change type to hash_table. Update dependent calls and types.
19091
19092 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19093 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19094 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19095
19096 * graphite-dependences.c: Include graphite-htab.h.
19097 (loop_is_parallel_p): Change hash table type of parameter.
19098
19099 * graphite-htab.h: New.
19100 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19101 (extern find_pbb_via_hash): Move from graphite-poly.h.
19102 (extern loop_is_parallel_p): Move from graphite-poly.h.
19103 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19104
19105 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19106 (extern loop_is_parallel_p): Move to graphite-htab.h.
19107 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19108
19109 * haifa-sched.c (htab_t delay_htab):
19110 Change type to hash_table. Update dependent calls and types.
19111 (htab_t delay_htab_i2): Likewise.
19112
19113 * ira-color.c (htab_t allocno_hard_regs_htab):
19114 Change type to hash_table. Update dependent calls and types.
19115
19116 * ira-costs.c (htab_t cost_classes_htab):
19117 Change type to hash_table. Update dependent calls and types.
19118
19119 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19120 Change type to hash_table. Update dependent calls and types.
19121
19122 * loop-iv.c (htab_t bivs):
19123 Change type to hash_table. Update dependent calls and types.
19124
19125 * loop-unroll.c (htab_t opt_info::insns_to_split):
19126 Change type to hash_table. Update dependent calls and types.
19127 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19128
19129 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19130 (struct string_slot_hasher): New.
19131 (htab_t output_block::string_hash_table):
19132 Change type to hash_table. Update dependent calls and types.
19133
19134 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19135 (htab_t file_name_hash_table):
19136 Change type to hash_table. Update dependent calls and types.
19137
19138 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19139 calls and types.
19140
19141 * lto-streamer.c (htab_t tree_htab):
19142 Change type to hash_table. Update dependent calls and types.
19143
19144 * omp-low.c: Include gimplify-ctx.h.
19145
19146 * passes.c (htab_t name_to_pass_map):
19147 Change type to hash_table. Update dependent calls and types.
19148 (pass_traverse): Rename to passes_pass_traverse.
19149
19150 * plugin.c (htab_t event_tab):
19151 Change type to hash_table. Update dependent calls and types.
19152
19153 * postreload-gcse.c (htab_t expr_table):
19154 Change type to hash_table. Update dependent calls and types.
19155 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19156
19157 * sese.c (debug_rename_map_1): Make extern.
19158 (htab_t copy_bb_and_scalar_dependences::rename_map):
19159 Change type to hash_table. Update dependent calls and types.
19160
19161 * sese.h (extern debug_rename_map): Move to .c file.
19162
19163 * store-motion.c (htab_t store_motion_mems_table):
19164 Change type to hash_table. Update dependent calls and types.
19165
19166 * trans-mem.c (htab_t tm_new_mem_hash):
19167 Change type to hash_table. Update dependent calls and types.
19168
19169 * tree-browser.c (htab_t TB_up_ht):
19170 Change type to hash_table. Update dependent calls and types.
19171
19172 * tree-cfg.c (htab_t discriminator_per_locus):
19173 Change type to hash_table. Update dependent calls and types.
19174
19175 * tree-complex.c: Include tree-hasher.h
19176 (htab_t complex_variable_components):
19177 Change type to hash_table. Update dependent calls and types.
19178
19179 * tree-eh.c (htab_t finally_tree):
19180 Change type to hash_table. Update dependent calls and types.
19181
19182 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19183 struct int_tree_hasher.
19184 (extern int_tree_map_eq): Likewise.
19185 (uid_decl_map_hash): Removed.
19186 (extern decl_tree_map_eq): Likewise.
19187
19188 * tree-hasher.h: New.
19189 (struct int_tree_hasher): New.
19190 (typedef int_tree_htab_type): New.
19191
19192 * tree-inline.c: Include gimplify-ctx.h.
19193
19194 * tree-mudflap.c: Include gimplify-ctx.h.
19195
19196 * tree-parloops.c: Include tree-hasher.h.
19197 (htab_t eliminate_local_variables_stmt::decl_address):
19198 Change type to hash_table. Update dependent calls and types.
19199 (htab_t separate_decls_in_region::decl_copies): Likewise.
19200
19201 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19202 Change type to hash_table. Update dependent calls and types.
19203
19204 * tree-sra.c (candidates):
19205 Change type to hash_table. Update dependent calls and types.
19206
19207 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19208 in tree-flow.h.
19209 (int_tree_map_hash): Likewise.
19210
19211 * tree-ssa-dom.c (htab_t avail_exprs):
19212 Change type to hash_table. Update dependent calls and types.
19213
19214 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19215 Change type to hash_table. Update dependent calls and types.
19216
19217 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19218 Change type to hash_table. Update dependent calls and types.
19219
19220 * tree-ssa-phiopt.c (seen_ssa_names):
19221 Change type to hash_table. Update dependent calls and types.
19222
19223 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19224 Change type to hash_table. Update dependent calls and types.
19225
19226 * tree-ssa-uncprop.c (equiv):
19227 Change type to hash_table. Update dependent calls and types.
19228
19229 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19230
19231 PR rtl-optimization/57003
19232 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19233 call note_stores with kill_clobbered_value callback again after
19234 killing regs_invalidated_by_call.
19235
19236 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19237
19238 * config/aarch64/aarch64-simd.md
19239 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19240 (aarch64_simd_bsl<mode>): Likewise.
19241 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19242
19243 2013-04-25 Marek Polacek <polacek@redhat.com>
19244
19245 PR tree-optimization/57066
19246 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19247
19248 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19249
19250 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19251
19252 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19253
19254 * config/aarch64/aarch64-builtins.c
19255 (aarch64_fold_builtin): New.
19256 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19257 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19258 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19259 * config/aarch64/arm_neon.h
19260 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19261
19262 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19263 Tejas Belagod <tejas.belagod@arm.com>
19264
19265 * config/aarch64/aarch64-builtins.c
19266 (aarch64_gimple_fold_builtin): New.
19267 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19268 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19269 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19270 (addvv4sf): Update.
19271 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19272
19273 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19274
19275 * config/aarch64/aarch64.md
19276 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19277
19278 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19279
19280 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19281 (*ngcsi_uxtw): New pattern.
19282
19283 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19284 Julian Brown <julian@codesourcery.com>
19285
19286 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19287 (TB_DREG): Add T_V4HF.
19288 (v4hf_UP): New macro.
19289 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19290 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19291 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19292 built-ins.
19293 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19294 (arm_vector_mode_supported_p): Handle V4HF.
19295 (arm_mangle_map): Handle V4HFmode.
19296 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19297 * config/arm/arm_neon_builtins.def: Add entries for
19298 vcvtv4hfv4sf, vcvtv4sfv4hf.
19299 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19300 (neon_vcvtv4hfv4sf): Likewise.
19301 * config/arm/neon-gen.ml: Handle half-precision floating point
19302 features.
19303 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19304 * config/arm/arm_neon.h: Regenerate.
19305 * config/arm/neon.ml (type elts): Add F16.
19306 (type vectype): Add T_float16x4, T_floatHF.
19307 (type vecmode): Add V4HF.
19308 (type features): Add Requires_FP_bit feature.
19309 (elt_width): Handle F16.
19310 (elt_class): Likewise.
19311 (elt_of_class_width): Likewise.
19312 (mode_of_elt): Refactor.
19313 (type_for_elt): Handle F16, fix error messages.
19314 (vectype_size): Handle T_float16x4.
19315 (vcvt_sh): New function.
19316 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19317 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19318 (string_of_mode): Handle V4HF.
19319 * doc/arm-neon-intrinsics.texi: Regenerate.
19320
19321 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19322
19323 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19324 format specifier in 'X' case.
19325
19326 2013-04-25 Alan Modra <amodra@gmail.com>
19327
19328 PR target/57052
19329 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19330 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19331 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19332 Repeat for many other rotate/shift and mask patterns using subregs.
19333 Name lshiftrt insns.
19334 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19335 on WORDS_BIG_ENDIAN.
19336
19337 2013-04-25 Alan Modra <amodra@gmail.com>
19338
19339 * config.gcc: Support little-endian powerpc-linux targets.
19340 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19341 (LINK_OS_LINUX_SPEC): Define.
19342 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19343 Preserve MASK_LITTLE_ENDIAN.
19344 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19345 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19346 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19347 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19348 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19349 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19350 Correct fp word order for little-endian. Don't shift toc entries
19351 smaller than a word for little-endian.
19352 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19353 (bswapdi2 splits): Correct low-part subreg for little-endian.
19354 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19355 low/high where such is correct only for be.
19356 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19357 little-endian for -mcall-aixdesc.
19358
19359 2013-04-25 Alan Modra <amodra@gmail.com>
19360
19361 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19362 replace_equiv_address_nv.
19363
19364 2013-04-25 Alan Modra <amodra@gmail.com>
19365
19366 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19367
19368 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19369
19370 Revert:
19371 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19372 * rtl.h (struct rtx_def): ...
19373
19374 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19375
19376 PR rtl-optimizations/57046
19377 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19378 for multi-reg splits.
19379
19380 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19381
19382 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19383
19384 2013-04-24 Sterling Augustine <saugustine@google.com>
19385
19386 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19387 (comp_dir_string, debug_str_dwo_section): New.
19388 (DEBUG_STR_DWO_SECTION): Rename to ...
19389 (DEBUG_DWO_STR_SECTION): ... this.
19390 (DEBUG_NORM_STR_SECTION): Delete.
19391 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19392 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19393 (find_AT_string): Move most logic to ...
19394 (find_AT_string_in_table): ... here. New.
19395 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19396 add_skeleton_AT_string. Delete logic.
19397 (output_skeleton_debug_sections): Remove call to
19398 add_top_level_skeleton_die_attrs.
19399 (add_comp_dir_attribute): Move logic to comp_dir_string.
19400 (dwarf2out_init): Initialize debug_str_dwo_section.
19401 (output_indirect_string): Call find_string_form.
19402 (output_indirect_strings): Rewrite.
19403 (prune_unused_types): Empty skeleton_debug_str_hash.
19404 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19405 (dwarf2out_finish): Call output_indirect_strings.
19406
19407 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19408
19409 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19410
19411 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19412
19413 * rtl.h (struct rtx_def): Add comment for field jump.
19414 (LRA_SUBREG_P): New macro.
19415 * recog.c (register_operand): Check LRA_SUBREG_P.
19416 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19417 stack frame.
19418 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19419 (lra_final_code_change): Skip subreg change for operators.
19420 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19421 if there are no operand changes.
19422 * lra-constraints.c (curr_insn_set): New.
19423 (match_reload): Set LRA_SUBREG_P.
19424 (emit_spill_move): Ditto.
19425 (check_and_process_move): Use curr_insn_set. Process only single
19426 set insns. Don't initialize sec_mem_p and change_p.
19427 (simplify_operand_subreg): Use LRA_SUBREG_P.
19428 (reg_in_class_p): New function.
19429 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19430 of #ifdef. Add code to remove cycling.
19431 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19432 non-null disp. Reload inner instead of disp when base and index
19433 are null. Try to put lo_sum into register.
19434 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19435 (check_and_process_move): Move code for move cost check to
19436 simple_move_p. Remove equiv_substitution.
19437 (simple_move_p): New function.
19438 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19439 curr_insn_set. Call check_and_process_move only for single set
19440 insns. Use the new function. Move call of check_and_process_move
19441 after operand equiv substitution and address process.
19442
19443 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19444
19445 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19446 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19447 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19448
19449 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19450
19451 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19452
19453 2013-04-24 Marek Polacek <polacek@redhat.com>
19454
19455 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19456 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19457 (select_loops_exit_conditions): Likewise.
19458 (number_of_iterations_for_all_loops): Likewise.
19459 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19460 (scev_analysis): Likewise.
19461
19462 2013-04-02 Catherine Moore <clm@codesourcery.com>
19463 Chao-ying Fu <fu@mips.com>
19464
19465 * config/mips/micromips.md (jraddiusp): New pattern.
19466 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19467 instruction if possible.
19468
19469 2013-04-24 Alan Modra <amodra@gmail.com>
19470
19471 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19472
19473 2013-04-24 Julian Brown <julian@codesourcery.com>
19474 Chung-Lin Tang <cltang@codesourcery.com>
19475
19476 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19477 dependency behavior in enumeration type DIE generation. Add TODO note
19478 to comments about future DW_FORM_sdata/udata re-work of related code.
19479
19480 2013-04-23 Lawrence Crowl <crowl@google.com>
19481
19482 * Makefile.in: Update as needed below.
19483
19484 * hash-table.h (class hash_table):
19485 Correct many methods with parameter types compare_type to the correct
19486 value_type. (Correct code was unlikely to notice the change.)
19487 (hash_table::elements_with_deleted) New.
19488 (class hashtable::iterator): New.
19489 (hashtable::begin()): New.
19490 (hashtable::end()): New.
19491 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19492
19493 * statistics.c (statistics_hashes):
19494 Change type to hash_table. Update dependent calls and types.
19495
19496 * tree-into-ssa.c (var_infos):
19497 Change type to hash_table. Update dependent calls and types.
19498
19499 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19500 Change type to hash_table. Update dependent calls and types.
19501
19502 * tree-ssa-loop-im.c (struct mem_ref.refs):
19503 Change type to hash_table. Update dependent calls and types.
19504
19505 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19506 Change type to hash_table. Update dependent calls and types.
19507
19508 * tree-ssa-sccvn.c (vn_tables_s::nary):
19509 Change type to hash_table. Update dependent calls and types.
19510 (vn_tables_s::phis): Likewise.
19511 (vn_tables_s::references): Likewise.
19512
19513 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19514 (vn_reference_eq): Update parameter and return types.
19515
19516 * tree-ssa-structalias.c (pointer_equiv_class_table):
19517 Change type to hash_table. Update dependent calls and types.
19518 (location_equiv_class_table): Likewise.
19519
19520 * tree-vect-data-refs.c: Consequential changes for making
19521 peeling a hash_table.
19522
19523 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19524 (destroy_loop_vec_info): Dependent hash_table update.
19525
19526 * tree-vectorizer.h (peeling_htab):
19527 Change type to hash_table. Update dependent calls and types.
19528
19529 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19530
19531 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19532 to check the register content is equal or not.
19533 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19534 to assign register content record.
19535 * lra-eliminations.c (update_reg_eliminate): Use
19536 lra_update_reg_val_offset to update register content offset.
19537 * lra-int.h (struct lra_reg): Add offset member.
19538 (lra_reg_val_equal_p): New static inline function.
19539 (lra_update_reg_val_offset): New static inline function.
19540 (lra_assign_reg_val): New static inline function.
19541 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19542 to assign register content record.
19543 (initialize_lra_reg_info_element): Initial offset to zero.
19544
19545 2013-04-23 Catherine Moore <clm@codesourcery.com>
19546
19547 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19548 operands. Record compression.
19549
19550 2013-04-23 Xinliang David Li <davidxl@google.com>
19551
19552 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19553
19554 2013-04-23 Richard Biener <rguenther@suse.de>
19555
19556 PR middle-end/57036
19557 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19558 parameter, only add abnormal goto edges from the copied body
19559 if the call could perform abnormal gotos.
19560 (copy_cfg_body): Adjust.
19561
19562 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19563
19564 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19565
19566 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19567
19568 * coretypes.h (gimple_stmt_iterator): Add struct to make
19569 compatible with C.
19570
19571 2013-04-23 Richard Biener <rguenther@suse.de>
19572
19573 PR tree-optimization/57026
19574 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19575 from SSA names occuring in abnormal PHI nodes.
19576
19577 2013-04-22 Andi Kleen <ak@linux.intel.com>
19578
19579 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19580
19581 2013-04-22 Andi Kleen <ak@linux.intel.com>
19582
19583 * lto/lto.c (print_lto_report_1): Declare early.
19584 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19585
19586 2013-04-22 Andi Kleen <ak@linux.intel.com>
19587
19588 * common.opt (-flto-report-wpa): Add.
19589 * doc/invoke.texi (-flto-report-wpa): Add.
19590 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19591 (lto_main): dito.
19592
19593 2013-04-22 Xinliang David Li <davidxl@google.com>
19594
19595 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19596 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19597 * Makefile.in: New dependency
19598
19599 David Daney <ddaney.cavm@gmail.com>
19600
19601 * configure.ac (gcc_cv_as_micromips_support): Use the
19602 --fatal-warnings option.
19603 * configure: Regenerate.
19604
19605 2013-04-22 Marek Polacek <polacek@redhat.com>
19606
19607 PR sanitizer/56990
19608 * tsan.c (instrument_expr): Don't instrument expression
19609 in case its size is zero.
19610
19611 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19612
19613 PR target/57032
19614 Revert:
19615 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19616
19617 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19618
19619 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19620
19621 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19622 (gimple_stmt_iterator): New typedef.
19623 * gimple.h (gimple_stmt_iterator): Rename to...
19624 (gimple_stmt_iterator_d): ... This.
19625 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19626 trees be valid for GIMPLE and GENERIC.
19627 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19628 * gimple-fold.c (gimple_fold_call): Call target hook
19629 gimple_fold_builtin.
19630 * hooks.c (hook_bool_gsiptr_false): New.
19631 * hooks.h (hook_bool_gsiptr_false): New.
19632 * target.def (fold_stmt): New.
19633 * doc/tm.texi: Regenerate.
19634
19635 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19636
19637 PR target/57018
19638 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19639 a set sp if no stack realignment.
19640
19641 2013-04-22 Nick Clifton <nickc@redhat.com>
19642
19643 * config.gcc (tilegx-linux): Extend extra_objs rather than
19644 overwriting it.
19645 (tilepro-linux): Likewise.
19646
19647 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19648
19649 * config/aarch64/aarch64-builtins.c
19650 (CF): Remove.
19651 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19652 (VAR<1-12>): Add MAP parameter.
19653 (BUILTIN_*): Likewise.
19654 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19655 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19656 (aarch64_ushl_n<mode>): Likewise.
19657 (aarch64_sshr_n<mode>): Likewise.
19658 (aarch64_ushr_n<mode>): Likewise.
19659 (aarch64_<maxmin><mode>): Likewise.
19660 (aarch64_sqrt<mode>): Likewise.
19661 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19662 (vshr<q>_n_*): Likewise.
19663
19664 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19665
19666 * config/aarch64/aarch64-builtins.c
19667 (aarch64_simd_builtin_type_mode): Handle SF types.
19668 (sf_UP): Define.
19669 (BUILTIN_GPF): Define.
19670 (aarch64_init_simd_builtins): Handle SF types.
19671 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19672 (frecps): Likewise.
19673 (frecpx): Likewise.
19674 * config/aarch64/aarch64-simd.md
19675 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19676 (aarch64_frecpe<mode>): New.
19677 (aarch64_frecps<mode>): Likewise.
19678 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19679 (v8type): Add frecp<esx>.
19680 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19681 (aarch64_frecps<mode>): Likewise.
19682 * config/aarch64/iterators.md (FRECP): New.
19683 (frecp_suffix): Likewise.
19684 * config/aarch64/arm_neon.h
19685 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19686
19687 2013-04-22 Christian Bruel <christian.bruel@st.com>
19688
19689 PR target/56995
19690 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19691 (REG_CLASS_NAMES): Idem.
19692 (REG_CLASS_CONTENTS): Idem.
19693 (REGCLASS_HAS_FP_REG): Idem.
19694 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19695 (sh_conditional_register_usage): Idem.
19696
19697 2013-04-21 Jeff Law <law@redhat.com>
19698
19699 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19700 (ssa_forward_propagate_and_combine): Use it.
19701
19702 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19703
19704 * lra.c: Update the flow chart diagram.
19705
19706 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19707
19708 PR rtl-optimization/56847
19709 * lra-constraints.c (process_alt_operands): Discourage alternative
19710 with non-matche doffsettable memory constraint fro memory with
19711 known offset.
19712
19713 2013-04-19 Richard Biener <rguenther@suse.de>
19714
19715 PR tree-optimization/56982
19716 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19717 function.
19718 * gimplify.c (gimplify_call_expr): Notice special calls.
19719 (gimplify_modify_expr): Likewise.
19720 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19721 abnormal control flow receivers.
19722 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19723 in the same way as cfun->has_nonlocal_labels.
19724 (gimple_purge_dead_abnormal_call_edges): Likewise.
19725 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19726 receivers start a basic-block.
19727
19728 2013-04-19 Richard Biener <rguenther@suse.de>
19729
19730 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19731 member ...
19732 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19733 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19734 (SLP_TREE_LOAD_PERMUTATION): Add.
19735 (vect_transform_slp_perm_load): Adjust prototype.
19736 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19737 (vect_free_slp_instance): Likewise.
19738 (vect_create_new_slp_node): Likewise.
19739 (vect_supported_slp_permutation_p): Remove.
19740 (vect_slp_rearrange_stmts): Adjust.
19741 (vect_supported_load_permutation_p): Likewise. Inline
19742 vect_supported_slp_permutation_p here.
19743 (vect_analyze_slp_instance): Compute load permutations per
19744 slp node instead of per instance.
19745 (vect_get_slp_defs): Adjust.
19746 (vect_transform_slp_perm_load): Likewise.
19747 (vect_schedule_slp_instance): Remove redundant code.
19748 (vect_schedule_slp): Remove hack for PR56270, add it ...
19749 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19750 CSE loads for SLP. Adjust.
19751
19752 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19753
19754 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19755 spelling in two comments.
19756
19757 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19758
19759 PR target/56797
19760 * config/arm/arm.c (load_multiple_sequence): Require SP
19761 as base register for loads if SP is in the register list.
19762
19763 2013-04-19 Martin Jambor <mjambor@suse.cz>
19764
19765 PR tree-optimization/56718
19766 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19767 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19768 and made public. Adjusted all callers.
19769 (ipa_intraprocedural_devirtualization): New function.
19770 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19771 (ipa_intraprocedural_devirtualization): Likewise.
19772 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19773
19774 2013-04-19 Richard Biener <rguenther@suse.de>
19775
19776 PR tree-optimization/57000
19777 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19778
19779 2013-04-19 Terry Guo <terry.guo@arm.com>
19780
19781 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19782 Replace with ...
19783 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19784 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19785 (cortex_m4_fmacs): Use new reservations.
19786 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19787
19788 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19789
19790 PR rtl-optimization/56999
19791 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19792 related code.
19793 (lra_coalesce): Remove split_origin_bitmap and related code.
19794 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19795 ranges if necessary.
19796
19797 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19798
19799 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19800 New array.
19801 (ix86_expand_call): Remove clobbered_registers array and use
19802 x86_64_ms_sysv_extra_clobbered_registers instead.
19803 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19804 Declare here.
19805 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19806 predicate.
19807 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19808 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19809 (*call_value_rex64_ms_sysv): Ditto.
19810
19811 2013-04-18 Cary Coutant <ccoutant@google.com>
19812
19813 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19814 parent instead of die_mark.
19815
19816 2013-04-18 Diego Novillo <dnovillo@google.com>
19817
19818 * gimple.c (create_gimple_tmp): New.
19819 (get_expr_type): New.
19820 (build_assign): New.
19821 (build_type_cast): New.
19822 * gimple.h (enum ssa_mode): Define.
19823 (gimple_seq_set_location): New.
19824 * asan.c (build_check_stmt): Change some gimple_build_* calls
19825 to use build_assign and build_type_cast.
19826
19827 2013-04-18 Richard Biener <rguenther@suse.de>
19828
19829 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19830 handle negative step. Remove redundant checks.
19831 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19832 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19833 for negative step and grouped loads fail to vectorize.
19834
19835 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19836
19837 * emit-rtl.c (reset_insn_used_flags): New function.
19838 (reset_all_used_flags): Use it.
19839 (verify_insn_sharing): New function.
19840 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19841
19842 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19843
19844 PR tree-optimization/56984
19845 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19846 and (x >> M) >= N don't register any assertion if N << M is the
19847 minimum value.
19848
19849 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19850
19851 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19852 do not delete_insn insns that have not yet been emitted, only
19853 unlink them with remove_insn.
19854 * df-scan.c (df_insn_delete): Revert r197492.
19855
19856 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19857
19858 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19859 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19860
19861 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19862
19863 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19864 define_insn_and_split.
19865 (and_scc,ior_scc,negscc): Likewise.
19866 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19867
19868 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19869
19870 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19871 can benefit from using a sequence of LDRD instructions in epilogue
19872 instead of a single LDM instruction.
19873
19874 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19875
19876 PR 45688
19877 * doc/extend.texi: Fix typo.
19878
19879 2013-04-17 Richard Biener <rguenther@suse.de>
19880
19881 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19882 (vect_build_slp_tree): ... here.
19883 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19884 match. Remove special-casing of mismatched complex loads.
19885 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19886 re-try the match with swapped commutative operands.
19887 (vect_supported_load_permutation_p): Remove special-casing of
19888 mismatched complex loads.
19889 (vect_analyze_slp_instance): Adjust.
19890
19891 2013-04-17 Richard Biener <rguenther@suse.de>
19892
19893 PR rtl-optimization/56921
19894 * cfgloop.h (struct loop): Add simple_loop_desc member.
19895 (struct niter_desc): Mark with GTY(()).
19896 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19897 * loop-iv.c (get_simple_loop_desc): Likewise.
19898 (free_simple_loop_desc): Likewise.
19899
19900 Revert
19901 2013-04-16 Richard Biener <rguenther@suse.de>
19902
19903 PR rtl-optimization/56921
19904 * loop-init.c (pass_rtl_move_loop_invariants): Add
19905 TODO_do_not_ggc_collect to todo_flags_finish.
19906 (pass_rtl_unswitch): Same.
19907 (pass_rtl_unroll_and_peel_loops): Same.
19908 (pass_rtl_doloop): Same.
19909
19910 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19911
19912 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19913 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19914 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19915 references.
19916 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19917 * tree-streamer.c (record_common_node): Adjust reference in comment.
19918
19919 2013-04-17 Terry Guo <terry.guo@arm.com>
19920
19921 * config/arm/cortex-m4.md: Add a new bypass.
19922
19923 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19924
19925 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19926 New pattern.
19927 (*subs_<optab><mode>_multp2): New pattern.
19928 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19929 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19930
19931 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19932
19933 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19934 (*subs_mul_imm_<mode>): New pattern.
19935
19936 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19937
19938 PR target/56948
19939 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19940 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19941 (vsx_movti_32bit): Change j->wa to O->wa.
19942
19943 2013-04-16 Richard Biener <rguenther@suse.de>
19944
19945 PR rtl-optimization/56921
19946 * loop-init.c (pass_rtl_move_loop_invariants): Add
19947 TODO_do_not_ggc_collect to todo_flags_finish.
19948 (pass_rtl_unswitch): Same.
19949 (pass_rtl_unroll_and_peel_loops): Same.
19950 (pass_rtl_doloop): Same.
19951
19952 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19953
19954 * config/arm/arm.c (emit_multi_reg_push): New declaration
19955 for an existing function.
19956 (arm_emit_strd_push): New function.
19957 (arm_expand_prologue): Used here.
19958 (arm_emit_ldrd_pop): New function.
19959 (arm_expand_epilogue): Used here.
19960 (arm_get_frame_offsets): Update condition.
19961 (arm_emit_multi_reg_pop): Add a special case for load of a single
19962 register with writeback.
19963
19964 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19965
19966 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19967 description.
19968
19969 2013-04-16 Richard Biener <rguenther@suse.de>
19970
19971 PR tree-optimization/56756
19972 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19973 (first_mem_ref_loc): New.
19974 (execute_sm): Place the load temporarily before a previous
19975 access instead of in the latch edge to ensure its SSA dependencies
19976 are defined at points dominating the load.
19977
19978 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19979
19980 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19981 correct fix by moving header and footer insn to the footer of
19982 the merged basic block. Clear BB_END of the merged-away block.
19983
19984 PR middle-end/43631
19985 * emit-rtl.c (make_note_raw): New function.
19986 (link_insn_into_chain): New static inline function.
19987 (add_insn): Use it.
19988 (add_insn_before, add_insn_after): Factor insn chain linking code...
19989 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19990 using link_insn_into_chain.
19991 (note_outside_basic_block_p): New helper function for emit_note_after
19992 and emit_note_before.
19993 (emit_note_after): Use nobb variant of add_insn_after if the note
19994 should not be contained in a basic block.
19995 (emit_note_before): Use nobb variant of add_insn_before if the note
19996 should not be contained in a basic block.
19997 (emit_note_copy): Use make_note_raw.
19998 (emit_note): Likewise.
19999 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20000 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20001 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20002 the moved barrier the tail of the basic block it follows.
20003 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20004
20005 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20006
20007 PR tree-optimization/56962
20008 * gimple-ssa-strength-reduction.c (record_increment): Only set
20009 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20010 either rhs1 or rhs2 is equal to c->base_expr.
20011
20012 2013-04-15 Richard Biener <rguenther@suse.de>
20013
20014 PR tree-optimization/56933
20015 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20016 member.
20017 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20018 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20019 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20020 dependence check ...
20021 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20022 ... here.
20023 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20024 GROUP_READ_WRITE_DEPENDENCE.
20025
20026 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20027
20028 * emit-rtl.c (reset_all_used_flags): New function.
20029 (verify_rtl_sharing): Call reset_all_used_flags before and after
20030 performing the checks.
20031
20032 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20033
20034 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20035 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20036 * config/arm/constraints.md (De): New constraint.
20037 * config/arm/neon.md (anddi3_neon): Delete.
20038 (neon_vand<mode>): Expand to standard anddi3 pattern.
20039 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20040 Move earlier in the file.
20041 (neon_inv_logic_op2): Likewise.
20042 (arm_anddi_operand_neon): New predicate.
20043
20044 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20045
20046 * configure.ac (gcc_cv_ld_as_needed): Set
20047 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20048 Use -z ignore, -z record on *-*-solaris2*.
20049 (HAVE_LD_AS_NEEDED): Update comment.
20050 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20051 * configure: Regenerate.
20052 * config.in: Regenerate.
20053 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20054 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20055 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20056 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20057 equivalents. Fix markup.
20058 * doc/tm.texi: Regenerate.
20059
20060 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20061
20062 * config/i386/i386.opt: New option mstack-protector-guard=.
20063 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20064 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20065 TARGET_SSP_TLS_GUARD.
20066 * config/i386/i386.c (ix86_option_override_internal): Set
20067 ix86_stack_protector_guard.
20068 * config/i386/i386.md (stack_protect_set): Enable for
20069 TARGET_SSP_TLS_GUARD only.
20070 (stack_protect_set_<mode>): Ditto.
20071 (stack_protect_test): Ditto.
20072 (stack_protect_test_<mode>): Ditto.
20073 * doc/invoke.texi (i386 Option): Document.
20074
20075 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20076
20077 PR target/56890
20078 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20079 (S_MODES): Set H_MODE bit.
20080 (SF_MODES): Set only S_MODE and SF_MODE bits.
