]> gcc.gnu.org Git - gcc.git/blob - gcc/ChangeLog
pa.c (pa_option_override): Remove auto increment FIXME.
[gcc.git] / gcc / ChangeLog
1 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
2
3 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
4
5 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
6 Uros Bizjak <ubizjak@gmail.com>
7
8 PR target/60151
9 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU
10 assembler.
11 * configure: Regenerated.
12
13 2014-02-12 Richard Biener <rguenther@suse.de>
14
15 * vec.c (vec_prefix::calculate_allocation): Move as
16 inline variant to vec.h.
17 (vec_prefix::calculate_allocation_1): New out-of-line version.
18 * vec.h (vec_prefix::calculate_allocation_1): Declare.
19 (vec_prefix::m_has_auto_buf): Rename to ...
20 (vec_prefix::m_using_auto_storage): ... this.
21 (vec_prefix::calculate_allocation): Inline the easy cases
22 and dispatch to calculate_allocation_1 which doesn't need the
23 prefix address.
24 (va_heap::reserve): Use gcc_checking_assert.
25 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
26 m_using_auto_storage.
27 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
28 member and adjust.
29 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
30 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
31 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
32
33 2014-02-12 Richard Biener <rguenther@suse.de>
34
35 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
36 when we found a dependence.
37
38 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
39
40 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
41 common code...
42 (maybe_fold_stmt): ... into this new function.
43 * omp-low.c (lower_omp): Update comment.
44
45 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
46 last use.
47
48 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
49 dereference.
50
51 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
52
53 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
54 identifiers in comments.
55 (cortexa53_extra_costs): Likewise.
56 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in
57 comments.
58 (cortexa7_extra_costs): Likewise.
59 (cortexa12_extra_costs): Likewise.
60 (cortexa15_extra_costs): Likewise.
61 (v7m_extra_costs): Likewise.
62
63 2014-02-12 Richard Biener <rguenther@suse.de>
64
65 PR middle-end/60092
66 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
67 of posix_memalign being successful.
68 (lower_stmt): Restrict lowering of posix_memalign to when
69 -ftree-bit-ccp is enabled.
70
71 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
72
73 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
74 arg_loc.
75 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
76
77 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
78
79 PR rtl-optimization/60116
80 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
81 other_insn once the combination has been validated.
82
83 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
84
85 PR lto/59468
86 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
87 and wrapper.
88 * ipa-devirt.c: Include demangle.h
89 (odr_violation_reported): New static variable.
90 (add_type_duplicate): Update odr_violations.
91 (maybe_record_node): Add completep parameter; update it.
92 (record_target_from_binfo): Add COMPLETEP parameter;
93 update it as needed.
94 (possible_polymorphic_call_targets_1): Likewise.
95 (struct polymorphic_call_target_d): Add nonconstruction_targets;
96 rename FINAL to COMPLETE.
97 (record_targets_from_bases): Sanity check we found the binfo;
98 fix COMPLETEP updating.
99 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
100 parameter, fix computing of COMPLETEP.
101 (dump_possible_polymorphic_call_targets): Imrove readability of dump; at
102 LTO time do demangling.
103 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
104 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
105 parameter.
106 (gimple_get_virt_method_for_binfo): Likewise.
107 * gimple-fold.h (gimple_get_virt_method_for_binfo,
108 gimple_get_virt_method_for_vtable): Update prototypes.
109
110 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
111
112 PR target/49008
113 * genautomata.c (add_presence_absence): Fix typo with
114 {final_}presence_list.
115
116 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
117
118 PR target/60137
119 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
120 for VSX/Altivec vectors that land in GPR registers.
121
122 2014-02-11 Richard Henderson <rth@redhat.com>
123 Jakub Jelinek <jakub@redhat.com>
124
125 PR debug/59776
126 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
127 around drhs if type conversion to lacc->type is not useless.
128
129 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
130
131 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
132 tuning struct.
133 (cortex-a57.cortex-a53): Likewise.
134 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
135
136 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
137
138 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
139 arm_restrict_it.
140
141 2014-02-11 Renlin Li <Renlin.Li@arm.com>
142
143 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
144 add_options_for_arm_vfp3.
145
146 2014-02-11 Jeff Law <law@redhat.com>
147
148 PR middle-end/54041
149 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
150 object with an undesirable mode.
151
152 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
153
154 PR libgomp/60107
155 * config/i386/sol2-9.h: New file.
156 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
157 *-*-solaris2.9*): Use it.
158
159 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
160
161 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
162 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
163
164 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
165
166 * config/microblaze/microblaze.c: Extend mcpu version format
167
168 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
169
170 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
171
172 2014-02-10 Richard Henderson <rth@redhat.com>
173
174 PR target/59927
175 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
176 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
177 ms-abi vs -mno-accumulate-outgoing-args.
178 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
179 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
180 respect to ms-abi.
181
182 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
183
184 PR middle-end/60080
185 * cfgexpand.c (expand_asm_operands): Attach source location to
186 ASM_INPUT rtx objects.
187 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
188
189 2014-02-10 Nick Clifton <nickc@redhat.com>
190
191 * config/mn10300/mn10300.c (popcount): New function.
192 (mn10300_expand_prologue): Include saved registers in stack usage
193 count.
194
195 2014-02-10 Jeff Law <law@redhat.com>
196
197 PR middle-end/52306
198 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
199 when changing the SET_DEST of a prior insn to avoid an input
200 reload.
201
202 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
203
204 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
205 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
206 -mcall-openbsd, or -mcall-linux.
207 (CC1_ENDIAN_BIG_SPEC): Remove.
208 (CC1_ENDIAN_LITTLE_SPEC): Remove.
209 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
210 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
211 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
212 and %cc1_endian_default.
213 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
214
215 2014-02-10 Richard Biener <rguenther@suse.de>
216
217 PR tree-optimization/60115
218 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
219 MEM_REF handling. Properly verify that the accesses are not
220 out of the objects bound.
221
222 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
223
224 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
225 coretex to cortex.
226
227 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
228
229 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
230 proper constants and fix formatting.
231 (possible_polymorphic_call_targets): Fix formatting.
232
233 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
234 Ilya Tocar <ilya.tocar@intel.com>
235
236 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
237 (_mm512_loadu_epi32): Renamed into...
238 (_mm512_loadu_si512): This.
239 (_mm512_storeu_epi32): Renamed into...
240 (_mm512_storeu_si512): This.
241 (_mm512_maskz_ceil_ps): Removed.
242 (_mm512_maskz_ceil_pd): Ditto.
243 (_mm512_maskz_floor_ps): Ditto.
244 (_mm512_maskz_floor_pd): Ditto.
245 (_mm512_floor_round_ps): Ditto.
246 (_mm512_floor_round_pd): Ditto.
247 (_mm512_ceil_round_ps): Ditto.
248 (_mm512_ceil_round_pd): Ditto.
249 (_mm512_mask_floor_round_ps): Ditto.
250 (_mm512_mask_floor_round_pd): Ditto.
251 (_mm512_mask_ceil_round_ps): Ditto.
252 (_mm512_mask_ceil_round_pd): Ditto.
253 (_mm512_maskz_floor_round_ps): Ditto.
254 (_mm512_maskz_floor_round_pd): Ditto.
255 (_mm512_maskz_ceil_round_ps): Ditto.
256 (_mm512_maskz_ceil_round_pd): Ditto.
257 (_mm512_expand_pd): Ditto.
258 (_mm512_expand_ps): Ditto.
259 * config/i386/i386.c (ix86_builtins): Remove
260 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
261 (bdesc_args): Ditto.
262 * config/i386/predicates.md (const1256_operand): New.
263 (const_1_to_2_operand): Ditto.
264 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
265 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
266 (*avx512pf_gatherpf<mode>sf): Ditto.
267 (avx512pf_gatherpf<mode>df): Ditto.
268 (*avx512pf_gatherpf<mode>df_mask): Ditto.
269 (*avx512pf_gatherpf<mode>df): Ditto.
270 (avx512pf_scatterpf<mode>sf): Ditto.
271 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
272 (*avx512pf_scatterpf<mode>sf): Ditto.
273 (avx512pf_scatterpf<mode>df): Ditto.
274 (*avx512pf_scatterpf<mode>df_mask): Ditto.
275 (*avx512pf_scatterpf<mode>df): Ditto.
276 (avx512f_expand<mode>): Removed.
277 (<shift_insn><mode>3<mask_name>): Change predicate type.
278
279 2014-02-08 Jakub Jelinek <jakub@redhat.com>
280
281 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
282 not at the end of datarefs vector use ordered_remove to avoid
283 reordering datarefs vector.
284
285 PR c/59984
286 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
287 mark local addressable non-static vars as GOVD_PRIVATE
288 instead of GOVD_LOCAL.
289 * omp-low.c (lower_omp_for): Move gimple_bind_vars
290 and BLOCK_VARS of gimple_bind_block to new_stmt rather
291 than copying them.
292
293 PR middle-end/60092
294 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
295 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
296 assume_aligned or alloc_align attributes.
297 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
298 arguments. Handle also assume_aligned and alloc_align attributes.
299 (evaluate_stmt): Adjust bit_value_assume_aligned caller.
300 Handle calls to functions with assume_aligned or alloc_align
301 attributes.
302 * doc/extend.texi: Document assume_aligned and alloc_align
303 attributes.
304
305 2014-02-08 Terry Guo <terry.guo@arm.com>
306
307 * doc/invoke.texi: Document ARM -march=armv7e-m.
308
309 2014-02-08 Jakub Jelinek <jakub@redhat.com>
310
311 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
312 flag on __cilkrts_rethrow builtin.
313
314 PR ipa/60026
315 * ipa-cp.c (determine_versionability): Fail at -O0
316 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
317 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
318
319 Revert:
320 2014-02-04 Jakub Jelinek <jakub@redhat.com>
321
322 PR ipa/60026
323 * tree-inline.c (copy_forbidden): Fail for
324 __attribute__((optimize (0))) functions.
325
326 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
327
328 * varpool.c: Include pointer-set.h.
329 (varpool_remove_unreferenced_decls): Variables in other partitions
330 will not be output; be however careful to not lose information
331 about partitioning.
332
333 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
334
335 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
336 lookup in the vtable constructor.
337
338 2014-02-07 Jeff Law <law@redhat.com>
339
340 PR target/40977
341 * config/m68k/m68k.md (ashldi_extsi): Turn into a
342 define_insn_and_split.
343
344 * ipa-inline.c (inline_small_functions): Fix typos.
345
346 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
347
348 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
349 (s390_can_use_return_insn): Declare.
350 * config/s390/s390.h (EPILOGUE_USES): Define.
351 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
352 instructions.
353 (s390_chunkify_start): Handle return JUMP_LABELs.
354 (s390_early_mach): Emit a main_pool instruction on the entry edge.
355 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
356 (s390_can_use_return_insn): New functions.
357 (s390_fix_long_loop_prediction): Handle conditional returns.
358 (TARGET_SET_UP_BY_PROLOGUE): Define.
359 * config/s390/s390.md (ANY_RETURN): New code iterator.
360 (*creturn, *csimple_return, return, simple_return): New patterns.
361
362 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
363
364 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
365 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
366 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
367 REG_CFA_RESTORE list when deciding not to restore a register.
368
369 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
370
371 * config/s390/s390.c: Include tree-pass.h and context.h.
372 (s390_early_mach): New function, split out from...
373 (s390_emit_prologue): ...here.
374 (pass_data_s390_early_mach): New pass structure.
375 (pass_s390_early_mach): New class.
376 (s390_option_override): Create and register early_mach pass.
377 Move to end of file.
378
379 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
380
381 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
382 to match for the exit block.
383
384 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
385
386 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
387 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
388 Reject misaligned operands.
389
390 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
391
392 * optabs.c (expand_atomic_compare_and_swap): Allow expander to
393 fail.
394
395 2014-02-07 Richard Biener <rguenther@suse.de>
396
397 PR middle-end/60092
398 * gimple-low.c (lower_builtin_posix_memalign): New function.
399 (lower_stmt): Call it to lower posix_memalign in a way
400 to make alignment info accessible.
401
402 2014-02-07 Jakub Jelinek <jakub@redhat.com>
403
404 PR c++/60082
405 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
406 __builtin_setjmp_receiver.
407
408 2014-02-07 Richard Biener <rguenther@suse.de>
409
410 PR middle-end/60092
411 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
412 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
413 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
414 Handle BUILT_IN_POSIX_MEMALIGN.
415 (find_func_clobbers): Likewise.
416 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
417 (call_may_clobber_ref_p_1): Likewise.
418
419 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
420
421 PR ipa/59918
422 * ipa-devirt.c (record_target_from_binfo): Remove overactive sanity check.
423
424 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
425
426 PR ipa/59469
427 * lto-cgraph.c (lto_output_node): Use
428 symtab_get_symbol_partitioning_class.
429 (lto_output_varpool_node): likewise.
430 (symtab_get_symbol_partitioning_class): Move here from
431 lto/lto-partition.c
432 * cgraph.h (symbol_partitioning_class): Likewise.
433 (symtab_get_symbol_partitioning_class): Declare.
434
435 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
436
437 * ggc.h (ggc_internal_cleared_alloc): New macro.
438 * vec.h (vec_safe_copy): Handle memory stats.
439 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
440 * target-globals.c (save_target_globals): Likewise.
441
442 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
443
444 PR target/60077
445 * expr.c (emit_move_resolve_push): Export; be bit more selective
446 on when to clear alias set.
447 * expr.h (emit_move_resolve_push): Declare.
