]> gcc.gnu.org Git - gcc.git/blob - gcc/ChangeLog
re PR target/60179 (target optimization attribute streaming is broken)
[gcc.git] / gcc / ChangeLog
1 2014-02-14 Richard Biener <rguenther@suse.de>
2
3 PR lto/60179
4 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5 DECL_FUNCTION_SPECIFIC_TARGET.
6 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
7 * tree-streamer-out.c (pack_ts_target_option): Remove.
8 (streamer_pack_tree_bitfields): Do not stream
9 TS_TARGET_OPTION.
10 (write_ts_function_decl_tree_pointers): Do not stream
11 DECL_FUNCTION_SPECIFIC_TARGET.
12 * tree-streamer-in.c (unpack_ts_target_option): Remove.
13 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
14 (lto_input_ts_function_decl_tree_pointers): Do not stream
15 DECL_FUNCTION_SPECIFIC_TARGET.
16
17 2014-02-14 Jakub Jelinek <jakub@redhat.com>
18
19 * tree-vect-loop.c (vect_is_slp_reduction): Don't set
20 use_stmt twice.
21 (get_initial_def_for_induction, vectorizable_induction): Ignore
22 debug stmts when looking for exit_phi.
23 (vectorizable_live_operation): Fix up condition.
24
25 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
26
27 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
28 nreverse() because it changes the content of original tree list.
29
30 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
31
32 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
33 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
34
35 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
36
37 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
38 GNU coding standards.
39
40 2014-02-13 Jakub Jelinek <jakub@redhat.com>
41
42 PR debug/60152
43 * dwarf2out.c (gen_subprogram_die): Don't call
44 add_calling_convention_attribute if subr_die is old_die.
45
46 2014-02-13 Sharad Singhai <singhai@google.com>
47
48 * doc/optinfo.texi: Fix order of nodes.
49
50 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
51
52 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
53 operands[2], not operands[3].
54
55 2014-02-13 Richard Biener <rguenther@suse.de>
56
57 PR bootstrap/59878
58 * doc/install.texi (ISL): Update recommended version to 0.12.2,
59 mention the possibility of an in-tree build.
60 (CLooG): Update recommended version to 0.18.1, mention the
61 possibility of an in-tree build and clarify that the ISL
62 bundled with CLooG does not work.
63
64 2014-02-13 Jakub Jelinek <jakub@redhat.com>
65
66 PR target/43546
67 * expr.c (compress_float_constant): If x is a hard register,
68 extend into a pseudo and then move to x.
69
70 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
71
72 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
73 caused by bad second argument to warning_at() with -mhotpatch and
74 nested functions (e.g. with gfortran).
75
76 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
77
78 * opts.c (option_name): Remove "enabled by default" rider.
79
80 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
81
82 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
83
84 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
85 Uros Bizjak <ubizjak@gmail.com>
86
87 PR target/60151
88 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
89 * configure: Regenerated.
90
91 2014-02-12 Richard Biener <rguenther@suse.de>
92
93 * vec.c (vec_prefix::calculate_allocation): Move as
94 inline variant to vec.h.
95 (vec_prefix::calculate_allocation_1): New out-of-line version.
96 * vec.h (vec_prefix::calculate_allocation_1): Declare.
97 (vec_prefix::m_has_auto_buf): Rename to ...
98 (vec_prefix::m_using_auto_storage): ... this.
99 (vec_prefix::calculate_allocation): Inline the easy cases
100 and dispatch to calculate_allocation_1 which doesn't need the
101 prefix address.
102 (va_heap::reserve): Use gcc_checking_assert.
103 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
104 m_using_auto_storage.
105 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
106 member and adjust.
107 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
108 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
109 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
110
111 2014-02-12 Richard Biener <rguenther@suse.de>
112
113 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
114 when we found a dependence.
115
116 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
117
118 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
119 common code...
120 (maybe_fold_stmt): ... into this new function.
121 * omp-low.c (lower_omp): Update comment.
122
123 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
124 last use.
125
126 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
127 dereference.
128
129 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
130
131 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
132 identifiers in comments.
133 (cortexa53_extra_costs): Likewise.
134 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
135 (cortexa7_extra_costs): Likewise.
136 (cortexa12_extra_costs): Likewise.
137 (cortexa15_extra_costs): Likewise.
138 (v7m_extra_costs): Likewise.
139
140 2014-02-12 Richard Biener <rguenther@suse.de>
141
142 PR middle-end/60092
143 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
144 of posix_memalign being successful.
145 (lower_stmt): Restrict lowering of posix_memalign to when
146 -ftree-bit-ccp is enabled.
147
148 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
149
150 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
151 arg_loc.
152 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
153
154 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
155
156 PR rtl-optimization/60116
157 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
158 other_insn once the combination has been validated.
159
160 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
161
162 PR lto/59468
163 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
164 and wrapper.
165 * ipa-devirt.c: Include demangle.h
166 (odr_violation_reported): New static variable.
167 (add_type_duplicate): Update odr_violations.
168 (maybe_record_node): Add completep parameter; update it.
169 (record_target_from_binfo): Add COMPLETEP parameter;
170 update it as needed.
171 (possible_polymorphic_call_targets_1): Likewise.
172 (struct polymorphic_call_target_d): Add nonconstruction_targets;
173 rename FINAL to COMPLETE.
174 (record_targets_from_bases): Sanity check we found the binfo;
175 fix COMPLETEP updating.
176 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
177 parameter, fix computing of COMPLETEP.
178 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
179 at LTO time do demangling.
180 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
181 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
182 parameter.
183 (gimple_get_virt_method_for_binfo): Likewise.
184 * gimple-fold.h (gimple_get_virt_method_for_binfo,
185 gimple_get_virt_method_for_vtable): Update prototypes.
186
187 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
188
189 PR target/49008
190 * genautomata.c (add_presence_absence): Fix typo with
191 {final_}presence_list.
192
193 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
194
195 PR target/60137
196 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
197 for VSX/Altivec vectors that land in GPR registers.
198
199 2014-02-11 Richard Henderson <rth@redhat.com>
200 Jakub Jelinek <jakub@redhat.com>
201
202 PR debug/59776
203 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
204 around drhs if type conversion to lacc->type is not useless.
205
206 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
207
208 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
209 tuning struct.
210 (cortex-a57.cortex-a53): Likewise.
211 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
212
213 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
214
215 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
216 arm_restrict_it.
217
218 2014-02-11 Renlin Li <Renlin.Li@arm.com>
219
220 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
221 add_options_for_arm_vfp3.
222
223 2014-02-11 Jeff Law <law@redhat.com>
224
225 PR middle-end/54041
226 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
227 object with an undesirable mode.
228
229 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
230
231 PR libgomp/60107
232 * config/i386/sol2-9.h: New file.
233 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
234 *-*-solaris2.9*): Use it.
235
236 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
237
238 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
239 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
240
241 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
242
243 * config/microblaze/microblaze.c: Extend mcpu version format
244
245 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
246
247 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
248
249 2014-02-10 Richard Henderson <rth@redhat.com>
250
251 PR target/59927
252 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
253 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
254 ms-abi vs -mno-accumulate-outgoing-args.
255 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
256 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
257 respect to ms-abi.
258
259 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
260
261 PR middle-end/60080
262 * cfgexpand.c (expand_asm_operands): Attach source location to
263 ASM_INPUT rtx objects.
264 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
265
266 2014-02-10 Nick Clifton <nickc@redhat.com>
267
268 * config/mn10300/mn10300.c (popcount): New function.
269 (mn10300_expand_prologue): Include saved registers in stack usage
270 count.
271
272 2014-02-10 Jeff Law <law@redhat.com>
273
274 PR middle-end/52306
275 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
276 when changing the SET_DEST of a prior insn to avoid an input reload.
277
278 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
279
280 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
281 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
282 -mcall-openbsd, or -mcall-linux.
283 (CC1_ENDIAN_BIG_SPEC): Remove.
284 (CC1_ENDIAN_LITTLE_SPEC): Remove.
285 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
286 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
287 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
288 and %cc1_endian_default.
289 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
290
291 2014-02-10 Richard Biener <rguenther@suse.de>
292
293 PR tree-optimization/60115
294 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
295 MEM_REF handling. Properly verify that the accesses are not
296 out of the objects bound.
297
298 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
299
300 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
301 coretex to cortex.
302
303 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
304
305 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
306 proper constants and fix formatting.
307 (possible_polymorphic_call_targets): Fix formatting.
308
309 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
310 Ilya Tocar <ilya.tocar@intel.com>
311
312 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
313 (_mm512_loadu_epi32): Renamed into...
314 (_mm512_loadu_si512): This.
315 (_mm512_storeu_epi32): Renamed into...
316 (_mm512_storeu_si512): This.
317 (_mm512_maskz_ceil_ps): Removed.
318 (_mm512_maskz_ceil_pd): Ditto.
319 (_mm512_maskz_floor_ps): Ditto.
320 (_mm512_maskz_floor_pd): Ditto.
321 (_mm512_floor_round_ps): Ditto.
322 (_mm512_floor_round_pd): Ditto.
323 (_mm512_ceil_round_ps): Ditto.
324 (_mm512_ceil_round_pd): Ditto.
325 (_mm512_mask_floor_round_ps): Ditto.
326 (_mm512_mask_floor_round_pd): Ditto.
327 (_mm512_mask_ceil_round_ps): Ditto.
328 (_mm512_mask_ceil_round_pd): Ditto.
329 (_mm512_maskz_floor_round_ps): Ditto.
330 (_mm512_maskz_floor_round_pd): Ditto.
331 (_mm512_maskz_ceil_round_ps): Ditto.
332 (_mm512_maskz_ceil_round_pd): Ditto.
333 (_mm512_expand_pd): Ditto.
334 (_mm512_expand_ps): Ditto.
335 * config/i386/i386.c (ix86_builtins): Remove
336 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
337 (bdesc_args): Ditto.
338 * config/i386/predicates.md (const1256_operand): New.
339 (const_1_to_2_operand): Ditto.
340 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
341 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
342 (*avx512pf_gatherpf<mode>sf): Ditto.
343 (avx512pf_gatherpf<mode>df): Ditto.
344 (*avx512pf_gatherpf<mode>df_mask): Ditto.
345 (*avx512pf_gatherpf<mode>df): Ditto.
346 (avx512pf_scatterpf<mode>sf): Ditto.
347 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
348 (*avx512pf_scatterpf<mode>sf): Ditto.
349 (avx512pf_scatterpf<mode>df): Ditto.
350 (*avx512pf_scatterpf<mode>df_mask): Ditto.
351 (*avx512pf_scatterpf<mode>df): Ditto.
352 (avx512f_expand<mode>): Removed.
353 (<shift_insn><mode>3<mask_name>): Change predicate type.
354
355 2014-02-08 Jakub Jelinek <jakub@redhat.com>
356
357 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
358 not at the end of datarefs vector use ordered_remove to avoid
359 reordering datarefs vector.
360
361 PR c/59984
362 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
363 mark local addressable non-static vars as GOVD_PRIVATE
364 instead of GOVD_LOCAL.
365 * omp-low.c (lower_omp_for): Move gimple_bind_vars
366 and BLOCK_VARS of gimple_bind_block to new_stmt rather
367 than copying them.
368
369 PR middle-end/60092
370 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
371 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
372 assume_aligned or alloc_align attributes.
373 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
374 arguments. Handle also assume_aligned and alloc_align attributes.
375 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
376 calls to functions with assume_aligned or alloc_align attributes.
377 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
378
379 2014-02-08 Terry Guo <terry.guo@arm.com>
380
381 * doc/invoke.texi: Document ARM -march=armv7e-m.
382
383 2014-02-08 Jakub Jelinek <jakub@redhat.com>
384
385 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
386 flag on __cilkrts_rethrow builtin.
387
388 PR ipa/60026
389 * ipa-cp.c (determine_versionability): Fail at -O0
390 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
391 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
392
393 Revert:
394 2014-02-04 Jakub Jelinek <jakub@redhat.com>
395
396 PR ipa/60026
397 * tree-inline.c (copy_forbidden): Fail for
398 __attribute__((optimize (0))) functions.
399
400 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
401
402 * varpool.c: Include pointer-set.h.
403 (varpool_remove_unreferenced_decls): Variables in other partitions
404 will not be output; be however careful to not lose information
405 about partitioning.
406
407 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
408
409 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
410 lookup in the vtable constructor.
411
412 2014-02-07 Jeff Law <law@redhat.com>
413
414 PR target/40977
415 * config/m68k/m68k.md (ashldi_extsi): Turn into a
416 define_insn_and_split.
417
418 * ipa-inline.c (inline_small_functions): Fix typos.
419
420 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
421
422 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
423 (s390_can_use_return_insn): Declare.
424 * config/s390/s390.h (EPILOGUE_USES): Define.
425 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
426 instructions.
427 (s390_chunkify_start): Handle return JUMP_LABELs.
428 (s390_early_mach): Emit a main_pool instruction on the entry edge.
429 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
430 (s390_can_use_return_insn): New functions.
431 (s390_fix_long_loop_prediction): Handle conditional returns.
432 (TARGET_SET_UP_BY_PROLOGUE): Define.
433 * config/s390/s390.md (ANY_RETURN): New code iterator.
434 (*creturn, *csimple_return, return, simple_return): New patterns.
435
436 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
437
438 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
439 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
440 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
441 REG_CFA_RESTORE list when deciding not to restore a register.
442
443 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
444
445 * config/s390/s390.c: Include tree-pass.h and context.h.
446 (s390_early_mach): New function, split out from...
447 (s390_emit_prologue): ...here.
448 (pass_data_s390_early_mach): New pass structure.
449 (pass_s390_early_mach): New class.
450 (s390_option_override): Create and register early_mach pass.