20081 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20082 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20083 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20084 <MODE_FLOAT>: Likewise.
20085
20086 2013-04-15 Joey Ye <joey.ye@arm.com>
20087
20088 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20089
20090 2013-04-15 Joey Ye <joey.ye@arm.com>
20091
20092 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20093 for real far jump.
20094 (thumb_far_jump_used_p): Count instruction size and set
20095 far_jump_used.
20096
20097 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20098
20099 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20100 * resource.c (mark_target_live_regs): Reformat conditional block.
20101
20102 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20103
20104 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20105 notes, they are emitted only just before final.
20106 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20107
20108 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20109
20110 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20111 * cfgrtl.c (delete_insn): Call it here instead.
20112 * lra-spills.c (lra_final_code_change): Use delete_insn.
20113 * haifa-sched.c (sched_remove_insn): Likewise.
20114 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20115 returning to the nop pool.
20116 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20117 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20118
20119 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20120
20121 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20122 * doc/tm.texi: Regenerated.
20123
20124 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20125
20126 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20127 QImode checks.
20128
20129 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20130
20131 * df-core.c (df_find_def): Compare register numbers.
20132 (df_find_use): Likewise.
20133
20134 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20135
20136 PR target/56903
20137 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20138 lra_in_progress for return.
20139
20140 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20141
20142 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20143 define_insn into define_insn_and_split and emit movsicc patterns.
20144
20145 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20146
20147 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20148
20149 2013-04-12 Richard Biener <rguenther@suse.de>
20150
20151 * tree-pass.h (TODO_do_not_ggc_collect): New.
20152 * passes.c (execute_one_ipa_transform_pass): Honor
20153 TODO_do_not_ggc_collect.
20154 (execute_one_pass): Likewise.
20155
20156 Revert
20157 2013-04-10 Richard Biener <rguenther@suse.de>
20158
20159 * passes.c (init_optimization_passes): Remove reload pass.
20160 * ira.c (do_reload): Merge into ...
20161 (ira): ... this.
20162 (rest_of_handle_reload): Remove.
20163 (pass_reload): Likewise.
20164 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20165 of reload for vzeroupper pass placement.
20166
20167 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20168
20169 PR tree-optimization/56918
20170 PR tree-optimization/56920
20171 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20172 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20173 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20174 use wide_mul_with_sign method.
20175
20176 2013-04-12 Richard Biener <rguenther@suse.de>
20177
20178 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20179 not be considered a gimple constant.
20180
20181 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20182
20183 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20184 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20185 operations.
20186
20187 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20188 Jakub Jelinek <jakub@redhat.com>
20189
20190 * opts.c: Include diagnostic-color.h.
20191 (common_handle_option): Handle OPT_fdiagnostics_color_.
20192 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20193 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20194 (diagnostic-color.o): New.
20195 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20196 (diagnostic_color_rule): New enum.
20197 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20198 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20199 the location string.
20200 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20201 either NULL, or color kind.
20202 * diagnostic-color.c: New file.
20203 * diagnostic-color.h: New file.
20204 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20205 arguments.
20206 * doc/invoke.texi (-fdiagnostics-color): Document.
20207 * pretty-print.h (pp_show_color): Define.
20208 (struct pretty_print_info): Add show_color field.
20209 * diagnostic.c: Include diagnostic-color.h.
20210 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20211 macros. Colorize error:, warning: etc. strings and also the location
20212 string.
20213 (diagnostic_show_locus): Colorize the caret line.
20214 * pretty-print.c: Include diagnostic-color.h.
20215 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20216 inside of %< %> quotes or quoted through q format modifier.
20217
20218 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20219
20220 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20221
20222 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20223
20224 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20225 code in CC_NZ mode.
20226 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20227 pattern.
20228
20229 2013-04-11 Marek Polacek <polacek@redhat.com>
20230
20231 PR tree-optimization/48184
20232 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20233
20234 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20235
20236 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20237 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20238 (skip_simple_arithmetic): Tidy up.
20239 * tree.h (skip_simple_constant_arithmetic): Declare.
20240
20241 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20242
20243 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20244
20245 2013-04-11 Richard Biener <rguenther@suse.de>
20246
20247 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20248 generate vector constants.
20249
20250 2013-04-11 Richard Biener <rguenther@suse.de>
20251
20252 PR tree-optimization/56878
20253 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20254 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20255 New function.
20256 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20257 Prefer to align the DR with the most invariant base address.
20258
20259 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20260
20261 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20262 comment.
20263
20264 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20265
20266 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20267 floating-point vector comparisons against 0.
20268
20269 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20270
20271 PR tree-optimization/56899
20272 * fold-const.c (extract_muldiv_1): Apply distributive law
20273 only if TYPE_OVERFLOW_WRAPS (ctype).
20274
20275 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20276
20277 PR target/56124
20278 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20279 loading has side effect.
20280
20281 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20282
20283 * config/sparc/sparc.c: Include tree-pass.h.
20284 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20285 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20286 head of file. Change return type. Split off gate function.
20287 (sparc_gate_work_around_errata): New function.
20288 (pass_work_around_errata): New pass definition.
20289 (insert_pass_work_around_errata) New pass insert definition to
20290 insert pass_work_around_errata just after delayed-branch scheduling.
20291 (sparc_option_override): Insert the pass.
20292 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20293
20294 2013-04-10 David S. Miller <davem@davemloft.net>
20295
20296 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20297 or -mcpu=hypersparc.
20298
20299 * target.def (cstore_mode): New hook.
20300 * target.h: Include insn-codes.h
20301 * targhooks.c: Likewise.
20302 (default_cstore_mode): New function.
20303 * targhooks.h: Declare it.
20304 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20305 * doc/tm.texi: Rebuild.
20306 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20307 target hook, rather than inspecting the insn_data.
20308 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20309 (TARGET_CSTORE_MODE): Redefine.
20310 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20311 result patterns.
20312 * config/sparc/predicates.md (cstore_result_operand): New special
20313 predicate.
20314 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20315 Use it for operand 0.
20316 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20317 (*snesi_special): Likewise.
20318 (*snesi_zero): Likewise.
20319 (*seqsi_zero): Likewise.
20320 (*sltu_insn): Likewise.
20321 (*sgeu_insn): Likewise.
20322 (*seqdi_special): Make operand 0 and comparison operation be of
20323 DImode.
20324 (*snedi_special): Likewise.
20325 (*snedi_special_vis3): Likewise.
20326 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20327 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20328 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20329 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20330 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20331 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20332 (*sltu_extend_sp64): Likewise.
20333 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20334 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20335 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20336 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20337 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20338
20339 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20340
20341 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20342 (aarch64_start_file): Use the new function.
20343
20344 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20345 Jason Merrill <jason@redhat.com>
20346
20347 * common.opt: Add -gdwarf.
20348 * opts.c (common_handle_option): Handle it.
20349 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20350
20351 2013-04-10 Richard Biener <rguenther@suse.de>
20352
20353 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20354 (execute_one_ipa_transform_pass): But unconditionally here.
20355 (execute_one_pass): And here.
20356 (init_optimization_passes): Remove reload pass.
20357 * tree-pass.h (TODO_ggc_collect): Remove.
20358 (pass_reload): Likewise.
20359 * ira.c (do_reload): Merge into ...
20360 (ira): ... this.
20361 (rest_of_handle_reload): Remove.
20362 (pass_reload): Likewise.
20363 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20364 of reload for vzeroupper pass placement.
20365 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20366 and todo_flags_finish of all passes.
20367
20368 2013-04-10 Richard Biener <rguenther@suse.de>
20369
20370 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20371 first_const_oprnd field, rename first_def_type to first_op_type.
20372 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20373 (vect_get_and_check_slp_defs): Always use the type of the
20374 operand. Allow mixed vect_external_def, vect_constant_def types.
20375 (vect_get_constant_vectors): Handle mixed vect_external_def,
20376 vect_constant_def types.
20377
20378 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20379
20380 PR tree-optimization/55524
20381 * tree-ssa-math-opts.c
20382 (convert_mult_to_fma): Don't use an fms construct
20383 when we don't have an fms operation, but fnma, and it looks
20384 likely that we'll be able to use the latter.
20385
20386 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20387
20388 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20389 function.
20390 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20391 inline fail caused by overwritable functions.
20392
20393 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20394
20395 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20396 unnecessary bits in the constant power of two case.
20397
20398 2013-04-10 Richard Biener <rguenther@suse.de>
20399
20400 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20401 broken code swapping operands.
20402 (vect_build_slp_tree): Do not compute load permutations here.
20403 (vect_analyze_slp_instance): Compute load permutations here,
20404 after building the SLP tree.
20405
20406 2013-04-09 Christian Bruel <christian.bruel@st.com>
20407
20408 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20409 of next/prev_real_insn.
20410
20411 2013-04-09 Jan Hubicka <jh@suse.cz>
20412
20413 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20414 Drop aliased parameter.
20415 (function_and_variable_visibility): Do not handle alias pairs.
20416 * cgraph.c (varpool_externally_visible_p): Update prototype.
20417 * varpool.c (varpool_add_new_variable): Update.
20418
20419 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20420
20421 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20422
20423 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20424
20425 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20426
20427 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20428
20429 2013-04-09 Marek Polacek <polacek@redhat.com>
20430
20431 PR tree-optimization/48762
20432 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20433
20434 2013-04-09 Richard Biener <rguenther@suse.de>
20435
20436 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20437 dealing with cost.
20438 (vect_build_slp_tree): Likewise.
20439 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20440 calculating the cost of a SLP instance.
20441 (vect_analyze_slp_instance): Use it from here, after building
20442 the SLP tree.
20443
20444 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20445
20446 PR middle-end/56883
20447 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20448 expand_omp_for_static_chunk): Use simple_p = true in
20449 force_gimple_operand_gsi calls when assigning to addressable decls.
20450
20451 2013-04-09 Jeff Law <law@redhat.com>
20452
20453 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20454 when the boolean was created by converting a wider object which
20455 had a boolean range.
20456
20457 2013-04-09 Richard Biener <rguenther@suse.de>
20458
20459 * tree-vectorizer.h (slp_void_p): Remove.
20460 (slp_tree): Typedef before _slp_tree declaration.
20461 (struct _slp_tree): Use a vector of slp_tree as children.
20462 (vect_get_place_in_interleaving_chain): Remove.
20463 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20464 Move ...
20465 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20466 and make static.
20467 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20468 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20469 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20470 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20471 Use slp_node instead of slp_void_p and adjust.
20472
20473 2013-04-09 Richard Biener <rguenther@suse.de>
20474
20475 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20476 work that is not necessary.
20477
20478 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20479
20480 PR tree-optimization/56854
20481 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20482 forward into clobber stmts if it would change MEM_REF lhs into
20483 non-MEM_REF.
20484
20485 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20486
20487 * tree.c (type_hash_lookup, type_hash_add): Make static.
20488 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20489
20490 2013-04-09 Richard Biener <rguenther@suse.de>
20491
20492 * tree.h (unsave_expr_now): Remove.
20493 * tree-inline.c (mark_local_for_remap_r): Remove.
20494 (unsave_expr_1): Likewise.
20495 (unsave_r): Likewise.
20496 (unsave_expr_now): Likewise.
20497 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20498 (propagate_tree_value): Likewise.
20499
20500 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20501
20502 * doc/rtl.texi (sequence): Rewrite documentation to match the
20503 current use of SEQUENCE rtl objects.
20504 * rtl.def (SEQUENCE): Likewise.
20505
20506 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20507 Update documentation.
20508 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20509 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20510
20511 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20512
20513 2013-04-08 Teresa Johnson <tejohnson@google.com>
20514
20515 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20516 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20517 methods.
20518 (estimate_edge_size_and_time): Add comment to suggest using rounding
20519 methods.
20520 (estimate_node_size_and_time): Ditto.
20521 (remap_edge_change_prob): Use helper rounding divide methods.
20522 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20523 (gimple_mod_pow2_value_transform): Ditto.
20524 (gimple_mod_subtract_transform): Ditto.
20525 (gimple_ic_transform): Ditto.
20526 (gimple_stringops_transform): Ditto.
20527 * stmt.c (conditional_probability): Ditto.
20528 (emit_case_dispatch_table): Ditto.
20529 * lto-cgraph.c (merge_profile_summaries): Ditto.
20530 * tree-optimize.c (execute_fixup_cfg): Ditto.
20531 * cfgcleanup.c (try_forward_edges): Ditto.
20532 * cfgloopmanip.c (scale_loop_profile): Ditto.
20533 (loopify): Ditto.
20534 (duplicate_loop_to_header_edge): Ditto.
20535 (lv_adjust_loop_entry_edge): Ditto.
20536 * tree-vect-loop.c (vect_transform_loop): Ditto.
20537 * profile.c (compute_branch_probabilities): Ditto.
20538 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20539 * lto-streamer-in.c (input_cfg): Ditto.
20540 * gimple-streamer-in.c (input_bb): Ditto.
20541 * ipa-cp.c (update_profiling_info): Ditto.
20542 (update_specialized_profile): Ditto.
20543 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20544 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20545 rounding methods.
20546 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20547 (compute_trg_info): Ditto.
20548 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20549 (purge_dead_edges): Ditto.
20550 * loop-unswitch.c (unswitch_loop): Ditto.
20551 * cgraphclones.c (cgraph_clone_edge): Ditto.
20552 (cgraph_clone_node): Ditto.
20553 * tree-inline.c (copy_bb): Ditto.
20554 (copy_edges_for_bb): Ditto.
20555 (initialize_cfun): Ditto.
20556 (copy_cfg_body): Ditto.
20557 (expand_call_inline): Ditto.
20558
20559 2013-04-08 Kai Tietz <ktietz@redhat.com>
20560
20561 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20562 TARGET_CYGWIN64 by TARGET_64BIT.
20563
20564 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20565
20566 * config/epiphany/epiphany.md (GPR_1): New constant.
20567 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20568 * config/epiphany/epiphany.c (gen_compare_reg):
20569 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20570 is already in place.
20571 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20572 Don't require being called during rtl expansion; If y operlaps r0,
20573 return 0.
20574 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20575 (epiphany_expand_epilogue): Likewise.
20576
20577 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20578 Don't use CC_FPmode for ORDERED / UNORDERED.
20579 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20580
20581 * config/epiphany/constraints.md (CnL): New constraint.
20582 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20583 * config/epiphany/predicates.md (add_operand): Allow 1024.
20584
20585 * config/epiphany/epiphany.md (logical_op): New code iterator.
20586 (op_mnc): New code attribute.
20587 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20588 (mov_f+1, mov_f+2): New peephole2 patterns.
20589
20590 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20591 (cstoresi4): Also allow re-use of zero result when doing a NE
20592 comparison to a non-zero operand.
20593 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20594
20595 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20596 Use gen_addsi3_i / gen_subsi3_i.
20597
20598 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20599
20600 PR c++/34949
20601 PR c++/50243
20602 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20603 contain anything but clobbers, at most one __builtin_stack_restore,
20604 optionally debug stmts and final resx, and if it has at least one
20605 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20606 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20607 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20608 which isn't defaut definition, remove them.
20609 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20610 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20611 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20612 with MEM_REF LHS with SSA_NAME address.
20613
20614 2013-04-08 Jeff Law <law@redhat.com>
20615
20616 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20617
20618 2013-04-08 Richard Biener <rguenther@suse.de>
20619
20620 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20621 extra newline.
20622 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20623 determined vector type.
20624 (vect_analyze_data_refs): Likewise.
20625 (vect_get_new_vect_var): Adjust.
20626 (vect_create_destination_var): Preserve SSA name versions.
20627 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20628 not dump anything here.
20629
20630 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20631
20632 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20633 Add member lr_slot_known.
20634 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20635 if necessary.
20636 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20637 Remove code that sets lr_slot_offset according to what a previous
20638 version of epiphany_emit_save_restore used to do.
20639 (epiphany_emit_save_restore): When doing an lr save or restore,
20640 set/verify lr_slot_known and lr_slot_offset.
20641
20642 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20643
20644 PR target/54338
20645 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20646 in ALL_REGS.
20647
20648 2013-04-08 Richard Biener <rguenther@suse.de>
20649
20650 * alias.c (find_base_term): Fix thinko in previous change.
20651
20652 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20653
20654 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20655 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20656 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20657 if possible to compute val.
20658 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20659 For QImode integers don't require anything about precision. Use
20660 const_with_all_bytes_same to find out if the constant doesn't have
20661 repeated bytes in it.
20662
20663 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20664
20665 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20666 within mode size.
20667
20668 2013-04-08 Marek Polacek <polacek@redhat.com>
20669
20670 PR rtl-optimization/48182
20671 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20672 value to 1.
20673
20674 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20675
20676 PR target/55487
20677 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20678 nuses, make sure we have a label.
20679
20680 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20681
20682 PR target/56843
20683 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20684 (rs6000_emit_swdiv_low_precision): Remove.
20685 (rs6000_emit_swdiv): Rewrite to handle between one and four
20686 iterations of Newton-Raphson generally; modify required number of
20687 iterations for some cases.
20688 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20689
20690 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20691
20692 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20693 set-but-unused variable.
20694
20695 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20696 basic blocks of released function bodies garbage-collectable.
20697
20698 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20699 (struct rtl_opt_pass): Add TODO_df_finish.
20700
20701 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20702
20703 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20704
20705 * config/arm/constraints.md (q): New constraint.
20706 * config/arm/ldrdstrd.md: New file.
20707 * config/arm/arm.md (ldrdstrd.md) New include.
20708 (arm_movdi): Use "q" instead of "r" constraint
20709 for double-word memory access.
20710 (movdf_soft_insn): Likewise.
20711 * config/arm/vfp.md (movdi_vfp): Likewise.
20712 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20713 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20714 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20715 (mem_ok_for_ldrd_strd): Likewise.
20716 (output_move_double): Update assertion.
20717
20718 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20719
20720 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20721
20722 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20723
20724 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20725 define_insn_and_split.
20726 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20727
20728 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20729
20730 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20731 define_insn_and_split.
20732 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20733 (shiftsi3_compare): New pattern.
20734 (rrx): New pattern.
20735 * config/arm/unspecs.md (UNSPEC_RRX): New.
20736
20737 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20738
20739 * config/arm/arm.md (negdi_extendsidi): New pattern.
20740 (negdi_zero_extendsidi): Likewise.
20741
20742 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20743
20744 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20745 define_insn_and_split.
20746 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20747 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20748
20749 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20750
20751 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20752 define_insn_and_split.
20753 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20754 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20755
20756 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20757
20758 * config/arm/arm.md (subsi3_carryin): New pattern.
20759 (subsi3_carryin_const): Likewise.
20760 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20761 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20762
20763 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20764
20765 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20766
20767 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20768
20769 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20770 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20771
20772 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20773
20774 * config/arm/arm.c (arm_expand_builtin): Change fcode
20775 type to unsigned int.
20776
20777 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20778
20779 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20780
20781 2013-04-04 Ian Lance Taylor <iant@google.com>
20782
20783 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20784 language standard.
20785
20786 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20787
20788 PR middle-end/56729
20789 * df-scan.c (df_insn_delete): Disable failing assert.
20790
20791 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20792
20793 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20794 New function prototype.
20795 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20796 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20797 (arm_builtin_vectorized_function): New function.
20798
20799 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20800
20801 * config/arm/arm_neon_builtins.def: New file.
20802 * config/arm/arm.c (neon_builtin_data): Move contents to
20803 arm_neon_builtins.def.
20804 (enum arm_builtins): Include neon builtin definitions.
20805 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20806 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20807
20808 2013-04-04 Marek Polacek <polacek@redhat.com>
20809
20810 PR tree-optimization/48186
20811 * predict.c (maybe_hot_frequency_p): Return false if
20812 HOT_BB_FREQUENCY_FRACTION is 0.
20813 (cgraph_maybe_hot_edge_p): Likewise.
20814
20815 2013-04-04 Richard Biener <rguenther@suse.de>
20816
20817 PR tree-optimization/56826
20818 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20819 more accurately.
20820
20821 2013-04-04 Richard Biener <rguenther@suse.de>
20822
20823 PR tree-optimization/56213
20824 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20825 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20826
20827 2013-04-04 Richard Biener <rguenther@suse.de>
20828
20829 PR tree-optimization/56837
20830 * tree-loop-distribution.c (classify_partition): For non-zero
20831 values require that the value has the same precision as its
20832 mode to be useful as memset value.
20833
20834 2013-04-03 Nick Clifton <nickc@redhat.com>
20835
20836 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20837 (fmssf4): Use fmsf.s on E3V5 architectures.
20838 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20839 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20840
20841 2013-04-03 Jeff Law <law@redhat.com>
20842
20843 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20844 (lra-eliminations.o): Likewise.
20845
20846 2013-04-03 Teresa Johnson <tejohnson@google.com>
20847
20848 * gcov-io.c (compute_working_sets): Moved most of body of old
20849 compute_working_sets here from profile.c.
20850 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20851 (gcov_working_set_t): Moved typedef here from basic-block.h
20852 (compute_working_set): Declare.
20853 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20854 (get_working_sets): Renamed from compute_working_set,
20855 replace most of body with call to new compute_working_sets.
20856 (get_exec_counts): Replace call to compute_working_sets
20857 to get_working_sets.
20858 * profile.h (get_working_sets): Renamed from compute_working_set.
20859 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20860 to get_working_sets.
20861 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20862 * gcov-dump.c (dump_working_sets): New function.
20863
20864 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20865
20866 * hwint.c (sext_hwi, zext_hwi): New functions.
20867 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20868 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20869 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20870 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20871 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20872 (sext_hwi, zext_hwi): New functions.
20873
20874 2013-04-03 Jeff Law <law@redhat.com>
20875
20876 PR tree-optimization/56799
20877 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20878 back test for widening conversion erroneously dropped in prior change.
20879
20880 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20881
20882 PR target/56809
20883 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20884 instead of next_real_insn.
20885
20886 2013-04-03 Marek Polacek <polacek@redhat.com>
20887
20888 PR sanitizer/55702
20889 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20890
20891 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20892
20893 PR target/56809
20894 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20895 next_real_insn.
20896 (thumb1_output_casesi): Likewise.
20897 (thumb2_output_casesi): Likewise.
20898
20899 2013-04-03 Richard Biener <rguenther@suse.de>
20900
20901 PR tree-optimization/56817
20902 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20903 Split out ...
20904 (tree_unroll_loops_completely_1): ... new function to manually
20905 walk the loop tree, properly defering outer loops of unrolled
20906 loops to later iterations.
20907
20908 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20909
20910 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20911 (vectorizable_load): Likewise.
20912 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20913 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20914
20915 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20916
20917 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20918 BIT_FIELD_REF.
20919
20920 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20921
20922 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20923
20924 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20925
20926 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20927
20928 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20929
20930 PR tree-optimization/56790
20931 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20932 folding.
20933
20934 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20935
20936 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20937 Handle VEC_MERGE.
20938 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20939 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20940 equal arguments.
20941
20942 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20943
20944 PR c/19449
20945 * tree.h (force_folding_builtin_constant_p): New decl.
20946 * builtins.c (force_folding_builtin_constant_p): New variable.
20947 (fold_builtin_constant_p): Fold immediately also if
20948 force_folding_builtin_constant_p.
20949
20950 2013-04-03 Richard Biener <rguenther@suse.de>
20951
20952 PR tree-optimization/56812
20953 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20954 DRs of the same interleaving chain are independent.
20955
20956 2013-04-02 Jason Merrill <jason@redhat.com>
20957
20958 * gdbinit.in (pbb): Use debug fn.
20959
20960 2013-04-02 Lawrence Crowl <crowl@google.com>
20961
20962 * sese.h (struct ivtype_map_elt_s): Remove unused.
20963 (extern debug_ivtype_map): Remove unused.
20964 (extern eq_ivtype_map_elts): Remove unused.
20965 * sese.c (debug_ivtype_map): Removed unused.
20966 (debug_ivtype_map_1): Removed unused.
20967 (debug_ivtype_elt): Remove unused.
20968 (eq_ivtype_map_elts): Remove unused.
20969
20970
20971 2013-04-02 Kai Tietz <ktietz@redhat.com>
20972
20973 PR target/52790
20974 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20975 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20976 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20977 function.
20978 (legitimize_pe_coff_symbol): Likewise.
20979 (is_imported_p): New helper-function.
20980 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20981 for Windows x64 targets.
20982 (ix86_expand_prologue): Optimize for pe-coff targets.
20983 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20984 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20985 medium/large code-model.
20986 (legitimize_pic_address): Likewise.
20987 (legitimize_tls_address): Likewise.
20988 (ix86_expand_call): Likewise.
20989 (x86_output_mi_thunk): Likewise.
20990 (get_dllimport_decl): Add new beimport argument.
20991 (construct_plt_address): Don't assert for x64 pe-coff targets.
20992 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20993 targets.
20994 (SYMBOL_FLAG_STUBVAR): New macro.
20995 (SYMBOL_REF_STUBVAR_P): Likewise.
20996 * config/i386/winnt.c (stub_list): New structure.
20997 (stub_head): New local variable.
20998 (i386_pe_record_stub): New function.
20999 (i386_pe_file_end): Emit refptr-stubs.
21000
21001 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21002
21003 PR rtl-optimization/56745
21004 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21005 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21006
21007 PR c++/34949
21008 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21009 and both of them are MEM_REFs, just compare first argument for
21010 equality and attempt to deal even with differing offsets.
21011
21012 PR c++/34949
21013 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21014 of gimple_clobber_p to be MEM_REF.
21015 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21016 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21017 after gimplification.
21018 * asan.c (get_mem_ref_of_assignment): Don't instrument
21019 gimple_clobber_p stmts.
21020 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21021 gimple_clobber_p stmt if they have MEM_REF lhs and
21022 are dead because of another gimple_clobber_p stmt.
21023 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21024 gimple_clobber_p stmts like debug stmts.
21025 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21026 that refer to unused VAR_DECLs or uninitialized values.
21027 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21028 gimple_clobber_p stmts if they refer to removed parameters.
21029 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21030 formatting.
21031
21032 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21033
21034 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21035 using SWI48 mode attribute.
21036
21037 2013-04-02 Wei Mi <wmi@google.com>
21038
21039 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21040 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21041 *<rotate_insn><mode>3_mask in i386.md.
21042
21043 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21044
21045 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21046
21047 2013-04-02 Richard Biener <rguenther@suse.de>
21048
21049 PR tree-optimization/56778
21050 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21051 Runtime alias tests are not supported for gather loads.
21052 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21053 stmts referenced from SSA operands before updating SSA form.
21054
21055 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21056 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21057
21058 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21059 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21060 * config/arm/cortex-a53.md: New file.
21061 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21062 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21063 * config/arm/arm.c (arm_issue_rate): Likewise.
21064 * config/arm/arm-tune.md: Regenerate
21065 * config/arm/arm-tables.opt: Regenerate.
21066 * config/arm/arm-cores.def: Add cortex-a53.
21067
21068 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21069
21070 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21071 non-static link.
21072
21073 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21074
21075 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21076 scalar load/store operations using B/H registers.
21077 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21078
21079 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21080
21081 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21082 scalar move.
21083 * config/aarch64/aarch64.c
21084 (aarch64_simd_scalar_immediate_valid_for_move): New.
21085 * config/aarch64/aarch64-protos.h
21086 (aarch64_simd_scalar_immediate_valid_for_move): New.
21087 * config/aarch64/constraints.md (Dh, Dq): New.
21088 * config/aarch64/iterators.md (hq): New.
21089
21090 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21091
21092 * reorg.c (get_branch_condition): Deal with conditional returns.
21093 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21094
21095 2013-04-01 Wei Mi <wmi@google.com>
21096
21097 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21098 Truncate operand 2 using %b asm operand modifier.
21099 (*<shift_insn><mode>3_mask): Ditto.
21100 (*<rotate_insn><mode>3_mask): Ditto.
21101
21102 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21103
21104 PR middle-end/56798
21105 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21106
21107 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21108
21109 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21110 of next_real_insn.
21111 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21112
21113 2013-03-30 Lawrence Crowl <crowl@google.com>
21114
21115 * dse.c (clear_alias_sets): Remove never set.
21116 (disqualified_clear_alias_sets): Remove never set.
21117 (clear_alias_mode_pool): Remove never set.
21118 (dse_step0): Remove condition that is never true.
21119 (canon_address): Remove condition that is never true.
21120 (dse_step7): Remove condition that is never true.
21121 (rest_of_handle_dse): Remove condition that is never true.
21122 (rest_of_handle_dse::did_global): Remove never read from above.
21123 (dse_step2_spill): Remove never called from above.
21124 (dse_step5_spill): Remove never called from above.
21125
21126 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21127
21128 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21129 JUMP_TABLE_DATA changes.
21130 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21131 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21132 (Insns) <jump_table_data>: New entry.
21133 * doc/tm.texi: Regenerate.
21134
21135 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21136
21137 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21138 for table jump at the end of a basic block using tablejump_p.
21139 * targhooks.c (default_invalid_within_doloop): Likewise.
21140 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21141 target hook implementation that is identical to the default hook.
21142 (rs6000_invalid_within_doloop): Remove.
21143
21144 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21145 unused variable from tablejump_p call.
21146
21147 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21148 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21149 (INSN_DELETED_P): Likewise.
21150 (emit_jump_table_data): New prototype.
21151 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21152 after 4th as unused.
21153 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21154 * sched-vis.c (print_insn): Likewise.
21155 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21156 insn for compatibility with back ends that use next_active_insn to
21157 identify jump table data.
21158 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21159 (remove_insn): Likewise.
21160 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21161 to be emitted.
21162 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21163 (emit_jump_table_data): New function.
21164
21165 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21166 basic block, a JUMP_TABLE_DATA never is.
21167 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21168 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21169 off from code handling real insns.
21170 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21171 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21172 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21173 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21174 is not a NONDEBUG_INSN_P.
21175 * ira-costs.c (scan_one_insn): Likewise.
21176 * jump.c (mark_all_labels): Likewise.
21177 (mark_jump_label_1): Likewise.
21178 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21179 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21180 a basic block.
21181 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21182 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21183 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21184 JUMP_TABLE_DATA_P insns.
21185 (calculate_elim_costs_all_insns): Likewise.
21186 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21187 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21188 (delete_output_reload): Code style fixups.
21189 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21190 insn flags on this non-insn.
21191 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21192 as scheduling barriers, for pre-change compatibility.
21193 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21194 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21195
21196 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21197 redundant JUMP_TABLE_DATA_P test.
21198 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21199 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21200 (frv_for_each_packet): Likewise.
21201 * config/i386/i386.c (min_insn_size): Likewise.
21202 (ix86_avoid_jump_mispredicts): Likewise.
21203 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21204 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21205 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21206 (mips16_insn_length): Robustify.
21207 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21208 (mips16_split_long_branches): Likewise.
