1 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
4 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
7 2014-01-29 DJ Delorie <dj@redhat.com>
9 * config/msp430/msp430.opt (-minrt): New.
10 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
12 (ENDFILE_SPEC): Likewise.
14 2014-01-29 Jan Hubicka <jh@suse.cz>
16 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
17 (estimate_function_body_sizes): Use it.
19 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
22 * dwarf2out.c (is_cxx_auto): New.
23 (is_base_type): Use it.
24 (gen_type_die_with_usage): Likewise.
26 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
28 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
29 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
30 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
31 -maltivec=be with LE targets.
32 (vsx_mergeh_<mode>): Likewise.
33 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New
35 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
36 (altivec_vmrghb): Replace with define_expand and new
37 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE
39 (altivec_vmrghb_direct): New define_insn.
40 (altivec_vmrghh): Replace with define_expand and new
41 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE
43 (altivec_vmrghh_direct): New define_insn.
44 (altivec_vmrghw): Replace with define_expand and new
45 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE
47 (altivec_vmrghw_direct): New define_insn.
48 (*altivec_vmrghsf): Adjust for endianness.
49 (altivec_vmrglb): Replace with define_expand and new
50 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE
52 (altivec_vmrglb_direct): New define_insn.
53 (altivec_vmrglh): Replace with define_expand and new
54 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE
56 (altivec_vmrglh_direct): New define_insn.
57 (altivec_vmrglw): Replace with define_expand and new
58 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE
60 (altivec_vmrglw_direct): New define_insn.
61 (*altivec_vmrglsf): Adjust for endianness.
62 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
63 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
64 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
65 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
66 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
67 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
68 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
69 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
71 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
73 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
74 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
77 2014-01-29 Richard Biener <rguenther@suse.de>
79 PR tree-optimization/58742
80 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
81 associate_pointerplus_align.
82 (associate_pointerplus_diff): New function.
83 (associate_pointerplus): Likewise. Call associate_pointerplus_align
84 and associate_pointerplus_diff.
86 2014-01-29 Richard Biener <rguenther@suse.de>
88 * lto-streamer.h (LTO_major_version): Bump to 3.
89 (LTO_minor_version): Reset to 0.
91 2014-01-29 Renlin Li <Renlin.Li@arm.com>
93 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
94 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
95 (arm_file_start): Generate correct asm header for armv7ve.
96 * config/arm/bpabi.h: Add multilib support for armv7ve.
97 * config/arm/driver-arm.c: Change the architectures of cortex-a7
98 and cortex-a15 to armv7ve.
99 * config/arm/t-aprofile: Add multilib support for armv7ve.
100 * doc/invoke.texi: Document -march=armv7ve.
102 2014-01-29 Richard Biener <rguenther@suse.de>
104 PR tree-optimization/58742
105 * tree-ssa-forwprop.c (associate_plusminus): Return true
106 if we changed sth, defer EH cleanup to ...
107 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
108 (simplify_mult): New function.
110 2014-01-29 Jakub Jelinek <jakub@redhat.com>
113 PR tree-optimization/59920
114 * tree.c (build_common_builtin_nodes): Remove
115 __builtin_setjmp_dispatcher initialization.
116 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
117 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
118 instead of gsi_after_labels + manually skipping debug stmts.
119 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
120 ignore bbs with IFN_ABNORMAL_DISPATCHER.
121 * tree-inline.c (copy_edges_for_bb): Remove
122 can_make_abnormal_goto argument, instead add abnormal_goto_dest
123 argument. Ignore computed_goto_p stmts. Don't call
124 make_abnormal_goto_edges. If a call might need abnormal edges
125 for non-local gotos, see if it already has an edge to
126 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
127 with true argument, don't do anything then, otherwise add
128 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
129 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
131 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
132 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
133 (lower_stmt): Don't set data->calls_builtin_setjmp.
134 (lower_builtin_setjmp): Adjust comment.
135 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
136 * tree-cfg.c (found_computed_goto): Remove.
137 (factor_computed_gotos): Remove.
138 (make_goto_expr_edges): Return bool, true for computed gotos.
139 Don't call make_abnormal_goto_edges.
140 (build_gimple_cfg): Don't set found_computed_goto, don't call
141 factor_computed_gotos.
142 (computed_goto_p): No longer static.
143 (make_blocks): Don't set found_computed_goto.
144 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
145 (make_edges): If make_goto_expr_edges returns true, push bb
146 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
147 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
148 vector. Record mapping between bbs and OpenMP regions if there
149 are any, adjust make_gimple_omp_edges caller. Call
150 handle_abnormal_edges.
151 (make_abnormal_goto_edges): Remove.
152 * tree-cfg.h (make_abnormal_goto_edges): Remove.
153 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
154 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
155 * builtins.c (expand_builtin): Don't handle
156 BUILT_IN_SETJMP_DISPATCHER.
157 * internal-fn.def (ABNORMAL_DISPATCHER): New.
158 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
159 filling *region also set *region_idx to (*region)->entry->index.
162 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
163 For REGs set ORIGINAL_REGNO.
165 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
167 * doc/md.texi: Mention that a target shouldn't implement
168 vec_widen_(s|u)mul_even/odd pair if it is less efficient
171 2014-01-29 Jakub Jelinek <jakub@redhat.com>
173 PR tree-optimization/59594
174 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
175 a copy of the datarefs vector rather than the vector itself.
177 2014-01-28 Jason Merrill <jason@redhat.com>
180 * dwarf2out.c (auto_die): New static.
181 (gen_type_die_with_usage): Handle C++1y 'auto'.
182 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
185 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
188 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
189 (SPEC_X32): Likewise.
191 * config/i386/i386.c (ix86_option_override_internal): Turn off
192 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
194 (x86_file_start): Output .code16gcc for TARGET_16BIT.
195 * config/i386/i386.h (TARGET_16BIT): New macro.
196 (TARGET_16BIT_P): Likewise.
197 * config/i386/i386.opt: Add m16.
198 * doc/invoke.texi: Document -m16.
200 2014-01-28 Jakub Jelinek <jakub@redhat.com>
202 PR preprocessor/59935
203 * input.c (location_get_source_line): Bail out on when line number
204 is zero, and test the return value of
205 lookup_or_add_file_to_cache_tab.
207 2014-01-28 Richard Biener <rguenther@suse.de>
209 PR tree-optimization/58742
210 * tree-ssa-forwprop.c (associate_plusminus): Handle
211 pointer subtraction of the form (T)(P + A) - (T)P.
213 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
215 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
218 2014-01-28 Richard Biener <rguenther@suse.de>
221 2014-01-28 Richard Biener <rguenther@suse.de>
223 PR rtl-optimization/45364
224 PR rtl-optimization/59890
225 * var-tracking.c (local_get_addr_clear_given_value): Handle
226 already cleared slot.
227 (val_reset): Handle not allocated local_get_addr_cache.
228 (vt_find_locations): Use post-order on the inverted CFG.
230 2014-01-28 Richard Biener <rguenther@suse.de>
232 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps):
235 2014-01-28 Richard Biener <rguenther@suse.de>
237 PR rtl-optimization/45364
238 PR rtl-optimization/59890
239 * var-tracking.c (local_get_addr_clear_given_value): Handle
240 already cleared slot.
241 (val_reset): Handle not allocated local_get_addr_cache.
242 (vt_find_locations): Use post-order on the inverted CFG.
244 2014-01-28 Alan Modra <amodra@gmail.com>
246 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
247 * configure.ac <recursive call for build != host>: Define
248 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
249 and LD_FOR_BUILD too.
250 * configure: Regenerate.
252 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
254 * config/i386/i386.c (get_builtin_code_for_version): Separate
255 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
256 Broadwell from Haswell.
258 2014-01-27 Steve Ellcey <sellcey@mips.com>
260 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
261 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
262 * config/mips/mips.c (mips_option_override): Change setting
264 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
265 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
266 Change from Mask to Var.
268 2014-01-27 Jeff Law <law@redhat.com>
270 * ipa-inline.c (inline_small_functions): Fix typo.
