| blob | 5dba715e4b72bd2d509bbe3b77ced901b1f29a4f |
1 2014-01-29 Jan Hubicka <jh@suse.cz>
3 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
4 (estimate_function_body_sizes): Use it.
6 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
8 PR c++/58561
9 * dwarf2out.c (is_cxx_auto): New.
10 (is_base_type): Use it.
11 (gen_type_die_with_usage): Likewise.
13 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
16 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
17 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
18 -maltivec=be with LE targets.
19 (vsx_mergeh_<mode>): Likewise.
20 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New
21 unspecs.
22 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
23 (altivec_vmrghb): Replace with define_expand and new
24 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE
25 targets.
26 (altivec_vmrghb_direct): New define_insn.
27 (altivec_vmrghh): Replace with define_expand and new
28 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE
29 targets.
30 (altivec_vmrghh_direct): New define_insn.
31 (altivec_vmrghw): Replace with define_expand and new
32 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE
33 targets.
34 (altivec_vmrghw_direct): New define_insn.
35 (*altivec_vmrghsf): Adjust for endianness.
36 (altivec_vmrglb): Replace with define_expand and new
37 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE
38 targets.
39 (altivec_vmrglb_direct): New define_insn.
40 (altivec_vmrglh): Replace with define_expand and new
41 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE
42 targets.
43 (altivec_vmrglh_direct): New define_insn.
44 (altivec_vmrglw): Replace with define_expand and new
45 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE
46 targets.
47 (altivec_vmrglw_direct): New define_insn.
48 (*altivec_vmrglsf): Adjust for endianness.
49 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
50 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
51 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
52 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
53 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
54 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
55 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
56 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
58 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
60 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
61 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
62 whitespace.
64 2014-01-29 Richard Biener <rguenther@suse.de>
66 PR tree-optimization/58742
67 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
68 associate_pointerplus_align.
69 (associate_pointerplus_diff): New function.
70 (associate_pointerplus): Likewise. Call associate_pointerplus_align
71 and associate_pointerplus_diff.
73 2014-01-29 Richard Biener <rguenther@suse.de>
75 * lto-streamer.h (LTO_major_version): Bump to 3.
76 (LTO_minor_version): Reset to 0.
78 2014-01-29 Renlin Li <Renlin.Li@arm.com>
80 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
81 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
82 (arm_file_start): Generate correct asm header for armv7ve.
83 * config/arm/bpabi.h: Add multilib support for armv7ve.
84 * config/arm/driver-arm.c: Change the architectures of cortex-a7
85 and cortex-a15 to armv7ve.
86 * config/arm/t-aprofile: Add multilib support for armv7ve.
87 * doc/invoke.texi: Document -march=armv7ve.
89 2014-01-29 Richard Biener <rguenther@suse.de>
91 PR tree-optimization/58742
92 * tree-ssa-forwprop.c (associate_plusminus): Return true
93 if we changed sth, defer EH cleanup to ...
94 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
95 (simplify_mult): New function.
97 2014-01-29 Jakub Jelinek <jakub@redhat.com>
99 PR middle-end/59917
100 PR tree-optimization/59920
101 * tree.c (build_common_builtin_nodes): Remove
102 __builtin_setjmp_dispatcher initialization.
103 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
104 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
105 instead of gsi_after_labels + manually skipping debug stmts.
106 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
107 ignore bbs with IFN_ABNORMAL_DISPATCHER.
108 * tree-inline.c (copy_edges_for_bb): Remove
109 can_make_abnormal_goto argument, instead add abnormal_goto_dest
110 argument. Ignore computed_goto_p stmts. Don't call
111 make_abnormal_goto_edges. If a call might need abnormal edges
112 for non-local gotos, see if it already has an edge to
113 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
114 with true argument, don't do anything then, otherwise add
115 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
116 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
117 caller.
118 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
119 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
120 (lower_stmt): Don't set data->calls_builtin_setjmp.
121 (lower_builtin_setjmp): Adjust comment.
122 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
123 * tree-cfg.c (found_computed_goto): Remove.
124 (factor_computed_gotos): Remove.
125 (make_goto_expr_edges): Return bool, true for computed gotos.
126 Don't call make_abnormal_goto_edges.
127 (build_gimple_cfg): Don't set found_computed_goto, don't call
128 factor_computed_gotos.
129 (computed_goto_p): No longer static.
130 (make_blocks): Don't set found_computed_goto.
131 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
132 (make_edges): If make_goto_expr_edges returns true, push bb
133 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
134 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
135 vector. Record mapping between bbs and OpenMP regions if there
136 are any, adjust make_gimple_omp_edges caller. Call
137 handle_abnormal_edges.
138 (make_abnormal_goto_edges): Remove.
139 * tree-cfg.h (make_abnormal_goto_edges): Remove.
140 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
141 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
142 * builtins.c (expand_builtin): Don't handle
143 BUILT_IN_SETJMP_DISPATCHER.
144 * internal-fn.def (ABNORMAL_DISPATCHER): New.
145 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
146 filling *region also set *region_idx to (*region)->entry->index.
148 PR other/58712
149 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
150 For REGs set ORIGINAL_REGNO.
152 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
154 * doc/md.texi: Mention that a target shouldn't implement
155 vec_widen_(s|u)mul_even/odd pair if it is less efficient
156 than hi/lo pair.
158 2014-01-29 Jakub Jelinek <jakub@redhat.com>
160 PR tree-optimization/59594
161 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
162 a copy of the datarefs vector rather than the vector itself.
164 2014-01-28 Jason Merrill <jason@redhat.com>
166 PR c++/53756
167 * dwarf2out.c (auto_die): New static.
168 (gen_type_die_with_usage): Handle C++1y 'auto'.
169 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
170 on definition.
172 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
174 PR target/59672
175 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
176 (SPEC_X32): Likewise.
177 (SPEC_64): Likewise.
178 * config/i386/i386.c (ix86_option_override_internal): Turn off
179 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
180 for TARGET_16BIT.
181 (x86_file_start): Output .code16gcc for TARGET_16BIT.
182 * config/i386/i386.h (TARGET_16BIT): New macro.
183 (TARGET_16BIT_P): Likewise.
184 * config/i386/i386.opt: Add m16.
185 * doc/invoke.texi: Document -m16.
187 2014-01-28 Jakub Jelinek <jakub@redhat.com>
189 PR preprocessor/59935
190 * input.c (location_get_source_line): Bail out on when line number
191 is zero, and test the return value of
192 lookup_or_add_file_to_cache_tab.
194 2014-01-28 Richard Biener <rguenther@suse.de>
196 PR tree-optimization/58742
197 * tree-ssa-forwprop.c (associate_plusminus): Handle
198 pointer subtraction of the form (T)(P + A) - (T)P.
200 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
202 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
203 at const_int_cost.
205 2014-01-28 Richard Biener <rguenther@suse.de>
207 Revert
208 2014-01-28 Richard Biener <rguenther@suse.de>
210 PR rtl-optimization/45364
211 PR rtl-optimization/59890
212 * var-tracking.c (local_get_addr_clear_given_value): Handle
213 already cleared slot.
214 (val_reset): Handle not allocated local_get_addr_cache.
215 (vt_find_locations): Use post-order on the inverted CFG.
217 2014-01-28 Richard Biener <rguenther@suse.de>
219 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps):
220 Remove.
222 2014-01-28 Richard Biener <rguenther@suse.de>
224 PR rtl-optimization/45364
225 PR rtl-optimization/59890
226 * var-tracking.c (local_get_addr_clear_given_value): Handle
227 already cleared slot.
228 (val_reset): Handle not allocated local_get_addr_cache.
229 (vt_find_locations): Use post-order on the inverted CFG.
231 2014-01-28 Alan Modra <amodra@gmail.com>
233 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
234 * configure.ac <recursive call for build != host>: Define
235 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
236 and LD_FOR_BUILD too.
237 * configure: Regenerate.
239 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
241 * config/i386/i386.c (get_builtin_code_for_version): Separate
242 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
243 Broadwell from Haswell.
245 2014-01-27 Steve Ellcey <sellcey@mips.com>
247 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
248 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
249 * config/mips/mips.c (mips_option_override): Change setting
250 of TARGET_DSP.
251 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
252 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
253 Change from Mask to Var.
255 2014-01-27 Jeff Law <law@redhat.com>
257 * ipa-inline.c (inline_small_functions): Fix typo.