21209 * config/pa/pa.c (pa_combine_instructions): Likewise.
21210 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21211 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21212 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21213 as contributing to pool range lengths.
21214 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21215 Remove redundant JUMP_TABLE_DATA_P test.
21216 (sh_loop_align): Likewise.
21217 (split_branches): Likewise.
21218 (sh_insn_length_adjustment): Likewise.
21219 * config/spu/spu.c (get_branch_target): Likewise.
21220
21221 2013-03-29 Jan Hubicka <jh@suse.cz>
21222
21223 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21224 gcov streaming; stream hot bb threshold to ltrans.
21225 * predict.c (get_hot_bb_threshold): Break out from ....
21226 (maybe_hot_count_p): ... here.
21227 (set_hot_bb_threshold): New function.
21228 * lto-section-in.c (lto_section_name): Add profile.
21229 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21230 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21231 and data-streamer.h
21232 (histogram_entry): New structure.
21233 (histogram, histogram_pool): New global vars.
21234 (histogram_hash): New structure.
21235 (histogram_hash::hash): New method.
21236 (histogram_hash::equal): Likewise.
21237 (account_time_size): New function.
21238 (cmp_counts): New function.
21239 (dump_histogram): New function.
21240 (ipa_profile_generate_summary): New function.
21241 (ipa_profile_write_summary): New function.
21242 (ipa_profile_read_summary): New function.
21243 (ipa_profile): Decide on threshold.
21244 (pass_ipa_profile): Add ipa_profile_write_summary and
21245 ipa_profile_read_summary.
21246 * Makefile.in (ipa.o): Update dependencies.
21247 * lto-streamer.h (LTO_section_ipa_profile): New section.
21248
21249 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21250
21251 * tree.h (VAR_P): New.
21252
21253 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21254
21255 PR lto/56777
21256 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21257
21258 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21259
21260 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21261 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21262 (control_flow_insn_p): Likewise.
21263 * cfgrtl.c (duplicate_insn_chain): Likewise.
21264 * final.c (get_attr_length_1): Likewise.
21265 (shorten_branches): Likewise.
21266 (final_scan_insn): Likewise.
21267 * function.c (instantiate_virtual_regs): Likewise.
21268 * gcse.c (insert_insn_end_basic_block): Likewise.
21269 * ira-costs.c (scan_one_insn): Likewise.
21270 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21271 * lra.c (check_rtl): Likewise.
21272 * reload1.c (elimination_costs_in_insn): Likewise.
21273 * reorg.c (follow_jumps): Likewise.
21274
21275 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21276 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21277 (thumb_far_jump_used_p): Likewise.
21278 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21279 (workaround_speculation): Likewise.
21280 (add_sched_insns_for_speculation): Likewise.
21281 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21282 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21283 (frv_for_each_packet): Likewise.
21284 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21285 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21286 (final_emit_insn_group_barriers): Likewise.
21287 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21288 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21289 (mips16_insn_length): Likewise.
21290 * config/pa/pa.c (pa_reorg): Likewise.
21291 (pa_combine_instructions): Likewise.
21292 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21293 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21294 (sh_reorg): Likewise.
21295 (split_branches): Likewise.
21296 * config/spu/spu.c (get_branch_target): Likewise.
21297
21298 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21299 JUMP_TABLE_DATA_P.
21300
21301 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21302
21303 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21304 Fix declaration name.
21305
21306 2013-03-28 Lawrence Crowl <crowl@google.com>
21307
21308 * graphds.h (struct graph.indicies): Remove unused.
21309 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21310 (SCOP_ORIGINAL_PDDRS): Remove unused.
21311 * sese.h (extern insert_loop_close_phis): Removed unused.
21312 (extern insert_guard_phis): Removed unused.
21313 (extern ivtype_map_elt_info): Removed unused.
21314 (new_ivtype_map_elt): Removed unused.
21315 * sese.c (ivtype_map_elt_info): Removed unused.
21316
21317 2013-03-28 Lawrence Crowl <crowl@google.com>
21318
21319 * Makefile.in: Add several missing include dependences.
21320 (DUMPFILE_H): New.
21321 (test-dump.o): New. This object is not added to any executable,
21322 but is present for ad-hoc testing.
21323 * bitmap.c
21324 (debug (const bitmap_head_def &)): New.
21325 (debug (const bitmap_head_def *)): New.
21326 * bitmap.h
21327 (extern debug (const bitmap_head_def &)): New.
21328 (extern debug (const bitmap_head_def *)): New.
21329 * cfg.c
21330 (debug (edge_def &)): New.
21331 (debug (edge_def *)): New.
21332 * cfghooks.c
21333 (debug (basic_block_def &)): New.
21334 (debug (basic_block_def *)): New.
21335 * dumpfile.h
21336 (dump_node (const_tree, int, FILE *)): Correct source file.
21337 * dwarf2out.c
21338 (debug (die_struct &)): New.
21339 (debug (die_struct *)): New.
21340 * dwarf2out.h
21341 (extern debug (die_struct &)): New.
21342 (extern debug (die_struct *)): New.
21343 * gimple-pretty-print.c
21344 (debug (gimple_statement_d &)): New.
21345 (debug (gimple_statement_d *)): New.
21346 * gimple-pretty-print.h
21347 (extern debug (gimple_statement_d &)): New.
21348 (extern debug (gimple_statement_d *)): New.
21349 * ira-build.c
21350 (debug (ira_allocno_copy &)): New.
21351 (debug (ira_allocno_copy *)): New.
21352 (debug (ira_allocno &)): New.
21353 (debug (ira_allocno *)): New.
21354 * ira-int.h
21355 (extern debug (ira_allocno_copy &)): New.
21356 (extern debug (ira_allocno_copy *)): New.
21357 (extern debug (ira_allocno &)): New.
21358 (extern debug (ira_allocno *)): New.
21359 * ira-lives.c
21360 (debug (live_range &)): New.
21361 (debug (live_range *)): New.
21362 * lra-int.h
21363 (debug (lra_live_range &)): New.
21364 (debug (lra_live_range *)): New.
21365 * lra-lives.c
21366 (debug (lra_live_range &)): New.
21367 (debug (lra_live_range *)): New.
21368 * omega.c
21369 (debug (omega_pb_d &)): New.
21370 (debug (omega_pb_d *)): New.
21371 * omega.h
21372 (extern debug (omega_pb_d &)): New.
21373 (extern debug (omega_pb_d *)): New.
21374 * print-rtl.c
21375 (debug (const rtx_def &)): New.
21376 (debug (const rtx_def *)): New.
21377 * print-tree.c
21378 (debug_tree (tree): Move within file.
21379 (debug_raw (const tree_node &)): New.
21380 (debug_raw (const tree_node *)): New.
21381 (dump_tree_via_hooks (const tree_node *, int)): New.
21382 (debug (const tree_node &)): New.
21383 (debug (const tree_node *)): New.
21384 (debug_verbose (const tree_node &)): New.
21385 (debug_verbose (const tree_node *)): New.
21386 (debug_head (const tree_node &)): New.
21387 (debug_head (const tree_node *)): New.
21388 (debug_body (const tree_node &)): New.
21389 (debug_body (const tree_node *)): New.
21390 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21391 (debug (vec<tree, va_gc> &)): New.
21392 (debug (vec<tree, va_gc> *)): New.
21393 * rtl.h
21394 (extern debug (const rtx_def &)): New.
21395 (extern debug (const rtx_def *)): New.
21396 * sbitmap.c
21397 (debug_raw (simple_bitmap_def &)): New.
21398 (debug_raw (simple_bitmap_def *)): New.
21399 (debug (simple_bitmap_def &)): New.
21400 (debug (simple_bitmap_def *)): New.
21401 * sbitmap.h
21402 (extern debug (simple_bitmap_def &)): New.
21403 (extern debug (simple_bitmap_def *)): New.
21404 (extern debug_raw (simple_bitmap_def &)): New.
21405 (extern debug_raw (simple_bitmap_def *)): New.
21406 * sel-sched-dump.c
21407 (debug (vinsn_def &)): New.
21408 (debug (vinsn_def *)): New.
21409 (debug_verbose (vinsn_def &)): New.
21410 (debug_verbose (vinsn_def *)): New.
21411 (debug (expr_def &)): New.
21412 (debug (expr_def *)): New.
21413 (debug_verbose (expr_def &)): New.
21414 (debug_verbose (expr_def *)): New.
21415 (debug (vec<rtx> &)): New.
21416 (debug (vec<rtx> *)): New.
21417 * sel-sched-dump.h
21418 (extern debug (vinsn_def &)): New.
21419 (extern debug (vinsn_def *)): New.
21420 (extern debug_verbose (vinsn_def &)): New.
21421 (extern debug_verbose (vinsn_def *)): New.
21422 (extern debug (expr_def &)): New.
21423 (extern debug (expr_def *)): New.
21424 (extern debug_verbose (expr_def &)): New.
21425 (extern debug_verbose (expr_def *)): New.
21426 (extern debug (vec<rtx> &)): New.
21427 (extern debug (vec<rtx> *)): New.
21428 * sel-sched-ir.h
21429 (_list_iter_cond_expr): Make inline instead of static.
21430 * sreal.c
21431 (debug (sreal &)): New.
21432 (debug (sreal *)): New.
21433 * sreal.h
21434 (extern debug (sreal &)): New.
21435 (extern debug (sreal *)): New.
21436 * tree.h
21437 (extern debug_raw (const tree_node &)): New.
21438 (extern debug_raw (const tree_node *)): New.
21439 (extern debug (const tree_node &)): New.
21440 (extern debug (const tree_node *)): New.
21441 (extern debug_verbose (const tree_node &)): New.
21442 (extern debug_verbose (const tree_node *)): New.
21443 (extern debug_head (const tree_node &)): New.
21444 (extern debug_head (const tree_node *)): New.
21445 (extern debug_body (const tree_node &)): New.
21446 (extern debug_body (const tree_node *)): New.
21447 (extern debug (vec<tree, va_gc> &)): New.
21448 (extern debug (vec<tree, va_gc> *)): New.
21449 * tree-cfg.c
21450 (debug (struct loop &)): New.
21451 (debug (struct loop *)): New.
21452 (debug_verbose (struct loop &)): New.
21453 (debug_verbose (struct loop *)): New.
21454 * tree-dump.c: Add header dependence.
21455 * tree-flow.h
21456 (extern debug (struct loop &)): New.
21457 (extern debug (struct loop *)): New.
21458 (extern debug_verbose (struct loop &)): New.
21459 (extern debug_verbose (struct loop *)): New.
21460 * tree-data-ref.c
21461 (debug (data_reference &)): New.
21462 (debug (data_reference *)): New.
21463 (debug (vec<data_reference_p> &)): New.
21464 (debug (vec<data_reference_p> *)): New.
21465 (debug (vec<ddr_p> &)): New.
21466 (debug (vec<ddr_p> *)): New.
21467 * tree-data-ref.h
21468 (extern debug (data_reference &)): New.
21469 (extern debug (data_reference *)): New.
21470 (extern debug (vec<data_reference_p> &)): New.
21471 (extern debug (vec<data_reference_p> *)): New.
21472 (extern debug (vec<ddr_p> &)): New.
21473 (extern debug (vec<ddr_p> *)): New.
21474 * tree-ssa-alias.c
21475 (debug (pt_solution &)): New.
21476 (debug (pt_solution *)): New.
21477 * tree-ssa-alias.h
21478 (extern debug (pt_solution &)): New.
21479 (extern debug (pt_solution *)): New.
21480 * tree-ssa-alias.c
21481 (debug (_var_map &)): New.
21482 (debug (_var_map *)): New.
21483 (debug (tree_live_info_d &)): New.
21484 (debug (tree_live_info_d *)): New.
21485 * tree-ssa-alias.h
21486 (extern debug (_var_map &)): New.
21487 (extern debug (_var_map *)): New.
21488 (extern debug (tree_live_info_d &)): New.
21489 (extern debug (tree_live_info_d *)): New.
21490
21491 2013-03-28 Jan Hubicka <jh@suse.cz>
21492
21493 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21494
21495 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21496
21497 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21498 record only when desired or required.
21499
21500 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21501
21502 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21503 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21504
21505 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21506
21507 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21508 (*andsi3_compare0_uxtw): New pattern.
21509 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21510 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21511
21512 2013-03-28 Jan Hubicka <jh@suse.cz>
21513
21514 * data-streamer-in.c (streamer_read_gcov_count): New function.
21515 * gimple-streamer-out.c: Include value-prof.h.
21516 (output_gimple_stmt): Output histogram.
21517 (output_bb): Use streamer_write_gcov_count.
21518 * value-prof.c: Include data-streamer.h
21519 (dump_histogram_value): Add HIST_TYPE_MAX.
21520 (stream_out_histogram_value): New function.
21521 (stream_in_histogram_value): New function.
21522 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21523 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21524 * data-streamer-out.c (streamer_write_gcov_count): New function.
21525 (streamer_write_gcov_count_stream): New function.
21526 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21527 (lto_output_node): Likewise.
21528 (input_node, input_edge): Likewise.
21529 * lto-streamer-out.c (output_cfg): Update streaming.
21530 * lto-streamer-in.c (input_cfg): Likewise.
21531 * data-streamer.h (streamer_write_gcov_count,
21532 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21533 * gimple-streamer-in.c: Include value-prof.h
21534 (input_gimple_stmt): Input histograms.
21535 (input_bb): Update profile streaming.
21536
21537 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21538
21539 * genmodes.c (emit_max_int): New function.
21540 (emit_insn_modes_h): Added call to emit_max_function.
21541 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21542 Added doc.
21543 * machmode.def: Fixed comment.
21544
21545 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21546
21547 * combine.c (try_combine): Removed useless assert.
21548 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21549
21550 2013-03-28 Marek Polacek <polacek@redhat.com>
21551 Richard Biener <rguenther@suse.de>
21552
21553 PR tree-optimization/56695
21554 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21555 build signed result of a vector comparison.
21556 * tree-cfg.c (verify_gimple_comparison): Check that a result
21557 of a vector comparison has signed type.
21558
21559 2013-03-28 Richard Biener <rguenther@suse.de>
21560
21561 PR tree-optimization/37021
21562 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21563 do not restrict gaps between groups.
21564 * tree-vect-stmts.c (vectorizable_load): Properly account for
21565 a gap between groups.
21566
21567 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21568
21569 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21570 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21571 is not enabled.
21572
21573 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21574
21575 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21576 * doc/extend.texi (Named Address Spaces): Ditto.
21577 (Variable Attributes): Ditto.
21578
21579 2013-03-27 Kai Tietz <ktietz@redhat.com>
21580
21581 * config.build: Add support for cygwin x64 target.
21582 * config.gcc: Likewise.
21583 * config.host: Likewise.
21584 * configure.ac: Likewise
21585 * configure: Regenerated.
21586
21587 2013-03-27 Kai Tietz <ktietz@redhat.com>
21588
21589 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21590 * config/i386/t-cygwin-w64: New file.
21591 * config/i386/cygwin-w64.h: New file.
21592 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21593 and add support for x64-cygwin target.
21594 (CPP_SPEC): Likewise.
21595 (CXX_WRAP_SPEC_LIST): Undefine before define.
21596 (LIBGCJ_SONAME): Use 15 as version.
21597
21598 2013-03-27 Richard Biener <rguenther@suse.de>
21599
21600 PR tree-optimization/56716
21601 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21602 dumping for ref nodes.
21603
21604 2013-03-27 Martin Jambor <mjambor@suse.cz>
21605
21606 PR tree-optimization/55334
21607 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21608 restricted pointers to arrays.
21609
21610 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21611
21612 * Makefile.in (.SUFFIXES): Add .cc.
21613 (.c.o): Apply same recipe for implicit rule .cc.o.
21614
21615 2013-03-27 Richard Biener <rguenther@suse.de>
21616
21617 PR tree-optimization/37021
21618 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21619 REALPART/IMAGPART_EXPRs around the supported refs.
21620 * tree-ssa-structalias.c (find_func_aliases): Assume that
21621 floating-point values are not used to transfer pointers.
21622
21623 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21624
21625 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21626 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21627 * doc/tm.texi: Regenerate.
21628 * targhooks.h (default_has_ifunc_p): New.
21629 * targhooks.c (default_has_ifunc_p): Ditto.
21630 * config/linux-protos.h: New file.
21631 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21632 hook for linux which disables support of indirect functions in android.
21633 * config/linux-android.c: New file.
21634 * config/t-linux-android.c: Ditto.
21635 * config.gcc: Added new object file linux-android.o.
21636 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21637 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21638 * varasm.c (do_assemble_alias): Likewise.
21639 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21640 doesn't support indirect functions.
21641 * configure: Regenerate.
21642
21643 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21644
21645 PR target/56102
21646 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21647 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21648 mult-word mode.
21649
21650 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21651
21652 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21653
21654 2013-03-27 Terry Guo <terry.guo@arm.com>
21655
21656 * config/arm/arm-cores.def: Added core cortex-r7.
21657 * config/arm/arm-tune.md: Regenerated.
21658 * config/arm/arm-tables.opt: Regenerated.
21659 * doc/invoke.texi: Added entry for core cortex-r7.
21660
21661 2013-03-27 Walter Lee <walt@tilera.com>
21662
21663 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21664 double-decrement of next_scratch_regno.
21665
21666 2013-03-27 Walter Lee <walt@tilera.com>
21667
21668 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21669 input operands.
21670 (insn_v1mulus): Ditto.
21671 (insn_v2muls): Ditto.
21672
21673 2013-03-27 Walter Lee <walt@tilera.com>
21674
21675 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21676 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21677
21678 2013-03-27 Walter Lee <walt@tilera.com>
21679
21680 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21681 (*sibcall_value): Ditto.
21682
21683 2013-03-27 Walter Lee <walt@tilera.com>
21684
21685 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21686 (insn_mnz_v8qi): ... this ...
21687 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21688 vector equivalent.
21689 (insn_v<n>mnz): Replaced by ...
21690 (insn_v1mnz): ... this ...
21691 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21692 equivalent.
21693 (insn_mz_<mode>): Replaced by ...
21694 (insn_mz_v8qi): ... this ...
21695 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21696 vector equivalent.
21697 (insn_v<n>mz): Replaced by ...
21698 (insn_v1mz): ... this ...
21699 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21700 equivalent.
21701
21702 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21703
21704 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21705
21706 2013-03-26 Roland McGrath <mcgrathr@google.com>
21707
21708 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21709 than fprintf with a non-constant, non-format string.
21710
21711 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21712
21713 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21714 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21715 operand 0 predicate.
21716 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21717 attribute. Use general_x64nomem_operand as operand 1 predicate.
21718 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21719 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21720 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21721 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21722 (mov<mode>_insv_1): Remove expander. Merge insn with
21723 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21724 Use general_x64nomem_operand as operand 1 predicate.
21725 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21726 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21727 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21728 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21729 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21730 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21731 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21732 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21733 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21734 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21735 (general_x64nomem_operand): Ditto.
21736
21737 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21738
21739 * config/rtems.opt: Add -pthread option.
21740
21741 2013-03-26 Richard Biener <rguenther@suse.de>
21742
21743 * alias.c (find_base_term): Avoid redundant and not used recursion.
21744 (base_alias_check): Get the initial base term from the caller.
21745 (true_dependence_1): Compute and pass base terms to base_alias_check.
21746 (write_dependence_p): Likewise.
21747 (may_alias_p): Likewise.
21748
21749 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21750
21751 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21752 PC-relative load in SI modes and above only.
21753
21754 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21755
21756 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21757 * config/arm/iwmmxt.md (WCGR0): Update.
21758 (WCGR1, WCGR2, WCGR3): Likewise.
21759
21760 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21761
21762 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21763 Use x64 and nox64 isa attributes.
21764
21765 2013-03-26 Richard Biener <rguenther@suse.de>
21766
21767 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21768 alignment computations and rely on get_object_alignment_1
21769 for the !TYPE_P case.
21770 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21771
21772 2013-03-26 Walter Lee <walt@tilera.com>
21773
21774 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21775 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21776
21777 2013-03-25 Jeff Law <law@redhat.com>
21778
21779 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21780 check for INTEGRAL_TYPE_P that was missing due to checking in
21781 wrong version of prior patch.
21782
21783 2013-03-25 Walter Lee <walt@tilera.com>
21784
21785 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21786 TILEGX_INSN_SHUFFLEBYTES1.
21787 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21788 shufflebytes1.
21789 (tilegx_builtins): Ditto.
21790 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21791
21792 2013-03-25 Walter Lee <walt@tilera.com>
21793
21794 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21795 (floatunssisf2): New pattern.
21796 (floatsidf2): New pattern.
21797 (floatunssidf2): New pattern.
21798
21799 2013-03-25 Walter Lee <walt@tilera.com>
21800
21801 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21802 tests for constraint J, K, N, P.
21803
21804 2013-03-25 Walter Lee <walt@tilera.com>
21805
21806 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21807 Use indirect/pcrel encoding.
21808 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21809 Ditto.
21810
21811 2013-03-25 Steve Ellcey <sellcey@mips.com>
21812
21813 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21814 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21815 * config/mips/mips.c (mips_option_override): Set IMADD default.
21816 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21817 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21818 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21819 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21820 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21821
21822 2013-03-25 Jeff Law <law@redhat.com>
21823
21824 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21825 slightly to avoid creating and folding useless trees. Simplify
21826 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21827
21828 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21829
21830 * config/i386/i386.md (*zero_extendsidi2): Merge with
21831 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21832 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21833 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21834 targets. Clarify comment.
21835
21836 2013-03-25 Martin Jambor <mjambor@suse.cz>
21837
21838 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21839 pass-through jump functions differently.
21840 (ipa_read_jump_function): Likewise. Also use setter functions to set
21841 up jump functions.
21842
21843 2013-03-25 Martin Jambor <mjambor@suse.cz>
21844
21845 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21846 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21847 process it.
21848 (ipa_get_indirect_edge_target): New function.
21849 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21850 ipa_get_indirect_edge_target. Update all callers.
21851 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21852 ipa_get_indirect_edge_target_1 instead of calling
21853 ipa_get_indirect_edge_target.
21854 (create_specialized_node): Pass aggvlas to
21855 ipcp_discover_new_direct_edges.
21856
21857 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21858
21859 * config/arm/arm.md (f_sels, f_seld): New types.
21860 (*cmov<mode>): New pattern.
21861 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21862 predicate.
21863
21864 2013-03-25 Kai Tietz <ktietz@redhat.com>
21865
21866 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21867 POSIX-printf for mingw-hosted builds.
21868
21869 2013-03-25 Richard Biener <rguenther@suse.de>
21870
21871 PR middle-end/56694
21872 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21873 must-not-throw stmt location.
21874
21875 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21876
21877 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21878 Emit load-acquire versions when acq is true.
21879 (arm_emit_store_exclusive): Add rel parameter.
21880 Emit store-release versions when rel is true.
21881 (arm_split_compare_and_swap): Use acquire-release instructions
21882 instead.
21883 of barriers when appropriate.
21884 (arm_split_atomic_op): Likewise.
21885 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21886 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21887 (VUNSPEC_SLX): Likewise.
21888 (VUNSPEC_LDA): Likewise.
21889 (VUNSPEC_STL): Likewise.
21890 * config/arm/sync.md (atomic_load<mode>): New pattern.
21891 (atomic_store<mode>): Likewise.
21892 (arm_load_acquire_exclusive<mode>): Likewise.
21893 (arm_load_acquire_exclusivesi): Likewise.
21894 (arm_load_acquire_exclusivedi): Likewise.
21895 (arm_store_release_exclusive<mode>): Likewise.
21896
21897 2013-03-25 Catherine Moore <clm@codesourcery.com>
21898
21899 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21900 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21901 * config/mip/predicates.md (lwsp_swsp_operand,
21902 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21903 sb16_operand, db4_operand, db7_operand, ib3_operand,
21904 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21905 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21906 andi16_operand): New predicates.
21907 * config/mips/mips.md (compression): New attribute.
21908 (enabled): New attribute.
21909 (length): Consider compression in computing length.
21910 (shift_compression): New code attribute.
21911 (*add<mode>3): New operands. Record compression.
21912 (sub<mode>3): Likewise.
21913 (one_cmpl<mode>2): Likewise.
21914 (*and<mode>3): Likewise.
21915 (*ior<mode>3): Likewise.
21916 (unnamed pattern for xor): Likewise.
21917 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21918 (*<optab><mode>3): Likewise.
21919 (*mov<mode>_internal: Likewise.
21920 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21921 (mips_unsigned_immediate_p): New.
21922 (umips_lwsp_swsp_address_p): New.
21923 (m16_based_address_p): New.
21924 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21925 (mips_unsigned_immediate_p): New prototype.
21926 (lwsp_swsp_address_p): New prototype.
21927 (m16_based_address_p): New prototype.
21928 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21929 (mips_signed_immediate_p): New function.
21930 (m16_based_address_p): New function.
21931 (lwsp_swsp_address_p): New function.
21932 (mips_print_operand_punctuation): Recognize short delay slot insns
21933 for microMIPS.add<mode>3"
21934
21935 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21936
21937 PR target/56720
21938 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21939 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21940
21941 2013-03-25 Richard Biener <rguenther@suse.de>
21942
21943 PR tree-optimization/56689
21944 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21945 any edge.
21946
21947 2013-03-25 Richard Biener <rguenther@suse.de>
21948
21949 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21950 of bitmap.
21951 (memory_references): Likewise.
21952 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21953 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21954 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21955 (gather_mem_refs_in_loops): Fold into ...
21956 (analyze_memory_references): ... this. Move initialization
21957 to tree_ssa_lim_initialize.
21958 (fill_always_executed_in): Rename to ...
21959 (fill_always_executed_in_1): ... this.
21960 (fill_always_executed_in): Move contains_call computation to
21961 this new function from ...
21962 (tree_ssa_lim_initialize): ... here.
21963 (tree_ssa_lim): Call fill_always_executed_in.
21964
21965 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21966
21967 * postreload.c (reload_combine): Fix code detecting returns.
21968
21969 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21970
21971 * function.c (emit_use_return_register_into_block): On cc0 targets,
21972 do not emit the sequence between cc0 setter and user.
21973
21974 2013-03-25 Kai Tietz <ktietz@redhat.com>
21975
21976 * config/i386/predicates.md (local_symbolic_operand): Interpret
21977 dll-imported symbols as none-local.
21978
21979 2013-03-25 Richard Biener <rguenther@suse.de>
21980
21981 * tree-ssa-loop-im.c (struct depend): Remove.
21982 (struct lim_aux_data): Make depends a vec of gimples.
21983 (free_lim_aux_data): Adjust.
21984 (add_dependency): Likewise.
21985 (set_level): Likewise.
21986
21987 2013-03-25 Richard Biener <rguenther@suse.de>
21988
21989 PR middle-end/56434
21990 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21991 the pointer returned by calls with ECF_MALLOC set.
21992
21993 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21994
21995 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21996
21997 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21998
21999 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22000 using MMXMODE mode iterator.
22001 (*move<mode>_internal): Merge with *movv2sf_internal and
22002 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22003
22004 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22005
22006 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22007 (record_last_mem_set_info): Likewise.
22008
22009 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22010 of XNEWVEC followed by memset.
22011 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22012
22013 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22014
22015 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22016 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22017 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22018 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22019 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22020 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22021 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22022 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22023 BARRIER_P instead of GET_CODE.
22024
22025 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22026
22027 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22028 inaccuracy in the probing code.
22029
22030 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22031 (ctrapdi4): Likewise.
22032
22033 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22034
22035 * calls.c (expand_call): Add missing guard to code handling return
22036 of non-BLKmode structures in MSB.
22037 * function.c (expand_function_end): Likewise.
22038
22039 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22040
22041 * combine.c (try_combine): Adjust comment. Do not add the set of
22042 insn #0 if the destination indirectly is set or dies in insn #2.
22043 Tidy up code to distribute a new note.
22044
22045 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22046
22047 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22048 also for alternatives 16 and 17.
22049
22050 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22051
22052 * config/i386/sse.md (*mov<mode>_internal): Merge with
22053 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22054 Emit insn template depending on type attribute. Use
22055 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22056 movd instead of movq mnemonic for interunit moves. Rewrite mode
22057 attribute calculation. Remove unit attribute calculation.
22058 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22059 Set prefix_data16 attribute for DImode ssemov types.
22060 Use Ym instead of y for SSE-MMX conversion alternatives.
22061 Reorder operand constraints.
22062
22063 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22064
22065 * df.h (df_insn_delete): Adjust prototype.
22066 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22067 and let it decide whether mark the basic block dirty.
22068 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22069 * df-scan.c (df_insn_info_delete): New helper function, split
22070 off from df_insn_delete.
22071 (df_scan_free_bb_info): Use it.
22072 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22073 Likewise.
22074 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22075 that the insn is actually an insn and it has a non-NULL basic block.
22076 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22077
22078 2013-03-22 Richard Biener <rguenther@suse.de>
22079
22080 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22081 dep_ref members.
22082 (mem_ref_alloc): Do not allocate them.
22083 (refs_independent_p): Do not query or maintain a cache.
22084
22085 2013-03-22 Richard Biener <rguenther@suse.de>
22086
22087 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22088 (gather_mem_refs_in_loops): Do not compute it.
22089 (analyze_memory_references): Do not allocate it.
22090 (tree_ssa_lim_finalize): Do not free it.
22091 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22092
22093 2013-03-22 Richard Biener <rguenther@suse.de>
22094
22095 * is-a.h (as_a): Use gcc_checking_assert.
22096
22097 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22098
22099 * config/aarch64/aarch64.c (aarch64_print_operand): New
22100 format specifier for printing a constant in hex.
22101 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22102 format specifier for printing second operand.
22103
22104 2013-03-22 Richard Biener <rguenther@suse.de>
22105
22106 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22107 bitmaps.
22108 (gather_mem_refs_in_loops): Perform store accumulation here.
22109 (create_vop_ref_mapping_loop): Remove.
22110 (create_vop_ref_mapping): Likewise.
22111 (analyze_memory_references): Initialize refs_stored_in_loop.
22112 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22113 (record_indep_loop): Remove.
22114 (record_dep_loop): New function.
22115 (ref_indep_loop_p_1): Adjust to only walk over references
22116 in the loop, not its subloops.
22117 (ref_indep_loop_p): Rename to ...
22118 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22119 maintaining a more fine-grained cache.
22120 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22121 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22122
22123 2013-03-22 Richard Biener <rguenther@suse.de>
22124
22125 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22126 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22127 aggregate mem_ref_loc.
22128 (free_mem_ref_locs): Inline into ...
22129 (memref_free): ... this and adjust.
22130 (mem_ref_alloc): Adjust.
22131 (mem_ref_locs_alloc): Remove.
22132 (record_mem_ref_loc): Adjust.
22133 (get_all_locs_in_loop): Rewrite into ...
22134 (for_all_locs_in_loop): ... this iterator.