272 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
274 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
275 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
276 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
277 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
278 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
279 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
280 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
281 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
282 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
283 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
284 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
285 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
286 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
287 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
288 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
289 (_mm512_storeu_epi64): Ditto.
290 (_mm512_cmpge_epi32_mask): Ditto.
291 (_mm512_cmpge_epu32_mask): Ditto.
292 (_mm512_cmpge_epi64_mask): Ditto.
293 (_mm512_cmpge_epu64_mask): Ditto.
294 (_mm512_cmple_epi32_mask): Ditto.
295 (_mm512_cmple_epu32_mask): Ditto.
296 (_mm512_cmple_epi64_mask): Ditto.
297 (_mm512_cmple_epu64_mask): Ditto.
298 (_mm512_cmplt_epi32_mask): Ditto.
299 (_mm512_cmplt_epu32_mask): Ditto.
300 (_mm512_cmplt_epi64_mask): Ditto.
301 (_mm512_cmplt_epu64_mask): Ditto.
302 (_mm512_cmpneq_epi32_mask): Ditto.
303 (_mm512_cmpneq_epu32_mask): Ditto.
304 (_mm512_cmpneq_epi64_mask): Ditto.
305 (_mm512_cmpneq_epu64_mask): Ditto.
306 (_mm512_expand_pd): Ditto.
307 (_mm512_expand_ps): Ditto.
308 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
309 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
310 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
311 * config/i386/i386.c (ix86_builtins): Add
312 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
313 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
314 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
315 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
316 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
317 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
318 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
319 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
320 IX86_BUILTIN_PMOVUSQW512_MEM.
321 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
322 __builtin_ia32_pmovsqd512mem_mask,
323 __builtin_ia32_pmovqd512mem_mask,
324 __builtin_ia32_pmovusqw512mem_mask,
325 __builtin_ia32_pmovsqw512mem_mask,
326 __builtin_ia32_pmovqw512mem_mask,
327 __builtin_ia32_pmovusdw512mem_mask,
328 __builtin_ia32_pmovsdw512mem_mask,
329 __builtin_ia32_pmovdw512mem_mask,
330 __builtin_ia32_pmovqb512mem_mask,
331 __builtin_ia32_pmovusqb512mem_mask,
332 __builtin_ia32_pmovsqb512mem_mask,
333 __builtin_ia32_pmovusdb512mem_mask,
334 __builtin_ia32_pmovsdb512mem_mask,
335 __builtin_ia32_pmovdb512mem_mask.
336 (bdesc_args): Add __builtin_ia32_expanddf512,
337 __builtin_ia32_expandsf512.
338 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
339 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
340 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
341 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
342 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
343 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
344 (avx512f_<code>v8div16qi2_mask_store): This.
345 (avx512f_expand<mode>): New.
347 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
349 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
351 (_mm512_mask_prefetch_i64gather_pd): Ditto.
352 (_mm512_prefetch_i32scatter_pd): Ditto.
353 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
354 (_mm512_prefetch_i64scatter_pd): Ditto.
355 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
356 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
357 (_mm512_mask_prefetch_i64gather_ps): Ditto.
358 (_mm512_prefetch_i32scatter_ps): Ditto.
359 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
360 (_mm512_prefetch_i64scatter_ps): Ditto.
361 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
362 * config/i386/i386-builtin-types.def: Define
363 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
364 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
365 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
366 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
367 IX86_BUILTIN_SCATTERPFQPD.
368 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
369 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
370 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
371 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
372 __builtin_ia32_scatterpfqps.
373 (ix86_expand_builtin): Expand new built-ins.
374 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
375 fix memory access data type.
376 (*avx512pf_gatherpf<mode>_mask): Ditto.
377 (*avx512pf_gatherpf<mode>): Ditto.
378 (avx512pf_scatterpf<mode>): Ditto.
379 (*avx512pf_scatterpf<mode>_mask): Ditto.
380 (*avx512pf_scatterpf<mode>): Ditto.
381 (GATHER_SCATTER_SF_MEM_MODE): New.
382 (avx512pf_gatherpf<mode>df): Ditto.
383 (*avx512pf_gatherpf<mode>df_mask): Ditto.
384 (*avx512pf_scatterpf<mode>df): Ditto.
386 2014-01-27 Jakub Jelinek <jakub@redhat.com>
389 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
390 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
393 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
395 * common/config/arm/arm-common.c
396 (arm_rewrite_mcpu): Handle multiple names.
398 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
400 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
402 * gimple-builder.h (create_gimple_tmp): Delete.
404 2014-01-27 Christian Bruel <christian.bruel@st.com>
406 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
409 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
411 * config/pa/pa.md (call): Generate indirect long calls to non-local
412 functions when outputing 32-bit code.
413 (call_value): Likewise except for special call to buggy powf function.
415 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
416 portable runtime and PIC indirect calls.
417 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
418 and PIC call sequences. Use ldo instead of blr to set return register
419 in PIC call sequence.
421 2014-01-25 Walter Lee <walt@tilera.com>
423 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
424 avoid clobbering a live register.
426 2014-01-25 Walter Lee <walt@tilera.com>
428 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
429 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
430 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
431 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
433 2014-01-25 Walter Lee <walt@tilera.com>
435 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
436 arguments on even registers.
437 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
439 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
440 (BIGGEST_ALIGNMENT): Ditto.
441 (BIGGEST_FIELD_ALIGNMENT): Ditto.
443 2014-01-25 Walter Lee <walt@tilera.com>
445 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
446 insns before bundling.
447 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
449 2014-01-25 Walter Lee <walt@tilera.com>
451 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
452 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
453 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
455 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
457 * config/mips/constraints.md (kl): Delete.
458 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
459 define expands, using...
460 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
461 instructions for MIPS16.
462 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
463 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
465 2014-01-25 Walter Lee <walt@tilera.com>
467 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
471 2014-01-25 Walter Lee <walt@tilera.com>
473 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
474 (TARGET_EXPAND_TO_RTL_HOOK): Define.
476 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
478 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
481 2014-01-25 Jakub Jelinek <jakub@redhat.com>
483 * print-rtl.c (in_call_function_usage): New var.
484 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
485 EXPR_LIST mode as mode and not as reg note name.
488 * cfgloopmanip.c (copy_loop_info): If
489 loop->warned_aggressive_loop_optimizations, make sure
490 the flag is set in target loop too.
492 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
494 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
496 * builtins.def: Likewise.
497 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
498 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
499 * ira.c (ira_setup_eliminable_regset): Likewise.
500 * omp-low.c (gate_expand_omp): Likewise.
501 (execute_lower_omp): Likewise.
502 (diagnose_sb_0): Likewise.
503 (gate_diagnose_omp_blocks): Likewise.
504 (simd_clone_clauses_extract): Likewise.
507 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
509 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
510 correction for little endian...
511 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
514 2014-01-24 Jeff Law <law@redhat.com>
516 PR tree-optimization/59919
517 * tree-vrp.c (find_assert_locations_1): Do not register asserts
518 for non-returning calls.
520 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
522 * common/config/aarch64/aarch64-common.c
523 (aarch64_rewrite_mcpu): Handle multiple names.
524 * config/aarch64/aarch64.h
525 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
527 2014-01-24 Dodji Seketeli <dodji@redhat.com>
529 * input.c (add_file_to_cache_tab): Handle the case where fopen
532 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
535 * config/i386/i386.md (pushsf splitter): Get stack adjustment
536 from push operand if code of push isn't PRE_DEC.
538 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
541 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
542 -mquad-memory-atomic. Update -mquad-memory documentation to say
543 it is only used for non-atomic loads/stores.
545 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
546 -mquad-memory or -mquad-memory-atomic switches.