259 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
261 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
262 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
263 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
264 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
265 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
266 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
267 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
268 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
269 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
270 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
271 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
272 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
273 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
274 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
275 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
276 (_mm512_storeu_epi64): Ditto.
277 (_mm512_cmpge_epi32_mask): Ditto.
278 (_mm512_cmpge_epu32_mask): Ditto.
279 (_mm512_cmpge_epi64_mask): Ditto.
280 (_mm512_cmpge_epu64_mask): Ditto.
281 (_mm512_cmple_epi32_mask): Ditto.
282 (_mm512_cmple_epu32_mask): Ditto.
283 (_mm512_cmple_epi64_mask): Ditto.
284 (_mm512_cmple_epu64_mask): Ditto.
285 (_mm512_cmplt_epi32_mask): Ditto.
286 (_mm512_cmplt_epu32_mask): Ditto.
287 (_mm512_cmplt_epi64_mask): Ditto.
288 (_mm512_cmplt_epu64_mask): Ditto.
289 (_mm512_cmpneq_epi32_mask): Ditto.
290 (_mm512_cmpneq_epu32_mask): Ditto.
291 (_mm512_cmpneq_epi64_mask): Ditto.
292 (_mm512_cmpneq_epu64_mask): Ditto.
293 (_mm512_expand_pd): Ditto.
294 (_mm512_expand_ps): Ditto.
295 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
296 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
297 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
298 * config/i386/i386.c (ix86_builtins): Add
299 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
300 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
301 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
302 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
303 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
304 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
305 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
306 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
307 IX86_BUILTIN_PMOVUSQW512_MEM.
308 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
309 __builtin_ia32_pmovsqd512mem_mask,
310 __builtin_ia32_pmovqd512mem_mask,
311 __builtin_ia32_pmovusqw512mem_mask,
312 __builtin_ia32_pmovsqw512mem_mask,
313 __builtin_ia32_pmovqw512mem_mask,
314 __builtin_ia32_pmovusdw512mem_mask,
315 __builtin_ia32_pmovsdw512mem_mask,
316 __builtin_ia32_pmovdw512mem_mask,
317 __builtin_ia32_pmovqb512mem_mask,
318 __builtin_ia32_pmovusqb512mem_mask,
319 __builtin_ia32_pmovsqb512mem_mask,
320 __builtin_ia32_pmovusdb512mem_mask,
321 __builtin_ia32_pmovsdb512mem_mask,
322 __builtin_ia32_pmovdb512mem_mask.
323 (bdesc_args): Add __builtin_ia32_expanddf512,
324 __builtin_ia32_expandsf512.
325 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
326 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
327 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
328 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
329 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
330 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
331 (avx512f_<code>v8div16qi2_mask_store): This.
332 (avx512f_expand<mode>): New.
334 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
336 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
337 New.
338 (_mm512_mask_prefetch_i64gather_pd): Ditto.
339 (_mm512_prefetch_i32scatter_pd): Ditto.
340 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
341 (_mm512_prefetch_i64scatter_pd): Ditto.
342 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
343 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
344 (_mm512_mask_prefetch_i64gather_ps): Ditto.
345 (_mm512_prefetch_i32scatter_ps): Ditto.
346 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
347 (_mm512_prefetch_i64scatter_ps): Ditto.
348 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
349 * config/i386/i386-builtin-types.def: Define
350 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
351 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
352 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
353 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
354 IX86_BUILTIN_SCATTERPFQPD.
355 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
356 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
357 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
358 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
359 __builtin_ia32_scatterpfqps.
360 (ix86_expand_builtin): Expand new built-ins.
361 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
362 fix memory access data type.
363 (*avx512pf_gatherpf<mode>_mask): Ditto.
364 (*avx512pf_gatherpf<mode>): Ditto.
365 (avx512pf_scatterpf<mode>): Ditto.
366 (*avx512pf_scatterpf<mode>_mask): Ditto.
367 (*avx512pf_scatterpf<mode>): Ditto.
368 (GATHER_SCATTER_SF_MEM_MODE): New.
369 (avx512pf_gatherpf<mode>df): Ditto.
370 (*avx512pf_gatherpf<mode>df_mask): Ditto.
371 (*avx512pf_scatterpf<mode>df): Ditto.
373 2014-01-27 Jakub Jelinek <jakub@redhat.com>
375 PR bootstrap/59934
376 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
377 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
378 reached.
380 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
382 * common/config/arm/arm-common.c
383 (arm_rewrite_mcpu): Handle multiple names.
384 * config/arm/arm.h
385 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
387 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
389 * gimple-builder.h (create_gimple_tmp): Delete.
391 2014-01-27 Christian Bruel <christian.bruel@st.com>
393 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
394 words comparisons.
396 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
398 * config/pa/pa.md (call): Generate indirect long calls to non-local
399 functions when outputing 32-bit code.
400 (call_value): Likewise except for special call to buggy powf function.
402 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
403 portable runtime and PIC indirect calls.
404 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
405 and PIC call sequences. Use ldo instead of blr to set return register
406 in PIC call sequence.
408 2014-01-25 Walter Lee <walt@tilera.com>
410 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
411 avoid clobbering a live register.
413 2014-01-25 Walter Lee <walt@tilera.com>
415 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
416 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
417 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
418 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
420 2014-01-25 Walter Lee <walt@tilera.com>
422 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
423 arguments on even registers.
424 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
425 STACK_BOUNDARY.
426 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
427 (BIGGEST_ALIGNMENT): Ditto.
428 (BIGGEST_FIELD_ALIGNMENT): Ditto.
430 2014-01-25 Walter Lee <walt@tilera.com>
432 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
433 insns before bundling.
434 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
436 2014-01-25 Walter Lee <walt@tilera.com>
438 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
439 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
440 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
442 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
444 * config/mips/constraints.md (kl): Delete.
445 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
446 define expands, using...
447 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
448 instructions for MIPS16.
449 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
450 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
452 2014-01-25 Walter Lee <walt@tilera.com>
454 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
455 (clzdi2): Ditto.
456 (ffsdi2): Ditto.
458 2014-01-25 Walter Lee <walt@tilera.com>
460 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
461 (TARGET_EXPAND_TO_RTL_HOOK): Define.
463 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
465 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
466 Handle XOR.
468 2014-01-25 Jakub Jelinek <jakub@redhat.com>
470 * print-rtl.c (in_call_function_usage): New var.
471 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
472 EXPR_LIST mode as mode and not as reg note name.
474 PR middle-end/59561
475 * cfgloopmanip.c (copy_loop_info): If
476 loop->warned_aggressive_loop_optimizations, make sure
477 the flag is set in target loop too.
479 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
481 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
482 flag_cilkplus.
483 * builtins.def: Likewise.
484 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
485 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
486 * ira.c (ira_setup_eliminable_regset): Likewise.
487 * omp-low.c (gate_expand_omp): Likewise.
488 (execute_lower_omp): Likewise.
489 (diagnose_sb_0): Likewise.
490 (gate_diagnose_omp_blocks): Likewise.
491 (simd_clone_clauses_extract): Likewise.
492 (gate): Likewise.
494 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
496 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
497 correction for little endian...
498 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
499 here.
501 2014-01-24 Jeff Law <law@redhat.com>
503 PR tree-optimization/59919
504 * tree-vrp.c (find_assert_locations_1): Do not register asserts
505 for non-returning calls.
507 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
509 * common/config/aarch64/aarch64-common.c
510 (aarch64_rewrite_mcpu): Handle multiple names.
511 * config/aarch64/aarch64.h
512 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
514 2014-01-24 Dodji Seketeli <dodji@redhat.com>
516 * input.c (add_file_to_cache_tab): Handle the case where fopen
517 returns NULL.
519 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
521 PR target/59929
522 * config/i386/i386.md (pushsf splitter): Get stack adjustment
523 from push operand if code of push isn't PRE_DEC.
525 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
527 PR target/59909
528 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
529 -mquad-memory-atomic. Update -mquad-memory documentation to say
530 it is only used for non-atomic loads/stores.
532 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
533 -mquad-memory or -mquad-memory-atomic switches.
535 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
536 -mquad-memory-atomic to ISA 2.07 support.
538 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
539 to separate support of normal quad word memory operations (ldq, stq)
540 from the atomic quad word memory operations.
