| blob | f3ebc0cb543173c544a84fc18946f92a8eebb72e |
1 2014-02-03 Cong Hou <congh@google.com>
3 PR tree-optimization/60000
4 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5 if the vectorized statement is a store. A store statement can only
6 appear at the end of pattern statements.
8 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
11 (ix86_option_override_internal): Default long double to 64-bit for
12 32-bit Bionic and to 128-bit for 64-bit Bionic.
14 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
15 TARGET_LONG_DOUBLE_128 is true.
16 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
18 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
19 (mlong-double-64): Negate -mlong-double-128.
20 (mlong-double-128): New option.
22 * config/i386/i386-c.c (ix86_target_macros): Define
23 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
25 * doc/invoke.texi: Document -mlong-double-128.
27 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
29 PR rtl-optimization/60024
30 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
32 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
34 * doc/invoke.texi: (fprofile-reorder-functions): Fix typo.
36 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
38 PR rtl-optimization/57662
39 * sel-sched.c (code_motion_path_driver): Do not mark already not
40 existing blocks in the visiting bitmap.
42 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
44 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
45 on the insn being emitted.
47 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
48 Will Deacon <will.deacon@arm.com>
50 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
52 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
54 * config/arm/arm-tables.opt: Regenerate.
56 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
58 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
59 for vector types other than V16QImode.
60 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
61 define_expand, and call altivec_expand_vec_perm_le when producing
62 code with little endian element order.
63 (*altivec_vperm_<mode>_internal): New insn having previous
64 behavior of altivec_vperm_<mode>.
65 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
66 altivec_expand_vec_perm_le when producing code with little endian
67 element order.
68 (*altivec_vperm_<mode>_uns_internal): New insn having previous
69 behavior of altivec_vperm_<mode>_uns.
71 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
73 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
74 (altivec_vsumsws): Add handling for -maltivec=be with a little
75 endian target.
76 (altivec_vsumsws_direct): New.
77 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
78 gen_altivec_vsumsws.
80 2014-02-02 Jan Hubicka <jh@suse.cz>
82 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
83 vtable_pointer_value_to_binfo): New functions.
84 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
85 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
87 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
89 * config/nios2/nios2.md (load_got_register): Initialize GOT
90 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
91 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
93 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
95 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
96 preserverd by passthrough, do not propagate the type.
98 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
100 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
101 (mips_atomic_assign_expand_fenv): New function.
102 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
104 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
106 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
107 (__builtin_mips_set_fcsr): Likewise.
108 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
109 MIPS_USI_FTYPE_VOID.
110 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
111 (mips16_expand_set_fcsr): Likewise.
112 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
113 (mips16_set_fcsr_stub): Likewise.
114 (mips16_get_fcsr_one_only_stub): New class.
115 (mips16_set_fcsr_one_only_stub): Likewise.
116 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
117 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
118 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
119 (hard_float): New availability predicate.
120 (mips_builtins): Add get_fcsr and set_fcsr.
121 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
122 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
123 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
124 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
125 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
126 patterns.
128 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
130 * config/mips/mips.c (mips_one_only_stub): New class.
131 (mips_need_mips16_rdhwr_p): Replace with...
132 (mips16_rdhwr_stub): ...this new variable.
133 (mips16_stub_call_address): New function.
134 (mips16_rdhwr_one_only_stub): New class.
135 (mips_expand_thread_pointer): Use mips16_stub_call_address.
136 (mips_output_mips16_rdhwr): Delete.
137 (mips_finish_stub): New function.
138 (mips_code_end): Use it to handle rdhwr stubs.
140 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
142 PR target/60017
143 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
144 when calculating size of integer atomic types.
146 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
148 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
150 2014-02-01 Jakub Jelinek <jakub@redhat.com>
152 PR tree-optimization/60003
153 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
154 * profile.c (branch_prob): Use gimple_call_builtin_p
155 to check for BUILT_IN_SETJMP_RECEIVER.
156 * tree-inline.c (copy_bb): Call notice_special_calls.
158 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
160 PR bootstrap/59985
161 * lra-constraints.c (process_alt_operands): Update reload_sum only
162 on the first pass.
164 2014-01-31 Richard Henderson <rth@redhat.com>
166 PR middle-end/60004
167 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
168 until after else_eh is processed.
170 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
172 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
173 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
174 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
175 in smmintrin.h, remove them.
176 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
177 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
178 * config/i386/i386.md (ROUND_SAE): Fix value.
179 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
180 (const48_operand): New.
181 * config/i386/subst.md (round), (round_expand): Use
182 const_4_or_8_to_11_operand.
183 (round_saeonly), (round_saeonly_expand): Use const48_operand.
185 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
187 * config/i386/constraints.md (Yk): Swap meaning with k.
188 * config/i386/i386.md (movhi_internal): Change Yk to k.
189 (movqi_internal): Ditto.
190 (*k<logic><mode>): Ditto.
191 (*andhi_1): Ditto.
192 (*andqi_1): Ditto.
193 (kandn<mode>): Ditto.
194 (*<code>hi_1): Ditto.
195 (*<code>qi_1): Ditto.
196 (kxnor<mode>): Ditto.
197 (kortestzhi): Ditto.
198 (kortestchi): Ditto.
199 (kunpckhi): Ditto.
200 (*one_cmplhi2_1): Ditto.
201 (*one_cmplqi2_1): Ditto.
202 * config/i386/sse.md (): Change k to Yk.
203 (avx512f_load<mode>_mask): Ditto.
204 (avx512f_blendm<mode>): Ditto.
205 (avx512f_store<mode>_mask): Ditto.
206 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
207 (avx512f_storedqu<mode>_mask): Ditto.
208 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
209 Ditto.
210 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
211 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
212 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
213 (avx512f_maskcmp<mode>3): Ditto.
214 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
215 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
216 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
217 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
218 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
219 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
220 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
221 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
222 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
223 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
224 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
225 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
226 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
227 (vec_extract_lo_<mode>_maskm): Ditto.
228 (vec_extract_hi_<mode>_maskm): Ditto.
229 (avx512f_vternlog<mode>_mask): Ditto.
230 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
231 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
232 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
233 (avx512f_<code>v8div16qi2_mask): Ditto.
234 (avx512f_<code>v8div16qi2_mask_store): Ditto.
235 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
236 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
237 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
238 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
239 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
240 (*avx512pf_gatherpf<mode>df_mask): Ditto.
241 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
242 (*avx512pf_scatterpf<mode>df_mask): Ditto.
243 (avx512cd_maskb_vec_dupv8di): Ditto.
244 (avx512cd_maskw_vec_dupv16si): Ditto.
245 (avx512f_vpermi2var<mode>3_maskz): Ditto.
246 (avx512f_vpermi2var<mode>3_mask): Ditto.
247 (avx512f_vpermi2var<mode>3_mask): Ditto.
248 (avx512f_vpermt2var<mode>3_maskz): Ditto.
249 (*avx512f_gathersi<mode>): Ditto.
250 (*avx512f_gathersi<mode>_2): Ditto.
251 (*avx512f_gatherdi<mode>): Ditto.
252 (*avx512f_gatherdi<mode>_2): Ditto.
253 (*avx512f_scattersi<mode>): Ditto.
254 (*avx512f_scatterdi<mode>): Ditto.
255 (avx512f_compress<mode>_mask): Ditto.
256 (avx512f_compressstore<mode>_mask): Ditto.
257 (avx512f_expand<mode>_mask): Ditto.
258 * config/i386/subst.md (mask): Change k to Yk.
259 (mask_scalar_merge): Ditto.
260 (sd): Ditto.
262 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
264 * doc/extend.texi (Vector Extensions): Document ?: in C++.
266 2014-01-31 Richard Biener <rguenther@suse.de>
268 PR middle-end/59990
269 * builtins.c (fold_builtin_memory_op): Make sure to not
270 use a floating-point mode or a boolean or enumeral type for
271 the copy operation.
273 2014-01-30 DJ Delorie <dj@redhat.com>
275 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
276 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
277 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
278 whenever main() has an epilogue.
280 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
282 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
283 unused variable "field".
284 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
285 (vsx_mergeh_<mode>): Likewise.
286 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
287 (altivec_vmrghh): Likewise.
288 (altivec_vmrghw): Likewise.
289 (altivec_vmrglb): Likewise.
290 (altivec_vmrglh): Likewise.
291 (altivec_vmrglw): Likewise.
292 (altivec_vspltb): Add missing uses.
293 (altivec_vsplth): Likewise.
294 (altivec_vspltw): Likewise.
295 (altivec_vspltsf): Likewise.