548 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
549 -mquad-memory-atomic to ISA 2.07 support.
551 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
552 to separate support of normal quad word memory operations (ldq, stq)
553 from the atomic quad word memory operations.
555 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
556 support to separate non-atomic quad word operations from atomic
557 quad word operations. Disable non-atomic quad word operations in
558 little endian mode so that we don't have to swap words after the
559 load and before the store.
560 (quad_load_store_p): Add comment about atomic quad word support.
561 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
562 options printed with -mdebug=reg.
564 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
565 -mquad-memory-atomic as the test for whether we have quad word
567 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
568 or -mp8-vector are used, allow byte/half-word atomic operations.
570 * config/rs6000/sync.md (load_lockedti): Insure that the address
571 is a proper indexed or indirect address for the lqarx instruction.
572 On little endian systems, swap the hi/lo registers after the lqarx
574 (load_lockedpti): Use indexed_or_indirect_operand predicate to
575 insure the address is valid for the lqarx instruction.
576 (store_conditionalti): Insure that the address is a proper indexed
577 or indirect address for the stqcrx. instruction. On little endian
578 systems, swap the hi/lo registers before doing the stqcrx.
580 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
581 insure the address is valid for the stqcrx. instruction.
583 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
584 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
585 type of quad memory support is available.
587 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
590 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
591 there is a danger of looping.
593 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
595 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
596 force flag_ira_loop_pressure if set via command line.
598 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
600 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
601 (ashr_simd): New builtin handling DI mode.
602 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
603 (aarch64_sshr_simddi): New match pattern.
604 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
605 (vshrd_n_s64): Likewise.
606 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
608 2014-01-23 Nick Clifton <nickc@redhat.com>
610 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
611 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
612 favour of mcu specific scripts.
613 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
616 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
617 Alex Velenko <Alex.Velenko@arm.com>
619 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
620 (vaddv_s16): Likewise.
621 (vaddv_s32): Likewise.
622 (vaddv_u8): Likewise.
623 (vaddv_u16): Likewise.
624 (vaddv_u32): Likewise.
625 (vaddvq_s8): Likewise.
626 (vaddvq_s16): Likewise.
627 (vaddvq_s32): Likewise.
628 (vaddvq_s64): Likewise.
629 (vaddvq_u8): Likewise.
630 (vaddvq_u16): Likewise.
631 (vaddvq_u32): Likewise.
632 (vaddvq_u64): Likewise.
633 (vaddv_f32): Likewise.
634 (vaddvq_f32): Likewise.
635 (vaddvq_f64): Likewise.
636 (vmaxv_f32): Likewise.
637 (vmaxv_s8): Likewise.
638 (vmaxv_s16): Likewise.
639 (vmaxv_s32): Likewise.
640 (vmaxv_u8): Likewise.
641 (vmaxv_u16): Likewise.
642 (vmaxv_u32): Likewise.
643 (vmaxvq_f32): Likewise.
644 (vmaxvq_f64): Likewise.
645 (vmaxvq_s8): Likewise.
646 (vmaxvq_s16): Likewise.
647 (vmaxvq_s32): Likewise.
648 (vmaxvq_u8): Likewise.
649 (vmaxvq_u16): Likewise.
650 (vmaxvq_u32): Likewise.
651 (vmaxnmv_f32): Likewise.
652 (vmaxnmvq_f32): Likewise.
653 (vmaxnmvq_f64): Likewise.
654 (vminv_f32): Likewise.
655 (vminv_s8): Likewise.
656 (vminv_s16): Likewise.
657 (vminv_s32): Likewise.
658 (vminv_u8): Likewise.
659 (vminv_u16): Likewise.
660 (vminv_u32): Likewise.
661 (vminvq_f32): Likewise.
662 (vminvq_f64): Likewise.
663 (vminvq_s8): Likewise.
664 (vminvq_s16): Likewise.
665 (vminvq_s32): Likewise.
666 (vminvq_u8): Likewise.
667 (vminvq_u16): Likewise.
668 (vminvq_u32): Likewise.
669 (vminnmv_f32): Likewise.
670 (vminnmvq_f32): Likewise.
671 (vminnmvq_f64): Likewise.
673 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
675 * config/aarch64/aarch64-simd.md
676 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
677 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
678 (*aarch64_mul3_elt<mode>): Likewise.
679 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
680 (*aarch64_mul3_elt_to_64v2df): Likewise.
681 (*aarch64_mla_elt<mode>): Likewise.
682 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
683 (*aarch64_mls_elt<mode>): Likewise.
684 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
685 (*aarch64_fma4_elt<mode>): Likewise.
686 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
687 (*aarch64_fma4_elt_to_64v2df): Likewise.
688 (*aarch64_fnma4_elt<mode>): Likewise.
689 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
690 (*aarch64_fnma4_elt_to_64v2df): Likewise.
691 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
692 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
693 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
694 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
695 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
696 (aarch64_sqdmull_lane<mode>_internal): Likewise.
697 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
699 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
701 * config/aarch64/aarch64-simd.md
702 (aarch64_be_checked_get_lane<mode>): New define_expand.
703 * config/aarch64/aarch64-simd-builtins.def
704 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
705 New builtin definition.
706 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
707 Use new safe be builtin.
709 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
711 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
713 (aarch64_be_st1<mode>): Likewise.
714 (aarch_ld1<VALL:mode>): Define_expand modified.
715 (aarch_st1<VALL:mode>): Likewise.
716 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
717 (UNSPEC_ST1): Likewise.
719 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
721 * config/microblaze/microblaze.md: Add trap insn and attribute
723 2014-01-23 Dodji Seketeli <dodji@redhat.com>
725 PR preprocessor/58580
726 * input.h (location_get_source_line): Take an additional line_size
728 (void diagnostics_file_cache_fini): Declare new function.
729 * input.c (struct fcache): New type.
730 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
731 New static constants.
732 (diagnostic_file_cache_init, total_lines_num)
733 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
734 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
735 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
736 (get_next_line, read_next_line, goto_next_line, read_line_num):
737 New static function definitions.
738 (diagnostic_file_cache_fini): New function.
739 (location_get_source_line): Take an additional output line_len
740 parameter. Re-write using lookup_or_add_file_to_cache_tab and
742 * diagnostic.c (diagnostic_finish): Call
743 diagnostic_file_cache_fini.
744 (adjust_line): Take an additional input parameter for the length
745 of the line, rather than calculating it with strlen.
746 (diagnostic_show_locus): Adjust the use of
747 location_get_source_line and adjust_line with respect to their new
748 signature. While displaying a line now, do not stop at the first
749 null byte. Rather, display the zero byte as a space and keep
750 going until we reach the size of the line.
751 * Makefile.in: Add vec.o to OBJS-libcommon
753 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
754 Ilya Tocar <ilya.tocar@intel.com>
756 * config/i386/avx512fintrin.h (_mm512_kmov): New.
757 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
758 (__builtin_ia32_kmov16): Ditto.
759 * config/i386/i386.md (UNSPEC_KMOV): New.
762 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
764 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
765 (_mm512_storeu_si512): Ditto.
767 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
770 * rtl.h (get_referenced_operands): Declare.
771 * recog.c (get_referenced_operands): New function.
772 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
773 operands have been referenced when recording LO_SUM references.
775 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
777 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
779 2014-01-22 Jan Hubicka <jh@suse.cz>
781 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
782 Enable for generic and recent AMD targets.
784 2014-01-22 Jan Hubicka <jh@suse.cz>
786 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
787 ARG_SIZE note when adjustment was eliminated.
789 2014-01-22 Jeff Law <law@redhat.com>
791 PR tree-optimization/59597
792 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
793 in file. Accept new argument REGISTERING and use it to modify
794 dump output appropriately.
795 (register_jump_thread): Corresponding changes.
796 (mark_threaded_blocks): Reinstate code to cancel unprofitable
797 thread paths involving joiner blocks. Add code to dump cancelled
798 jump threading paths.