542 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
543 support to separate non-atomic quad word operations from atomic
544 quad word operations. Disable non-atomic quad word operations in
545 little endian mode so that we don't have to swap words after the
546 load and before the store.
547 (quad_load_store_p): Add comment about atomic quad word support.
548 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
549 options printed with -mdebug=reg.
551 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
552 -mquad-memory-atomic as the test for whether we have quad word
553 atomic instructions.
554 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
555 or -mp8-vector are used, allow byte/half-word atomic operations.
557 * config/rs6000/sync.md (load_lockedti): Insure that the address
558 is a proper indexed or indirect address for the lqarx instruction.
559 On little endian systems, swap the hi/lo registers after the lqarx
560 instruction.
561 (load_lockedpti): Use indexed_or_indirect_operand predicate to
562 insure the address is valid for the lqarx instruction.
563 (store_conditionalti): Insure that the address is a proper indexed
564 or indirect address for the stqcrx. instruction. On little endian
565 systems, swap the hi/lo registers before doing the stqcrx.
566 instruction.
567 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
568 insure the address is valid for the stqcrx. instruction.
570 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
571 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
572 type of quad memory support is available.
574 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
576 PR regression/59915
577 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
578 there is a danger of looping.
580 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
582 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
583 force flag_ira_loop_pressure if set via command line.
585 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
587 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
588 (ashr_simd): New builtin handling DI mode.
589 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
590 (aarch64_sshr_simddi): New match pattern.
591 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
592 (vshrd_n_s64): Likewise.
593 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
595 2014-01-23 Nick Clifton <nickc@redhat.com>
597 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
598 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
599 favour of mcu specific scripts.
600 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
601 430x multilibs.
603 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
604 Alex Velenko <Alex.Velenko@arm.com>
606 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
607 (vaddv_s16): Likewise.
608 (vaddv_s32): Likewise.
609 (vaddv_u8): Likewise.
610 (vaddv_u16): Likewise.
611 (vaddv_u32): Likewise.
612 (vaddvq_s8): Likewise.
613 (vaddvq_s16): Likewise.
614 (vaddvq_s32): Likewise.
615 (vaddvq_s64): Likewise.
616 (vaddvq_u8): Likewise.
617 (vaddvq_u16): Likewise.
618 (vaddvq_u32): Likewise.
619 (vaddvq_u64): Likewise.
620 (vaddv_f32): Likewise.
621 (vaddvq_f32): Likewise.
622 (vaddvq_f64): Likewise.
623 (vmaxv_f32): Likewise.
624 (vmaxv_s8): Likewise.
625 (vmaxv_s16): Likewise.
626 (vmaxv_s32): Likewise.
627 (vmaxv_u8): Likewise.
628 (vmaxv_u16): Likewise.
629 (vmaxv_u32): Likewise.
630 (vmaxvq_f32): Likewise.
631 (vmaxvq_f64): Likewise.
632 (vmaxvq_s8): Likewise.
633 (vmaxvq_s16): Likewise.
634 (vmaxvq_s32): Likewise.
635 (vmaxvq_u8): Likewise.
636 (vmaxvq_u16): Likewise.
637 (vmaxvq_u32): Likewise.
638 (vmaxnmv_f32): Likewise.
639 (vmaxnmvq_f32): Likewise.
640 (vmaxnmvq_f64): Likewise.
641 (vminv_f32): Likewise.
642 (vminv_s8): Likewise.
643 (vminv_s16): Likewise.
644 (vminv_s32): Likewise.
645 (vminv_u8): Likewise.
646 (vminv_u16): Likewise.
647 (vminv_u32): Likewise.
648 (vminvq_f32): Likewise.
649 (vminvq_f64): Likewise.
650 (vminvq_s8): Likewise.
651 (vminvq_s16): Likewise.
652 (vminvq_s32): Likewise.
653 (vminvq_u8): Likewise.
654 (vminvq_u16): Likewise.
655 (vminvq_u32): Likewise.
656 (vminnmv_f32): Likewise.
657 (vminnmvq_f32): Likewise.
658 (vminnmvq_f64): Likewise.
660 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
662 * config/aarch64/aarch64-simd.md
663 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
664 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
665 (*aarch64_mul3_elt<mode>): Likewise.
666 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
667 (*aarch64_mul3_elt_to_64v2df): Likewise.
668 (*aarch64_mla_elt<mode>): Likewise.
669 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
670 (*aarch64_mls_elt<mode>): Likewise.
671 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
672 (*aarch64_fma4_elt<mode>): Likewise.
673 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
674 (*aarch64_fma4_elt_to_64v2df): Likewise.
675 (*aarch64_fnma4_elt<mode>): Likewise.
676 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
677 (*aarch64_fnma4_elt_to_64v2df): Likewise.
678 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
679 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
680 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
681 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
682 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
683 (aarch64_sqdmull_lane<mode>_internal): Likewise.
684 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
686 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
688 * config/aarch64/aarch64-simd.md
689 (aarch64_be_checked_get_lane<mode>): New define_expand.
690 * config/aarch64/aarch64-simd-builtins.def
691 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
692 New builtin definition.
693 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
694 Use new safe be builtin.
696 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
698 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
699 New define_insn.
700 (aarch64_be_st1<mode>): Likewise.
701 (aarch_ld1<VALL:mode>): Define_expand modified.
702 (aarch_st1<VALL:mode>): Likewise.
703 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
704 (UNSPEC_ST1): Likewise.
706 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
708 * config/microblaze/microblaze.md: Add trap insn and attribute
710 2014-01-23 Dodji Seketeli <dodji@redhat.com>
712 PR preprocessor/58580
713 * input.h (location_get_source_line): Take an additional line_size
714 parameter.
715 (void diagnostics_file_cache_fini): Declare new function.
716 * input.c (struct fcache): New type.
717 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
718 New static constants.
719 (diagnostic_file_cache_init, total_lines_num)
720 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
721 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
722 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
723 (get_next_line, read_next_line, goto_next_line, read_line_num):
724 New static function definitions.
725 (diagnostic_file_cache_fini): New function.
726 (location_get_source_line): Take an additional output line_len
727 parameter. Re-write using lookup_or_add_file_to_cache_tab and
728 read_line_num.
729 * diagnostic.c (diagnostic_finish): Call
730 diagnostic_file_cache_fini.
731 (adjust_line): Take an additional input parameter for the length
732 of the line, rather than calculating it with strlen.
733 (diagnostic_show_locus): Adjust the use of
734 location_get_source_line and adjust_line with respect to their new
735 signature. While displaying a line now, do not stop at the first
736 null byte. Rather, display the zero byte as a space and keep
737 going until we reach the size of the line.
738 * Makefile.in: Add vec.o to OBJS-libcommon
740 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
741 Ilya Tocar <ilya.tocar@intel.com>
743 * config/i386/avx512fintrin.h (_mm512_kmov): New.
744 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
745 (__builtin_ia32_kmov16): Ditto.
746 * config/i386/i386.md (UNSPEC_KMOV): New.
747 (kmovw): Ditto.
749 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
751 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
752 (_mm512_storeu_si512): Ditto.
754 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
756 PR target/52125
757 * rtl.h (get_referenced_operands): Declare.
758 * recog.c (get_referenced_operands): New function.
759 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
760 operands have been referenced when recording LO_SUM references.
762 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
764 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
766 2014-01-22 Jan Hubicka <jh@suse.cz>
768 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
769 Enable for generic and recent AMD targets.
771 2014-01-22 Jan Hubicka <jh@suse.cz>
773 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
774 ARG_SIZE note when adjustment was eliminated.
776 2014-01-22 Jeff Law <law@redhat.com>
778 PR tree-optimization/59597
779 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
780 in file. Accept new argument REGISTERING and use it to modify
781 dump output appropriately.
782 (register_jump_thread): Corresponding changes.
783 (mark_threaded_blocks): Reinstate code to cancel unprofitable
784 thread paths involving joiner blocks. Add code to dump cancelled
785 jump threading paths.
787 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
789 PR rtl-optimization/59477
790 * lra-constraints.c (inherit_in_ebb): Process call for living hard
791 regs. Update reloads_num and potential_reload_hard_regs for all insns.
793 2014-01-22 Tom Tromey <tromey@redhat.com>
795 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
796 PARAMS.
797 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
799 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
801 PR rtl-optimization/59896
802 * lra-constraints.c (process_alt_operands): Check unused note for
803 matched operands of insn with no output reloads.