297 2014-01-30 Jakub Jelinek <jakub@redhat.com>
299 PR target/59923
300 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
301 frame related instructions.
303 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
305 PR rtl-optimization/59959
306 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
307 any reload of register whose subreg is invalid.
309 2014-01-30 Jakub Jelinek <jakub@redhat.com>
311 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
312 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
313 Add missing return type - void.
315 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
317 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
318 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
319 remove element index adjustment for endian (now handled in vsx.md
320 and altivec.md).
321 (altivec_expand_vec_perm_const): Use
322 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
323 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
324 (vsx_xxspltw_<mode>): Adjust element index for little endian.
325 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
326 define_expand and a new define_insn *altivec_vspltb_internal;
327 adjust for -maltivec=be on a little endian target.
328 (altivec_vspltb_direct): New.
329 (altivec_vsplth): Divide into a define_expand and a new
330 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
331 little endian target.
332 (altivec_vsplth_direct): New.
333 (altivec_vspltw): Divide into a define_expand and a new
334 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
335 little endian target.
336 (altivec_vspltw_direct): New.
337 (altivec_vspltsf): Divide into a define_expand and a new
338 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
339 a little endian target.
341 2014-01-30 Richard Biener <rguenther@suse.de>
343 PR tree-optimization/59993
344 * tree-ssa-forwprop.c (associate_pointerplus): Check we
345 can propagate form the earlier stmt and avoid the transform
346 when the intermediate result is needed.
348 2014-01-30 Alangi Derick <alangiderick@gmail.com>
350 * README.Portability: Fix typo.
352 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
354 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
355 comparison_operator with ordered_comparison_operator.
357 2014-01-30 Nick Clifton <nickc@redhat.com>
359 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
360 Rename to mn10300_store_multiple_regs.
361 * config/mn10300/mn10300.c: Likewise.
362 * config/mn10300/mn10300.md (store_movm): Fix typo: call
363 store_multiple_regs.
364 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
365 Call mn10300_store_multiple_regs.
367 2014-01-30 Nick Clifton <nickc@redhat.com>
368 DJ Delorie <dj@redhat.com>
370 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
371 %fp 2 to keep registers after it properly word-aligned.
372 (rl78_alloc_physical_registers_umul): Handle the case where both
373 input operands are the same.
375 2014-01-30 Richard Biener <rguenther@suse.de>
377 PR tree-optimization/59903
378 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
379 check properly.
381 2014-01-30 Jason Merrill <jason@redhat.com>
383 PR c++/59633
384 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
386 PR c++/59645
387 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
389 2014-01-30 Richard Biener <rguenther@suse.de>
391 PR tree-optimization/59951
392 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
394 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
396 PR target/59784
397 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
398 SFmode to DFmode case.
400 2014-01-29 DJ Delorie <dj@redhat.com>
402 * config/msp430/msp430.opt (-minrt): New.
403 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
404 if -minrt given.
405 (ENDFILE_SPEC): Likewise.
407 2014-01-29 Jan Hubicka <jh@suse.cz>
409 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
410 (estimate_function_body_sizes): Use it.
412 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
414 PR c++/58561
415 * dwarf2out.c (is_cxx_auto): New.
416 (is_base_type): Use it.
417 (gen_type_die_with_usage): Likewise.
419 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
421 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
422 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
423 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
424 -maltivec=be with LE targets.
425 (vsx_mergeh_<mode>): Likewise.
426 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
427 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
428 (altivec_vmrghb): Replace with define_expand and new
429 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
430 (altivec_vmrghb_direct): New define_insn.
431 (altivec_vmrghh): Replace with define_expand and new
432 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
433 (altivec_vmrghh_direct): New define_insn.
434 (altivec_vmrghw): Replace with define_expand and new
435 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
436 (altivec_vmrghw_direct): New define_insn.
437 (*altivec_vmrghsf): Adjust for endianness.
438 (altivec_vmrglb): Replace with define_expand and new
439 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
440 (altivec_vmrglb_direct): New define_insn.
441 (altivec_vmrglh): Replace with define_expand and new
442 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
443 (altivec_vmrglh_direct): New define_insn.
444 (altivec_vmrglw): Replace with define_expand and new
445 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
446 (altivec_vmrglw_direct): New define_insn.
447 (*altivec_vmrglsf): Adjust for endianness.
448 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
449 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
450 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
451 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
452 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
453 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
454 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
455 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
457 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
459 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
460 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
461 whitespace.
463 2014-01-29 Richard Biener <rguenther@suse.de>
465 PR tree-optimization/58742
466 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
467 associate_pointerplus_align.
468 (associate_pointerplus_diff): New function.
469 (associate_pointerplus): Likewise. Call associate_pointerplus_align
470 and associate_pointerplus_diff.
472 2014-01-29 Richard Biener <rguenther@suse.de>
474 * lto-streamer.h (LTO_major_version): Bump to 3.
475 (LTO_minor_version): Reset to 0.
477 2014-01-29 Renlin Li <Renlin.Li@arm.com>
479 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
480 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
481 (arm_file_start): Generate correct asm header for armv7ve.
482 * config/arm/bpabi.h: Add multilib support for armv7ve.
483 * config/arm/driver-arm.c: Change the architectures of cortex-a7
484 and cortex-a15 to armv7ve.
485 * config/arm/t-aprofile: Add multilib support for armv7ve.
486 * doc/invoke.texi: Document -march=armv7ve.
488 2014-01-29 Richard Biener <rguenther@suse.de>
490 PR tree-optimization/58742
491 * tree-ssa-forwprop.c (associate_plusminus): Return true
492 if we changed sth, defer EH cleanup to ...
493 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
494 (simplify_mult): New function.
496 2014-01-29 Jakub Jelinek <jakub@redhat.com>
498 PR middle-end/59917
499 PR tree-optimization/59920
500 * tree.c (build_common_builtin_nodes): Remove
501 __builtin_setjmp_dispatcher initialization.
502 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
503 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
504 instead of gsi_after_labels + manually skipping debug stmts.
505 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
506 ignore bbs with IFN_ABNORMAL_DISPATCHER.
507 * tree-inline.c (copy_edges_for_bb): Remove
508 can_make_abnormal_goto argument, instead add abnormal_goto_dest
509 argument. Ignore computed_goto_p stmts. Don't call
510 make_abnormal_goto_edges. If a call might need abnormal edges
511 for non-local gotos, see if it already has an edge to
512 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
513 with true argument, don't do anything then, otherwise add
514 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
515 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
516 caller.
517 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
518 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
519 (lower_stmt): Don't set data->calls_builtin_setjmp.
520 (lower_builtin_setjmp): Adjust comment.
521 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
522 * tree-cfg.c (found_computed_goto): Remove.
523 (factor_computed_gotos): Remove.
524 (make_goto_expr_edges): Return bool, true for computed gotos.
525 Don't call make_abnormal_goto_edges.
526 (build_gimple_cfg): Don't set found_computed_goto, don't call
527 factor_computed_gotos.
528 (computed_goto_p): No longer static.
529 (make_blocks): Don't set found_computed_goto.
530 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
531 (make_edges): If make_goto_expr_edges returns true, push bb
532 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
533 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
534 vector. Record mapping between bbs and OpenMP regions if there
535 are any, adjust make_gimple_omp_edges caller. Call
536 handle_abnormal_edges.
537 (make_abnormal_goto_edges): Remove.
538 * tree-cfg.h (make_abnormal_goto_edges): Remove.
539 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
540 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
541 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
542 * internal-fn.def (ABNORMAL_DISPATCHER): New.
543 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
544 filling *region also set *region_idx to (*region)->entry->index.
546 PR other/58712
547 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
548 For REGs set ORIGINAL_REGNO.
550 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
552 * doc/md.texi: Mention that a target shouldn't implement
553 vec_widen_(s|u)mul_even/odd pair if it is less efficient
554 than hi/lo pair.
556 2014-01-29 Jakub Jelinek <jakub@redhat.com>
558 PR tree-optimization/59594
559 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
560 a copy of the datarefs vector rather than the vector itself.
562 2014-01-28 Jason Merrill <jason@redhat.com>
564 PR c++/53756
565 * dwarf2out.c (auto_die): New static.
566 (gen_type_die_with_usage): Handle C++1y 'auto'.
567 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
568 on definition.
570 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
572 PR target/59672
573 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
574 (SPEC_X32): Likewise.
575 (SPEC_64): Likewise.
576 * config/i386/i386.c (ix86_option_override_internal): Turn off
577 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
578 for TARGET_16BIT.
579 (x86_file_start): Output .code16gcc for TARGET_16BIT.