448 * function.h (struct function): Add tail_call_marked.
449 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
450 * config/i386/i386-protos.h (ix86_expand_push): Remove.
451 * config/i386/i386.md (TImode move expander): De not call
452 ix86_expand_push.
453 (FP push expanders): Preserve memory attributes.
454 * config/i386/sse.md (push<mode>1): Remove.
455 * config/i386/i386.c (ix86_expand_vector_move): Handle push
456 operation.
457 (ix86_expand_push): Remove.
458 * config/i386/mmx.md (push<mode>1): Remove.
459
460 2014-02-06 Jakub Jelinek <jakub@redhat.com>
461
462 PR rtl-optimization/60030
463 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
464 lopart with paradoxical subreg before shifting it up by hprec.
465
466 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
467
468 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
469 Remove extra newline at end of file.
470 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
471 (arm_issue_rate): Handle cortexa57.
472 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
473 (cortex-a57.cortex-a53): Likewise.
474
475 2014-02-06 Jakub Jelinek <jakub@redhat.com>
476
477 PR target/59575
478 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
479 don't record in REG_FRAME_RELATED_EXPR registers not set in that
480 bitmask.
481 (arm_expand_prologue): Adjust all callers.
482 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
483 info, registers also at the lowest numbered registers side. Use
484 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
485 XEXP.
486
487 PR debug/59992
488 * var-tracking.c (adjust_mems): Before adding a SET
489 to amd->side_effects, adjust it's SET_SRC using
490 simplify_replace_fn_rtx.
491
492 2014-02-06 Alan Modra <amodra@gmail.com>
493
494 PR target/60032
495 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
496 change SDmode to DDmode when lra_in_progress.
497
498 2014-02-06 Jakub Jelinek <jakub@redhat.com>
499
500 PR middle-end/59150
501 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
502 free_data_ref on the dr first, and before goto again also set dr
503 to the next dr. For simd_lane_access, free old datarefs[i] before
504 overwriting it. For get_vectype_for_scalar_type failure, don't
505 free_data_ref if simd_lane_access.
506
507 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
508
509 PR target/60062
510 * tree.h (opts_for_fn): New inline function.
511 (opt_for_fn): Define.
512 * config/i386/i386.c (ix86_function_regparm): Use
513 opt_for_fn (decl, optimize) instead of optimize.
514
515 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
516
517 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
518 for SYMBOL_REF in large memory model.
519
520 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
521
522 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
523 and crypto support.
524 (cortex-a57): Likewise.
525 (cortex-a57.cortex-a53): Likewise.
526
527 2014-02-06 Yury Gribov <y.gribov@samsung.com>
528 Kugan Vivekanandarajah <kuganv@linaro.org>
529
530 * config/arm/arm.c (arm_vector_alignment_reachable): Check
531 unaligned_access.
532 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
533
534 2014-02-06 Richard Biener <rguenther@suse.de>
535
536 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
537 set_loop_copy and initialize_original_copy_tables.
538
539 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
540
541 * config/aarch64/aarch64-simd.md
542 (aarch64_ashr_simddi): Change QI to SI.
543
544 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
545 Jakub Jelinek <jakub@redhat.com>
546
547 PR middle-end/60013
548 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
549 of the dataflow.
550
551 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
552
553 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
554 CODE_FOR_altivec_vpku[hw]um to
555 CODE_FOR_altivec_vpku[hw]um_direct.
556 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
557 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
558 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
559 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
560
561 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
562
563 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
564 generation for -maltivec=be.
565 (altivec_vsumsws): Simplify redundant test.
566
567 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
568
569 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
570 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
571 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
572 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
573 gen_altivec_vpkuwum.
574 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
575 BYTES_BIG_ENDIAN.
576 (altivec_vpks<VI_char>ss): Likewise.
577 (altivec_vpks<VI_char>us): Likewise.
578 (altivec_vpku<VI_char>us): Likewise.
579 (altivec_vpku<VI_char>um): Likewise.
580 (altivec_vpku<VI_char>um_direct): New (copy of
581 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
582 internal use).
583 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
584 target is little endian and -maltivec=be is not specified.
585 (*altivec_vupkhs<VU_char>_direct): New (copy of
586 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal
587 use).
588 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
589 target is little endian and -maltivec=be is not specified.
590 (*altivec_vupkls<VU_char>_direct): New (copy of
591 altivec_vupkls<VU_char> that always emits vupkls*, for internal
592 use).
593 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
594 little endian and -maltivec=be is not specified.
595 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
596 little endian and -maltivec=be is not specified.
597
598 2014-02-05 Richard Henderson <rth@redhat.com>
599
600 PR debug/52727
601 * combine-stack-adj.c: Revert r206943.
602 * sched-int.h (struct deps_desc): Add last_args_size.
603 * sched-deps.c (init_deps): Initialize it.
604 (sched_analyze_insn): Add OUTPUT dependencies between insns that
605 contain REG_ARGS_SIZE notes.
606
607 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
608
609 * lto-cgraph.c (asm_nodes_output): Make global.
610 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
611 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA
612 parameter
613 (driver_handle_option): Handle OPT_fwpa.
614
615 2014-02-05 Jakub Jelinek <jakub@redhat.com>
616
617 PR ipa/59947
618 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
619 a comment typo and formatting issue. If odr_hash hasn't been
620 created, return vNULL and set *completep to false.
621
622 PR middle-end/57499
623 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
624 bb with no successors.
625
626 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
627
628 PR target/59718
629 * doc/invoke.texi (-march): Clarify documentation for ARM.
630 (-mtune): Likewise.
631 (-mcpu): Likewise.
632
633 2014-02-05 Richard Biener <rguenther@suse.de>
634
635 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
636 when not vectorizing because of too many alias checks.
637 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
638 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
639
640 2014-02-05 Nick Clifton <nickc@redhat.com>
641
642 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
643 accept extended registers in any mode when compiling for the
644 MN10300.
645
646 2014-02-05 Yury Gribov <y.gribov@samsung.com>
647
648 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
649 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
650 sanitization attributes.
651 (can_inline_edge_p): Likewise.
652 (sanitize_attrs_match_for_inline_p): New function.
653
654 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
655
656 * ipa-prop.c (detect_type_change): Shor circuit testing of
657 type changes on THIS pointer.
658
659 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
660
661 PR target/59777
662 * config/pa/pa.c (legitimize_tls_address): Return original address
663 if not passed a SYMBOL_REF rtx.
664 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
665 addresses.
666 (pa_emit_move_sequence): Simplify TLS source operands.
667 (pa_legitimate_constant_p): Reject all TLS constants.
668 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
669 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
670
671 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
672
673 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
674 groups when we know they are controlled by LTO.
675 * varasm.c (default_binds_local_p_1): If object is in other partition,
676 it will be resolved locally.
677
678 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
679
680 * config/host-linux.c (linux_gt_pch_use_address): Don't
681 use SSIZE_MAX because it is not always defined.
682
683 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
684
685 PR bootstrap/59913
686 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
687 threshold for pseudo splitting.
688 (update_ebb_live_info): Process call argument hard registers and
689 hard registers from insn definition too.
690 (max_small_class_regs_num): New constant.
691 (inherit_in_ebb): Update live hard regs through EBBs. Update
692 reloads_num only for small register classes. Don't split for
693 outputs of jumps.
694
695 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
696
697 PR ipa/60058
698 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
699 is non-null.
700
701 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
702
703 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default visibility is safe.
704
705 2014-02-04 Marek Polacek <polacek@redhat.com>
706
707 * gdbinit.in (pel): Define.
708
709 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
710
711 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
712 behavior.
713
714 2014-02-04 Richard Biener <rguenther@suse.de>
715
716 PR lto/59723
717 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
718 in function context local.
719 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
720 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
721 similar to LTO_imported_decl_ref.
722
723 2014-02-04 Jakub Jelinek <jakub@redhat.com>
724
725 PR tree-optimization/60002
726 * cgraphclones.c (build_function_decl_skip_args): Clear
727 DECL_LANG_SPECIFIC.
728
729 PR tree-optimization/60023
730 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
731 false to gsi_replace.
732 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
733 has been in some EH region and vec_stmt could throw, add
734 vec_stmt into the same EH region.
735 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
736 has no lhs, ignore it.
737 * internal-fn.c (expand_MASK_LOAD): Likewise.
738
739 PR ipa/60026
740 * tree-inline.c (copy_forbidden): Fail for
741 __attribute__((optimize (0))) functions.
742
743 PR other/58712
744 * omp-low.c (simd_clone_struct_copy): If from->inbranch
745 is set, copy one less argument.
746 (expand_simd_clones): Don't subtract clone_info->inbranch
747 from simd_clone_struct_alloc argument.
748
749 PR rtl-optimization/57915
750 * recog.c (simplify_while_replacing): If all unary/binary/relational
751 operation arguments are constant, attempt to simplify those.
752
753 PR middle-end/59261
754 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
755 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
756
757 2014-02-04 Richard Biener <rguenther@suse.de>
758
759 PR tree-optimization/60012
760 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
761 TBAA disambiguation to all DDRs.
762
763 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
764
765 PR target/59788
766 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
767 (LINK_SPEC): Use it for -shared, -shared-libgcc.
768
769 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
770
771 PR ipa/59882
772 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
773
774 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
775
776 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
777 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
778
779 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
780
781 PR ipa/59831
782 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
783 to figure out targets of polymorphic calls with known decl.
784 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
785 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
786 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
787 (get_polymorphic_call_info): ... here.
788 (get_polymorphic_call_info_from_invariant): New function.
789
790 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
791
792 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
793 lookup via vtable pointer; check for type consistency
794 and turn inconsitent facts into UNREACHABLE.
795 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
796 * gimple-fold.c (gimple_get_virt_method_for_vtable):
797 Do not ICE on type inconsistent querries; return UNREACHABLE
798 instead.
799
800 2014-02-03 Richard Henderson <rth@twiddle.net>
801
802 PR tree-opt/59924
803 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
804 already processed this node.
805 (normalize_one_pred_1): Pass along mark_set.
806 (normalize_one_pred): Create and destroy a pointer_set_t.
807 (normalize_one_pred_chain): Likewise.
808
809 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
810
811 PR gcov-profile/58602
812 * gcc/gcov-io.c (gcov_open): Open with truncation when mode<0
813
814 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
815
816 PR ipa/59831
817 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on -fno-devirtualize;
818 Try to devirtualize by the knowledge of virtual table pointer given by
819 aggregate propagation.
820 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
821 ipa_print_node_jump_functions): Dump also offset that
822 is relevant for polymorphic calls.
823 (determine_known_aggregate_parts): Add arg_type parameter; use it
824 instead of determining the type from pointer type.
825 (ipa_compute_jump_functions_for_edge): Update call of
826 determine_known_aggregate_parts.
827 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
828 (gimple_get_virt_method_for_binfo): ... here; simplify using
829 vtable_pointer_value_to_vtable.
830 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
831 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
832 to unsigned HOST_WIDE_INT; Use vtable_pointer_value_to_vtable.
833 (vtable_pointer_value_to_vtable): Break out from ...; handle also
834 POINTER_PLUS_EXPR.
835 (vtable_pointer_value_to_binfo): ... here.
836 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
837
838 2014-02-03 Teresa Johnson <tejohnson@google.com>
839
840 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
841 redef of outer loop index variable.
842
843 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
844
845 PR c++/53017
846 PR c++/59211
847 * doc/extend.texi (Function Attributes): Typo.
848
849 2014-02-03 Cong Hou <congh@google.com>
850
851 PR tree-optimization/60000
852 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
853 if the vectorized statement is a store. A store statement can only
854 appear at the end of pattern statements.
855
856 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
857
858 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
859 (ix86_option_override_internal): Default long double to 64-bit for
860 32-bit Bionic and to 128-bit for 64-bit Bionic.
861
862 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
863 TARGET_LONG_DOUBLE_128 is true.
864 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
865
866 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
867 (mlong-double-64): Negate -mlong-double-128.
868 (mlong-double-128): New option.
869
870 * config/i386/i386-c.c (ix86_target_macros): Define
871 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
872
873 * doc/invoke.texi: Document -mlong-double-128.
874
875 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
876
877 PR rtl-optimization/60024
878 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
879
880 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
881
882 * doc/invoke.texi: (fprofile-reorder-functions): Fix typo.
883
884 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
885
886 PR rtl-optimization/57662
887 * sel-sched.c (code_motion_path_driver): Do not mark already not
888 existing blocks in the visiting bitmap.
889
890 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
891
892 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
893 on the insn being emitted.
894
895 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
896 Will Deacon <will.deacon@arm.com>
897
898 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
899
900 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
901
902 * config/arm/arm-tables.opt: Regenerate.
903
904 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
905
906 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
907 for vector types other than V16QImode.
908 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
909 define_expand, and call altivec_expand_vec_perm_le when producing
910 code with little endian element order.
911 (*altivec_vperm_<mode>_internal): New insn having previous
912 behavior of altivec_vperm_<mode>.
913 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
914 altivec_expand_vec_perm_le when producing code with little endian
915 element order.
916 (*altivec_vperm_<mode>_uns_internal): New insn having previous
917 behavior of altivec_vperm_<mode>_uns.
918
919 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
920
921 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
922 (altivec_vsumsws): Add handling for -maltivec=be with a little
923 endian target.
924 (altivec_vsumsws_direct): New.
925 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
926 gen_altivec_vsumsws.
927
928 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
929
930 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
931 vtable_pointer_value_to_binfo): New functions.