805 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
807 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
808 (mips_move_from_gpr_cost): Likewise.
810 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
812 PR rtl-optimization/59858
813 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
814 ira_class_hard_regs_num.
815 (process_alt_operands): Increase reject for dying matched operand.
817 2014-01-21 Jakub Jelinek <jakub@redhat.com>
819 PR target/59003
820 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
821 smaller than size, perform several stores or loads and stores
822 at dst + count - size to store or copy all of size bytes, rather
823 than just last modesize bytes.
825 2014-01-20 DJ Delorie <dj@redhat.com>
827 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
828 that CLOBBERs are REGs before propogating their values.
830 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
832 PR middle-end/59789
833 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
834 (cgraph_inline_failed_type): New function.
835 * cgraph.h (DEFCIFCODE): Add type.
836 (cgraph_inline_failed_type_t): New enum.
837 (cgraph_inline_failed_type): New prototype.
838 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
839 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
840 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
841 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
842 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
843 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
844 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
845 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
846 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
847 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
848 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
849 OPTIMIZATION_MISMATCH.
850 * tree-inline.c (expand_call_inline): Emit errors during
851 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
853 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
855 PR target/59685
856 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
857 mode attribute in insn output.
859 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
861 * output.h (output_constant): Delete.
862 * varasm.c (output_constant): Make private.
864 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
866 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
868 2014-01-20 Jakub Jelinek <jakub@redhat.com>
870 PR middle-end/59860
871 * tree.h (fold_builtin_strcat): New prototype.
872 * builtins.c (fold_builtin_strcat): No longer static. Add len
873 argument, if non-NULL, don't call c_strlen. Optimize
874 directly into __builtin_memcpy instead of __builtin_strcpy.
875 (fold_builtin_2): Adjust fold_builtin_strcat caller.
876 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
878 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
880 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
881 for SImode_address_operand operands, having only a REG argument.
883 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
885 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
886 loader name using mbig-endian.
887 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
889 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
891 * doc/invoke.texi (-march): Clarify documentation for AArch64.
892 (-mtune): Likewise.
893 (-mcpu): Likewise.
895 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
897 * config/aarch64/aarch64-protos.h
898 (aarch64_cannot_change_mode_class_ptr): Declare.
899 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
900 aarch64_cannot_change_mode_class_ptr): New.
901 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
902 backend hook aarch64_cannot_change_mode_class.
904 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
906 * common/config/aarch64/aarch64-common.c
907 (aarch64_handle_option): Don't handle any option order logic here.
908 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
909 selected_cpu, warn on architecture version mismatch.
910 (aarch64_override_options): Fix parsing order for option strings.
912 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
913 Iain Sandoe <iain@codesourcery.com>
915 PR bootstrap/59496
916 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
917 warning. Amend comment to reflect current functionality.
919 2014-01-20 Richard Biener <rguenther@suse.de>
921 PR middle-end/59860
922 * builtins.c (fold_builtin_strcat): Remove case better handled
923 by tree-ssa-strlen.c.
925 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
927 * config/aarch64/aarch64.opt
928 (mcpu, march, mtune): Make case-insensitive.
930 2014-01-20 Jakub Jelinek <jakub@redhat.com>
932 PR target/59880
933 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
934 if operands[1] is a REG or ZERO_EXTEND of a REG.
936 2014-01-19 Jan Hubicka <jh@suse.cz>
938 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
940 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
942 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
943 long non-pic millicode calls.
945 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
947 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
949 2014-01-19 Kito Cheng <kito@0xlab.org>
951 * builtins.c (expand_movstr): Check movstr expand done or fail.
953 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
954 H.J. Lu <hongjiu.lu@intel.com>
956 PR target/59379
957 * config/i386/i386.md (*lea<mode>): Zero-extend return register
958 to DImode for zero-extended addresses.
960 2014-01-19 Jakub Jelinek <jakub@redhat.com>
962 PR rtl-optimization/57763
963 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
964 on the new indirect jump_insn and increment LABEL_NUSES (label).
966 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
968 PR bootstrap/59580
969 PR bootstrap/59583
970 * config.gcc (x86_archs): New variable.
971 (x86_64_archs): Likewise.
972 (x86_cpus): Likewise.
973 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
974 --with-arch/--with-cpu= options.
975 Support --with-arch=/--with-cpu={nehalem,westmere,
976 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
978 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
980 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
981 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
983 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
985 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
987 2014-01-18 Jakub Jelinek <jakub@redhat.com>
989 PR target/58944
990 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
991 clear cpp_get_options (parse_in)->warn_unused_macros for
992 ix86_target_macros_internal with cpp_define.
994 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
996 * jump.c (delete_related_insns): Keep (use (insn))s.
997 * reorg.c (redundant_insn): Check for barriers too.
999 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1001 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
1003 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
1005 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
1006 call to $$dyncall when TARGET_LONG_CALLS is true.
1008 2014-01-17 Jeff Law <law@redhat.com>
1010 * ree.c (combine_set_extension): Temporarily disable test for
1011 changing number of hard registers.
1013 2014-01-17 Jan Hubicka <jh@suse.cz>
1015 PR middle-end/58125
1016 * ipa-inline-analysis.c (inline_free_summary):
1017 Do not free summary of aliases.
1019 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1021 PR middle-end/59706
1022 * gimplify.c (gimplify_expr): Use create_tmp_var
1023 instead of create_tmp_var_raw. If cond doesn't have
1024 integral type, don't add the IFN_ANNOTATE builtin at all.
1026 2014-01-17 Martin Jambor <mjambor@suse.cz>
1028 PR ipa/59736
1029 * ipa-cp.c (prev_edge_clone): New variable.
1030 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
1031 Also resize prev_edge_clone vector.
1032 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
1033 (ipcp_edge_removal_hook): New function.
1034 (ipcp_driver): Register ipcp_edge_removal_hook.
1036 2014-01-17 Andrew Pinski <apinski@cavium.com>
1037 Steve Ellcey <sellcey@mips.com>
1039 PR target/59462
1040 * config/mips/mips.c (mips_print_operand): Check operand mode instead
1041 of operator mode.
1043 2014-01-17 Jeff Law <law@redhat.com>
1045 PR middle-end/57904
1046 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
1047 so that pass_ccp runs first.
1049 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1051 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
1052 (ix86_adjust_cost): Use !TARGET_XXX.
1053 (do_reorder_for_imul): Likewise.
1054 (swap_top_of_ready_list): Likewise.
1055 (ix86_sched_reorder): Likewise.
1057 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1059 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1060 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
1061 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
1062 (intel_memset): New. Duplicate slm_memset.
1063 (intel_cost): New. Duplicate slm_cost.
1064 (m_INTEL): New macro.
1065 (processor_target_table): Add "intel".
1066 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
1067 with PROCESSOR_INTEL for "intel".
1068 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
1069 PROCESSOR_SILVERMONT.
1070 (ix86_issue_rate): Likewise.
1071 (ix86_adjust_cost): Likewise.
1072 (ia32_multipass_dfa_lookahead): Likewise.
1073 (swap_top_of_ready_list): Likewise.
1074 (ix86_sched_reorder): Likewise.
1075 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
1076 instead of TARGET_OPT_AGU.
1077 * config/i386/i386.h (TARGET_INTEL): New.
1078 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
1079 (processor_type): Add PROCESSOR_INTEL.
1080 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
1081 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
1083 2014-01-17 Marek Polacek <polacek@redhat.com>
1085 PR c/58346
1086 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
1087 size is zero.
1089 2014-01-17 Richard Biener <rguenther@suse.de>
1091 PR tree-optimization/46590
1092 * opts.c (default_options_table): Add entries for
1093 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
1094 all enabled at -O1 but not for -Og.
1095 * common.opt (fbranch-count-reg): Remove Init(1).
1096 (fmove-loop-invariants): Likewise.
1097 (ftree-pta): Likewise.
1099 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1101 * config/i386/i386.c (ix86_data_alignment): For compatibility with
1102 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
1103 decls to at least the GCC 4.8 used alignments.
1105 PR fortran/59440
1106 * tree-nested.c (convert_nonlocal_reference_stmt,
1107 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
1108 of GIMPLE_BIND stmts, adjust associated decls.
1110 2014-01-17 Richard Biener <rguenther@suse.de>
1112 PR tree-optimization/46590
1113 * vec.h (vec<>::bseach): New member function implementing
1114 binary search according to C89 bsearch.