800 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
802 PR rtl-optimization/59477
803 * lra-constraints.c (inherit_in_ebb): Process call for living hard
804 regs. Update reloads_num and potential_reload_hard_regs for all insns.
806 2014-01-22 Tom Tromey <tromey@redhat.com>
808 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
810 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
812 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
814 PR rtl-optimization/59896
815 * lra-constraints.c (process_alt_operands): Check unused note for
816 matched operands of insn with no output reloads.
818 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
820 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
821 (mips_move_from_gpr_cost): Likewise.
823 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
825 PR rtl-optimization/59858
826 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
827 ira_class_hard_regs_num.
828 (process_alt_operands): Increase reject for dying matched operand.
830 2014-01-21 Jakub Jelinek <jakub@redhat.com>
833 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
834 smaller than size, perform several stores or loads and stores
835 at dst + count - size to store or copy all of size bytes, rather
836 than just last modesize bytes.
838 2014-01-20 DJ Delorie <dj@redhat.com>
840 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
841 that CLOBBERs are REGs before propogating their values.
843 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
846 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
847 (cgraph_inline_failed_type): New function.
848 * cgraph.h (DEFCIFCODE): Add type.
849 (cgraph_inline_failed_type_t): New enum.
850 (cgraph_inline_failed_type): New prototype.
851 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
852 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
853 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
854 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
855 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
856 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
857 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
858 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
859 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
860 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
861 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
862 OPTIMIZATION_MISMATCH.
863 * tree-inline.c (expand_call_inline): Emit errors during
864 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
866 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
869 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
870 mode attribute in insn output.
872 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
874 * output.h (output_constant): Delete.
875 * varasm.c (output_constant): Make private.
877 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
879 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
881 2014-01-20 Jakub Jelinek <jakub@redhat.com>
884 * tree.h (fold_builtin_strcat): New prototype.
885 * builtins.c (fold_builtin_strcat): No longer static. Add len
886 argument, if non-NULL, don't call c_strlen. Optimize
887 directly into __builtin_memcpy instead of __builtin_strcpy.
888 (fold_builtin_2): Adjust fold_builtin_strcat caller.
889 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
891 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
893 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
894 for SImode_address_operand operands, having only a REG argument.
896 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
898 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
899 loader name using mbig-endian.
900 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
902 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
904 * doc/invoke.texi (-march): Clarify documentation for AArch64.
908 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
910 * config/aarch64/aarch64-protos.h
911 (aarch64_cannot_change_mode_class_ptr): Declare.
912 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
913 aarch64_cannot_change_mode_class_ptr): New.
914 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
915 backend hook aarch64_cannot_change_mode_class.
917 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
919 * common/config/aarch64/aarch64-common.c
920 (aarch64_handle_option): Don't handle any option order logic here.
921 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
922 selected_cpu, warn on architecture version mismatch.
923 (aarch64_override_options): Fix parsing order for option strings.
925 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
926 Iain Sandoe <iain@codesourcery.com>
929 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
930 warning. Amend comment to reflect current functionality.
932 2014-01-20 Richard Biener <rguenther@suse.de>
935 * builtins.c (fold_builtin_strcat): Remove case better handled
936 by tree-ssa-strlen.c.
938 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
940 * config/aarch64/aarch64.opt
941 (mcpu, march, mtune): Make case-insensitive.
943 2014-01-20 Jakub Jelinek <jakub@redhat.com>
946 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
947 if operands[1] is a REG or ZERO_EXTEND of a REG.
949 2014-01-19 Jan Hubicka <jh@suse.cz>
951 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
953 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
955 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
956 long non-pic millicode calls.
958 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
960 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
962 2014-01-19 Kito Cheng <kito@0xlab.org>
964 * builtins.c (expand_movstr): Check movstr expand done or fail.
966 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
967 H.J. Lu <hongjiu.lu@intel.com>
970 * config/i386/i386.md (*lea<mode>): Zero-extend return register
971 to DImode for zero-extended addresses.
973 2014-01-19 Jakub Jelinek <jakub@redhat.com>
975 PR rtl-optimization/57763
976 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
977 on the new indirect jump_insn and increment LABEL_NUSES (label).
979 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
983 * config.gcc (x86_archs): New variable.
984 (x86_64_archs): Likewise.
985 (x86_cpus): Likewise.
986 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
987 --with-arch/--with-cpu= options.
988 Support --with-arch=/--with-cpu={nehalem,westmere,
989 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
991 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
993 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
994 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
996 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
998 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
1000 2014-01-18 Jakub Jelinek <jakub@redhat.com>
1003 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
1004 clear cpp_get_options (parse_in)->warn_unused_macros for
1005 ix86_target_macros_internal with cpp_define.
1007 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
1009 * jump.c (delete_related_insns): Keep (use (insn))s.
1010 * reorg.c (redundant_insn): Check for barriers too.
1012 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1014 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
1016 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
1018 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
1019 call to $$dyncall when TARGET_LONG_CALLS is true.
1021 2014-01-17 Jeff Law <law@redhat.com>
1023 * ree.c (combine_set_extension): Temporarily disable test for
1024 changing number of hard registers.
1026 2014-01-17 Jan Hubicka <jh@suse.cz>
1029 * ipa-inline-analysis.c (inline_free_summary):
1030 Do not free summary of aliases.
1032 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1035 * gimplify.c (gimplify_expr): Use create_tmp_var
1036 instead of create_tmp_var_raw. If cond doesn't have
1037 integral type, don't add the IFN_ANNOTATE builtin at all.
1039 2014-01-17 Martin Jambor <mjambor@suse.cz>
1042 * ipa-cp.c (prev_edge_clone): New variable.
1043 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
1044 Also resize prev_edge_clone vector.
1045 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
1046 (ipcp_edge_removal_hook): New function.
1047 (ipcp_driver): Register ipcp_edge_removal_hook.
1049 2014-01-17 Andrew Pinski <apinski@cavium.com>
1050 Steve Ellcey <sellcey@mips.com>
1053 * config/mips/mips.c (mips_print_operand): Check operand mode instead
1056 2014-01-17 Jeff Law <law@redhat.com>
1059 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
1060 so that pass_ccp runs first.
1062 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1064 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
1065 (ix86_adjust_cost): Use !TARGET_XXX.
1066 (do_reorder_for_imul): Likewise.
1067 (swap_top_of_ready_list): Likewise.
1068 (ix86_sched_reorder): Likewise.
1070 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1072 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1073 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
1074 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
1075 (intel_memset): New. Duplicate slm_memset.
1076 (intel_cost): New. Duplicate slm_cost.
1077 (m_INTEL): New macro.
1078 (processor_target_table): Add "intel".
1079 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
1080 with PROCESSOR_INTEL for "intel".
1081 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
1082 PROCESSOR_SILVERMONT.
1083 (ix86_issue_rate): Likewise.
1084 (ix86_adjust_cost): Likewise.
1085 (ia32_multipass_dfa_lookahead): Likewise.
1086 (swap_top_of_ready_list): Likewise.
1087 (ix86_sched_reorder): Likewise.
1088 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
1089 instead of TARGET_OPT_AGU.
1090 * config/i386/i386.h (TARGET_INTEL): New.
1091 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
1092 (processor_type): Add PROCESSOR_INTEL.
1093 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
1094 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
1096 2014-01-17 Marek Polacek <polacek@redhat.com>
1099 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
1102 2014-01-17 Richard Biener <rguenther@suse.de>
1104 PR tree-optimization/46590
1105 * opts.c (default_options_table): Add entries for
1106 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
1107 all enabled at -O1 but not for -Og.
1108 * common.opt (fbranch-count-reg): Remove Init(1).
1109 (fmove-loop-invariants): Likewise.
1110 (ftree-pta): Likewise.