932 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
933 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
934
935 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
936
937 * config/nios2/nios2.md (load_got_register): Initialize GOT
938 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
939 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
940
941 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
942
943 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
944 preserverd by passthrough, do not propagate the type.
945
946 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
947
948 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
949 (mips_atomic_assign_expand_fenv): New function.
950 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
951
952 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
953
954 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
955 (__builtin_mips_set_fcsr): Likewise.
956 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
957 MIPS_USI_FTYPE_VOID.
958 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
959 (mips16_expand_set_fcsr): Likewise.
960 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
961 (mips16_set_fcsr_stub): Likewise.
962 (mips16_get_fcsr_one_only_stub): New class.
963 (mips16_set_fcsr_one_only_stub): Likewise.
964 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
965 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
966 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
967 (hard_float): New availability predicate.
968 (mips_builtins): Add get_fcsr and set_fcsr.
969 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
970 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
971 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
972 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
973 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
974 patterns.
975
976 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
977
978 * config/mips/mips.c (mips_one_only_stub): New class.
979 (mips_need_mips16_rdhwr_p): Replace with...
980 (mips16_rdhwr_stub): ...this new variable.
981 (mips16_stub_call_address): New function.
982 (mips16_rdhwr_one_only_stub): New class.
983 (mips_expand_thread_pointer): Use mips16_stub_call_address.
984 (mips_output_mips16_rdhwr): Delete.
985 (mips_finish_stub): New function.
986 (mips_code_end): Use it to handle rdhwr stubs.
987
988 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
989
990 PR target/60017
991 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
992 when calculating size of integer atomic types.
993
994 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
995
996 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
997
998 2014-02-01 Jakub Jelinek <jakub@redhat.com>
999
1000 PR tree-optimization/60003
1001 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
1002 * profile.c (branch_prob): Use gimple_call_builtin_p
1003 to check for BUILT_IN_SETJMP_RECEIVER.
1004 * tree-inline.c (copy_bb): Call notice_special_calls.
1005
1006 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
1007
1008 PR bootstrap/59985
1009 * lra-constraints.c (process_alt_operands): Update reload_sum only
1010 on the first pass.
1011
1012 2014-01-31 Richard Henderson <rth@redhat.com>
1013
1014 PR middle-end/60004
1015 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
1016 until after else_eh is processed.
1017
1018 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
1019
1020 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
1021 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
1022 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
1023 in smmintrin.h, remove them.
1024 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
1025 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
1026 * config/i386/i386.md (ROUND_SAE): Fix value.
1027 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
1028 (const48_operand): New.
1029 * config/i386/subst.md (round), (round_expand): Use
1030 const_4_or_8_to_11_operand.
1031 (round_saeonly), (round_saeonly_expand): Use const48_operand.
1032
1033 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
1034
1035 * config/i386/constraints.md (Yk): Swap meaning with k.
1036 * config/i386/i386.md (movhi_internal): Change Yk to k.
1037 (movqi_internal): Ditto.
1038 (*k<logic><mode>): Ditto.
1039 (*andhi_1): Ditto.
1040 (*andqi_1): Ditto.
1041 (kandn<mode>): Ditto.
1042 (*<code>hi_1): Ditto.
1043 (*<code>qi_1): Ditto.
1044 (kxnor<mode>): Ditto.
1045 (kortestzhi): Ditto.
1046 (kortestchi): Ditto.
1047 (kunpckhi): Ditto.
1048 (*one_cmplhi2_1): Ditto.
1049 (*one_cmplqi2_1): Ditto.
1050 * config/i386/sse.md (): Change k to Yk.
1051 (avx512f_load<mode>_mask): Ditto.
1052 (avx512f_blendm<mode>): Ditto.
1053 (avx512f_store<mode>_mask): Ditto.
1054 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
1055 (avx512f_storedqu<mode>_mask): Ditto.
1056 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
1057 Ditto.
1058 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
1059 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
1060 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
1061 (avx512f_maskcmp<mode>3): Ditto.
1062 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
1063 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
1064 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
1065 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
1066 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
1067 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
1068 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
1069 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
1070 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
1071 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
1072 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
1073 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
1074 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
1075 (vec_extract_lo_<mode>_maskm): Ditto.
1076 (vec_extract_hi_<mode>_maskm): Ditto.
1077 (avx512f_vternlog<mode>_mask): Ditto.
1078 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
1079 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
1080 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
1081 (avx512f_<code>v8div16qi2_mask): Ditto.
1082 (avx512f_<code>v8div16qi2_mask_store): Ditto.
1083 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
1084 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
1085 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
1086 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
1087 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
1088 (*avx512pf_gatherpf<mode>df_mask): Ditto.
1089 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
1090 (*avx512pf_scatterpf<mode>df_mask): Ditto.
1091 (avx512cd_maskb_vec_dupv8di): Ditto.
1092 (avx512cd_maskw_vec_dupv16si): Ditto.
1093 (avx512f_vpermi2var<mode>3_maskz): Ditto.
1094 (avx512f_vpermi2var<mode>3_mask): Ditto.
1095 (avx512f_vpermi2var<mode>3_mask): Ditto.
1096 (avx512f_vpermt2var<mode>3_maskz): Ditto.
1097 (*avx512f_gathersi<mode>): Ditto.
1098 (*avx512f_gathersi<mode>_2): Ditto.
1099 (*avx512f_gatherdi<mode>): Ditto.
1100 (*avx512f_gatherdi<mode>_2): Ditto.
1101 (*avx512f_scattersi<mode>): Ditto.
1102 (*avx512f_scatterdi<mode>): Ditto.
1103 (avx512f_compress<mode>_mask): Ditto.
1104 (avx512f_compressstore<mode>_mask): Ditto.
1105 (avx512f_expand<mode>_mask): Ditto.
1106 * config/i386/subst.md (mask): Change k to Yk.
1107 (mask_scalar_merge): Ditto.
1108 (sd): Ditto.
1109
1110 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
1111
1112 * doc/extend.texi (Vector Extensions): Document ?: in C++.
1113
1114 2014-01-31 Richard Biener <rguenther@suse.de>
1115
1116 PR middle-end/59990
1117 * builtins.c (fold_builtin_memory_op): Make sure to not
1118 use a floating-point mode or a boolean or enumeral type for
1119 the copy operation.
1120
1121 2014-01-30 DJ Delorie <dj@redhat.com>
1122
1123 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
1124 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
1125 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
1126 whenever main() has an epilogue.
1127
1128 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1129
1130 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
1131 unused variable "field".
1132 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
1133 (vsx_mergeh_<mode>): Likewise.
1134 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
1135 (altivec_vmrghh): Likewise.
1136 (altivec_vmrghw): Likewise.
1137 (altivec_vmrglb): Likewise.
1138 (altivec_vmrglh): Likewise.
1139 (altivec_vmrglw): Likewise.
1140 (altivec_vspltb): Add missing uses.
1141 (altivec_vsplth): Likewise.
1142 (altivec_vspltw): Likewise.
1143 (altivec_vspltsf): Likewise.
1144
1145 2014-01-30 Jakub Jelinek <jakub@redhat.com>
1146
1147 PR target/59923
1148 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
1149 frame related instructions.
1150
1151 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
1152
1153 PR rtl-optimization/59959
1154 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
1155 any reload of register whose subreg is invalid.
1156
1157 2014-01-30 Jakub Jelinek <jakub@redhat.com>
1158
1159 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
1160 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
1161 Add missing return type - void.
1162
1163 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1164
1165 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
1166 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
1167 remove element index adjustment for endian (now handled in vsx.md
1168 and altivec.md).
1169 (altivec_expand_vec_perm_const): Use
1170 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
1171 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
1172 (vsx_xxspltw_<mode>): Adjust element index for little endian.
1173 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
1174 define_expand and a new define_insn *altivec_vspltb_internal;
1175 adjust for -maltivec=be on a little endian target.
1176 (altivec_vspltb_direct): New.
1177 (altivec_vsplth): Divide into a define_expand and a new
1178 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
1179 little endian target.
1180 (altivec_vsplth_direct): New.
1181 (altivec_vspltw): Divide into a define_expand and a new
1182 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
1183 little endian target.
1184 (altivec_vspltw_direct): New.
1185 (altivec_vspltsf): Divide into a define_expand and a new
1186 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
1187 a little endian target.
1188
1189 2014-01-30 Richard Biener <rguenther@suse.de>
1190
1191 PR tree-optimization/59993
1192 * tree-ssa-forwprop.c (associate_pointerplus): Check we
1193 can propagate form the earlier stmt and avoid the transform
1194 when the intermediate result is needed.
1195
1196 2014-01-30 Alangi Derick <alangiderick@gmail.com>
1197
1198 * README.Portability: Fix typo.
1199
1200 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
1201
1202 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
1203 comparison_operator with ordered_comparison_operator.
1204
1205 2014-01-30 Nick Clifton <nickc@redhat.com>
1206
1207 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
1208 Rename to mn10300_store_multiple_regs.
1209 * config/mn10300/mn10300.c: Likewise.
1210 * config/mn10300/mn10300.md (store_movm): Fix typo: call
1211 store_multiple_regs.
1212 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
1213 Call mn10300_store_multiple_regs.
1214
1215 2014-01-30 Nick Clifton <nickc@redhat.com>
1216 DJ Delorie <dj@redhat.com>
1217
1218 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
1219 %fp 2 to keep registers after it properly word-aligned.
1220 (rl78_alloc_physical_registers_umul): Handle the case where both
1221 input operands are the same.
1222
1223 2014-01-30 Richard Biener <rguenther@suse.de>
1224
1225 PR tree-optimization/59903
1226 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
1227 check properly.
1228
1229 2014-01-30 Jason Merrill <jason@redhat.com>
1230
1231 PR c++/59633
1232 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
1233
1234 PR c++/59645
1235 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
1236
1237 2014-01-30 Richard Biener <rguenther@suse.de>
1238
1239 PR tree-optimization/59951
1240 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
1241
1242 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
1243
1244 PR target/59784
1245 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
1246 SFmode to DFmode case.
1247
1248 2014-01-29 DJ Delorie <dj@redhat.com>
1249
1250 * config/msp430/msp430.opt (-minrt): New.
1251 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
1252 if -minrt given.
1253 (ENDFILE_SPEC): Likewise.
1254
1255 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
1256
1257 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
1258 (estimate_function_body_sizes): Use it.
1259
1260 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
1261
1262 PR c++/58561
1263 * dwarf2out.c (is_cxx_auto): New.
1264 (is_base_type): Use it.
1265 (gen_type_die_with_usage): Likewise.
1266
1267 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1268
1269 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
1270 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
1271 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
1272 -maltivec=be with LE targets.
1273 (vsx_mergeh_<mode>): Likewise.
1274 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
1275 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
1276 (altivec_vmrghb): Replace with define_expand and new
1277 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
1278 (altivec_vmrghb_direct): New define_insn.
1279 (altivec_vmrghh): Replace with define_expand and new
1280 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
1281 (altivec_vmrghh_direct): New define_insn.
1282 (altivec_vmrghw): Replace with define_expand and new
1283 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
1284 (altivec_vmrghw_direct): New define_insn.
1285 (*altivec_vmrghsf): Adjust for endianness.
1286 (altivec_vmrglb): Replace with define_expand and new
1287 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
1288 (altivec_vmrglb_direct): New define_insn.
1289 (altivec_vmrglh): Replace with define_expand and new
1290 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
1291 (altivec_vmrglh_direct): New define_insn.
1292 (altivec_vmrglw): Replace with define_expand and new
1293 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
1294 (altivec_vmrglw_direct): New define_insn.
1295 (*altivec_vmrglsf): Adjust for endianness.
1296 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1297 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1298 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
1299 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
1300 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1301 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1302 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
1303 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
1304
1305 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
1306
1307 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
1308 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
1309 whitespace.
1310
1311 2014-01-29 Richard Biener <rguenther@suse.de>
1312
1313 PR tree-optimization/58742
1314 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
1315 associate_pointerplus_align.
1316 (associate_pointerplus_diff): New function.
1317 (associate_pointerplus): Likewise. Call associate_pointerplus_align
1318 and associate_pointerplus_diff.
1319
1320 2014-01-29 Richard Biener <rguenther@suse.de>
1321
1322 * lto-streamer.h (LTO_major_version): Bump to 3.
1323 (LTO_minor_version): Reset to 0.
1324
1325 2014-01-29 Renlin Li <Renlin.Li@arm.com>
1326
1327 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
1328 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
1329 (arm_file_start): Generate correct asm header for armv7ve.
1330 * config/arm/bpabi.h: Add multilib support for armv7ve.
1331 * config/arm/driver-arm.c: Change the architectures of cortex-a7
1332 and cortex-a15 to armv7ve.
1333 * config/arm/t-aprofile: Add multilib support for armv7ve.
1334 * doc/invoke.texi: Document -march=armv7ve.
1335
1336 2014-01-29 Richard Biener <rguenther@suse.de>
1337
1338 PR tree-optimization/58742
1339 * tree-ssa-forwprop.c (associate_plusminus): Return true
1340 if we changed sth, defer EH cleanup to ...
1341 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
1342 (simplify_mult): New function.
1343
1344 2014-01-29 Jakub Jelinek <jakub@redhat.com>
1345
1346 PR middle-end/59917
1347 PR tree-optimization/59920
1348 * tree.c (build_common_builtin_nodes): Remove
1349 __builtin_setjmp_dispatcher initialization.
1350 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
1351 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
1352 instead of gsi_after_labels + manually skipping debug stmts.
1353 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
1354 ignore bbs with IFN_ABNORMAL_DISPATCHER.
