1 2014-01-30 Alangi Derick <alangiderick@gmail.com>
3 * README.Portability: Fix typo.
5 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
7 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
8 comparison_operator with ordered_comparison_operator.
10 2014-01-30 Nick Clifton <nickc@redhat.com>
12 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
13 Rename to mn10300_store_multiple_regs.
14 * config/mn10300/mn10300.c: Likewise.
15 * config/mn10300/mn10300.md (store_movm): Fix typo: call
17 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
18 Call mn10300_store_multiple_regs.
20 2014-01-30 Nick Clifton <nickc@redhat.com>
21 DJ Delorie <dj@redhat.com>
23 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
24 %fp 2 to keep registers after it properly word-aligned.
25 (rl78_alloc_physical_registers_umul): Handle the case where both
26 input operands are the same.
28 2014-01-30 Richard Biener <rguenther@suse.de>
30 PR tree-optimization/59903
31 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
34 2014-01-30 Jason Merrill <jason@redhat.com>
37 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
40 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
42 2014-01-30 Richard Biener <rguenther@suse.de>
44 PR tree-optimization/59951
45 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug
48 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
51 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
52 SFmode to DFmode case.
54 2014-01-29 DJ Delorie <dj@redhat.com>
56 * config/msp430/msp430.opt (-minrt): New.
57 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
59 (ENDFILE_SPEC): Likewise.
61 2014-01-29 Jan Hubicka <jh@suse.cz>
63 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
64 (estimate_function_body_sizes): Use it.
66 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
69 * dwarf2out.c (is_cxx_auto): New.
70 (is_base_type): Use it.
71 (gen_type_die_with_usage): Likewise.
73 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
75 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
76 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
77 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
78 -maltivec=be with LE targets.
79 (vsx_mergeh_<mode>): Likewise.
80 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New
82 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
83 (altivec_vmrghb): Replace with define_expand and new
84 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE
86 (altivec_vmrghb_direct): New define_insn.
87 (altivec_vmrghh): Replace with define_expand and new
88 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE
90 (altivec_vmrghh_direct): New define_insn.
91 (altivec_vmrghw): Replace with define_expand and new
92 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE
94 (altivec_vmrghw_direct): New define_insn.
95 (*altivec_vmrghsf): Adjust for endianness.
96 (altivec_vmrglb): Replace with define_expand and new
97 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE
99 (altivec_vmrglb_direct): New define_insn.
100 (altivec_vmrglh): Replace with define_expand and new
101 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE
103 (altivec_vmrglh_direct): New define_insn.
104 (altivec_vmrglw): Replace with define_expand and new
105 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE
107 (altivec_vmrglw_direct): New define_insn.
108 (*altivec_vmrglsf): Adjust for endianness.
109 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
110 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
111 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
112 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
113 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
114 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
115 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
116 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
118 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
120 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
121 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
124 2014-01-29 Richard Biener <rguenther@suse.de>
126 PR tree-optimization/58742
127 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
128 associate_pointerplus_align.
129 (associate_pointerplus_diff): New function.
130 (associate_pointerplus): Likewise. Call associate_pointerplus_align
131 and associate_pointerplus_diff.
133 2014-01-29 Richard Biener <rguenther@suse.de>
135 * lto-streamer.h (LTO_major_version): Bump to 3.
136 (LTO_minor_version): Reset to 0.
138 2014-01-29 Renlin Li <Renlin.Li@arm.com>
140 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
141 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
142 (arm_file_start): Generate correct asm header for armv7ve.
143 * config/arm/bpabi.h: Add multilib support for armv7ve.
144 * config/arm/driver-arm.c: Change the architectures of cortex-a7
145 and cortex-a15 to armv7ve.
146 * config/arm/t-aprofile: Add multilib support for armv7ve.
147 * doc/invoke.texi: Document -march=armv7ve.
149 2014-01-29 Richard Biener <rguenther@suse.de>
151 PR tree-optimization/58742
152 * tree-ssa-forwprop.c (associate_plusminus): Return true
153 if we changed sth, defer EH cleanup to ...
154 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
155 (simplify_mult): New function.
157 2014-01-29 Jakub Jelinek <jakub@redhat.com>
160 PR tree-optimization/59920
161 * tree.c (build_common_builtin_nodes): Remove
162 __builtin_setjmp_dispatcher initialization.
163 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
164 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
165 instead of gsi_after_labels + manually skipping debug stmts.
166 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
167 ignore bbs with IFN_ABNORMAL_DISPATCHER.
168 * tree-inline.c (copy_edges_for_bb): Remove
169 can_make_abnormal_goto argument, instead add abnormal_goto_dest
170 argument. Ignore computed_goto_p stmts. Don't call
171 make_abnormal_goto_edges. If a call might need abnormal edges
172 for non-local gotos, see if it already has an edge to
173 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
174 with true argument, don't do anything then, otherwise add
175 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
176 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
178 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
179 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
180 (lower_stmt): Don't set data->calls_builtin_setjmp.
181 (lower_builtin_setjmp): Adjust comment.
182 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
183 * tree-cfg.c (found_computed_goto): Remove.
184 (factor_computed_gotos): Remove.
185 (make_goto_expr_edges): Return bool, true for computed gotos.
186 Don't call make_abnormal_goto_edges.
187 (build_gimple_cfg): Don't set found_computed_goto, don't call
188 factor_computed_gotos.
189 (computed_goto_p): No longer static.
190 (make_blocks): Don't set found_computed_goto.
191 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
192 (make_edges): If make_goto_expr_edges returns true, push bb
193 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
194 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
195 vector. Record mapping between bbs and OpenMP regions if there
196 are any, adjust make_gimple_omp_edges caller. Call
197 handle_abnormal_edges.
198 (make_abnormal_goto_edges): Remove.
199 * tree-cfg.h (make_abnormal_goto_edges): Remove.
200 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
201 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
202 * builtins.c (expand_builtin): Don't handle
203 BUILT_IN_SETJMP_DISPATCHER.
204 * internal-fn.def (ABNORMAL_DISPATCHER): New.
205 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
206 filling *region also set *region_idx to (*region)->entry->index.
209 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
210 For REGs set ORIGINAL_REGNO.
212 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
214 * doc/md.texi: Mention that a target shouldn't implement
215 vec_widen_(s|u)mul_even/odd pair if it is less efficient
218 2014-01-29 Jakub Jelinek <jakub@redhat.com>
220 PR tree-optimization/59594
221 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
222 a copy of the datarefs vector rather than the vector itself.
224 2014-01-28 Jason Merrill <jason@redhat.com>
227 * dwarf2out.c (auto_die): New static.
228 (gen_type_die_with_usage): Handle C++1y 'auto'.
229 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
232 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
235 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
236 (SPEC_X32): Likewise.
238 * config/i386/i386.c (ix86_option_override_internal): Turn off
239 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
241 (x86_file_start): Output .code16gcc for TARGET_16BIT.
242 * config/i386/i386.h (TARGET_16BIT): New macro.
243 (TARGET_16BIT_P): Likewise.
244 * config/i386/i386.opt: Add m16.
245 * doc/invoke.texi: Document -m16.
247 2014-01-28 Jakub Jelinek <jakub@redhat.com>
249 PR preprocessor/59935
250 * input.c (location_get_source_line): Bail out on when line number
251 is zero, and test the return value of
252 lookup_or_add_file_to_cache_tab.
254 2014-01-28 Richard Biener <rguenther@suse.de>
256 PR tree-optimization/58742
257 * tree-ssa-forwprop.c (associate_plusminus): Handle
258 pointer subtraction of the form (T)(P + A) - (T)P.
260 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
262 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
265 2014-01-28 Richard Biener <rguenther@suse.de>
268 2014-01-28 Richard Biener <rguenther@suse.de>
270 PR rtl-optimization/45364
271 PR rtl-optimization/59890
272 * var-tracking.c (local_get_addr_clear_given_value): Handle
273 already cleared slot.
274 (val_reset): Handle not allocated local_get_addr_cache.