451 Move to end of file.
452
453 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
454
455 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
456 to match for the exit block.
457
458 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
459
460 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
461 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
462 Reject misaligned operands.
463
464 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
465
466 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
467
468 2014-02-07 Richard Biener <rguenther@suse.de>
469
470 PR middle-end/60092
471 * gimple-low.c (lower_builtin_posix_memalign): New function.
472 (lower_stmt): Call it to lower posix_memalign in a way
473 to make alignment info accessible.
474
475 2014-02-07 Jakub Jelinek <jakub@redhat.com>
476
477 PR c++/60082
478 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
479 __builtin_setjmp_receiver.
480
481 2014-02-07 Richard Biener <rguenther@suse.de>
482
483 PR middle-end/60092
484 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
485 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
486 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
487 Handle BUILT_IN_POSIX_MEMALIGN.
488 (find_func_clobbers): Likewise.
489 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
490 (call_may_clobber_ref_p_1): Likewise.
491
492 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
493
494 PR ipa/59918
495 * ipa-devirt.c (record_target_from_binfo): Remove overactive
496 sanity check.
497
498 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
499
500 PR ipa/59469
501 * lto-cgraph.c (lto_output_node): Use
502 symtab_get_symbol_partitioning_class.
503 (lto_output_varpool_node): likewise.
504 (symtab_get_symbol_partitioning_class): Move here from
505 lto/lto-partition.c
506 * cgraph.h (symbol_partitioning_class): Likewise.
507 (symtab_get_symbol_partitioning_class): Declare.
508
509 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
510
511 * ggc.h (ggc_internal_cleared_alloc): New macro.
512 * vec.h (vec_safe_copy): Handle memory stats.
513 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
514 * target-globals.c (save_target_globals): Likewise.
515
516 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
517
518 PR target/60077
519 * expr.c (emit_move_resolve_push): Export; be bit more selective
520 on when to clear alias set.
521 * expr.h (emit_move_resolve_push): Declare.
522 * function.h (struct function): Add tail_call_marked.
523 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
524 * config/i386/i386-protos.h (ix86_expand_push): Remove.
525 * config/i386/i386.md (TImode move expander): De not call
526 ix86_expand_push.
527 (FP push expanders): Preserve memory attributes.
528 * config/i386/sse.md (push<mode>1): Remove.
529 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
530 (ix86_expand_push): Remove.
531 * config/i386/mmx.md (push<mode>1): Remove.
532
533 2014-02-06 Jakub Jelinek <jakub@redhat.com>
534
535 PR rtl-optimization/60030
536 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
537 lopart with paradoxical subreg before shifting it up by hprec.
538
539 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
540
541 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
542 Remove extra newline at end of file.
543 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
544 (arm_issue_rate): Handle cortexa57.
545 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
546 (cortex-a57.cortex-a53): Likewise.
547
548 2014-02-06 Jakub Jelinek <jakub@redhat.com>
549
550 PR target/59575
551 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
552 don't record in REG_FRAME_RELATED_EXPR registers not set in that
553 bitmask.
554 (arm_expand_prologue): Adjust all callers.
555 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
556 info, registers also at the lowest numbered registers side. Use
557 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
558 XEXP.
559
560 PR debug/59992
561 * var-tracking.c (adjust_mems): Before adding a SET to
562 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
563
564 2014-02-06 Alan Modra <amodra@gmail.com>
565
566 PR target/60032
567 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
568 change SDmode to DDmode when lra_in_progress.
569
570 2014-02-06 Jakub Jelinek <jakub@redhat.com>
571
572 PR middle-end/59150
573 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
574 free_data_ref on the dr first, and before goto again also set dr
575 to the next dr. For simd_lane_access, free old datarefs[i] before
576 overwriting it. For get_vectype_for_scalar_type failure, don't
577 free_data_ref if simd_lane_access.
578
579 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
580
581 PR target/60062
582 * tree.h (opts_for_fn): New inline function.
583 (opt_for_fn): Define.
584 * config/i386/i386.c (ix86_function_regparm): Use
585 opt_for_fn (decl, optimize) instead of optimize.
586
587 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
588
589 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
590 for SYMBOL_REF in large memory model.
591
592 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
593
594 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
595 and crypto support.
596 (cortex-a57): Likewise.
597 (cortex-a57.cortex-a53): Likewise.
598
599 2014-02-06 Yury Gribov <y.gribov@samsung.com>
600 Kugan Vivekanandarajah <kuganv@linaro.org>
601
602 * config/arm/arm.c (arm_vector_alignment_reachable): Check
603 unaligned_access.
604 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
605
606 2014-02-06 Richard Biener <rguenther@suse.de>
607
608 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
609 set_loop_copy and initialize_original_copy_tables.
610
611 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
612
613 * config/aarch64/aarch64-simd.md
614 (aarch64_ashr_simddi): Change QI to SI.
615
616 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
617 Jakub Jelinek <jakub@redhat.com>
618
619 PR middle-end/60013
620 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
621 of the dataflow.
622
623 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
624
625 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
626 CODE_FOR_altivec_vpku[hw]um to
627 CODE_FOR_altivec_vpku[hw]um_direct.
628 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
629 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
630 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
631 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
632
633 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
634
635 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
636 generation for -maltivec=be.
637 (altivec_vsumsws): Simplify redundant test.
638
639 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
640
641 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
642 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
643 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
644 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
645 gen_altivec_vpkuwum.
646 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
647 BYTES_BIG_ENDIAN.
648 (altivec_vpks<VI_char>ss): Likewise.
649 (altivec_vpks<VI_char>us): Likewise.
650 (altivec_vpku<VI_char>us): Likewise.
651 (altivec_vpku<VI_char>um): Likewise.
652 (altivec_vpku<VI_char>um_direct): New (copy of
653 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
654 internal use).
655 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
656 target is little endian and -maltivec=be is not specified.
657 (*altivec_vupkhs<VU_char>_direct): New (copy of
658 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
659 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
660 target is little endian and -maltivec=be is not specified.
661 (*altivec_vupkls<VU_char>_direct): New (copy of
662 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
663 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
664 little endian and -maltivec=be is not specified.
665 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
666 little endian and -maltivec=be is not specified.
667
668 2014-02-05 Richard Henderson <rth@redhat.com>
669
670 PR debug/52727
671 * combine-stack-adj.c: Revert r206943.
672 * sched-int.h (struct deps_desc): Add last_args_size.
673 * sched-deps.c (init_deps): Initialize it.
674 (sched_analyze_insn): Add OUTPUT dependencies between insns that
675 contain REG_ARGS_SIZE notes.
676
677 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
678
679 * lto-cgraph.c (asm_nodes_output): Make global.
680 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
681 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
682 (driver_handle_option): Handle OPT_fwpa.
683
684 2014-02-05 Jakub Jelinek <jakub@redhat.com>
685
686 PR ipa/59947
687 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
688 a comment typo and formatting issue. If odr_hash hasn't been
689 created, return vNULL and set *completep to false.
690
691 PR middle-end/57499
692 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
693 bb with no successors.
694
695 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
696
697 PR target/59718
698 * doc/invoke.texi (-march): Clarify documentation for ARM.
699 (-mtune): Likewise.
700 (-mcpu): Likewise.
701
702 2014-02-05 Richard Biener <rguenther@suse.de>
703
704 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
705 when not vectorizing because of too many alias checks.
706 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
707 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
708
709 2014-02-05 Nick Clifton <nickc@redhat.com>
710
711 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
712 accept extended registers in any mode when compiling for the MN10300.
713
714 2014-02-05 Yury Gribov <y.gribov@samsung.com>
715
716 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
717 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
718 sanitization attributes.
719 (can_inline_edge_p): Likewise.
720 (sanitize_attrs_match_for_inline_p): New function.
721
722 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
723
724 * ipa-prop.c (detect_type_change): Shor circuit testing of
725 type changes on THIS pointer.
726
727 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
728
729 PR target/59777
730 * config/pa/pa.c (legitimize_tls_address): Return original address
731 if not passed a SYMBOL_REF rtx.
732 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
733 addresses.
734 (pa_emit_move_sequence): Simplify TLS source operands.
735 (pa_legitimate_constant_p): Reject all TLS constants.
736 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
737 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
738
739 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
740
741 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
742 groups when we know they are controlled by LTO.
743 * varasm.c (default_binds_local_p_1): If object is in other partition,
744 it will be resolved locally.
745
746 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
747
748 * config/host-linux.c (linux_gt_pch_use_address): Don't
749 use SSIZE_MAX because it is not always defined.
750
751 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
752
753 PR bootstrap/59913
754 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
755 threshold for pseudo splitting.
756 (update_ebb_live_info): Process call argument hard registers and
757 hard registers from insn definition too.
758 (max_small_class_regs_num): New constant.
759 (inherit_in_ebb): Update live hard regs through EBBs. Update
760 reloads_num only for small register classes. Don't split for
761 outputs of jumps.
762
763 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
764
765 PR ipa/60058
766 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
767 is non-null.
768
769 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
770
771 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
772 visibility is safe.
773
774 2014-02-04 Marek Polacek <polacek@redhat.com>
775
776 * gdbinit.in (pel): Define.
777
778 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
779
780 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
781 behavior.
782
783 2014-02-04 Richard Biener <rguenther@suse.de>
784
785 PR lto/59723
786 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
787 in function context local.
788 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
789 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
790 similar to LTO_imported_decl_ref.
791
792 2014-02-04 Jakub Jelinek <jakub@redhat.com>
793
794 PR tree-optimization/60002
795 * cgraphclones.c (build_function_decl_skip_args): Clear
796 DECL_LANG_SPECIFIC.
797
798 PR tree-optimization/60023
799 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
800 false to gsi_replace.
801 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
802 has been in some EH region and vec_stmt could throw, add
803 vec_stmt into the same EH region.
804 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
805 has no lhs, ignore it.
806 * internal-fn.c (expand_MASK_LOAD): Likewise.
807
808 PR ipa/60026
809 * tree-inline.c (copy_forbidden): Fail for
810 __attribute__((optimize (0))) functions.
811
812 PR other/58712
813 * omp-low.c (simd_clone_struct_copy): If from->inbranch
814 is set, copy one less argument.
815 (expand_simd_clones): Don't subtract clone_info->inbranch
816 from simd_clone_struct_alloc argument.
817
818 PR rtl-optimization/57915
819 * recog.c (simplify_while_replacing): If all unary/binary/relational
820 operation arguments are constant, attempt to simplify those.
821
822 PR middle-end/59261
823 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
824 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
825
826 2014-02-04 Richard Biener <rguenther@suse.de>
827
828 PR tree-optimization/60012
829 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
830 TBAA disambiguation to all DDRs.
831
832 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
833
834 PR target/59788
835 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
836 (LINK_SPEC): Use it for -shared, -shared-libgcc.
837
838 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
839
840 PR ipa/59882
841 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
842
843 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
844
845 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
846 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
847
848 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
849
850 PR ipa/59831
851 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
852 to figure out targets of polymorphic calls with known decl.
853 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
854 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
855 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
856 (get_polymorphic_call_info): ... here.
857 (get_polymorphic_call_info_from_invariant): New function.
858
859 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
860
861 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
862 lookup via vtable pointer; check for type consistency
863 and turn inconsitent facts into UNREACHABLE.
864 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
865 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
866 type inconsistent querries; return UNREACHABLE instead.
867
868 2014-02-03 Richard Henderson <rth@twiddle.net>
869
870 PR tree-opt/59924
871 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
872 already processed this node.
873 (normalize_one_pred_1): Pass along mark_set.
874 (normalize_one_pred): Create and destroy a pointer_set_t.
875 (normalize_one_pred_chain): Likewise.
876
877 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
878
879 PR gcov-profile/58602
880 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
881
882 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
883
884 PR ipa/59831
885 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
886 -fno-devirtualize; try to devirtualize by the knowledge of
887 virtual table pointer given by aggregate propagation.
888 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
889 (ipa_print_node_jump_functions): Dump also offset that
890 is relevant for polymorphic calls.
891 (determine_known_aggregate_parts): Add arg_type parameter; use it
892 instead of determining the type from pointer type.
893 (ipa_compute_jump_functions_for_edge): Update call of
894 determine_known_aggregate_parts.
895 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
896 (gimple_get_virt_method_for_binfo): ... here; simplify using
897 vtable_pointer_value_to_vtable.
898 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
899 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
900 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
901 (vtable_pointer_value_to_vtable): Break out from ...; handle also
902 POINTER_PLUS_EXPR.
903 (vtable_pointer_value_to_binfo): ... here.
904 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
905
906 2014-02-03 Teresa Johnson <tejohnson@google.com>
907
908 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
909 redef of outer loop index variable.
910
911 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
912
913 PR c++/53017
914 PR c++/59211
915 * doc/extend.texi (Function Attributes): Typo.
916
917 2014-02-03 Cong Hou <congh@google.com>
918
919 PR tree-optimization/60000
920 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
921 if the vectorized statement is a store. A store statement can only
922 appear at the end of pattern statements.
923
924 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
925
926 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
927 (ix86_option_override_internal): Default long double to 64-bit for
928 32-bit Bionic and to 128-bit for 64-bit Bionic.
929
930 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
931 TARGET_LONG_DOUBLE_128 is true.
932 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
933
934 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
935 (mlong-double-64): Negate -mlong-double-128.
936 (mlong-double-128): New option.
937
938 * config/i386/i386-c.c (ix86_target_macros): Define
939 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
940
941 * doc/invoke.texi: Document -mlong-double-128.
942
943 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
944
945 PR rtl-optimization/60024
946 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
947
948 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
949
950 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
951
952 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
953
954 PR rtl-optimization/57662
955 * sel-sched.c (code_motion_path_driver): Do not mark already not
956 existing blocks in the visiting bitmap.