580 * config/i386/i386.h (TARGET_16BIT): New macro.
581 (TARGET_16BIT_P): Likewise.
582 * config/i386/i386.opt: Add m16.
583 * doc/invoke.texi: Document -m16.
585 2014-01-28 Jakub Jelinek <jakub@redhat.com>
587 PR preprocessor/59935
588 * input.c (location_get_source_line): Bail out on when line number
589 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
591 2014-01-28 Richard Biener <rguenther@suse.de>
593 PR tree-optimization/58742
594 * tree-ssa-forwprop.c (associate_plusminus): Handle
595 pointer subtraction of the form (T)(P + A) - (T)P.
597 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
599 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
600 at const_int_cost.
602 2014-01-28 Richard Biener <rguenther@suse.de>
604 Revert
605 2014-01-28 Richard Biener <rguenther@suse.de>
607 PR rtl-optimization/45364
608 PR rtl-optimization/59890
609 * var-tracking.c (local_get_addr_clear_given_value): Handle
610 already cleared slot.
611 (val_reset): Handle not allocated local_get_addr_cache.
612 (vt_find_locations): Use post-order on the inverted CFG.
614 2014-01-28 Richard Biener <rguenther@suse.de>
616 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
618 2014-01-28 Richard Biener <rguenther@suse.de>
620 PR rtl-optimization/45364
621 PR rtl-optimization/59890
622 * var-tracking.c (local_get_addr_clear_given_value): Handle
623 already cleared slot.
624 (val_reset): Handle not allocated local_get_addr_cache.
625 (vt_find_locations): Use post-order on the inverted CFG.
627 2014-01-28 Alan Modra <amodra@gmail.com>
629 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
630 * configure.ac <recursive call for build != host>: Define
631 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
632 and LD_FOR_BUILD too.
633 * configure: Regenerate.
635 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
637 * config/i386/i386.c (get_builtin_code_for_version): Separate
638 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
639 Broadwell from Haswell.
641 2014-01-27 Steve Ellcey <sellcey@mips.com>
643 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
644 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
645 * config/mips/mips.c (mips_option_override): Change setting
646 of TARGET_DSP.
647 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
648 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
649 Change from Mask to Var.
651 2014-01-27 Jeff Law <law@redhat.com>
653 * ipa-inline.c (inline_small_functions): Fix typo.
655 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
657 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
658 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
659 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
660 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
661 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
662 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
663 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
664 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
665 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
666 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
667 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
668 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
669 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
670 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
671 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
672 (_mm512_storeu_epi64): Ditto.
673 (_mm512_cmpge_epi32_mask): Ditto.
674 (_mm512_cmpge_epu32_mask): Ditto.
675 (_mm512_cmpge_epi64_mask): Ditto.
676 (_mm512_cmpge_epu64_mask): Ditto.
677 (_mm512_cmple_epi32_mask): Ditto.
678 (_mm512_cmple_epu32_mask): Ditto.
679 (_mm512_cmple_epi64_mask): Ditto.
680 (_mm512_cmple_epu64_mask): Ditto.
681 (_mm512_cmplt_epi32_mask): Ditto.
682 (_mm512_cmplt_epu32_mask): Ditto.
683 (_mm512_cmplt_epi64_mask): Ditto.
684 (_mm512_cmplt_epu64_mask): Ditto.
685 (_mm512_cmpneq_epi32_mask): Ditto.
686 (_mm512_cmpneq_epu32_mask): Ditto.
687 (_mm512_cmpneq_epi64_mask): Ditto.
688 (_mm512_cmpneq_epu64_mask): Ditto.
689 (_mm512_expand_pd): Ditto.
690 (_mm512_expand_ps): Ditto.
691 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
692 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
693 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
694 * config/i386/i386.c (ix86_builtins): Add
695 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
696 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
697 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
698 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
699 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
700 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
701 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
702 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
703 IX86_BUILTIN_PMOVUSQW512_MEM.
704 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
705 __builtin_ia32_pmovsqd512mem_mask,
706 __builtin_ia32_pmovqd512mem_mask,
707 __builtin_ia32_pmovusqw512mem_mask,
708 __builtin_ia32_pmovsqw512mem_mask,
709 __builtin_ia32_pmovqw512mem_mask,
710 __builtin_ia32_pmovusdw512mem_mask,
711 __builtin_ia32_pmovsdw512mem_mask,
712 __builtin_ia32_pmovdw512mem_mask,
713 __builtin_ia32_pmovqb512mem_mask,
714 __builtin_ia32_pmovusqb512mem_mask,
715 __builtin_ia32_pmovsqb512mem_mask,
716 __builtin_ia32_pmovusdb512mem_mask,
717 __builtin_ia32_pmovsdb512mem_mask,
718 __builtin_ia32_pmovdb512mem_mask.
719 (bdesc_args): Add __builtin_ia32_expanddf512,
720 __builtin_ia32_expandsf512.
721 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
722 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
723 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
724 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
725 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
726 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
727 (avx512f_<code>v8div16qi2_mask_store): This.
728 (avx512f_expand<mode>): New.
730 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
732 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
733 New.
734 (_mm512_mask_prefetch_i64gather_pd): Ditto.
735 (_mm512_prefetch_i32scatter_pd): Ditto.
736 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
737 (_mm512_prefetch_i64scatter_pd): Ditto.
738 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
739 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
740 (_mm512_mask_prefetch_i64gather_ps): Ditto.
741 (_mm512_prefetch_i32scatter_ps): Ditto.
742 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
743 (_mm512_prefetch_i64scatter_ps): Ditto.
744 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
745 * config/i386/i386-builtin-types.def: Define
746 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
747 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
748 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
749 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
750 IX86_BUILTIN_SCATTERPFQPD.
751 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
752 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
753 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
754 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
755 __builtin_ia32_scatterpfqps.
756 (ix86_expand_builtin): Expand new built-ins.
757 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
758 fix memory access data type.
759 (*avx512pf_gatherpf<mode>_mask): Ditto.
760 (*avx512pf_gatherpf<mode>): Ditto.
761 (avx512pf_scatterpf<mode>): Ditto.
762 (*avx512pf_scatterpf<mode>_mask): Ditto.
763 (*avx512pf_scatterpf<mode>): Ditto.
764 (GATHER_SCATTER_SF_MEM_MODE): New.
765 (avx512pf_gatherpf<mode>df): Ditto.
766 (*avx512pf_gatherpf<mode>df_mask): Ditto.
767 (*avx512pf_scatterpf<mode>df): Ditto.
769 2014-01-27 Jakub Jelinek <jakub@redhat.com>
771 PR bootstrap/59934
772 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
773 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
774 reached.
776 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
778 * common/config/arm/arm-common.c
779 (arm_rewrite_mcpu): Handle multiple names.
780 * config/arm/arm.h
781 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
783 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
785 * gimple-builder.h (create_gimple_tmp): Delete.
787 2014-01-27 Christian Bruel <christian.bruel@st.com>
789 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
790 words comparisons.
792 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
794 * config/pa/pa.md (call): Generate indirect long calls to non-local
795 functions when outputing 32-bit code.
796 (call_value): Likewise except for special call to buggy powf function.
798 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
799 portable runtime and PIC indirect calls.
800 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
801 and PIC call sequences. Use ldo instead of blr to set return register
802 in PIC call sequence.
804 2014-01-25 Walter Lee <walt@tilera.com>
806 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
807 avoid clobbering a live register.
809 2014-01-25 Walter Lee <walt@tilera.com>
811 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
812 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
813 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
814 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
816 2014-01-25 Walter Lee <walt@tilera.com>
818 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
819 arguments on even registers.
820 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
821 STACK_BOUNDARY.
822 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
823 (BIGGEST_ALIGNMENT): Ditto.
824 (BIGGEST_FIELD_ALIGNMENT): Ditto.
826 2014-01-25 Walter Lee <walt@tilera.com>
828 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
829 insns before bundling.
830 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
832 2014-01-25 Walter Lee <walt@tilera.com>
834 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
835 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
836 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
838 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
840 * config/mips/constraints.md (kl): Delete.
841 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
842 define expands, using...
843 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
844 instructions for MIPS16.
845 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
846 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
848 2014-01-25 Walter Lee <walt@tilera.com>
850 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
851 (clzdi2): Ditto.
852 (ffsdi2): Ditto.
854 2014-01-25 Walter Lee <walt@tilera.com>
856 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
857 (TARGET_EXPAND_TO_RTL_HOOK): Define.
859 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
861 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
862 Handle XOR.
864 2014-01-25 Jakub Jelinek <jakub@redhat.com>
866 * print-rtl.c (in_call_function_usage): New var.