275 (vt_find_locations): Use post-order on the inverted CFG.
277 2014-01-28 Richard Biener <rguenther@suse.de>
279 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps):
282 2014-01-28 Richard Biener <rguenther@suse.de>
284 PR rtl-optimization/45364
285 PR rtl-optimization/59890
286 * var-tracking.c (local_get_addr_clear_given_value): Handle
287 already cleared slot.
288 (val_reset): Handle not allocated local_get_addr_cache.
289 (vt_find_locations): Use post-order on the inverted CFG.
291 2014-01-28 Alan Modra <amodra@gmail.com>
293 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
294 * configure.ac <recursive call for build != host>: Define
295 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
296 and LD_FOR_BUILD too.
297 * configure: Regenerate.
299 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
301 * config/i386/i386.c (get_builtin_code_for_version): Separate
302 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
303 Broadwell from Haswell.
305 2014-01-27 Steve Ellcey <sellcey@mips.com>
307 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
308 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
309 * config/mips/mips.c (mips_option_override): Change setting
311 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
312 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
313 Change from Mask to Var.
315 2014-01-27 Jeff Law <law@redhat.com>
317 * ipa-inline.c (inline_small_functions): Fix typo.
319 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
321 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
322 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
323 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
324 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
325 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
326 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
327 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
328 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
329 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
330 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
331 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
332 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
333 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
334 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
335 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
336 (_mm512_storeu_epi64): Ditto.
337 (_mm512_cmpge_epi32_mask): Ditto.
338 (_mm512_cmpge_epu32_mask): Ditto.
339 (_mm512_cmpge_epi64_mask): Ditto.
340 (_mm512_cmpge_epu64_mask): Ditto.
341 (_mm512_cmple_epi32_mask): Ditto.
342 (_mm512_cmple_epu32_mask): Ditto.
343 (_mm512_cmple_epi64_mask): Ditto.
344 (_mm512_cmple_epu64_mask): Ditto.
345 (_mm512_cmplt_epi32_mask): Ditto.
346 (_mm512_cmplt_epu32_mask): Ditto.
347 (_mm512_cmplt_epi64_mask): Ditto.
348 (_mm512_cmplt_epu64_mask): Ditto.
349 (_mm512_cmpneq_epi32_mask): Ditto.
350 (_mm512_cmpneq_epu32_mask): Ditto.
351 (_mm512_cmpneq_epi64_mask): Ditto.
352 (_mm512_cmpneq_epu64_mask): Ditto.
353 (_mm512_expand_pd): Ditto.
354 (_mm512_expand_ps): Ditto.
355 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
356 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
357 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
358 * config/i386/i386.c (ix86_builtins): Add
359 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
360 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
361 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
362 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
363 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
364 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
365 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
366 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
367 IX86_BUILTIN_PMOVUSQW512_MEM.
368 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
369 __builtin_ia32_pmovsqd512mem_mask,
370 __builtin_ia32_pmovqd512mem_mask,
371 __builtin_ia32_pmovusqw512mem_mask,
372 __builtin_ia32_pmovsqw512mem_mask,
373 __builtin_ia32_pmovqw512mem_mask,
374 __builtin_ia32_pmovusdw512mem_mask,
375 __builtin_ia32_pmovsdw512mem_mask,
376 __builtin_ia32_pmovdw512mem_mask,
377 __builtin_ia32_pmovqb512mem_mask,
378 __builtin_ia32_pmovusqb512mem_mask,
379 __builtin_ia32_pmovsqb512mem_mask,
380 __builtin_ia32_pmovusdb512mem_mask,
381 __builtin_ia32_pmovsdb512mem_mask,
382 __builtin_ia32_pmovdb512mem_mask.
383 (bdesc_args): Add __builtin_ia32_expanddf512,
384 __builtin_ia32_expandsf512.
385 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
386 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
387 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
388 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
389 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
390 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
391 (avx512f_<code>v8div16qi2_mask_store): This.
392 (avx512f_expand<mode>): New.
394 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
396 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
398 (_mm512_mask_prefetch_i64gather_pd): Ditto.
399 (_mm512_prefetch_i32scatter_pd): Ditto.
400 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
401 (_mm512_prefetch_i64scatter_pd): Ditto.
402 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
403 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
404 (_mm512_mask_prefetch_i64gather_ps): Ditto.
405 (_mm512_prefetch_i32scatter_ps): Ditto.
406 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
407 (_mm512_prefetch_i64scatter_ps): Ditto.
408 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
409 * config/i386/i386-builtin-types.def: Define
410 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
411 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
412 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
413 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
414 IX86_BUILTIN_SCATTERPFQPD.
415 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
416 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
417 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
418 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
419 __builtin_ia32_scatterpfqps.
420 (ix86_expand_builtin): Expand new built-ins.
421 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
422 fix memory access data type.
423 (*avx512pf_gatherpf<mode>_mask): Ditto.
424 (*avx512pf_gatherpf<mode>): Ditto.
425 (avx512pf_scatterpf<mode>): Ditto.
426 (*avx512pf_scatterpf<mode>_mask): Ditto.
427 (*avx512pf_scatterpf<mode>): Ditto.
428 (GATHER_SCATTER_SF_MEM_MODE): New.
429 (avx512pf_gatherpf<mode>df): Ditto.
430 (*avx512pf_gatherpf<mode>df_mask): Ditto.
431 (*avx512pf_scatterpf<mode>df): Ditto.
433 2014-01-27 Jakub Jelinek <jakub@redhat.com>
436 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
437 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
440 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
442 * common/config/arm/arm-common.c
443 (arm_rewrite_mcpu): Handle multiple names.
445 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
447 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
449 * gimple-builder.h (create_gimple_tmp): Delete.
451 2014-01-27 Christian Bruel <christian.bruel@st.com>
453 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
456 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
458 * config/pa/pa.md (call): Generate indirect long calls to non-local
459 functions when outputing 32-bit code.
460 (call_value): Likewise except for special call to buggy powf function.
462 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
463 portable runtime and PIC indirect calls.
464 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
465 and PIC call sequences. Use ldo instead of blr to set return register
466 in PIC call sequence.
468 2014-01-25 Walter Lee <walt@tilera.com>
470 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
471 avoid clobbering a live register.
473 2014-01-25 Walter Lee <walt@tilera.com>
475 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
476 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
477 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
478 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
480 2014-01-25 Walter Lee <walt@tilera.com>
482 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
483 arguments on even registers.
484 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
486 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
487 (BIGGEST_ALIGNMENT): Ditto.
488 (BIGGEST_FIELD_ALIGNMENT): Ditto.
490 2014-01-25 Walter Lee <walt@tilera.com>
492 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
493 insns before bundling.
494 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
496 2014-01-25 Walter Lee <walt@tilera.com>
498 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
499 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
500 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
502 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
504 * config/mips/constraints.md (kl): Delete.
505 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
506 define expands, using...
507 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
508 instructions for MIPS16.
509 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
510 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
512 2014-01-25 Walter Lee <walt@tilera.com>
514 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
518 2014-01-25 Walter Lee <walt@tilera.com>
520 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
521 (TARGET_EXPAND_TO_RTL_HOOK): Define.
523 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
525 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
528 2014-01-25 Jakub Jelinek <jakub@redhat.com>
530 * print-rtl.c (in_call_function_usage): New var.
531 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
532 EXPR_LIST mode as mode and not as reg note name.
535 * cfgloopmanip.c (copy_loop_info): If
536 loop->warned_aggressive_loop_optimizations, make sure
537 the flag is set in target loop too.
539 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
541 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
543 * builtins.def: Likewise.
544 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
545 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
546 * ira.c (ira_setup_eliminable_regset): Likewise.
547 * omp-low.c (gate_expand_omp): Likewise.
548 (execute_lower_omp): Likewise.
549 (diagnose_sb_0): Likewise.
550 (gate_diagnose_omp_blocks): Likewise.
551 (simd_clone_clauses_extract): Likewise.
554 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
556 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
557 correction for little endian...