1355 * tree-inline.c (copy_edges_for_bb): Remove
1356 can_make_abnormal_goto argument, instead add abnormal_goto_dest
1357 argument. Ignore computed_goto_p stmts. Don't call
1358 make_abnormal_goto_edges. If a call might need abnormal edges
1359 for non-local gotos, see if it already has an edge to
1360 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
1361 with true argument, don't do anything then, otherwise add
1362 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
1363 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
1364 caller.
1365 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
1366 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
1367 (lower_stmt): Don't set data->calls_builtin_setjmp.
1368 (lower_builtin_setjmp): Adjust comment.
1369 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
1370 * tree-cfg.c (found_computed_goto): Remove.
1371 (factor_computed_gotos): Remove.
1372 (make_goto_expr_edges): Return bool, true for computed gotos.
1373 Don't call make_abnormal_goto_edges.
1374 (build_gimple_cfg): Don't set found_computed_goto, don't call
1375 factor_computed_gotos.
1376 (computed_goto_p): No longer static.
1377 (make_blocks): Don't set found_computed_goto.
1378 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
1379 (make_edges): If make_goto_expr_edges returns true, push bb
1380 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
1381 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
1382 vector. Record mapping between bbs and OpenMP regions if there
1383 are any, adjust make_gimple_omp_edges caller. Call
1384 handle_abnormal_edges.
1385 (make_abnormal_goto_edges): Remove.
1386 * tree-cfg.h (make_abnormal_goto_edges): Remove.
1387 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
1388 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
1389 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
1390 * internal-fn.def (ABNORMAL_DISPATCHER): New.
1391 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
1392 filling *region also set *region_idx to (*region)->entry->index.
1393
1394 PR other/58712
1395 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
1396 For REGs set ORIGINAL_REGNO.
1397
1398 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
1399
1400 * doc/md.texi: Mention that a target shouldn't implement
1401 vec_widen_(s|u)mul_even/odd pair if it is less efficient
1402 than hi/lo pair.
1403
1404 2014-01-29 Jakub Jelinek <jakub@redhat.com>
1405
1406 PR tree-optimization/59594
1407 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
1408 a copy of the datarefs vector rather than the vector itself.
1409
1410 2014-01-28 Jason Merrill <jason@redhat.com>
1411
1412 PR c++/53756
1413 * dwarf2out.c (auto_die): New static.
1414 (gen_type_die_with_usage): Handle C++1y 'auto'.
1415 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
1416 on definition.
1417
1418 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
1419
1420 PR target/59672
1421 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
1422 (SPEC_X32): Likewise.
1423 (SPEC_64): Likewise.
1424 * config/i386/i386.c (ix86_option_override_internal): Turn off
1425 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
1426 for TARGET_16BIT.
1427 (x86_file_start): Output .code16gcc for TARGET_16BIT.
1428 * config/i386/i386.h (TARGET_16BIT): New macro.
1429 (TARGET_16BIT_P): Likewise.
1430 * config/i386/i386.opt: Add m16.
1431 * doc/invoke.texi: Document -m16.
1432
1433 2014-01-28 Jakub Jelinek <jakub@redhat.com>
1434
1435 PR preprocessor/59935
1436 * input.c (location_get_source_line): Bail out on when line number
1437 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
1438
1439 2014-01-28 Richard Biener <rguenther@suse.de>
1440
1441 PR tree-optimization/58742
1442 * tree-ssa-forwprop.c (associate_plusminus): Handle
1443 pointer subtraction of the form (T)(P + A) - (T)P.
1444
1445 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1446
1447 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
1448 at const_int_cost.
1449
1450 2014-01-28 Richard Biener <rguenther@suse.de>
1451
1452 Revert
1453 2014-01-28 Richard Biener <rguenther@suse.de>
1454
1455 PR rtl-optimization/45364
1456 PR rtl-optimization/59890
1457 * var-tracking.c (local_get_addr_clear_given_value): Handle
1458 already cleared slot.
1459 (val_reset): Handle not allocated local_get_addr_cache.
1460 (vt_find_locations): Use post-order on the inverted CFG.
1461
1462 2014-01-28 Richard Biener <rguenther@suse.de>
1463
1464 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
1465
1466 2014-01-28 Richard Biener <rguenther@suse.de>
1467
1468 PR rtl-optimization/45364
1469 PR rtl-optimization/59890
1470 * var-tracking.c (local_get_addr_clear_given_value): Handle
1471 already cleared slot.
1472 (val_reset): Handle not allocated local_get_addr_cache.
1473 (vt_find_locations): Use post-order on the inverted CFG.
1474
1475 2014-01-28 Alan Modra <amodra@gmail.com>
1476
1477 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
1478 * configure.ac <recursive call for build != host>: Define
1479 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
1480 and LD_FOR_BUILD too.
1481 * configure: Regenerate.
1482
1483 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
1484
1485 * config/i386/i386.c (get_builtin_code_for_version): Separate
1486 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
1487 Broadwell from Haswell.
1488
1489 2014-01-27 Steve Ellcey <sellcey@mips.com>
1490
1491 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
1492 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
1493 * config/mips/mips.c (mips_option_override): Change setting
1494 of TARGET_DSP.
1495 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
1496 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
1497 Change from Mask to Var.
1498
1499 2014-01-27 Jeff Law <law@redhat.com>
1500
1501 * ipa-inline.c (inline_small_functions): Fix typo.
1502
1503 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
1504
1505 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
1506 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
1507 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
1508 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
1509 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
1510 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
1511 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
1512 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
1513 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
1514 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
1515 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
1516 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
1517 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
1518 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
1519 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
1520 (_mm512_storeu_epi64): Ditto.
1521 (_mm512_cmpge_epi32_mask): Ditto.
1522 (_mm512_cmpge_epu32_mask): Ditto.
1523 (_mm512_cmpge_epi64_mask): Ditto.
1524 (_mm512_cmpge_epu64_mask): Ditto.
1525 (_mm512_cmple_epi32_mask): Ditto.
1526 (_mm512_cmple_epu32_mask): Ditto.
1527 (_mm512_cmple_epi64_mask): Ditto.
1528 (_mm512_cmple_epu64_mask): Ditto.
1529 (_mm512_cmplt_epi32_mask): Ditto.
1530 (_mm512_cmplt_epu32_mask): Ditto.
1531 (_mm512_cmplt_epi64_mask): Ditto.
1532 (_mm512_cmplt_epu64_mask): Ditto.
1533 (_mm512_cmpneq_epi32_mask): Ditto.
1534 (_mm512_cmpneq_epu32_mask): Ditto.
1535 (_mm512_cmpneq_epi64_mask): Ditto.
1536 (_mm512_cmpneq_epu64_mask): Ditto.
1537 (_mm512_expand_pd): Ditto.
1538 (_mm512_expand_ps): Ditto.
1539 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
1540 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
1541 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
1542 * config/i386/i386.c (ix86_builtins): Add
1543 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
1544 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
1545 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
1546 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
1547 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
1548 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
1549 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
1550 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
1551 IX86_BUILTIN_PMOVUSQW512_MEM.
1552 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
1553 __builtin_ia32_pmovsqd512mem_mask,
1554 __builtin_ia32_pmovqd512mem_mask,
1555 __builtin_ia32_pmovusqw512mem_mask,
1556 __builtin_ia32_pmovsqw512mem_mask,
1557 __builtin_ia32_pmovqw512mem_mask,
1558 __builtin_ia32_pmovusdw512mem_mask,
1559 __builtin_ia32_pmovsdw512mem_mask,
1560 __builtin_ia32_pmovdw512mem_mask,
1561 __builtin_ia32_pmovqb512mem_mask,
1562 __builtin_ia32_pmovusqb512mem_mask,
1563 __builtin_ia32_pmovsqb512mem_mask,
1564 __builtin_ia32_pmovusdb512mem_mask,
1565 __builtin_ia32_pmovsdb512mem_mask,
1566 __builtin_ia32_pmovdb512mem_mask.
1567 (bdesc_args): Add __builtin_ia32_expanddf512,
1568 __builtin_ia32_expandsf512.
1569 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
1570 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
1571 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
1572 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
1573 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
1574 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
1575 (avx512f_<code>v8div16qi2_mask_store): This.
1576 (avx512f_expand<mode>): New.
1577
1578 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
1579
1580 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
1581 New.
1582 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1583 (_mm512_prefetch_i32scatter_pd): Ditto.
1584 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1585 (_mm512_prefetch_i64scatter_pd): Ditto.
1586 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1587 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
1588 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1589 (_mm512_prefetch_i32scatter_ps): Ditto.
1590 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1591 (_mm512_prefetch_i64scatter_ps): Ditto.
1592 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1593 * config/i386/i386-builtin-types.def: Define
1594 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
1595 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
1596 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
1597 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
1598 IX86_BUILTIN_SCATTERPFQPD.
1599 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
1600 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
1601 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
1602 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
1603 __builtin_ia32_scatterpfqps.
1604 (ix86_expand_builtin): Expand new built-ins.
1605 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
1606 fix memory access data type.
1607 (*avx512pf_gatherpf<mode>_mask): Ditto.
1608 (*avx512pf_gatherpf<mode>): Ditto.
1609 (avx512pf_scatterpf<mode>): Ditto.
1610 (*avx512pf_scatterpf<mode>_mask): Ditto.
1611 (*avx512pf_scatterpf<mode>): Ditto.
1612 (GATHER_SCATTER_SF_MEM_MODE): New.
1613 (avx512pf_gatherpf<mode>df): Ditto.
1614 (*avx512pf_gatherpf<mode>df_mask): Ditto.
1615 (*avx512pf_scatterpf<mode>df): Ditto.
1616
1617 2014-01-27 Jakub Jelinek <jakub@redhat.com>
1618
1619 PR bootstrap/59934
1620 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
1621 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
1622 reached.
1623
1624 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
1625
1626 * common/config/arm/arm-common.c
1627 (arm_rewrite_mcpu): Handle multiple names.
1628 * config/arm/arm.h
1629 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
1630
1631 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
1632
1633 * gimple-builder.h (create_gimple_tmp): Delete.
1634
1635 2014-01-27 Christian Bruel <christian.bruel@st.com>
1636
1637 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
1638 words comparisons.
1639
1640 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
1641
1642 * config/pa/pa.md (call): Generate indirect long calls to non-local
1643 functions when outputing 32-bit code.
1644 (call_value): Likewise except for special call to buggy powf function.
1645
1646 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
1647 portable runtime and PIC indirect calls.
1648 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
1649 and PIC call sequences. Use ldo instead of blr to set return register
1650 in PIC call sequence.
1651
1652 2014-01-25 Walter Lee <walt@tilera.com>
1653
1654 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
1655 avoid clobbering a live register.
1656
1657 2014-01-25 Walter Lee <walt@tilera.com>
1658
1659 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
1660 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
1661 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
1662 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
1663
1664 2014-01-25 Walter Lee <walt@tilera.com>
1665
1666 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
1667 arguments on even registers.
1668 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
1669 STACK_BOUNDARY.
1670 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
1671 (BIGGEST_ALIGNMENT): Ditto.
1672 (BIGGEST_FIELD_ALIGNMENT): Ditto.
1673
1674 2014-01-25 Walter Lee <walt@tilera.com>
1675
1676 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
1677 insns before bundling.
1678 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
1679
1680 2014-01-25 Walter Lee <walt@tilera.com>
1681
1682 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
1683 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
1684 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
1685
1686 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
1687
1688 * config/mips/constraints.md (kl): Delete.
1689 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
1690 define expands, using...
1691 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
1692 instructions for MIPS16.
1693 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
1694 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
1695
1696 2014-01-25 Walter Lee <walt@tilera.com>
1697
1698 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
1699 (clzdi2): Ditto.
1700 (ffsdi2): Ditto.
1701
1702 2014-01-25 Walter Lee <walt@tilera.com>
1703
1704 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
1705 (TARGET_EXPAND_TO_RTL_HOOK): Define.
1706
1707 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
1708
1709 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
1710 Handle XOR.
1711
1712 2014-01-25 Jakub Jelinek <jakub@redhat.com>
1713
1714 * print-rtl.c (in_call_function_usage): New var.
1715 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
1716 EXPR_LIST mode as mode and not as reg note name.
1717
1718 PR middle-end/59561
1719 * cfgloopmanip.c (copy_loop_info): If
1720 loop->warned_aggressive_loop_optimizations, make sure
1721 the flag is set in target loop too.
1722
1723 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
1724
1725 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
1726 flag_cilkplus.
1727 * builtins.def: Likewise.
1728 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
1729 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
1730 * ira.c (ira_setup_eliminable_regset): Likewise.
1731 * omp-low.c (gate_expand_omp): Likewise.
1732 (execute_lower_omp): Likewise.
1733 (diagnose_sb_0): Likewise.
1734 (gate_diagnose_omp_blocks): Likewise.
1735 (simd_clone_clauses_extract): Likewise.
1736 (gate): Likewise.
1737
1738 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1739
1740 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
1741 correction for little endian...
1742 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
1743 here.
1744
1745 2014-01-24 Jeff Law <law@redhat.com>
1746
1747 PR tree-optimization/59919
1748 * tree-vrp.c (find_assert_locations_1): Do not register asserts
1749 for non-returning calls.
1750
1751 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
1752
1753 * common/config/aarch64/aarch64-common.c
1754 (aarch64_rewrite_mcpu): Handle multiple names.
1755 * config/aarch64/aarch64.h
1756 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
1757
1758 2014-01-24 Dodji Seketeli <dodji@redhat.com>
1759
1760 * input.c (add_file_to_cache_tab): Handle the case where fopen
1761 returns NULL.