867 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
868 EXPR_LIST mode as mode and not as reg note name.
870 PR middle-end/59561
871 * cfgloopmanip.c (copy_loop_info): If
872 loop->warned_aggressive_loop_optimizations, make sure
873 the flag is set in target loop too.
875 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
877 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
878 flag_cilkplus.
879 * builtins.def: Likewise.
880 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
881 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
882 * ira.c (ira_setup_eliminable_regset): Likewise.
883 * omp-low.c (gate_expand_omp): Likewise.
884 (execute_lower_omp): Likewise.
885 (diagnose_sb_0): Likewise.
886 (gate_diagnose_omp_blocks): Likewise.
887 (simd_clone_clauses_extract): Likewise.
888 (gate): Likewise.
890 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
892 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
893 correction for little endian...
894 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
895 here.
897 2014-01-24 Jeff Law <law@redhat.com>
899 PR tree-optimization/59919
900 * tree-vrp.c (find_assert_locations_1): Do not register asserts
901 for non-returning calls.
903 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
905 * common/config/aarch64/aarch64-common.c
906 (aarch64_rewrite_mcpu): Handle multiple names.
907 * config/aarch64/aarch64.h
908 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
910 2014-01-24 Dodji Seketeli <dodji@redhat.com>
912 * input.c (add_file_to_cache_tab): Handle the case where fopen
913 returns NULL.
915 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
917 PR target/59929
918 * config/i386/i386.md (pushsf splitter): Get stack adjustment
919 from push operand if code of push isn't PRE_DEC.
921 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
923 PR target/59909
924 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
925 -mquad-memory-atomic. Update -mquad-memory documentation to say
926 it is only used for non-atomic loads/stores.
928 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
929 -mquad-memory or -mquad-memory-atomic switches.
931 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
932 -mquad-memory-atomic to ISA 2.07 support.
934 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
935 to separate support of normal quad word memory operations (ldq, stq)
936 from the atomic quad word memory operations.
938 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
939 support to separate non-atomic quad word operations from atomic
940 quad word operations. Disable non-atomic quad word operations in
941 little endian mode so that we don't have to swap words after the
942 load and before the store.
943 (quad_load_store_p): Add comment about atomic quad word support.
944 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
945 options printed with -mdebug=reg.
947 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
948 -mquad-memory-atomic as the test for whether we have quad word
949 atomic instructions.
950 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
951 or -mp8-vector are used, allow byte/half-word atomic operations.
953 * config/rs6000/sync.md (load_lockedti): Insure that the address
954 is a proper indexed or indirect address for the lqarx instruction.
955 On little endian systems, swap the hi/lo registers after the lqarx
956 instruction.
957 (load_lockedpti): Use indexed_or_indirect_operand predicate to
958 insure the address is valid for the lqarx instruction.
959 (store_conditionalti): Insure that the address is a proper indexed
960 or indirect address for the stqcrx. instruction. On little endian
961 systems, swap the hi/lo registers before doing the stqcrx.
962 instruction.
963 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
964 insure the address is valid for the stqcrx. instruction.
966 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
967 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
968 type of quad memory support is available.
970 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
972 PR regression/59915
973 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
974 there is a danger of looping.
976 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
978 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
979 force flag_ira_loop_pressure if set via command line.
981 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
983 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
984 (ashr_simd): New builtin handling DI mode.
985 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
986 (aarch64_sshr_simddi): New match pattern.
987 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
988 (vshrd_n_s64): Likewise.
989 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
991 2014-01-23 Nick Clifton <nickc@redhat.com>
993 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
994 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
995 favour of mcu specific scripts.
996 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
997 430x multilibs.
999 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1000 Alex Velenko <Alex.Velenko@arm.com>
1002 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
1003 (vaddv_s16): Likewise.
1004 (vaddv_s32): Likewise.
1005 (vaddv_u8): Likewise.
1006 (vaddv_u16): Likewise.
1007 (vaddv_u32): Likewise.
1008 (vaddvq_s8): Likewise.
1009 (vaddvq_s16): Likewise.
1010 (vaddvq_s32): Likewise.
1011 (vaddvq_s64): Likewise.
1012 (vaddvq_u8): Likewise.
1013 (vaddvq_u16): Likewise.
1014 (vaddvq_u32): Likewise.
1015 (vaddvq_u64): Likewise.
1016 (vaddv_f32): Likewise.
1017 (vaddvq_f32): Likewise.
1018 (vaddvq_f64): Likewise.
1019 (vmaxv_f32): Likewise.
1020 (vmaxv_s8): Likewise.
1021 (vmaxv_s16): Likewise.
1022 (vmaxv_s32): Likewise.
1023 (vmaxv_u8): Likewise.
1024 (vmaxv_u16): Likewise.
1025 (vmaxv_u32): Likewise.
1026 (vmaxvq_f32): Likewise.
1027 (vmaxvq_f64): Likewise.
1028 (vmaxvq_s8): Likewise.
1029 (vmaxvq_s16): Likewise.
1030 (vmaxvq_s32): Likewise.
1031 (vmaxvq_u8): Likewise.
1032 (vmaxvq_u16): Likewise.
1033 (vmaxvq_u32): Likewise.
1034 (vmaxnmv_f32): Likewise.
1035 (vmaxnmvq_f32): Likewise.
1036 (vmaxnmvq_f64): Likewise.
1037 (vminv_f32): Likewise.
1038 (vminv_s8): Likewise.
1039 (vminv_s16): Likewise.
1040 (vminv_s32): Likewise.
1041 (vminv_u8): Likewise.
1042 (vminv_u16): Likewise.
1043 (vminv_u32): Likewise.
1044 (vminvq_f32): Likewise.
1045 (vminvq_f64): Likewise.
1046 (vminvq_s8): Likewise.
1047 (vminvq_s16): Likewise.
1048 (vminvq_s32): Likewise.
1049 (vminvq_u8): Likewise.
1050 (vminvq_u16): Likewise.
1051 (vminvq_u32): Likewise.
1052 (vminnmv_f32): Likewise.
1053 (vminnmvq_f32): Likewise.
1054 (vminnmvq_f64): Likewise.
1056 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
1058 * config/aarch64/aarch64-simd.md
1059 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
1060 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
1061 (*aarch64_mul3_elt<mode>): Likewise.
1062 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
1063 (*aarch64_mul3_elt_to_64v2df): Likewise.
1064 (*aarch64_mla_elt<mode>): Likewise.
1065 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1066 (*aarch64_mls_elt<mode>): Likewise.
1067 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1068 (*aarch64_fma4_elt<mode>): Likewise.
1069 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1070 (*aarch64_fma4_elt_to_64v2df): Likewise.
1071 (*aarch64_fnma4_elt<mode>): Likewise.
1072 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1073 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1074 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1075 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1076 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1077 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1078 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1079 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1080 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1082 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
1084 * config/aarch64/aarch64-simd.md
1085 (aarch64_be_checked_get_lane<mode>): New define_expand.
1086 * config/aarch64/aarch64-simd-builtins.def
1087 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
1088 New builtin definition.
1089 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
1090 Use new safe be builtin.
1092 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
1094 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
1095 New define_insn.
1096 (aarch64_be_st1<mode>): Likewise.
1097 (aarch_ld1<VALL:mode>): Define_expand modified.
1098 (aarch_st1<VALL:mode>): Likewise.
1099 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
1100 (UNSPEC_ST1): Likewise.
1102 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
1104 * config/microblaze/microblaze.md: Add trap insn and attribute
1106 2014-01-23 Dodji Seketeli <dodji@redhat.com>
1108 PR preprocessor/58580
1109 * input.h (location_get_source_line): Take an additional line_size
1110 parameter.
1111 (void diagnostics_file_cache_fini): Declare new function.
1112 * input.c (struct fcache): New type.
1113 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
1114 New static constants.
1115 (diagnostic_file_cache_init, total_lines_num)
1116 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
1117 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
1118 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
1119 (get_next_line, read_next_line, goto_next_line, read_line_num):
1120 New static function definitions.
1121 (diagnostic_file_cache_fini): New function.
1122 (location_get_source_line): Take an additional output line_len
1123 parameter. Re-write using lookup_or_add_file_to_cache_tab and
1124 read_line_num.
1125 * diagnostic.c (diagnostic_finish): Call
1126 diagnostic_file_cache_fini.
1127 (adjust_line): Take an additional input parameter for the length
1128 of the line, rather than calculating it with strlen.