558 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
561 2014-01-24 Jeff Law <law@redhat.com>
563 PR tree-optimization/59919
564 * tree-vrp.c (find_assert_locations_1): Do not register asserts
565 for non-returning calls.
567 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
569 * common/config/aarch64/aarch64-common.c
570 (aarch64_rewrite_mcpu): Handle multiple names.
571 * config/aarch64/aarch64.h
572 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
574 2014-01-24 Dodji Seketeli <dodji@redhat.com>
576 * input.c (add_file_to_cache_tab): Handle the case where fopen
579 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
582 * config/i386/i386.md (pushsf splitter): Get stack adjustment
583 from push operand if code of push isn't PRE_DEC.
585 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
588 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
589 -mquad-memory-atomic. Update -mquad-memory documentation to say
590 it is only used for non-atomic loads/stores.
592 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
593 -mquad-memory or -mquad-memory-atomic switches.
595 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
596 -mquad-memory-atomic to ISA 2.07 support.
598 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
599 to separate support of normal quad word memory operations (ldq, stq)
600 from the atomic quad word memory operations.
602 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
603 support to separate non-atomic quad word operations from atomic
604 quad word operations. Disable non-atomic quad word operations in
605 little endian mode so that we don't have to swap words after the
606 load and before the store.
607 (quad_load_store_p): Add comment about atomic quad word support.
608 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
609 options printed with -mdebug=reg.
611 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
612 -mquad-memory-atomic as the test for whether we have quad word
614 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
615 or -mp8-vector are used, allow byte/half-word atomic operations.
617 * config/rs6000/sync.md (load_lockedti): Insure that the address
618 is a proper indexed or indirect address for the lqarx instruction.
619 On little endian systems, swap the hi/lo registers after the lqarx
621 (load_lockedpti): Use indexed_or_indirect_operand predicate to
622 insure the address is valid for the lqarx instruction.
623 (store_conditionalti): Insure that the address is a proper indexed
624 or indirect address for the stqcrx. instruction. On little endian
625 systems, swap the hi/lo registers before doing the stqcrx.
627 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
628 insure the address is valid for the stqcrx. instruction.
630 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
631 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
632 type of quad memory support is available.
634 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
637 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
638 there is a danger of looping.
640 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
642 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
643 force flag_ira_loop_pressure if set via command line.
645 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
647 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
648 (ashr_simd): New builtin handling DI mode.
649 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
650 (aarch64_sshr_simddi): New match pattern.
651 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
652 (vshrd_n_s64): Likewise.
653 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
655 2014-01-23 Nick Clifton <nickc@redhat.com>
657 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
658 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
659 favour of mcu specific scripts.
660 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
663 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
664 Alex Velenko <Alex.Velenko@arm.com>
666 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
667 (vaddv_s16): Likewise.
668 (vaddv_s32): Likewise.
669 (vaddv_u8): Likewise.
670 (vaddv_u16): Likewise.
671 (vaddv_u32): Likewise.
672 (vaddvq_s8): Likewise.
673 (vaddvq_s16): Likewise.
674 (vaddvq_s32): Likewise.
675 (vaddvq_s64): Likewise.
676 (vaddvq_u8): Likewise.
677 (vaddvq_u16): Likewise.
678 (vaddvq_u32): Likewise.
679 (vaddvq_u64): Likewise.
680 (vaddv_f32): Likewise.
681 (vaddvq_f32): Likewise.
682 (vaddvq_f64): Likewise.
683 (vmaxv_f32): Likewise.
684 (vmaxv_s8): Likewise.
685 (vmaxv_s16): Likewise.
686 (vmaxv_s32): Likewise.
687 (vmaxv_u8): Likewise.
688 (vmaxv_u16): Likewise.
689 (vmaxv_u32): Likewise.
690 (vmaxvq_f32): Likewise.
691 (vmaxvq_f64): Likewise.
692 (vmaxvq_s8): Likewise.
693 (vmaxvq_s16): Likewise.
694 (vmaxvq_s32): Likewise.
695 (vmaxvq_u8): Likewise.
696 (vmaxvq_u16): Likewise.
697 (vmaxvq_u32): Likewise.
698 (vmaxnmv_f32): Likewise.
699 (vmaxnmvq_f32): Likewise.
700 (vmaxnmvq_f64): Likewise.
701 (vminv_f32): Likewise.
702 (vminv_s8): Likewise.
703 (vminv_s16): Likewise.
704 (vminv_s32): Likewise.
705 (vminv_u8): Likewise.
706 (vminv_u16): Likewise.
707 (vminv_u32): Likewise.
708 (vminvq_f32): Likewise.
709 (vminvq_f64): Likewise.
710 (vminvq_s8): Likewise.
711 (vminvq_s16): Likewise.
712 (vminvq_s32): Likewise.
713 (vminvq_u8): Likewise.
714 (vminvq_u16): Likewise.
715 (vminvq_u32): Likewise.
716 (vminnmv_f32): Likewise.
717 (vminnmvq_f32): Likewise.
718 (vminnmvq_f64): Likewise.
720 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
722 * config/aarch64/aarch64-simd.md
723 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
724 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
725 (*aarch64_mul3_elt<mode>): Likewise.
726 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
727 (*aarch64_mul3_elt_to_64v2df): Likewise.
728 (*aarch64_mla_elt<mode>): Likewise.
729 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
730 (*aarch64_mls_elt<mode>): Likewise.
731 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
732 (*aarch64_fma4_elt<mode>): Likewise.
733 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
734 (*aarch64_fma4_elt_to_64v2df): Likewise.
735 (*aarch64_fnma4_elt<mode>): Likewise.
736 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
737 (*aarch64_fnma4_elt_to_64v2df): Likewise.
738 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
739 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
740 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
741 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
742 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
743 (aarch64_sqdmull_lane<mode>_internal): Likewise.
744 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
746 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
748 * config/aarch64/aarch64-simd.md
749 (aarch64_be_checked_get_lane<mode>): New define_expand.
750 * config/aarch64/aarch64-simd-builtins.def
751 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
752 New builtin definition.
753 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
754 Use new safe be builtin.
756 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
758 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
760 (aarch64_be_st1<mode>): Likewise.
761 (aarch_ld1<VALL:mode>): Define_expand modified.
762 (aarch_st1<VALL:mode>): Likewise.
763 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
764 (UNSPEC_ST1): Likewise.
766 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
768 * config/microblaze/microblaze.md: Add trap insn and attribute
770 2014-01-23 Dodji Seketeli <dodji@redhat.com>
772 PR preprocessor/58580
773 * input.h (location_get_source_line): Take an additional line_size
775 (void diagnostics_file_cache_fini): Declare new function.
776 * input.c (struct fcache): New type.
777 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
778 New static constants.
779 (diagnostic_file_cache_init, total_lines_num)
780 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
781 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
782 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
783 (get_next_line, read_next_line, goto_next_line, read_line_num):
784 New static function definitions.
785 (diagnostic_file_cache_fini): New function.
786 (location_get_source_line): Take an additional output line_len
787 parameter. Re-write using lookup_or_add_file_to_cache_tab and
789 * diagnostic.c (diagnostic_finish): Call
790 diagnostic_file_cache_fini.
791 (adjust_line): Take an additional input parameter for the length
792 of the line, rather than calculating it with strlen.
793 (diagnostic_show_locus): Adjust the use of
794 location_get_source_line and adjust_line with respect to their new
795 signature. While displaying a line now, do not stop at the first
796 null byte. Rather, display the zero byte as a space and keep
797 going until we reach the size of the line.
798 * Makefile.in: Add vec.o to OBJS-libcommon
800 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
801 Ilya Tocar <ilya.tocar@intel.com>
803 * config/i386/avx512fintrin.h (_mm512_kmov): New.
804 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
805 (__builtin_ia32_kmov16): Ditto.
806 * config/i386/i386.md (UNSPEC_KMOV): New.
809 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
811 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
812 (_mm512_storeu_si512): Ditto.
814 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
817 * rtl.h (get_referenced_operands): Declare.
818 * recog.c (get_referenced_operands): New function.