1112 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1114 * config/i386/i386.c (ix86_data_alignment): For compatibility with
1115 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
1116 decls to at least the GCC 4.8 used alignments.
1119 * tree-nested.c (convert_nonlocal_reference_stmt,
1120 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
1121 of GIMPLE_BIND stmts, adjust associated decls.
1123 2014-01-17 Richard Biener <rguenther@suse.de>
1125 PR tree-optimization/46590
1126 * vec.h (vec<>::bseach): New member function implementing
1127 binary search according to C89 bsearch.
1128 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
1129 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
1130 bitmap pointer again. Make accesses_in_loop a flat array.
1131 (mem_ref_obstack): New global.
1132 (outermost_indep_loop): Adjust for mem_ref->stored changes.
1133 (mark_ref_stored): Likewise.
1134 (ref_indep_loop_p_2): Likewise.
1135 (set_ref_stored_in_loop): New helper function.
1136 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
1137 (memref_free): Adjust.
1138 (record_mem_ref_loc): Simplify.
1139 (gather_mem_refs_stmt): Adjust.
1140 (sort_locs_in_loop_postorder_cmp): New function.
1141 (analyze_memory_references): Sort accesses_in_loop after
1142 loop postorder number.
1143 (find_ref_loc_in_loop_cmp): New function.
1144 (for_all_locs_in_loop): Find relevant cluster of locs in
1145 accesses_in_loop and iterate without recursion.
1146 (execute_sm): Avoid uninit warning.
1147 (struct ref_always_accessed): Simplify.
1148 (ref_always_accessed::operator ()): Likewise.
1149 (ref_always_accessed_p): Likewise.
1150 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
1151 loop postorder numbers here.
1152 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
1155 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
1158 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
1159 on decls for which assemble_alias has been called.
1161 2014-01-17 Nick Clifton <nickc@redhat.com>
1163 * config/msp430/msp430.opt: (mcpu): New option.
1164 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
1165 (msp430_option_override): Parse target_cpu. If the MCU name
1166 matches a generic string, clear target_mcu.
1167 (msp430_attr): Allow numeric interrupt values up to 63.
1168 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
1169 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
1171 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
1173 * config/msp430/msp430.md (popm): Use %J rather than %I.
1174 (addsi3): Use msp430_nonimmediate_operand for operand 2.
1175 (addhi_cy_i): Use immediate_operand for operand 2.
1176 * doc/invoke.texi: Document -mcpu option.
1178 2014-01-17 Richard Biener <rguenther@suse.de>
1180 PR rtl-optimization/38518
1181 * df.h (df_analyze_loop): Declare.
1182 * df-core.c: Include cfgloop.h.
1183 (df_analyze_1): Split out main part of df_analyze.
1184 (df_analyze): Adjust.
1185 (loop_inverted_post_order_compute): New function.
1186 (loop_post_order_compute): Likewise.
1187 (df_analyze_loop): New function avoiding whole-function
1189 * loop-invariant.c (find_defs): Use df_analyze_loop.
1190 (find_invariants): Adjust.
1191 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
1193 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
1195 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
1196 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
1198 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
1200 * ipa-ref.c (ipa_remove_stmt_references): Fix references
1201 traversal when removing references.
1203 2014-01-16 Jan Hubicka <jh@suse.cz>
1206 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
1208 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
1211 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
1212 pushing a reload for an autoinc when we had previously reloaded an
1213 inner part of the address.
1215 2014-01-16 Jakub Jelinek <jakub@redhat.com>
1217 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
1219 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
1220 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
1221 when not giving up or versioning for alias only because of
1223 (vect_analyze_data_ref_dependences): Set to true.
1224 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
1226 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
1227 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
1231 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
1234 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
1235 operand 0 predicate for gathers, use a new pseudo as subtarget.
1237 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1240 * lra-constraints.c (process_alt_operands): Add printing debug info.
1241 Check absence of input/output reloads for matched operands too.
1243 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1245 PR rtl-optimization/59835
1246 * ira.c (ira_init_register_move_cost): Increase cost for
1249 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
1251 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
1253 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
1256 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
1257 non-register objects. Use gen_(high/low)part more consistently.
1260 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1263 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
1264 endian support, remove tests for WORDS_BIG_ENDIAN.
1265 (p8_mfvsrd_3_<mode>): Likewise.
1266 (reload_gpr_from_vsx<mode>): Likewise.
1267 (reload_gpr_from_vsxsf): Likewise.
1268 (p8_mfvsrd_4_disf): Likewise.
1270 2014-01-16 Richard Biener <rguenther@suse.de>
1272 PR rtl-optimization/46590
1273 * lcm.c (compute_antinout_edge): Use postorder iteration.
1274 (compute_laterin): Use inverted postorder iteration.
1276 2014-01-16 Nick Clifton <nickc@redhat.com>
1279 * varasm.c (output_constant): Return the number of bytes actually
1281 (output_constructor_array_range): Update the field size with the
1282 number of bytes emitted by output_constant.
1283 (output_constructor_regular_field): Likewise. Also do not
1284 complain if the total number of bytes emitted is now greater
1285 than the expected fieldpos.
1286 * output.h (output_constant): Update prototype and descriptive comment.
1288 2014-01-16 Marek Polacek <polacek@redhat.com>
1291 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
1292 it is error_mark_node.
1294 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
1296 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
1297 VALID_AVX256_REG_OR_OI_MODE.
1299 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
1301 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
1302 current procedure should be profiled.
1304 2014-01-15 Andrew Pinski <apinski@cavium.com>
1306 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
1307 of moving from/to the STACK_REG register class.
1309 2014-01-15 Richard Henderson <rth@redhat.com>
1312 * reginfo.c (global_regs_decl): Globalize.
1313 * rtl.h (global_regs_decl): Declare.
1314 * ira.c (do_reload): Diagnose frame_pointer_needed and it
1315 reserved via global_regs.
1317 2014-01-15 Teresa Johnson <tejohnson@google.com>
1319 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
1321 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1323 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
1324 and vmulosh rather than call gen_vec_widen_smult_*.
1325 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
1326 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
1327 (vec_widen_smult_even_v16qi): Likewise.
1328 (vec_widen_umult_even_v8hi): Likewise.
1329 (vec_widen_smult_even_v8hi): Likewise.
1330 (vec_widen_umult_odd_v16qi): Likewise.
1331 (vec_widen_smult_odd_v16qi): Likewise.
1332 (vec_widen_umult_odd_v8hi): Likewise.
1333 (vec_widen_smult_odd_v8hi): Likewise.
1334 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
1335 vmuloub rather than call gen_vec_widen_umult_*.
1336 (vec_widen_umult_lo_v16qi): Likewise.
1337 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
1338 vmulosb rather than call gen_vec_widen_smult_*.
1339 (vec_widen_smult_lo_v16qi): Likewise.
1340 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
1341 rather than call gen_vec_widen_umult_*.
1342 (vec_widen_umult_lo_v8hi): Likewise.
1343 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
1344 rather than call gen_vec_widen_smult_*.
1345 (vec_widen_smult_lo_v8hi): Likewise.
1347 2014-01-15 Jeff Law <law@redhat.com>
1349 PR tree-optimization/59747
1350 * ree.c (find_and_remove_re): Properly handle case where a second
1351 eliminated extension requires widening a copy created for elimination
1352 of a prior extension.
1353 (combine_set_extension): Ensure that the number of hard regs needed
1354 for a destination register does not change when we widen it.
1356 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
1358 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
1359 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
1360 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
1361 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
1362 (avr-*-rtems*): Likewise.
1363 (bfin*-rtems*): Likewise.
1364 (moxie-*-rtems*): Likewise.
1365 (h8300-*-rtems*): Likewise.
1366 (i[34567]86-*-rtems*): Likewise.
1367 (lm32-*-rtems*): Likewise.
1368 (m32r-*-rtems*): Likewise.
1369 (m68k-*-rtems*): Likewise.