22135 (rewrite_mem_ref_loc): New functor.
22136 (rewrite_mem_refs): Use for_all_locs_in_loop.
22137 (sm_set_flag_if_changed): New functor.
22138 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22139 (ref_always_accessed): New functor.
22140 (ref_always_accessed_p): Use for_all_locs_in_loop.
22141
22142 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22143
22144 * tree-pass.h (PROP_gimple_lvec): New.
22145 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22146 (init_optimization_passes): Move pass_lower_vector.
22147 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22148 PROP_gimple_lvec.
22149 (pass_lower_vector): Provide PROP_gimple_lvec.
22150 (pass_lower_vector_ssa): Likewise.
22151 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22152
22153 2013-03-21 Mark Wielaard <mjw@redhat.com>
22154
22155 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22156
22157 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22158
22159 * config/i386/i386.md (*movdi_internal): Disparage slightly
22160 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22161 conversion alternatives.
22162
22163 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22164
22165 PR middle-end/48087
22166 * diagnostic.def (DK_WERROR): New kind.
22167 * diagnostic.h (werrorcount): Define.
22168 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22169 promoted to DK_ERROR, increment DK_WERROR counter instead of
22170 DK_ERROR counter.
22171 * toplev.c (toplev_main): Call print_ignored_options even if
22172 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22173 even if just werrorcount is non-zero.
22174
22175 PR debug/55608
22176 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22177 on failure.
22178 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22179 (string_cst_pool_decl): New function.
22180 (optimize_one_addr_into_implicit_ptr): New function.
22181 (resolve_addr_in_expr): Optimize DWARF location expression
22182 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22183 which doesn't live in memory, but has DW_AT_location or
22184 DW_AT_const_value, or refers to a string literal, into
22185 DW_OP_GNU_implicit_pointer.
22186 (optimize_location_into_implicit_ptr): New function.
22187 (resolve_addr): If removing DW_AT_location of a variable because
22188 it was DW_OP_addr of address of the variable, but the variable doesn't
22189 live in memory, try to emit const value attribute for the initializer.
22190
22191 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22192
22193 * tree.h (VECTOR_TYPE_P): New macro.
22194 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22195 TYPE_MODE): Use it.
22196 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22197 VEC_COND_EXPR cannot be lvalues.
22198 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22199
22200 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22201
22202 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22203 Restrict the transformation to equal modes.
22204
22205 2013-03-21 Richard Biener <rguenther@suse.de>
22206
22207 PR tree-optimization/39326
22208 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22209 (MEM_ANALYZABLE): Adjust.
22210 (record_mem_ref_loc): Move bitmap ops ...
22211 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22212 unanalyzable refs, do not record locations for it.
22213 (analyze_memory_references): Allocate ref zero as shared
22214 unanalyzable ref.
22215 (refs_independent_p): Do not test for unanalyzed mems here.
22216 (ref_indep_loop_p_1): Special-case disambiguation against
22217 the unanalyzed ref.
22218 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22219
22220 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22221
22222 * config/arm/arm-protos.h (tune_params): Add
22223 prefer_neon_for_64bits field.
22224 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22225 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22226 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22227 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22228 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22229 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22230 (arm_option_override): Handle -mneon-for-64bits new option.
22231 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22232 (prefer_neon_for_64bits): Declare new variable.
22233 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22234 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22235 (arch_enabled): Handle new arch types. Remove support for onlya8
22236 and nota8.
22237 (one_cmpldi2): Use new arch names.
22238 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22239 * config/arm/arm.opt (mneon-for-64bits): Add option.
22240 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22241 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22242 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22243 of onlya8.
22244 * doc/invoke.texi (-mneon-for-64bits): Document.
22245
22246 2013-03-21 Richard Biener <rguenther@suse.de>
22247
22248 PR tree-optimization/39326
22249 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22250 (sort_bbs_in_loop_postorder_cmp): New function.
22251 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22252
22253 2013-03-21 Richard Biener <rguenther@suse.de>
22254
22255 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22256 (vect_insert_into_interleaving_chain): Likewise.
22257 (vect_drs_dependent_in_basic_block): Inline ...
22258 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22259 split out from ...
22260 (vect_analyze_data_ref_dependence): ... here. Simplify.
22261 (vect_check_interleaving): Simplify.
22262 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22263 (vect_slp_analyze_data_ref_dependences): ... this new function.
22264 (dr_group_sort_cmp): New function.
22265 (vect_analyze_data_ref_accesses): Compute data-reference groups
22266 here instead of in vect_analyze_data_ref_dependence. Use
22267 a more efficient algorithm.
22268 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22269 vect_slp_analyze_data_ref_dependences. Call
22270 vect_analyze_data_ref_accesses earlier.
22271 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22272 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22273 (vect_slp_analyze_data_ref_dependences): New prototype.
22274
22275 2013-03-21 Richard Biener <rguenther@suse.de>
22276
22277 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22278 ref is stored in the loop.
22279 (find_refs_for_sm): Walk only over all stores.
22280 (store_motion_loop): Allocate from lim_bitmap_obstack.
22281 (store_motion): Likewise.
22282
22283 2013-03-21 Richard Biener <rguenther@suse.de>
22284
22285 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22286 Update virtual SSA form.
22287
22288 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22289
22290 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22291 * configure: Regenerate.
22292 * config.in: Regenerate.
22293 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22294 if !HAVE_LD_EH_FRAME_CIEV3.
22295
22296 2013-03-21 Richard Biener <rguenther@suse.de>
22297
22298 * tree-cfg.c (verify_expr_no_block): New function.
22299 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22300 nor DECL_VALUE_EXPR have locations with associated blocks.
22301 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22302 (clear_unused_block_pointer): Remove code dealing with
22303 blocks in DECL_DEBUG_EXPR locations.
22304
22305 2013-03-21 Richard Biener <rguenther@suse.de>
22306
22307 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22308 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22309 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22310 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22311 instead of DECL_DEBUG_EXPR_IS_FROM.
22312 * gimplify.c (gimplify_modify_expr): Likewise.
22313 * tree-cfg.c (verify_expr_location_1): Likewise.
22314 * tree-complex.c (create_one_component_var): Likewise.
22315 * tree-sra.c (create_access_replacement): Likewise.
22316 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22317 (clear_unused_block_pointer): Likewise.
22318 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22319 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22320 * var-tracking.c (var_debug_decl): Likewise.
22321 (track_expr_p): Likewise.
22322 * tree-inline.c (add_local_variables): Likewise. Set
22323 DECL_HAS_DEBUG_EXPR_P after copying it.
22324 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22325 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22326
22327 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22328
22329 PR bootstrap/56656
22330 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22331 * configure: Regenerate.
22332 * config.in: Regenerate.
22333 * config/i386/i386.md (*movdf_internal): Use
22334 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22335 movd instead of movq mnemonic for interunit moves.
22336 (*movdi_internal): Ditto.
22337
22338 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22339
22340 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22341 (abd<mode>_3): New pattern.
22342 (aba<mode>_3): New pattern.
22343 (fabd<mode>_3): New pattern.
22344
22345 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22346
22347 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22348 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22349 occurrence of REGISTER_PREFIX as its empty string.
22350
22351 2013-03-20 Jeff Law <law@redhat.com>
22352
22353 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22354 addititional equivalences for equality comparisons between an SSA_NAME
22355 and a constant where the SSA_NAME was set from a widening conversion.
22356
22357 2013-03-20 Walter Lee <walt@tilera.com>
22358
22359 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22360
22361 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22362
22363 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22364 depending on type attribute.
22365 (*movti_internal): Ditto.
22366 (*movtf_internal): Ditto.
22367 (*movxf_internal): Ditto.
22368 (*movdf_internal): Ditto.
22369 (*movsf_internal): Ditto.
22370
22371 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22372
22373 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22374 maybe_vex for sselog1 and ssemov types.
22375 (*movdi_internal): Reorder operand constraints.
22376 (*movsi_internal): Ditto. Set prefix attribute to
22377 maybe_vex for sselog1 and ssemov types.
22378 (*movtf_internal): Set prefix attribute to maybe_vex
22379 for sselog1 and ssemov types.
22380 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22381 DImode ssemov types. Reorder operand constraints.
22382 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22383 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22384 attribute for SImode ssemov types. Reorder operand constraints.
22385
22386 2013-03-20 Martin Jambor <mjambor@suse.cz>
22387
22388 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22389 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22390
22391 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22392
22393 * config/rs6000/predicates.md (indexed_address, update_address_mem
22394 update_indexed_address_mem): New predicates.
22395 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22396 attribute for load/store instructions.
22397 * config/rs6000/dfp.md (movsd_store): Likewise.
22398 (movsd_load): Likewise.
22399 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22400 (unnamed HI->DI extend define_insn): Likewise.
22401 (unnamed SI->DI extend define_insn): Likewise.
22402 (unnamed QI->SI extend define_insn): Likewise.
22403 (unnamed QI->HI extend define_insn): Likewise.
22404 (unnamed HI->SI extend define_insn): Likewise.
22405 (unnamed HI->SI extend define_insn): Likewise.
22406 (extendsfdf2_fpr): Likewise.
22407 (movsi_internal1): Likewise.
22408 (movsi_internal1_single): Likewise.
22409 (movhi_internal): Likewise.
22410 (movqi_internal): Likewise.
22411 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22412 attribute for load/store instructions.
22413 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22414 instructions.
22415 (mov<mode>_softfloat): Likewise.
22416 (mov<mode>_hardfloat32): Likewise.
22417 (mov<mode>_hardfloat64): Likewise.
22418 (mov<mode>_softfloat64): Likewise.
22419 (movdi_internal32): Likewise.
22420 (movdi_internal64): Likewise.
22421 (probe_stack_<mode>): Likewise.
22422
22423 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22424
22425 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22426 floating point, and decimal floating point to reload iterator.
22427
22428 * config/rs6000/constraints.md (wl constraint): New constraints to
22429 return FLOAT_REGS if certain options are used to reduce the number
22430 of separate patterns that exist in the file.
22431 (wx constraint): Likewise.
22432 (wz constraint): Likewise.
22433
22434 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22435 -mdebug=reg, print wg, wl, wx, and wz constraints.
22436 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22437 Initialize the reload functions for 64-bit binary/decimal floating
22438 point types.
22439 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22440 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22441 create the buffer on the stack to overcome not having a 32-bit
22442 load and store.
22443 (rs6000_emit_move): Likewise.
22444 (rs6000_secondary_memory_needed_rtx): Likewise.
22445 (rs6000_alloc_sdmode_stack_slot): Likewise.
22446 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22447 via xxlxor, just like DFmode 0.0.
22448
22449 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22450 define as 1 if we are running on a power7 or newer.
22451 (enum r6000_reg_class_enum): Add new constraints.
22452
22453 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22454 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22455 with other moves by using conditional constraits (wg). Use LFIWZX
22456 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22457 (movsd splitter): Likewise.
22458 (movsd_hardfloat): Likewise.
22459 (movsd_softfloat): Likewise.
22460
22461 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22462 binary and decimal floating point moves.
22463 (fmove_ok): New attributes to combine binary and decimal floating
22464 point moves, and to combine power6x (mfpgpr) moves along normal
22465 floating moves.
22466 (real_value_to_target): Likewise.
22467 (f32_lr): Likewise.
22468 (f32_lm): Likewise.
22469 (f32_li): Likewise.
22470 (f32_sr): Likewise.
22471 (f32_sm): Likewise.
22472 (f32_si): Likewise.
22473 (movsf): Combine binary and decimal floating point moves. Combine
22474 power6x (mfpgpr) moves with other moves by using conditional
22475 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22476 (mov<mode> for SFmode/SDmode); Likewise.
22477 (SFmode/SDmode splitters): Likewise.
22478 (movsf_hardfloat): Likewise.
22479 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22480 (movsf_softfloat): Likewise.
22481 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22482
22483 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22484 wx and wz constraints.
22485
22486 * config/rs6000/constraints.md (wg constraint): New constraint to
22487 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22488
22489 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22490 constraint.
22491
22492 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22493 -mdebug=reg, print wg, wl, wx, and wz constraints.
22494 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22495 Initialize the reload functions for 64-bit binary/decimal floating
22496 point types.
22497 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22498 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22499 create the buffer on the stack to overcome not having a 32-bit
22500 load and store.
22501 (rs6000_emit_move): Likewise.
22502 (rs6000_secondary_memory_needed_rtx): Likewise.
22503 (rs6000_alloc_sdmode_stack_slot): Likewise.
22504 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22505 via xxlxor, just like DFmode 0.0.
22506
22507 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22508 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22509 with other moves by using conditional constraits (wg). Use LFIWZX
22510 and STFIWX for loading SDmode on power7.
22511 (movdd splitters): Likewise.
22512 (movdd_hardfloat32): Likewise.
22513 (movdd_softfloat32): Likewise.
22514 (movdd_hardfloat64_mfpgpr): Likewise.
22515 (movdd_hardfloat64): Likewise.
22516 (movdd_softfloat64): Likewise.
22517
22518 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22519 64-bit binary and decimal floating point moves.
22520 (FMOVE64X): Likewise.
22521 (movdf): Combine 64-bit binary and decimal floating point moves.
22522 Combine power6x (mfpgpr) moves with other moves by using
22523 conditional constraits (wg).
22524 (mov<mode> for DFmode/DDmode): Likewise.
22525 (DFmode/DDmode splitters): Likewise.
22526 (movdf_hardfloat32): Likewise.
22527 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22528 (movdf_softfloat32): Likewise.
22529 (movdf_hardfloat64_mfpgpr): Likewise.
22530 (movdf_hardfloat64): Likewise.
22531 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22532 (movdf_softfloat64): Likewise.
22533 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22534 (reload_<mode>_load): Move to later in the file so they aren't in
22535 the middle of the floating point move insns.
22536 (reload_<mode>_store): Likewise.
22537
22538 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22539 constraint.
22540
22541 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22542 constraint if -mdebug=reg.
22543 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22544 Enable using dd reload support if needed.
22545
22546 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22547 binary and decimal floating point moves in rs6000.md.
22548 (movtd_internal): Likewise.
22549
22550 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22551 decimal floating point moves.
22552 (movtf): Likewise.
22553 (movtf_internal): Likewise.
22554 (mov<mode>_internal, TDmode/TFmode): Likewise.
22555 (movtf_softfloat): Likewise.
22556 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22557
22558 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22559 movdi_internal64, using wg constraint for move direct operations.
22560 (movdi_internal64): Likewise.
22561
22562 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22563 MODES_TIEABLE_P for selected modes. Print the numerical value of
22564 the various virtual registers. Use GPR/FPR first/last values,
22565 instead of hard coding the register numbers. Print which modes
22566 have reload functions registered.
22567 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22568 settings before/after setting cpu, target and subtarget settings.
22569 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22570 and for secondary reload failures in rs6000_secondary_reload_inner.
22571 (rs6000_secondary_reload_fail): Likewise.
22572 (rs6000_secondary_reload_inner): Likewise.
22573
22574 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22575 macros for first/last GPR and FPR registers.
22576 (LAST_GPR_REGNO): Likewise.
22577 (FIRST_FPR_REGNO): Likewise.
22578 (LAST_FPR_REGNO): Likewise.
22579
22580 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22581 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22582 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22583 (vcond<mode><mode>): Likewise.
22584 (vcondu<mode><mode>): Likewise.
22585 (vector_gtu<mode>): Likewise.
22586 (vector_gte<mode>): Likewise.
22587 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22588 to prevent the compiler from converting DImode operations to TImode.
22589 (ior<mode>3): Likewise.
22590 (and<mode>3): Likewise.
22591 (one_cmpl<mode>2): Likewise.
22592 (nor<mode>3): Likewise.
22593 (andc<mode>3): Likewise.
22594
22595 * config/rs6000/constraints.md (wt constraint): New constraint
22596 that returns VSX_REGS if TImode is allowed in VSX registers.
22597
22598 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22599 constant under VSX.
22600
22601 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22602 similar to TImode, but it is restricted to being in the GPRs.
22603
22604 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22605 TImode to occupy a single VSX register.
22606
22607 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22608 -mvsx-timode for power7/power8.
22609 (power7 cpu): Likewise.
22610 (power8 cpu): Likewise.
22611
22612 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22613 sure that TFmode/TDmode take up two registers if they are ever
22614 allowed in the upper VSX registers.
22615 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22616 registers.
22617 (rs6000_init_hard_regno_mode_ok): Likewise.
22618 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22619 constraint. Print if LRA is turned on.
22620 (rs6000_option_override_internal): Give an error if -mvsx-timode
22621 and VSX is not enabled.
22622 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22623 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22624 to reg+offset addressing. Use PTImode when checking offset
22625 addresses for validity.
22626 (reg_offset_addressing_ok_p): Likewise.
22627 (rs6000_legitimate_offset_address_p): Likewise.
22628 (rs6000_legitimize_address): Likewise.
22629 (rs6000_legitimize_reload_address): Likewise.
22630 (rs6000_legitimate_address_p): Likewise.
22631 (rs6000_eliminate_indexed_memrefs): Likewise.
22632 (rs6000_emit_move): Likewise.
22633 (rs6000_secondary_reload): Likewise.
22634 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22635 reloads to fpr registers to continue to use reg+offset addressing,
22636 but 64-bit reloads to altivec registers need reg+reg addressing.
22637 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22638 it. Treat LO_SUM like a PLUS operation.
22639 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22640 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22641 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22642 registers to share a register with a smaller sized type, since VSX
22643 puts scalars in the upper 64-bits.
22644 (print_operand): Add support for PTImode.
22645 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22646 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22647 registers, but don't have arithmetic support.
22648 (rs6000_memory_move_cost): Add test for VSX.
22649 (rs6000_opt_masks): Add -mvsx-timode.
22650
22651 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22652 for TImode.
22653 (VSs): Likewise.
22654 (VSr): Use wt constraint for TImode.
22655 (VSv): Drop TImode support.
22656 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22657 (vsx_movti_64bit): Likewise.
22658 (vsx_movti_32bit): Likewise.
22659 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22660 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22661 one '?' on the appropriate output constraint. Do not allow TImode
22662 logical operations on 32-bit systems.
22663 (vsx_ior<mode>3): Likewise.
22664 (vsx_xor<mode>3): Likewise.
22665 (vsx_one_cmpl<mode>2): Likewise.
22666 (vsx_nor<mode>3): Likewise.
22667 (vsx_andc<mode>3): Likewise.
22668 (vsx_concat_<mode>): Likewise.
22669 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22670
22671 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22672 OPTION_MASK_VSX_TIMODE.
22673 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22674 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22675
22676 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22677 (TI2 iterator): New iterator for TImode, PTImode.
22678 (wd mode attribute): Add values for vector types.
22679 (movti_string): Replace TI move operations with operations for TImode
22680 and PTImode. Add support for TImode being allowed in VSX registers.
22681 (mov<mode>_string, TImode/PTImode): Likewise.
22682 (movti_ppc64): Likewise.
22683 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22684 (TI mode splitters): Likewise.
22685
22686 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22687 constraint.
22688
22689 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22690
22691 PR tree-optimization/56355
22692 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22693 Also handle integers with undefined overflow.
22694
22695 2013-03-20 Catherine Moore <clm@codesourcery.com>
22696 Maciej W. Rozycki <macro@codesourcery.com>
22697 Tom de Vries <tom@codesourcery.com>
22698 Nathan Sidwell <nathan@codesourcery.com>
22699 Iain Sandoe <iain@codesourcery.com>
22700 Nathan Froyd <froydnj@codesourcery.com>
22701 Chao-ying Fu <fu@mips.com>
22702
22703 * doc/extend.texi: (micromips, nomicromips, nocompression):
22704 Document new function attributes.
22705 * doc/invoke.texi (minterlink-compressed, mmicromips,
22706 m14k, m14ke, m14kec): Document new options.
22707 (minterlink-mips16): Update documentation.
22708 * doc/md.texi (ZC, ZD): Document new constraints.
22709 * configure.ac (gcc_cv_as_micromips): Check if linker
22710 supports the .set micromips directive.
22711 * configure: Regenerate.
22712 * config.in: Regenerate.
22713 * config/mips/mips-tables.opt: Regenerate.
22714 * config/mips/micromips.md: New file.
22715 * constraints.md (ZC, ZD): New constraints.
22716 * config/mips/predicates.md (movep_src_register): New predicate.
22717 (movep_src_operand): New predicate.
22718 (non_volatile_mem_operand): New predicate.
22719 * config/mips/mips.md (multimem): New type.
22720 (length): Differentiate between 17-bit and 18-bit branch offsets.
22721 (MOVEP1, MOVEP2): New mode iterator.
22722 (mov_<load>l): Use ZC constraint.
22723 (mov_<load>r): Likewise.
22724 (mov_<store>l): Likewise.
22725 (mov_<store>r): Likewise.
22726 (*branch_equality<mode>_inverted): Add microMIPS support.
22727 (*branch_equality<mode>): Likewise.
22728 (*jump_absolute): Likewise.
22729 (indirect_jump_<mode>): Likewise.
22730 (tablejump_<mode>): Likewise.
22731 (<optab>_internal): Likewise.
22732 (sibcall_internal): Likewise.
22733 (sibcall_value_internal): Likewise.
22734 (prefetch): Use constraint ZD.
22735 * config/mips/mips.opt (minterlink-compressed): New option.
22736 (minterlink-mips16): Now an alias for minterlink-compressed.
22737 (mmicromips): New option.
22738 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22739 (compare_and_swap_12): Likewise.
22740 (sync_add<mode>): Likewise.
22741 (sync_<optab>_12): Likewise.
22742 (sync_old_<optab>_12): Likewise.
22743 (sync_new_<optab>_12): Likewise.
22744 (sync_nand_12): Likewise.
22745 (sync_old_nand_12): Likewise.
22746 (sync_new_nand_12): Likewise.
22747 (sync_sub<mode>): Likewise.
22748 (sync_old_add<mode>): Likewise.
22749 (sync_old_sub<mode>): Likewise.
22750 (sync_new_add<mode>): Likewise.
22751 (sync_new_sub<mode>): Likewise.
22752 (sync_<optab><mode>): Likewise.
22753 (sync_old_<optab><mode>): Likewise.
22754 (sync_new_<optab><mode>): Likewise.
22755 (sync_nand<mode>): Likewise.
22756 (sync_old_nand<mode>): Likewise.
22757 (sync_new_nand<mode>): Likewise.
22758 (sync_lock_test_and_set<mode>): Likewise.
22759 (test_and_set_12): Likewise.
22760 (atomic_compare_and_swap<mode>): Likewise.
22761 (atomic_exchange<mode>_llsc): Likewise.
22762 (atomic_fetch_add<mode>_llsc): Likewise.
22763 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22764 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22765 (umips_save_restore_pattern_p): Likewise.
22766 (umips_load_store_pair_p): Likewise.
22767 (umips_output_load_store_pair): Likewise.
22768 (umips_movep_target_p): Likewise.
22769 (umips_12bit_offset_address_p): Likewise.
22770 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22771 (mips_base_mips16): Rename this...
22772 (mips_base_compression_flags): ...to this. Update all uses.
22773 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22774 (mips_mips16_decl_p): Delete.
22775 (mips_nomips16_decl_p): Delete.
22776 (mips_get_compress_on_flags): New function.
22777 (mips_get_compress_off_flags): New function.
22778 (mips_get_compress_mode): New function.
22779 (mips_get_compress_on_name): New function.
22780 (mips_get_compress_off_name): New function.
22781 (mips_insert_attributes): Support multiple compression types.
22782 (mips_merge_decl_attributes): Likewise.
22783 (umips_12bit_offset_address_p): New function.
22784 (mips_start_function_definition): Emit .set micromips directive.
22785 (mips_call_may_need_jalx_p): New function.
22786 (mips_function_ok_for_sibcall): Add microMIPS support.
22787 (mips_print_operand_punctuation): Support short delay slots and
22788 compact jumps.
22789 (umips_swm_mask, umips_swm_encoding): New.
22790 (umips_build_save_restore): New function.
22791 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22792 (was_mips16_p): Remove.
22793 (old_compression_mode): New.
22794 (mips_set_compression_mode): New function.
22795 (mips_set_current_function): Add microMIPS support.
22796 (mips_option_override): Likewise.
22797 (umips_save_restore_pattern_p): New function.
22798 (umips_output_save_restore): New function.
22799 (umips_load_store_pair_p_1): New function.
22800 (umips_load_store_pair_p): New function.
22801 (umips_output_load_store_pair_1): New function.
22802 (umips_output_load_store_pair): New function.
22803 (umips_movep_target_p) New function.
22804 (mips_prepare_pch_save): Add microMIPS support.
22805 * config/mips/mips.h (TARGET_COMPRESSION): New.
22806 (TARGET_CPU_CPP_BUILTINS): Update macro
22807 to use new compression flags and to support microMIPS.
22808 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22809 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22810 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22811 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22812 (ASM_SPEC): Support mmicromips and mno-micromips.
22813 (M16STORE_REG_P): New macro.
22814 (MIPS_CALL): Support TARGET_MICROMIPS.
22815 (MICROMIPS_J): New macro.
22816 (mips_base_mips16): Rename this...
22817 (mips_base_compression_flags): ...to this.
22818 (UMIPS_12BIT_OFFSET_P): New macro.
22819 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22820 (MULTILIB_DIRNAMES): Likewise.
22821 2013-03-20 Richard Biener <rguenther@suse.de>
22822
22823 PR tree-optimization/56661
22824 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22825 the result does not have to be distinct.
22826
22827 2013-03-20 Richard Biener <rguenther@suse.de>
22828
22829 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22830 remap_gimple_op_r.
22831
22832 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22833 Steven Bosscher <steven@gcc.gnu.org>
22834
22835 PR rtl-optimization/56605
22836 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22837
22838 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22839
22840 PR bootstrap/56656
22841 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22842 that require movd instead of movq.
22843
22844 2013-03-20 Richard Biener <rguenther@suse.de>
22845
22846 * tree-ssa-structalias.c (struct variable_info): Add pointer
22847 to the first field of an aggregate with sub-vars. Make
22848 this and the pointer to the next subfield its ID.
22849 (vi_next): New function.
22850 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22851 storedanything_id, integer_id): Increment by one.
22852 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22853 get_call_clobber_vi): Adjust.
22854 (solution_set_expand): Simplify and speedup.
22855 (solution_set_add): Inline into ...
22856 (set_union_with_increment): ... this. Adjust accordingly.
22857 (do_sd_constraint): Likewise.
22858 (do_ds_constraint): Likewise.
22859 (do_complex_constraint): Simplify.
22860 (build_pred_graph): Adjust.
22861 (solve_graph): Likewise. Simplify and speedup.
22862 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22863 get_constraint_for_component_ref, get_constraint_for_1,
22864 first_vi_for_offset, first_or_preceding_vi_for_offset,
22865 create_function_info_for, create_variable_info_for_1,
22866 create_variable_info_for, intra_create_variable_infos): Adjust.
22867 (init_base_vars): Push NULL for ID zero.
22868 (compute_points_to_sets): Adjust.
22869
22870 2013-03-20 Richard Biener <rguenther@suse.de>
22871
22872 * cfgloop.c (verify_loop_structure): Streamline and avoid
22873 ICEing on corrupt loop tree.
22874 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22875 loop tree.
22876
22877 2013-03-20 Richard Biener <rguenther@suse.de>
22878
22879 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22880 check whether an SSA update is needed.
22881
22882 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22883
22884 * config/mips/constraints.md (T): Rename to...
22885 (Yf): ...this.
22886 (U): Rename to...
22887 (Yd): ...this.
22888 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22889 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22890
22891 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22892
22893 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22894 (*subsi3_carryin_uxtw): Likewise.
22895
22896 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22897
22898 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22899 (*rorsi3_insn_uxtw): Likewise.
22900
22901 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22902
22903 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22904 (*extrsi5_insn_uxtw): Likewise.
22905
22906 2013-03-19 Richard Biener <rguenther@suse.de>
22907
22908 PR tree-optimization/56273
22909 * passes.c (init_optimization_passes): Move second VRP after DOM.
22910
22911 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22912
22913 * config/i386/i386.md (*movti_internal): Merge from
22914 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22915 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22916 nox64 isa attributes.
22917
22918 2013-03-18 Richard Biener <rguenther@suse.de>
22919
22920 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22921 (unite): Likewise.
22922 (merge_node_constraints): Likewise.
22923 (build_succ_graph): Likewise.
22924 (valid_graph_edge): Inline into single caller.
22925 (unify_nodes): Likewise. Use bitmap_set_bit return value
22926 and cache varinfo.
22927 (scc_visit): Fix formatting and variable use.
22928 (do_sd_constraint): Use gcc_checking_assert.
22929 (do_ds_constraint): Likewise.
22930 (do_complex_constraint): Likewise.
22931 (condense_visit): Likewise. Cleanup.
22932 (dump_pred_graph): New function.
22933 (perform_var_substitution): Dump the pred-graph before
22934 variable substitution.
22935 (find_equivalent_node): Use gcc_checking_assert.
22936 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22937
22938 2013-03-18 Richard Biener <rguenther@suse.de>
22939
22940 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22941 Remove cond_expr_stmt_list argument and do not gimplify the
22942 built expression.
22943 (vect_loop_versioning): Adjust.
22944 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22945 Cleanup to use less temporaries.
22946 (vect_create_data_ref_ptr): Cleanup.
22947
22948 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22949
22950 PR tree-optimization/56635
22951 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22952 require types_compatible_p types.
22953
22954 2013-03-18 Nick Clifton <nickc@redhat.com>
22955
22956 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22957 spurious backslash.
22958
22959 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22960 Add missing line to comment describing function.
22961
22962 2013-03-18 Richard Biener <rguenther@suse.de>
22963
22964 PR tree-optimization/56210
22965 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22966 Handle string / character search functions.
22967 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22968
22969 2013-03-18 Richard Biener <rguenther@suse.de>
22970
22971 PR middle-end/56483
22972 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22973 and implement properly.
22974 * gimple.h (gimple_cond_single_var_p): Remove.
22975
22976 2013-03-18 Richard Biener <rguenther@suse.de>
22977
22978 * tree-data-ref.h (find_data_references_in_loop): Declare.
22979 * tree-data-ref.c (get_references_in_stmt): Use a stack
22980 vector pre-allocated in the callers.
22981 (find_data_references_in_stmt): Adjust.
22982 (graphite_find_data_references_in_stmt): Likewise.
22983 (create_rdg_vertices): Likewise.
22984 (find_data_references_in_loop): Export.
22985 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22986 Compute dependences here...
22987 (vect_analyze_data_refs): ...not here. When we encounter
22988 a non-vectorizable data reference in basic-block vectorization
22989 truncate the data reference vector. Do not bother to
22990 fixup data-dependence information for gather loads.
22991 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22992 of data references, as reported.