1115 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
1116 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
1117 bitmap pointer again. Make accesses_in_loop a flat array.
1118 (mem_ref_obstack): New global.
1119 (outermost_indep_loop): Adjust for mem_ref->stored changes.
1120 (mark_ref_stored): Likewise.
1121 (ref_indep_loop_p_2): Likewise.
1122 (set_ref_stored_in_loop): New helper function.
1123 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
1124 (memref_free): Adjust.
1125 (record_mem_ref_loc): Simplify.
1126 (gather_mem_refs_stmt): Adjust.
1127 (sort_locs_in_loop_postorder_cmp): New function.
1128 (analyze_memory_references): Sort accesses_in_loop after
1129 loop postorder number.
1130 (find_ref_loc_in_loop_cmp): New function.
1131 (for_all_locs_in_loop): Find relevant cluster of locs in
1132 accesses_in_loop and iterate without recursion.
1133 (execute_sm): Avoid uninit warning.
1134 (struct ref_always_accessed): Simplify.
1135 (ref_always_accessed::operator ()): Likewise.
1136 (ref_always_accessed_p): Likewise.
1137 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
1138 loop postorder numbers here.
1139 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
1140 numbers.
1142 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
1144 PR c++/57945
1145 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
1146 on decls for which assemble_alias has been called.
1148 2014-01-17 Nick Clifton <nickc@redhat.com>
1150 * config/msp430/msp430.opt: (mcpu): New option.
1151 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
1152 (msp430_option_override): Parse target_cpu. If the MCU name
1153 matches a generic string, clear target_mcu.
1154 (msp430_attr): Allow numeric interrupt values up to 63.
1155 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
1156 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
1157 option.
1158 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
1159 Add mcpu matches.
1160 * config/msp430/msp430.md (popm): Use %J rather than %I.
1161 (addsi3): Use msp430_nonimmediate_operand for operand 2.
1162 (addhi_cy_i): Use immediate_operand for operand 2.
1163 * doc/invoke.texi: Document -mcpu option.
1165 2014-01-17 Richard Biener <rguenther@suse.de>
1167 PR rtl-optimization/38518
1168 * df.h (df_analyze_loop): Declare.
1169 * df-core.c: Include cfgloop.h.
1170 (df_analyze_1): Split out main part of df_analyze.
1171 (df_analyze): Adjust.
1172 (loop_inverted_post_order_compute): New function.
1173 (loop_post_order_compute): Likewise.
1174 (df_analyze_loop): New function avoiding whole-function
1175 postorder computes.
1176 * loop-invariant.c (find_defs): Use df_analyze_loop.
1177 (find_invariants): Adjust.
1178 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
1180 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
1182 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
1183 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
1185 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
1187 * ipa-ref.c (ipa_remove_stmt_references): Fix references
1188 traversal when removing references.
1190 2014-01-16 Jan Hubicka <jh@suse.cz>
1192 PR ipa/59775
1193 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
1195 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
1197 PR middle-end/56791
1198 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
1199 pushing a reload for an autoinc when we had previously reloaded an
1200 inner part of the address.
1202 2014-01-16 Jakub Jelinek <jakub@redhat.com>
1204 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
1205 field.
1206 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
1207 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
1208 when not giving up or versioning for alias only because of
1209 loop->safelen.
1210 (vect_analyze_data_ref_dependences): Set to true.
1211 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
1212 is a GIMPLE_PHI.
1213 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
1214 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
1215 to the condition.
1217 PR middle-end/58344
1218 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
1220 PR target/59839
1221 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
1222 operand 0 predicate for gathers, use a new pseudo as subtarget.
1224 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1226 PR middle-end/59609
1227 * lra-constraints.c (process_alt_operands): Add printing debug info.
1228 Check absence of input/output reloads for matched operands too.
1230 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1232 PR rtl-optimization/59835
1233 * ira.c (ira_init_register_move_cost): Increase cost for
1234 impossible modes.
1236 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
1238 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
1240 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
1242 PR target/59780
1243 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
1244 non-register objects. Use gen_(high/low)part more consistently.
1245 Fix assertions.
1247 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1249 PR target/59844
1250 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
1251 endian support, remove tests for WORDS_BIG_ENDIAN.
1252 (p8_mfvsrd_3_<mode>): Likewise.
1253 (reload_gpr_from_vsx<mode>): Likewise.
1254 (reload_gpr_from_vsxsf): Likewise.
1255 (p8_mfvsrd_4_disf): Likewise.
1257 2014-01-16 Richard Biener <rguenther@suse.de>
1259 PR rtl-optimization/46590
1260 * lcm.c (compute_antinout_edge): Use postorder iteration.
1261 (compute_laterin): Use inverted postorder iteration.
1263 2014-01-16 Nick Clifton <nickc@redhat.com>
1265 PR middle-end/28865
1266 * varasm.c (output_constant): Return the number of bytes actually
1267 emitted.
1268 (output_constructor_array_range): Update the field size with the
1269 number of bytes emitted by output_constant.
1270 (output_constructor_regular_field): Likewise. Also do not
1271 complain if the total number of bytes emitted is now greater
1272 than the expected fieldpos.
1273 * output.h (output_constant): Update prototype and descriptive comment.
1275 2014-01-16 Marek Polacek <polacek@redhat.com>
1277 PR middle-end/59827
1278 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
1279 it is error_mark_node.
1281 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
1283 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
1284 VALID_AVX256_REG_OR_OI_MODE.
1286 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
1288 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
1289 current procedure should be profiled.
1291 2014-01-15 Andrew Pinski <apinski@cavium.com>
1293 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
1294 of moving from/to the STACK_REG register class.
1296 2014-01-15 Richard Henderson <rth@redhat.com>
1298 PR debug/54694
1299 * reginfo.c (global_regs_decl): Globalize.
1300 * rtl.h (global_regs_decl): Declare.
1301 * ira.c (do_reload): Diagnose frame_pointer_needed and it
1302 reserved via global_regs.
1304 2014-01-15 Teresa Johnson <tejohnson@google.com>
1306 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
1308 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1310 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
1311 and vmulosh rather than call gen_vec_widen_smult_*.
1312 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
1313 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
1314 (vec_widen_smult_even_v16qi): Likewise.
1315 (vec_widen_umult_even_v8hi): Likewise.
1316 (vec_widen_smult_even_v8hi): Likewise.
1317 (vec_widen_umult_odd_v16qi): Likewise.
1318 (vec_widen_smult_odd_v16qi): Likewise.
1319 (vec_widen_umult_odd_v8hi): Likewise.
1320 (vec_widen_smult_odd_v8hi): Likewise.
1321 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
1322 vmuloub rather than call gen_vec_widen_umult_*.
1323 (vec_widen_umult_lo_v16qi): Likewise.
1324 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
1325 vmulosb rather than call gen_vec_widen_smult_*.
1326 (vec_widen_smult_lo_v16qi): Likewise.
1327 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
1328 rather than call gen_vec_widen_umult_*.
1329 (vec_widen_umult_lo_v8hi): Likewise.
1330 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
1331 rather than call gen_vec_widen_smult_*.
1332 (vec_widen_smult_lo_v8hi): Likewise.
1334 2014-01-15 Jeff Law <law@redhat.com>
1336 PR tree-optimization/59747
1337 * ree.c (find_and_remove_re): Properly handle case where a second
1338 eliminated extension requires widening a copy created for elimination
1339 of a prior extension.
1340 (combine_set_extension): Ensure that the number of hard regs needed
1341 for a destination register does not change when we widen it.
1343 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
1345 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
1346 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
1347 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
1348 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
1349 (avr-*-rtems*): Likewise.
1350 (bfin*-rtems*): Likewise.
1351 (moxie-*-rtems*): Likewise.
1352 (h8300-*-rtems*): Likewise.
1353 (i[34567]86-*-rtems*): Likewise.
1354 (lm32-*-rtems*): Likewise.
1355 (m32r-*-rtems*): Likewise.
1356 (m68k-*-rtems*): Likewise.
1357 (microblaze*-*-rtems*): Likewise.
1358 (mips*-*-rtems*): Likewise.
1359 (powerpc-*-rtems*): Likewise.
1360 (sh-*-rtems*): Likewise.
1361 (sparc-*-rtems*): Likewise.
1362 (sparc64-*-rtems*): Likewise.
1363 (v850-*-rtems*): Likewise.