1370 (microblaze*-*-rtems*): Likewise.
1371 (mips*-*-rtems*): Likewise.
1372 (powerpc-*-rtems*): Likewise.
1373 (sh-*-rtems*): Likewise.
1374 (sparc-*-rtems*): Likewise.
1375 (sparc64-*-rtems*): Likewise.
1376 (v850-*-rtems*): Likewise.
1377 (m32c-*-rtems*): Likewise.
1379 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
1381 PR rtl-optimization/59511
1382 * ira.c (ira_init_register_move_cost): Use memory costs for some
1383 cases of register move cost calculations.
1384 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
1385 instead of BB frequency.
1386 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
1387 * lra-assigns.c (find_hard_regno_for): Ditto.
1389 2014-01-15 Richard Biener <rguenther@suse.de>
1391 PR tree-optimization/59822
1392 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
1393 (vectorizable_load): Use it to hoist defs of uses of invariant
1394 loads out of the loop.
1396 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1397 Kugan Vivekanandarajah <kuganv@linaro.org>
1400 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
1403 2014-01-15 Richard Biener <rguenther@suse.de>
1405 PR rtl-optimization/59802
1406 * lcm.c (compute_available): Use inverted postorder to seed
1407 the initial worklist.
1409 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1412 * config/s390/s390.c (s390_preferred_reload_class): Don't return
1413 ADDR_REGS for invalid symrefs in non-PIC code.
1415 2014-01-15 Jakub Jelinek <jakub@redhat.com>
1418 * builtins.c (determine_block_size): Initialize *probable_max_size
1419 even if len_rtx is CONST_INT.
1421 2014-01-14 Andrew Pinski <apinski@cavium.com>
1423 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
1424 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
1425 (cortexa53_tunings): Likewise.
1426 (aarch64_sched_issue_rate): New function.
1427 (TARGET_SCHED_ISSUE_RATE): Define.
1429 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1431 * ira-costs.c (find_costs_and_classes): Add missed
1432 ira_init_register_move_cost_if_necessary.
1434 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1437 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
1439 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
1442 * config/i386/i386.c (type_natural_mode): Add a bool parameter
1443 to indicate if type is used for function return value. Warn ABI
1444 change if the vector mode isn't available for function return value.
1445 (ix86_function_arg_advance): Pass false to type_natural_mode.
1446 (ix86_function_arg): Likewise.
1447 (ix86_gimplify_va_arg): Likewise.
1448 (function_arg_32): Don't warn ABI change.
1449 (ix86_function_value): Pass true to type_natural_mode.
1450 (ix86_return_in_memory): Likewise.
1451 (ix86_struct_value_rtx): Removed.
1452 (TARGET_STRUCT_VALUE_RTX): Likewise.
1454 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1456 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
1457 converting a conditional jump into a conditional return.
1459 2014-01-14 Richard Biener <rguenther@suse.de>
1461 PR tree-optimization/58921
1462 PR tree-optimization/59006
1463 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
1464 hoisting invariant stmts.
1465 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
1466 invariant loads on the preheader edge if possible.
1468 2014-01-14 Joey Ye <joey.ye@arm.com>
1470 * doc/plugin.texi (Building GCC plugins): Update to C++.
1472 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
1474 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
1475 (_mm_rcp28_round_ss): Ditto.
1476 (_mm_rsqrt28_round_sd): Ditto.
1477 (_mm_rsqrt28_round_ss): Ditto.
1478 (_mm_rcp28_sd): Ditto.
1479 (_mm_rcp28_ss): Ditto.
1480 (_mm_rsqrt28_sd): Ditto.
1481 (_mm_rsqrt28_ss): Ditto.
1482 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
1483 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
1484 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
1485 (IX86_BUILTIN_RCP28SD): Ditto.
1486 (IX86_BUILTIN_RCP28SS): Ditto.
1487 (IX86_BUILTIN_RSQRT28SD): Ditto.
1488 (IX86_BUILTIN_RSQRT28SS): Ditto.
1489 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
1490 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
1491 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
1492 (ix86_expand_special_args_builtin): Expand new FTYPE.
1493 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
1494 (srcp14<mode>): Make insn unary.
1495 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
1496 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
1497 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
1498 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
1499 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
1500 Fix rounding: make it SAE only.
1501 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
1503 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
1505 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
1506 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
1507 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
1508 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
1509 (round_saeonly_mask_scalar_operand4): Ditto.
1510 (round_saeonly_mask_scalar_op3): Ditto.
1511 (round_saeonly_mask_scalar_op4): Ditto.
1513 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1515 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
1516 Implement -maltivec=be for vec_insert and vec_extract.
1518 2014-01-10 DJ Delorie <dj@redhat.com>
1520 * config/msp430/msp430.md (call_internal): Don't allow memory
1521 references with SP as the base register.
1522 (call_value_internal): Likewise.
1523 * config/msp430/constraints.md (Yc): New. For memory references
1524 that don't use SP as a base register.
1526 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
1527 "an integer without a # prefix"
1528 * config/msp430/msp430.md (epilogue_helper): Use it.
1530 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1533 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
1534 AVX512F gather builtins.
1535 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
1536 on gather decls with INTEGER_TYPE masktype.
1537 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
1538 directly into the builtin rather than hoisting it before loop.
1540 PR tree-optimization/59387
1541 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
1542 (scev_const_prop): If folded_casts and type has undefined overflow,
1543 use force_gimple_operand instead of force_gimple_operand_gsi and
1544 for each added stmt if it is assign with
1545 arith_code_with_undefined_signed_overflow, call
1546 rewrite_to_defined_overflow.
1547 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
1548 gimple-fold.h instead.
1549 (arith_code_with_undefined_signed_overflow,
1550 rewrite_to_defined_overflow): Moved to ...
1551 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
1552 rewrite_to_defined_overflow): ... here. No longer static.
1553 Include gimplify-me.h.
1554 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
1555 rewrite_to_defined_overflow): New prototypes.
1557 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1559 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
1561 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
1563 * builtins.c (get_object_alignment_2): Minor tweak.
1564 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
1566 2014-01-13 Christian Bruel <christian.bruel@st.com>
1568 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
1569 optimized non constant lengths.
1571 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1574 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
1575 load as __atomic_load_N if possible.
1577 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1579 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
1581 (rs6000_expand_builtin): Adjust call.
1583 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1586 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
1587 * config/rs6000/rs6000.c: Include target-globals.h.
1588 (rs6000_set_current_function): Instead of doing target_reinit
1589 unconditionally, use save_target_globals_default_opts and
1590 restore_target_globals.
1592 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
1594 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
1595 (rs6000_expand_builtin): Handle mffs and mtfsf.
1596 (rs6000_init_builtins): Define mffs and mtfsf.
1597 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
1598 (rs6000_mffs): New pattern.
1599 (rs6000_mtfsf): New pattern.
1601 2014-01-11 Bin Cheng <bin.cheng@arm.com>
1603 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
1604 Start narrowing with START. Apply candidate-use pair
1605 and check overall cost in narrowing.
1606 (iv_ca_prune): Pass new argument.
1608 2014-01-10 Jeff Law <law@redhat.com>
1611 * ree.c (combine_reaching_defs): Ensure the defining statement
1612 occurs before the extension when optimizing extensions with
1613 different source and destination hard registers.
1615 2014-01-10 Jan Hubicka <jh@suse.cz>
1618 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
1619 vtables into the type inheritance graph.
1621 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1623 PR rtl-optimization/59754
1624 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
1625 modes in the REGNO != REGNO case.
1627 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1629 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
1631 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1633 PR tree-optimization/59745
1634 * tree-predcom.c (tree_predictive_commoning_loop): Call
1635 free_affine_expand_cache if giving up because components is NULL.
1637 * target-globals.c (save_target_globals): Allocate < 4KB structs using
1638 GC in payload of target_globals struct instead of allocating them on
1639 the heap and the larger structs separately using GC.