819 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
820 operands have been referenced when recording LO_SUM references.
822 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
824 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
826 2014-01-22 Jan Hubicka <jh@suse.cz>
828 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
829 Enable for generic and recent AMD targets.
831 2014-01-22 Jan Hubicka <jh@suse.cz>
833 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
834 ARG_SIZE note when adjustment was eliminated.
836 2014-01-22 Jeff Law <law@redhat.com>
838 PR tree-optimization/59597
839 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
840 in file. Accept new argument REGISTERING and use it to modify
841 dump output appropriately.
842 (register_jump_thread): Corresponding changes.
843 (mark_threaded_blocks): Reinstate code to cancel unprofitable
844 thread paths involving joiner blocks. Add code to dump cancelled
845 jump threading paths.
847 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
849 PR rtl-optimization/59477
850 * lra-constraints.c (inherit_in_ebb): Process call for living hard
851 regs. Update reloads_num and potential_reload_hard_regs for all insns.
853 2014-01-22 Tom Tromey <tromey@redhat.com>
855 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
857 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
859 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
861 PR rtl-optimization/59896
862 * lra-constraints.c (process_alt_operands): Check unused note for
863 matched operands of insn with no output reloads.
865 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
867 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
868 (mips_move_from_gpr_cost): Likewise.
870 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
872 PR rtl-optimization/59858
873 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
874 ira_class_hard_regs_num.
875 (process_alt_operands): Increase reject for dying matched operand.
877 2014-01-21 Jakub Jelinek <jakub@redhat.com>
880 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
881 smaller than size, perform several stores or loads and stores
882 at dst + count - size to store or copy all of size bytes, rather
883 than just last modesize bytes.
885 2014-01-20 DJ Delorie <dj@redhat.com>
887 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
888 that CLOBBERs are REGs before propogating their values.
890 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
893 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
894 (cgraph_inline_failed_type): New function.
895 * cgraph.h (DEFCIFCODE): Add type.
896 (cgraph_inline_failed_type_t): New enum.
897 (cgraph_inline_failed_type): New prototype.
898 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
899 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
900 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
901 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
902 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
903 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
904 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
905 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
906 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
907 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
908 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
909 OPTIMIZATION_MISMATCH.
910 * tree-inline.c (expand_call_inline): Emit errors during
911 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
913 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
916 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
917 mode attribute in insn output.
919 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
921 * output.h (output_constant): Delete.
922 * varasm.c (output_constant): Make private.
924 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
926 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
928 2014-01-20 Jakub Jelinek <jakub@redhat.com>
931 * tree.h (fold_builtin_strcat): New prototype.
932 * builtins.c (fold_builtin_strcat): No longer static. Add len
933 argument, if non-NULL, don't call c_strlen. Optimize
934 directly into __builtin_memcpy instead of __builtin_strcpy.
935 (fold_builtin_2): Adjust fold_builtin_strcat caller.
936 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
938 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
940 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
941 for SImode_address_operand operands, having only a REG argument.
943 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
945 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
946 loader name using mbig-endian.
947 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
949 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
951 * doc/invoke.texi (-march): Clarify documentation for AArch64.
955 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
957 * config/aarch64/aarch64-protos.h
958 (aarch64_cannot_change_mode_class_ptr): Declare.
959 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
960 aarch64_cannot_change_mode_class_ptr): New.
961 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
962 backend hook aarch64_cannot_change_mode_class.
964 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
966 * common/config/aarch64/aarch64-common.c
967 (aarch64_handle_option): Don't handle any option order logic here.
968 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
969 selected_cpu, warn on architecture version mismatch.
970 (aarch64_override_options): Fix parsing order for option strings.
972 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
973 Iain Sandoe <iain@codesourcery.com>
976 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
977 warning. Amend comment to reflect current functionality.
979 2014-01-20 Richard Biener <rguenther@suse.de>
982 * builtins.c (fold_builtin_strcat): Remove case better handled
983 by tree-ssa-strlen.c.
985 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
987 * config/aarch64/aarch64.opt
988 (mcpu, march, mtune): Make case-insensitive.
990 2014-01-20 Jakub Jelinek <jakub@redhat.com>
993 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
994 if operands[1] is a REG or ZERO_EXTEND of a REG.
996 2014-01-19 Jan Hubicka <jh@suse.cz>
998 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
1000 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
1002 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
1003 long non-pic millicode calls.
1005 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1007 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
1009 2014-01-19 Kito Cheng <kito@0xlab.org>
1011 * builtins.c (expand_movstr): Check movstr expand done or fail.
1013 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
1014 H.J. Lu <hongjiu.lu@intel.com>
1017 * config/i386/i386.md (*lea<mode>): Zero-extend return register
1018 to DImode for zero-extended addresses.
1020 2014-01-19 Jakub Jelinek <jakub@redhat.com>
1022 PR rtl-optimization/57763
1023 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
1024 on the new indirect jump_insn and increment LABEL_NUSES (label).
1026 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
1030 * config.gcc (x86_archs): New variable.
1031 (x86_64_archs): Likewise.
1032 (x86_cpus): Likewise.
1033 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
1034 --with-arch/--with-cpu= options.
1035 Support --with-arch=/--with-cpu={nehalem,westmere,
1036 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
1038 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
1040 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
1041 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
1043 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
1045 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
1047 2014-01-18 Jakub Jelinek <jakub@redhat.com>
1050 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
1051 clear cpp_get_options (parse_in)->warn_unused_macros for
1052 ix86_target_macros_internal with cpp_define.
1054 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
1056 * jump.c (delete_related_insns): Keep (use (insn))s.
1057 * reorg.c (redundant_insn): Check for barriers too.
1059 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1061 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
1063 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
1065 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
1066 call to $$dyncall when TARGET_LONG_CALLS is true.
1068 2014-01-17 Jeff Law <law@redhat.com>
1070 * ree.c (combine_set_extension): Temporarily disable test for
1071 changing number of hard registers.
1073 2014-01-17 Jan Hubicka <jh@suse.cz>
1076 * ipa-inline-analysis.c (inline_free_summary):
1077 Do not free summary of aliases.
1079 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1082 * gimplify.c (gimplify_expr): Use create_tmp_var
1083 instead of create_tmp_var_raw. If cond doesn't have
1084 integral type, don't add the IFN_ANNOTATE builtin at all.
1086 2014-01-17 Martin Jambor <mjambor@suse.cz>
1089 * ipa-cp.c (prev_edge_clone): New variable.
1090 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
1091 Also resize prev_edge_clone vector.
1092 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
1093 (ipcp_edge_removal_hook): New function.
1094 (ipcp_driver): Register ipcp_edge_removal_hook.
1096 2014-01-17 Andrew Pinski <apinski@cavium.com>
1097 Steve Ellcey <sellcey@mips.com>
1100 * config/mips/mips.c (mips_print_operand): Check operand mode instead
1103 2014-01-17 Jeff Law <law@redhat.com>
1106 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
1107 so that pass_ccp runs first.
1109 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1111 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
1112 (ix86_adjust_cost): Use !TARGET_XXX.
1113 (do_reorder_for_imul): Likewise.
1114 (swap_top_of_ready_list): Likewise.
1115 (ix86_sched_reorder): Likewise.
1117 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1119 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1120 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
1121 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
1122 (intel_memset): New. Duplicate slm_memset.
1123 (intel_cost): New. Duplicate slm_cost.
1124 (m_INTEL): New macro.
1125 (processor_target_table): Add "intel".
1126 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
1127 with PROCESSOR_INTEL for "intel".
1128 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
1129 PROCESSOR_SILVERMONT.
1130 (ix86_issue_rate): Likewise.
1131 (ix86_adjust_cost): Likewise.
1132 (ia32_multipass_dfa_lookahead): Likewise.
1133 (swap_top_of_ready_list): Likewise.
1134 (ix86_sched_reorder): Likewise.
1135 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
1136 instead of TARGET_OPT_AGU.
1137 * config/i386/i386.h (TARGET_INTEL): New.