1129 (diagnostic_show_locus): Adjust the use of
1130 location_get_source_line and adjust_line with respect to their new
1131 signature. While displaying a line now, do not stop at the first
1132 null byte. Rather, display the zero byte as a space and keep
1133 going until we reach the size of the line.
1134 * Makefile.in: Add vec.o to OBJS-libcommon
1136 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
1137 Ilya Tocar <ilya.tocar@intel.com>
1139 * config/i386/avx512fintrin.h (_mm512_kmov): New.
1140 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
1141 (__builtin_ia32_kmov16): Ditto.
1142 * config/i386/i386.md (UNSPEC_KMOV): New.
1143 (kmovw): Ditto.
1145 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
1147 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
1148 (_mm512_storeu_si512): Ditto.
1150 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
1152 PR target/52125
1153 * rtl.h (get_referenced_operands): Declare.
1154 * recog.c (get_referenced_operands): New function.
1155 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
1156 operands have been referenced when recording LO_SUM references.
1158 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
1160 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
1162 2014-01-22 Jan Hubicka <jh@suse.cz>
1164 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
1165 Enable for generic and recent AMD targets.
1167 2014-01-22 Jan Hubicka <jh@suse.cz>
1169 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
1170 ARG_SIZE note when adjustment was eliminated.
1172 2014-01-22 Jeff Law <law@redhat.com>
1174 PR tree-optimization/59597
1175 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
1176 in file. Accept new argument REGISTERING and use it to modify
1177 dump output appropriately.
1178 (register_jump_thread): Corresponding changes.
1179 (mark_threaded_blocks): Reinstate code to cancel unprofitable
1180 thread paths involving joiner blocks. Add code to dump cancelled
1181 jump threading paths.
1183 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
1185 PR rtl-optimization/59477
1186 * lra-constraints.c (inherit_in_ebb): Process call for living hard
1187 regs. Update reloads_num and potential_reload_hard_regs for all insns.
1189 2014-01-22 Tom Tromey <tromey@redhat.com>
1191 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
1192 PARAMS.
1193 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
1195 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
1197 PR rtl-optimization/59896
1198 * lra-constraints.c (process_alt_operands): Check unused note for
1199 matched operands of insn with no output reloads.
1201 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
1203 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
1204 (mips_move_from_gpr_cost): Likewise.
1206 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
1208 PR rtl-optimization/59858
1209 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
1210 ira_class_hard_regs_num.
1211 (process_alt_operands): Increase reject for dying matched operand.
1213 2014-01-21 Jakub Jelinek <jakub@redhat.com>
1215 PR target/59003
1216 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
1217 smaller than size, perform several stores or loads and stores
1218 at dst + count - size to store or copy all of size bytes, rather
1219 than just last modesize bytes.
1221 2014-01-20 DJ Delorie <dj@redhat.com>
1223 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
1224 that CLOBBERs are REGs before propogating their values.
1226 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
1228 PR middle-end/59789
1229 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
1230 (cgraph_inline_failed_type): New function.
1231 * cgraph.h (DEFCIFCODE): Add type.
1232 (cgraph_inline_failed_type_t): New enum.
1233 (cgraph_inline_failed_type): New prototype.
1234 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
1235 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
1236 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
1237 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
1238 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
1239 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
1240 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
1241 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
1242 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
1243 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
1244 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
1245 OPTIMIZATION_MISMATCH.
1246 * tree-inline.c (expand_call_inline): Emit errors during
1247 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
1249 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
1251 PR target/59685
1252 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
1253 mode attribute in insn output.
1255 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
1257 * output.h (output_constant): Delete.
1258 * varasm.c (output_constant): Make private.
1260 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
1262 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
1264 2014-01-20 Jakub Jelinek <jakub@redhat.com>
1266 PR middle-end/59860
1267 * tree.h (fold_builtin_strcat): New prototype.
1268 * builtins.c (fold_builtin_strcat): No longer static. Add len
1269 argument, if non-NULL, don't call c_strlen. Optimize
1270 directly into __builtin_memcpy instead of __builtin_strcpy.
1271 (fold_builtin_2): Adjust fold_builtin_strcat caller.
1272 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
1274 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
1276 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
1277 for SImode_address_operand operands, having only a REG argument.
1279 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
1281 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
1282 loader name using mbig-endian.
1283 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
1285 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
1287 * doc/invoke.texi (-march): Clarify documentation for AArch64.
1288 (-mtune): Likewise.
1289 (-mcpu): Likewise.
1291 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
1293 * config/aarch64/aarch64-protos.h
1294 (aarch64_cannot_change_mode_class_ptr): Declare.
1295 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
1296 aarch64_cannot_change_mode_class_ptr): New.
1297 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
1298 backend hook aarch64_cannot_change_mode_class.
1300 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
1302 * common/config/aarch64/aarch64-common.c
1303 (aarch64_handle_option): Don't handle any option order logic here.
1304 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
1305 selected_cpu, warn on architecture version mismatch.
1306 (aarch64_override_options): Fix parsing order for option strings.
1308 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1309 Iain Sandoe <iain@codesourcery.com>
1311 PR bootstrap/59496
1312 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
1313 warning. Amend comment to reflect current functionality.
1315 2014-01-20 Richard Biener <rguenther@suse.de>
1317 PR middle-end/59860
1318 * builtins.c (fold_builtin_strcat): Remove case better handled
1319 by tree-ssa-strlen.c.
1321 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
1323 * config/aarch64/aarch64.opt
1324 (mcpu, march, mtune): Make case-insensitive.
1326 2014-01-20 Jakub Jelinek <jakub@redhat.com>
1328 PR target/59880
1329 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
1330 if operands[1] is a REG or ZERO_EXTEND of a REG.
1332 2014-01-19 Jan Hubicka <jh@suse.cz>
1334 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
1336 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
1338 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
1339 long non-pic millicode calls.
1341 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1343 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
1345 2014-01-19 Kito Cheng <kito@0xlab.org>
1347 * builtins.c (expand_movstr): Check movstr expand done or fail.
1349 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
1350 H.J. Lu <hongjiu.lu@intel.com>
1352 PR target/59379
1353 * config/i386/i386.md (*lea<mode>): Zero-extend return register
1354 to DImode for zero-extended addresses.
1356 2014-01-19 Jakub Jelinek <jakub@redhat.com>
1358 PR rtl-optimization/57763
1359 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
1360 on the new indirect jump_insn and increment LABEL_NUSES (label).
1362 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
1364 PR bootstrap/59580
1365 PR bootstrap/59583
1366 * config.gcc (x86_archs): New variable.
1367 (x86_64_archs): Likewise.
1368 (x86_cpus): Likewise.
1369 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
1370 --with-arch/--with-cpu= options.
1371 Support --with-arch=/--with-cpu={nehalem,westmere,
1372 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
1374 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
1376 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
1377 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
1379 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
1381 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
1383 2014-01-18 Jakub Jelinek <jakub@redhat.com>
1385 PR target/58944
1386 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
1387 clear cpp_get_options (parse_in)->warn_unused_macros for
1388 ix86_target_macros_internal with cpp_define.
1390 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
1392 * jump.c (delete_related_insns): Keep (use (insn))s.
1393 * reorg.c (redundant_insn): Check for barriers too.
1395 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1397 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
1399 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
1401 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
1402 call to $$dyncall when TARGET_LONG_CALLS is true.
1404 2014-01-17 Jeff Law <law@redhat.com>
1406 * ree.c (combine_set_extension): Temporarily disable test for
1407 changing number of hard registers.
1409 2014-01-17 Jan Hubicka <jh@suse.cz>
1411 PR middle-end/58125
1412 * ipa-inline-analysis.c (inline_free_summary):
1413 Do not free summary of aliases.
1415 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1417 PR middle-end/59706
1418 * gimplify.c (gimplify_expr): Use create_tmp_var
1419 instead of create_tmp_var_raw. If cond doesn't have
1420 integral type, don't add the IFN_ANNOTATE builtin at all.
1422 2014-01-17 Martin Jambor <mjambor@suse.cz>
1424 PR ipa/59736
1425 * ipa-cp.c (prev_edge_clone): New variable.
1426 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
1427 Also resize prev_edge_clone vector.
1428 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
1429 (ipcp_edge_removal_hook): New function.
1430 (ipcp_driver): Register ipcp_edge_removal_hook.
1432 2014-01-17 Andrew Pinski <apinski@cavium.com>
1433 Steve Ellcey <sellcey@mips.com>
1435 PR target/59462
1436 * config/mips/mips.c (mips_print_operand): Check operand mode instead
1437 of operator mode.
1439 2014-01-17 Jeff Law <law@redhat.com>
1441 PR middle-end/57904
1442 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
1443 so that pass_ccp runs first.