1640 * target-globals.h (struct target_globals): Make regs, hard_regs,
1641 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
1642 of GTY((skip)) and change type to void *.
1643 (reset_target_globals): Cast loads from those fields to corresponding
1646 2014-01-10 Steve Ellcey <sellcey@mips.com>
1649 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
1650 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
1651 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
1653 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
1656 * aarch64-modes.def (CC_Zmode): New flags mode.
1657 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
1658 represents an equality.
1659 (aarch64_get_condition_code): Handle CC_Zmode.
1660 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
1662 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1664 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
1665 extraction in good case.
1667 2014-01-10 Richard Biener <rguenther@suse.de>
1669 PR tree-optimization/59374
1670 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
1671 checking after SLP discovery. Mark stmts not participating
1672 in any SLP instance properly.
1674 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1676 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
1677 when handling a SET rtx.
1679 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1681 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
1682 (cortex-a57): Likewise.
1683 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
1685 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1687 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
1688 non-iwmmxt builtins.
1690 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
1694 * ipa-devirt.c record_target_from_binfo): Take as argument
1695 stack of binfos and lookup matching one for virtual inheritance.
1696 (possible_polymorphic_call_targets_1): Update.
1698 2014-01-10 Huacai Chen <chenhc@lemote.com>
1700 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
1701 kernel strings for Loongson-2E/2F/3A.
1703 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1706 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
1707 is_gimple_call before calling gimple_call_internal_p.
1709 2014-01-09 Steve Ellcey <sellcey@mips.com>
1711 * Makefile.in (TREE_FLOW_H): Remove.
1712 (TREE_SSA_H): Add file names from tree-flow.h.
1713 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
1714 * tree.h: Remove tree-flow.h reference.
1715 * hash-table.h: Remove tree-flow.h reference.
1716 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
1717 reference with tree-ssa-loop.h.
1719 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1721 * doc/invoke.texi: Add -maltivec={be,le} options, and document
1722 default element-order behavior for -maltivec.
1723 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
1724 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
1725 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
1726 when targeting big endian, at least for now.
1727 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
1729 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1732 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
1733 var satisfies use_register_for_decl, just take into account type
1734 alignment, rather than decl alignment.
1736 PR tree-optimization/59622
1737 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
1738 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
1739 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
1740 Don't devirtualize for inplace at all. For targets.length () == 1,
1741 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
1743 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1745 * config/i386/i386.md (cpu): Remove the unused btver1.
1747 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1749 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
1751 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1754 * tree-core.h (struct target_globals): New forward declaration.
1755 (struct tree_target_option): Add globals field.
1756 * tree.h (TREE_TARGET_GLOBALS): Define.
1757 (prepare_target_option_nodes_for_pch): New prototype.
1758 * target-globals.h (struct target_globals): Define even if
1760 * tree.c (prepare_target_option_node_for_pch,
1761 prepare_target_option_nodes_for_pch): New functions.
1762 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
1763 * config/i386/i386.c: Include target-globals.h.
1764 (ix86_set_current_function): Instead of doing target_reinit
1765 unconditionally, use save_target_globals_default_opts and
1766 restore_target_globals.
1768 2014-01-09 Richard Biener <rguenther@suse.de>
1770 PR tree-optimization/59715
1771 * tree-cfg.h (split_critical_edges): Declare.
1772 * tree-cfg.c (split_critical_edges): Export.
1773 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
1775 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
1777 * cfgexpand.c (expand_stack_vars): Optionally disable
1778 asan stack protection.
1779 (expand_used_vars): Likewise.
1780 (partition_stack_vars): Likewise.
1781 * asan.c (asan_emit_stack_protection): Optionally disable
1782 after return stack usage.
1783 (instrument_derefs): Optionally disable memory access instrumentation.
1784 (instrument_builtin_call): Likewise.
1785 (instrument_strlen_call): Likewise.
1786 (asan_protect_global): Optionally disable global variables protection.
1787 * doc/invoke.texi: Added doc for new options.
1788 * params.def: Added new options.
1789 * params.h: Likewise.
1791 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1793 PR rtl-optimization/59724
1794 * ifcvt.c (cond_exec_process_if_block): Don't call
1795 flow_find_head_matching_sequence with 0 longest_match.
1796 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
1797 non-active insns if !stop_after.
1798 (try_head_merge_bb): Revert 2014-01-07 changes.
1800 2014-01-08 Jeff Law <law@redhat.com>
1802 * ree.c (get_sub_rtx): New function, extracted from...
1803 (merge_def_and_ext): Here.
1804 (combine_reaching_defs): Use get_sub_rtx.
1806 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
1808 * cgraph.h (varpool_variable_node): Do not choke on null node.
1810 2014-01-08 Catherine Moore <clm@codesourcery.com>
1812 * config/mips/mips.md (simple_return): Attempt to use JRC
1814 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
1816 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1818 PR rtl-optimization/59137
1819 * reorg.c (steal_delay_list_from_target): Call update_block for
1821 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
1823 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1825 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
1826 two duplicate entries.
1828 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1831 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
1833 * config/mips/mips.c (mips_truncated_op_cost): New function.
1834 (mips_rtx_costs): Adjust test for BADDU.
1835 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
1837 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
1839 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
1840 (*baddu_si): ...this new pattern.
1842 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1845 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
1847 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1850 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
1852 (expand_expr, expand_normal): Adjust.
1853 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
1854 inner_reference_p. Use inner_reference_p to expand inner references.
1855 (store_expr): Adjust.
1856 * cfgexpand.c (expand_call_stmt): Adjust.
1858 2014-01-08 Rong Xu <xur@google.com>
1860 * gcov-io.c (gcov_var): Move from gcov-io.h.
1861 (gcov_position): Ditto.
1862 (gcov_is_error): Ditto.
1863 (gcov_rewrite): Ditto.
1864 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
1865 only part to libgcc/libgcov.h.
1867 2014-01-08 Marek Polacek <polacek@redhat.com>
1870 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
1872 2014-01-08 Marek Polacek <polacek@redhat.com>
1875 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
1877 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1879 PR rtl-optimization/59649
1880 * stor-layout.c (get_mode_bounds): For BImode return
1881 0 and STORE_FLAG_VALUE.
1883 2014-01-08 Richard Biener <rguenther@suse.de>
1886 * gimple.h (is_gimple_builtin_call): Remove.
1887 (gimple_builtin_call_types_compatible_p): New.
1888 (gimple_call_builtin_p): New overload.
1889 * gimple.c (is_gimple_builtin_call): Remove.
1890 (validate_call): Rename to ...
1891 (gimple_builtin_call_types_compatible_p): ... this and export. Also
1893 (validate_type): New static function.
1894 (gimple_call_builtin_p): New overload and adjust.
1895 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
1896 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
1897 (gimple_fold_stmt_to_constant_1): Likewise.
1898 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
1900 2014-01-08 Richard Biener <rguenther@suse.de>
1903 * gimplify.c (gimplify_expr): Gimplify register-register type
1904 VIEW_CONVERT_EXPRs to separate stmts.
1906 2014-01-07 Jeff Law <law@redhat.com>
1909 * ree.c (combine_set_extension): Handle case where source
1910 and destination registers in an extension insn are different.
1911 (combine_reaching_defs): Allow source and destination registers
1912 in extension to be different under limited circumstances.
1913 (add_removable_extension): Remove restriction that the
1914 source and destination registers in the extension are the same.
1915 (find_and_remove_re): Emit a copy from the extension's
1916 destination to its source after the defining insn if
1917 the source and destination registers are different.
1920 * ifcvt.c (merge_if_block): If we are merging a block with more than
1921 one successor with a block with no successors, remove any BARRIER
1922 after the second block.
1924 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
1926 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
1928 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
1931 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
1932 for 14-bit register offsets when INT14_OK_STRICT is false.