1762
1763 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
1764
1765 PR target/59929
1766 * config/i386/i386.md (pushsf splitter): Get stack adjustment
1767 from push operand if code of push isn't PRE_DEC.
1768
1769 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1770
1771 PR target/59909
1772 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
1773 -mquad-memory-atomic. Update -mquad-memory documentation to say
1774 it is only used for non-atomic loads/stores.
1775
1776 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
1777 -mquad-memory or -mquad-memory-atomic switches.
1778
1779 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
1780 -mquad-memory-atomic to ISA 2.07 support.
1781
1782 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
1783 to separate support of normal quad word memory operations (ldq, stq)
1784 from the atomic quad word memory operations.
1785
1786 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
1787 support to separate non-atomic quad word operations from atomic
1788 quad word operations. Disable non-atomic quad word operations in
1789 little endian mode so that we don't have to swap words after the
1790 load and before the store.
1791 (quad_load_store_p): Add comment about atomic quad word support.
1792 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
1793 options printed with -mdebug=reg.
1794
1795 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
1796 -mquad-memory-atomic as the test for whether we have quad word
1797 atomic instructions.
1798 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
1799 or -mp8-vector are used, allow byte/half-word atomic operations.
1800
1801 * config/rs6000/sync.md (load_lockedti): Insure that the address
1802 is a proper indexed or indirect address for the lqarx instruction.
1803 On little endian systems, swap the hi/lo registers after the lqarx
1804 instruction.
1805 (load_lockedpti): Use indexed_or_indirect_operand predicate to
1806 insure the address is valid for the lqarx instruction.
1807 (store_conditionalti): Insure that the address is a proper indexed
1808 or indirect address for the stqcrx. instruction. On little endian
1809 systems, swap the hi/lo registers before doing the stqcrx.
1810 instruction.
1811 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
1812 insure the address is valid for the stqcrx. instruction.
1813
1814 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
1815 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
1816 type of quad memory support is available.
1817
1818 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
1819
1820 PR regression/59915
1821 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
1822 there is a danger of looping.
1823
1824 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
1825
1826 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
1827 force flag_ira_loop_pressure if set via command line.
1828
1829 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
1830
1831 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
1832 (ashr_simd): New builtin handling DI mode.
1833 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
1834 (aarch64_sshr_simddi): New match pattern.
1835 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
1836 (vshrd_n_s64): Likewise.
1837 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
1838
1839 2014-01-23 Nick Clifton <nickc@redhat.com>
1840
1841 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
1842 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
1843 favour of mcu specific scripts.
1844 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
1845 430x multilibs.
1846
1847 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1848 Alex Velenko <Alex.Velenko@arm.com>
1849
1850 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
1851 (vaddv_s16): Likewise.
1852 (vaddv_s32): Likewise.
1853 (vaddv_u8): Likewise.
1854 (vaddv_u16): Likewise.
1855 (vaddv_u32): Likewise.
1856 (vaddvq_s8): Likewise.
1857 (vaddvq_s16): Likewise.
1858 (vaddvq_s32): Likewise.
1859 (vaddvq_s64): Likewise.
1860 (vaddvq_u8): Likewise.
1861 (vaddvq_u16): Likewise.
1862 (vaddvq_u32): Likewise.
1863 (vaddvq_u64): Likewise.
1864 (vaddv_f32): Likewise.
1865 (vaddvq_f32): Likewise.
1866 (vaddvq_f64): Likewise.
1867 (vmaxv_f32): Likewise.
1868 (vmaxv_s8): Likewise.
1869 (vmaxv_s16): Likewise.
1870 (vmaxv_s32): Likewise.
1871 (vmaxv_u8): Likewise.
1872 (vmaxv_u16): Likewise.
1873 (vmaxv_u32): Likewise.
1874 (vmaxvq_f32): Likewise.
1875 (vmaxvq_f64): Likewise.
1876 (vmaxvq_s8): Likewise.
1877 (vmaxvq_s16): Likewise.
1878 (vmaxvq_s32): Likewise.
1879 (vmaxvq_u8): Likewise.
1880 (vmaxvq_u16): Likewise.
1881 (vmaxvq_u32): Likewise.
1882 (vmaxnmv_f32): Likewise.
1883 (vmaxnmvq_f32): Likewise.
1884 (vmaxnmvq_f64): Likewise.
1885 (vminv_f32): Likewise.
1886 (vminv_s8): Likewise.
1887 (vminv_s16): Likewise.
1888 (vminv_s32): Likewise.
1889 (vminv_u8): Likewise.
1890 (vminv_u16): Likewise.
1891 (vminv_u32): Likewise.
1892 (vminvq_f32): Likewise.
1893 (vminvq_f64): Likewise.
1894 (vminvq_s8): Likewise.
1895 (vminvq_s16): Likewise.
1896 (vminvq_s32): Likewise.
1897 (vminvq_u8): Likewise.
1898 (vminvq_u16): Likewise.
1899 (vminvq_u32): Likewise.
1900 (vminnmv_f32): Likewise.
1901 (vminnmvq_f32): Likewise.
1902 (vminnmvq_f64): Likewise.
1903
1904 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1905
1906 * config/aarch64/aarch64-simd.md
1907 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
1908 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
1909 (*aarch64_mul3_elt<mode>): Likewise.
1910 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
1911 (*aarch64_mul3_elt_to_64v2df): Likewise.
1912 (*aarch64_mla_elt<mode>): Likewise.
1913 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1914 (*aarch64_mls_elt<mode>): Likewise.
1915 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1916 (*aarch64_fma4_elt<mode>): Likewise.
1917 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1918 (*aarch64_fma4_elt_to_64v2df): Likewise.
1919 (*aarch64_fnma4_elt<mode>): Likewise.
1920 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1921 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1922 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1923 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1924 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1925 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1926 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1927 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1928 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1929
1930 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
1931
1932 * config/aarch64/aarch64-simd.md
1933 (aarch64_be_checked_get_lane<mode>): New define_expand.
1934 * config/aarch64/aarch64-simd-builtins.def
1935 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
1936 New builtin definition.
1937 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
1938 Use new safe be builtin.
1939
1940 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
1941
1942 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
1943 New define_insn.
1944 (aarch64_be_st1<mode>): Likewise.
1945 (aarch_ld1<VALL:mode>): Define_expand modified.
1946 (aarch_st1<VALL:mode>): Likewise.
1947 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
1948 (UNSPEC_ST1): Likewise.
1949
1950 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
1951
1952 * config/microblaze/microblaze.md: Add trap insn and attribute
1953
1954 2014-01-23 Dodji Seketeli <dodji@redhat.com>
1955
1956 PR preprocessor/58580
1957 * input.h (location_get_source_line): Take an additional line_size
1958 parameter.
1959 (void diagnostics_file_cache_fini): Declare new function.
1960 * input.c (struct fcache): New type.
1961 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
1962 New static constants.
1963 (diagnostic_file_cache_init, total_lines_num)
1964 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
1965 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
1966 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
1967 (get_next_line, read_next_line, goto_next_line, read_line_num):
1968 New static function definitions.
1969 (diagnostic_file_cache_fini): New function.
1970 (location_get_source_line): Take an additional output line_len
1971 parameter. Re-write using lookup_or_add_file_to_cache_tab and
1972 read_line_num.
1973 * diagnostic.c (diagnostic_finish): Call
1974 diagnostic_file_cache_fini.
1975 (adjust_line): Take an additional input parameter for the length
1976 of the line, rather than calculating it with strlen.
1977 (diagnostic_show_locus): Adjust the use of
1978 location_get_source_line and adjust_line with respect to their new
1979 signature. While displaying a line now, do not stop at the first
1980 null byte. Rather, display the zero byte as a space and keep
1981 going until we reach the size of the line.
1982 * Makefile.in: Add vec.o to OBJS-libcommon
1983
1984 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
1985 Ilya Tocar <ilya.tocar@intel.com>
1986
1987 * config/i386/avx512fintrin.h (_mm512_kmov): New.
1988 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
1989 (__builtin_ia32_kmov16): Ditto.
1990 * config/i386/i386.md (UNSPEC_KMOV): New.
1991 (kmovw): Ditto.
1992
1993 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
1994
1995 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
1996 (_mm512_storeu_si512): Ditto.
1997
1998 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
1999
2000 PR target/52125
2001 * rtl.h (get_referenced_operands): Declare.
2002 * recog.c (get_referenced_operands): New function.
2003 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
2004 operands have been referenced when recording LO_SUM references.
2005
2006 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
2007
2008 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
2009
2010 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
2011
2012 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
2013 Enable for generic and recent AMD targets.
2014
2015 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
2016
2017 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
2018 ARG_SIZE note when adjustment was eliminated.
2019
2020 2014-01-22 Jeff Law <law@redhat.com>
2021
2022 PR tree-optimization/59597
2023 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
2024 in file. Accept new argument REGISTERING and use it to modify
2025 dump output appropriately.
2026 (register_jump_thread): Corresponding changes.
2027 (mark_threaded_blocks): Reinstate code to cancel unprofitable
2028 thread paths involving joiner blocks. Add code to dump cancelled
2029 jump threading paths.
2030
2031 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
2032
2033 PR rtl-optimization/59477
2034 * lra-constraints.c (inherit_in_ebb): Process call for living hard
2035 regs. Update reloads_num and potential_reload_hard_regs for all insns.
2036
2037 2014-01-22 Tom Tromey <tromey@redhat.com>
2038
2039 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
2040 PARAMS.
2041 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
2042
2043 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
2044
2045 PR rtl-optimization/59896
2046 * lra-constraints.c (process_alt_operands): Check unused note for
2047 matched operands of insn with no output reloads.
2048
2049 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
2050
2051 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
2052 (mips_move_from_gpr_cost): Likewise.
2053
2054 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
2055
2056 PR rtl-optimization/59858
2057 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
2058 ira_class_hard_regs_num.
2059 (process_alt_operands): Increase reject for dying matched operand.
2060
2061 2014-01-21 Jakub Jelinek <jakub@redhat.com>
2062
2063 PR target/59003
2064 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
2065 smaller than size, perform several stores or loads and stores
2066 at dst + count - size to store or copy all of size bytes, rather
2067 than just last modesize bytes.
2068
2069 2014-01-20 DJ Delorie <dj@redhat.com>
2070
2071 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
2072 that CLOBBERs are REGs before propogating their values.
2073
2074 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
2075
2076 PR middle-end/59789
2077 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
2078 (cgraph_inline_failed_type): New function.
2079 * cgraph.h (DEFCIFCODE): Add type.
2080 (cgraph_inline_failed_type_t): New enum.
2081 (cgraph_inline_failed_type): New prototype.
2082 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
2083 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
2084 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
2085 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
2086 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
2087 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
2088 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
2089 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
2090 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
2091 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
2092 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
2093 OPTIMIZATION_MISMATCH.
2094 * tree-inline.c (expand_call_inline): Emit errors during
2095 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
2096
2097 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
2098
2099 PR target/59685
2100 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
2101 mode attribute in insn output.
2102
2103 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
2104
2105 * output.h (output_constant): Delete.
2106 * varasm.c (output_constant): Make private.
2107
2108 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
2109
2110 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
2111
2112 2014-01-20 Jakub Jelinek <jakub@redhat.com>
2113
2114 PR middle-end/59860
2115 * tree.h (fold_builtin_strcat): New prototype.
2116 * builtins.c (fold_builtin_strcat): No longer static. Add len
2117 argument, if non-NULL, don't call c_strlen. Optimize
2118 directly into __builtin_memcpy instead of __builtin_strcpy.
2119 (fold_builtin_2): Adjust fold_builtin_strcat caller.
2120 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
2121
2122 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
2123
2124 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2125 for SImode_address_operand operands, having only a REG argument.
2126
2127 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
2128
2129 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
2130 loader name using mbig-endian.
2131 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
2132
2133 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
2134
2135 * doc/invoke.texi (-march): Clarify documentation for AArch64.
2136 (-mtune): Likewise.
2137 (-mcpu): Likewise.
2138
2139 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
2140
2141 * config/aarch64/aarch64-protos.h
2142 (aarch64_cannot_change_mode_class_ptr): Declare.
2143 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
2144 aarch64_cannot_change_mode_class_ptr): New.
2145 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
2146 backend hook aarch64_cannot_change_mode_class.
2147
2148 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
2149
2150 * common/config/aarch64/aarch64-common.c
2151 (aarch64_handle_option): Don't handle any option order logic here.
2152 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
2153 selected_cpu, warn on architecture version mismatch.
2154 (aarch64_override_options): Fix parsing order for option strings.
2155
2156 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2157 Iain Sandoe <iain@codesourcery.com>
2158
2159 PR bootstrap/59496
2160 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
2161 warning. Amend comment to reflect current functionality.
2162
2163 2014-01-20 Richard Biener <rguenther@suse.de>
2164
2165 PR middle-end/59860
2166 * builtins.c (fold_builtin_strcat): Remove case better handled
2167 by tree-ssa-strlen.c.
2168
2169 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
2170
2171 * config/aarch64/aarch64.opt
2172 (mcpu, march, mtune): Make case-insensitive.
2173
2174 2014-01-20 Jakub Jelinek <jakub@redhat.com>
2175
2176 PR target/59880
2177 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
2178 if operands[1] is a REG or ZERO_EXTEND of a REG.
2179
2180 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
2181
2182 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
2183
2184 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
2185
2186 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
2187 long non-pic millicode calls.