1364 (m32c-*-rtems*): Likewise.
1366 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
1368 PR rtl-optimization/59511
1369 * ira.c (ira_init_register_move_cost): Use memory costs for some
1370 cases of register move cost calculations.
1371 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
1372 instead of BB frequency.
1373 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
1374 * lra-assigns.c (find_hard_regno_for): Ditto.
1376 2014-01-15 Richard Biener <rguenther@suse.de>
1378 PR tree-optimization/59822
1379 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
1380 (vectorizable_load): Use it to hoist defs of uses of invariant
1381 loads out of the loop.
1383 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1384 Kugan Vivekanandarajah <kuganv@linaro.org>
1386 PR target/59695
1387 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
1388 truncation.
1390 2014-01-15 Richard Biener <rguenther@suse.de>
1392 PR rtl-optimization/59802
1393 * lcm.c (compute_available): Use inverted postorder to seed
1394 the initial worklist.
1396 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1398 PR target/59803
1399 * config/s390/s390.c (s390_preferred_reload_class): Don't return
1400 ADDR_REGS for invalid symrefs in non-PIC code.
1402 2014-01-15 Jakub Jelinek <jakub@redhat.com>
1404 PR other/58712
1405 * builtins.c (determine_block_size): Initialize *probable_max_size
1406 even if len_rtx is CONST_INT.
1408 2014-01-14 Andrew Pinski <apinski@cavium.com>
1410 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
1411 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
1412 (cortexa53_tunings): Likewise.
1413 (aarch64_sched_issue_rate): New function.
1414 (TARGET_SCHED_ISSUE_RATE): Define.
1416 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1418 * ira-costs.c (find_costs_and_classes): Add missed
1419 ira_init_register_move_cost_if_necessary.
1421 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1423 PR target/59787
1424 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
1426 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
1428 PR target/59794
1429 * config/i386/i386.c (type_natural_mode): Add a bool parameter
1430 to indicate if type is used for function return value. Warn ABI
1431 change if the vector mode isn't available for function return value.
1432 (ix86_function_arg_advance): Pass false to type_natural_mode.
1433 (ix86_function_arg): Likewise.
1434 (ix86_gimplify_va_arg): Likewise.
1435 (function_arg_32): Don't warn ABI change.
1436 (ix86_function_value): Pass true to type_natural_mode.
1437 (ix86_return_in_memory): Likewise.
1438 (ix86_struct_value_rtx): Removed.
1439 (TARGET_STRUCT_VALUE_RTX): Likewise.
1441 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1443 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
1444 converting a conditional jump into a conditional return.
1446 2014-01-14 Richard Biener <rguenther@suse.de>
1448 PR tree-optimization/58921
1449 PR tree-optimization/59006
1450 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
1451 hoisting invariant stmts.
1452 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
1453 invariant loads on the preheader edge if possible.
1455 2014-01-14 Joey Ye <joey.ye@arm.com>
1457 * doc/plugin.texi (Building GCC plugins): Update to C++.
1459 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
1461 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
1462 (_mm_rcp28_round_ss): Ditto.
1463 (_mm_rsqrt28_round_sd): Ditto.
1464 (_mm_rsqrt28_round_ss): Ditto.
1465 (_mm_rcp28_sd): Ditto.
1466 (_mm_rcp28_ss): Ditto.
1467 (_mm_rsqrt28_sd): Ditto.
1468 (_mm_rsqrt28_ss): Ditto.
1469 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
1470 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
1471 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
1472 (IX86_BUILTIN_RCP28SD): Ditto.
1473 (IX86_BUILTIN_RCP28SS): Ditto.
1474 (IX86_BUILTIN_RSQRT28SD): Ditto.
1475 (IX86_BUILTIN_RSQRT28SS): Ditto.
1476 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
1477 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
1478 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
1479 (ix86_expand_special_args_builtin): Expand new FTYPE.
1480 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
1481 (srcp14<mode>): Make insn unary.
1482 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
1483 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
1484 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
1485 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
1486 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
1487 Fix rounding: make it SAE only.
1488 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
1489 Ditto.
1490 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
1491 Ditto.
1492 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
1493 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
1494 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
1495 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
1496 (round_saeonly_mask_scalar_operand4): Ditto.
1497 (round_saeonly_mask_scalar_op3): Ditto.
1498 (round_saeonly_mask_scalar_op4): Ditto.
1500 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1502 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
1503 Implement -maltivec=be for vec_insert and vec_extract.
1505 2014-01-10 DJ Delorie <dj@redhat.com>
1507 * config/msp430/msp430.md (call_internal): Don't allow memory
1508 references with SP as the base register.
1509 (call_value_internal): Likewise.
1510 * config/msp430/constraints.md (Yc): New. For memory references
1511 that don't use SP as a base register.
1513 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
1514 "an integer without a # prefix"
1515 * config/msp430/msp430.md (epilogue_helper): Use it.
1517 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1519 PR target/59617
1520 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
1521 AVX512F gather builtins.
1522 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
1523 on gather decls with INTEGER_TYPE masktype.
1524 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
1525 directly into the builtin rather than hoisting it before loop.
1527 PR tree-optimization/59387
1528 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
1529 (scev_const_prop): If folded_casts and type has undefined overflow,
1530 use force_gimple_operand instead of force_gimple_operand_gsi and
1531 for each added stmt if it is assign with
1532 arith_code_with_undefined_signed_overflow, call
1533 rewrite_to_defined_overflow.
1534 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
1535 gimple-fold.h instead.
1536 (arith_code_with_undefined_signed_overflow,
1537 rewrite_to_defined_overflow): Moved to ...
1538 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
1539 rewrite_to_defined_overflow): ... here. No longer static.
1540 Include gimplify-me.h.
1541 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
1542 rewrite_to_defined_overflow): New prototypes.
1544 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1546 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
1548 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
1550 * builtins.c (get_object_alignment_2): Minor tweak.
1551 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
1553 2014-01-13 Christian Bruel <christian.bruel@st.com>
1555 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
1556 optimized non constant lengths.
1558 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1560 PR libgomp/59194
1561 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
1562 load as __atomic_load_N if possible.
1564 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1566 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
1567 target parameter.
1568 (rs6000_expand_builtin): Adjust call.
1570 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1572 PR target/58115
1573 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
1574 * config/rs6000/rs6000.c: Include target-globals.h.
1575 (rs6000_set_current_function): Instead of doing target_reinit
1576 unconditionally, use save_target_globals_default_opts and
1577 restore_target_globals.
1579 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
1580 FPSCR.
1581 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
1582 (rs6000_expand_builtin): Handle mffs and mtfsf.
1583 (rs6000_init_builtins): Define mffs and mtfsf.
1584 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
1585 (rs6000_mffs): New pattern.
1586 (rs6000_mtfsf): New pattern.
1588 2014-01-11 Bin Cheng <bin.cheng@arm.com>
1590 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
1591 Start narrowing with START. Apply candidate-use pair
1592 and check overall cost in narrowing.
1593 (iv_ca_prune): Pass new argument.
1595 2014-01-10 Jeff Law <law@redhat.com>
1597 PR middle-end/59743
1598 * ree.c (combine_reaching_defs): Ensure the defining statement
1599 occurs before the extension when optimizing extensions with
1600 different source and destination hard registers.
1602 2014-01-10 Jan Hubicka <jh@suse.cz>
1604 PR ipa/58585
1605 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
1606 vtables into the type inheritance graph.
1608 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1610 PR rtl-optimization/59754
1611 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
1612 modes in the REGNO != REGNO case.
1614 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1616 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
1618 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1620 PR tree-optimization/59745
1621 * tree-predcom.c (tree_predictive_commoning_loop): Call
1622 free_affine_expand_cache if giving up because components is NULL.
1624 * target-globals.c (save_target_globals): Allocate < 4KB structs using
1625 GC in payload of target_globals struct instead of allocating them on
1626 the heap and the larger structs separately using GC.
1627 * target-globals.h (struct target_globals): Make regs, hard_regs,
1628 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
1629 of GTY((skip)) and change type to void *.
1630 (reset_target_globals): Cast loads from those fields to corresponding
1631 types.
1633 2014-01-10 Steve Ellcey <sellcey@mips.com>
1635 PR plugins/59335
1636 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
1637 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
1638 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
1640 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
1642 PR target/59744
1643 * aarch64-modes.def (CC_Zmode): New flags mode.