957
958 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
959
960 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
961 on the insn being emitted.
962
963 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
964 Will Deacon <will.deacon@arm.com>
965
966 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
967
968 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
969
970 * config/arm/arm-tables.opt: Regenerate.
971
972 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
973
974 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
975 for vector types other than V16QImode.
976 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
977 define_expand, and call altivec_expand_vec_perm_le when producing
978 code with little endian element order.
979 (*altivec_vperm_<mode>_internal): New insn having previous
980 behavior of altivec_vperm_<mode>.
981 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
982 altivec_expand_vec_perm_le when producing code with little endian
983 element order.
984 (*altivec_vperm_<mode>_uns_internal): New insn having previous
985 behavior of altivec_vperm_<mode>_uns.
986
987 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
988
989 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
990 (altivec_vsumsws): Add handling for -maltivec=be with a little
991 endian target.
992 (altivec_vsumsws_direct): New.
993 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
994 gen_altivec_vsumsws.
995
996 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
997
998 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
999 vtable_pointer_value_to_binfo): New functions.
1000 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
1001 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
1002
1003 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
1004
1005 * config/nios2/nios2.md (load_got_register): Initialize GOT
1006 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
1007 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
1008
1009 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
1010
1011 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
1012 preserverd by passthrough, do not propagate the type.
1013
1014 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
1015
1016 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
1017 (mips_atomic_assign_expand_fenv): New function.
1018 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
1019
1020 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
1021
1022 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
1023 (__builtin_mips_set_fcsr): Likewise.
1024 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
1025 MIPS_USI_FTYPE_VOID.
1026 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
1027 (mips16_expand_set_fcsr): Likewise.
1028 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
1029 (mips16_set_fcsr_stub): Likewise.
1030 (mips16_get_fcsr_one_only_stub): New class.
1031 (mips16_set_fcsr_one_only_stub): Likewise.
1032 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
1033 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
1034 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
1035 (hard_float): New availability predicate.
1036 (mips_builtins): Add get_fcsr and set_fcsr.
1037 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
1038 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
1039 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
1040 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
1041 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
1042 patterns.
1043
1044 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
1045
1046 * config/mips/mips.c (mips_one_only_stub): New class.
1047 (mips_need_mips16_rdhwr_p): Replace with...
1048 (mips16_rdhwr_stub): ...this new variable.
1049 (mips16_stub_call_address): New function.
1050 (mips16_rdhwr_one_only_stub): New class.
1051 (mips_expand_thread_pointer): Use mips16_stub_call_address.
1052 (mips_output_mips16_rdhwr): Delete.
1053 (mips_finish_stub): New function.
1054 (mips_code_end): Use it to handle rdhwr stubs.
1055
1056 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
1057
1058 PR target/60017
1059 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
1060 when calculating size of integer atomic types.
1061
1062 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
1063
1064 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
1065
1066 2014-02-01 Jakub Jelinek <jakub@redhat.com>
1067
1068 PR tree-optimization/60003
1069 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
1070 * profile.c (branch_prob): Use gimple_call_builtin_p
1071 to check for BUILT_IN_SETJMP_RECEIVER.
1072 * tree-inline.c (copy_bb): Call notice_special_calls.
1073
1074 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
1075
1076 PR bootstrap/59985
1077 * lra-constraints.c (process_alt_operands): Update reload_sum only
1078 on the first pass.
1079
1080 2014-01-31 Richard Henderson <rth@redhat.com>
1081
1082 PR middle-end/60004
1083 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
1084 until after else_eh is processed.
1085
1086 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
1087
1088 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
1089 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
1090 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
1091 in smmintrin.h, remove them.
1092 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
1093 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
1094 * config/i386/i386.md (ROUND_SAE): Fix value.
1095 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
1096 (const48_operand): New.
1097 * config/i386/subst.md (round), (round_expand): Use
1098 const_4_or_8_to_11_operand.
1099 (round_saeonly), (round_saeonly_expand): Use const48_operand.
1100
1101 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
1102
1103 * config/i386/constraints.md (Yk): Swap meaning with k.
1104 * config/i386/i386.md (movhi_internal): Change Yk to k.
1105 (movqi_internal): Ditto.
1106 (*k<logic><mode>): Ditto.
1107 (*andhi_1): Ditto.
1108 (*andqi_1): Ditto.
1109 (kandn<mode>): Ditto.
1110 (*<code>hi_1): Ditto.
1111 (*<code>qi_1): Ditto.
1112 (kxnor<mode>): Ditto.
1113 (kortestzhi): Ditto.
1114 (kortestchi): Ditto.
1115 (kunpckhi): Ditto.
1116 (*one_cmplhi2_1): Ditto.
1117 (*one_cmplqi2_1): Ditto.
1118 * config/i386/sse.md (): Change k to Yk.
1119 (avx512f_load<mode>_mask): Ditto.
1120 (avx512f_blendm<mode>): Ditto.
1121 (avx512f_store<mode>_mask): Ditto.
1122 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1123 (avx512f_storedqu<mode>_mask): Ditto.
1124 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
1125 Ditto.
1126 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
1127 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
1128 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
1129 (avx512f_maskcmp<mode>3): Ditto.
1130 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
1131 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
1132 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
1133 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
1134 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
1135 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
1136 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
1137 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
1138 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
1139 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
1140 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
1141 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
1142 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1143 (vec_extract_lo_<mode>_maskm): Ditto.
1144 (vec_extract_hi_<mode>_maskm): Ditto.
1145 (avx512f_vternlog<mode>_mask): Ditto.
1146 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
1147 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
1148 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1149 (avx512f_<code>v8div16qi2_mask): Ditto.
1150 (avx512f_<code>v8div16qi2_mask_store): Ditto.
1151 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
1152 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
1153 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
1154 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
1155 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
1156 (*avx512pf_gatherpf<mode>df_mask): Ditto.
1157 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
1158 (*avx512pf_scatterpf<mode>df_mask): Ditto.
1159 (avx512cd_maskb_vec_dupv8di): Ditto.
1160 (avx512cd_maskw_vec_dupv16si): Ditto.
1161 (avx512f_vpermi2var<mode>3_maskz): Ditto.
1162 (avx512f_vpermi2var<mode>3_mask): Ditto.
1163 (avx512f_vpermi2var<mode>3_mask): Ditto.
1164 (avx512f_vpermt2var<mode>3_maskz): Ditto.
1165 (*avx512f_gathersi<mode>): Ditto.
1166 (*avx512f_gathersi<mode>_2): Ditto.
1167 (*avx512f_gatherdi<mode>): Ditto.
1168 (*avx512f_gatherdi<mode>_2): Ditto.
1169 (*avx512f_scattersi<mode>): Ditto.
1170 (*avx512f_scatterdi<mode>): Ditto.
1171 (avx512f_compress<mode>_mask): Ditto.
1172 (avx512f_compressstore<mode>_mask): Ditto.
1173 (avx512f_expand<mode>_mask): Ditto.
1174 * config/i386/subst.md (mask): Change k to Yk.
1175 (mask_scalar_merge): Ditto.
1176 (sd): Ditto.
1177
1178 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
1179
1180 * doc/extend.texi (Vector Extensions): Document ?: in C++.
1181
1182 2014-01-31 Richard Biener <rguenther@suse.de>
1183
1184 PR middle-end/59990
1185 * builtins.c (fold_builtin_memory_op): Make sure to not
1186 use a floating-point mode or a boolean or enumeral type for
1187 the copy operation.
1188
1189 2014-01-30 DJ Delorie <dj@redhat.com>
1190
1191 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
1192 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
1193 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
1194 whenever main() has an epilogue.
1195
1196 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1197
1198 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
1199 unused variable "field".
1200 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
1201 (vsx_mergeh_<mode>): Likewise.
1202 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
1203 (altivec_vmrghh): Likewise.
1204 (altivec_vmrghw): Likewise.
1205 (altivec_vmrglb): Likewise.
1206 (altivec_vmrglh): Likewise.
1207 (altivec_vmrglw): Likewise.
1208 (altivec_vspltb): Add missing uses.
1209 (altivec_vsplth): Likewise.
1210 (altivec_vspltw): Likewise.
1211 (altivec_vspltsf): Likewise.
1212
1213 2014-01-30 Jakub Jelinek <jakub@redhat.com>
1214
1215 PR target/59923
1216 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
1217 frame related instructions.
1218
1219 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
1220
1221 PR rtl-optimization/59959
1222 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
1223 any reload of register whose subreg is invalid.
1224
1225 2014-01-30 Jakub Jelinek <jakub@redhat.com>
1226
1227 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
1228 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
1229 Add missing return type - void.
1230
1231 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1232
1233 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
1234 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
1235 remove element index adjustment for endian (now handled in vsx.md
1236 and altivec.md).
1237 (altivec_expand_vec_perm_const): Use
1238 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
1239 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
1240 (vsx_xxspltw_<mode>): Adjust element index for little endian.
1241 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
1242 define_expand and a new define_insn *altivec_vspltb_internal;
1243 adjust for -maltivec=be on a little endian target.
1244 (altivec_vspltb_direct): New.
1245 (altivec_vsplth): Divide into a define_expand and a new
1246 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
1247 little endian target.
1248 (altivec_vsplth_direct): New.
1249 (altivec_vspltw): Divide into a define_expand and a new
1250 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
1251 little endian target.
1252 (altivec_vspltw_direct): New.
1253 (altivec_vspltsf): Divide into a define_expand and a new
1254 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
1255 a little endian target.
1256
1257 2014-01-30 Richard Biener <rguenther@suse.de>
1258
1259 PR tree-optimization/59993
1260 * tree-ssa-forwprop.c (associate_pointerplus): Check we
1261 can propagate form the earlier stmt and avoid the transform
1262 when the intermediate result is needed.
1263
1264 2014-01-30 Alangi Derick <alangiderick@gmail.com>
1265
1266 * README.Portability: Fix typo.
1267
1268 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
1269
1270 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
1271 comparison_operator with ordered_comparison_operator.
1272
1273 2014-01-30 Nick Clifton <nickc@redhat.com>
1274
1275 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
1276 Rename to mn10300_store_multiple_regs.
1277 * config/mn10300/mn10300.c: Likewise.
1278 * config/mn10300/mn10300.md (store_movm): Fix typo: call
1279 store_multiple_regs.
1280 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
1281 Call mn10300_store_multiple_regs.
1282
1283 2014-01-30 Nick Clifton <nickc@redhat.com>
1284 DJ Delorie <dj@redhat.com>
1285
1286 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
1287 %fp 2 to keep registers after it properly word-aligned.
1288 (rl78_alloc_physical_registers_umul): Handle the case where both
1289 input operands are the same.
1290
1291 2014-01-30 Richard Biener <rguenther@suse.de>
1292
1293 PR tree-optimization/59903
1294 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
1295 check properly.
1296
1297 2014-01-30 Jason Merrill <jason@redhat.com>
1298
1299 PR c++/59633
1300 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
1301
1302 PR c++/59645
1303 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
1304
1305 2014-01-30 Richard Biener <rguenther@suse.de>
1306
1307 PR tree-optimization/59951
1308 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
1309
1310 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
1311
1312 PR target/59784
1313 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
1314 SFmode to DFmode case.
1315
1316 2014-01-29 DJ Delorie <dj@redhat.com>
1317
1318 * config/msp430/msp430.opt (-minrt): New.
1319 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
1320 if -minrt given.
1321 (ENDFILE_SPEC): Likewise.
1322
1323 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
1324
1325 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
1326 (estimate_function_body_sizes): Use it.
1327
1328 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
1329
1330 PR c++/58561
1331 * dwarf2out.c (is_cxx_auto): New.
1332 (is_base_type): Use it.
1333 (gen_type_die_with_usage): Likewise.
1334
1335 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1336
1337 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
1338 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
1339 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
1340 -maltivec=be with LE targets.
1341 (vsx_mergeh_<mode>): Likewise.
1342 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
1343 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
1344 (altivec_vmrghb): Replace with define_expand and new
1345 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
1346 (altivec_vmrghb_direct): New define_insn.
1347 (altivec_vmrghh): Replace with define_expand and new
1348 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
1349 (altivec_vmrghh_direct): New define_insn.
1350 (altivec_vmrghw): Replace with define_expand and new
1351 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
1352 (altivec_vmrghw_direct): New define_insn.
1353 (*altivec_vmrghsf): Adjust for endianness.
1354 (altivec_vmrglb): Replace with define_expand and new
1355 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
1356 (altivec_vmrglb_direct): New define_insn.
1357 (altivec_vmrglh): Replace with define_expand and new
1358 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
1359 (altivec_vmrglh_direct): New define_insn.
1360 (altivec_vmrglw): Replace with define_expand and new
1361 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
1362 (altivec_vmrglw_direct): New define_insn.
1363 (*altivec_vmrglsf): Adjust for endianness.
1364 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1365 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1366 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1367 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1368 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1369 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1370 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1371 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1372
1373 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
1374
1375 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
1376 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
1377 whitespace.
1378
1379 2014-01-29 Richard Biener <rguenther@suse.de>
1380
1381 PR tree-optimization/58742
1382 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
1383 associate_pointerplus_align.
1384 (associate_pointerplus_diff): New function.
1385 (associate_pointerplus): Likewise. Call associate_pointerplus_align
1386 and associate_pointerplus_diff.
1387
1388 2014-01-29 Richard Biener <rguenther@suse.de>
1389
1390 * lto-streamer.h (LTO_major_version): Bump to 3.
1391 (LTO_minor_version): Reset to 0.
1392
1393 2014-01-29 Renlin Li <Renlin.Li@arm.com>
1394
1395 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
1396 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
1397 (arm_file_start): Generate correct asm header for armv7ve.
1398 * config/arm/bpabi.h: Add multilib support for armv7ve.