1445 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1447 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
1448 (ix86_adjust_cost): Use !TARGET_XXX.
1449 (do_reorder_for_imul): Likewise.
1450 (swap_top_of_ready_list): Likewise.
1451 (ix86_sched_reorder): Likewise.
1453 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
1455 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1456 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
1457 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
1458 (intel_memset): New. Duplicate slm_memset.
1459 (intel_cost): New. Duplicate slm_cost.
1460 (m_INTEL): New macro.
1461 (processor_target_table): Add "intel".
1462 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
1463 with PROCESSOR_INTEL for "intel".
1464 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
1465 PROCESSOR_SILVERMONT.
1466 (ix86_issue_rate): Likewise.
1467 (ix86_adjust_cost): Likewise.
1468 (ia32_multipass_dfa_lookahead): Likewise.
1469 (swap_top_of_ready_list): Likewise.
1470 (ix86_sched_reorder): Likewise.
1471 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
1472 instead of TARGET_OPT_AGU.
1473 * config/i386/i386.h (TARGET_INTEL): New.
1474 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
1475 (processor_type): Add PROCESSOR_INTEL.
1476 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
1477 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
1479 2014-01-17 Marek Polacek <polacek@redhat.com>
1481 PR c/58346
1482 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
1483 size is zero.
1485 2014-01-17 Richard Biener <rguenther@suse.de>
1487 PR tree-optimization/46590
1488 * opts.c (default_options_table): Add entries for
1489 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
1490 all enabled at -O1 but not for -Og.
1491 * common.opt (fbranch-count-reg): Remove Init(1).
1492 (fmove-loop-invariants): Likewise.
1493 (ftree-pta): Likewise.
1495 2014-01-17 Jakub Jelinek <jakub@redhat.com>
1497 * config/i386/i386.c (ix86_data_alignment): For compatibility with
1498 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
1499 decls to at least the GCC 4.8 used alignments.
1501 PR fortran/59440
1502 * tree-nested.c (convert_nonlocal_reference_stmt,
1503 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
1504 of GIMPLE_BIND stmts, adjust associated decls.
1506 2014-01-17 Richard Biener <rguenther@suse.de>
1508 PR tree-optimization/46590
1509 * vec.h (vec<>::bseach): New member function implementing
1510 binary search according to C89 bsearch.
1511 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
1512 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
1513 bitmap pointer again. Make accesses_in_loop a flat array.
1514 (mem_ref_obstack): New global.
1515 (outermost_indep_loop): Adjust for mem_ref->stored changes.
1516 (mark_ref_stored): Likewise.
1517 (ref_indep_loop_p_2): Likewise.
1518 (set_ref_stored_in_loop): New helper function.
1519 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
1520 (memref_free): Adjust.
1521 (record_mem_ref_loc): Simplify.
1522 (gather_mem_refs_stmt): Adjust.
1523 (sort_locs_in_loop_postorder_cmp): New function.
1524 (analyze_memory_references): Sort accesses_in_loop after
1525 loop postorder number.
1526 (find_ref_loc_in_loop_cmp): New function.
1527 (for_all_locs_in_loop): Find relevant cluster of locs in
1528 accesses_in_loop and iterate without recursion.
1529 (execute_sm): Avoid uninit warning.
1530 (struct ref_always_accessed): Simplify.
1531 (ref_always_accessed::operator ()): Likewise.
1532 (ref_always_accessed_p): Likewise.
1533 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
1534 loop postorder numbers here.
1535 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
1536 numbers.
1538 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
1540 PR c++/57945
1541 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
1542 on decls for which assemble_alias has been called.
1544 2014-01-17 Nick Clifton <nickc@redhat.com>
1546 * config/msp430/msp430.opt: (mcpu): New option.
1547 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
1548 (msp430_option_override): Parse target_cpu. If the MCU name
1549 matches a generic string, clear target_mcu.
1550 (msp430_attr): Allow numeric interrupt values up to 63.
1551 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
1552 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
1553 option.
1554 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
1555 Add mcpu matches.
1556 * config/msp430/msp430.md (popm): Use %J rather than %I.
1557 (addsi3): Use msp430_nonimmediate_operand for operand 2.
1558 (addhi_cy_i): Use immediate_operand for operand 2.
1559 * doc/invoke.texi: Document -mcpu option.
1561 2014-01-17 Richard Biener <rguenther@suse.de>
1563 PR rtl-optimization/38518
1564 * df.h (df_analyze_loop): Declare.
1565 * df-core.c: Include cfgloop.h.
1566 (df_analyze_1): Split out main part of df_analyze.
1567 (df_analyze): Adjust.
1568 (loop_inverted_post_order_compute): New function.
1569 (loop_post_order_compute): Likewise.
1570 (df_analyze_loop): New function avoiding whole-function
1571 postorder computes.
1572 * loop-invariant.c (find_defs): Use df_analyze_loop.
1573 (find_invariants): Adjust.
1574 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
1576 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
1578 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
1579 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
1581 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
1583 * ipa-ref.c (ipa_remove_stmt_references): Fix references
1584 traversal when removing references.
1586 2014-01-16 Jan Hubicka <jh@suse.cz>
1588 PR ipa/59775
1589 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
1591 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
1593 PR middle-end/56791
1594 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
1595 pushing a reload for an autoinc when we had previously reloaded an
1596 inner part of the address.
1598 2014-01-16 Jakub Jelinek <jakub@redhat.com>
1600 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
1601 field.
1602 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
1603 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
1604 when not giving up or versioning for alias only because of
1605 loop->safelen.
1606 (vect_analyze_data_ref_dependences): Set to true.
1607 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
1608 is a GIMPLE_PHI.
1609 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
1610 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
1611 to the condition.
1613 PR middle-end/58344
1614 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
1616 PR target/59839
1617 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
1618 operand 0 predicate for gathers, use a new pseudo as subtarget.
1620 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1622 PR middle-end/59609
1623 * lra-constraints.c (process_alt_operands): Add printing debug info.
1624 Check absence of input/output reloads for matched operands too.
1626 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
1628 PR rtl-optimization/59835
1629 * ira.c (ira_init_register_move_cost): Increase cost for
1630 impossible modes.
1632 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
1634 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
1636 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
1638 PR target/59780
1639 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
1640 non-register objects. Use gen_(high/low)part more consistently.
1641 Fix assertions.
1643 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1645 PR target/59844
1646 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
1647 endian support, remove tests for WORDS_BIG_ENDIAN.
1648 (p8_mfvsrd_3_<mode>): Likewise.
1649 (reload_gpr_from_vsx<mode>): Likewise.
1650 (reload_gpr_from_vsxsf): Likewise.
1651 (p8_mfvsrd_4_disf): Likewise.
1653 2014-01-16 Richard Biener <rguenther@suse.de>
1655 PR rtl-optimization/46590
1656 * lcm.c (compute_antinout_edge): Use postorder iteration.
1657 (compute_laterin): Use inverted postorder iteration.
1659 2014-01-16 Nick Clifton <nickc@redhat.com>
1661 PR middle-end/28865
1662 * varasm.c (output_constant): Return the number of bytes actually
1663 emitted.
1664 (output_constructor_array_range): Update the field size with the
1665 number of bytes emitted by output_constant.
1666 (output_constructor_regular_field): Likewise. Also do not
1667 complain if the total number of bytes emitted is now greater
1668 than the expected fieldpos.
1669 * output.h (output_constant): Update prototype and descriptive comment.
1671 2014-01-16 Marek Polacek <polacek@redhat.com>
1673 PR middle-end/59827
1674 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
1675 it is error_mark_node.
1677 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
1679 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
1680 VALID_AVX256_REG_OR_OI_MODE.
1682 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
1684 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
1685 current procedure should be profiled.
1687 2014-01-15 Andrew Pinski <apinski@cavium.com>
1689 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
1690 of moving from/to the STACK_REG register class.
1692 2014-01-15 Richard Henderson <rth@redhat.com>
1694 PR debug/54694
1695 * reginfo.c (global_regs_decl): Globalize.
1696 * rtl.h (global_regs_decl): Declare.
1697 * ira.c (do_reload): Diagnose frame_pointer_needed and it
1698 reserved via global_regs.
1700 2014-01-15 Teresa Johnson <tejohnson@google.com>
1702 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
1704 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1706 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
1707 and vmulosh rather than call gen_vec_widen_smult_*.
1708 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
1709 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
1710 (vec_widen_smult_even_v16qi): Likewise.
1711 (vec_widen_umult_even_v8hi): Likewise.