22993
22994 2013-03-18 Richard Biener <rguenther@suse.de>
22995
22996 PR tree-optimization/3713
22997 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22998 has_constants and expr.
22999 (stmt_has_constants): Properly valueize SSA names when deciding
23000 whether the stmt has constants.
23001
23002 2013-03-18 Richard Biener <rguenther@suse.de>
23003
23004 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23005 whole function when there is nothing to do.
23006 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23007 * tree-vectorizer.c (vectorize_loops): Update virtual and
23008 loop-closed SSA once.
23009 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23010
23011 2013-03-18 Richard Biener <rguenther@suse.de>
23012
23013 PR middle-end/56113
23014 * domwalk.c (bb_postorder): New global static.
23015 (cmp_bb_postorder): New function.
23016 (walk_dominator_tree): Replace scheme imposing an order for
23017 visiting dominator sons by one sorting them at the time they
23018 are pushed on the stack.
23019
23020 2013-03-18 Richard Biener <rguenther@suse.de>
23021
23022 PR tree-optimization/39326
23023 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23024 (struct mem_ref): Replace mem member with ao_ref typed member.
23025 (MEM_ANALYZABLE): Adjust.
23026 (memref_eq): Likewise.
23027 (mem_ref_alloc): Likewise.
23028 (gather_mem_refs_stmt): Likewise.
23029 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23030 (execute_sm_if_changed_flag_set): Adjust.
23031 (execute_sm): Likewise.
23032 (ref_always_accessed_p): Likewise.
23033 (refs_independent_p): Likewise.
23034 (can_sm_ref_p): Likewise.
23035
23036 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23037
23038 PR c/56566
23039 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23040 return 1 even for !unsignedp.
23041
23042 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23043
23044 * config/i386/i386.md (isa): Add x64 and nox64.
23045 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23046 (*pushtf): Enable *roF alternative for x64 isa only.
23047 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23048 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23049 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23050 (*movtf_internal): Merge from *movtf_internal_rex64 and
23051 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23052 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23053 nox64 isa attributes.
23054 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23055 nox64 isa attributes.
23056 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23057
23058 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23059
23060 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23061
23062 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23063
23064 PR target/56640
23065 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23066 class names. Remove trailing comma after "ALL_REGS".
23067
23068 2013-03-16 Jan Hubicka <jh@suse.cz>
23069
23070 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23071 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23072 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23073 of cgraph_get_create_node.
23074 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23075
23076 2013-03-16 Jason Merrill <jason@redhat.com>
23077
23078 PR debug/49090
23079 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23080 with DW_AT_default_value.
23081
23082 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23083
23084 * BASE-VER: Set to 4.9.0.
23085
23086 2013-03-14 Andi Kleen <ak@linux.intel.com>
23087
23088 PR target/56619
23089 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23090 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23091 Document _x* TSX intrinsics.
23092
23093 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23094 David Holsgrove <david.holsgrove@xilinx.com>
23095
23096 * configure.ac: Add MicroBlaze TLS support detection.
23097 * configure: Regenerate.
23098 * config/microblaze/microblaze-protos.h
23099 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23100 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23101 * config/microblaze/microblaze.c (microblaze_address_type): Add
23102 ADDRESS_TLS and tls_reloc address types.
23103 (microblaze_address_info): Add tls_reloc.
23104 (TARGET_HAVE_TLS): Define.
23105 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23106 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23107 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23108 load_tls_operand, microblaze_call_tls_get_addr,
23109 microblaze_legitimize_tls_address): New functions.
23110 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23111 (get_base_reg): Use microblaze_tls_symbol_p.
23112 (microblaze_classify_address): Handle TLS.
23113 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23114 label_mentioned_p and microblaze_tls_referenced_p.
23115 (microblaze_legitimize_address): Handle TLS.
23116 (microblaze_address_insns): Handle ADDRESS_TLS.
23117 (pic_address_needs_scratch): Handle TLS.
23118 (print_operand_address): Handle TLS.
23119 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23120 (microblaze_expand_move): Handle TLS.
23121 (microblaze_legitimate_constant_p): Check
23122 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23123 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23124 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23125 (PIC_OFFSET_TABLE_REGNUM): Set.
23126 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23127 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23128 (addsi3, movsi_internal2, movdf_internal): Update constraints
23129 * config/microblaze/predicates.md (arith_plus_operand): Define
23130 (move_operand): Redefine as move_src_operand,
23131 check microblaze_tls_referenced_p.
23132
23133 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23134
23135 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23136 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23137
23138 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23139
23140 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23141 CC mode for AND.
23142
23143 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23144
23145 PR tree-optimization/53265
23146 * common.opt (Waggressive-loop-optimizations): New option.
23147 * tree-ssa-loop-niter.c: Include tree-pass.h.
23148 (do_warn_aggressive_loop_optimizations): New function.
23149 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23150 if number_of_latch_executions returned constant.
23151 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23152 early. If number_of_latch_executions returned constant, set
23153 nb_iterations_upper_bound back to it.
23154 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23155 field.
23156 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23157 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23158
23159 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23160 (MULTILIB_OSDIRNAMES): Set.
23161 * genmultilib: If defaultosdirname doesn't start with :: , set
23162 defaultosdirname2 instead, clear it and emit two . multilib_raw
23163 entries instead of just one.
23164
23165 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23166
23167 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23168 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23169 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23170 (SUBTARGET_OVERRIDE_OPTIONS): New.
23171
23172 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23173
23174 PR target/49880
23175 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23176 (musermode): Convert to Var(TARGET_USERMODE).
23177 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23178 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23179 * config/sh/sh.c (sh_option_override): Use
23180 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23181 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23182 condition.
23183 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23184 TARGET_SH4.
23185 (udivsi3_i4_single, divsi3_i4_single): Use
23186 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23187
23188 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23189
23190 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23191 default setting.
23192
23193 2013-03-13 Richard Biener <rguenther@suse.de>
23194
23195 PR tree-optimization/56608
23196 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23197 calls when vectorizing basic-blocks.
23198
23199 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23200
23201 PR plugins/45078
23202 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23203 tm_file.
23204
23205 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23206
23207 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23208
23209 2013-03-11 Jan Hubicka <jh@suse.cz>
23210
23211 PR lto/56557
23212 * lto-streamer-out.c (output_symbol_p): Skip references from
23213 constructors of external variables.
23214
23215 2013-03-11 Jan Hubicka <jh@suse.cz>
23216
23217 PR middle-end/56571
23218 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23219 from pseudos.
23220 * emit-rtl.c (verify_rtx_sharing): Likewise.
23221 (copy_insn_1): Likewise.
23222 * rtl.c (copy_rtx): Likewise.
23223
23224 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23225
23226 PR target/56591
23227 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23228 output_operand_lossage message.
23229
23230 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23231
23232 PR target/56470
23233 * arm.c (shift_op): Validate RTL pattern on the fly.
23234 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23235 the RTL.
23236
23237 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23238
23239 PR target/56347
23240 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23241 new call patterns that clobber %fr12.
23242 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23243 split and postreload patterns.
23244 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23245 registers %fr12 and %fr12R as call used.
23246
23247 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23248
23249 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23250 (canon_address, record_store, replace_read, check_mem_read_rtx,
23251 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23252 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23253 rest_of_handle_dse): Likewise.
23254
23255 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23256
23257 PR middle-end/56524
23258 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23259 Add base_optabs.
23260 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23261 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23262 (save_optabs_if_changed): Replace with...
23263 (init_tree_optimization_optabs): ...this.
23264 * optabs.c (save_optabs_if_changed): Rename to...
23265 (init_tree_optimization_optabs): ...this. Take the optimization node
23266 as argument. Do nothing if the base optabs are already correct.
23267 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23268 to recompute optabs.
23269 * function.h (function): Remove optabs field.
23270 * function.c (invoke_set_current_function_hook): Call
23271 init_tree_optimization_optabs. Use the result to initialize
23272 this_fn_optabs.
23273
23274 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23275
23276 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23277 if GTMA_HAS_NO_INSTRUMENTATION.
23278 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23279 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23280 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23281 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23282 GTMA_HAS_NO_INSTRUMENTATION.
23283
23284 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23285
23286 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23287 libasan_preinit.o.
23288
23289 2013-03-08 Marek Polacek <polacek@redhat.com>
23290 Jakub Jelinek <jakub@redhat.com>
23291
23292 PR tree-optimization/56478
23293 * predict.c (is_comparison_with_loop_invariant_p): Change the
23294 type of loop_step to tree.
23295 (predict_loops): Adjust.
23296 (predict_iv_comparison): Perform the computations on double_ints.
23297
23298 2013-03-08 Richard Biener <rguenther@suse.de>
23299
23300 PR tree-optimization/56570
23301 * tree-cfg.c (verify_expr_location_1): Verify locations for
23302 DECL_DEBUG_EXPR.
23303 * tree-sra.c (create_access_replacement): Strip locations
23304 from DECL_DEBUG_EXPRs.
23305
23306 2013-03-08 Richard Biener <rguenther@suse.de>
23307
23308 * tree-inline.c (expand_call_inline): Do not associate
23309 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23310 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23311
23312 2013-03-08 Richard Biener <rguenther@suse.de>
23313
23314 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23315 or block changes with -Og. Fix for location / block encoding
23316 changes and PHI arguments with locations.
23317
23318 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23319
23320 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23321 for all counters.
23322 (struct output_info): Likewise.
23323 (register_overhead): Remove bad gcc_assert.
23324 (bitmap_find_bit): If there is only a single bitmap element, do not
23325 count a miss as a search.
23326 (print_statistics): Update for counter type changes.
23327 (dump_bitmap_statistics): Likewise. Print headers such that they
23328 are properly lined up with the printed counters.
23329
23330 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23331
23332 PR tree-optimization/56559
23333 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23334 check that it has only a single use.
23335
23336 2013-03-07 Richard Biener <rguenther@suse.de>
23337
23338 * doc/invoke.texi (fwhole-program): Discourage use in combination
23339 with -flto.
23340
23341 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23342
23343 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23344
23345 PR tree-optimization/56539
23346 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23347 instead of GSI_CONTINUE_LINKING as last argument to
23348 force_gimple_operand_gsi. Adjust function comment.
23349
23350 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23351 aarch64-cores.def.
23352
23353 PR middle-end/56548
23354 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23355 promoted mode, convert the result back to the original mode.
23356
23357 2013-03-06 Richard Biener <rguenther@suse.de>
23358
23359 PR middle-end/56294
23360 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23361 (insert_updated_phi_nodes_compare_uids): New function.
23362 (update_ssa): Sort symbols_to_rename after UID before
23363 traversing it to insert PHI nodes.
23364
23365 2013-03-06 Richard Biener <rguenther@suse.de>
23366
23367 PR middle-end/50494
23368 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23369 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23370
23371 Revert
23372 2013-02-13 Richard Biener <rguenther@suse.de>
23373
23374 PR lto/50494
23375 * varasm.c (output_constant_def_1): Get the decl representing
23376 the constant as argument.
23377 (output_constant_def): Wrap output_constant_def_1.
23378 (make_decl_rtl): Use output_constant_def_1 with the decl
23379 representing the constant.
23380 (build_constant_desc): Optionally re-use a decl already
23381 representing the constant.
23382 (tree_output_constant_def): Adjust.
23383
23384 2013-03-06 Joey Ye <joey.ye@arm.com>
23385
23386 PR lto/50293
23387 * gcc.c (convert_white_space): New function.
23388 (main): Handles white space in function name.
23389
23390 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23391
23392 PR target/56529
23393 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23394 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23395 to SH_DIV_CALL_TABLE for TARGET_SH2.
23396 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23397 list.
23398 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23399 call-table options.
23400
23401 2013-03-05 Sterling Augustine <saugustine@google.com>
23402 Cary Coutant <ccoutant@google.com>
23403
23404 PR debug/55364
23405 * dwarf2out.c (resolve_addr): Don't call
23406 remove_loc_list_addr_table_entries a second time for the same
23407 expression.
23408
23409 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23410
23411 PR debug/56510
23412 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23413 (avoid_complex_debug_insns): New function.
23414 (expand_debug_locations): Call it.
23415
23416 PR rtl-optimization/56484
23417 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23418 lifetimes of hard registers on small register class machines.
23419
23420 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23421
23422 * config/microblaze/microblaze-protos.h: Rename
23423 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23424 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23425 fast_interrupt.
23426 (microblaze_fast_interrupt_function_p): New function.
23427 (microblaze_is_interrupt_handler): Rename to
23428 microblaze_is_interrupt_variant and add fast_interrupt check.
23429 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23430 (save_restore_insns): Likewise.
23431 (compute_frame_size): Likewise.
23432 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23433 (microblaze_globalize_label): Likewise.
23434 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23435 * config/microblaze/microblaze.md: Use wrapper
23436 microblaze_is_interrupt_variant.
23437
23438 2013-03-05 Kai Tietz <ktietz@redhat.com>
23439
23440 * sdbout.c (sdbout_one_type): Switch to current function's section
23441 supporting cold/hot.
23442
23443 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23444
23445 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23446 -mxl-reorder.
23447
23448 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23449
23450 PR middle-end/56461
23451 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23452 if VALGRIND_GET_VBITS is defined, temporarily make object
23453 memory all defined, and restore previous valgrind addressability
23454 and definability afterwards. Free this_object at the end.
23455
23456 PR middle-end/56461
23457 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23458 right before calling lra_create_live_ranges, also call it
23459 when clearing live_p. Only call lra_clear_live_ranges
23460 at the end if live_p.
23461
23462 PR middle-end/56461
23463 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23464
23465 2013-03-05 Richard Biener <rguenther@suse.de>
23466
23467 PR tree-optimization/56521
23468 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23469 value-id.
23470
23471 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23472
23473 PR c++/55135
23474 * except.h (remove_unreachable_eh_regions): New prototype.
23475 * except.c (remove_eh_handler_splicer): New function, split out
23476 of remove_eh_handler.
23477 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23478 warning about running it on many EH regions one at a time.
23479 (remove_unreachable_eh_regions_worker): New function, walk the
23480 EH tree in depth-first order and remove non-marked regions.
23481 (remove_unreachable_eh_regions): New function.
23482 * tree-eh.c (mark_reachable_handlers): New function, split out
23483 from remove_unreachable_handlers.
23484 (remove_unreachable_handlers): Use mark_reachable_handlers and
23485 remove_unreachable_eh_regions.
23486 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23487 and remove_unreachable_eh_regions.
23488
23489 2013-03-05 Richard Biener <rguenther@suse.de>
23490
23491 PR middle-end/56525
23492 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23493 not freeing them until the end.
23494
23495 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23496
23497 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23498
23499 2013-03-05 Richard Biener <rguenther@suse.de>
23500
23501 PR tree-optimization/56270
23502 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23503 of loads after scheduling an SLP instance.
23504
23505 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23506
23507 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23508 tic6x.exp.
23509 (check_gcc_parallelize): Run guality.exp as a separate job from
23510 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23511 struct-layout-1.exp with stackalign.exp.
23512
23513 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23514
23515 PR middle-end/56461
23516 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23517 load_index sbitmap even if some bit in it isn't set.
23518
23519 PR middle-end/56461
23520 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23521 (discover_iteration_bound_by_body_walk): Change queues to
23522 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23523 spelling in comment. Call safe_push on queues[bound_index] directly.
23524 Release queues[queue_index] in every iteration unconditionally.
23525 Release bounds vector.
23526
23527 PR middle-end/56461
23528 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23529 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23530 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23531 inner_phis vector.
23532
23533 2013-03-05 Richard Biener <rguenther@suse.de>
23534
23535 PR lto/56515
23536 * tree-inline.c (remap_blocks_to_null): New function.
23537 (expand_call_inline): When expanding a call stmt without
23538 an associated block inline remap all callee blocks to NULL.
23539
23540 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23541
23542 PR rtl-optimization/56494
23543 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23544 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23545 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23546
23547 PR middle-end/56461
23548 * sel-sched-ir.c (free_sched_pools): Release
23549 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23550 if succs_info_pool.max_top isn't -1.
23551
23552 PR bootstrap/56509
23553 * opts.c (opts_obstack, opts_concat): Moved to...
23554 * opts-common.c (opts_obstack, opts_concat): ... here.
23555
23556 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23557
23558 PR middle-end/56461
23559 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23560
23561 2013-03-04 Martin Jambor <mjambor@suse.cz>
23562
23563 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23564 all appropriate places.
23565
23566 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23567
23568 PR tree-optimization/56424
23569 * ipa-split.c (split_function): Do not set the RSO flag if result is
23570 not by reference and its type is a register type.
23571
23572 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23573
23574 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23575 (microblaze_legitimate_pic_operand): Likewise
23576 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23577 new function microblaze_legitimate_pic_operand
23578 * config/microblaze/microblaze-protos.h
23579 (microblaze_legitimate_pic_operand): Declare.
23580
23581 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23582
23583 * config/microblaze/predicates.md (call_insn_simple_operand):
23584 New predicate for supported rtx code types.
23585 * config/microblaze/microblaze.md (call_internal1): Use
23586 call_insn_simple_operand predicate.
23587
23588 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23589
23590 PR middle-end/56461
23591 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23592 partitions.ordered_remove.
23593
23594 PR middle-end/56461
23595 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23596 vec_oprnds0.create (1) for modifier == NONE.
23597
23598 PR middle-end/56461
23599 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23600 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23601 vec_oprnds1 right before pushing anything to it for
23602 scalar_shift_arg.
23603
23604 PR middle-end/56461
23605 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23606 set nbbs to 0 instead of having separate code path.
23607 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23608 instead of false as last argument if returning NULL.
23609
23610 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23611
23612 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23613 the attribute is now called "target" instead of "option".
23614 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23615 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23616 attribute/pragma name for TARGET_OPTION_VALID_P and
23617 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23618 * doc/tm.texi: Regenerated.
23619
23620 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23621
23622 * config/microblaze/microblaze.c:
23623 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23624 * config/microblaze/microblaze.h: Add -mxl-reorder to
23625 DRIVER_SELF_SPECS.
23626 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23627 instructions emitted if TARGET_REORDER.
23628 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23629 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23630 separately.
23631
23632 2013-03-01 Xinliang David Li <davidxl@google.com>
23633
23634 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23635 walk length.
23636
23637 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23638
23639 PR middle-end/56461
23640 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23641 vector even when returning true. Fix up function comment formatting.
23642
23643 PR middle-end/56461
23644 * ira-build.c (ira_loop_nodes_count): New variable.
23645 (create_loop_tree_nodes): Initialize it.
23646 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23647
23648 PR middle-end/56461
23649 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23650 method on dr_chain and result_chain.
23651 * tree-vect-stmts.c (vectorizable_store): Only call
23652 result_chain.create if j == 0.
23653
23654 PR middle-end/56461
23655 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23656 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23657 before overwriting it.
23658
23659 2013-03-01 Tobias Burnus <burnus@net-b.de>
23660
23661 * doc/extended.texi (C Extensions): Change order in @menu
23662 to match @node.
23663 (Other MIPS Built-in Functions): Move last MIPS entry before
23664 "picoChip Built-in Functions".
23665 (SH Built-in Functions): Move after RX Built-in Functions.
23666 * doc/gcc.texi (Introduction): Change order in @menu
23667 to match @node.
23668 * doc/md.texi (Constraints): Ditto.
23669 * gty.texi (Type Information): Ditto.
23670 (User-provided marking routines for template types): Make
23671 subsection.
23672 * doc/invoke.texi (AArch64 Options): Move before
23673 "Adapteva Epiphany Options".
23674
23675 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23676 Jakub Jelinek <jakub@redhat.com>
23677
23678 PR sanitizer/56454
23679 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23680 no_address_safety_analysis attribute.
23681 * doc/extend.texi (no_address_safety_attribute): Rename to
23682 no_sanitize_address attribute, mention no_address_safety_analysis
23683 attribute as deprecated alias.
23684
23685 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23686
23687 PR middle-end/56461
23688 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23689 type to vec<vec<tree> > *.
23690 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23691 to be vec<tree> instead of vec<tree> *, set vec_defs
23692 to vNULL and call vec_defs.create (number_of_vects), adjust other
23693 uses of vec_defs.
23694 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23695 vectorizable_condition): Adjust vect_get_slp_defs callers.
23696
23697 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23698
23699 * config/aarch64/aarch64.c
23700 (aarch64_float_const_representable): Remove unused variable.
23701
23702 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23703
23704 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23705
23706 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23707
23708 * config/aarch64/aarch64-builtins.c
23709 (aarch64_init_simd_builtins): Make static.
23710
23711 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23712
23713 * config/aarch64/aarch64.c
23714 (aarch64_simd_make_constant): Make static.
23715
23716 2013-02-28 Martin Jambor <mjambor@suse.cz>
23717
23718 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23719 with no initialization to the RHS of debug statements.
23720
23721 2013-02-28 Martin Jambor <mjambor@suse.cz>
23722
23723 PR tree-optimization/56294
23724 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23725 Adjust dumping.
23726 (get_access_replacement): Do not call create_access_replacement.
23727 Assert a replacement exists.
23728 (get_repl_default_def_ssa_name): Create the replacement declaration
23729 itself.
23730
23731 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23732
23733 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23734 final_end_function.
23735
23736 2013-02-28 Marek Polacek <polacek@redhat.com>
23737
23738 PR rtl-optimization/56466
23739 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23740 if we're changing a loop.
23741 (peel_loops_completely): Likewise.
23742
23743 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23744
23745 PR c++/55813
23746 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23747
23748 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23749
23750 PR target/56445
23751 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23752 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23753 INTX_FTYPE_FX, FX_FTYPE_INTX.
23754 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23755
23756 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23757
23758 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23759 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23760 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23761 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23762 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23763 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23764 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23765 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23766 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23767 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23768 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23769 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23770 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23771 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23772 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23773 (avrxmega6): Increase max flash segments from 5 to 6.
23774 * config/avr/t-multilib: Regenerate.
23775 * config/avr/avr-tables.opt: Regenerate.
23776 * doc/avr-mmcu.texi: Regenerate.
23777
23778 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23779
23780 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23781 (avr_device_to_arch): Rename to avr_device_to_ld.
23782 (avr_device_to_as): New prototype.
23783 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23784 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23785 * config/avr/driver-avr.c (avr_device_to_as): New.
23786 (avr_device_to_arch): Rename to avr_device_to_ld.
23787
23788 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23789
23790 PR middle-end/56461
23791 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23792 method on dr_chain and result_chain.
23793
23794 PR middle-end/56461
23795 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23796 pointer_set_destroy on not_executed_last_iteration.
23797
23798 PR middle-end/56461
23799 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23800
23801 PR middle-end/56461
23802 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23803 FOR_EACH_DEFINED_FUNCTION when freeing state.
23804
23805 PR middle-end/56461
23806 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23807 pool_free.
23808 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23809 overwriting it.
23810
23811 PR middle-end/56461
23812 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23813 known_aggs[i].items and release known_aggs vector.
23814
23815 PR middle-end/56461
23816 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23817
23818 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23819
23820 * config/microblaze/microblaze.c (microblaze_emit_compare):
23821 Use xor for EQ/NE comparisions.
23822 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23823 (cbranchsf4): Adjust operator to comparison_operator.
23824
23825 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23826
23827 PR middle-end/56461
23828 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23829 vector.
23830 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23831 vec_safe_push, always update *slot.
23832 (redirect_edge_var_map_clear): Use vec_free.
23833 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23834 (free_var_map_entry): Use vec_free.
23835 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23836 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23837
23838 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23839
23840 PR middle-end/45472
23841 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23842 when the may_trap_p bit of the exprs being merged differs.
23843 Reorder tests for speculativeness in the logical and operator.
23844
23845 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23846
23847 * incpath.c (add_standard_paths): Use reconcat instead of concat
23848 where appropriate and avoid leaking memory.
23849
23850 * opts.h: Include obstack.h.
23851 (opts_concat): New prototype.
23852 (opts_obstack): New declaration.
23853 * opts.c (opts_concat): New function.
23854 (opts_obstack): New variable.
23855 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23856 (finish_options): Use opts_concat instead of concat
23857 and XOBNEWVEC instead of XNEWVEC.
23858 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23859 generate_option): Likewise.
23860 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23861 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23862
23863 PR target/56455
23864 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23865 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23866
23867 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23868
23869 PR middle-end/56461
23870 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23871
23872 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23873
23874 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23875 (arm_block_move_unaligned_straight): Likewise.
23876 (arm_adjust_block_mem): Likewise.
23877
23878 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23879
23880 PR target/48901
23881 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23882 temp, cond and label.
23883 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23884
23885 PR target/52500
23886 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23887 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23888
23889 PR target/52501
23890 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23891 of prologue/epilogue functions.
23892
23893 PR target/52550
23894 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23895 Remove unused variable cfa_offset.
23896 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23897
23898 PR target/54639
23899 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23900 type promotion to unsigned.
23901
23902 PR target/54640
23903 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23904 for HOST_WIDE_INT of 32 bit / same size as int.
23905 (arm_block_move_unaligned_straight): Likewise.
23906 (arm_adjust_block_mem): Likewise.
23907
23908 PR target/54662
23909 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23910 ALL_CFLAGS.
23911
23912 2013-02-26 Marek Polacek <polacek@redhat.com>
23913
23914 PR tree-optimization/56426
23915 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23916
23917 2013-02-26 Richard Biener <rguenther@suse.de>
23918
23919 PR target/56444
23920 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23921 unused variable loops.
23922
23923 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23924
23925 PR tree-optimization/56448
23926 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23927 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23928 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23929 later operands of the references, or even first operand for
23930 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23931
23932 PR tree-optimization/56443
23933 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23934 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23935 to type_for_mode langhook.
23936
23937 2013-02-25 Matt Turner <mattst88@gmail.com>
23938
23939 * doc/invoke.texi: Document r4700.
23940
23941 2013-02-25 Richard Biener <rguenther@suse.de>
23942
23943 PR tree-optimization/56175
23944 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23945 split out from ...
23946 (simplify_bitwise_binary): ... here. Also guard the conversion
23947 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23948
23949 2013-02-25 Catherine Moore <clm@codesourcery.com>
23950
23951 Revert:
23952 2013-02-24 Catherine Moore <clm@codesourcery.com>
23953 Maciej W. Rozycki <macro@codesourcery.com>
23954 Tom de Vries <tom@codesourcery.com>
23955 Nathan Sidwell <nathan@codesourcery.com>
23956 Iain Sandoe <iain@codesourcery.com>
23957 Nathan Froyd <froydnj@codesourcery.com>
23958 Chao-ying Fu <fu@mips.com>
23959
23960 * doc/extend.texi: (micromips, nomicromips, nocompression):
23961 Document new function attributes.
23962 * doc/invoke.texi (minterlink-compressed, mmicromips,
23963 m14k, m14ke, m14kec): Document new options.
23964 (minterlink-mips16): Update documentation.
23965 * doc/md.texi (ZC, ZD): Document new constraints.
23966 * configure.ac (gcc_cv_as_micromips): Check if linker
23967 supports the .set micromips directive.
23968 * configure: Regenerate.
23969 * config.in: Regenerate.
23970 * config/mips/mips-tables.opt: Regenerate.
23971 * config/mips/micromips.md: New file.
23972 * constraints.md (ZC, AD): New constraints.
23973 * config/mips/predicates.md (movep_src_register): New predicate.
23974 (movep_src_operand): New predicate.
23975 (non_volatile_mem_operand): New predicate.
23976 * config/mips/mips.md (multimem): New type.
23977 (length): Differentiate between 17-bit and 18-bit branch offsets.
23978 (MOVEP1, MOVEP2): New mode iterator.
23979 (mov_<load>l): Use ZC constraint.
23980 (mov_<load>r): Likewise.
23981 (mov_<store>l): Likewise.
23982 (mov_<store>r): Likewise.
23983 (*branch_equality<mode>_inverted): Add microMIPS support.
23984 (*branch_equality<mode>): Likewise.
23985 (*jump_absolute): Likewise.
23986 (indirect_jump_<mode>): Likewise.
23987 (tablejump_<mode>): Likewise.
23988 (<optab>_internal): Likewise.
23989 (sibcall_internal): Likewise.
23990 (sibcall_value_internal): Likewise.
23991 (prefetch): Use constraint ZD.
23992 * config/mips/mips.opt (minterlink-compressed): New option.
23993 (minterlink-mips16): Now an alias for minterlink-compressed.
23994 (mmicromips): New option.
23995 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23996 (compare_and_swap_12): Likewise.
23997 (sync_add<mode>): Likewise.
23998 (sync_<optab>_12): Likewise.
23999 (sync_old_<optab>_12): Likewise.
24000 (sync_new_<optab>_12): Likewise.
24001 (sync_nand_12): Likewise.
24002 (sync_old_nand_12): Likewise.
24003 (sync_new_nand_12): Likewise.
24004 (sync_sub<mode>): Likewise.
24005 (sync_old_add<mode>): Likewise.
24006 (sync_old_sub<mode>): Likewise.
24007 (sync_new_add<mode>): Likewise.
24008 (sync_new_sub<mode>): Likewise.
24009 (sync_<optab><mode>): Likewise.
24010 (sync_old_<optab><mode>): Likewise.
24011 (sync_new_<optab><mode>): Likewise.
24012 (sync_nand<mode>): Likewise.
24013 (sync_old_nand<mode>): Likewise.
24014 (sync_new_nand<mode>): Likewise.
24015 (sync_lock_test_and_set<mode>): Likewise.
24016 (test_and_set_12): Likewise.
24017 (atomic_compare_and_swap<mode>): Likewise.
24018 (atomic_exchange<mode>_llsc): Likewise.
24019 (atomic_fetch_add<mode>_llsc): Likewise.
24020 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24021 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24022 (umips_save_restore_pattern_p): Likewise.
24023 (umips_load_store_pair_p): Likewise.
24024 (umips_output_load_store_pair): Likewise.
24025 (umips_movep_target_p): Likewise.
24026 (umips_12bit_offset_address_p): Likewise.
24027 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24028 (mips_base_mips16): Rename this...
24029 (mips_base_compression_flags): ...to this. Update all uses.
24030 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24031 (mips_mips16_decl_p): Delete.
24032 (mips_nomips16_decl_p): Delete.
24033 (mips_get_compress_on_flags): New function.
24034 (mips_get_compress_off_flags): New function.
24035 (mips_get_compress_mode): New function.
24036 (mips_get_compress_on_name): New function.
24037 (mips_get_compress_off_name): New function.
24038 (mips_insert_attributes): Support multiple compression types.
24039 (mips_merge_decl_attributes): Likewise.
24040 (umips_12bit_offset_address_p): New function.
24041 (mips_start_function_definition): Emit .set micromips directive.
24042 (mips_call_may_need_jalx_p): New function.
24043 (mips_function_ok_for_sibcall): Add microMIPS support.
24044 (mips_print_operand_punctuation): Support short delay slots and
24045 compact jumps.
24046 (umips_swm_mask, umips_swm_encoding): New.