1399 * config/arm/driver-arm.c: Change the architectures of cortex-a7
1400 and cortex-a15 to armv7ve.
1401 * config/arm/t-aprofile: Add multilib support for armv7ve.
1402 * doc/invoke.texi: Document -march=armv7ve.
1403
1404 2014-01-29 Richard Biener <rguenther@suse.de>
1405
1406 PR tree-optimization/58742
1407 * tree-ssa-forwprop.c (associate_plusminus): Return true
1408 if we changed sth, defer EH cleanup to ...
1409 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
1410 (simplify_mult): New function.
1411
1412 2014-01-29 Jakub Jelinek <jakub@redhat.com>
1413
1414 PR middle-end/59917
1415 PR tree-optimization/59920
1416 * tree.c (build_common_builtin_nodes): Remove
1417 __builtin_setjmp_dispatcher initialization.
1418 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
1419 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
1420 instead of gsi_after_labels + manually skipping debug stmts.
1421 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
1422 ignore bbs with IFN_ABNORMAL_DISPATCHER.
1423 * tree-inline.c (copy_edges_for_bb): Remove
1424 can_make_abnormal_goto argument, instead add abnormal_goto_dest
1425 argument. Ignore computed_goto_p stmts. Don't call
1426 make_abnormal_goto_edges. If a call might need abnormal edges
1427 for non-local gotos, see if it already has an edge to
1428 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
1429 with true argument, don't do anything then, otherwise add
1430 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
1431 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
1432 caller.
1433 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
1434 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
1435 (lower_stmt): Don't set data->calls_builtin_setjmp.
1436 (lower_builtin_setjmp): Adjust comment.
1437 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
1438 * tree-cfg.c (found_computed_goto): Remove.
1439 (factor_computed_gotos): Remove.
1440 (make_goto_expr_edges): Return bool, true for computed gotos.
1441 Don't call make_abnormal_goto_edges.
1442 (build_gimple_cfg): Don't set found_computed_goto, don't call
1443 factor_computed_gotos.
1444 (computed_goto_p): No longer static.
1445 (make_blocks): Don't set found_computed_goto.
1446 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
1447 (make_edges): If make_goto_expr_edges returns true, push bb
1448 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
1449 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
1450 vector. Record mapping between bbs and OpenMP regions if there
1451 are any, adjust make_gimple_omp_edges caller. Call
1452 handle_abnormal_edges.
1453 (make_abnormal_goto_edges): Remove.
1454 * tree-cfg.h (make_abnormal_goto_edges): Remove.
1455 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
1456 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
1457 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
1458 * internal-fn.def (ABNORMAL_DISPATCHER): New.
1459 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
1460 filling *region also set *region_idx to (*region)->entry->index.
1461
1462 PR other/58712
1463 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
1464 For REGs set ORIGINAL_REGNO.
1465
1466 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
1467
1468 * doc/md.texi: Mention that a target shouldn't implement
1469 vec_widen_(s|u)mul_even/odd pair if it is less efficient
1470 than hi/lo pair.
1471
1472 2014-01-29 Jakub Jelinek <jakub@redhat.com>
1473
1474 PR tree-optimization/59594
1475 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
1476 a copy of the datarefs vector rather than the vector itself.
1477
1478 2014-01-28 Jason Merrill <jason@redhat.com>
1479
1480 PR c++/53756
1481 * dwarf2out.c (auto_die): New static.
1482 (gen_type_die_with_usage): Handle C++1y 'auto'.
1483 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
1484 on definition.
1485
1486 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
1487
1488 PR target/59672
1489 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
1490 (SPEC_X32): Likewise.
1491 (SPEC_64): Likewise.
1492 * config/i386/i386.c (ix86_option_override_internal): Turn off
1493 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
1494 for TARGET_16BIT.
1495 (x86_file_start): Output .code16gcc for TARGET_16BIT.
1496 * config/i386/i386.h (TARGET_16BIT): New macro.
1497 (TARGET_16BIT_P): Likewise.
1498 * config/i386/i386.opt: Add m16.
1499 * doc/invoke.texi: Document -m16.
1500
1501 2014-01-28 Jakub Jelinek <jakub@redhat.com>
1502
1503 PR preprocessor/59935
1504 * input.c (location_get_source_line): Bail out on when line number
1505 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
1506
1507 2014-01-28 Richard Biener <rguenther@suse.de>
1508
1509 PR tree-optimization/58742
1510 * tree-ssa-forwprop.c (associate_plusminus): Handle
1511 pointer subtraction of the form (T)(P + A) - (T)P.
1512
1513 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1514
1515 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
1516 at const_int_cost.
1517
1518 2014-01-28 Richard Biener <rguenther@suse.de>
1519
1520 Revert
1521 2014-01-28 Richard Biener <rguenther@suse.de>
1522
1523 PR rtl-optimization/45364
1524 PR rtl-optimization/59890
1525 * var-tracking.c (local_get_addr_clear_given_value): Handle
1526 already cleared slot.
1527 (val_reset): Handle not allocated local_get_addr_cache.
1528 (vt_find_locations): Use post-order on the inverted CFG.
1529
1530 2014-01-28 Richard Biener <rguenther@suse.de>
1531
1532 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
1533
1534 2014-01-28 Richard Biener <rguenther@suse.de>
1535
1536 PR rtl-optimization/45364
1537 PR rtl-optimization/59890
1538 * var-tracking.c (local_get_addr_clear_given_value): Handle
1539 already cleared slot.
1540 (val_reset): Handle not allocated local_get_addr_cache.
1541 (vt_find_locations): Use post-order on the inverted CFG.
1542
1543 2014-01-28 Alan Modra <amodra@gmail.com>
1544
1545 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
1546 * configure.ac <recursive call for build != host>: Define
1547 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
1548 and LD_FOR_BUILD too.
1549 * configure: Regenerate.
1550
1551 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
1552
1553 * config/i386/i386.c (get_builtin_code_for_version): Separate
1554 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
1555 Broadwell from Haswell.
1556
1557 2014-01-27 Steve Ellcey <sellcey@mips.com>
1558
1559 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
1560 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
1561 * config/mips/mips.c (mips_option_override): Change setting
1562 of TARGET_DSP.
1563 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
1564 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
1565 Change from Mask to Var.
1566
1567 2014-01-27 Jeff Law <law@redhat.com>
1568
1569 * ipa-inline.c (inline_small_functions): Fix typo.
1570
1571 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
1572
1573 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
1574 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
1575 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
1576 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
1577 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
1578 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
1579 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
1580 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
1581 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
1582 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
1583 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
1584 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
1585 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
1586 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
1587 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
1588 (_mm512_storeu_epi64): Ditto.
1589 (_mm512_cmpge_epi32_mask): Ditto.
1590 (_mm512_cmpge_epu32_mask): Ditto.
1591 (_mm512_cmpge_epi64_mask): Ditto.
1592 (_mm512_cmpge_epu64_mask): Ditto.
1593 (_mm512_cmple_epi32_mask): Ditto.
1594 (_mm512_cmple_epu32_mask): Ditto.
1595 (_mm512_cmple_epi64_mask): Ditto.
1596 (_mm512_cmple_epu64_mask): Ditto.
1597 (_mm512_cmplt_epi32_mask): Ditto.
1598 (_mm512_cmplt_epu32_mask): Ditto.
1599 (_mm512_cmplt_epi64_mask): Ditto.
1600 (_mm512_cmplt_epu64_mask): Ditto.
1601 (_mm512_cmpneq_epi32_mask): Ditto.
1602 (_mm512_cmpneq_epu32_mask): Ditto.
1603 (_mm512_cmpneq_epi64_mask): Ditto.
1604 (_mm512_cmpneq_epu64_mask): Ditto.
1605 (_mm512_expand_pd): Ditto.
1606 (_mm512_expand_ps): Ditto.
1607 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
1608 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
1609 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
1610 * config/i386/i386.c (ix86_builtins): Add
1611 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
1612 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
1613 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
1614 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
1615 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
1616 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
1617 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
1618 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
1619 IX86_BUILTIN_PMOVUSQW512_MEM.
1620 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
1621 __builtin_ia32_pmovsqd512mem_mask,
1622 __builtin_ia32_pmovqd512mem_mask,
1623 __builtin_ia32_pmovusqw512mem_mask,
1624 __builtin_ia32_pmovsqw512mem_mask,
1625 __builtin_ia32_pmovqw512mem_mask,
1626 __builtin_ia32_pmovusdw512mem_mask,
1627 __builtin_ia32_pmovsdw512mem_mask,
1628 __builtin_ia32_pmovdw512mem_mask,
1629 __builtin_ia32_pmovqb512mem_mask,
1630 __builtin_ia32_pmovusqb512mem_mask,
1631 __builtin_ia32_pmovsqb512mem_mask,
1632 __builtin_ia32_pmovusdb512mem_mask,
1633 __builtin_ia32_pmovsdb512mem_mask,
1634 __builtin_ia32_pmovdb512mem_mask.
1635 (bdesc_args): Add __builtin_ia32_expanddf512,
1636 __builtin_ia32_expandsf512.
1637 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
1638 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
1639 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
1640 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
1641 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
1642 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
1643 (avx512f_<code>v8div16qi2_mask_store): This.
1644 (avx512f_expand<mode>): New.
1645
1646 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
1647
1648 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
1649 New.
1650 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1651 (_mm512_prefetch_i32scatter_pd): Ditto.
1652 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1653 (_mm512_prefetch_i64scatter_pd): Ditto.
1654 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1655 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
1656 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1657 (_mm512_prefetch_i32scatter_ps): Ditto.
1658 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1659 (_mm512_prefetch_i64scatter_ps): Ditto.
1660 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1661 * config/i386/i386-builtin-types.def: Define
1662 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
1663 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
1664 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
1665 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
1666 IX86_BUILTIN_SCATTERPFQPD.
1667 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
1668 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
1669 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
1670 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
1671 __builtin_ia32_scatterpfqps.
1672 (ix86_expand_builtin): Expand new built-ins.
1673 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
1674 fix memory access data type.
1675 (*avx512pf_gatherpf<mode>_mask): Ditto.
1676 (*avx512pf_gatherpf<mode>): Ditto.
1677 (avx512pf_scatterpf<mode>): Ditto.
1678 (*avx512pf_scatterpf<mode>_mask): Ditto.
1679 (*avx512pf_scatterpf<mode>): Ditto.
1680 (GATHER_SCATTER_SF_MEM_MODE): New.
1681 (avx512pf_gatherpf<mode>df): Ditto.
1682 (*avx512pf_gatherpf<mode>df_mask): Ditto.
1683 (*avx512pf_scatterpf<mode>df): Ditto.
1684
1685 2014-01-27 Jakub Jelinek <jakub@redhat.com>
1686
1687 PR bootstrap/59934
1688 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
1689 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
1690 reached.
1691
1692 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
1693
1694 * common/config/arm/arm-common.c
1695 (arm_rewrite_mcpu): Handle multiple names.
1696 * config/arm/arm.h
1697 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
1698
1699 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
1700
1701 * gimple-builder.h (create_gimple_tmp): Delete.
1702
1703 2014-01-27 Christian Bruel <christian.bruel@st.com>
1704
1705 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
1706 words comparisons.
1707
1708 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
1709
1710 * config/pa/pa.md (call): Generate indirect long calls to non-local
1711 functions when outputing 32-bit code.
1712 (call_value): Likewise except for special call to buggy powf function.
1713
1714 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
1715 portable runtime and PIC indirect calls.
1716 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
1717 and PIC call sequences. Use ldo instead of blr to set return register
1718 in PIC call sequence.
1719
1720 2014-01-25 Walter Lee <walt@tilera.com>
1721
1722 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
1723 avoid clobbering a live register.
1724
1725 2014-01-25 Walter Lee <walt@tilera.com>
1726
1727 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
1728 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
1729 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
1730 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
1731
1732 2014-01-25 Walter Lee <walt@tilera.com>
1733
1734 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
1735 arguments on even registers.
1736 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
1737 STACK_BOUNDARY.
1738 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
1739 (BIGGEST_ALIGNMENT): Ditto.
1740 (BIGGEST_FIELD_ALIGNMENT): Ditto.
1741
1742 2014-01-25 Walter Lee <walt@tilera.com>
1743
1744 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
1745 insns before bundling.
1746 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
1747
1748 2014-01-25 Walter Lee <walt@tilera.com>
1749
1750 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
1751 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
1752 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
1753
1754 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
1755
1756 * config/mips/constraints.md (kl): Delete.
1757 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
1758 define expands, using...
1759 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
1760 instructions for MIPS16.
1761 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
1762 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
1763
1764 2014-01-25 Walter Lee <walt@tilera.com>
1765
1766 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
1767 (clzdi2): Ditto.
1768 (ffsdi2): Ditto.
1769
1770 2014-01-25 Walter Lee <walt@tilera.com>
1771
1772 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
1773 (TARGET_EXPAND_TO_RTL_HOOK): Define.
1774
1775 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
1776
1777 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
1778 Handle XOR.
1779
1780 2014-01-25 Jakub Jelinek <jakub@redhat.com>
1781
1782 * print-rtl.c (in_call_function_usage): New var.
1783 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
1784 EXPR_LIST mode as mode and not as reg note name.
1785
1786 PR middle-end/59561
1787 * cfgloopmanip.c (copy_loop_info): If
1788 loop->warned_aggressive_loop_optimizations, make sure
1789 the flag is set in target loop too.
1790
1791 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
1792
1793 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
1794 flag_cilkplus.
1795 * builtins.def: Likewise.
1796 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
1797 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
1798 * ira.c (ira_setup_eliminable_regset): Likewise.
1799 * omp-low.c (gate_expand_omp): Likewise.
1800 (execute_lower_omp): Likewise.