1712 (vec_widen_smult_even_v8hi): Likewise.
1713 (vec_widen_umult_odd_v16qi): Likewise.
1714 (vec_widen_smult_odd_v16qi): Likewise.
1715 (vec_widen_umult_odd_v8hi): Likewise.
1716 (vec_widen_smult_odd_v8hi): Likewise.
1717 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
1718 vmuloub rather than call gen_vec_widen_umult_*.
1719 (vec_widen_umult_lo_v16qi): Likewise.
1720 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
1721 vmulosb rather than call gen_vec_widen_smult_*.
1722 (vec_widen_smult_lo_v16qi): Likewise.
1723 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
1724 rather than call gen_vec_widen_umult_*.
1725 (vec_widen_umult_lo_v8hi): Likewise.
1726 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
1727 rather than call gen_vec_widen_smult_*.
1728 (vec_widen_smult_lo_v8hi): Likewise.
1730 2014-01-15 Jeff Law <law@redhat.com>
1732 PR tree-optimization/59747
1733 * ree.c (find_and_remove_re): Properly handle case where a second
1734 eliminated extension requires widening a copy created for elimination
1735 of a prior extension.
1736 (combine_set_extension): Ensure that the number of hard regs needed
1737 for a destination register does not change when we widen it.
1739 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
1741 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
1742 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
1743 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
1744 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
1745 (avr-*-rtems*): Likewise.
1746 (bfin*-rtems*): Likewise.
1747 (moxie-*-rtems*): Likewise.
1748 (h8300-*-rtems*): Likewise.
1749 (i[34567]86-*-rtems*): Likewise.
1750 (lm32-*-rtems*): Likewise.
1751 (m32r-*-rtems*): Likewise.
1752 (m68k-*-rtems*): Likewise.
1753 (microblaze*-*-rtems*): Likewise.
1754 (mips*-*-rtems*): Likewise.
1755 (powerpc-*-rtems*): Likewise.
1756 (sh-*-rtems*): Likewise.
1757 (sparc-*-rtems*): Likewise.
1758 (sparc64-*-rtems*): Likewise.
1759 (v850-*-rtems*): Likewise.
1760 (m32c-*-rtems*): Likewise.
1762 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
1764 PR rtl-optimization/59511
1765 * ira.c (ira_init_register_move_cost): Use memory costs for some
1766 cases of register move cost calculations.
1767 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
1768 instead of BB frequency.
1769 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
1770 * lra-assigns.c (find_hard_regno_for): Ditto.
1772 2014-01-15 Richard Biener <rguenther@suse.de>
1774 PR tree-optimization/59822
1775 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
1776 (vectorizable_load): Use it to hoist defs of uses of invariant
1777 loads out of the loop.
1779 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1780 Kugan Vivekanandarajah <kuganv@linaro.org>
1782 PR target/59695
1783 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
1784 truncation.
1786 2014-01-15 Richard Biener <rguenther@suse.de>
1788 PR rtl-optimization/59802
1789 * lcm.c (compute_available): Use inverted postorder to seed
1790 the initial worklist.
1792 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1794 PR target/59803
1795 * config/s390/s390.c (s390_preferred_reload_class): Don't return
1796 ADDR_REGS for invalid symrefs in non-PIC code.
1798 2014-01-15 Jakub Jelinek <jakub@redhat.com>
1800 PR other/58712
1801 * builtins.c (determine_block_size): Initialize *probable_max_size
1802 even if len_rtx is CONST_INT.
1804 2014-01-14 Andrew Pinski <apinski@cavium.com>
1806 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
1807 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
1808 (cortexa53_tunings): Likewise.
1809 (aarch64_sched_issue_rate): New function.
1810 (TARGET_SCHED_ISSUE_RATE): Define.
1812 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1814 * ira-costs.c (find_costs_and_classes): Add missed
1815 ira_init_register_move_cost_if_necessary.
1817 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
1819 PR target/59787
1820 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
1822 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
1824 PR target/59794
1825 * config/i386/i386.c (type_natural_mode): Add a bool parameter
1826 to indicate if type is used for function return value. Warn ABI
1827 change if the vector mode isn't available for function return value.
1828 (ix86_function_arg_advance): Pass false to type_natural_mode.
1829 (ix86_function_arg): Likewise.
1830 (ix86_gimplify_va_arg): Likewise.
1831 (function_arg_32): Don't warn ABI change.
1832 (ix86_function_value): Pass true to type_natural_mode.
1833 (ix86_return_in_memory): Likewise.
1834 (ix86_struct_value_rtx): Removed.
1835 (TARGET_STRUCT_VALUE_RTX): Likewise.
1837 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1839 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
1840 converting a conditional jump into a conditional return.
1842 2014-01-14 Richard Biener <rguenther@suse.de>
1844 PR tree-optimization/58921
1845 PR tree-optimization/59006
1846 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
1847 hoisting invariant stmts.
1848 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
1849 invariant loads on the preheader edge if possible.
1851 2014-01-14 Joey Ye <joey.ye@arm.com>
1853 * doc/plugin.texi (Building GCC plugins): Update to C++.
1855 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
1857 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
1858 (_mm_rcp28_round_ss): Ditto.
1859 (_mm_rsqrt28_round_sd): Ditto.
1860 (_mm_rsqrt28_round_ss): Ditto.
1861 (_mm_rcp28_sd): Ditto.
1862 (_mm_rcp28_ss): Ditto.
1863 (_mm_rsqrt28_sd): Ditto.
1864 (_mm_rsqrt28_ss): Ditto.
1865 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
1866 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
1867 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
1868 (IX86_BUILTIN_RCP28SD): Ditto.
1869 (IX86_BUILTIN_RCP28SS): Ditto.
1870 (IX86_BUILTIN_RSQRT28SD): Ditto.
1871 (IX86_BUILTIN_RSQRT28SS): Ditto.
1872 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
1873 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
1874 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
1875 (ix86_expand_special_args_builtin): Expand new FTYPE.
1876 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
1877 (srcp14<mode>): Make insn unary.
1878 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
1879 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
1880 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
1881 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
1882 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
1883 Fix rounding: make it SAE only.
1884 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
1885 Ditto.
1886 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
1887 Ditto.
1888 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
1889 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
1890 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
1891 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
1892 (round_saeonly_mask_scalar_operand4): Ditto.
1893 (round_saeonly_mask_scalar_op3): Ditto.
1894 (round_saeonly_mask_scalar_op4): Ditto.
1896 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1898 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
1899 Implement -maltivec=be for vec_insert and vec_extract.
1901 2014-01-10 DJ Delorie <dj@redhat.com>
1903 * config/msp430/msp430.md (call_internal): Don't allow memory
1904 references with SP as the base register.
1905 (call_value_internal): Likewise.
1906 * config/msp430/constraints.md (Yc): New. For memory references
1907 that don't use SP as a base register.
1909 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
1910 "an integer without a # prefix"
1911 * config/msp430/msp430.md (epilogue_helper): Use it.
1913 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1915 PR target/59617
1916 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
1917 AVX512F gather builtins.
1918 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
1919 on gather decls with INTEGER_TYPE masktype.
1920 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
1921 directly into the builtin rather than hoisting it before loop.
1923 PR tree-optimization/59387
1924 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
1925 (scev_const_prop): If folded_casts and type has undefined overflow,
1926 use force_gimple_operand instead of force_gimple_operand_gsi and
1927 for each added stmt if it is assign with
1928 arith_code_with_undefined_signed_overflow, call
1929 rewrite_to_defined_overflow.
1930 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
1931 gimple-fold.h instead.
1932 (arith_code_with_undefined_signed_overflow,
1933 rewrite_to_defined_overflow): Moved to ...
1934 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
1935 rewrite_to_defined_overflow): ... here. No longer static.
1936 Include gimplify-me.h.
1937 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
1938 rewrite_to_defined_overflow): New prototypes.
1940 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1942 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
1944 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
1946 * builtins.c (get_object_alignment_2): Minor tweak.
1947 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
1949 2014-01-13 Christian Bruel <christian.bruel@st.com>
1951 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
1952 optimized non constant lengths.
1954 2014-01-13 Jakub Jelinek <jakub@redhat.com>
1956 PR libgomp/59194
1957 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
1958 load as __atomic_load_N if possible.
1960 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1962 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
1963 target parameter.
1964 (rs6000_expand_builtin): Adjust call.
1966 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
1968 PR target/58115
1969 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
1970 * config/rs6000/rs6000.c: Include target-globals.h.
1971 (rs6000_set_current_function): Instead of doing target_reinit
1972 unconditionally, use save_target_globals_default_opts and
1973 restore_target_globals.