1138 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
1139 (processor_type): Add PROCESSOR_INTEL.
1140 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
1141 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
1143 2014-01-17 Marek Polacek <polacek@redhat.com>
1146 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
1149 2014-01-17 Richard Biener <rguenther@suse.de>
1151 PR tree-optimization/46590
1152 * opts.c (default_options_table): Add entries for
1153 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
1154 all enabled at -O1 but not for -Og.
1155 * common.opt (fbranch-count-reg): Remove Init(1).
1156 (fmove-loop-invariants): Likewise.
1157 (ftree-pta): Likewise.
1159 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1161 * config/i386/i386.c (ix86_data_alignment): For compatibility with
1162 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
1163 decls to at least the GCC 4.8 used alignments.
1166 * tree-nested.c (convert_nonlocal_reference_stmt,
1167 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
1168 of GIMPLE_BIND stmts, adjust associated decls.
1170 2014-01-17 Richard Biener <rguenther@suse.de>
1172 PR tree-optimization/46590
1173 * vec.h (vec<>::bseach): New member function implementing
1174 binary search according to C89 bsearch.
1175 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
1176 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
1177 bitmap pointer again. Make accesses_in_loop a flat array.
1178 (mem_ref_obstack): New global.
1179 (outermost_indep_loop): Adjust for mem_ref->stored changes.
1180 (mark_ref_stored): Likewise.
1181 (ref_indep_loop_p_2): Likewise.
1182 (set_ref_stored_in_loop): New helper function.
1183 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
1184 (memref_free): Adjust.
1185 (record_mem_ref_loc): Simplify.
1186 (gather_mem_refs_stmt): Adjust.
1187 (sort_locs_in_loop_postorder_cmp): New function.
1188 (analyze_memory_references): Sort accesses_in_loop after
1189 loop postorder number.
1190 (find_ref_loc_in_loop_cmp): New function.
1191 (for_all_locs_in_loop): Find relevant cluster of locs in
1192 accesses_in_loop and iterate without recursion.
1193 (execute_sm): Avoid uninit warning.
1194 (struct ref_always_accessed): Simplify.
1195 (ref_always_accessed::operator ()): Likewise.
1196 (ref_always_accessed_p): Likewise.
1197 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
1198 loop postorder numbers here.
1199 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
1202 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
1205 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
1206 on decls for which assemble_alias has been called.
1208 2014-01-17 Nick Clifton <nickc@redhat.com>
1210 * config/msp430/msp430.opt: (mcpu): New option.
1211 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
1212 (msp430_option_override): Parse target_cpu. If the MCU name
1213 matches a generic string, clear target_mcu.
1214 (msp430_attr): Allow numeric interrupt values up to 63.
1215 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
1216 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
1218 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
1220 * config/msp430/msp430.md (popm): Use %J rather than %I.
1221 (addsi3): Use msp430_nonimmediate_operand for operand 2.
1222 (addhi_cy_i): Use immediate_operand for operand 2.
1223 * doc/invoke.texi: Document -mcpu option.
1225 2014-01-17 Richard Biener <rguenther@suse.de>
1227 PR rtl-optimization/38518
1228 * df.h (df_analyze_loop): Declare.
1229 * df-core.c: Include cfgloop.h.
1230 (df_analyze_1): Split out main part of df_analyze.
1231 (df_analyze): Adjust.
1232 (loop_inverted_post_order_compute): New function.
1233 (loop_post_order_compute): Likewise.
1234 (df_analyze_loop): New function avoiding whole-function
1236 * loop-invariant.c (find_defs): Use df_analyze_loop.
1237 (find_invariants): Adjust.
1238 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
1240 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
1242 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
1243 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
1245 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
1247 * ipa-ref.c (ipa_remove_stmt_references): Fix references
1248 traversal when removing references.
1250 2014-01-16 Jan Hubicka <jh@suse.cz>
1253 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
1255 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
1258 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
1259 pushing a reload for an autoinc when we had previously reloaded an
1260 inner part of the address.
1262 2014-01-16 Jakub Jelinek <jakub@redhat.com>
1264 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
1266 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
1267 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
1268 when not giving up or versioning for alias only because of
1270 (vect_analyze_data_ref_dependences): Set to true.
1271 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
1273 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
1274 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
1278 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
1281 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
1282 operand 0 predicate for gathers, use a new pseudo as subtarget.
1284 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1287 * lra-constraints.c (process_alt_operands): Add printing debug info.
1288 Check absence of input/output reloads for matched operands too.
1290 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1292 PR rtl-optimization/59835
1293 * ira.c (ira_init_register_move_cost): Increase cost for
1296 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
1298 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
1300 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
1303 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
1304 non-register objects. Use gen_(high/low)part more consistently.
1307 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1310 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
1311 endian support, remove tests for WORDS_BIG_ENDIAN.
1312 (p8_mfvsrd_3_<mode>): Likewise.
1313 (reload_gpr_from_vsx<mode>): Likewise.
1314 (reload_gpr_from_vsxsf): Likewise.
1315 (p8_mfvsrd_4_disf): Likewise.
1317 2014-01-16 Richard Biener <rguenther@suse.de>
1319 PR rtl-optimization/46590
1320 * lcm.c (compute_antinout_edge): Use postorder iteration.
1321 (compute_laterin): Use inverted postorder iteration.
1323 2014-01-16 Nick Clifton <nickc@redhat.com>
1326 * varasm.c (output_constant): Return the number of bytes actually
1328 (output_constructor_array_range): Update the field size with the
1329 number of bytes emitted by output_constant.
1330 (output_constructor_regular_field): Likewise. Also do not
1331 complain if the total number of bytes emitted is now greater
1332 than the expected fieldpos.
1333 * output.h (output_constant): Update prototype and descriptive comment.
1335 2014-01-16 Marek Polacek <polacek@redhat.com>
1338 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
1339 it is error_mark_node.
1341 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
1343 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
1344 VALID_AVX256_REG_OR_OI_MODE.
1346 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
1348 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
1349 current procedure should be profiled.
1351 2014-01-15 Andrew Pinski <apinski@cavium.com>
1353 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
1354 of moving from/to the STACK_REG register class.
1356 2014-01-15 Richard Henderson <rth@redhat.com>
1359 * reginfo.c (global_regs_decl): Globalize.
1360 * rtl.h (global_regs_decl): Declare.
1361 * ira.c (do_reload): Diagnose frame_pointer_needed and it
1362 reserved via global_regs.
1364 2014-01-15 Teresa Johnson <tejohnson@google.com>
1366 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
1368 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1370 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
1371 and vmulosh rather than call gen_vec_widen_smult_*.
1372 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
1373 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
1374 (vec_widen_smult_even_v16qi): Likewise.
1375 (vec_widen_umult_even_v8hi): Likewise.
1376 (vec_widen_smult_even_v8hi): Likewise.
1377 (vec_widen_umult_odd_v16qi): Likewise.
1378 (vec_widen_smult_odd_v16qi): Likewise.
1379 (vec_widen_umult_odd_v8hi): Likewise.
1380 (vec_widen_smult_odd_v8hi): Likewise.
1381 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
1382 vmuloub rather than call gen_vec_widen_umult_*.
1383 (vec_widen_umult_lo_v16qi): Likewise.
1384 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
1385 vmulosb rather than call gen_vec_widen_smult_*.
1386 (vec_widen_smult_lo_v16qi): Likewise.
1387 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
1388 rather than call gen_vec_widen_umult_*.
1389 (vec_widen_umult_lo_v8hi): Likewise.
1390 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
1391 rather than call gen_vec_widen_smult_*.
1392 (vec_widen_smult_lo_v8hi): Likewise.
1394 2014-01-15 Jeff Law <law@redhat.com>
1396 PR tree-optimization/59747
1397 * ree.c (find_and_remove_re): Properly handle case where a second
1398 eliminated extension requires widening a copy created for elimination
1399 of a prior extension.
1400 (combine_set_extension): Ensure that the number of hard regs needed
1401 for a destination register does not change when we widen it.