1801 (diagnose_sb_0): Likewise.
1802 (gate_diagnose_omp_blocks): Likewise.
1803 (simd_clone_clauses_extract): Likewise.
1804 (gate): Likewise.
1805
1806 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1807
1808 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
1809 correction for little endian...
1810 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
1811 here.
1812
1813 2014-01-24 Jeff Law <law@redhat.com>
1814
1815 PR tree-optimization/59919
1816 * tree-vrp.c (find_assert_locations_1): Do not register asserts
1817 for non-returning calls.
1818
1819 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
1820
1821 * common/config/aarch64/aarch64-common.c
1822 (aarch64_rewrite_mcpu): Handle multiple names.
1823 * config/aarch64/aarch64.h
1824 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
1825
1826 2014-01-24 Dodji Seketeli <dodji@redhat.com>
1827
1828 * input.c (add_file_to_cache_tab): Handle the case where fopen
1829 returns NULL.
1830
1831 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
1832
1833 PR target/59929
1834 * config/i386/i386.md (pushsf splitter): Get stack adjustment
1835 from push operand if code of push isn't PRE_DEC.
1836
1837 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1838
1839 PR target/59909
1840 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
1841 -mquad-memory-atomic. Update -mquad-memory documentation to say
1842 it is only used for non-atomic loads/stores.
1843
1844 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
1845 -mquad-memory or -mquad-memory-atomic switches.
1846
1847 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
1848 -mquad-memory-atomic to ISA 2.07 support.
1849
1850 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
1851 to separate support of normal quad word memory operations (ldq, stq)
1852 from the atomic quad word memory operations.
1853
1854 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
1855 support to separate non-atomic quad word operations from atomic
1856 quad word operations. Disable non-atomic quad word operations in
1857 little endian mode so that we don't have to swap words after the
1858 load and before the store.
1859 (quad_load_store_p): Add comment about atomic quad word support.
1860 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
1861 options printed with -mdebug=reg.
1862
1863 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
1864 -mquad-memory-atomic as the test for whether we have quad word
1865 atomic instructions.
1866 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
1867 or -mp8-vector are used, allow byte/half-word atomic operations.
1868
1869 * config/rs6000/sync.md (load_lockedti): Insure that the address
1870 is a proper indexed or indirect address for the lqarx instruction.
1871 On little endian systems, swap the hi/lo registers after the lqarx
1872 instruction.
1873 (load_lockedpti): Use indexed_or_indirect_operand predicate to
1874 insure the address is valid for the lqarx instruction.
1875 (store_conditionalti): Insure that the address is a proper indexed
1876 or indirect address for the stqcrx. instruction. On little endian
1877 systems, swap the hi/lo registers before doing the stqcrx.
1878 instruction.
1879 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
1880 insure the address is valid for the stqcrx. instruction.
1881
1882 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
1883 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
1884 type of quad memory support is available.
1885
1886 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
1887
1888 PR regression/59915
1889 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
1890 there is a danger of looping.
1891
1892 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
1893
1894 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
1895 force flag_ira_loop_pressure if set via command line.
1896
1897 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
1898
1899 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
1900 (ashr_simd): New builtin handling DI mode.
1901 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
1902 (aarch64_sshr_simddi): New match pattern.
1903 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
1904 (vshrd_n_s64): Likewise.
1905 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
1906
1907 2014-01-23 Nick Clifton <nickc@redhat.com>
1908
1909 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
1910 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
1911 favour of mcu specific scripts.
1912 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
1913 430x multilibs.
1914
1915 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1916 Alex Velenko <Alex.Velenko@arm.com>
1917
1918 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
1919 (vaddv_s16): Likewise.
1920 (vaddv_s32): Likewise.
1921 (vaddv_u8): Likewise.
1922 (vaddv_u16): Likewise.
1923 (vaddv_u32): Likewise.
1924 (vaddvq_s8): Likewise.
1925 (vaddvq_s16): Likewise.
1926 (vaddvq_s32): Likewise.
1927 (vaddvq_s64): Likewise.
1928 (vaddvq_u8): Likewise.
1929 (vaddvq_u16): Likewise.
1930 (vaddvq_u32): Likewise.
1931 (vaddvq_u64): Likewise.
1932 (vaddv_f32): Likewise.
1933 (vaddvq_f32): Likewise.
1934 (vaddvq_f64): Likewise.
1935 (vmaxv_f32): Likewise.
1936 (vmaxv_s8): Likewise.
1937 (vmaxv_s16): Likewise.
1938 (vmaxv_s32): Likewise.
1939 (vmaxv_u8): Likewise.
1940 (vmaxv_u16): Likewise.
1941 (vmaxv_u32): Likewise.
1942 (vmaxvq_f32): Likewise.
1943 (vmaxvq_f64): Likewise.
1944 (vmaxvq_s8): Likewise.
1945 (vmaxvq_s16): Likewise.
1946 (vmaxvq_s32): Likewise.
1947 (vmaxvq_u8): Likewise.
1948 (vmaxvq_u16): Likewise.
1949 (vmaxvq_u32): Likewise.
1950 (vmaxnmv_f32): Likewise.
1951 (vmaxnmvq_f32): Likewise.
1952 (vmaxnmvq_f64): Likewise.
1953 (vminv_f32): Likewise.
1954 (vminv_s8): Likewise.
1955 (vminv_s16): Likewise.
1956 (vminv_s32): Likewise.
1957 (vminv_u8): Likewise.
1958 (vminv_u16): Likewise.
1959 (vminv_u32): Likewise.
1960 (vminvq_f32): Likewise.
1961 (vminvq_f64): Likewise.
1962 (vminvq_s8): Likewise.
1963 (vminvq_s16): Likewise.
1964 (vminvq_s32): Likewise.
1965 (vminvq_u8): Likewise.
1966 (vminvq_u16): Likewise.
1967 (vminvq_u32): Likewise.
1968 (vminnmv_f32): Likewise.
1969 (vminnmvq_f32): Likewise.
1970 (vminnmvq_f64): Likewise.
1971
1972 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1973
1974 * config/aarch64/aarch64-simd.md
1975 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
1976 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
1977 (*aarch64_mul3_elt<mode>): Likewise.
1978 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
1979 (*aarch64_mul3_elt_to_64v2df): Likewise.
1980 (*aarch64_mla_elt<mode>): Likewise.
1981 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1982 (*aarch64_mls_elt<mode>): Likewise.
1983 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1984 (*aarch64_fma4_elt<mode>): Likewise.
1985 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1986 (*aarch64_fma4_elt_to_64v2df): Likewise.
1987 (*aarch64_fnma4_elt<mode>): Likewise.
1988 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1989 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1990 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1991 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1992 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1993 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1994 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1995 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1996 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1997
1998 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
1999
2000 * config/aarch64/aarch64-simd.md
2001 (aarch64_be_checked_get_lane<mode>): New define_expand.
2002 * config/aarch64/aarch64-simd-builtins.def
2003 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
2004 New builtin definition.
2005 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
2006 Use new safe be builtin.
2007
2008 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
2009
2010 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
2011 New define_insn.
2012 (aarch64_be_st1<mode>): Likewise.
2013 (aarch_ld1<VALL:mode>): Define_expand modified.
2014 (aarch_st1<VALL:mode>): Likewise.
2015 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
2016 (UNSPEC_ST1): Likewise.
2017
2018 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
2019
2020 * config/microblaze/microblaze.md: Add trap insn and attribute
2021
2022 2014-01-23 Dodji Seketeli <dodji@redhat.com>
2023
2024 PR preprocessor/58580
2025 * input.h (location_get_source_line): Take an additional line_size
2026 parameter.
2027 (void diagnostics_file_cache_fini): Declare new function.
2028 * input.c (struct fcache): New type.
2029 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
2030 New static constants.
2031 (diagnostic_file_cache_init, total_lines_num)
2032 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
2033 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
2034 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
2035 (get_next_line, read_next_line, goto_next_line, read_line_num):
2036 New static function definitions.
2037 (diagnostic_file_cache_fini): New function.
2038 (location_get_source_line): Take an additional output line_len
2039 parameter. Re-write using lookup_or_add_file_to_cache_tab and
2040 read_line_num.
2041 * diagnostic.c (diagnostic_finish): Call
2042 diagnostic_file_cache_fini.
2043 (adjust_line): Take an additional input parameter for the length
2044 of the line, rather than calculating it with strlen.
2045 (diagnostic_show_locus): Adjust the use of
2046 location_get_source_line and adjust_line with respect to their new
2047 signature. While displaying a line now, do not stop at the first
2048 null byte. Rather, display the zero byte as a space and keep
2049 going until we reach the size of the line.
2050 * Makefile.in: Add vec.o to OBJS-libcommon
2051
2052 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
2053 Ilya Tocar <ilya.tocar@intel.com>
2054
2055 * config/i386/avx512fintrin.h (_mm512_kmov): New.
2056 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
2057 (__builtin_ia32_kmov16): Ditto.
2058 * config/i386/i386.md (UNSPEC_KMOV): New.
2059 (kmovw): Ditto.
2060
2061 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
2062
2063 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
2064 (_mm512_storeu_si512): Ditto.
2065
2066 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
2067
2068 PR target/52125
2069 * rtl.h (get_referenced_operands): Declare.
2070 * recog.c (get_referenced_operands): New function.
2071 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
2072 operands have been referenced when recording LO_SUM references.
2073
2074 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
2075
2076 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
2077
2078 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
2079
2080 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
2081 Enable for generic and recent AMD targets.
2082
2083 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
2084
2085 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
2086 ARG_SIZE note when adjustment was eliminated.
2087
2088 2014-01-22 Jeff Law <law@redhat.com>
2089
2090 PR tree-optimization/59597
2091 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
2092 in file. Accept new argument REGISTERING and use it to modify
2093 dump output appropriately.
2094 (register_jump_thread): Corresponding changes.
2095 (mark_threaded_blocks): Reinstate code to cancel unprofitable
2096 thread paths involving joiner blocks. Add code to dump cancelled
2097 jump threading paths.
2098
2099 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
2100
2101 PR rtl-optimization/59477
2102 * lra-constraints.c (inherit_in_ebb): Process call for living hard
2103 regs. Update reloads_num and potential_reload_hard_regs for all insns.
2104
2105 2014-01-22 Tom Tromey <tromey@redhat.com>
2106
2107 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
2108 PARAMS.
2109 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
2110
2111 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
2112
2113 PR rtl-optimization/59896
2114 * lra-constraints.c (process_alt_operands): Check unused note for
2115 matched operands of insn with no output reloads.
2116
2117 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
2118
2119 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
2120 (mips_move_from_gpr_cost): Likewise.
2121
2122 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
2123
2124 PR rtl-optimization/59858
2125 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
2126 ira_class_hard_regs_num.
2127 (process_alt_operands): Increase reject for dying matched operand.
2128
2129 2014-01-21 Jakub Jelinek <jakub@redhat.com>
2130
2131 PR target/59003
2132 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
2133 smaller than size, perform several stores or loads and stores
2134 at dst + count - size to store or copy all of size bytes, rather
2135 than just last modesize bytes.
2136
2137 2014-01-20 DJ Delorie <dj@redhat.com>
2138
2139 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
2140 that CLOBBERs are REGs before propogating their values.
2141
2142 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
2143
2144 PR middle-end/59789
2145 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
2146 (cgraph_inline_failed_type): New function.
2147 * cgraph.h (DEFCIFCODE): Add type.
2148 (cgraph_inline_failed_type_t): New enum.
2149 (cgraph_inline_failed_type): New prototype.
2150 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
2151 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
2152 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
2153 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
2154 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
2155 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
2156 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
2157 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
2158 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
2159 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
2160 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
2161 OPTIMIZATION_MISMATCH.
2162 * tree-inline.c (expand_call_inline): Emit errors during
2163 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
2164
2165 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
2166
2167 PR target/59685
2168 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
2169 mode attribute in insn output.
2170
2171 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
2172
2173 * output.h (output_constant): Delete.
2174 * varasm.c (output_constant): Make private.
2175
2176 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
2177
2178 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
2179
2180 2014-01-20 Jakub Jelinek <jakub@redhat.com>
2181
2182 PR middle-end/59860
2183 * tree.h (fold_builtin_strcat): New prototype.
2184 * builtins.c (fold_builtin_strcat): No longer static. Add len
2185 argument, if non-NULL, don't call c_strlen. Optimize
2186 directly into __builtin_memcpy instead of __builtin_strcpy.
2187 (fold_builtin_2): Adjust fold_builtin_strcat caller.
2188 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
2189
2190 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
2191
2192 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2193 for SImode_address_operand operands, having only a REG argument.
2194
2195 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
2196
2197 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
2198 loader name using mbig-endian.
2199 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
2200
2201 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
2202
2203 * doc/invoke.texi (-march): Clarify documentation for AArch64.
2204 (-mtune): Likewise.
2205 (-mcpu): Likewise.
2206
2207 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
2208
2209 * config/aarch64/aarch64-protos.h
2210 (aarch64_cannot_change_mode_class_ptr): Declare.
2211 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
2212 aarch64_cannot_change_mode_class_ptr): New.
2213 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
2214 backend hook aarch64_cannot_change_mode_class.
2215
2216 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
2217
2218 * common/config/aarch64/aarch64-common.c
2219 (aarch64_handle_option): Don't handle any option order logic here.
2220 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
2221 selected_cpu, warn on architecture version mismatch.
2222 (aarch64_override_options): Fix parsing order for option strings.
2223
2224 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2225 Iain Sandoe <iain@codesourcery.com>
2226
2227 PR bootstrap/59496
2228 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
2229 warning. Amend comment to reflect current functionality.
2230
2231 2014-01-20 Richard Biener <rguenther@suse.de>
2232
2233 PR middle-end/59860
2234 * builtins.c (fold_builtin_strcat): Remove case better handled
2235 by tree-ssa-strlen.c.