2188
2189 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2190
2191 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
2192
2193 2014-01-19 Kito Cheng <kito@0xlab.org>
2194
2195 * builtins.c (expand_movstr): Check movstr expand done or fail.
2196
2197 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2198 H.J. Lu <hongjiu.lu@intel.com>
2199
2200 PR target/59379
2201 * config/i386/i386.md (*lea<mode>): Zero-extend return register
2202 to DImode for zero-extended addresses.
2203
2204 2014-01-19 Jakub Jelinek <jakub@redhat.com>
2205
2206 PR rtl-optimization/57763
2207 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
2208 on the new indirect jump_insn and increment LABEL_NUSES (label).
2209
2210 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
2211
2212 PR bootstrap/59580
2213 PR bootstrap/59583
2214 * config.gcc (x86_archs): New variable.
2215 (x86_64_archs): Likewise.
2216 (x86_cpus): Likewise.
2217 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
2218 --with-arch/--with-cpu= options.
2219 Support --with-arch=/--with-cpu={nehalem,westmere,
2220 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
2221
2222 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2223
2224 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
2225 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
2226
2227 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
2228
2229 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
2230
2231 2014-01-18 Jakub Jelinek <jakub@redhat.com>
2232
2233 PR target/58944
2234 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
2235 clear cpp_get_options (parse_in)->warn_unused_macros for
2236 ix86_target_macros_internal with cpp_define.
2237
2238 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
2239
2240 * jump.c (delete_related_insns): Keep (use (insn))s.
2241 * reorg.c (redundant_insn): Check for barriers too.
2242
2243 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2244
2245 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
2246
2247 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
2248
2249 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
2250 call to $$dyncall when TARGET_LONG_CALLS is true.
2251
2252 2014-01-17 Jeff Law <law@redhat.com>
2253
2254 * ree.c (combine_set_extension): Temporarily disable test for
2255 changing number of hard registers.
2256
2257 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
2258
2259 PR middle-end/58125
2260 * ipa-inline-analysis.c (inline_free_summary):
2261 Do not free summary of aliases.
2262
2263 2014-01-17 Jakub Jelinek <jakub@redhat.com>
2264
2265 PR middle-end/59706
2266 * gimplify.c (gimplify_expr): Use create_tmp_var
2267 instead of create_tmp_var_raw. If cond doesn't have
2268 integral type, don't add the IFN_ANNOTATE builtin at all.
2269
2270 2014-01-17 Martin Jambor <mjambor@suse.cz>
2271
2272 PR ipa/59736
2273 * ipa-cp.c (prev_edge_clone): New variable.
2274 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
2275 Also resize prev_edge_clone vector.
2276 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
2277 (ipcp_edge_removal_hook): New function.
2278 (ipcp_driver): Register ipcp_edge_removal_hook.
2279
2280 2014-01-17 Andrew Pinski <apinski@cavium.com>
2281 Steve Ellcey <sellcey@mips.com>
2282
2283 PR target/59462
2284 * config/mips/mips.c (mips_print_operand): Check operand mode instead
2285 of operator mode.
2286
2287 2014-01-17 Jeff Law <law@redhat.com>
2288
2289 PR middle-end/57904
2290 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
2291 so that pass_ccp runs first.
2292
2293 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2294
2295 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
2296 (ix86_adjust_cost): Use !TARGET_XXX.
2297 (do_reorder_for_imul): Likewise.
2298 (swap_top_of_ready_list): Likewise.
2299 (ix86_sched_reorder): Likewise.
2300
2301 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
2302
2303 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2304 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
2305 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
2306 (intel_memset): New. Duplicate slm_memset.
2307 (intel_cost): New. Duplicate slm_cost.
2308 (m_INTEL): New macro.
2309 (processor_target_table): Add "intel".
2310 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
2311 with PROCESSOR_INTEL for "intel".
2312 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
2313 PROCESSOR_SILVERMONT.
2314 (ix86_issue_rate): Likewise.
2315 (ix86_adjust_cost): Likewise.
2316 (ia32_multipass_dfa_lookahead): Likewise.
2317 (swap_top_of_ready_list): Likewise.
2318 (ix86_sched_reorder): Likewise.
2319 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
2320 instead of TARGET_OPT_AGU.
2321 * config/i386/i386.h (TARGET_INTEL): New.
2322 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
2323 (processor_type): Add PROCESSOR_INTEL.
2324 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
2325 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
2326
2327 2014-01-17 Marek Polacek <polacek@redhat.com>
2328
2329 PR c/58346
2330 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
2331 size is zero.
2332
2333 2014-01-17 Richard Biener <rguenther@suse.de>
2334
2335 PR tree-optimization/46590
2336 * opts.c (default_options_table): Add entries for
2337 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
2338 all enabled at -O1 but not for -Og.
2339 * common.opt (fbranch-count-reg): Remove Init(1).
2340 (fmove-loop-invariants): Likewise.
2341 (ftree-pta): Likewise.
2342
2343 2014-01-17 Jakub Jelinek <jakub@redhat.com>
2344
2345 * config/i386/i386.c (ix86_data_alignment): For compatibility with
2346 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
2347 decls to at least the GCC 4.8 used alignments.
2348
2349 PR fortran/59440
2350 * tree-nested.c (convert_nonlocal_reference_stmt,
2351 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
2352 of GIMPLE_BIND stmts, adjust associated decls.
2353
2354 2014-01-17 Richard Biener <rguenther@suse.de>
2355
2356 PR tree-optimization/46590
2357 * vec.h (vec<>::bseach): New member function implementing
2358 binary search according to C89 bsearch.
2359 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
2360 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
2361 bitmap pointer again. Make accesses_in_loop a flat array.
2362 (mem_ref_obstack): New global.
2363 (outermost_indep_loop): Adjust for mem_ref->stored changes.
2364 (mark_ref_stored): Likewise.
2365 (ref_indep_loop_p_2): Likewise.
2366 (set_ref_stored_in_loop): New helper function.
2367 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
2368 (memref_free): Adjust.
2369 (record_mem_ref_loc): Simplify.
2370 (gather_mem_refs_stmt): Adjust.
2371 (sort_locs_in_loop_postorder_cmp): New function.
2372 (analyze_memory_references): Sort accesses_in_loop after
2373 loop postorder number.
2374 (find_ref_loc_in_loop_cmp): New function.
2375 (for_all_locs_in_loop): Find relevant cluster of locs in
2376 accesses_in_loop and iterate without recursion.
2377 (execute_sm): Avoid uninit warning.
2378 (struct ref_always_accessed): Simplify.
2379 (ref_always_accessed::operator ()): Likewise.
2380 (ref_always_accessed_p): Likewise.
2381 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
2382 loop postorder numbers here.
2383 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
2384 numbers.
2385
2386 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
2387
2388 PR c++/57945
2389 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
2390 on decls for which assemble_alias has been called.
2391
2392 2014-01-17 Nick Clifton <nickc@redhat.com>
2393
2394 * config/msp430/msp430.opt: (mcpu): New option.
2395 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
2396 (msp430_option_override): Parse target_cpu. If the MCU name
2397 matches a generic string, clear target_mcu.
2398 (msp430_attr): Allow numeric interrupt values up to 63.
2399 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
2400 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
2401 option.
2402 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
2403 Add mcpu matches.
2404 * config/msp430/msp430.md (popm): Use %J rather than %I.
2405 (addsi3): Use msp430_nonimmediate_operand for operand 2.
2406 (addhi_cy_i): Use immediate_operand for operand 2.
2407 * doc/invoke.texi: Document -mcpu option.
2408
2409 2014-01-17 Richard Biener <rguenther@suse.de>
2410
2411 PR rtl-optimization/38518
2412 * df.h (df_analyze_loop): Declare.
2413 * df-core.c: Include cfgloop.h.
2414 (df_analyze_1): Split out main part of df_analyze.
2415 (df_analyze): Adjust.
2416 (loop_inverted_post_order_compute): New function.
2417 (loop_post_order_compute): Likewise.
2418 (df_analyze_loop): New function avoiding whole-function
2419 postorder computes.
2420 * loop-invariant.c (find_defs): Use df_analyze_loop.
2421 (find_invariants): Adjust.
2422 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
2423
2424 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
2425
2426 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
2427 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
2428
2429 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
2430
2431 * ipa-ref.c (ipa_remove_stmt_references): Fix references
2432 traversal when removing references.
2433
2434 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
2435
2436 PR ipa/59775
2437 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
2438
2439 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
2440
2441 PR middle-end/56791
2442 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
2443 pushing a reload for an autoinc when we had previously reloaded an
2444 inner part of the address.
2445
2446 2014-01-16 Jakub Jelinek <jakub@redhat.com>
2447
2448 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
2449 field.
2450 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
2451 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
2452 when not giving up or versioning for alias only because of
2453 loop->safelen.
2454 (vect_analyze_data_ref_dependences): Set to true.
2455 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
2456 is a GIMPLE_PHI.
2457 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
2458 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
2459 to the condition.
2460
2461 PR middle-end/58344
2462 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
2463
2464 PR target/59839
2465 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
2466 operand 0 predicate for gathers, use a new pseudo as subtarget.
2467
2468 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
2469
2470 PR middle-end/59609
2471 * lra-constraints.c (process_alt_operands): Add printing debug info.
2472 Check absence of input/output reloads for matched operands too.
2473
2474 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
2475
2476 PR rtl-optimization/59835
2477 * ira.c (ira_init_register_move_cost): Increase cost for
2478 impossible modes.
2479
2480 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
2481
2482 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
2483
2484 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
2485
2486 PR target/59780
2487 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
2488 non-register objects. Use gen_(high/low)part more consistently.
2489 Fix assertions.
2490
2491 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2492
2493 PR target/59844
2494 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
2495 endian support, remove tests for WORDS_BIG_ENDIAN.
2496 (p8_mfvsrd_3_<mode>): Likewise.
2497 (reload_gpr_from_vsx<mode>): Likewise.
2498 (reload_gpr_from_vsxsf): Likewise.
2499 (p8_mfvsrd_4_disf): Likewise.
2500
2501 2014-01-16 Richard Biener <rguenther@suse.de>
2502
2503 PR rtl-optimization/46590
2504 * lcm.c (compute_antinout_edge): Use postorder iteration.
2505 (compute_laterin): Use inverted postorder iteration.
2506
2507 2014-01-16 Nick Clifton <nickc@redhat.com>
2508
2509 PR middle-end/28865
2510 * varasm.c (output_constant): Return the number of bytes actually
2511 emitted.
2512 (output_constructor_array_range): Update the field size with the
2513 number of bytes emitted by output_constant.
2514 (output_constructor_regular_field): Likewise. Also do not
2515 complain if the total number of bytes emitted is now greater
2516 than the expected fieldpos.
2517 * output.h (output_constant): Update prototype and descriptive comment.
2518
2519 2014-01-16 Marek Polacek <polacek@redhat.com>
2520
2521 PR middle-end/59827
2522 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
2523 it is error_mark_node.
2524
2525 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
2526
2527 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
2528 VALID_AVX256_REG_OR_OI_MODE.
2529
2530 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
2531
2532 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
2533 current procedure should be profiled.
2534
2535 2014-01-15 Andrew Pinski <apinski@cavium.com>
2536
2537 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
2538 of moving from/to the STACK_REG register class.
2539
2540 2014-01-15 Richard Henderson <rth@redhat.com>
2541
2542 PR debug/54694
2543 * reginfo.c (global_regs_decl): Globalize.
2544 * rtl.h (global_regs_decl): Declare.
2545 * ira.c (do_reload): Diagnose frame_pointer_needed and it
2546 reserved via global_regs.
2547
2548 2014-01-15 Teresa Johnson <tejohnson@google.com>
2549
2550 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
2551
2552 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2553
2554 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
2555 and vmulosh rather than call gen_vec_widen_smult_*.
2556 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
2557 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
2558 (vec_widen_smult_even_v16qi): Likewise.
2559 (vec_widen_umult_even_v8hi): Likewise.
2560 (vec_widen_smult_even_v8hi): Likewise.
2561 (vec_widen_umult_odd_v16qi): Likewise.
2562 (vec_widen_smult_odd_v16qi): Likewise.
2563 (vec_widen_umult_odd_v8hi): Likewise.
2564 (vec_widen_smult_odd_v8hi): Likewise.
2565 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
2566 vmuloub rather than call gen_vec_widen_umult_*.
2567 (vec_widen_umult_lo_v16qi): Likewise.
2568 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
2569 vmulosb rather than call gen_vec_widen_smult_*.
2570 (vec_widen_smult_lo_v16qi): Likewise.
2571 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
2572 rather than call gen_vec_widen_umult_*.
2573 (vec_widen_umult_lo_v8hi): Likewise.
2574 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
2575 rather than call gen_vec_widen_smult_*.
2576 (vec_widen_smult_lo_v8hi): Likewise.
2577
2578 2014-01-15 Jeff Law <law@redhat.com>
2579
2580 PR tree-optimization/59747
2581 * ree.c (find_and_remove_re): Properly handle case where a second
2582 eliminated extension requires widening a copy created for elimination
2583 of a prior extension.
2584 (combine_set_extension): Ensure that the number of hard regs needed
2585 for a destination register does not change when we widen it.
2586
2587 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
2588
2589 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
2590 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
2591 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
2592 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
2593 (avr-*-rtems*): Likewise.
2594 (bfin*-rtems*): Likewise.
2595 (moxie-*-rtems*): Likewise.
2596 (h8300-*-rtems*): Likewise.
2597 (i[34567]86-*-rtems*): Likewise.
2598 (lm32-*-rtems*): Likewise.