24047 (umips_build_save_restore): New function.
24048 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24049 (was_mips16_p): Remove.
24050 (old_compression_mode): New.
24051 (mips_set_compression_mode): New function.
24052 (mips_set_current_function): Add microMIPS support.
24053 (mips_option_override): Likewise.
24054 (umips_save_restore_pattern_p): New function.
24055 (umips_output_save_restore): New function.
24056 (umips_load_store_pair_p_1): New function.
24057 (umips_load_store_pair_p): New function.
24058 (umips_output_load_store_pair_1): New function.
24059 (umips_output_load_store_pair): New function.
24060 (umips_movep_target_p) New function.
24061 (mips_prepare_pch_save): Add microMIPS support.
24062 * config/mips/mips.h (TARGET_COMPRESSION): New.
24063 (TARGET_CPU_CPP_BUILTINS): Update macro
24064 to use new compression flags and to support microMIPS.
24065 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24066 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24067 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24068 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24069 (ASM_SPEC): Support mmicromips and mno-micromips.
24070 (M16STORE_REG_P): New macro.
24071 (MIPS_CALL): Support TARGET_MICROMIPS.
24072 (MICROMIPS_J): New macro.
24073 (mips_base_mips16): Rename this...
24074 (mips_base_compression_flags): ...to this.
24075 (UMIPS_12BIT_OFFSET_P): New macro.
24076 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24077 (MULTILIB_DIRNAMES): Likewise.
24078
24079 2013-02-25 Tom de Vries <tom@codesourcery.com>
24080
24081 PR rtl-optimization/56131
24082 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24083 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24084 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24085
24086 2013-02-25 Tobias Burnus <burnus@net-b.de>
24087
24088 * doc/invoke.texi (-fsanitize=): Move from optimization
24089 to debugging options.
24090
24091 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24092
24093 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24094
24095 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24096 Alexander Monakov <amonakov@ispras.ru>
24097
24098 PR middle-end/56077
24099 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24100 flush pending lists also on non-jumps. Adjust comment.
24101
24102 2013-02-24 Catherine Moore <clm@codesourcery.com>
24103 Maciej W. Rozycki <macro@codesourcery.com>
24104 Tom de Vries <tom@codesourcery.com>
24105 Nathan Sidwell <nathan@codesourcery.com>
24106 Iain Sandoe <iain@codesourcery.com>
24107 Nathan Froyd <froydnj@codesourcery.com>
24108 Chao-ying Fu <fu@mips.com>
24109
24110 * doc/extend.texi: (micromips, nomicromips, nocompression):
24111 Document new function attributes.
24112 * doc/invoke.texi (minterlink-compressed, mmicromips,
24113 m14k, m14ke, m14kec): Document new options.
24114 (minterlink-mips16): Update documentation.
24115 * doc/md.texi (ZC, ZD): Document new constraints.
24116 * configure.ac (gcc_cv_as_micromips): Check if linker
24117 supports the .set micromips directive.
24118 * configure: Regenerate.
24119 * config.in: Regenerate.
24120 * config/mips/mips-tables.opt: Regenerate.
24121 * config/mips/micromips.md: New file.
24122 * constraints.md (ZC, AD): New constraints.
24123 * config/mips/predicates.md (movep_src_register): New predicate.
24124 (movep_src_operand): New predicate.
24125 (non_volatile_mem_operand): New predicate.
24126 * config/mips/mips.md (multimem): New type.
24127 (length): Differentiate between 17-bit and 18-bit branch offsets.
24128 (MOVEP1, MOVEP2): New mode iterator.
24129 (mov_<load>l): Use ZC constraint.
24130 (mov_<load>r): Likewise.
24131 (mov_<store>l): Likewise.
24132 (mov_<store>r): Likewise.
24133 (*branch_equality<mode>_inverted): Add microMIPS support.
24134 (*branch_equality<mode>): Likewise.
24135 (*jump_absolute): Likewise.
24136 (indirect_jump_<mode>): Likewise.
24137 (tablejump_<mode>): Likewise.
24138 (<optab>_internal): Likewise.
24139 (sibcall_internal): Likewise.
24140 (sibcall_value_internal): Likewise.
24141 (prefetch): Use constraint ZD.
24142 * config/mips/mips.opt (minterlink-compressed): New option.
24143 (minterlink-mips16): Now an alias for minterlink-compressed.
24144 (mmicromips): New option.
24145 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24146 (compare_and_swap_12): Likewise.
24147 (sync_add<mode>): Likewise.
24148 (sync_<optab>_12): Likewise.
24149 (sync_old_<optab>_12): Likewise.
24150 (sync_new_<optab>_12): Likewise.
24151 (sync_nand_12): Likewise.
24152 (sync_old_nand_12): Likewise.
24153 (sync_new_nand_12): Likewise.
24154 (sync_sub<mode>): Likewise.
24155 (sync_old_add<mode>): Likewise.
24156 (sync_old_sub<mode>): Likewise.
24157 (sync_new_add<mode>): Likewise.
24158 (sync_new_sub<mode>): Likewise.
24159 (sync_<optab><mode>): Likewise.
24160 (sync_old_<optab><mode>): Likewise.
24161 (sync_new_<optab><mode>): Likewise.
24162 (sync_nand<mode>): Likewise.
24163 (sync_old_nand<mode>): Likewise.
24164 (sync_new_nand<mode>): Likewise.
24165 (sync_lock_test_and_set<mode>): Likewise.
24166 (test_and_set_12): Likewise.
24167 (atomic_compare_and_swap<mode>): Likewise.
24168 (atomic_exchange<mode>_llsc): Likewise.
24169 (atomic_fetch_add<mode>_llsc): Likewise.
24170 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24171 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24172 (umips_save_restore_pattern_p): Likewise.
24173 (umips_load_store_pair_p): Likewise.
24174 (umips_output_load_store_pair): Likewise.
24175 (umips_movep_target_p): Likewise.
24176 (umips_12bit_offset_address_p): Likewise.
24177 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24178 (mips_base_mips16): Rename this...
24179 (mips_base_compression_flags): ...to this. Update all uses.
24180 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24181 (mips_mips16_decl_p): Delete.
24182 (mips_nomips16_decl_p): Delete.
24183 (mips_get_compress_on_flags): New function.
24184 (mips_get_compress_off_flags): New function.
24185 (mips_get_compress_mode): New function.
24186 (mips_get_compress_on_name): New function.
24187 (mips_get_compress_off_name): New function.
24188 (mips_insert_attributes): Support multiple compression types.
24189 (mips_merge_decl_attributes): Likewise.
24190 (umips_12bit_offset_address_p): New function.
24191 (mips_start_function_definition): Emit .set micromips directive.
24192 (mips_call_may_need_jalx_p): New function.
24193 (mips_function_ok_for_sibcall): Add microMIPS support.
24194 (mips_print_operand_punctuation): Support short delay slots and
24195 compact jumps.
24196 (umips_swm_mask, umips_swm_encoding): New.
24197 (umips_build_save_restore): New function.
24198 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24199 (was_mips16_p): Remove.
24200 (old_compression_mode): New.
24201 (mips_set_compression_mode): New function.
24202 (mips_set_current_function): Add microMIPS support.
24203 (mips_option_override): Likewise.
24204 (umips_save_restore_pattern_p): New function.
24205 (umips_output_save_restore): New function.
24206 (umips_load_store_pair_p_1): New function.
24207 (umips_load_store_pair_p): New function.
24208 (umips_output_load_store_pair_1): New function.
24209 (umips_output_load_store_pair): New function.
24210 (umips_movep_target_p) New function.
24211 (mips_prepare_pch_save): Add microMIPS support.
24212 * config/mips/mips.h (TARGET_COMPRESSION): New.
24213 (TARGET_CPU_CPP_BUILTINS): Update macro
24214 to use new compression flags and to support microMIPS.
24215 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24216 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24217 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24218 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24219 (ASM_SPEC): Support mmicromips and mno-micromips.
24220 (M16STORE_REG_P): New macro.
24221 (MIPS_CALL): Support TARGET_MICROMIPS.
24222 (MICROMIPS_J): New macro.
24223 (mips_base_mips16): Rename this...
24224 (mips_base_compression_flags): ...to this.
24225 (UMIPS_12BIT_OFFSET_P): New macro.
24226 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24227 (MULTILIB_DIRNAMES): Likewise.
24228
24229 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24230
24231 PR target/52555
24232 * target-globals.c (save_target_globals): For init_reg_sets and
24233 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24234
24235 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24236
24237 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24238 * config/aarch64/t-aarch64
24239 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24240
24241 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24242
24243 PR inline-asm/56148
24244 * lra-constraints.c (process_alt_operands): Reload operand
24245 conflicting with earlier clobber only if no more other conflicting
24246 operands.
24247
24248 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24249
24250 PR sanitizer/56393
24251 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24252 if not linking a shared library.
24253
24254 2013-02-22 Seth LaForge <sethml@google.com>
24255
24256 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24257
24258 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24259
24260 * config/arm/arm.md (split for extendsidi): Update condition.
24261 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24262 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24263 (qhs_zextenddi_cstr): Likewise.
24264
24265 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24266
24267 PR middle-end/56420
24268 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24269 avoid signed wrapping.
24270 (expand_mult): Handle properly multiplication by
24271 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24272 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24273 in the compiler if coeff is HOST_WIDE_INT_MIN.
24274 (expand_divmod): Don't make ext_op1 static, change it's type to
24275 uhwi. Avoid undefined behavior in -INTVAL (op1).
24276
24277 PR rtl-optimization/50339
24278 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24279 field.
24280 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24281 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24282 into splitting_ashiftrt field.
24283 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24284 ASHIFTRT.
24285 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24286 choices.
24287
24288 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24289
24290 PR middle-end/56108
24291 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24292 are sure to go irrevocable.
24293
24294 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24295
24296 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24297 scalars are valid operands.
24298
24299 2013-02-21 Martin Jambor <mjambor@suse.cz>
24300
24301 PR tree-optimization/56310
24302 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24303 only matching indices and non-negative final offsets.
24304 (intersect_aggregates_with_edge): Pass src_idx to
24305 agg_replacements_to_vector. Pass src_idx insstead of index to
24306 intersect_with_agg_replacements.
24307
24308 2013-02-21 Martin Jambor <mjambor@suse.cz>
24309
24310 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24311 instead of hard-wired defaults.
24312
24313 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24314
24315 * doc/invoke.texi (MIPS Options): Update documentation of the
24316 floating-point multiply-accumulate instruction restrictions.
24317
24318 2013-02-21 Kostya Serebryany <kcc@google.com>
24319
24320 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24321 asan_shadow_offset on x86_64 linux.
24322
24323 2013-02-21 Richard Biener <rguenther@suse.de>
24324
24325 PR tree-optimization/56415
24326 Revert
24327 2013-02-11 Richard Biener <rguenther@suse.de>
24328
24329 PR tree-optimization/56273
24330 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24331 first VRP run.
24332
24333 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24334
24335 PR bootstrap/56258
24336 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24337 instead of @itemx.
24338
24339 PR inline-asm/56405
24340 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24341 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24342
24343 2013-02-20 Jan Hubicka <jh@suse.cz>
24344
24345 PR tree-optimization/56265
24346 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24347 when target is referenced for first time.
24348
24349 2013-02-20 Richard Biener <rguenther@suse.de>
24350
24351 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24352 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24353 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24354 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24355 not return anything.
24356 (rename_ssa_copies): Do not remove unused locals.
24357 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24358 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24359 * passes.c (execute_function_todo): Do not schedule unused locals
24360 removal if cleanup_tree_cfg did something.
24361 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24362 about the number of removed locals.
24363
24364 2013-02-20 Richard Biener <rguenther@suse.de>
24365
24366 PR tree-optimization/56398
24367 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24368
24369 2013-02-20 Martin Jambor <mjambor@suse.cz>
24370
24371 PR tree-optimization/55334
24372 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24373 restricted pointers to arrays.
24374
24375 2013-02-20 Richard Biener <rguenther@suse.de>
24376 Jakub Jelinek <jakub@redhat.com>
24377
24378 PR tree-optimization/56396
24379 * tree-ssa-ccp.c (n_const_val): New static variable.
24380 (get_value): Return NULL for SSA names we don't have a lattice
24381 entry for.
24382 (ccp_initialize): Initialize n_const_val.
24383 * tree-ssa-copy.c (n_copy_of): New static variable.
24384 (init_copy_prop): Initialize n_copy_of.
24385 (get_value): Return NULL_TREE for SSA names we don't have a
24386 lattice entry for.
24387
24388 2013-02-20 Martin Jambor <mjambor@suse.cz>
24389
24390 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24391
24392 2013-02-20 Richard Biener <rguenther@suse.de>
24393
24394 * genpreds.c (write_lookup_constraint): Do not compare first
24395 letter of the constraint again.
24396
24397 2013-02-20 Richard Biener <rguenther@suse.de>
24398
24399 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24400 and ceil_log2.
24401 (get_use_iv_cost): Terminate hashtable walk when coming across
24402 an empty entry.
24403
24404 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24405
24406 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24407 reassociation for avx2 targets.
24408
24409 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24410
24411 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24412 Add version check for v8.10.a to enable microblaze_has_clz
24413 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24414 version and TARGET_PATTERN_COMPARE check
24415 * config/microblaze/microblaze.md: New clzsi2 instruction
24416
24417 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24418
24419 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24420 function before branching.
24421
24422 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24423
24424 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24425 DUMP_INSN_RTX_UID.
24426 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24427
24428 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24429
24430 PR middle-end/55889
24431 * sel-sched.c: Include ira.h.
24432 (implicit_clobber_conflict_p): New function.
24433 (moveup_expr): Use it.
24434 * Makefile.in (sel-sched.o): Depend on ira.h.
24435
24436 2013-02-19 Richard Biener <rguenther@suse.de>
24437
24438 PR tree-optimization/56384
24439 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24440 (vn_hash_type): Split out from ...
24441 (vn_hash_constant_with_type): ... here.
24442 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24443 (vn_phi_eq): Compare types from vn_phi_s structure.
24444 (vn_phi_lookup): Populate vn_phi_s type.
24445 (vn_phi_insert): Likewise.
24446
24447 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24448
24449 PR tree-optimization/56350
24450 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24451 if haven't found reduction or nested cycle operand, rather than
24452 asserting we must find it.
24453
24454 PR tree-optimization/56381
24455 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24456 to fold_build3.
24457
24458 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24459 Jakub Jelinek <jakub@redhat.com>
24460
24461 PR target/52555
24462 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24463 (swap_optab_enable): Same.
24464 (init_all_optabs): Use argument instead of global.
24465 * tree.h (struct tree_optimization_option): New field target_optabs.
24466 * expr.h (init_all_optabs): Add argument to prototype.
24467 (TREE_OPTIMIZATION_OPTABS): New.
24468 (save_optabs_if_changed): Protoize.
24469 * optabs.h: Declare this_fn_optabs.
24470 * optabs.c (save_optabs_if_changed): New.
24471 Declare this_fn_optabs.
24472 (init_optabs): Add argument to init_all_optabs() call.
24473 * function.c (invoke_set_current_function_hook): Handle per
24474 function optabs.
24475 * function.h (struct function): New field optabs.
24476 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24477 optimization_current_node has changed.
24478 * target-globals.h (save_target_globals_default_opts): Protoize.
24479 * target-globals.c (save_target_globals_default_opts): New.
24480
24481 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24482
24483 PR target/56347
24484 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24485 registers %fr12 and %fr12R as call used.
24486
24487 PR target/56214
24488 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24489 and HImode, require all displacements to be an integer multiple of
24490 their mode size.
24491 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24492 only allow QImode and HImode when reload is in progress and strict is
24493 true. Likewise for symbolic addresses. Use base14_operand to check
24494 displacements in REG+BASE addresses.
24495
24496 2013-02-18 Richard Biener <rguenther@suse.de>
24497
24498 PR tree-optimization/56366
24499 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24500 handle sign-conversion of outer-loop initial induction value.
24501
24502 2013-02-18 Richard Biener <rguenther@suse.de>
24503
24504 PR middle-end/56349
24505 * cfghooks.c (merge_blocks): If we merge a latch into another
24506 block adjust references to it.
24507 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24508 (verify_loop_structure): Verify that a recorded latch is in fact
24509 a latch.
24510
24511 2013-02-18 Richard Biener <rguenther@suse.de>
24512
24513 PR tree-optimization/56321
24514 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24515 order SSA name release and virtual operand unlinking.
24516
24517 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24518
24519 * config/microblaze/microblaze.md (save_stack_block): Define.
24520 (restore_stack_block): Likewise.
24521
24522 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24523
24524 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24525 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24526 * config/microblaze/microblaze.c (microblaze_option_override):
24527 Bail out early for PIC modes when target does not support PIC.
24528
24529 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24530
24531 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24532 Replace with a microblaze version.
24533 (microblaze_trampoline_init): Adapt for microblaze.
24534 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24535 microblaze.
24536
24537 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24538 Dodji Seketeli <dodji@redhat.com>
24539
24540 PR asan/56330
24541 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24542 (instrument_mem_region_access): Do not forget to always put
24543 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24544 0) statement, even for cases where either 'base' or 'base + len'
24545 are not instrumented -- because they have been previously
24546 instrumented. Simplify the logic by putting all the statements
24547 instrument 'base + len' inside a sequence, and then insert that
24548 sequence right before the current insertion point. Then, to
24549 instrument 'base + len', just get an iterator on that statement.
24550 And do not forget to update the pointer to iterator the function
24551 received as argument.
24552
24553 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24554
24555 PR rtl-optimization/56348
24556 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24557
24558 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24559
24560 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24561 (clean_graph_dump_file): Pass base to start_graph_dump.
24562
24563 2013-02-14 Richard Henderson <rth@redhat.com>
24564
24565 PR target/55941
24566 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24567
24568 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24569
24570 * collect2-aix.h: Define F_LOADONLY.
24571
24572 2013-02-14 Richard Biener <rguenther@suse.de>
24573
24574 PR lto/50494
24575 * varasm.c (output_constant_def_1): Get the decl representing
24576 the constant as argument.
24577 (output_constant_def): Wrap output_constant_def_1.
24578 (make_decl_rtl): Use output_constant_def_1 with the decl
24579 representing the constant.
24580 (build_constant_desc): Optionally re-use a decl already
24581 representing the constant.
24582 (tree_output_constant_def): Adjust.
24583
24584 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24585
24586 Fix an asan crash
24587 * asan.c (instrument_builtin_call): Really put the length of the
24588 second source argument into src1_len.
24589
24590 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24591
24592 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24593 argument. If it is false, don't create edge from then_bb to
24594 fallthru_bb.
24595 (insert_if_then_before_iter): Pass true to it.
24596 (build_check_stmt): Pass false to it.
24597 (transform_statements): Flush hash table only on extended basic
24598 block boundaries, rather than at the beginning of every bb.
24599 Don't flush hash table on nonfreeing_call_p calls.
24600 * tree-flow.h (nonfreeing_call_p): New prototype.
24601 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24602
24603 2013-02-13 David S. Miller <davem@davemloft.net>
24604
24605 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24606
24607 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24608
24609 PR target/56184
24610 * ira.c (max_regno_before_ira): Move from ...
24611 (ira): ... here.
24612 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24613 vec_safe_length.
24614
24615 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24616
24617 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24618
24619 2013-02-13 Richard Biener <rguenther@suse.de>
24620
24621 PR lto/56295
24622 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24623 globals in MEM_REFs.
24624
24625 2013-02-13 Richard Biener <rguenther@suse.de>
24626
24627 * loop-init.c (loop_optimizer_init): Clear loop state when
24628 re-initializing preserved loops.
24629 * loop-unswitch.c (unswitch_single_loop): Return whether
24630 we unswitched the loop. Do not verify loop state here.
24631 (unswitch_loops): When we unswitched a loop discover new loops.
24632
24633 2013-02-13 Kostya Serebryany <kcc@google.com>
24634
24635 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24636 on x86_64 linux.
24637 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24638
24639 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24640
24641 Avoid instrumenting duplicated memory access in the same basic block
24642 * Makefile.in (asan.o): Add new dependency on hash-table.h
24643 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24644 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24645 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24646 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24647 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24648 (get_mem_ref_of_assignment): New functions.
24649 (get_mem_refs_of_builtin_call): Extract from
24650 instrument_builtin_call and tweak a little bit to make it fit with
24651 the new signature.
24652 (instrument_builtin_call): Use the new
24653 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24654 of is_gimple_builtin_call.
24655 (instrument_derefs, instrument_mem_region_access): Insert the
24656 instrumented memory reference into the hash table.
24657 (maybe_instrument_assignment): Renamed instrument_assignment into
24658 this, and change it to advance the iterator when instrumentation
24659 actually happened and return true in that case. This makes it
24660 homogeneous with maybe_instrument_assignment, and thus give a
24661 chance to callers to be more 'regular'.
24662 (transform_statements): Clear the memory reference hash table
24663 whenever we enter a new BB, when we cross a function call, or when
24664 we are done transforming statements. Use
24665 maybe_instrument_assignment instead of instrumentation. No more
24666 need to special case maybe_instrument_assignment and advance the
24667 iterator after calling it; it's now handled just like
24668 maybe_instrument_call. Update comment.
24669
24670 2013-02-13 Richard Biener <rguenther@suse.de>
24671
24672 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24673 Fix loop discovery code.
24674
24675 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24676
24677 PR inline-asm/56148
24678 * lra-constraints.c (process_alt_operands): Match early clobber
24679 operand with itself. Check conflicts with earlyclobber only if
24680 the operand is not reloaded. Prefer to reload conflicting operand
24681 if earlyclobber and matching operands are the same.
24682
24683 2013-02-12 Richard Biener <rguenther@suse.de>
24684
24685 PR lto/56297
24686 * lto-streamer-out.c (write_symbol): Do not output symbols
24687 for hard register variables.
24688
24689 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24690
24691 PR target/54222
24692 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24693 (umulsidi3_insn, mulsidi3_insn): New insns.
24694
24695 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24696
24697 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24698 (struct tune_params): Add vec_costs field.
24699 * config/arm/arm.c (arm_builtin_vectorization_cost)
24700 (arm_add_stmt_cost): New functions.
24701 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24702 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24703 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24704 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24705 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24706 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24707 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24708
24709 2013-02-12 Richard Biener <rguenther@suse.de>
24710
24711 PR lto/56295
24712 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24713 decls again if possible.
24714
24715 2013-02-12 Richard Biener <rguenther@suse.de>
24716
24717 PR middle-end/56288
24718 * tree-ssa.c (verify_ssa_name): Fix check, move
24719 SSA_NAME_IN_FREE_LIST check up.
24720
24721 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24722 Steven Bosscher <steven@gcc.gnu.org>
24723
24724 PR rtl-optimization/56151
24725 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24726 equal to op0 or op1, and last_insn pattern is CODE operation
24727 with MEM dest and one of the operands matches that MEM.
24728
24729 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24730
24731 * doc/extend.texi: Document Function Multiversioning and "default"
24732 parameter string to target attribute.
24733 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24734 target attribute parameter is "default".
24735 (ix86_compare_version_priority): Remove checks for target attribute.
24736 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24737 Remove check for target attribute equal to NULL. Add assert.
24738 (ix86_generate_version_dispatcher_body): Change error to sorry.
24739
24740 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24741 Jack Howarth <howarth@bromo.med.uc.edu>
24742 Patrick Marlier <patrick.marlier@gmail.com>
24743
24744 PR libitm/55693
24745 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24746 define ENDFILE_SPEC as TM_DESTRUCTOR.
24747 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24748
24749 2013-02-11 Alexander Potapenko <glider@google.com>
24750 Jack Howarth <howarth@bromo.med.uc.edu>
24751 Jakub Jelinek <jakub@redhat.com>
24752
24753 PR sanitizer/55617
24754 * config/darwin.c (cdtor_record): Rename ctor_record.
24755 (sort_cdtor_records): Rename sort_ctor_records.
24756 (finalize_dtors): New routine to sort destructors by
24757 priority before use in assemble_integer.
24758 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24759
24760 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24761
24762 PR rtl-optimization/56275
24763 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24764 offset is non-negative and less than cmode size before
24765 calling simplify_subreg.
24766
24767 2013-02-11 Richard Biener <rguenther@suse.de>
24768
24769 PR tree-optimization/56264
24770 * cfgloop.h (fix_loop_structure): Adjust prototype.
24771 * loop-init.c (fix_loop_structure): Return the number of
24772 newly discovered loops.
24773 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24774 are discovered, do a full loop-closed SSA rewrite.
24775
24776 2013-02-11 Richard Biener <rguenther@suse.de>
24777
24778 PR tree-optimization/56273
24779 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24780 first VRP run.
24781 (check_array_ref): Fix missing newline in dumps.
24782 (search_for_addr_array): Likewise.
24783
24784 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24785
24786 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24787
24788 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24789
24790 PR target/56256
24791 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24792
24793 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24794
24795 PR rtl-optimization/56246
24796 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24797 reload pseudo.
24798 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24799 constraints are satisfied.
24800
24801 2013-02-08 Jeff Law <law@redhat.com>
24802
24803 PR debug/53948
24804 * emit-rtl.c (reg_is_parm_p): New function.
24805 * regs.h (reg_is_parm_p): New prototype.
24806 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24807 callee-clobbered registers.
24808
24809 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24810
24811 PR target/56043
24812 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24813 If there is no implicit builtin declaration, just return NULL.
24814
24815 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24816
24817 * config/i386/sse.md (FMAMODEM): New mode iterator.
24818 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24819 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24820
24821 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24822
24823 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24824 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24825 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24826
24827 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24828
24829 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24830 (microblaze*-*-elf): Likewise.
24831 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24832 LINK_SPEC.
24833 * config/microblaze/microblaze-c.c: Add builtin defines for
24834 _LITTLE_ENDIAN and _BIG_ENDIAN.
24835 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24836 add to TARGET_DEFAULT flags.
24837 Expand ASM_SPEC and LINK_SPEC.
24838 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24839 * config/microblaze/microblaze.md: Update extendsidi2 and
24840 movdi_internal instructions to use low-order / high-order reg
24841 print_operands.
24842 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24843 options and inversemask / mask of LITTLE_ENDIAN.
24844 * config/microblaze/t-microblaze: Expand multilib options to
24845 include mlittle-endian (le) and update exceptions patterns.
24846
24847 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24848
24849 PR rtl-optimization/56195
24850 * lra-constraints.c (get_reload_reg): Don't reuse regs
24851 if they have smaller mode than requested, if they have
24852 wider mode than requested, try to return a SUBREG.
24853
24854 PR tree-optimization/56250
24855 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24856 if type is unsigned and code isn't MULT_EXPR.
24857
24858 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24859
24860 PR tree-optimization/56064
24861 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24862 bits according to mode.
24863 * fixed-value.h (fixed_from_double_int)
24864 (const_fixed_from_double_int): Adjust comments.
24865
24866 2013-02-08 Richard Biener <rguenther@suse.de>
24867
24868 PR lto/56231
24869 * lto-streamer.h (struct data_in): Remove current_file, current_line
24870 and current_col members.
24871 * lto-streamer-out.c (lto_output_location): Stream changed bits
24872 en-block for efficiency.
24873 * lto-streamer-in.c (clear_line_info): Remove.
24874 (lto_input_location): Cache current file, line and column
24875 globally via local statics. Read changed bits en-block.
24876 (input_function): Do not call clear_line_info.
24877 (lto_read_body): Likewise.
24878 (lto_input_toplevel_asms): Likewise.
24879
24880 2013-02-08 Michael Matz <matz@suse.de>
24881
24882 PR tree-optimization/52448
24883 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24884 (nt_call_phase): New static.
24885 (add_or_mark_expr): Only mark accesses with newer phase than any
24886 call seen.
24887 (nonfreeing_call_p): New.
24888 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24889 (nt_fini_block): Keep blocks marked as visited.
24890 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24891
24892 2013-02-08 Richard Biener <rguenther@suse.de>
24893
24894 * ira.c (ira): Free broken dominator information.
24895
24896 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24897
24898 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24899
24900 2013-02-08 Marek Polacek <polacek@redhat.com>
24901
24902 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24903
24904 2013-02-08 Richard Biener <rguenther@suse.de>
24905
24906 PR middle-end/56181
24907 * cfgloop.h (flow_loops_find): Adjust.
24908 (bb_loop_header_p): Declare.
24909 * cfgloop.c (bb_loop_header_p): New function split out from ...
24910 (flow_loops_find): ... here. Adjust function signature,
24911 support incremental loop structure update.
24912 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24913 * cfgloopmanip.c (fix_loop_structure): Move ...
24914 * loop-init.c (fix_loop_structure): ... here.
24915 (apply_loop_flags): Split out from ...
24916 (loop_optimizer_init): ... here.
24917 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24918 in incremental mode, only remove dead loops here.
24919
24920 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24921
24922 PR target/54222
24923 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24924 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24925 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24926 (*round<mode>3.libgcc): New insns for fixed-modes.
24927 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24928 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24929 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24930 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24931 implementations. Define to __builtin_avr_absFX,
24932 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24933 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24934 __builtin_avr_countlsFX, respectively.
24935 * config/avr/avr-c.c (target.h): Include it.
24936 (enum avr_builtin_id): New enum.
24937 (avr_resolve_overloaded_builtin): New static function.
24938 (avr_register_target_pragmas): Use it to set
24939 targetm.resolve_overloaded_builtin.
24940 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24941 tree nodes used by DEF_BUILTIN.
24942 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24943 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24944 <AVR_BUILTIN_xxBITS>: Same.
24945
24946 2013-02-08 Richard Biener <rguenther@suse.de>
24947
24948 * cfgloop.c (verify_loop_structure): Properly handle
24949 a loop exiting to another loop header.
24950 * ira-int.h (ira_loops): Remove.
24951 * ira.c (ira_loops): Remove.
24952 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24953 (do_reload): Use loop_optimizer_finalize.
24954 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24955 number_of_loops to access the loop tree.
24956 (more_one_region_p): Likewise.
24957 (finish_loop_tree_nodes): Likewise.
24958 (rebuild_regno_allocno_maps): Likewise.
24959 (mark_loops_for_removal): Likewise.
24960 (mark_all_loops_for_removal): Likewise.
24961 (remove_unnecessary_regions): Likewise.
24962 (ira_build): Likewise.
24963 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24964
24965 2013-02-08 Richard Biener <rguenther@suse.de>
24966
24967 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24968 * ipa-pure-const.c (analyze_function): Avoid calling
24969 mark_irreducible_loops twice.
24970 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24971
24972 2013-02-07 David S. Miller <davem@davemloft.net>
24973
24974 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24975 on 'reg'.
24976 * var-tracking.c (vt_add_function_parameter): Test the presence of
24977 HAVE_window_save properly and do not remap argument registers when
24978 we have a leaf function.