2236
2237 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
2238
2239 * config/aarch64/aarch64.opt
2240 (mcpu, march, mtune): Make case-insensitive.
2241
2242 2014-01-20 Jakub Jelinek <jakub@redhat.com>
2243
2244 PR target/59880
2245 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2246 if operands[1] is a REG or ZERO_EXTEND of a REG.
2247
2248 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
2249
2250 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
2251
2252 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
2253
2254 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
2255 long non-pic millicode calls.
2256
2257 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2258
2259 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
2260
2261 2014-01-19 Kito Cheng <kito@0xlab.org>
2262
2263 * builtins.c (expand_movstr): Check movstr expand done or fail.
2264
2265 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2266 H.J. Lu <hongjiu.lu@intel.com>
2267
2268 PR target/59379
2269 * config/i386/i386.md (*lea<mode>): Zero-extend return register
2270 to DImode for zero-extended addresses.
2271
2272 2014-01-19 Jakub Jelinek <jakub@redhat.com>
2273
2274 PR rtl-optimization/57763
2275 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
2276 on the new indirect jump_insn and increment LABEL_NUSES (label).
2277
2278 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
2279
2280 PR bootstrap/59580
2281 PR bootstrap/59583
2282 * config.gcc (x86_archs): New variable.
2283 (x86_64_archs): Likewise.
2284 (x86_cpus): Likewise.
2285 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
2286 --with-arch/--with-cpu= options.
2287 Support --with-arch=/--with-cpu={nehalem,westmere,
2288 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
2289
2290 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2291
2292 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
2293 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
2294
2295 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2296
2297 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
2298
2299 2014-01-18 Jakub Jelinek <jakub@redhat.com>
2300
2301 PR target/58944
2302 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
2303 clear cpp_get_options (parse_in)->warn_unused_macros for
2304 ix86_target_macros_internal with cpp_define.
2305
2306 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
2307
2308 * jump.c (delete_related_insns): Keep (use (insn))s.
2309 * reorg.c (redundant_insn): Check for barriers too.
2310
2311 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2312
2313 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
2314
2315 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
2316
2317 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
2318 call to $$dyncall when TARGET_LONG_CALLS is true.
2319
2320 2014-01-17 Jeff Law <law@redhat.com>
2321
2322 * ree.c (combine_set_extension): Temporarily disable test for
2323 changing number of hard registers.
2324
2325 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
2326
2327 PR middle-end/58125
2328 * ipa-inline-analysis.c (inline_free_summary):
2329 Do not free summary of aliases.
2330
2331 2014-01-17 Jakub Jelinek <jakub@redhat.com>
2332
2333 PR middle-end/59706
2334 * gimplify.c (gimplify_expr): Use create_tmp_var
2335 instead of create_tmp_var_raw. If cond doesn't have
2336 integral type, don't add the IFN_ANNOTATE builtin at all.
2337
2338 2014-01-17 Martin Jambor <mjambor@suse.cz>
2339
2340 PR ipa/59736
2341 * ipa-cp.c (prev_edge_clone): New variable.
2342 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
2343 Also resize prev_edge_clone vector.
2344 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
2345 (ipcp_edge_removal_hook): New function.
2346 (ipcp_driver): Register ipcp_edge_removal_hook.
2347
2348 2014-01-17 Andrew Pinski <apinski@cavium.com>
2349 Steve Ellcey <sellcey@mips.com>
2350
2351 PR target/59462
2352 * config/mips/mips.c (mips_print_operand): Check operand mode instead
2353 of operator mode.
2354
2355 2014-01-17 Jeff Law <law@redhat.com>
2356
2357 PR middle-end/57904
2358 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
2359 so that pass_ccp runs first.
2360
2361 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2362
2363 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
2364 (ix86_adjust_cost): Use !TARGET_XXX.
2365 (do_reorder_for_imul): Likewise.
2366 (swap_top_of_ready_list): Likewise.
2367 (ix86_sched_reorder): Likewise.
2368
2369 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2370
2371 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2372 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
2373 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
2374 (intel_memset): New. Duplicate slm_memset.
2375 (intel_cost): New. Duplicate slm_cost.
2376 (m_INTEL): New macro.
2377 (processor_target_table): Add "intel".
2378 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
2379 with PROCESSOR_INTEL for "intel".
2380 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
2381 PROCESSOR_SILVERMONT.
2382 (ix86_issue_rate): Likewise.
2383 (ix86_adjust_cost): Likewise.
2384 (ia32_multipass_dfa_lookahead): Likewise.
2385 (swap_top_of_ready_list): Likewise.
2386 (ix86_sched_reorder): Likewise.
2387 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
2388 instead of TARGET_OPT_AGU.
2389 * config/i386/i386.h (TARGET_INTEL): New.
2390 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
2391 (processor_type): Add PROCESSOR_INTEL.
2392 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
2393 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
2394
2395 2014-01-17 Marek Polacek <polacek@redhat.com>
2396
2397 PR c/58346
2398 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
2399 size is zero.
2400
2401 2014-01-17 Richard Biener <rguenther@suse.de>
2402
2403 PR tree-optimization/46590
2404 * opts.c (default_options_table): Add entries for
2405 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
2406 all enabled at -O1 but not for -Og.
2407 * common.opt (fbranch-count-reg): Remove Init(1).
2408 (fmove-loop-invariants): Likewise.
2409 (ftree-pta): Likewise.
2410
2411 2014-01-17 Jakub Jelinek <jakub@redhat.com>
2412
2413 * config/i386/i386.c (ix86_data_alignment): For compatibility with
2414 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
2415 decls to at least the GCC 4.8 used alignments.
2416
2417 PR fortran/59440
2418 * tree-nested.c (convert_nonlocal_reference_stmt,
2419 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
2420 of GIMPLE_BIND stmts, adjust associated decls.
2421
2422 2014-01-17 Richard Biener <rguenther@suse.de>
2423
2424 PR tree-optimization/46590
2425 * vec.h (vec<>::bseach): New member function implementing
2426 binary search according to C89 bsearch.
2427 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
2428 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
2429 bitmap pointer again. Make accesses_in_loop a flat array.
2430 (mem_ref_obstack): New global.
2431 (outermost_indep_loop): Adjust for mem_ref->stored changes.
2432 (mark_ref_stored): Likewise.
2433 (ref_indep_loop_p_2): Likewise.
2434 (set_ref_stored_in_loop): New helper function.
2435 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
2436 (memref_free): Adjust.
2437 (record_mem_ref_loc): Simplify.
2438 (gather_mem_refs_stmt): Adjust.
2439 (sort_locs_in_loop_postorder_cmp): New function.
2440 (analyze_memory_references): Sort accesses_in_loop after
2441 loop postorder number.
2442 (find_ref_loc_in_loop_cmp): New function.
2443 (for_all_locs_in_loop): Find relevant cluster of locs in
2444 accesses_in_loop and iterate without recursion.
2445 (execute_sm): Avoid uninit warning.
2446 (struct ref_always_accessed): Simplify.
2447 (ref_always_accessed::operator ()): Likewise.
2448 (ref_always_accessed_p): Likewise.
2449 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
2450 loop postorder numbers here.
2451 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
2452 numbers.
2453
2454 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
2455
2456 PR c++/57945
2457 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
2458 on decls for which assemble_alias has been called.
2459
2460 2014-01-17 Nick Clifton <nickc@redhat.com>
2461
2462 * config/msp430/msp430.opt: (mcpu): New option.
2463 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
2464 (msp430_option_override): Parse target_cpu. If the MCU name
2465 matches a generic string, clear target_mcu.
2466 (msp430_attr): Allow numeric interrupt values up to 63.
2467 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
2468 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
2469 option.
2470 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
2471 Add mcpu matches.
2472 * config/msp430/msp430.md (popm): Use %J rather than %I.
2473 (addsi3): Use msp430_nonimmediate_operand for operand 2.
2474 (addhi_cy_i): Use immediate_operand for operand 2.
2475 * doc/invoke.texi: Document -mcpu option.
2476
2477 2014-01-17 Richard Biener <rguenther@suse.de>
2478
2479 PR rtl-optimization/38518
2480 * df.h (df_analyze_loop): Declare.
2481 * df-core.c: Include cfgloop.h.
2482 (df_analyze_1): Split out main part of df_analyze.
2483 (df_analyze): Adjust.
2484 (loop_inverted_post_order_compute): New function.
2485 (loop_post_order_compute): Likewise.
2486 (df_analyze_loop): New function avoiding whole-function
2487 postorder computes.
2488 * loop-invariant.c (find_defs): Use df_analyze_loop.
2489 (find_invariants): Adjust.
2490 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
2491
2492 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
2493
2494 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
2495 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
2496
2497 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
2498
2499 * ipa-ref.c (ipa_remove_stmt_references): Fix references
2500 traversal when removing references.
2501
2502 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
2503
2504 PR ipa/59775
2505 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
2506
2507 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
2508
2509 PR middle-end/56791
2510 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
2511 pushing a reload for an autoinc when we had previously reloaded an
2512 inner part of the address.
2513
2514 2014-01-16 Jakub Jelinek <jakub@redhat.com>
2515
2516 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
2517 field.
2518 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
2519 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
2520 when not giving up or versioning for alias only because of
2521 loop->safelen.
2522 (vect_analyze_data_ref_dependences): Set to true.
2523 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
2524 is a GIMPLE_PHI.
2525 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
2526 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
2527 to the condition.
2528
2529 PR middle-end/58344
2530 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
2531
2532 PR target/59839
2533 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
2534 operand 0 predicate for gathers, use a new pseudo as subtarget.
2535
2536 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
2537
2538 PR middle-end/59609
2539 * lra-constraints.c (process_alt_operands): Add printing debug info.
2540 Check absence of input/output reloads for matched operands too.
2541
2542 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
2543
2544 PR rtl-optimization/59835
2545 * ira.c (ira_init_register_move_cost): Increase cost for
2546 impossible modes.
2547
2548 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
2549
2550 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
2551
2552 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
2553
2554 PR target/59780
2555 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
2556 non-register objects. Use gen_(high/low)part more consistently.
2557 Fix assertions.
2558
2559 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2560
2561 PR target/59844
2562 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
2563 endian support, remove tests for WORDS_BIG_ENDIAN.
2564 (p8_mfvsrd_3_<mode>): Likewise.
2565 (reload_gpr_from_vsx<mode>): Likewise.
2566 (reload_gpr_from_vsxsf): Likewise.
2567 (p8_mfvsrd_4_disf): Likewise.
2568
2569 2014-01-16 Richard Biener <rguenther@suse.de>
2570
2571 PR rtl-optimization/46590
2572 * lcm.c (compute_antinout_edge): Use postorder iteration.
2573 (compute_laterin): Use inverted postorder iteration.
2574
2575 2014-01-16 Nick Clifton <nickc@redhat.com>
2576
2577 PR middle-end/28865
2578 * varasm.c (output_constant): Return the number of bytes actually
2579 emitted.
2580 (output_constructor_array_range): Update the field size with the
2581 number of bytes emitted by output_constant.
2582 (output_constructor_regular_field): Likewise. Also do not
2583 complain if the total number of bytes emitted is now greater
2584 than the expected fieldpos.
2585 * output.h (output_constant): Update prototype and descriptive comment.
2586
2587 2014-01-16 Marek Polacek <polacek@redhat.com>
2588
2589 PR middle-end/59827
2590 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
2591 it is error_mark_node.
2592
2593 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
2594
2595 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
2596 VALID_AVX256_REG_OR_OI_MODE.
2597
2598 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
2599
2600 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
2601 current procedure should be profiled.
2602
2603 2014-01-15 Andrew Pinski <apinski@cavium.com>
2604
2605 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
2606 of moving from/to the STACK_REG register class.
2607
2608 2014-01-15 Richard Henderson <rth@redhat.com>
2609
2610 PR debug/54694
2611 * reginfo.c (global_regs_decl): Globalize.
2612 * rtl.h (global_regs_decl): Declare.
2613 * ira.c (do_reload): Diagnose frame_pointer_needed and it
2614 reserved via global_regs.
2615
2616 2014-01-15 Teresa Johnson <tejohnson@google.com>
2617
2618 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
2619
2620 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2621
2622 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
2623 and vmulosh rather than call gen_vec_widen_smult_*.
2624 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
2625 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
2626 (vec_widen_smult_even_v16qi): Likewise.
2627 (vec_widen_umult_even_v8hi): Likewise.
2628 (vec_widen_smult_even_v8hi): Likewise.
2629 (vec_widen_umult_odd_v16qi): Likewise.
2630 (vec_widen_smult_odd_v16qi): Likewise.
2631 (vec_widen_umult_odd_v8hi): Likewise.
2632 (vec_widen_smult_odd_v8hi): Likewise.
2633 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
2634 vmuloub rather than call gen_vec_widen_umult_*.
2635 (vec_widen_umult_lo_v16qi): Likewise.
2636 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
2637 vmulosb rather than call gen_vec_widen_smult_*.
2638 (vec_widen_smult_lo_v16qi): Likewise.
2639 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
2640 rather than call gen_vec_widen_umult_*.
2641 (vec_widen_umult_lo_v8hi): Likewise.
2642 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
2643 rather than call gen_vec_widen_smult_*.
2644 (vec_widen_smult_lo_v8hi): Likewise.
2645
2646 2014-01-15 Jeff Law <law@redhat.com>
2647
2648 PR tree-optimization/59747
2649 * ree.c (find_and_remove_re): Properly handle case where a second
2650 eliminated extension requires widening a copy created for elimination
2651 of a prior extension.
2652 (combine_set_extension): Ensure that the number of hard regs needed
2653 for a destination register does not change when we widen it.