1934 2014-01-07 Roland Stigge <stigge@antcom.de>
1935 Michael Meissner <meissner@linux.vnet.ibm.com>
1938 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
1939 Only check TFmode for SPE constants. Don't check TImode or TDmode.
1941 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
1943 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
1946 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
1948 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
1949 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
1950 rtx is const0_rtx or not.
1952 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
1955 * target-globals.c (save_target_globals): Remove this_fn_optab
1957 * toplev.c: Include optabs.h.
1958 (target_reinit): Temporarily restore the global options if another
1959 set of options are in force.
1961 2014-01-07 Jakub Jelinek <jakub@redhat.com>
1963 PR rtl-optimization/58668
1964 * cfgcleanup.c (flow_find_cross_jump): Don't count
1965 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
1966 to determine what is counted.
1967 (flow_find_head_matching_sequence): Use active_insn_p to determine
1969 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
1971 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
1972 determine what is counted.
1974 PR tree-optimization/59643
1975 * tree-predcom.c (split_data_refs_to_components): If one dr is
1976 read and one write, determine_offset fails and the write isn't
1977 in the bad component, just put the read into the bad component.
1979 2014-01-07 Mike Stump <mikestump@comcast.net>
1980 Jakub Jelinek <jakub@redhat.com>
1983 * tree-core.h (struct tree_optimization_option): Change optabs
1984 type from unsigned char * to void *.
1985 * optabs.c (init_tree_optimization_optabs): Adjust
1986 TREE_OPTIMIZATION_OPTABS initialization.
1988 2014-01-06 Jakub Jelinek <jakub@redhat.com>
1991 * config/i386/i386.h (struct machine_function): Add
1992 no_drap_save_restore field.
1993 * config/i386/i386.c (ix86_save_reg): Use
1994 !cfun->machine->no_drap_save_restore instead of
1995 crtl->stack_realign_needed.
1996 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
1997 this function clears frame_pointer_needed. Set
1998 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
1999 and DRAP reg is needed.
2001 2014-01-06 Marek Polacek <polacek@redhat.com>
2004 * doc/implement-c.texi: Mention that other integer types are
2005 permitted as bit-field types in strictly conforming mode.
2007 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
2009 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
2012 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
2014 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
2016 2014-01-06 Martin Jambor <mjambor@suse.cz>
2019 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
2021 * ipa-prop.c (ipa_print_node_params): Fix indentation.
2023 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
2027 * var-tracking.c (add_stores): Preserve the value of the source even if
2028 we don't record the store.
2030 2014-01-06 Terry Guo <terry.guo@arm.com>
2032 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
2034 2014-01-05 Iain Sandoe <iain@codesourcery.com>
2037 * config/darwin.c (darwin_function_section): Adjust return values to
2038 correspond to optimisation changes made in r206070.
2040 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
2042 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
2043 from prefetch_block tune setting.
2044 (nocona_cost): Correct size of prefetch block to 64.
2046 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
2048 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
2049 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
2050 used to save the static chain register in the computation of the offset
2051 from which the FP registers need to be restored.
2053 2014-01-04 Jakub Jelinek <jakub@redhat.com>
2055 PR tree-optimization/59519
2056 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
2057 ICE if get_current_def (current_new_name) is already non-NULL, as long
2058 as it is a phi result of some other phi in *new_exit_bb that has
2061 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
2062 or vmovdqu* for misaligned_operand.
2063 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
2064 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
2065 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
2066 aligned_mem for AVX512F masked aligned load and store builtins and for
2069 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
2071 PR tree-optimization/59651
2072 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
2073 Address range for negative step should be added by TYPE_SIZE_UNIT.
2075 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
2077 * config/m68k/m68k.c (handle_move_double): Handle pushes with
2078 overlapping registers also for registers other than the stack pointer.
2080 2014-01-03 Marek Polacek <polacek@redhat.com>
2083 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
2086 2014-01-03 Jakub Jelinek <jakub@redhat.com>
2089 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
2092 * config/i386/i386.md (MODE_SIZE): New mode attribute.
2093 (push splitter): Use <P:MODE_SIZE> instead of
2094 GET_MODE_SIZE (<P:MODE>mode).
2095 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
2096 (mov -1, reg peephole2): Likewise.
2097 * config/i386/sse.md (*mov<mode>_internal,
2098 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
2099 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
2100 *<code><mode>3, *andnot<mode>3<mask_name>,
2101 <mask_codefor><code><mode>3<mask_name>): Likewise.
2102 * config/i386/subst.md (mask_mode512bit_condition,
2103 sd_mask_mode512bit_condition): Likewise.
2105 2014-01-02 Xinliang David Li <davidxl@google.com>
2107 PR tree-optimization/59303
2108 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
2109 (dump_predicates): Better output format.
2110 (pred_equal_p): New function.
2111 (is_neq_relop_p): Ditto.
2112 (is_neq_zero_form_p): Ditto.
2113 (pred_expr_equal_p): Ditto.
2114 (pred_neg_p): Ditto.
2115 (simplify_pred): Ditto.
2116 (simplify_preds_2): Ditto.
2117 (simplify_preds_3): Ditto.
2118 (simplify_preds_4): Ditto.
2119 (simplify_preds): Ditto.
2121 (push_to_worklist): Ditto.
2122 (get_pred_info_from_cmp): Ditto.
2123 (is_degenerated_phi): Ditto.
2124 (normalize_one_pred_1): Ditto.
2125 (normalize_one_pred): Ditto.
2126 (normalize_one_pred_chain): Ditto.
2127 (normalize_preds): Ditto.
2128 (normalize_cond_1): Remove function.
2129 (normalize_cond): Ditto.
2130 (is_gcond_subset_of): Ditto.
2131 (is_subset_of_any): Ditto.
2132 (is_or_set_subset_of): Ditto.
2133 (is_and_set_subset_of): Ditto.
2134 (is_norm_cond_subset_of): Ditto.
2135 (pred_chain_length_cmp): Ditto.
2136 (convert_control_dep_chain_into_preds): Type change.
2137 (find_predicates): Ditto.
2138 (find_def_preds): Ditto.
2139 (destroy_predicates_vecs): Ditto.
2140 (find_matching_predicates_in_rest_chains): Ditto.
2141 (use_pred_not_overlap_with_undef_path_pred): Ditto.
2142 (is_pred_expr_subset): Ditto.
2143 (is_pred_chain_subset_of): Ditto.
2144 (is_included_in): Ditto.
2145 (is_superset_of): Ditto.
2147 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2149 Update copyright years.
2151 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2153 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
2154 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
2155 config/arc/arc.md, config/arc/arc.opt,
2156 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
2157 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
2158 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
2159 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
2160 config/linux-protos.h, config/linux.c, config/winnt-c.c,
2161 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
2162 vtable-verify.c, vtable-verify.h: Use the standard form for the
2165 2014-01-02 Tobias Burnus <burnus@net-b.de>
2167 * gcc.c (process_command): Update copyright notice dates.
2168 * gcov-dump.c: Ditto.
2170 * doc/cpp.texi: Bump @copying's copyright year.
2171 * doc/cppinternals.texi: Ditto.
2172 * doc/gcc.texi: Ditto.
2173 * doc/gccint.texi: Ditto.
2174 * doc/gcov.texi: Ditto.
2175 * doc/install.texi: Ditto.
2176 * doc/invoke.texi: Ditto.
2178 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2180 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
2182 2014-01-01 Jakub Jelinek <jakub@redhat.com>
2184 * config/i386/sse.md (*mov<mode>_internal): Guard
2185 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
2187 PR rtl-optimization/59647
2188 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
2189 new_rtx into UNSIGNED_FLOAT rtxes.
2191 Copyright (C) 2014 Free Software Foundation, Inc.
2193 Copying and distribution of this file, with or without modification,
2194 are permitted in any medium without royalty provided the copyright
2195 notice and this notice are preserved.