1644 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
1645 represents an equality.
1646 (aarch64_get_condition_code): Handle CC_Zmode.
1647 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
1649 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1651 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
1652 extraction in good case.
1654 2014-01-10 Richard Biener <rguenther@suse.de>
1656 PR tree-optimization/59374
1657 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
1658 checking after SLP discovery. Mark stmts not participating
1659 in any SLP instance properly.
1661 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1663 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
1664 when handling a SET rtx.
1666 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1668 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
1669 (cortex-a57): Likewise.
1670 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
1672 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1674 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
1675 non-iwmmxt builtins.
1677 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
1679 PR ipa/58252
1680 PR ipa/59226
1681 * ipa-devirt.c record_target_from_binfo): Take as argument
1682 stack of binfos and lookup matching one for virtual inheritance.
1683 (possible_polymorphic_call_targets_1): Update.
1685 2014-01-10 Huacai Chen <chenhc@lemote.com>
1687 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
1688 kernel strings for Loongson-2E/2F/3A.
1690 2014-01-10 Jakub Jelinek <jakub@redhat.com>
1692 PR middle-end/59670
1693 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
1694 is_gimple_call before calling gimple_call_internal_p.
1696 2014-01-09 Steve Ellcey <sellcey@mips.com>
1698 * Makefile.in (TREE_FLOW_H): Remove.
1699 (TREE_SSA_H): Add file names from tree-flow.h.
1700 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
1701 * tree.h: Remove tree-flow.h reference.
1702 * hash-table.h: Remove tree-flow.h reference.
1703 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
1704 reference with tree-ssa-loop.h.
1706 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1708 * doc/invoke.texi: Add -maltivec={be,le} options, and document
1709 default element-order behavior for -maltivec.
1710 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
1711 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
1712 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
1713 when targeting big endian, at least for now.
1714 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
1716 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1718 PR middle-end/47735
1719 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
1720 var satisfies use_register_for_decl, just take into account type
1721 alignment, rather than decl alignment.
1723 PR tree-optimization/59622
1724 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
1725 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
1726 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
1727 Don't devirtualize for inplace at all. For targets.length () == 1,
1728 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
1730 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1732 * config/i386/i386.md (cpu): Remove the unused btver1.
1734 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
1736 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
1738 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1740 PR target/58115
1741 * tree-core.h (struct target_globals): New forward declaration.
1742 (struct tree_target_option): Add globals field.
1743 * tree.h (TREE_TARGET_GLOBALS): Define.
1744 (prepare_target_option_nodes_for_pch): New prototype.
1745 * target-globals.h (struct target_globals): Define even if
1746 !SWITCHABLE_TARGET.
1747 * tree.c (prepare_target_option_node_for_pch,
1748 prepare_target_option_nodes_for_pch): New functions.
1749 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
1750 * config/i386/i386.c: Include target-globals.h.
1751 (ix86_set_current_function): Instead of doing target_reinit
1752 unconditionally, use save_target_globals_default_opts and
1753 restore_target_globals.
1755 2014-01-09 Richard Biener <rguenther@suse.de>
1757 PR tree-optimization/59715
1758 * tree-cfg.h (split_critical_edges): Declare.
1759 * tree-cfg.c (split_critical_edges): Export.
1760 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
1762 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
1764 * cfgexpand.c (expand_stack_vars): Optionally disable
1765 asan stack protection.
1766 (expand_used_vars): Likewise.
1767 (partition_stack_vars): Likewise.
1768 * asan.c (asan_emit_stack_protection): Optionally disable
1769 after return stack usage.
1770 (instrument_derefs): Optionally disable memory access instrumentation.
1771 (instrument_builtin_call): Likewise.
1772 (instrument_strlen_call): Likewise.
1773 (asan_protect_global): Optionally disable global variables protection.
1774 * doc/invoke.texi: Added doc for new options.
1775 * params.def: Added new options.
1776 * params.h: Likewise.
1778 2014-01-09 Jakub Jelinek <jakub@redhat.com>
1780 PR rtl-optimization/59724
1781 * ifcvt.c (cond_exec_process_if_block): Don't call
1782 flow_find_head_matching_sequence with 0 longest_match.
1783 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
1784 non-active insns if !stop_after.
1785 (try_head_merge_bb): Revert 2014-01-07 changes.
1787 2014-01-08 Jeff Law <law@redhat.com>
1789 * ree.c (get_sub_rtx): New function, extracted from...
1790 (merge_def_and_ext): Here.
1791 (combine_reaching_defs): Use get_sub_rtx.
1793 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
1795 * cgraph.h (varpool_variable_node): Do not choke on null node.
1797 2014-01-08 Catherine Moore <clm@codesourcery.com>
1799 * config/mips/mips.md (simple_return): Attempt to use JRC
1800 for microMIPS.
1801 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
1803 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1805 PR rtl-optimization/59137
1806 * reorg.c (steal_delay_list_from_target): Call update_block for
1807 elided insns.
1808 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
1810 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1812 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
1813 two duplicate entries.
1815 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
1817 Revert:
1818 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
1820 * config/mips/mips.c (mips_truncated_op_cost): New function.
1821 (mips_rtx_costs): Adjust test for BADDU.
1822 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
1824 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
1826 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
1827 (*baddu_si): ...this new pattern.
1829 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1831 PR ipa/59722
1832 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
1834 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1836 PR middle-end/57748
1837 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
1838 inner_reference_p.
1839 (expand_expr, expand_normal): Adjust.
1840 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
1841 inner_reference_p. Use inner_reference_p to expand inner references.
1842 (store_expr): Adjust.
1843 * cfgexpand.c (expand_call_stmt): Adjust.
1845 2014-01-08 Rong Xu <xur@google.com>
1847 * gcov-io.c (gcov_var): Move from gcov-io.h.
1848 (gcov_position): Ditto.
1849 (gcov_is_error): Ditto.
1850 (gcov_rewrite): Ditto.
1851 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
1852 only part to libgcc/libgcov.h.
1854 2014-01-08 Marek Polacek <polacek@redhat.com>
1856 PR middle-end/59669
1857 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
1859 2014-01-08 Marek Polacek <polacek@redhat.com>
1861 PR sanitizer/59667
1862 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
1864 2014-01-08 Jakub Jelinek <jakub@redhat.com>
1866 PR rtl-optimization/59649
1867 * stor-layout.c (get_mode_bounds): For BImode return
1868 0 and STORE_FLAG_VALUE.
1870 2014-01-08 Richard Biener <rguenther@suse.de>
1872 PR middle-end/59630
1873 * gimple.h (is_gimple_builtin_call): Remove.
1874 (gimple_builtin_call_types_compatible_p): New.
1875 (gimple_call_builtin_p): New overload.
1876 * gimple.c (is_gimple_builtin_call): Remove.
1877 (validate_call): Rename to ...
1878 (gimple_builtin_call_types_compatible_p): ... this and export. Also
1879 check return types.
1880 (validate_type): New static function.
1881 (gimple_call_builtin_p): New overload and adjust.
1882 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
1883 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
1884 (gimple_fold_stmt_to_constant_1): Likewise.
1885 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
1887 2014-01-08 Richard Biener <rguenther@suse.de>
1889 PR middle-end/59471
1890 * gimplify.c (gimplify_expr): Gimplify register-register type
1891 VIEW_CONVERT_EXPRs to separate stmts.
1893 2014-01-07 Jeff Law <law@redhat.com>
1895 PR middle-end/53623
1896 * ree.c (combine_set_extension): Handle case where source
1897 and destination registers in an extension insn are different.
1898 (combine_reaching_defs): Allow source and destination registers
1899 in extension to be different under limited circumstances.
1900 (add_removable_extension): Remove restriction that the
1901 source and destination registers in the extension are the same.
1902 (find_and_remove_re): Emit a copy from the extension's
1903 destination to its source after the defining insn if
1904 the source and destination registers are different.
1906 PR middle-end/59285
1907 * ifcvt.c (merge_if_block): If we are merging a block with more than
1908 one successor with a block with no successors, remove any BARRIER
1909 after the second block.
1911 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
1913 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
1915 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
1917 PR target/59652
1918 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
1919 for 14-bit register offsets when INT14_OK_STRICT is false.