2599 (m32r-*-rtems*): Likewise.
2600 (m68k-*-rtems*): Likewise.
2601 (microblaze*-*-rtems*): Likewise.
2602 (mips*-*-rtems*): Likewise.
2603 (powerpc-*-rtems*): Likewise.
2604 (sh-*-rtems*): Likewise.
2605 (sparc-*-rtems*): Likewise.
2606 (sparc64-*-rtems*): Likewise.
2607 (v850-*-rtems*): Likewise.
2608 (m32c-*-rtems*): Likewise.
2609
2610 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
2611
2612 PR rtl-optimization/59511
2613 * ira.c (ira_init_register_move_cost): Use memory costs for some
2614 cases of register move cost calculations.
2615 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
2616 instead of BB frequency.
2617 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
2618 * lra-assigns.c (find_hard_regno_for): Ditto.
2619
2620 2014-01-15 Richard Biener <rguenther@suse.de>
2621
2622 PR tree-optimization/59822
2623 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
2624 (vectorizable_load): Use it to hoist defs of uses of invariant
2625 loads out of the loop.
2626
2627 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
2628 Kugan Vivekanandarajah <kuganv@linaro.org>
2629
2630 PR target/59695
2631 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
2632 truncation.
2633
2634 2014-01-15 Richard Biener <rguenther@suse.de>
2635
2636 PR rtl-optimization/59802
2637 * lcm.c (compute_available): Use inverted postorder to seed
2638 the initial worklist.
2639
2640 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2641
2642 PR target/59803
2643 * config/s390/s390.c (s390_preferred_reload_class): Don't return
2644 ADDR_REGS for invalid symrefs in non-PIC code.
2645
2646 2014-01-15 Jakub Jelinek <jakub@redhat.com>
2647
2648 PR other/58712
2649 * builtins.c (determine_block_size): Initialize *probable_max_size
2650 even if len_rtx is CONST_INT.
2651
2652 2014-01-14 Andrew Pinski <apinski@cavium.com>
2653
2654 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
2655 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
2656 (cortexa53_tunings): Likewise.
2657 (aarch64_sched_issue_rate): New function.
2658 (TARGET_SCHED_ISSUE_RATE): Define.
2659
2660 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
2661
2662 * ira-costs.c (find_costs_and_classes): Add missed
2663 ira_init_register_move_cost_if_necessary.
2664
2665 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
2666
2667 PR target/59787
2668 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
2669
2670 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
2671
2672 PR target/59794
2673 * config/i386/i386.c (type_natural_mode): Add a bool parameter
2674 to indicate if type is used for function return value. Warn ABI
2675 change if the vector mode isn't available for function return value.
2676 (ix86_function_arg_advance): Pass false to type_natural_mode.
2677 (ix86_function_arg): Likewise.
2678 (ix86_gimplify_va_arg): Likewise.
2679 (function_arg_32): Don't warn ABI change.
2680 (ix86_function_value): Pass true to type_natural_mode.
2681 (ix86_return_in_memory): Likewise.
2682 (ix86_struct_value_rtx): Removed.
2683 (TARGET_STRUCT_VALUE_RTX): Likewise.
2684
2685 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2686
2687 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
2688 converting a conditional jump into a conditional return.
2689
2690 2014-01-14 Richard Biener <rguenther@suse.de>
2691
2692 PR tree-optimization/58921
2693 PR tree-optimization/59006
2694 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
2695 hoisting invariant stmts.
2696 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
2697 invariant loads on the preheader edge if possible.
2698
2699 2014-01-14 Joey Ye <joey.ye@arm.com>
2700
2701 * doc/plugin.texi (Building GCC plugins): Update to C++.
2702
2703 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
2704
2705 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
2706 (_mm_rcp28_round_ss): Ditto.
2707 (_mm_rsqrt28_round_sd): Ditto.
2708 (_mm_rsqrt28_round_ss): Ditto.
2709 (_mm_rcp28_sd): Ditto.
2710 (_mm_rcp28_ss): Ditto.
2711 (_mm_rsqrt28_sd): Ditto.
2712 (_mm_rsqrt28_ss): Ditto.
2713 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
2714 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
2715 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
2716 (IX86_BUILTIN_RCP28SD): Ditto.
2717 (IX86_BUILTIN_RCP28SS): Ditto.
2718 (IX86_BUILTIN_RSQRT28SD): Ditto.
2719 (IX86_BUILTIN_RSQRT28SS): Ditto.
2720 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
2721 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
2722 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
2723 (ix86_expand_special_args_builtin): Expand new FTYPE.
2724 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
2725 (srcp14<mode>): Make insn unary.
2726 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
2727 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
2728 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
2729 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
2730 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
2731 Fix rounding: make it SAE only.
2732 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
2733 Ditto.
2734 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
2735 Ditto.
2736 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
2737 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
2738 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
2739 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
2740 (round_saeonly_mask_scalar_operand4): Ditto.
2741 (round_saeonly_mask_scalar_op3): Ditto.
2742 (round_saeonly_mask_scalar_op4): Ditto.
2743
2744 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2745
2746 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
2747 Implement -maltivec=be for vec_insert and vec_extract.
2748
2749 2014-01-10 DJ Delorie <dj@redhat.com>
2750
2751 * config/msp430/msp430.md (call_internal): Don't allow memory
2752 references with SP as the base register.
2753 (call_value_internal): Likewise.
2754 * config/msp430/constraints.md (Yc): New. For memory references
2755 that don't use SP as a base register.
2756
2757 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
2758 "an integer without a # prefix"
2759 * config/msp430/msp430.md (epilogue_helper): Use it.
2760
2761 2014-01-13 Jakub Jelinek <jakub@redhat.com>
2762
2763 PR target/59617
2764 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
2765 AVX512F gather builtins.
2766 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
2767 on gather decls with INTEGER_TYPE masktype.
2768 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
2769 directly into the builtin rather than hoisting it before loop.
2770
2771 PR tree-optimization/59387
2772 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
2773 (scev_const_prop): If folded_casts and type has undefined overflow,
2774 use force_gimple_operand instead of force_gimple_operand_gsi and
2775 for each added stmt if it is assign with
2776 arith_code_with_undefined_signed_overflow, call
2777 rewrite_to_defined_overflow.
2778 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
2779 gimple-fold.h instead.
2780 (arith_code_with_undefined_signed_overflow,
2781 rewrite_to_defined_overflow): Moved to ...
2782 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
2783 rewrite_to_defined_overflow): ... here. No longer static.
2784 Include gimplify-me.h.
2785 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
2786 rewrite_to_defined_overflow): New prototypes.
2787
2788 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2789
2790 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
2791
2792 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
2793
2794 * builtins.c (get_object_alignment_2): Minor tweak.
2795 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
2796
2797 2014-01-13 Christian Bruel <christian.bruel@st.com>
2798
2799 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
2800 optimized non constant lengths.
2801
2802 2014-01-13 Jakub Jelinek <jakub@redhat.com>
2803
2804 PR libgomp/59194
2805 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
2806 load as __atomic_load_N if possible.
2807
2808 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
2809
2810 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
2811 target parameter.
2812 (rs6000_expand_builtin): Adjust call.
2813
2814 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
2815
2816 PR target/58115
2817 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
2818 * config/rs6000/rs6000.c: Include target-globals.h.
2819 (rs6000_set_current_function): Instead of doing target_reinit
2820 unconditionally, use save_target_globals_default_opts and
2821 restore_target_globals.
2822
2823 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
2824 FPSCR.
2825 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
2826 (rs6000_expand_builtin): Handle mffs and mtfsf.
2827 (rs6000_init_builtins): Define mffs and mtfsf.
2828 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
2829 (rs6000_mffs): New pattern.
2830 (rs6000_mtfsf): New pattern.
2831
2832 2014-01-11 Bin Cheng <bin.cheng@arm.com>
2833
2834 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
2835 Start narrowing with START. Apply candidate-use pair
2836 and check overall cost in narrowing.
2837 (iv_ca_prune): Pass new argument.
2838
2839 2014-01-10 Jeff Law <law@redhat.com>
2840
2841 PR middle-end/59743
2842 * ree.c (combine_reaching_defs): Ensure the defining statement
2843 occurs before the extension when optimizing extensions with
2844 different source and destination hard registers.
2845
2846 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
2847
2848 PR ipa/58585
2849 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
2850 vtables into the type inheritance graph.
2851
2852 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2853
2854 PR rtl-optimization/59754
2855 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
2856 modes in the REGNO != REGNO case.
2857
2858 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2859
2860 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
2861
2862 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2863
2864 PR tree-optimization/59745
2865 * tree-predcom.c (tree_predictive_commoning_loop): Call
2866 free_affine_expand_cache if giving up because components is NULL.
2867
2868 * target-globals.c (save_target_globals): Allocate < 4KB structs using
2869 GC in payload of target_globals struct instead of allocating them on
2870 the heap and the larger structs separately using GC.
2871 * target-globals.h (struct target_globals): Make regs, hard_regs,
2872 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
2873 of GTY((skip)) and change type to void *.
2874 (reset_target_globals): Cast loads from those fields to corresponding
2875 types.
2876
2877 2014-01-10 Steve Ellcey <sellcey@mips.com>
2878
2879 PR plugins/59335
2880 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
2881 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
2882 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
2883
2884 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
2885
2886 PR target/59744
2887 * aarch64-modes.def (CC_Zmode): New flags mode.
2888 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
2889 represents an equality.
2890 (aarch64_get_condition_code): Handle CC_Zmode.
2891 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
2892
2893 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2894
2895 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
2896 extraction in good case.
2897
2898 2014-01-10 Richard Biener <rguenther@suse.de>
2899
2900 PR tree-optimization/59374
2901 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
2902 checking after SLP discovery. Mark stmts not participating
2903 in any SLP instance properly.
2904
2905 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2906
2907 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
2908 when handling a SET rtx.
2909
2910 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2911
2912 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
2913 (cortex-a57): Likewise.
2914 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
2915
2916 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2917
2918 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
2919 non-iwmmxt builtins.
2920
2921 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
2922
2923 PR ipa/58252
2924 PR ipa/59226
2925 * ipa-devirt.c record_target_from_binfo): Take as argument
2926 stack of binfos and lookup matching one for virtual inheritance.
2927 (possible_polymorphic_call_targets_1): Update.
2928
2929 2014-01-10 Huacai Chen <chenhc@lemote.com>
2930
2931 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
2932 kernel strings for Loongson-2E/2F/3A.
2933
2934 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2935
2936 PR middle-end/59670
2937 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
2938 is_gimple_call before calling gimple_call_internal_p.
2939
2940 2014-01-09 Steve Ellcey <sellcey@mips.com>
2941
2942 * Makefile.in (TREE_FLOW_H): Remove.
2943 (TREE_SSA_H): Add file names from tree-flow.h.
2944 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
2945 * tree.h: Remove tree-flow.h reference.
2946 * hash-table.h: Remove tree-flow.h reference.
2947 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
2948 reference with tree-ssa-loop.h.
2949
2950 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2951
2952 * doc/invoke.texi: Add -maltivec={be,le} options, and document
2953 default element-order behavior for -maltivec.
2954 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
2955 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
2956 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
2957 when targeting big endian, at least for now.
2958 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
2959
2960 2014-01-09 Jakub Jelinek <jakub@redhat.com>
2961
2962 PR middle-end/47735
2963 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
2964 var satisfies use_register_for_decl, just take into account type
2965 alignment, rather than decl alignment.
2966
2967 PR tree-optimization/59622
2968 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
2969 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
2970 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
2971 Don't devirtualize for inplace at all. For targets.length () == 1,
2972 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
2973
2974 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
2975
2976 * config/i386/i386.md (cpu): Remove the unused btver1.
2977
2978 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
2979
2980 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
2981
2982 2014-01-09 Jakub Jelinek <jakub@redhat.com>
2983
2984 PR target/58115
2985 * tree-core.h (struct target_globals): New forward declaration.
2986 (struct tree_target_option): Add globals field.
2987 * tree.h (TREE_TARGET_GLOBALS): Define.
2988 (prepare_target_option_nodes_for_pch): New prototype.
2989 * target-globals.h (struct target_globals): Define even if
2990 !SWITCHABLE_TARGET.
2991 * tree.c (prepare_target_option_node_for_pch,
2992 prepare_target_option_nodes_for_pch): New functions.
2993 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
2994 * config/i386/i386.c: Include target-globals.h.
2995 (ix86_set_current_function): Instead of doing target_reinit
2996 unconditionally, use save_target_globals_default_opts and
2997 restore_target_globals.
2998
2999 2014-01-09 Richard Biener <rguenther@suse.de>
3000
3001 PR tree-optimization/59715
3002 * tree-cfg.h (split_critical_edges): Declare.
3003 * tree-cfg.c (split_critical_edges): Export.
3004 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
3005
3006 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
3007
3008 * cfgexpand.c (expand_stack_vars): Optionally disable
3009 asan stack protection.
3010 (expand_used_vars): Likewise.
3011 (partition_stack_vars): Likewise.
3012 * asan.c (asan_emit_stack_protection): Optionally disable
3013 after return stack usage.
3014 (instrument_derefs): Optionally disable memory access instrumentation.
3015 (instrument_builtin_call): Likewise.
3016 (instrument_strlen_call): Likewise.
3017 (asan_protect_global): Optionally disable global variables protection.
3018 * doc/invoke.texi: Added doc for new options.
3019 * params.def: Added new options.
3020 * params.h: Likewise.