2654
2655 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
2656
2657 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
2658 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
2659 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
2660 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
2661 (avr-*-rtems*): Likewise.
2662 (bfin*-rtems*): Likewise.
2663 (moxie-*-rtems*): Likewise.
2664 (h8300-*-rtems*): Likewise.
2665 (i[34567]86-*-rtems*): Likewise.
2666 (lm32-*-rtems*): Likewise.
2667 (m32r-*-rtems*): Likewise.
2668 (m68k-*-rtems*): Likewise.
2669 (microblaze*-*-rtems*): Likewise.
2670 (mips*-*-rtems*): Likewise.
2671 (powerpc-*-rtems*): Likewise.
2672 (sh-*-rtems*): Likewise.
2673 (sparc-*-rtems*): Likewise.
2674 (sparc64-*-rtems*): Likewise.
2675 (v850-*-rtems*): Likewise.
2676 (m32c-*-rtems*): Likewise.
2677
2678 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
2679
2680 PR rtl-optimization/59511
2681 * ira.c (ira_init_register_move_cost): Use memory costs for some
2682 cases of register move cost calculations.
2683 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
2684 instead of BB frequency.
2685 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
2686 * lra-assigns.c (find_hard_regno_for): Ditto.
2687
2688 2014-01-15 Richard Biener <rguenther@suse.de>
2689
2690 PR tree-optimization/59822
2691 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
2692 (vectorizable_load): Use it to hoist defs of uses of invariant
2693 loads out of the loop.
2694
2695 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
2696 Kugan Vivekanandarajah <kuganv@linaro.org>
2697
2698 PR target/59695
2699 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
2700 truncation.
2701
2702 2014-01-15 Richard Biener <rguenther@suse.de>
2703
2704 PR rtl-optimization/59802
2705 * lcm.c (compute_available): Use inverted postorder to seed
2706 the initial worklist.
2707
2708 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2709
2710 PR target/59803
2711 * config/s390/s390.c (s390_preferred_reload_class): Don't return
2712 ADDR_REGS for invalid symrefs in non-PIC code.
2713
2714 2014-01-15 Jakub Jelinek <jakub@redhat.com>
2715
2716 PR other/58712
2717 * builtins.c (determine_block_size): Initialize *probable_max_size
2718 even if len_rtx is CONST_INT.
2719
2720 2014-01-14 Andrew Pinski <apinski@cavium.com>
2721
2722 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
2723 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
2724 (cortexa53_tunings): Likewise.
2725 (aarch64_sched_issue_rate): New function.
2726 (TARGET_SCHED_ISSUE_RATE): Define.
2727
2728 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
2729
2730 * ira-costs.c (find_costs_and_classes): Add missed
2731 ira_init_register_move_cost_if_necessary.
2732
2733 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
2734
2735 PR target/59787
2736 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
2737
2738 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
2739
2740 PR target/59794
2741 * config/i386/i386.c (type_natural_mode): Add a bool parameter
2742 to indicate if type is used for function return value. Warn ABI
2743 change if the vector mode isn't available for function return value.
2744 (ix86_function_arg_advance): Pass false to type_natural_mode.
2745 (ix86_function_arg): Likewise.
2746 (ix86_gimplify_va_arg): Likewise.
2747 (function_arg_32): Don't warn ABI change.
2748 (ix86_function_value): Pass true to type_natural_mode.
2749 (ix86_return_in_memory): Likewise.
2750 (ix86_struct_value_rtx): Removed.
2751 (TARGET_STRUCT_VALUE_RTX): Likewise.
2752
2753 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2754
2755 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
2756 converting a conditional jump into a conditional return.
2757
2758 2014-01-14 Richard Biener <rguenther@suse.de>
2759
2760 PR tree-optimization/58921
2761 PR tree-optimization/59006
2762 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
2763 hoisting invariant stmts.
2764 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
2765 invariant loads on the preheader edge if possible.
2766
2767 2014-01-14 Joey Ye <joey.ye@arm.com>
2768
2769 * doc/plugin.texi (Building GCC plugins): Update to C++.
2770
2771 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
2772
2773 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
2774 (_mm_rcp28_round_ss): Ditto.
2775 (_mm_rsqrt28_round_sd): Ditto.
2776 (_mm_rsqrt28_round_ss): Ditto.
2777 (_mm_rcp28_sd): Ditto.
2778 (_mm_rcp28_ss): Ditto.
2779 (_mm_rsqrt28_sd): Ditto.
2780 (_mm_rsqrt28_ss): Ditto.
2781 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
2782 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
2783 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
2784 (IX86_BUILTIN_RCP28SD): Ditto.
2785 (IX86_BUILTIN_RCP28SS): Ditto.
2786 (IX86_BUILTIN_RSQRT28SD): Ditto.
2787 (IX86_BUILTIN_RSQRT28SS): Ditto.
2788 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
2789 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
2790 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
2791 (ix86_expand_special_args_builtin): Expand new FTYPE.
2792 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
2793 (srcp14<mode>): Make insn unary.
2794 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
2795 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
2796 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
2797 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
2798 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
2799 Fix rounding: make it SAE only.
2800 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
2801 Ditto.
2802 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
2803 Ditto.
2804 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
2805 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
2806 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
2807 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
2808 (round_saeonly_mask_scalar_operand4): Ditto.
2809 (round_saeonly_mask_scalar_op3): Ditto.
2810 (round_saeonly_mask_scalar_op4): Ditto.
2811
2812 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2813
2814 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
2815 Implement -maltivec=be for vec_insert and vec_extract.
2816
2817 2014-01-10 DJ Delorie <dj@redhat.com>
2818
2819 * config/msp430/msp430.md (call_internal): Don't allow memory
2820 references with SP as the base register.
2821 (call_value_internal): Likewise.
2822 * config/msp430/constraints.md (Yc): New. For memory references
2823 that don't use SP as a base register.
2824
2825 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
2826 "an integer without a # prefix"
2827 * config/msp430/msp430.md (epilogue_helper): Use it.
2828
2829 2014-01-13 Jakub Jelinek <jakub@redhat.com>
2830
2831 PR target/59617
2832 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
2833 AVX512F gather builtins.
2834 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
2835 on gather decls with INTEGER_TYPE masktype.
2836 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
2837 directly into the builtin rather than hoisting it before loop.
2838
2839 PR tree-optimization/59387
2840 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
2841 (scev_const_prop): If folded_casts and type has undefined overflow,
2842 use force_gimple_operand instead of force_gimple_operand_gsi and
2843 for each added stmt if it is assign with
2844 arith_code_with_undefined_signed_overflow, call
2845 rewrite_to_defined_overflow.
2846 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
2847 gimple-fold.h instead.
2848 (arith_code_with_undefined_signed_overflow,
2849 rewrite_to_defined_overflow): Moved to ...
2850 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
2851 rewrite_to_defined_overflow): ... here. No longer static.
2852 Include gimplify-me.h.
2853 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
2854 rewrite_to_defined_overflow): New prototypes.
2855
2856 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2857
2858 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
2859
2860 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
2861
2862 * builtins.c (get_object_alignment_2): Minor tweak.
2863 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
2864
2865 2014-01-13 Christian Bruel <christian.bruel@st.com>
2866
2867 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
2868 optimized non constant lengths.
2869
2870 2014-01-13 Jakub Jelinek <jakub@redhat.com>
2871
2872 PR libgomp/59194
2873 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
2874 load as __atomic_load_N if possible.
2875
2876 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
2877
2878 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
2879 target parameter.
2880 (rs6000_expand_builtin): Adjust call.
2881
2882 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
2883
2884 PR target/58115
2885 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
2886 * config/rs6000/rs6000.c: Include target-globals.h.
2887 (rs6000_set_current_function): Instead of doing target_reinit
2888 unconditionally, use save_target_globals_default_opts and
2889 restore_target_globals.
2890
2891 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
2892 FPSCR.
2893 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
2894 (rs6000_expand_builtin): Handle mffs and mtfsf.
2895 (rs6000_init_builtins): Define mffs and mtfsf.
2896 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
2897 (rs6000_mffs): New pattern.
2898 (rs6000_mtfsf): New pattern.
2899
2900 2014-01-11 Bin Cheng <bin.cheng@arm.com>
2901
2902 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
2903 Start narrowing with START. Apply candidate-use pair
2904 and check overall cost in narrowing.
2905 (iv_ca_prune): Pass new argument.
2906
2907 2014-01-10 Jeff Law <law@redhat.com>
2908
2909 PR middle-end/59743
2910 * ree.c (combine_reaching_defs): Ensure the defining statement
2911 occurs before the extension when optimizing extensions with
2912 different source and destination hard registers.
2913
2914 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
2915
2916 PR ipa/58585
2917 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
2918 vtables into the type inheritance graph.
2919
2920 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2921
2922 PR rtl-optimization/59754
2923 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
2924 modes in the REGNO != REGNO case.
2925
2926 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2927
2928 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
2929
2930 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2931
2932 PR tree-optimization/59745
2933 * tree-predcom.c (tree_predictive_commoning_loop): Call
2934 free_affine_expand_cache if giving up because components is NULL.
2935
2936 * target-globals.c (save_target_globals): Allocate < 4KB structs using
2937 GC in payload of target_globals struct instead of allocating them on
2938 the heap and the larger structs separately using GC.
2939 * target-globals.h (struct target_globals): Make regs, hard_regs,
2940 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
2941 of GTY((skip)) and change type to void *.
2942 (reset_target_globals): Cast loads from those fields to corresponding
2943 types.
2944
2945 2014-01-10 Steve Ellcey <sellcey@mips.com>
2946
2947 PR plugins/59335
2948 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
2949 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
2950 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
2951
2952 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
2953
2954 PR target/59744
2955 * aarch64-modes.def (CC_Zmode): New flags mode.
2956 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
2957 represents an equality.
2958 (aarch64_get_condition_code): Handle CC_Zmode.
2959 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
2960
2961 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2962
2963 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
2964 extraction in good case.
2965
2966 2014-01-10 Richard Biener <rguenther@suse.de>
2967
2968 PR tree-optimization/59374
2969 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
2970 checking after SLP discovery. Mark stmts not participating
2971 in any SLP instance properly.
2972
2973 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2974
2975 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
2976 when handling a SET rtx.
2977
2978 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2979
2980 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
2981 (cortex-a57): Likewise.
2982 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
2983
2984 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2985
2986 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
2987 non-iwmmxt builtins.
2988
2989 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
2990
2991 PR ipa/58252
2992 PR ipa/59226
2993 * ipa-devirt.c record_target_from_binfo): Take as argument
2994 stack of binfos and lookup matching one for virtual inheritance.
2995 (possible_polymorphic_call_targets_1): Update.
2996
2997 2014-01-10 Huacai Chen <chenhc@lemote.com>
2998
2999 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
3000 kernel strings for Loongson-2E/2F/3A.
3001
3002 2014-01-10 Jakub Jelinek <jakub@redhat.com>
3003
3004 PR middle-end/59670
3005 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
3006 is_gimple_call before calling gimple_call_internal_p.
3007
3008 2014-01-09 Steve Ellcey <sellcey@mips.com>
3009
3010 * Makefile.in (TREE_FLOW_H): Remove.
3011 (TREE_SSA_H): Add file names from tree-flow.h.
3012 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
3013 * tree.h: Remove tree-flow.h reference.
3014 * hash-table.h: Remove tree-flow.h reference.
3015 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
3016 reference with tree-ssa-loop.h.
3017
3018 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3019
3020 * doc/invoke.texi: Add -maltivec={be,le} options, and document
3021 default element-order behavior for -maltivec.
3022 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
3023 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
3024 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
3025 when targeting big endian, at least for now.
3026 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
3027
3028 2014-01-09 Jakub Jelinek <jakub@redhat.com>
3029
3030 PR middle-end/47735
3031 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
3032 var satisfies use_register_for_decl, just take into account type
3033 alignment, rather than decl alignment.
3034
3035 PR tree-optimization/59622
3036 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
3037 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
3038 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
3039 Don't devirtualize for inplace at all. For targets.length () == 1,
3040 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
3041
3042 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
3043
3044 * config/i386/i386.md (cpu): Remove the unused btver1.
3045
3046 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
3047
3048 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
3049
3050 2014-01-09 Jakub Jelinek <jakub@redhat.com>
3051
3052 PR target/58115
3053 * tree-core.h (struct target_globals): New forward declaration.
3054 (struct tree_target_option): Add globals field.
3055 * tree.h (TREE_TARGET_GLOBALS): Define.
3056 (prepare_target_option_nodes_for_pch): New prototype.
3057 * target-globals.h (struct target_globals): Define even if
3058 !SWITCHABLE_TARGET.
3059 * tree.c (prepare_target_option_node_for_pch,
3060 prepare_target_option_nodes_for_pch): New functions.
3061 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
3062 * config/i386/i386.c: Include target-globals.h.
3063 (ix86_set_current_function): Instead of doing target_reinit
3064 unconditionally, use save_target_globals_default_opts and
3065 restore_target_globals.
3066
3067 2014-01-09 Richard Biener <rguenther@suse.de>
3068
3069 PR tree-optimization/59715
3070 * tree-cfg.h (split_critical_edges): Declare.
3071 * tree-cfg.c (split_critical_edges): Export.
3072 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
3073
3074 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
3075
3076 * cfgexpand.c (expand_stack_vars): Optionally disable
3077 asan stack protection.
3078 (expand_used_vars): Likewise.
3079 (partition_stack_vars): Likewise.
3080 * asan.c (asan_emit_stack_protection): Optionally disable
3081 after return stack usage.
3082 (instrument_derefs): Optionally disable memory access instrumentation.
3083 (instrument_builtin_call): Likewise.
3084 (instrument_strlen_call): Likewise.