1921 2014-01-07 Roland Stigge <stigge@antcom.de>
1922 Michael Meissner <meissner@linux.vnet.ibm.com>
1924 PR 57386/target
1925 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
1926 Only check TFmode for SPE constants. Don't check TImode or TDmode.
1928 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
1930 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
1931 -mcpu.
1933 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
1935 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
1936 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
1937 rtx is const0_rtx or not.
1939 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
1941 PR target/58115
1942 * target-globals.c (save_target_globals): Remove this_fn_optab
1943 handling.
1944 * toplev.c: Include optabs.h.
1945 (target_reinit): Temporarily restore the global options if another
1946 set of options are in force.
1948 2014-01-07 Jakub Jelinek <jakub@redhat.com>
1950 PR rtl-optimization/58668
1951 * cfgcleanup.c (flow_find_cross_jump): Don't count
1952 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
1953 to determine what is counted.
1954 (flow_find_head_matching_sequence): Use active_insn_p to determine
1955 what is counted.
1956 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
1957 counting change.
1958 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
1959 determine what is counted.
1961 PR tree-optimization/59643
1962 * tree-predcom.c (split_data_refs_to_components): If one dr is
1963 read and one write, determine_offset fails and the write isn't
1964 in the bad component, just put the read into the bad component.
1966 2014-01-07 Mike Stump <mikestump@comcast.net>
1967 Jakub Jelinek <jakub@redhat.com>
1969 PR pch/59436
1970 * tree-core.h (struct tree_optimization_option): Change optabs
1971 type from unsigned char * to void *.
1972 * optabs.c (init_tree_optimization_optabs): Adjust
1973 TREE_OPTIMIZATION_OPTABS initialization.
1975 2014-01-06 Jakub Jelinek <jakub@redhat.com>
1977 PR target/59644
1978 * config/i386/i386.h (struct machine_function): Add
1979 no_drap_save_restore field.
1980 * config/i386/i386.c (ix86_save_reg): Use
1981 !cfun->machine->no_drap_save_restore instead of
1982 crtl->stack_realign_needed.
1983 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
1984 this function clears frame_pointer_needed. Set
1985 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
1986 and DRAP reg is needed.
1988 2014-01-06 Marek Polacek <polacek@redhat.com>
1990 PR c/57773
1991 * doc/implement-c.texi: Mention that other integer types are
1992 permitted as bit-field types in strictly conforming mode.
1994 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
1996 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
1997 is newly allocated.
1999 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
2001 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
2003 2014-01-06 Martin Jambor <mjambor@suse.cz>
2005 PR ipa/59008
2006 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
2007 to int.
2008 * ipa-prop.c (ipa_print_node_params): Fix indentation.
2010 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
2012 PR debug/59350
2013 PR debug/59510
2014 * var-tracking.c (add_stores): Preserve the value of the source even if
2015 we don't record the store.
2017 2014-01-06 Terry Guo <terry.guo@arm.com>
2019 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
2021 2014-01-05 Iain Sandoe <iain@codesourcery.com>
2023 PR bootstrap/59541
2024 * config/darwin.c (darwin_function_section): Adjust return values to
2025 correspond to optimisation changes made in r206070.
2027 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
2029 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
2030 from prefetch_block tune setting.
2031 (nocona_cost): Correct size of prefetch block to 64.
2033 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
2035 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
2036 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
2037 used to save the static chain register in the computation of the offset
2038 from which the FP registers need to be restored.
2040 2014-01-04 Jakub Jelinek <jakub@redhat.com>
2042 PR tree-optimization/59519
2043 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
2044 ICE if get_current_def (current_new_name) is already non-NULL, as long
2045 as it is a phi result of some other phi in *new_exit_bb that has
2046 the same argument.
2048 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
2049 or vmovdqu* for misaligned_operand.
2050 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
2051 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
2052 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
2053 aligned_mem for AVX512F masked aligned load and store builtins and for
2054 non-temporal moves.
2056 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
2058 PR tree-optimization/59651
2059 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
2060 Address range for negative step should be added by TYPE_SIZE_UNIT.
2062 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
2064 * config/m68k/m68k.c (handle_move_double): Handle pushes with
2065 overlapping registers also for registers other than the stack pointer.
2067 2014-01-03 Marek Polacek <polacek@redhat.com>
2069 PR other/59661
2070 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
2071 __builtin_FILE.
2073 2014-01-03 Jakub Jelinek <jakub@redhat.com>
2075 PR target/59625
2076 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
2077 asm goto as jump.
2079 * config/i386/i386.md (MODE_SIZE): New mode attribute.
2080 (push splitter): Use <P:MODE_SIZE> instead of
2081 GET_MODE_SIZE (<P:MODE>mode).
2082 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
2083 (mov -1, reg peephole2): Likewise.
2084 * config/i386/sse.md (*mov<mode>_internal,
2085 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
2086 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
2087 *<code><mode>3, *andnot<mode>3<mask_name>,
2088 <mask_codefor><code><mode>3<mask_name>): Likewise.
2089 * config/i386/subst.md (mask_mode512bit_condition,
2090 sd_mask_mode512bit_condition): Likewise.
2092 2014-01-02 Xinliang David Li <davidxl@google.com>
2094 PR tree-optimization/59303
2095 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
2096 (dump_predicates): Better output format.
2097 (pred_equal_p): New function.
2098 (is_neq_relop_p): Ditto.
2099 (is_neq_zero_form_p): Ditto.
2100 (pred_expr_equal_p): Ditto.
2101 (pred_neg_p): Ditto.
2102 (simplify_pred): Ditto.
2103 (simplify_preds_2): Ditto.
2104 (simplify_preds_3): Ditto.
2105 (simplify_preds_4): Ditto.
2106 (simplify_preds): Ditto.
2107 (push_pred): Ditto.
2108 (push_to_worklist): Ditto.
2109 (get_pred_info_from_cmp): Ditto.
2110 (is_degenerated_phi): Ditto.
2111 (normalize_one_pred_1): Ditto.
2112 (normalize_one_pred): Ditto.
2113 (normalize_one_pred_chain): Ditto.
2114 (normalize_preds): Ditto.
2115 (normalize_cond_1): Remove function.
2116 (normalize_cond): Ditto.
2117 (is_gcond_subset_of): Ditto.
2118 (is_subset_of_any): Ditto.
2119 (is_or_set_subset_of): Ditto.
2120 (is_and_set_subset_of): Ditto.
2121 (is_norm_cond_subset_of): Ditto.
2122 (pred_chain_length_cmp): Ditto.
2123 (convert_control_dep_chain_into_preds): Type change.
2124 (find_predicates): Ditto.
2125 (find_def_preds): Ditto.
2126 (destroy_predicates_vecs): Ditto.
2127 (find_matching_predicates_in_rest_chains): Ditto.
2128 (use_pred_not_overlap_with_undef_path_pred): Ditto.
2129 (is_pred_expr_subset): Ditto.
2130 (is_pred_chain_subset_of): Ditto.
2131 (is_included_in): Ditto.
2132 (is_superset_of): Ditto.
2134 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2136 Update copyright years.
2138 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2140 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
2141 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
2142 config/arc/arc.md, config/arc/arc.opt,
2143 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
2144 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
2145 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
2146 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
2147 config/linux-protos.h, config/linux.c, config/winnt-c.c,
2148 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
2149 vtable-verify.c, vtable-verify.h: Use the standard form for the
2150 copyright notice.
2152 2014-01-02 Tobias Burnus <burnus@net-b.de>
2154 * gcc.c (process_command): Update copyright notice dates.
2155 * gcov-dump.c: Ditto.
2156 * gcov.c: Ditto.
2157 * doc/cpp.texi: Bump @copying's copyright year.
2158 * doc/cppinternals.texi: Ditto.
2159 * doc/gcc.texi: Ditto.
2160 * doc/gccint.texi: Ditto.
2161 * doc/gcov.texi: Ditto.
2162 * doc/install.texi: Ditto.
2163 * doc/invoke.texi: Ditto.
2165 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2167 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
2169 2014-01-01 Jakub Jelinek <jakub@redhat.com>
2171 * config/i386/sse.md (*mov<mode>_internal): Guard
2172 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
2174 PR rtl-optimization/59647
2175 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
2176 new_rtx into UNSIGNED_FLOAT rtxes.
2177 \f
2178 Copyright (C) 2014 Free Software Foundation, Inc.
2180 Copying and distribution of this file, with or without modification,
2181 are permitted in any medium without royalty provided the copyright
2182 notice and this notice are preserved.