3021
3022 2014-01-09 Jakub Jelinek <jakub@redhat.com>
3023
3024 PR rtl-optimization/59724
3025 * ifcvt.c (cond_exec_process_if_block): Don't call
3026 flow_find_head_matching_sequence with 0 longest_match.
3027 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
3028 non-active insns if !stop_after.
3029 (try_head_merge_bb): Revert 2014-01-07 changes.
3030
3031 2014-01-08 Jeff Law <law@redhat.com>
3032
3033 * ree.c (get_sub_rtx): New function, extracted from...
3034 (merge_def_and_ext): Here.
3035 (combine_reaching_defs): Use get_sub_rtx.
3036
3037 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
3038
3039 * cgraph.h (varpool_variable_node): Do not choke on null node.
3040
3041 2014-01-08 Catherine Moore <clm@codesourcery.com>
3042
3043 * config/mips/mips.md (simple_return): Attempt to use JRC
3044 for microMIPS.
3045 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
3046
3047 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
3048
3049 PR rtl-optimization/59137
3050 * reorg.c (steal_delay_list_from_target): Call update_block for
3051 elided insns.
3052 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
3053
3054 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3055
3056 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
3057 two duplicate entries.
3058
3059 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
3060
3061 Revert:
3062 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
3063
3064 * config/mips/mips.c (mips_truncated_op_cost): New function.
3065 (mips_rtx_costs): Adjust test for BADDU.
3066 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
3067
3068 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
3069
3070 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
3071 (*baddu_si): ...this new pattern.
3072
3073 2014-01-08 Jakub Jelinek <jakub@redhat.com>
3074
3075 PR ipa/59722
3076 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
3077
3078 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
3079
3080 PR middle-end/57748
3081 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
3082 inner_reference_p.
3083 (expand_expr, expand_normal): Adjust.
3084 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
3085 inner_reference_p. Use inner_reference_p to expand inner references.
3086 (store_expr): Adjust.
3087 * cfgexpand.c (expand_call_stmt): Adjust.
3088
3089 2014-01-08 Rong Xu <xur@google.com>
3090
3091 * gcov-io.c (gcov_var): Move from gcov-io.h.
3092 (gcov_position): Ditto.
3093 (gcov_is_error): Ditto.
3094 (gcov_rewrite): Ditto.
3095 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
3096 only part to libgcc/libgcov.h.
3097
3098 2014-01-08 Marek Polacek <polacek@redhat.com>
3099
3100 PR middle-end/59669
3101 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
3102
3103 2014-01-08 Marek Polacek <polacek@redhat.com>
3104
3105 PR sanitizer/59667
3106 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
3107
3108 2014-01-08 Jakub Jelinek <jakub@redhat.com>
3109
3110 PR rtl-optimization/59649
3111 * stor-layout.c (get_mode_bounds): For BImode return
3112 0 and STORE_FLAG_VALUE.
3113
3114 2014-01-08 Richard Biener <rguenther@suse.de>
3115
3116 PR middle-end/59630
3117 * gimple.h (is_gimple_builtin_call): Remove.
3118 (gimple_builtin_call_types_compatible_p): New.
3119 (gimple_call_builtin_p): New overload.
3120 * gimple.c (is_gimple_builtin_call): Remove.
3121 (validate_call): Rename to ...
3122 (gimple_builtin_call_types_compatible_p): ... this and export. Also
3123 check return types.
3124 (validate_type): New static function.
3125 (gimple_call_builtin_p): New overload and adjust.
3126 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
3127 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
3128 (gimple_fold_stmt_to_constant_1): Likewise.
3129 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
3130
3131 2014-01-08 Richard Biener <rguenther@suse.de>
3132
3133 PR middle-end/59471
3134 * gimplify.c (gimplify_expr): Gimplify register-register type
3135 VIEW_CONVERT_EXPRs to separate stmts.
3136
3137 2014-01-07 Jeff Law <law@redhat.com>
3138
3139 PR middle-end/53623
3140 * ree.c (combine_set_extension): Handle case where source
3141 and destination registers in an extension insn are different.
3142 (combine_reaching_defs): Allow source and destination registers
3143 in extension to be different under limited circumstances.
3144 (add_removable_extension): Remove restriction that the
3145 source and destination registers in the extension are the same.
3146 (find_and_remove_re): Emit a copy from the extension's
3147 destination to its source after the defining insn if
3148 the source and destination registers are different.
3149
3150 PR middle-end/59285
3151 * ifcvt.c (merge_if_block): If we are merging a block with more than
3152 one successor with a block with no successors, remove any BARRIER
3153 after the second block.
3154
3155 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
3156
3157 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
3158
3159 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
3160
3161 PR target/59652
3162 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
3163 for 14-bit register offsets when INT14_OK_STRICT is false.
3164
3165 2014-01-07 Roland Stigge <stigge@antcom.de>
3166 Michael Meissner <meissner@linux.vnet.ibm.com>
3167
3168 PR 57386/target
3169 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
3170 Only check TFmode for SPE constants. Don't check TImode or TDmode.
3171
3172 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
3173
3174 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
3175 -mcpu.
3176
3177 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
3178
3179 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
3180 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
3181 rtx is const0_rtx or not.
3182
3183 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
3184
3185 PR target/58115
3186 * target-globals.c (save_target_globals): Remove this_fn_optab
3187 handling.
3188 * toplev.c: Include optabs.h.
3189 (target_reinit): Temporarily restore the global options if another
3190 set of options are in force.
3191
3192 2014-01-07 Jakub Jelinek <jakub@redhat.com>
3193
3194 PR rtl-optimization/58668
3195 * cfgcleanup.c (flow_find_cross_jump): Don't count
3196 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
3197 to determine what is counted.
3198 (flow_find_head_matching_sequence): Use active_insn_p to determine
3199 what is counted.
3200 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
3201 counting change.
3202 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
3203 determine what is counted.
3204
3205 PR tree-optimization/59643
3206 * tree-predcom.c (split_data_refs_to_components): If one dr is
3207 read and one write, determine_offset fails and the write isn't
3208 in the bad component, just put the read into the bad component.
3209
3210 2014-01-07 Mike Stump <mikestump@comcast.net>
3211 Jakub Jelinek <jakub@redhat.com>
3212
3213 PR pch/59436
3214 * tree-core.h (struct tree_optimization_option): Change optabs
3215 type from unsigned char * to void *.
3216 * optabs.c (init_tree_optimization_optabs): Adjust
3217 TREE_OPTIMIZATION_OPTABS initialization.
3218
3219 2014-01-06 Jakub Jelinek <jakub@redhat.com>
3220
3221 PR target/59644
3222 * config/i386/i386.h (struct machine_function): Add
3223 no_drap_save_restore field.
3224 * config/i386/i386.c (ix86_save_reg): Use
3225 !cfun->machine->no_drap_save_restore instead of
3226 crtl->stack_realign_needed.
3227 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
3228 this function clears frame_pointer_needed. Set
3229 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
3230 and DRAP reg is needed.
3231
3232 2014-01-06 Marek Polacek <polacek@redhat.com>
3233
3234 PR c/57773
3235 * doc/implement-c.texi: Mention that other integer types are
3236 permitted as bit-field types in strictly conforming mode.
3237
3238 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
3239
3240 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
3241 is newly allocated.
3242
3243 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
3244
3245 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
3246
3247 2014-01-06 Martin Jambor <mjambor@suse.cz>
3248
3249 PR ipa/59008
3250 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
3251 to int.
3252 * ipa-prop.c (ipa_print_node_params): Fix indentation.
3253
3254 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
3255
3256 PR debug/59350
3257 PR debug/59510
3258 * var-tracking.c (add_stores): Preserve the value of the source even if
3259 we don't record the store.
3260
3261 2014-01-06 Terry Guo <terry.guo@arm.com>
3262
3263 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
3264
3265 2014-01-05 Iain Sandoe <iain@codesourcery.com>
3266
3267 PR bootstrap/59541
3268 * config/darwin.c (darwin_function_section): Adjust return values to
3269 correspond to optimisation changes made in r206070.
3270
3271 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
3272
3273 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
3274 from prefetch_block tune setting.
3275 (nocona_cost): Correct size of prefetch block to 64.
3276
3277 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
3278
3279 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
3280 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
3281 used to save the static chain register in the computation of the offset
3282 from which the FP registers need to be restored.
3283
3284 2014-01-04 Jakub Jelinek <jakub@redhat.com>
3285
3286 PR tree-optimization/59519
3287 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
3288 ICE if get_current_def (current_new_name) is already non-NULL, as long
3289 as it is a phi result of some other phi in *new_exit_bb that has
3290 the same argument.
3291
3292 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
3293 or vmovdqu* for misaligned_operand.
3294 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
3295 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
3296 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
3297 aligned_mem for AVX512F masked aligned load and store builtins and for
3298 non-temporal moves.
3299
3300 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
3301
3302 PR tree-optimization/59651
3303 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
3304 Address range for negative step should be added by TYPE_SIZE_UNIT.
3305
3306 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
3307
3308 * config/m68k/m68k.c (handle_move_double): Handle pushes with
3309 overlapping registers also for registers other than the stack pointer.
3310
3311 2014-01-03 Marek Polacek <polacek@redhat.com>
3312
3313 PR other/59661
3314 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
3315 __builtin_FILE.
3316
3317 2014-01-03 Jakub Jelinek <jakub@redhat.com>
3318
3319 PR target/59625
3320 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
3321 asm goto as jump.
3322
3323 * config/i386/i386.md (MODE_SIZE): New mode attribute.
3324 (push splitter): Use <P:MODE_SIZE> instead of
3325 GET_MODE_SIZE (<P:MODE>mode).
3326 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
3327 (mov -1, reg peephole2): Likewise.
3328 * config/i386/sse.md (*mov<mode>_internal,
3329 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
3330 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
3331 *<code><mode>3, *andnot<mode>3<mask_name>,
3332 <mask_codefor><code><mode>3<mask_name>): Likewise.
3333 * config/i386/subst.md (mask_mode512bit_condition,
3334 sd_mask_mode512bit_condition): Likewise.
3335
3336 2014-01-02 Xinliang David Li <davidxl@google.com>
3337
3338 PR tree-optimization/59303
3339 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
3340 (dump_predicates): Better output format.
3341 (pred_equal_p): New function.
3342 (is_neq_relop_p): Ditto.
3343 (is_neq_zero_form_p): Ditto.
3344 (pred_expr_equal_p): Ditto.
3345 (pred_neg_p): Ditto.
3346 (simplify_pred): Ditto.
3347 (simplify_preds_2): Ditto.
3348 (simplify_preds_3): Ditto.
3349 (simplify_preds_4): Ditto.
3350 (simplify_preds): Ditto.
3351 (push_pred): Ditto.
3352 (push_to_worklist): Ditto.
3353 (get_pred_info_from_cmp): Ditto.
3354 (is_degenerated_phi): Ditto.
3355 (normalize_one_pred_1): Ditto.
3356 (normalize_one_pred): Ditto.
3357 (normalize_one_pred_chain): Ditto.
3358 (normalize_preds): Ditto.
3359 (normalize_cond_1): Remove function.
3360 (normalize_cond): Ditto.
3361 (is_gcond_subset_of): Ditto.
3362 (is_subset_of_any): Ditto.
3363 (is_or_set_subset_of): Ditto.
3364 (is_and_set_subset_of): Ditto.
3365 (is_norm_cond_subset_of): Ditto.
3366 (pred_chain_length_cmp): Ditto.
3367 (convert_control_dep_chain_into_preds): Type change.
3368 (find_predicates): Ditto.
3369 (find_def_preds): Ditto.
3370 (destroy_predicates_vecs): Ditto.
3371 (find_matching_predicates_in_rest_chains): Ditto.
3372 (use_pred_not_overlap_with_undef_path_pred): Ditto.
3373 (is_pred_expr_subset): Ditto.
3374 (is_pred_chain_subset_of): Ditto.
3375 (is_included_in): Ditto.
3376 (is_superset_of): Ditto.
3377
3378 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3379
3380 Update copyright years.
3381
3382 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
3383
3384 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
3385 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
3386 config/arc/arc.md, config/arc/arc.opt,
3387 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
3388 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
3389 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
3390 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
3391 config/linux-protos.h, config/linux.c, config/winnt-c.c,
3392 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
3393 vtable-verify.c, vtable-verify.h: Use the standard form for the
3394 copyright notice.
3395
3396 2014-01-02 Tobias Burnus <burnus@net-b.de>
3397
3398 * gcc.c (process_command): Update copyright notice dates.
3399 * gcov-dump.c: Ditto.
3400 * gcov.c: Ditto.
3401 * doc/cpp.texi: Bump @copying's copyright year.
3402 * doc/cppinternals.texi: Ditto.
3403 * doc/gcc.texi: Ditto.
3404 * doc/gccint.texi: Ditto.
3405 * doc/gcov.texi: Ditto.
3406 * doc/install.texi: Ditto.
3407 * doc/invoke.texi: Ditto.
3408
3409 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3410
3411 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
3412
3413 2014-01-01 Jakub Jelinek <jakub@redhat.com>
3414
3415 * config/i386/sse.md (*mov<mode>_internal): Guard
3416 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
3417
3418 PR rtl-optimization/59647
3419 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
3420 new_rtx into UNSIGNED_FLOAT rtxes.
3421 \f
3422 Copyright (C) 2014 Free Software Foundation, Inc.
3423
3424 Copying and distribution of this file, with or without modification,
3425 are permitted in any medium without royalty provided the copyright
3426 notice and this notice are preserved.
This page took 0.185803 seconds and 6 git commands to generate.