3085 (asan_protect_global): Optionally disable global variables protection.
3086 * doc/invoke.texi: Added doc for new options.
3087 * params.def: Added new options.
3088 * params.h: Likewise.
3089
3090 2014-01-09 Jakub Jelinek <jakub@redhat.com>
3091
3092 PR rtl-optimization/59724
3093 * ifcvt.c (cond_exec_process_if_block): Don't call
3094 flow_find_head_matching_sequence with 0 longest_match.
3095 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
3096 non-active insns if !stop_after.
3097 (try_head_merge_bb): Revert 2014-01-07 changes.
3098
3099 2014-01-08 Jeff Law <law@redhat.com>
3100
3101 * ree.c (get_sub_rtx): New function, extracted from...
3102 (merge_def_and_ext): Here.
3103 (combine_reaching_defs): Use get_sub_rtx.
3104
3105 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
3106
3107 * cgraph.h (varpool_variable_node): Do not choke on null node.
3108
3109 2014-01-08 Catherine Moore <clm@codesourcery.com>
3110
3111 * config/mips/mips.md (simple_return): Attempt to use JRC
3112 for microMIPS.
3113 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
3114
3115 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
3116
3117 PR rtl-optimization/59137
3118 * reorg.c (steal_delay_list_from_target): Call update_block for
3119 elided insns.
3120 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
3121
3122 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3123
3124 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
3125 two duplicate entries.
3126
3127 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
3128
3129 Revert:
3130 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
3131
3132 * config/mips/mips.c (mips_truncated_op_cost): New function.
3133 (mips_rtx_costs): Adjust test for BADDU.
3134 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
3135
3136 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
3137
3138 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
3139 (*baddu_si): ...this new pattern.
3140
3141 2014-01-08 Jakub Jelinek <jakub@redhat.com>
3142
3143 PR ipa/59722
3144 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
3145
3146 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
3147
3148 PR middle-end/57748
3149 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
3150 inner_reference_p.
3151 (expand_expr, expand_normal): Adjust.
3152 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
3153 inner_reference_p. Use inner_reference_p to expand inner references.
3154 (store_expr): Adjust.
3155 * cfgexpand.c (expand_call_stmt): Adjust.
3156
3157 2014-01-08 Rong Xu <xur@google.com>
3158
3159 * gcov-io.c (gcov_var): Move from gcov-io.h.
3160 (gcov_position): Ditto.
3161 (gcov_is_error): Ditto.
3162 (gcov_rewrite): Ditto.
3163 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
3164 only part to libgcc/libgcov.h.
3165
3166 2014-01-08 Marek Polacek <polacek@redhat.com>
3167
3168 PR middle-end/59669
3169 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
3170
3171 2014-01-08 Marek Polacek <polacek@redhat.com>
3172
3173 PR sanitizer/59667
3174 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
3175
3176 2014-01-08 Jakub Jelinek <jakub@redhat.com>
3177
3178 PR rtl-optimization/59649
3179 * stor-layout.c (get_mode_bounds): For BImode return
3180 0 and STORE_FLAG_VALUE.
3181
3182 2014-01-08 Richard Biener <rguenther@suse.de>
3183
3184 PR middle-end/59630
3185 * gimple.h (is_gimple_builtin_call): Remove.
3186 (gimple_builtin_call_types_compatible_p): New.
3187 (gimple_call_builtin_p): New overload.
3188 * gimple.c (is_gimple_builtin_call): Remove.
3189 (validate_call): Rename to ...
3190 (gimple_builtin_call_types_compatible_p): ... this and export. Also
3191 check return types.
3192 (validate_type): New static function.
3193 (gimple_call_builtin_p): New overload and adjust.
3194 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
3195 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
3196 (gimple_fold_stmt_to_constant_1): Likewise.
3197 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
3198
3199 2014-01-08 Richard Biener <rguenther@suse.de>
3200
3201 PR middle-end/59471
3202 * gimplify.c (gimplify_expr): Gimplify register-register type
3203 VIEW_CONVERT_EXPRs to separate stmts.
3204
3205 2014-01-07 Jeff Law <law@redhat.com>
3206
3207 PR middle-end/53623
3208 * ree.c (combine_set_extension): Handle case where source
3209 and destination registers in an extension insn are different.
3210 (combine_reaching_defs): Allow source and destination registers
3211 in extension to be different under limited circumstances.
3212 (add_removable_extension): Remove restriction that the
3213 source and destination registers in the extension are the same.
3214 (find_and_remove_re): Emit a copy from the extension's
3215 destination to its source after the defining insn if
3216 the source and destination registers are different.
3217
3218 PR middle-end/59285
3219 * ifcvt.c (merge_if_block): If we are merging a block with more than
3220 one successor with a block with no successors, remove any BARRIER
3221 after the second block.
3222
3223 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
3224
3225 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
3226
3227 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
3228
3229 PR target/59652
3230 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
3231 for 14-bit register offsets when INT14_OK_STRICT is false.
3232
3233 2014-01-07 Roland Stigge <stigge@antcom.de>
3234 Michael Meissner <meissner@linux.vnet.ibm.com>
3235
3236 PR 57386/target
3237 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
3238 Only check TFmode for SPE constants. Don't check TImode or TDmode.
3239
3240 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
3241
3242 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
3243 -mcpu.
3244
3245 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
3246
3247 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
3248 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
3249 rtx is const0_rtx or not.
3250
3251 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
3252
3253 PR target/58115
3254 * target-globals.c (save_target_globals): Remove this_fn_optab
3255 handling.
3256 * toplev.c: Include optabs.h.
3257 (target_reinit): Temporarily restore the global options if another
3258 set of options are in force.
3259
3260 2014-01-07 Jakub Jelinek <jakub@redhat.com>
3261
3262 PR rtl-optimization/58668
3263 * cfgcleanup.c (flow_find_cross_jump): Don't count
3264 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
3265 to determine what is counted.
3266 (flow_find_head_matching_sequence): Use active_insn_p to determine
3267 what is counted.
3268 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
3269 counting change.
3270 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
3271 determine what is counted.
3272
3273 PR tree-optimization/59643
3274 * tree-predcom.c (split_data_refs_to_components): If one dr is
3275 read and one write, determine_offset fails and the write isn't
3276 in the bad component, just put the read into the bad component.
3277
3278 2014-01-07 Mike Stump <mikestump@comcast.net>
3279 Jakub Jelinek <jakub@redhat.com>
3280
3281 PR pch/59436
3282 * tree-core.h (struct tree_optimization_option): Change optabs
3283 type from unsigned char * to void *.
3284 * optabs.c (init_tree_optimization_optabs): Adjust
3285 TREE_OPTIMIZATION_OPTABS initialization.
3286
3287 2014-01-06 Jakub Jelinek <jakub@redhat.com>
3288
3289 PR target/59644
3290 * config/i386/i386.h (struct machine_function): Add
3291 no_drap_save_restore field.
3292 * config/i386/i386.c (ix86_save_reg): Use
3293 !cfun->machine->no_drap_save_restore instead of
3294 crtl->stack_realign_needed.
3295 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
3296 this function clears frame_pointer_needed. Set
3297 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
3298 and DRAP reg is needed.
3299
3300 2014-01-06 Marek Polacek <polacek@redhat.com>
3301
3302 PR c/57773
3303 * doc/implement-c.texi: Mention that other integer types are
3304 permitted as bit-field types in strictly conforming mode.
3305
3306 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
3307
3308 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
3309 is newly allocated.
3310
3311 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
3312
3313 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
3314
3315 2014-01-06 Martin Jambor <mjambor@suse.cz>
3316
3317 PR ipa/59008
3318 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
3319 to int.
3320 * ipa-prop.c (ipa_print_node_params): Fix indentation.
3321
3322 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
3323
3324 PR debug/59350
3325 PR debug/59510
3326 * var-tracking.c (add_stores): Preserve the value of the source even if
3327 we don't record the store.
3328
3329 2014-01-06 Terry Guo <terry.guo@arm.com>
3330
3331 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
3332
3333 2014-01-05 Iain Sandoe <iain@codesourcery.com>
3334
3335 PR bootstrap/59541
3336 * config/darwin.c (darwin_function_section): Adjust return values to
3337 correspond to optimisation changes made in r206070.
3338
3339 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
3340
3341 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
3342 from prefetch_block tune setting.
3343 (nocona_cost): Correct size of prefetch block to 64.
3344
3345 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
3346
3347 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
3348 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
3349 used to save the static chain register in the computation of the offset
3350 from which the FP registers need to be restored.
3351
3352 2014-01-04 Jakub Jelinek <jakub@redhat.com>
3353
3354 PR tree-optimization/59519
3355 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
3356 ICE if get_current_def (current_new_name) is already non-NULL, as long
3357 as it is a phi result of some other phi in *new_exit_bb that has
3358 the same argument.
3359
3360 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
3361 or vmovdqu* for misaligned_operand.
3362 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
3363 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
3364 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
3365 aligned_mem for AVX512F masked aligned load and store builtins and for
3366 non-temporal moves.
3367
3368 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
3369
3370 PR tree-optimization/59651
3371 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
3372 Address range for negative step should be added by TYPE_SIZE_UNIT.
3373
3374 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
3375
3376 * config/m68k/m68k.c (handle_move_double): Handle pushes with
3377 overlapping registers also for registers other than the stack pointer.
3378
3379 2014-01-03 Marek Polacek <polacek@redhat.com>
3380
3381 PR other/59661
3382 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
3383 __builtin_FILE.
3384
3385 2014-01-03 Jakub Jelinek <jakub@redhat.com>
3386
3387 PR target/59625
3388 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
3389 asm goto as jump.
3390
3391 * config/i386/i386.md (MODE_SIZE): New mode attribute.
3392 (push splitter): Use <P:MODE_SIZE> instead of
3393 GET_MODE_SIZE (<P:MODE>mode).
3394 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
3395 (mov -1, reg peephole2): Likewise.
3396 * config/i386/sse.md (*mov<mode>_internal,
3397 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
3398 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
3399 *<code><mode>3, *andnot<mode>3<mask_name>,
3400 <mask_codefor><code><mode>3<mask_name>): Likewise.
3401 * config/i386/subst.md (mask_mode512bit_condition,
3402 sd_mask_mode512bit_condition): Likewise.
3403
3404 2014-01-02 Xinliang David Li <davidxl@google.com>
3405
3406 PR tree-optimization/59303
3407 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
3408 (dump_predicates): Better output format.
3409 (pred_equal_p): New function.
3410 (is_neq_relop_p): Ditto.
3411 (is_neq_zero_form_p): Ditto.
3412 (pred_expr_equal_p): Ditto.
3413 (pred_neg_p): Ditto.
3414 (simplify_pred): Ditto.
3415 (simplify_preds_2): Ditto.
3416 (simplify_preds_3): Ditto.
3417 (simplify_preds_4): Ditto.
3418 (simplify_preds): Ditto.
3419 (push_pred): Ditto.
3420 (push_to_worklist): Ditto.
3421 (get_pred_info_from_cmp): Ditto.
3422 (is_degenerated_phi): Ditto.
3423 (normalize_one_pred_1): Ditto.
3424 (normalize_one_pred): Ditto.
3425 (normalize_one_pred_chain): Ditto.
3426 (normalize_preds): Ditto.
3427 (normalize_cond_1): Remove function.
3428 (normalize_cond): Ditto.
3429 (is_gcond_subset_of): Ditto.
3430 (is_subset_of_any): Ditto.
3431 (is_or_set_subset_of): Ditto.
3432 (is_and_set_subset_of): Ditto.
3433 (is_norm_cond_subset_of): Ditto.
3434 (pred_chain_length_cmp): Ditto.
3435 (convert_control_dep_chain_into_preds): Type change.
3436 (find_predicates): Ditto.
3437 (find_def_preds): Ditto.
3438 (destroy_predicates_vecs): Ditto.
3439 (find_matching_predicates_in_rest_chains): Ditto.
3440 (use_pred_not_overlap_with_undef_path_pred): Ditto.
3441 (is_pred_expr_subset): Ditto.
3442 (is_pred_chain_subset_of): Ditto.
3443 (is_included_in): Ditto.
3444 (is_superset_of): Ditto.
3445
3446 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3447
3448 Update copyright years.
3449
3450 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3451
3452 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
3453 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
3454 config/arc/arc.md, config/arc/arc.opt,
3455 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
3456 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
3457 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
3458 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
3459 config/linux-protos.h, config/linux.c, config/winnt-c.c,
3460 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
3461 vtable-verify.c, vtable-verify.h: Use the standard form for the
3462 copyright notice.
3463
3464 2014-01-02 Tobias Burnus <burnus@net-b.de>
3465
3466 * gcc.c (process_command): Update copyright notice dates.
3467 * gcov-dump.c: Ditto.
3468 * gcov.c: Ditto.
3469 * doc/cpp.texi: Bump @copying's copyright year.
3470 * doc/cppinternals.texi: Ditto.
3471 * doc/gcc.texi: Ditto.
3472 * doc/gccint.texi: Ditto.
3473 * doc/gcov.texi: Ditto.
3474 * doc/install.texi: Ditto.
3475 * doc/invoke.texi: Ditto.
3476
3477 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3478
3479 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
3480
3481 2014-01-01 Jakub Jelinek <jakub@redhat.com>
3482
3483 * config/i386/sse.md (*mov<mode>_internal): Guard
3484 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
3485
3486 PR rtl-optimization/59647
3487 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
3488 new_rtx into UNSIGNED_FLOAT rtxes.
3489 \f
3490 Copyright (C) 2014 Free Software Foundation, Inc.
3491
3492 Copying and distribution of this file, with or without modification,
3493 are permitted in any medium without royalty provided the copyright
3494 notice and this notice are preserved.
This page took 0.192573 seconds and 6 git commands to generate.