1403 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
1405 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
1406 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
1407 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
1408 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
1409 (avr-*-rtems*): Likewise.
1410 (bfin*-rtems*): Likewise.
1411 (moxie-*-rtems*): Likewise.
1412 (h8300-*-rtems*): Likewise.
1413 (i[34567]86-*-rtems*): Likewise.
1414 (lm32-*-rtems*): Likewise.
1415 (m32r-*-rtems*): Likewise.
1416 (m68k-*-rtems*): Likewise.
1417 (microblaze*-*-rtems*): Likewise.
1418 (mips*-*-rtems*): Likewise.
1419 (powerpc-*-rtems*): Likewise.
1420 (sh-*-rtems*): Likewise.
1421 (sparc-*-rtems*): Likewise.
1422 (sparc64-*-rtems*): Likewise.
1423 (v850-*-rtems*): Likewise.
1424 (m32c-*-rtems*): Likewise.
1426 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
1428 PR rtl-optimization/59511
1429 * ira.c (ira_init_register_move_cost): Use memory costs for some
1430 cases of register move cost calculations.
1431 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
1432 instead of BB frequency.
1433 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
1434 * lra-assigns.c (find_hard_regno_for): Ditto.
1436 2014-01-15 Richard Biener <rguenther@suse.de>
1438 PR tree-optimization/59822
1439 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
1440 (vectorizable_load): Use it to hoist defs of uses of invariant
1441 loads out of the loop.
1443 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1444 Kugan Vivekanandarajah <kuganv@linaro.org>
1447 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
1450 2014-01-15 Richard Biener <rguenther@suse.de>
1452 PR rtl-optimization/59802
1453 * lcm.c (compute_available): Use inverted postorder to seed
1454 the initial worklist.
1456 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1459 * config/s390/s390.c (s390_preferred_reload_class): Don't return
1460 ADDR_REGS for invalid symrefs in non-PIC code.
1462 2014-01-15 Jakub Jelinek <jakub@redhat.com>
1465 * builtins.c (determine_block_size): Initialize *probable_max_size
1466 even if len_rtx is CONST_INT.
1468 2014-01-14 Andrew Pinski <apinski@cavium.com>
1470 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
1471 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
1472 (cortexa53_tunings): Likewise.
1473 (aarch64_sched_issue_rate): New function.
1474 (TARGET_SCHED_ISSUE_RATE): Define.
1476 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1478 * ira-costs.c (find_costs_and_classes): Add missed
1479 ira_init_register_move_cost_if_necessary.
1481 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1484 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
1486 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
1489 * config/i386/i386.c (type_natural_mode): Add a bool parameter
1490 to indicate if type is used for function return value. Warn ABI
1491 change if the vector mode isn't available for function return value.
1492 (ix86_function_arg_advance): Pass false to type_natural_mode.
1493 (ix86_function_arg): Likewise.
1494 (ix86_gimplify_va_arg): Likewise.
1495 (function_arg_32): Don't warn ABI change.
1496 (ix86_function_value): Pass true to type_natural_mode.
1497 (ix86_return_in_memory): Likewise.
1498 (ix86_struct_value_rtx): Removed.
1499 (TARGET_STRUCT_VALUE_RTX): Likewise.
1501 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1503 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
1504 converting a conditional jump into a conditional return.
1506 2014-01-14 Richard Biener <rguenther@suse.de>
1508 PR tree-optimization/58921
1509 PR tree-optimization/59006
1510 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
1511 hoisting invariant stmts.
1512 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
1513 invariant loads on the preheader edge if possible.
1515 2014-01-14 Joey Ye <joey.ye@arm.com>
1517 * doc/plugin.texi (Building GCC plugins): Update to C++.
1519 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
1521 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
1522 (_mm_rcp28_round_ss): Ditto.
1523 (_mm_rsqrt28_round_sd): Ditto.
1524 (_mm_rsqrt28_round_ss): Ditto.
1525 (_mm_rcp28_sd): Ditto.
1526 (_mm_rcp28_ss): Ditto.
1527 (_mm_rsqrt28_sd): Ditto.
1528 (_mm_rsqrt28_ss): Ditto.
1529 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
1530 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
1531 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
1532 (IX86_BUILTIN_RCP28SD): Ditto.
1533 (IX86_BUILTIN_RCP28SS): Ditto.
1534 (IX86_BUILTIN_RSQRT28SD): Ditto.
1535 (IX86_BUILTIN_RSQRT28SS): Ditto.
1536 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
1537 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
1538 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
1539 (ix86_expand_special_args_builtin): Expand new FTYPE.
1540 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
1541 (srcp14<mode>): Make insn unary.
1542 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
1543 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
1544 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
1545 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
1546 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
1547 Fix rounding: make it SAE only.
1548 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
1550 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
1552 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
1553 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
1554 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
1555 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
1556 (round_saeonly_mask_scalar_operand4): Ditto.
1557 (round_saeonly_mask_scalar_op3): Ditto.
1558 (round_saeonly_mask_scalar_op4): Ditto.
1560 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1562 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
1563 Implement -maltivec=be for vec_insert and vec_extract.
1565 2014-01-10 DJ Delorie <dj@redhat.com>
1567 * config/msp430/msp430.md (call_internal): Don't allow memory
1568 references with SP as the base register.
1569 (call_value_internal): Likewise.
1570 * config/msp430/constraints.md (Yc): New. For memory references
1571 that don't use SP as a base register.
1573 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
1574 "an integer without a # prefix"
1575 * config/msp430/msp430.md (epilogue_helper): Use it.
1577 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1580 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
1581 AVX512F gather builtins.
1582 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
1583 on gather decls with INTEGER_TYPE masktype.
1584 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
1585 directly into the builtin rather than hoisting it before loop.
1587 PR tree-optimization/59387
1588 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
1589 (scev_const_prop): If folded_casts and type has undefined overflow,
1590 use force_gimple_operand instead of force_gimple_operand_gsi and
1591 for each added stmt if it is assign with
1592 arith_code_with_undefined_signed_overflow, call
1593 rewrite_to_defined_overflow.
1594 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
1595 gimple-fold.h instead.
1596 (arith_code_with_undefined_signed_overflow,
1597 rewrite_to_defined_overflow): Moved to ...
1598 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
1599 rewrite_to_defined_overflow): ... here. No longer static.
1600 Include gimplify-me.h.
1601 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
1602 rewrite_to_defined_overflow): New prototypes.
1604 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1606 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
1608 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
1610 * builtins.c (get_object_alignment_2): Minor tweak.
1611 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
1613 2014-01-13 Christian Bruel <christian.bruel@st.com>
1615 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
1616 optimized non constant lengths.
1618 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1621 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
1622 load as __atomic_load_N if possible.
1624 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1626 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
1628 (rs6000_expand_builtin): Adjust call.
1630 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1633 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
1634 * config/rs6000/rs6000.c: Include target-globals.h.
1635 (rs6000_set_current_function): Instead of doing target_reinit
1636 unconditionally, use save_target_globals_default_opts and
1637 restore_target_globals.
1639 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
1641 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
1642 (rs6000_expand_builtin): Handle mffs and mtfsf.
1643 (rs6000_init_builtins): Define mffs and mtfsf.
1644 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
1645 (rs6000_mffs): New pattern.
1646 (rs6000_mtfsf): New pattern.
1648 2014-01-11 Bin Cheng <bin.cheng@arm.com>
1650 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
1651 Start narrowing with START. Apply candidate-use pair
1652 and check overall cost in narrowing.
1653 (iv_ca_prune): Pass new argument.
1655 2014-01-10 Jeff Law <law@redhat.com>
1658 * ree.c (combine_reaching_defs): Ensure the defining statement
1659 occurs before the extension when optimizing extensions with
1660 different source and destination hard registers.
1662 2014-01-10 Jan Hubicka <jh@suse.cz>
1665 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
1666 vtables into the type inheritance graph.
1668 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1670 PR rtl-optimization/59754
1671 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
1672 modes in the REGNO != REGNO case.