24979
24980 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24981
24982 PR bootstrap/56227
24983 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24984 instead of "ll".
24985 * config/i386/i386.c (ix86_print_operand): Ditto.
24986
24987 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24988
24989 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24990
24991 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24992
24993 PR rtl-optimization/56225
24994 * lra-constraints.c (process_alt_operands): Check that reload hard
24995 reg can hold value for strict_low_part.
24996
24997 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24998
24999 PR debug/56154
25000 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25001 dwarf2out_end_function.
25002 (in_first_function_p, maybe_at_text_label_p,
25003 first_loclabel_num_not_at_text_label): New variables.
25004 (dwarf2out_var_location): In the first function find out
25005 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25006 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25007 functions.
25008
25009 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25010
25011 PR rtl-optimization/56178
25012 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25013 SUBREG of a register. Tidy up related block of code.
25014 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25015 note if the source is a register or a SUBREG of a register.
25016
25017 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25018
25019 PR target/56228
25020 * config/rs6000/rs6000.md (ptrm): New mode attr.
25021 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25022 call_value_indirect_aix<pttrsize>,
25023 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25024 m in constraints.
25025
25026 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25027
25028 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25029 if -bnortl. Convert to strcmp and strncmp.
25030
25031 2013-02-07 Alan Modra <amodra@gmail.com>
25032
25033 PR target/54009
25034 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25035 addresses won't wrap when offsetting.
25036 (rs6000_secondary_reload): Provide secondary reloads needed for
25037 wrapping LO_SUM addresses.
25038
25039 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25040
25041 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25042 MACH, just __MACH__.
25043
25044 2013-02-06 Richard Biener <rguenther@suse.de>
25045
25046 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25047 instead of calling fix_loop_structure.
25048
25049 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25050
25051 PR middle-end/56217
25052 * omp-low.c (use_pointer_for_field): Return false if
25053 lower_send_shared_vars doesn't generate any copy-out code.
25054
25055 2013-02-06 Tom de Vries <tom@codesourcery.com>
25056
25057 PR rtl-optimization/56131
25058 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25059 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25060 of the label is NULL. Add comment.
25061
25062 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25063
25064 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25065
25066 PR sanitizer/55374
25067 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25068 (STATIC_LIBTSAN_LIBS): Likewise.
25069 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25070 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25071 is defined, don't add anything else beyond that.
25072 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25073 (LINK_COMMAND_SPEC): Use them.
25074
25075 PR tree-optimization/56205
25076 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25077 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25078 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25079
25080 2013-02-05 Richard Biener <rguenther@suse.de>
25081
25082 PR tree-optimization/53342
25083 PR tree-optimization/53185
25084 * tree-vectorizer.h (vect_check_strided_load): Remove.
25085 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25086 not disallow peeling for vectorized strided loads.
25087 (vect_check_strided_load): Make static and simplify.
25088 (vect_analyze_data_refs): Adjust.
25089 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25090 correctly when vectorizing strided loads.
25091
25092 2013-02-05 Richard Biener <rguenther@suse.de>
25093
25094 * doc/install.texi: Refer to ISL, not PPL.
25095
25096 2013-02-05 Jan Hubicka <jh@suse.cz>
25097
25098 PR tree-optimization/55789
25099 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25100
25101 2013-02-05 Jan Hubicka <jh@suse.cz>
25102
25103 PR tree-optimization/55789
25104 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25105 the dead call anyway.
25106
25107 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25108
25109 PR sanitizer/55374
25110 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25111
25112 2013-02-04 Alexander Potapenko <glider@google.com>
25113 Jack Howarth <howarth@bromo.med.uc.edu>
25114 Jakub Jelinek <jakub@redhat.com>
25115
25116 PR sanitizer/55617
25117 * config/darwin.c (sort_ctor_records): Stabilized qsort
25118 on constructor priority by using original position.
25119 (finalize_ctors): New routine to sort constructors by
25120 priority before use in assemble_integer.
25121 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25122
25123 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25124
25125 PR libstdc++/54314
25126 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25127 about visibility on artificial decls.
25128 * config/sol2.c (solaris_assemble_visibility): Likewise.
25129
25130 2013-02-04 Kai Tietz <ktietz@redhat.com>
25131
25132 PR target/56186
25133 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25134 argument and improve checking of return-argument types for 16-byte
25135 modes.
25136 (ix86_function_value_1): Add additional valtype argument on call
25137 of function_value_64.
25138 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25139 handling infunction_value_64 function.
25140
25141 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25142
25143 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25144
25145 2013-02-04 Richard Biener <rguenther@suse.de>
25146
25147 PR tree-optimization/56188
25148 * tree-ssa-structalias.c (label_visit): Consider case with
25149 initially non-empty points-to set.
25150 (perform_var_substitution): Dump node mapping and clean up.
25151
25152 2013-02-04 Richard Guenther <rguenther@suse.de>
25153
25154 PR lto/56168
25155 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25156 node prevail as last resort.
25157 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25158 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25159
25160 2013-02-04 Richard Biener <rguenther@suse.de>
25161
25162 PR tree-optimization/56113
25163 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25164 Merge into ...
25165 (equiv_class_lookup_or_add): ... this.
25166 (label_visit): Adjust and fix error in previous patch.
25167 (perform_var_substitution): Adjust.
25168
25169 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25170
25171 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25172 * config/sh/sh4-300.md: Likewise.
25173 * config/sh/sh4a.md: Likewise.
25174 * config/sh/constraints.md: Likewise.
25175 * config/sh/sh.md: Likewise.
25176 * config/sh/netbsd-elf.h: Likewise.
25177 * config/sh/predicates.md: Likewise.
25178 * config/sh/sh-protos.h: Likewise.
25179 * config/sh/ushmedia.h: Likewise.
25180 * config/sh/linux.h: Likewise.
25181 * config/sh/sh.c: Likewise.
25182 * config/sh/superh.h: Likewise.
25183 * config/sh/elf.h: Likewise.
25184 * config/sh/sh4.md: Likewise.
25185 * config/sh/sh.h: Likewise.
25186
25187 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25188
25189 * config/pa/constraints.md: Adjust unused letters. Change "T"
25190 constraint to match_test floating_point_store_memory_operand().
25191 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25192 (base14_operand): New.
25193 (floating_point_store_memory_operand): New.
25194 (integer_store_memory_operand): Revise to use base14_operand and
25195 reg_plus_base_memory_operand.
25196 (move_dest_operand): Allow symbolic_memory_operands.
25197 (symbolic_memory_operand): Check for LO_SOM.
25198 (symbolic_operand): Change default case to break.
25199 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25200 CONST_DOUBLE values to be reloaded by putting them into memory when
25201 the destination is a floating point register.
25202 (movdf): Remove code to handle CONST_DOUBLE.
25203 (movsf): Likewise.
25204 (reload_indf_r1): New.
25205 (reload_insf_r1): New.
25206 Consistently use "Q" and "T" constraints with integer and floating
25207 point move instructions, respectively.
25208 (movdi): Remove FAIL.
25209 Change predicate for source operand unamed DImode move from
25210 general_operand to move_src_operand.
25211 (umulsidi3): Change predicate for destination operand to
25212 register_operand.
25213 Likewise for similar unamed patterns.
25214 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25215 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25216 (hppa_legitimize_address): Simplify mask calculation.
25217 (pa_emit_move_sequence): Revised handling of secondary reloads from
25218 REG+D addresses for floating point loads and stores. Directly handle
25219 loading CONST0_RTX (mode) to a floating point register.
25220 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25221 to floating point registers. Don't restrict secondary reloads to
25222 floating point registers to integer modes. Revise some comments and
25223 cleanup some code.
25224 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25225 (pa_legitimate_address_p): New.
25226 (pa_legitimize_reload_address): New.
25227 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25228 (STRICT_REG_OK_FOR_BASE_P): New.
25229 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25230 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25231
25232 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25233 Andrew Dixie <andrewd@gentrack.com>
25234
25235 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25236 flag set.
25237
25238 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25239
25240 * expmed.c (extract_bit_field_1): Pass the full width of the
25241 structure to get_best_reg_extraction_insn.
25242
25243 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25244
25245 PR target/54601
25246 * configure.ac (use_cxa_atexit): Add AIX.
25247 * configure: Regenerate.
25248
25249 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25250
25251 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25252
25253 PR debug/54793
25254 * final.c (need_profile_function): New variable.
25255 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25256 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25257 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25258 notes, targetm.asm_out.function_prologue doesn't emit anything,
25259 HAVE_prologue and profiler should be emitted before prologue,
25260 set need_profile_function instead of emitting it.
25261 (final_scan_insn): If need_profile_function, emit
25262 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25263 NOTE_INSN_FUNCTION_BEG note.
25264
25265 2013-02-01 Richard Henderson <rth@redhat.com>
25266
25267 * config/rs6000/rs6000.md (smulditi3): New.
25268 (umulditi3): New.
25269
25270 * config/alpha/alpha.md (umulditi3): New.
25271
25272 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25273
25274 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25275 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25276
25277 2013-02-01 Richard Biener <rguenther@suse.de>
25278
25279 PR tree-optimization/56113
25280 * tree-ssa-structalias.c (label_visit): Reduce work for
25281 single-predecessor nodes.
25282
25283 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25284
25285 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25286 range isn't testing for zero.
25287
25288 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25289
25290 PR middle-end/56113
25291 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25292
25293 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25294 Nick Clifton <nickc@redhat.com>
25295
25296 * config/v850/constraints.md (Q): Define as a memory constraint.
25297 * config/v850/predicates.md (label_ref_operand): New predicate.
25298 (e3v5_shift_operand): New predicate.
25299 (ior_operator): New predicate.
25300 * config/v850/t-v850: Add e3v5 multilib.
25301 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25302 (v850_gen_movdi): Prototype.
25303 * config/v850/v850.c: Add support for e3v5 architecture.
25304 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25305 TARGET_V850E_UP.
25306 (construct_save_jarl): Add e3v5 long JARL support.
25307 (v850_adjust_insn_length): New function. Adjust length of call
25308 insns when using e3v5 instructions.
25309 (v850_gen_movdi): New function: Generate instructions to move a
25310 DImode value.
25311 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25312 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25313 (TARGET_USE_FPU): Enable for e3v5.
25314 (CONST_OK_FOR_W): New macro.
25315 (ADJUST_INSN_LENGTH): Define.
25316 * config/v850/v850.md (UNSPEC_LOOP): Define.
25317 (attr cpu): Add v850e3v5.
25318 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25319 (movdi): New pattern.
25320 (movdi_internal): New pattern.
25321 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25322 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25323 (cstoresf4): Likewise.
25324 (cstoredf4): Likewise.
25325 (insv): New pattern.
25326 (rotlso3_a): New pattern.
25327 (rotlsi3_b): New pattern
25328 (rotlsi3_v850e3v5): New pattern.
25329 (doloop_begin): New pattern.
25330 (fix_loop_counter): New pattern.
25331 (doloop_end): New pattern.
25332 (branch_normal): Add e3v5 long branch support.
25333 (branch_invert): Likewise.
25334 (branch_z_normal): Likewise.
25335 (branch_z_invert): Likewise.
25336 (branch_nz_normal): Likewise.
25337 (branch_nz_invert): Likewise.
25338 (call_internal_short): Add e3v5 register-indirect JARL support.
25339 (call_internal_long): Likewise.
25340 (call_value_internal_short): Likewise.
25341 (call_value_internal_long): Likewise.
25342 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25343 (mloop): New option.
25344 * config.gcc: Add support for configuring v840e3v5 target.
25345 * doc/invoke.texi: Document new v850 specific command line options.
25346
25347 2013-01-31 Paul Koning <ni1d@arrl.net>
25348
25349 PR debug/55059
25350 PR debug/54508
25351 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25352 children if parent is a class.
25353 (prune_unused_types_prune): Don't add DW_AT_declaration.
25354
25355 2013-01-31 Richard Biener <rguenther@suse.de>
25356
25357 PR tree-optimization/56157
25358 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25359 match up operand with SLP child.
25360
25361 2013-01-31 Jason Merrill <jason@redhat.com>
25362
25363 PR debug/54410
25364 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25365 parameters the first time.
25366 (gen_scheduled_generic_parms_dies): Check completeness here.
25367
25368 2013-01-31 Richard Biener <rguenther@suse.de>
25369
25370 PR middle-end/53073
25371 * common.opt (faggressive-loop-optimizations): New flag,
25372 enabled by default.
25373 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25374 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25375 infer_loop_bounds_from_undefined by it.
25376
25377 2013-01-31 Richard Biener <rguenther@suse.de>
25378
25379 PR tree-optimization/56150
25380 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25381 visit virtual operands.
25382 (find_uses_to_rename_bb): Likewise.
25383
25384 2013-01-31 Richard Biener <rguenther@suse.de>
25385
25386 PR tree-optimization/56150
25387 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25388 mixed store non-store stmts.
25389
25390 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25391
25392 PR sanitizer/55374
25393 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25394 LIBASAN_EARLY_SPEC is defined.
25395 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25396 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25397 before %o.
25398 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25399
25400 PR c++/55742
25401 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25402 invalid args instead of ICEing on it.
25403 (ix86_valid_target_attribute_tree): Return error_mark_node if
25404 ix86_valid_target_attribute_inner_p failed.
25405 (ix86_valid_target_attribute_p): Return false only if
25406 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25407 target("default") attribute.
25408 (sorted_attr_string): Change argument from const char * to tree,
25409 merge in all target attribute arguments rather than just one.
25410 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25411 instead of free. Avoid using strcat.
25412 (ix86_mangle_function_version_assembler_name): Mangle
25413 target("default") as if no target attribute is present. Adjust
25414 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25415 instead of xmalloc and XDELETEVEC instead of free.
25416 (ix86_function_versions): Don't return true if one of the decls
25417 doesn't have target attribute. If they don't and one of the decls
25418 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25419 sorted_attr_string caller. Use XDELETEVEC instead of free.
25420 (ix86_supports_function_versions): Remove.
25421 (make_name): Fix up formatting.
25422 (make_dispatcher_decl): Remove resolver_name and its initialization.
25423 Avoid leaking memory.
25424 (is_function_default_version): Return true if there is
25425 target("default") attribute rather than no target attribute at all.
25426 (make_resolver_func): Avoid leaking memory.
25427 (ix86_generate_version_dispatcher_body): Likewise.
25428 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25429 * target.def (supports_function_versions): Remove.
25430 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25431 * doc/tm.texi: Regenerated.
25432
25433 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25434
25435 PR rtl-optimization/56144
25436 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25437 for values with side effects.
25438
25439 2013-01-30 Richard Biener <rguenther@suse.de>
25440
25441 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25442 (sparseset_pop): Likewise.
25443 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25444 to be able to use quick_push in the worker loop.
25445
25446 2013-01-30 Marek Polacek <polacek@redhat.com>
25447
25448 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25449
25450 2013-01-30 Richard Biener <rguenther@suse.de>
25451
25452 PR lto/56147
25453 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25454
25455 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25456
25457 PR tree-optimization/56064
25458 * fixed-value.c (fixed_from_double_int): New function.
25459 * fixed-value.h (fixed_from_double_int): New prototype.
25460 (const_fixed_from_double_int): New static inline function.
25461 * fold-const.c (native_interpret_fixed): New static function.
25462 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25463 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25464 (native_encode_fixed): New static function.
25465 (native_encode_expr) <FIXED_CST>: Use it.
25466 (native_interpret_int): Move double_int worker code to...
25467 * double-int.c (double_int::from_buffer): ...this new static method.
25468 * double-int.h (double_int::from_buffer): Prototype it.
25469
25470 2013-01-30 Richard Biener <rguenther@suse.de>
25471
25472 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25473 New pointer-map and obstack.
25474 (init_alias_vars): Allocate pointer-map and obstack.
25475 (delete_points_to_sets): Free them.
25476 (find_what_var_points_to): Cache result.
25477 (find_what_p_points_to): Adjust for changed interface of
25478 find_what_var_points_to.
25479 (compute_points_to_sets): Likewise.
25480 (ipa_pta_execute): Likewise.
25481
25482 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25483
25484 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25485 * configure: Regenerate.
25486 * config.in: Regenerate.
25487 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25488 #nobits/#progbits if supported.
25489
25490 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25491
25492 PR target/56121
25493 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25494 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25495 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25496
25497 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25498
25499 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25500 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25501
25502 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25503
25504 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25505 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25506
25507 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25508
25509 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25510 declaration.
25511 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25512 * config/arm/cortex-a7.md: New bypasses using
25513 arm_mac_accumulator_is_result.
25514
25515 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25516
25517 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25518 (cortex_a7_neon_mla): Likewise.
25519 (cortex_a7_fpfmad): New reservation.
25520 (cortex_a7_fpmacs): Use ffmas and update required units.
25521 (cortex_a7_fpmuld): Update required units and latency.
25522 (cortex_a7_fpmacd): Likewise.
25523 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25524 (cortex_a7_neon). Likewise.
25525 (bypass) Update participating units.
25526
25527 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25528
25529 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25530 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25531 from fmac to ffma.
25532 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25533 (vfp_fmul): Use ffmad.
25534 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25535 (cortex_r4_fmacd): Use ffmad.
25536 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25537 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25538 (cortex_a9_fmacd): Use ffmad.
25539 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25540 (cortex_a8_vfp_macd): Use ffmad.
25541 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25542 (cortex_a5_fpmacd): Use ffmad.
25543 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25544 (cortex_a15_vfp_macd): Use ffmad.
25545 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25546
25547 2013-01-29 Jason Merrill <jason@redhat.com>
25548
25549 PR libstdc++/54314
25550 * varasm.c (default_assemble_visibility): Don't warn about
25551 visibility on artificial decls.
25552
25553 2013-01-29 Richard Biener <rguenther@suse.de>
25554
25555 PR tree-optimization/56113
25556 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25557 the bitmap leader.
25558 (label_visit): Free duplicate bitmaps and record the leader instead.
25559 (perform_var_substitution): Adjust.
25560
25561 2013-01-29 Richard Biener <rguenther@suse.de>
25562
25563 PR tree-optimization/55270
25564 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25565 the CFG, schedule loops for fixup.
25566
25567 2013-01-29 Nick Clifton <nickc@redhat.com>
25568
25569 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25570 SP_REG.
25571
25572 2013-01-28 Leif Ekblad <leif@rdos.net>
25573
25574 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25575 * config/i386/i386.h (TARGET_RDOS): New macro.
25576 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25577 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25578 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25579 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25580 DEFAULT_LARGE_SECTION_THRESHOLD.
25581 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25582 * config/i386/rdos.h: New file.
25583 * config/i386/rdos64.h: New file.
25584
25585 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25586
25587 PR other/54814
25588 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25589 TEST_HARD_REG_BIT.
25590
25591 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25592
25593 PR rtl-optimization/56117
25594 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25595 call cselib_lookup_from_insn on the MEM before calling
25596 add_insn_mem_dependence.
25597
25598 2013-01-28 Richard Biener <rguenther@suse.de>
25599
25600 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25601 to a stmt that didn't have one.
25602 (copy_phis_for_bb): Likewise for PHI arguments.
25603 (copy_debug_stmt): Likewise for debug stmts.
25604
25605 2013-01-28 Richard Biener <rguenther@suse.de>
25606
25607 PR tree-optimization/56034
25608 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25609 (partition_builtin_p): Adjust.
25610 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25611 it is the last partition.
25612 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25613 up the vertex for the definition.
25614 (classify_partition): Classify whether a partition is a
25615 PKIND_REDUCTION, thus has uses outside of the loop.
25616 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25617 Merge all PKIND_REDUCTION partitions into the last partition.
25618 (tree_loop_distribution): Seed partitions from reductions as well.
25619
25620 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25621
25622 PR tree-optimization/56125
25623 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25624 pow(x,c) into sqrt(x) * powi(x, n/2) or
25625 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25626 optimizing for size.
25627 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25628 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25629 integer.
25630
25631 PR tree-optimization/56094
25632 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25633 to UNKNOWN_LOCATION while gimplifying expr.
25634
25635 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25636
25637 PR target/56114
25638 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25639 operand 0 in movabs insn template for -masm=intel asm alternative.
25640 (*movabs<mode>_2): Ditto for operand 1.
25641
25642 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25643
25644 PR target/54663
25645 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25646 of microblaze-c.o
25647
25648 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25649
25650 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25651 tm_file.
25652
25653 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25654
25655 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25656 Undef to avoid warning.
25657
25658 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25659
25660 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25661 * configure: Regenerate.
25662
25663 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25664
25665 PR tree-optimization/56098
25666 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25667 for stmts with volatile ops.
25668 (cond_store_replacement): Don't optimize if assign has volatile ops.
25669 (cond_if_else_store_replacement_1): Don't optimize if either
25670 then_assign or else_assign have volatile ops.
25671 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25672 volatile ops.
25673
25674 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25675
25676 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25677
25678 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25679
25680 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25681 missing ':' in asm example.
25682
25683 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25684
25685 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25686 entries into lane and laneq entries.
25687 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25688 Remove AdvSIMD scalar modes.
25689 (aarch64_sq<r>dmulh_laneq<mode>): New.
25690 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25691 modes.
25692 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25693 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25694 * config/aarch64/iterators.md (VCOND): New.
25695 (VCONQ): New.
25696
25697 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25698
25699 PR target/54222
25700 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25701 Add NULL LIBNAME argument to existing definitions.
25702 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25703 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25704 * config/avr/avr.c (DEF_BUILTIN): Same.
25705 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25706 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25707 implementation is available (DECL_ASSEMBLER_NAME is set).
25708 (avr_fold_absfx): New static function.
25709 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25710 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25711 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25712 AVR_BUILTIN_ABSLLK.
25713 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25714 (abshk, absk, abslk, absllk): Provide as static inline functions.
25715
25716 2013-01-25 Marek Polacek <polacek@redhat.com>
25717
25718 PR tree-optimization/56035
25719 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25720
25721 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25722
25723 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25724 (*movtf_internal_rex64): Add (!o,C) alternative
25725 (*movxf_internal_rex64): Ditto.
25726 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25727
25728 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25729
25730 * doc/invoke.texi: fix typo.
25731 * doc/objc.texi: fix typo.
25732
25733 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25734
25735 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25736 for the first two alternatives.
25737
25738 2013-01-24 Diego Novillo <dnovillo@google.com>
25739
25740 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25741 (ggc-zone.o): Remove.
25742 * configure.ac: Remove option --with-gc.
25743 * configure: Re-generate.
25744 * doc/install.texi: Remove documentation for --with-gc.
25745 * gengtype.c (write_enum_defn): Remove. Update all users.
25746 (write_Types_process_field): Remove generation of gt_e_* argument.
25747 (output_type_enum): Remove. Update all users.
25748 (write_enum_defn): Remove. Update all users.
25749 (enum alloc_zone): Remove. Update all users.
25750 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25751 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25752 Update all callers.
25753 (struct ptr_data): Remove field TYPE. Update all users.
25754 (gt_pch_note_object): Remove argument TYPE. Update all users.
25755 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25756 Update all users.
25757 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25758 (struct alloc_zone): Remove.
25759 (ggc_internal_alloc_zone_stat): Remove.
25760 (ggc_internal_cleared_alloc_zone_stat): Remove.
25761 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25762 (ggc_pch_count_object): Remove last argument. Update all users.
25763 (ggc_pch_alloc_object): Remove last argument. Update all users.
25764 (struct alloc_zone): Remove.
25765 * ggc-zone.c: Remove.
25766 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25767 (struct alloc_zone): Remove.
25768 (ggc_alloc_typed_stat): Remove.
25769 (ggc_alloc_typed): Remove.
25770 (ggc_splay_alloc): Remove first argument.
25771 (rtl_zone): Remove. Update all users.
25772 (tree_zone): Remove. Update all users.
25773 (tree_id_zone): Remove. Update all users.
25774 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25775 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25776 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25777 * tree-ssanames.c: Remove references to zone allocator in comments.
25778
25779 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25780
25781 * config/avr/avr.c (avr_out_fract): Make register numbers that
25782 might be outside of source operand signed.
25783
25784 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25785
25786 * config/i386/constraints.md (Yf): New constraint.
25787 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25788 of f constraint to conditionaly disable x87 register preferences.
25789 (*movdf_internal): Ditto.
25790 (*movsf_internal): Ditto.
25791
25792 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25793
25794 PR inline-asm/55934
25795 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25796 that have operands with impossible constraints.
25797 Add a FIXME for a speed-up opportunity.
25798 * lra-constraints.c (process_alt_operands): Verify that a class
25799 selected from constraints on asms is valid for the operand mode.
25800 (curr_insn_transform): Remove incorrect comment.
25801
25802 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25803
25804 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25805 TOC operand is a valid symbol ref in the constant pool.
25806
25807 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25808
25809 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25810
25811 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25812
25813 PR target/54222
25814 * config/avr/stdfix.h: New file.
25815 * t-avr (stdfix-gcc.h): New rule to build it.
25816 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25817
25818 2013-01-23 Kostya Serebryany <kcc@google.com>
25819
25820 * config/darwin.h: remove dependency on
25821 CoreFoundation (asan on Mac OS).
25822
25823 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25824
25825 PR target/49069
25826 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25827 instead of cmpdi_operand for first comparison operand.
25828 Don't assert that comparison operands aren't both constants.
25829
25830 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25831
25832 * doc/install.texi (Downloading the Source): Update references to
25833 downloading separate components.
25834
25835 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25836
25837 * doc/extend.texi (__int128): Improve grammar.
25838
25839 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25840
25841 PR target/56028
25842 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25843 alternative to (o,r).
25844 (*movdi_internal_rex64): Remove (!o,n) alternative.
25845 (DImode immediate->memory splitter): Remove.
25846 (DImode immediate->memory peephole2): Remove.
25847 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25848 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25849 alternative to (!o,*r).
25850 (*movtf_internal_sse): New pattern.
25851 (*movxf_internal_rex64): New pattern.
25852 (*movxf_internal): Disable for TARGET_64BIT.
25853 (*movdf_internal_rex64): Remove (!o,F) alternative.
25854
25855 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25856
25857 PR middle-end/56074
25858 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25859 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25860 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25861 stmt locations where LOCATION_LOCUS of the stmt location is
25862 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25863
25864 PR target/55686
25865 * config/i386/i386.md (UNSPEC_STOS): New.
25866 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25867 *strsetqi_1): Add UNSPEC_STOS.
25868
25869 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25870
25871 PR c++/56067
25872 * doc/invoke.texi: Remove left over -Wsynth example.
25873
25874 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25875
25876 PR tree-optimization/56051
25877 * fold-const.c (fold_binary_loc): Don't fold
25878 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25879 a narrowing conversion, or widening conversion from signed
25880 to unsigned.
25881
25882 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25883
25884 PR rtl-optimization/56023
25885 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25886 dependent on debug instruction.
25887
25888 2013-01-21 Martin Jambor <mjambor@suse.cz>
25889
25890 PR middle-end/56022
25891 * function.c (allocate_struct_function): Call
25892 invoke_set_current_function_hook earlier.
25893
25894 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25895
25896 * reload1.c (init_reload): Only initialize reload_obstack
25897 during the first call.
25898
25899 2013-01-21 Marek Polacek <polacek@redhat.com>
25900
25901 * cfgloop.c (verify_loop_structure): Fix up grammar.
25902
25903 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25904
25905 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25906 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25907
25908 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25909
25910 PR target/56058
25911 * config/arm/marvell-pj4.md: Update copyright year.
25912 Fix up use of alu to alu_reg and simple_alu_imm.
25913
25914 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25915
25916 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25917
25918 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25919
25920 PR target/55433
25921 * lra-constraints.c (curr_insn_transform): Don't reuse original
25922 insn for secondary memory move when memory mode should be different.
25923
25924 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25925
25926 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25927 atomic_storedi_1): New patterns.
25928
25929 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25930
25931 btver2 pipeline descriptions.
25932 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25933 descriptions.
25934 * config/i386/i386.md (btver2_decode): New type attributes.
25935 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25936 type attributes.
25937 * config/i386/btver2.md: New file describing btver2 pipelines.
25938
25939 2013-01-19 Andrew Pinski <apinski@cavium.com>
25940
25941 PR tree-optimization/52631
25942 * tree-ssa-sccvn (visit_use): Before looking up the original
25943 statement, try looking up the simplified expression.
25944
25945 2013-01-19 Anthony Green <green@moxielogic.com>
25946
25947 * config/moxie/moxie.c (moxie_expand_prologue): Set
25948 current_function_static_stack_size.
25949
25950 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25951
25952 PR tree-optimization/56029
25953 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25954 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25955
25956 2013-01-18 Sharad Singhai <singhai@google.com>
25957
25958 PR tree-optimization/55995
25959 * dumpfile.c (dump_loc): Print location only if available.
25960 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25961
25962 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25963
25964 PR target/55433
25965 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25966 secondary memory move.
25967 (inherit_reload_reg): Use rclass instead of cl for
25968 check_secondary_memory_needed_p.
25969
25970 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25971
25972 PR middle-end/56015
25973 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25974 the case where writing real complex part of target modifies op1.
25975
25976 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25977
25978 * config/aarch64/aarch64-simd.md
25979 (aarch64_vcond_internal<mode>): Handle unordered cases.
25980 * config/aarch64/iterators.md (v_cmp_result): New.
25981
25982 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25983 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25984
25985 * config/arm/marvell-pj4.md: New file.
25986 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25987 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25988 (generic_vfp): Likewise.
25989 * config/arm/arm-cores.def: Add marvell-pj4.
25990 * config/arm/arm-tune.md: Regenerate.
25991 * config/arm/arm-tables.opt: Regenerate.
25992 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25993 * doc/invoke.texi: Document marvell-pj4.
25994
25995 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25996
25997 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25998
25999 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26000
26001 PR debug/54114
26002 PR debug/54402
26003 PR debug/49888
26004 * var-tracking.c (negative_power_of_two_p): New.
26005 (global_get_addr_cache, local_get_addr_cache): New.
26006 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26007 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26008 heading comment.
26009 (vt_stack_offset_p): Remove.
26010 (vt_canon_true_dep): Always canonicalize loc's address.
26011 (clobber_overlapping_mems): Make sure we have a MEM.
26012 (local_get_addr_clear_given_value): New.
26013 (val_reset): Clear local cached entries.
26014 (compute_bb_dataflow): Create and release the local cache.
26015 Disable duplicate MEMs clobbering.
26016 (emit_notes_in_bb): Clobber MEMs likewise.
26017 (vt_emit_notes): Create and release the local cache.
26018 (vt_initialize, vt_finalize): Create and release the global
26019 cache, respectively.
26020 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26021
26022 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26023
26024 PR libmudflap/53359
26025 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26026 not found in the symtab.
26027
26028 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26029
26030 PR debug/56006
26031 PR rtl-optimization/55547
26032 PR rtl-optimization/53827
26033 PR debug/53671
26034 PR debug/49888
26035 * alias.c (offset_overlap_p): New, factored out of...