1674 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1676 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
1678 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1680 PR tree-optimization/59745
1681 * tree-predcom.c (tree_predictive_commoning_loop): Call
1682 free_affine_expand_cache if giving up because components is NULL.
1684 * target-globals.c (save_target_globals): Allocate < 4KB structs using
1685 GC in payload of target_globals struct instead of allocating them on
1686 the heap and the larger structs separately using GC.
1687 * target-globals.h (struct target_globals): Make regs, hard_regs,
1688 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
1689 of GTY((skip)) and change type to void *.
1690 (reset_target_globals): Cast loads from those fields to corresponding
1693 2014-01-10 Steve Ellcey <sellcey@mips.com>
1696 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
1697 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
1698 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
1700 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
1703 * aarch64-modes.def (CC_Zmode): New flags mode.
1704 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
1705 represents an equality.
1706 (aarch64_get_condition_code): Handle CC_Zmode.
1707 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
1709 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1711 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
1712 extraction in good case.
1714 2014-01-10 Richard Biener <rguenther@suse.de>
1716 PR tree-optimization/59374
1717 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
1718 checking after SLP discovery. Mark stmts not participating
1719 in any SLP instance properly.
1721 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1723 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
1724 when handling a SET rtx.
1726 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1728 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
1729 (cortex-a57): Likewise.
1730 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
1732 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1734 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
1735 non-iwmmxt builtins.
1737 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
1741 * ipa-devirt.c record_target_from_binfo): Take as argument
1742 stack of binfos and lookup matching one for virtual inheritance.
1743 (possible_polymorphic_call_targets_1): Update.
1745 2014-01-10 Huacai Chen <chenhc@lemote.com>
1747 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
1748 kernel strings for Loongson-2E/2F/3A.
1750 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1753 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
1754 is_gimple_call before calling gimple_call_internal_p.
1756 2014-01-09 Steve Ellcey <sellcey@mips.com>
1758 * Makefile.in (TREE_FLOW_H): Remove.
1759 (TREE_SSA_H): Add file names from tree-flow.h.
1760 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
1761 * tree.h: Remove tree-flow.h reference.
1762 * hash-table.h: Remove tree-flow.h reference.
1763 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
1764 reference with tree-ssa-loop.h.
1766 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1768 * doc/invoke.texi: Add -maltivec={be,le} options, and document
1769 default element-order behavior for -maltivec.
1770 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
1771 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
1772 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
1773 when targeting big endian, at least for now.
1774 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
1776 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1779 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
1780 var satisfies use_register_for_decl, just take into account type
1781 alignment, rather than decl alignment.
1783 PR tree-optimization/59622
1784 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
1785 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
1786 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
1787 Don't devirtualize for inplace at all. For targets.length () == 1,
1788 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
1790 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1792 * config/i386/i386.md (cpu): Remove the unused btver1.
1794 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1796 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
1798 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1801 * tree-core.h (struct target_globals): New forward declaration.
1802 (struct tree_target_option): Add globals field.
1803 * tree.h (TREE_TARGET_GLOBALS): Define.
1804 (prepare_target_option_nodes_for_pch): New prototype.
1805 * target-globals.h (struct target_globals): Define even if
1807 * tree.c (prepare_target_option_node_for_pch,
1808 prepare_target_option_nodes_for_pch): New functions.
1809 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
1810 * config/i386/i386.c: Include target-globals.h.
1811 (ix86_set_current_function): Instead of doing target_reinit
1812 unconditionally, use save_target_globals_default_opts and
1813 restore_target_globals.
1815 2014-01-09 Richard Biener <rguenther@suse.de>
1817 PR tree-optimization/59715
1818 * tree-cfg.h (split_critical_edges): Declare.
1819 * tree-cfg.c (split_critical_edges): Export.
1820 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
1822 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
1824 * cfgexpand.c (expand_stack_vars): Optionally disable
1825 asan stack protection.
1826 (expand_used_vars): Likewise.
1827 (partition_stack_vars): Likewise.
1828 * asan.c (asan_emit_stack_protection): Optionally disable
1829 after return stack usage.
1830 (instrument_derefs): Optionally disable memory access instrumentation.
1831 (instrument_builtin_call): Likewise.
1832 (instrument_strlen_call): Likewise.
1833 (asan_protect_global): Optionally disable global variables protection.
1834 * doc/invoke.texi: Added doc for new options.
1835 * params.def: Added new options.
1836 * params.h: Likewise.
1838 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1840 PR rtl-optimization/59724
1841 * ifcvt.c (cond_exec_process_if_block): Don't call
1842 flow_find_head_matching_sequence with 0 longest_match.
1843 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
1844 non-active insns if !stop_after.
1845 (try_head_merge_bb): Revert 2014-01-07 changes.
1847 2014-01-08 Jeff Law <law@redhat.com>
1849 * ree.c (get_sub_rtx): New function, extracted from...
1850 (merge_def_and_ext): Here.
1851 (combine_reaching_defs): Use get_sub_rtx.
1853 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
1855 * cgraph.h (varpool_variable_node): Do not choke on null node.
1857 2014-01-08 Catherine Moore <clm@codesourcery.com>
1859 * config/mips/mips.md (simple_return): Attempt to use JRC
1861 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
1863 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1865 PR rtl-optimization/59137
1866 * reorg.c (steal_delay_list_from_target): Call update_block for
1868 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
1870 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1872 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
1873 two duplicate entries.
1875 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1878 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
1880 * config/mips/mips.c (mips_truncated_op_cost): New function.
1881 (mips_rtx_costs): Adjust test for BADDU.
1882 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
1884 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
1886 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
1887 (*baddu_si): ...this new pattern.
1889 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1892 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
1894 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1897 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
1899 (expand_expr, expand_normal): Adjust.
1900 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
1901 inner_reference_p. Use inner_reference_p to expand inner references.
1902 (store_expr): Adjust.
1903 * cfgexpand.c (expand_call_stmt): Adjust.
1905 2014-01-08 Rong Xu <xur@google.com>
1907 * gcov-io.c (gcov_var): Move from gcov-io.h.
1908 (gcov_position): Ditto.
1909 (gcov_is_error): Ditto.
1910 (gcov_rewrite): Ditto.
1911 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
1912 only part to libgcc/libgcov.h.
1914 2014-01-08 Marek Polacek <polacek@redhat.com>
1917 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
1919 2014-01-08 Marek Polacek <polacek@redhat.com>
1922 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
1924 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1926 PR rtl-optimization/59649
1927 * stor-layout.c (get_mode_bounds): For BImode return
1928 0 and STORE_FLAG_VALUE.
1930 2014-01-08 Richard Biener <rguenther@suse.de>
1933 * gimple.h (is_gimple_builtin_call): Remove.
1934 (gimple_builtin_call_types_compatible_p): New.
1935 (gimple_call_builtin_p): New overload.
1936 * gimple.c (is_gimple_builtin_call): Remove.
1937 (validate_call): Rename to ...
1938 (gimple_builtin_call_types_compatible_p): ... this and export. Also
1940 (validate_type): New static function.
1941 (gimple_call_builtin_p): New overload and adjust.
1942 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
1943 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
1944 (gimple_fold_stmt_to_constant_1): Likewise.
1945 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
1947 2014-01-08 Richard Biener <rguenther@suse.de>
1950 * gimplify.c (gimplify_expr): Gimplify register-register type
1951 VIEW_CONVERT_EXPRs to separate stmts.
1953 2014-01-07 Jeff Law <law@redhat.com>
1956 * ree.c (combine_set_extension): Handle case where source
1957 and destination registers in an extension insn are different.
1958 (combine_reaching_defs): Allow source and destination registers
1959 in extension to be different under limited circumstances.
1960 (add_removable_extension): Remove restriction that the
1961 source and destination registers in the extension are the same.
1962 (find_and_remove_re): Emit a copy from the extension's
1963 destination to its source after the defining insn if
1964 the source and destination registers are different.
1967 * ifcvt.c (merge_if_block): If we are merging a block with more than
1968 one successor with a block with no successors, remove any BARRIER
1969 after the second block.