26036 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26037 the conservative special case for symbolic constants. Don't
26038 adjust zero sizes on alignment.
26039
26040 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26041
26042 PR rtl-optimization/52573
26043 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26044 REG_UNUSED for the same register.
26045
26046 2013-01-17 Richard Biener <rguenther@suse.de>
26047 Marek Polacek <polacek@redhat.com>
26048
26049 PR rtl-optimization/55833
26050 * loop-unswitch.c (unswitch_loops): Move loop verification...
26051 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26052 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26053 Set it to true when we're removing a loop from hierarchy tree in
26054 an irreducible region.
26055 (fix_bb_placements): Adjust caller.
26056 (fix_loop_placements): Likewise.
26057
26058 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26059
26060 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26061 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26062 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26063 Remove ID. Adjust comments.
26064 * config/avr/avr-c.c (avr_builtin_name): Remove.
26065 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26066 * config/avr/avr.c (avr_tolower): New static function.
26067 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26068 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26069 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26070 default expansion.
26071
26072 2013-01-17 Jan Hubicka <jh@suse.cz>
26073
26074 PR tree-optimization/55273
26075 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26076
26077 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26078
26079 PR target/55981
26080 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26081 store through atomic_store<mode>_1.
26082 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26083
26084 2013-01-17 Martin Jambor <mjambor@suse.cz>
26085
26086 PR tree-optimizations/55264
26087 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26088 for virtual methods.
26089 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26090 virtual methods before inlining is over.
26091 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26092 virtual functions.
26093 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26094 non-virtual.
26095
26096 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26097
26098 PR rtl-optimization/56005
26099 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26100 pending reads for prefetch.
26101
26102 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26103
26104 * config/aarch64/aarch64.md
26105 (*cstoresi_neg_uxtw): New pattern.
26106 (*cmovsi_insn_uxtw): New pattern.
26107 (*<optab>si3_uxtw): New pattern.
26108 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26109 (*<optab>si3_insn_uxtw): New pattern.
26110 (*bswapsi2_uxtw): New pattern.
26111
26112 2013-01-16 Richard Biener <rguenther@suse.de>
26113
26114 * tree-inline.c (tree_function_versioning): Remove set but
26115 never used variable.
26116
26117 2013-01-16 Richard Biener <rguenther@suse.de>
26118
26119 PR tree-optimization/55964
26120 * tree-flow.h (rename_variables_in_loop): Remove.
26121 (rename_variables_in_bb): Likewise.
26122 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26123 (copy_loop_before): Adjust and delete update-ssa status.
26124 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26125 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26126 (rename_variables_in_loop): Remove.
26127 (slpeel_update_phis_for_duplicate_loop): Likewise.
26128 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26129 use available cfg machinery instead of duplicating it.
26130 Update PHI nodes and perform poor-mans SSA update here.
26131 (slpeel_tree_peel_loop_to_edge): Adjust.
26132
26133 2013-01-16 Richard Biener <rguenther@suse.de>
26134
26135 PR tree-optimization/54767
26136 PR tree-optimization/53465
26137 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26138 (vrp_visit_phi_node): For PHI arguments coming via backedges
26139 drop all symbolical range information.
26140 (execute_vrp): Compute backedges.
26141
26142 2013-01-16 Richard Biener <rguenther@suse.de>
26143
26144 * doc/install.texi: Update CLooG and ISL requirements to
26145 0.18.0 and 0.11.1.
26146
26147 2013-01-16 Christian Bruel <christian.bruel@st.com>
26148
26149 PR target/55301
26150 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26151 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26152 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26153
26154 2013-01-16 DJ Delorie <dj@redhat.com>
26155
26156 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26157 (UNSPECV_SP_SWITCH_E): New.
26158 (sp_switch_1): Change to an unspec.
26159 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26160 replace $r15.
26161
26162 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26163
26164 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26165 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26166 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26167 (expand_mem_thread_fence): Ditto.
26168 (expand_mem_signal_fence): Ditto.
26169 (expand_atomic_load): Ditto.
26170 (expand_atomic_store): Ditto.
26171
26172 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26173
26174 PR rtl-optimization/55547
26175 PR rtl-optimization/53827
26176 PR debug/53671
26177 PR debug/49888
26178 * alias.c (memrefs_conflict_p): Set sizes to negative after
26179 AND adjustments.
26180
26181 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26182
26183 PR target/55940
26184 * function.c (thread_prologue_and_epilogue_insns): Always
26185 add crtl->drap_reg to set_up_by_prologue.set, even if
26186 stack_realign_drap is false.
26187
26188 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26189
26190 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26191 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26192 *call): Fix indention.
26193
26194 2013-01-15 Tom de Vries <tom@codesourcery.com>
26195
26196 PR target/55876
26197 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26198 Update comment.
26199
26200 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26201
26202 PR rtl-optimization/55153
26203 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26204
26205 2013-01-15 Martin Jambor <mjambor@suse.cz>
26206
26207 PR tree-optimization/55920
26208 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26209 accesses as grp_to_be_debug_replaced.
26210
26211 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26212
26213 PR tree-optimization/55920
26214 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26215 there is non-useless type conversion needed from debug rhs to lhs,
26216 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26217
26218 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26219 Mikael Pettersson <mikpe@it.uu.se>
26220
26221 PR target/43961
26222 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26223 Thumb.
26224 (ASM_OUTPUT_CASE_LABEL): Remove.
26225 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26226 * final.c (shorten_branches): Update alignment of labels before
26227 jump tables if CASE_VECTOR_SHORTEN_MODE.
26228
26229 2013-01-15 Richard Biener <rguenther@suse.de>
26230
26231 PR bootstrap/55961
26232 * system.h: Do not include gmp.h for building host tools.
26233
26234 2013-01-15 Richard Biener <rguenther@suse.de>
26235
26236 PR middle-end/55882
26237 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26238 account for bitpos when computing alignment.
26239
26240 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26241
26242 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26243 (ix86_target_macros_internal): Likewise.
26244
26245 * config/i386/i386.c (m_CORE2I7): Removed.
26246 (m_CORE_HASWELL): New macro.
26247 (m_CORE_ALL): Likewise.
26248 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26249 (initial_ix86_arch_features): Likewise.
26250 (processor_target_table): Initializations for Core avx2.
26251 (cpu_names): New names "core-avx2".
26252 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26253 PROCESSOR_CORE_HASWELL.
26254 (ix86_issue_rate): New case.
26255 (ia32_multipass_dfa_lookahead): Likewise.
26256 (ix86_sched_init_global): Likewise.
26257
26258 * config/i386/i386.h (TARGET_HASWELL): New macro.
26259 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26260 (processor_type): New PROCESSOR_HASWELL.
26261
26262 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26263
26264 PR tree-optimization/55955
26265 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26266 *SHIFT_EXPR and *ROTATE_EXPR codes.
26267
26268 PR tree-optimization/48766
26269 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26270 -ftrapv disable -fwrapv.
26271
26272 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26273
26274 PR target/55974
26275 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26276 etc. to 1 and not to __flash.
26277 Use LL suffix for __INT24_MAX__ with -mint8.
26278 Use ULL suffix for __UINT24_MAX__ with -mint8.
26279
26280 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26281
26282 * config/avr/avr-arch.h
26283 (struct base_arch_s): Use typedef avr_arch_t instead.
26284 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26285 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26286 * config/avr/avr.c: Same.
26287 * config/avr/avr-devices.c: Same.
26288 * config/avr/driver-avr.c: Same.
26289 * config/avr/gen-avr-mmcu-texi.c: Same.
26290 * config/avr/avr-mcus.def: Adjust comment.
26291
26292 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26293
26294 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26295 * config/aarch64/iterators.md (VALLDI): New.
26296
26297 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26298 Andi Kleen <ak@linux.intel.com>
26299
26300 PR target/55948
26301 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26302 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26303 memmodel flag.
26304
26305 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26306
26307 * config/avr/avr-stdint.h: Remove trailing blanks.
26308 * config/avr/avr-log.h: Same.
26309 * config/avr/avr-arch.h: Same.
26310 * config/avr/avr-devices.c: Same.
26311 * config/avr/avr-dimode.md: Same.
26312 * config/avr/predicates.md: Same.
26313 * config/avr/avr-c.c: Same. And fix typo.
26314
26315 * config/avr/avr-protos.h: Same. And:
26316 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26317 (init_cumulative_args): Rename to avr_init_cumulative_args.
26318 (expand_prologue): Rename to avr_expand_prologue.
26319 (expand_epilogue): Rename to avr_expand_epilogue.
26320 (adjust_insn_length): Rename to avr_adjust_insn_length.
26321 (notice_update_cc): Rename to avr_notice_update_cc.
26322 (final_prescan_insn): Rename to avr_final_prescan_insn.
26323 * config/avr/avr.c: Same.
26324 * config/avr/avr.h: Same.
26325 * config/avr/avr.md: Remove trailing blanks.
26326 (prologue): Use avr_expand_prologue.
26327 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26328
26329 2013-01-14 Richard Biener <rguenther@suse.de>
26330
26331 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26332 verify_location, collect_subblocks): New functions.
26333 (verify_gimple_in_cfg): Verify that locations only reference
26334 BLOCKs in the functions BLOCK tree.
26335
26336 2013-01-14 Richard Biener <rguenther@suse.de>
26337
26338 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26339 PHI argument.
26340 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26341 unshare reference.
26342 (insert_out_of_ssa_copy_on_edge): Likewise.
26343 (rewrite_close_phi_out_of_ssa): Likewise.
26344 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26345 debug expressions.
26346 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26347 propagated constants.
26348 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26349 can not be shared.
26350
26351 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26352
26353 * config/avr/avr-modes.def: Add GPL copyright notice.
26354
26355 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26356
26357 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26358 MEMMODEL_MASK to determine memory model.
26359 (atomic_store<mode>): Ditto from operands[2].
26360 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26361
26362 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26363
26364 PR fortran/55935
26365 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26366 (fold_gimple_assign): Don't call unshare_expr here.
26367 (fold_ctor_reference): Call unshare_expr.
26368
26369 2013-01-13 Terry Guo <terry.guo@arm.com>
26370
26371 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26372 * doc/fragments.texi: Document MULTILIB_REUSE.
26373 * gcc.c (multilib_reuse): New internal spec.
26374 (set_multilib_dir): Also search multilib from multilib_reuse.
26375 * genmultilib (tmpmultilib3): Refactor code.
26376 (tmpmultilib4): Ditto.
26377 (multilib_reuse): New multilib argument.
26378
26379 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26380
26381 * Makefile.in: Update copyright.
26382
26383 2013-01-12 Tom de Vries <tom@codesourcery.com>
26384
26385 PR middle-end/55890
26386 * calls.c (expand_call): Check if arg_nr is valid.
26387
26388 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26389
26390 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26391 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26392 documentation. Add missing '__' in front of
26393 __builtin_ia32_packssdw256.
26394
26395 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26396
26397 PR target/55719
26398 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26399 NO_REGS for larl operands.
26400 (s390_reload_larl_operand): Use s390_load_address instead of
26401 emit_move_insn.
26402
26403 2013-01-11 Richard Biener <rguenther@suse.de>
26404
26405 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26406 (verify_node_sharing): ... here.
26407 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26408
26409 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26410
26411 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26412 Substitute TREECHECKING.
26413 * configure: Regenerate.
26414 * Makefile.in (TREECHECKING): New.
26415
26416 2013-01-11 Richard Guenther <rguenther@suse.de>
26417
26418 PR tree-optimization/44061
26419 * tree-vrp.c (extract_range_basic): Compute zero as
26420 value-range for __builtin_constant_p of function parameters.
26421
26422 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26423
26424 Update copyright years.
26425
26426 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26427
26428 PR rtl-optimization/55672
26429 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26430 const to be eliminable.
26431
26432 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26433
26434 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26435 * configure: Regenerate.
26436
26437 2013-01-10 Richard Biener <rguenther@suse.de>
26438
26439 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26440
26441 2013-01-10 Richard Biener <rguenther@suse.de>
26442
26443 PR bootstrap/55792
26444 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26445 locations for virtual PHI arguments.
26446 (rewrite_update_phi_arguments): Likewise.
26447
26448 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26449
26450 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26451 on to assembler.
26452
26453 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26454
26455 PR tree-optimization/55921
26456 * tree-complex.c (expand_complex_asm): New function.
26457 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26458
26459 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26460
26461 PR target/55718
26462 * config/s390/s390.c (s390_symref_operand_p)
26463 (s390_loadrelative_operand_p): Merge the two functions.
26464 (s390_check_qrst_address, print_operand_address): Add parameters
26465 to s390_loadrelative_operand_p invokation.
26466 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26467 (s390_reload_larl_operand, s390_secondary_reload): Use
26468 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26469 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26470
26471 2013-01-09 Mike Stump <mikestump@comcast.net>
26472
26473 * dse.c (record_store): Remove unnecessary assert.
26474
26475 2013-01-09 Jan Hubicka <jh@suse.cz>
26476
26477 PR tree-optimization/55569
26478 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26479 * cfgloop.h (scale_loop_profile): Likewise.
26480
26481 2013-01-09 Jan Hubicka <jh@suse.cz>
26482
26483 PR lto/45375
26484 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26485 functions.
26486 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26487
26488 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26489
26490 PR middle-end/55114
26491 * expr.h (maybe_emit_group_store): Declare.
26492 * expr.c (maybe_emit_group_store): New function.
26493 * builtins.c (expand_builtin_int_roundingfn): Call it.
26494 (expand_builtin_int_roundingfn_2): Likewise.
26495
26496 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26497
26498 PR rtl-optimization/55829
26499 * lra-constraints.c (match_reload): Add code for absent output.
26500 (curr_insn_transform): Add code for reloads of matched inputs
26501 without output.
26502
26503 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26504
26505 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26506 attribute of movddup insn to DF.
26507 (*vec_interleave_lowv2df): Ditto.
26508 (vec_dupv2df): Ditto.
26509
26510 2013-01-09 Jan Hubicka <jh@suse.cz>
26511
26512 PR tree-optimiation/55875
26513 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26514 EVERY_ITERATION parameter.
26515 (number_of_iterations_exit): Check if exit is executed every iteration.
26516 (idx_infer_loop_bounds): Similarly here.
26517 (n_of_executions_at_most): Simplify
26518 to only test for cases where statement is dominated by the
26519 particular bound; handle correctly the "postdominance" test.
26520 (scev_probably_wraps_p): Use max loop iterations info
26521 as a global bound first.
26522
26523 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26524 Nick Clifton <nickc@redhat.com>
26525
26526 * config/v850/v850.md (cbranchsf4): New pattern.
26527 (cstoresf4): New pattern.
26528 (cbranchdf4): New pattern.
26529 (cstoredf4): New pattern.
26530 (movsicc): Disallow floating point comparisons.
26531 (cmpsf_le_insn): Fix order of operators.
26532 (cmpsf_lt_insn): Likewise.
26533 (cmpsf_eq_insn): Likewise.
26534 (cmpdf_le_insn): Likewise.
26535 (cmpdf_lt_insn): Likewise.
26536 (cmpdf_eq_insn): Likewise.
26537 (cmpsf_ge_insn): Use LE comparison.
26538 (cmpdf_ge_insn): Likewise.
26539 (cmpsf_gt_insn): Use LT comparison.
26540 (cmpdf_gt_insn): Likewise.
26541 (cmpsf_ne_insn): Delete pattern.
26542 (cmpdf_ne_insn): Delete pattern.
26543 * config/v850/v850.c (v850_gen_float_compare): Use
26544 gen_cmpdf_eq_insn for NE comparison.
26545 (v850_float_z_comparison_operator)
26546 (v850_float_nz_comparison_operator): Move from here ...
26547 * config/v850/predicates.md: ... to here. Move GT and GE
26548 comparisons into v850_float_z_comparison_operator.
26549 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26550 Delete prototype.
26551 (v850_float_nz_comparison_operator): Likewise.
26552
26553 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26554
26555 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26556 with calls to gen_insvsi/gen_insvdi.
26557
26558 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26559
26560 * config/i386/i386.c (initial_ix86_tune_features): Set up
26561 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26562
26563 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26564 Jakub Jelinek <jakub@redhat.com>
26565
26566 PR tree-optimization/48189
26567 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26568 If nitercst is 0, don't predict the exit edge.
26569
26570 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26571
26572 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26573 in asm_fprintf with reg_names.
26574 (aarch64_print_operand_address): Likewise.
26575 (aarch64_return_addr): Likewise.
26576 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26577
26578 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26579
26580 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26581 (INT_U6_BITS): Likewise.
26582 * config/pa/predicates.md (uint6_operand): New predicate.
26583 (shift5_operand, shift6_operand): Likewise.
26584 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26585 arith32_operand.
26586 (lshrdi3): Use shift6_operand.
26587 (shrpsi4, shrpdi4): New insn patterns.
26588 (extzv): Delete expander.
26589 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26590 predicates in unamed zero extract patterns. Tighten common constraint.
26591 (extv): Delete expander.
26592 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26593 predicates in unamed sign extract patterns. Tighten common constraint.
26594 (insv): Delete expander.
26595 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26596 predicates in unamed insert patterns. Tighten common constraint.
26597 Change uint32_operand predicate to uint6_operand predicate in unamed
26598 DImode pattern to insert constant values of type 1...1xxxx.
26599
26600 2013-01-04 Jan Hubicka <jh@suse.cz>
26601
26602 PR tree-optimization/55823
26603 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26604 issue.
26605
26606 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26607 Uros Bizjak <ubizjak@gmail.com>
26608
26609 PR rtl-optimization/55845
26610 * df-problems.c (can_move_insns_across): Stop scanning at
26611 volatile_insn_p source instruction or give up if
26612 across_from .. across_to range contains any volatile_insn_p
26613 instructions.
26614
26615 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26616
26617 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26618 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26619 Declare.
26620 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26621 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26622
26623 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26624
26625 PR fortran/55341
26626 * asan.c (asan_clear_shadow): New function.
26627 (asan_emit_stack_protection): Use it.
26628
26629 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26630
26631 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26632 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26633 with tab instead of space.
26634
26635 2013-01-08 Nick Clifton <nickc@redhat.com>
26636
26637 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26638 register bank 0 at the start of an interrupt handler.
26639 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26640 MDBH registers.
26641
26642 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26643
26644 * config/aarch64/aarch64-simd.md
26645 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26646 (aarch64_simd_bsl): Likewise.
26647 (aarch64_vcond_internal<mode>): Likewise.
26648 (vcond<mode><mode>): Likewise.
26649 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26650 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26651
26652 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26653
26654 * config/aarch64/aarch64-builtins.c
26655 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26656
26657 2013-01-08 Martin Jambor <mjambor@suse.cz>
26658
26659 PR debug/55579
26660 * tree-sra.c (analyze_access_subtree): Return true also after
26661 potentially creating a debug-only replacement.
26662
26663 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26664
26665 PR middle-end/55890
26666 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26667
26668 PR tree-optimization/54120
26669 * tree-vrp.c (range_fits_type_p): Don't allow
26670 src_precision < precision from signed vr to unsigned_p
26671 if vr->min or vr->max is negative.
26672 (simplify_float_conversion_using_ranges): Test can_float_p
26673 against CODE_FOR_nothing.
26674
26675 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26676 Richard Biener <rguenther@suse.de>
26677
26678 PR middle-end/55851
26679 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26680 types instead of just INTEGER_TYPE types.
26681
26682 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26683
26684 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26685 TF_SIZE): Define.
26686
26687 2013-01-07 Steve Ellcey <sellcey@mips.com>
26688
26689 PR target/42661
26690 * config/mips/mips.opt: Change mad to mmad to match documentation.
26691
26692 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26693
26694 PR target/55897
26695 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26696 .progmemx.data now.
26697
26698 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26699
26700 PR target/55897
26701 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26702 (avr_addrspace_t): Add .section_name field.
26703 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26704 array size.
26705 (avr_addrspace): Same. Initialize .section_name. Remove last
26706 NULL entry. Put __memx into .progmemx.data.
26707 (progmem_section_prefix): Remove.
26708 (avr_asm_init_sections): No need to initialize progmem_section.
26709 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26710 section name prefix.
26711 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26712 retrieve the progmem section.
26713 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26714 boundary to run over avr_addrspace[].
26715 (avr_register_target_pragmas): Ditto.
26716
26717 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26718
26719 * varasm.c (output_constant_def_contents): For asan_protect_global
26720 protected strings, adjust DECL_ALIGN if needed, before testing for
26721 anchored symbols.
26722 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26723 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26724 normal decls.
26725 (output_object_block): For asan protected decls, emit asan padding
26726 after their contents.
26727 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26728 (asan_finish_file): Test it here instead.
26729
26730 2013-01-07 Nick Clifton <nickc@redhat.com>
26731 Matthias Klose <doko@debian.org>
26732 Doug Kwan <dougkwan@google.com>
26733 H.J. Lu <hongjiu.lu@intel.com>
26734
26735 PR driver/55470
26736 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26737
26738 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26739
26740 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26741
26742 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26743
26744 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26745
26746 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26747
26748 PR target/54461
26749 * doc/install.texi (Cross-Compiler-Specific Options): Document
26750 --with-avrlibc.
26751
26752 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26753
26754 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26755 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26756 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26757 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26758 vqmovun_high_s64): Fix source operand number and update copyright.
26759
26760 2013-01-07 Richard Biener <rguenther@suse.de>
26761
26762 PR middle-end/55890
26763 * gimple.h (gimple_call_builtin_p): New overload.
26764 * gimple.c (validate_call): New function.
26765 (gimple_call_builtin_p): Likewise.
26766 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26767 Use gimple_call_builtin_p.
26768 (find_func_clobbers): Likewise.
26769 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26770 (strlen_optimize_stmt): Likewise.
26771
26772 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26773
26774 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26775 (vld1q_dup_*): Likewise.
26776 (vld1_*): Likewise.
26777 (vld1q_*): Likewise.
26778 (vld1_lane_*): Likewise.
26779 (vld1q_lane_*): Likewise.
26780
26781 2013-01-07 Richard Biener <rguenther@suse.de>
26782
26783 * lto-streamer.h (LTO_minor_version): Bump to 2.
26784
26785 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26786
26787 * config/aarch64/aarch64-protos.h
26788 (aarch64_const_double_zero_rtx_p): Rename to...
26789 (aarch64_float_const_zero_rtx_p): ...this.
26790 (aarch64_float_const_representable_p): New.
26791 (aarch64_output_simd_mov_immediate): Likewise.
26792 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26793 move immediate case.
26794 * config/aarch64/aarch64.c
26795 (aarch64_const_double_zero_rtx_p): Rename to...
26796 (aarch64_float_const_zero_rtx_p): ...this.
26797 (aarch64_print_operand): Allow printing of new constants.
26798 (aarch64_valid_floating_const): New.
26799 (aarch64_legitimate_constant_p): Check for valid floating-point
26800 constants.
26801 (aarch64_simd_valid_immediate): Likewise.
26802 (aarch64_vect_float_const_representable_p): New.
26803 (aarch64_float_const_representable_p): Likewise.
26804 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26805 (aarch64_output_simd_mov_immediate): New.
26806 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26807 (*movdf_aarch64): Likewise.
26808 * config/aarch64/constraints.md (Ufc): New.
26809 (Y): call aarch64_float_const_zero_rtx.
26810 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26811
26812 2013-01-07 Richard Biener <rguenther@suse.de>
26813
26814 PR tree-optimization/55888
26815 PR tree-optimization/55862
26816 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26817 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26818 not if it is contained therein.
26819
26820 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26821
26822 * config/avr/t-avr: Typo.
26823
26824 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26825
26826 PR55243
26827 * config/avr/t-avr: Don't automatically rebuild
26828 $(srcdir)/config/avr/t-multilib
26829 $(srcdir)/config/avr/avr-tables.opt
26830 $(srcdir)/doc/avr-mmcu.texi
26831 (avr-mcus): New phony target to build them on request.
26832 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26833 * avr/avr-mcus.def: Adjust comments.
26834
26835 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26836
26837 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26838
26839 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26840
26841 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26842
26843 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26844
26845 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26846
26847 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26848
26849 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26850 to generate profiling.
26851 * config/rs6000/aix64.h (LIB_SPEC): Same.
26852
26853 2013-01-04 Andrew Pinski <apinski@cavium.com>
26854
26855 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26856 New function.
26857 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26858
26859 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26860
26861 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26862 unconditionally.
26863 (ix86_expand_move): Ditto.
26864 (ix86_zero_extend_to_Pmode): Ditto.
26865 (ix86_expand_call): Ditto.
26866 (ix86_expand_special_args_builtin): Ditto.
26867 (ix86_expand_builtin): Ditto.
26868
26869 2013-01-04 Richard Biener <rguenther@suse.de>
26870
26871 PR tree-optimization/55862
26872 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26873 translating them through PHI nodes.
26874
26875 2013-01-04 Martin Jambor <mjambor@suse.cz>
26876
26877 PR tree-optimization/55755
26878 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26879 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26880
26881 2013-01-04 Marek Polacek <polacek@redhat.com>
26882
26883 PR middle-end/55859
26884 * opts.c (default_options_optimization): Clarify error message.
26885
26886 2013-01-04 Richard Biener <rguenther@suse.de>
26887
26888 PR middle-end/55863
26889 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26890 reassociation.
26891
26892 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26893
26894 PR target/53789
26895 * config/pa/pa.md (movsi): Revert previous change.
26896 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26897 references.
26898
26899 2013-01-03 Richard Henderson <rth@redhat.com>
26900
26901 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26902 after eliminating TLS symbols.
26903
26904 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26905
26906 PR bootstrap/50167
26907 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26908 * graphite-poly.c (debug_gmp_value): Likewise.
26909
26910 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26911
26912 PR target/55712
26913 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26914 selected code model, define __code_mode_small__, __code_model_medium__,
26915 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26916 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26917 xchg temporary register with %k. Declare temporary register as
26918 early clobbered.
26919 [__x86_64__]: For medium and large code models, preserve %rbx register.
26920
26921 2013-01-03 Richard Biener <rguenther@suse.de>
26922
26923 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26924 (dump_subscript): Adjust.
26925 (finalize_ddr_dependent): Do not dump redundant info.
26926 (analyze_siv_subscript): Adjust.
26927 (subscript_dependence_tester): Likewise.
26928 (compute_affine_dependence): Likewise.
26929
26930 2013-01-03 Richard Biener <rguenther@suse.de>
26931
26932 Revert
26933 2013-01-03 Richard Biener <rguenther@suse.de>
26934
26935 PR tree-optimization/55857
26936 * tree-vect-stmts.c (vectorizable_load): Do not setup
26937 re-alignment for invariant loads.
26938
26939 2013-01-02 Richard Biener <rguenther@suse.de>
26940
26941 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26942 invariant load do not generate a vector load from the scalar location.
26943
26944 2013-01-03 Richard Biener <rguenther@suse.de>
26945
26946 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26947 for not vectorizing.
26948 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26949 not build INDIRECT_REFs, call get_name once only.
26950 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26951 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26952
26953 2013-01-03 Richard Biener <rguenther@suse.de>
26954
26955 PR tree-optimization/55857
26956 * tree-vect-stmts.c (vectorizable_load): Do not setup
26957 re-alignment for invariant loads.
26958
26959 2013-01-03 Richard Biener <rguenther@suse.de>
26960
26961 PR lto/55848
26962 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26963 prefer a built-in decl.
26964
26965 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26966
26967 * gcc.c (process_command): Update copyright notice dates.
26968 * gcov.c (print_version): Likewise.
26969 * gcov-dump.c (print_version): Likewise.
26970
26971 PR rtl-optimization/55838
26972 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26973 iv0.step, iv1.step and step.
26974
26975 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26976 Marc Glisse <marc.glisse@inria.fr>
26977
26978 PR tree-optimization/55832
26979 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26980 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26981 integer_{one,zero}_node.
26982
26983 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26984
26985 PR debug/54402
26986 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26987 * var-tracking.c (reverse_op): Don't add reverse ops to
26988 VALUEs that have already
26989 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26990
26991 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26992
26993 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26994
26995 2013-01-02 Teresa Johnson <tejohnson@google.com>
26996
26997 * dumpfile.c (dump_loc): Print filename with location.
26998 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26999 new location_t parameter to emit complete unroll message with
27000 new dump framework.
27001 (canonicalize_loop_induction_variables): Compute loops location
27002 and pass to try_unroll_loop_completely.
27003 * loop-unroll.c (report_unroll_peel): New function.
27004 (peel_loops_completely): Use new dump format with location
27005 for main dumpfile message, and invoke report_unroll_peel on success.
27006 (decide_unrolling_and_peeling): Ditto.
27007 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27008 by report_unroll_peel.
27009 (decide_peel_completely): Ditto.
27010 (decide_unroll_constant_iterations): Ditto.
27011 (decide_unroll_runtime_iterations): Ditto.
27012 (decide_peel_simple): Ditto.
27013 (decide_unroll_stupid): Ditto.
27014 * cfgloop.c (get_loop_location): New function.
27015 * cfgloop.h (get_loop_location): Declare.
27016
27017 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27018
27019 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27020 NULL.
27021
27022 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27023
27024 PR middle-end/55198
27025 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27026 BLKmode objects when EXPAND_MEMORY is specified.
27027
27028 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27029
27030 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27031 in loop predicate.
27032 (fold_builtin_cpu): Do not share cpu model decls across statements.
27033
27034 2013-01-02 Jason Merrill <jason@redhat.com>
27035
27036 PR c++/55804
27037 * tree.c (build_array_type_1): Revert earlier change.
27038
27039 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27040
27041 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27042 "cortex-a57".
27043 * config/aarch64/aarch64-tune.md: Re-generate.
27044
27045 2013-01-02 Richard Biener <rguenther@suse.de>
27046
27047 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27048 invariant load do not generate a vector load from the scalar location.
27049
27050 2013-01-02 Richard Biener <rguenther@suse.de>
27051
27052 PR bootstrap/55784
27053 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27054 * configure: Regenerate.
27055
27056 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27057
27058 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27059 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27060 (expand_builtin_int_roundingfn_2): Keep the original target around
27061 for the fallback case.
27062
27063 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27064
27065 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27066 to be clear for sign changes.
27067
27068 2013-01-01 Jan Hubicka <jh@suse.cz>
27069
27070 * ipa-inline-analysis.c: Fix formatting.
27071
27072 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27073
27074 PR tree-optimization/55831
27075 * tree-vect-loop.c (get_initial_def_for_induction): Use
27076 gsi_after_labels instead of gsi_start_bb.
27077 \f
27078 Copyright (C) 2013 Free Software Foundation, Inc.
27079
27080 Copying and distribution of this file, with or without modification,
27081 are permitted in any medium without royalty provided the copyright
27082 notice and this notice are preserved.
This page took 1.340899 seconds and 5 git commands to generate.