1971 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
1973 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
1975 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
1978 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
1979 for 14-bit register offsets when INT14_OK_STRICT is false.
1981 2014-01-07 Roland Stigge <stigge@antcom.de>
1982 Michael Meissner <meissner@linux.vnet.ibm.com>
1985 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
1986 Only check TFmode for SPE constants. Don't check TImode or TDmode.
1988 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
1990 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
1993 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
1995 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
1996 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
1997 rtx is const0_rtx or not.
1999 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
2002 * target-globals.c (save_target_globals): Remove this_fn_optab
2004 * toplev.c: Include optabs.h.
2005 (target_reinit): Temporarily restore the global options if another
2006 set of options are in force.
2008 2014-01-07 Jakub Jelinek <jakub@redhat.com>
2010 PR rtl-optimization/58668
2011 * cfgcleanup.c (flow_find_cross_jump): Don't count
2012 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
2013 to determine what is counted.
2014 (flow_find_head_matching_sequence): Use active_insn_p to determine
2016 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
2018 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
2019 determine what is counted.
2021 PR tree-optimization/59643
2022 * tree-predcom.c (split_data_refs_to_components): If one dr is
2023 read and one write, determine_offset fails and the write isn't
2024 in the bad component, just put the read into the bad component.
2026 2014-01-07 Mike Stump <mikestump@comcast.net>
2027 Jakub Jelinek <jakub@redhat.com>
2030 * tree-core.h (struct tree_optimization_option): Change optabs
2031 type from unsigned char * to void *.
2032 * optabs.c (init_tree_optimization_optabs): Adjust
2033 TREE_OPTIMIZATION_OPTABS initialization.
2035 2014-01-06 Jakub Jelinek <jakub@redhat.com>
2038 * config/i386/i386.h (struct machine_function): Add
2039 no_drap_save_restore field.
2040 * config/i386/i386.c (ix86_save_reg): Use
2041 !cfun->machine->no_drap_save_restore instead of
2042 crtl->stack_realign_needed.
2043 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
2044 this function clears frame_pointer_needed. Set
2045 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
2046 and DRAP reg is needed.
2048 2014-01-06 Marek Polacek <polacek@redhat.com>
2051 * doc/implement-c.texi: Mention that other integer types are
2052 permitted as bit-field types in strictly conforming mode.
2054 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
2056 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
2059 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
2061 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
2063 2014-01-06 Martin Jambor <mjambor@suse.cz>
2066 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
2068 * ipa-prop.c (ipa_print_node_params): Fix indentation.
2070 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
2074 * var-tracking.c (add_stores): Preserve the value of the source even if
2075 we don't record the store.
2077 2014-01-06 Terry Guo <terry.guo@arm.com>
2079 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
2081 2014-01-05 Iain Sandoe <iain@codesourcery.com>
2084 * config/darwin.c (darwin_function_section): Adjust return values to
2085 correspond to optimisation changes made in r206070.
2087 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
2089 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
2090 from prefetch_block tune setting.
2091 (nocona_cost): Correct size of prefetch block to 64.
2093 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
2095 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
2096 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
2097 used to save the static chain register in the computation of the offset
2098 from which the FP registers need to be restored.
2100 2014-01-04 Jakub Jelinek <jakub@redhat.com>
2102 PR tree-optimization/59519
2103 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
2104 ICE if get_current_def (current_new_name) is already non-NULL, as long
2105 as it is a phi result of some other phi in *new_exit_bb that has
2108 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
2109 or vmovdqu* for misaligned_operand.
2110 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
2111 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
2112 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
2113 aligned_mem for AVX512F masked aligned load and store builtins and for
2116 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
2118 PR tree-optimization/59651
2119 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
2120 Address range for negative step should be added by TYPE_SIZE_UNIT.
2122 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
2124 * config/m68k/m68k.c (handle_move_double): Handle pushes with
2125 overlapping registers also for registers other than the stack pointer.
2127 2014-01-03 Marek Polacek <polacek@redhat.com>
2130 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
2133 2014-01-03 Jakub Jelinek <jakub@redhat.com>
2136 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
2139 * config/i386/i386.md (MODE_SIZE): New mode attribute.
2140 (push splitter): Use <P:MODE_SIZE> instead of
2141 GET_MODE_SIZE (<P:MODE>mode).
2142 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
2143 (mov -1, reg peephole2): Likewise.
2144 * config/i386/sse.md (*mov<mode>_internal,
2145 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
2146 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
2147 *<code><mode>3, *andnot<mode>3<mask_name>,
2148 <mask_codefor><code><mode>3<mask_name>): Likewise.
2149 * config/i386/subst.md (mask_mode512bit_condition,
2150 sd_mask_mode512bit_condition): Likewise.
2152 2014-01-02 Xinliang David Li <davidxl@google.com>
2154 PR tree-optimization/59303
2155 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
2156 (dump_predicates): Better output format.
2157 (pred_equal_p): New function.
2158 (is_neq_relop_p): Ditto.
2159 (is_neq_zero_form_p): Ditto.
2160 (pred_expr_equal_p): Ditto.
2161 (pred_neg_p): Ditto.
2162 (simplify_pred): Ditto.
2163 (simplify_preds_2): Ditto.
2164 (simplify_preds_3): Ditto.
2165 (simplify_preds_4): Ditto.
2166 (simplify_preds): Ditto.
2168 (push_to_worklist): Ditto.
2169 (get_pred_info_from_cmp): Ditto.
2170 (is_degenerated_phi): Ditto.
2171 (normalize_one_pred_1): Ditto.
2172 (normalize_one_pred): Ditto.
2173 (normalize_one_pred_chain): Ditto.
2174 (normalize_preds): Ditto.
2175 (normalize_cond_1): Remove function.
2176 (normalize_cond): Ditto.
2177 (is_gcond_subset_of): Ditto.
2178 (is_subset_of_any): Ditto.
2179 (is_or_set_subset_of): Ditto.
2180 (is_and_set_subset_of): Ditto.
2181 (is_norm_cond_subset_of): Ditto.
2182 (pred_chain_length_cmp): Ditto.
2183 (convert_control_dep_chain_into_preds): Type change.
2184 (find_predicates): Ditto.
2185 (find_def_preds): Ditto.
2186 (destroy_predicates_vecs): Ditto.
2187 (find_matching_predicates_in_rest_chains): Ditto.
2188 (use_pred_not_overlap_with_undef_path_pred): Ditto.
2189 (is_pred_expr_subset): Ditto.
2190 (is_pred_chain_subset_of): Ditto.
2191 (is_included_in): Ditto.
2192 (is_superset_of): Ditto.
2194 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2196 Update copyright years.
2198 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2200 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
2201 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
2202 config/arc/arc.md, config/arc/arc.opt,
2203 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
2204 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
2205 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
2206 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
2207 config/linux-protos.h, config/linux.c, config/winnt-c.c,
2208 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
2209 vtable-verify.c, vtable-verify.h: Use the standard form for the
2212 2014-01-02 Tobias Burnus <burnus@net-b.de>
2214 * gcc.c (process_command): Update copyright notice dates.
2215 * gcov-dump.c: Ditto.
2217 * doc/cpp.texi: Bump @copying's copyright year.
2218 * doc/cppinternals.texi: Ditto.
2219 * doc/gcc.texi: Ditto.
2220 * doc/gccint.texi: Ditto.
2221 * doc/gcov.texi: Ditto.
2222 * doc/install.texi: Ditto.
2223 * doc/invoke.texi: Ditto.
2225 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2227 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
2229 2014-01-01 Jakub Jelinek <jakub@redhat.com>
2231 * config/i386/sse.md (*mov<mode>_internal): Guard
2232 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
2234 PR rtl-optimization/59647
2235 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
2236 new_rtx into UNSIGNED_FLOAT rtxes.
2238 Copyright (C) 2014 Free Software Foundation, Inc.
2240 Copying and distribution of this file, with or without modification,
2241 are permitted in any medium without royalty provided the copyright
2242 notice and this notice are preserved.