1975 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
1976 FPSCR.
1977 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
1978 (rs6000_expand_builtin): Handle mffs and mtfsf.
1979 (rs6000_init_builtins): Define mffs and mtfsf.
1980 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
1981 (rs6000_mffs): New pattern.
1982 (rs6000_mtfsf): New pattern.
1984 2014-01-11 Bin Cheng <bin.cheng@arm.com>
1986 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
1987 Start narrowing with START. Apply candidate-use pair
1988 and check overall cost in narrowing.
1989 (iv_ca_prune): Pass new argument.
1991 2014-01-10 Jeff Law <law@redhat.com>
1993 PR middle-end/59743
1994 * ree.c (combine_reaching_defs): Ensure the defining statement
1995 occurs before the extension when optimizing extensions with
1996 different source and destination hard registers.
1998 2014-01-10 Jan Hubicka <jh@suse.cz>
2000 PR ipa/58585
2001 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
2002 vtables into the type inheritance graph.
2004 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2006 PR rtl-optimization/59754
2007 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
2008 modes in the REGNO != REGNO case.
2010 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2012 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
2014 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2016 PR tree-optimization/59745
2017 * tree-predcom.c (tree_predictive_commoning_loop): Call
2018 free_affine_expand_cache if giving up because components is NULL.
2020 * target-globals.c (save_target_globals): Allocate < 4KB structs using
2021 GC in payload of target_globals struct instead of allocating them on
2022 the heap and the larger structs separately using GC.
2023 * target-globals.h (struct target_globals): Make regs, hard_regs,
2024 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
2025 of GTY((skip)) and change type to void *.
2026 (reset_target_globals): Cast loads from those fields to corresponding
2027 types.
2029 2014-01-10 Steve Ellcey <sellcey@mips.com>
2031 PR plugins/59335
2032 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
2033 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
2034 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
2036 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
2038 PR target/59744
2039 * aarch64-modes.def (CC_Zmode): New flags mode.
2040 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
2041 represents an equality.
2042 (aarch64_get_condition_code): Handle CC_Zmode.
2043 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
2045 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2047 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
2048 extraction in good case.
2050 2014-01-10 Richard Biener <rguenther@suse.de>
2052 PR tree-optimization/59374
2053 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
2054 checking after SLP discovery. Mark stmts not participating
2055 in any SLP instance properly.
2057 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2059 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
2060 when handling a SET rtx.
2062 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2064 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
2065 (cortex-a57): Likewise.
2066 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
2068 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2070 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
2071 non-iwmmxt builtins.
2073 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
2075 PR ipa/58252
2076 PR ipa/59226
2077 * ipa-devirt.c record_target_from_binfo): Take as argument
2078 stack of binfos and lookup matching one for virtual inheritance.
2079 (possible_polymorphic_call_targets_1): Update.
2081 2014-01-10 Huacai Chen <chenhc@lemote.com>
2083 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
2084 kernel strings for Loongson-2E/2F/3A.
2086 2014-01-10 Jakub Jelinek <jakub@redhat.com>
2088 PR middle-end/59670
2089 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
2090 is_gimple_call before calling gimple_call_internal_p.
2092 2014-01-09 Steve Ellcey <sellcey@mips.com>
2094 * Makefile.in (TREE_FLOW_H): Remove.
2095 (TREE_SSA_H): Add file names from tree-flow.h.
2096 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
2097 * tree.h: Remove tree-flow.h reference.
2098 * hash-table.h: Remove tree-flow.h reference.
2099 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
2100 reference with tree-ssa-loop.h.
2102 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2104 * doc/invoke.texi: Add -maltivec={be,le} options, and document
2105 default element-order behavior for -maltivec.
2106 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
2107 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
2108 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
2109 when targeting big endian, at least for now.
2110 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
2112 2014-01-09 Jakub Jelinek <jakub@redhat.com>
2114 PR middle-end/47735
2115 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
2116 var satisfies use_register_for_decl, just take into account type
2117 alignment, rather than decl alignment.
2119 PR tree-optimization/59622
2120 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
2121 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
2122 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
2123 Don't devirtualize for inplace at all. For targets.length () == 1,
2124 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
2126 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
2128 * config/i386/i386.md (cpu): Remove the unused btver1.
2130 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
2132 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
2134 2014-01-09 Jakub Jelinek <jakub@redhat.com>
2136 PR target/58115
2137 * tree-core.h (struct target_globals): New forward declaration.
2138 (struct tree_target_option): Add globals field.
2139 * tree.h (TREE_TARGET_GLOBALS): Define.
2140 (prepare_target_option_nodes_for_pch): New prototype.
2141 * target-globals.h (struct target_globals): Define even if
2142 !SWITCHABLE_TARGET.
2143 * tree.c (prepare_target_option_node_for_pch,
2144 prepare_target_option_nodes_for_pch): New functions.
2145 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
2146 * config/i386/i386.c: Include target-globals.h.
2147 (ix86_set_current_function): Instead of doing target_reinit
2148 unconditionally, use save_target_globals_default_opts and
2149 restore_target_globals.
2151 2014-01-09 Richard Biener <rguenther@suse.de>
2153 PR tree-optimization/59715
2154 * tree-cfg.h (split_critical_edges): Declare.
2155 * tree-cfg.c (split_critical_edges): Export.
2156 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
2158 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
2160 * cfgexpand.c (expand_stack_vars): Optionally disable
2161 asan stack protection.
2162 (expand_used_vars): Likewise.
2163 (partition_stack_vars): Likewise.
2164 * asan.c (asan_emit_stack_protection): Optionally disable
2165 after return stack usage.
2166 (instrument_derefs): Optionally disable memory access instrumentation.
2167 (instrument_builtin_call): Likewise.
2168 (instrument_strlen_call): Likewise.
2169 (asan_protect_global): Optionally disable global variables protection.
2170 * doc/invoke.texi: Added doc for new options.
2171 * params.def: Added new options.
2172 * params.h: Likewise.
2174 2014-01-09 Jakub Jelinek <jakub@redhat.com>
2176 PR rtl-optimization/59724
2177 * ifcvt.c (cond_exec_process_if_block): Don't call
2178 flow_find_head_matching_sequence with 0 longest_match.
2179 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
2180 non-active insns if !stop_after.
2181 (try_head_merge_bb): Revert 2014-01-07 changes.
2183 2014-01-08 Jeff Law <law@redhat.com>
2185 * ree.c (get_sub_rtx): New function, extracted from...
2186 (merge_def_and_ext): Here.
2187 (combine_reaching_defs): Use get_sub_rtx.
2189 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
2191 * cgraph.h (varpool_variable_node): Do not choke on null node.
2193 2014-01-08 Catherine Moore <clm@codesourcery.com>
2195 * config/mips/mips.md (simple_return): Attempt to use JRC
2196 for microMIPS.
2197 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
2199 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
2201 PR rtl-optimization/59137
2202 * reorg.c (steal_delay_list_from_target): Call update_block for
2203 elided insns.
2204 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
2206 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2208 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
2209 two duplicate entries.
2211 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
2213 Revert:
2214 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
2216 * config/mips/mips.c (mips_truncated_op_cost): New function.
2217 (mips_rtx_costs): Adjust test for BADDU.
2218 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
2220 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
2222 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
2223 (*baddu_si): ...this new pattern.
2225 2014-01-08 Jakub Jelinek <jakub@redhat.com>
2227 PR ipa/59722
2228 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
2230 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2232 PR middle-end/57748
2233 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
2234 inner_reference_p.
2235 (expand_expr, expand_normal): Adjust.
2236 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
2237 inner_reference_p. Use inner_reference_p to expand inner references.
2238 (store_expr): Adjust.
2239 * cfgexpand.c (expand_call_stmt): Adjust.
2241 2014-01-08 Rong Xu <xur@google.com>
2243 * gcov-io.c (gcov_var): Move from gcov-io.h.
2244 (gcov_position): Ditto.
2245 (gcov_is_error): Ditto.
2246 (gcov_rewrite): Ditto.
2247 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
2248 only part to libgcc/libgcov.h.
2250 2014-01-08 Marek Polacek <polacek@redhat.com>
2252 PR middle-end/59669
2253 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
2255 2014-01-08 Marek Polacek <polacek@redhat.com>
2257 PR sanitizer/59667
2258 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
2260 2014-01-08 Jakub Jelinek <jakub@redhat.com>
2262 PR rtl-optimization/59649
2263 * stor-layout.c (get_mode_bounds): For BImode return
2264 0 and STORE_FLAG_VALUE.
2266 2014-01-08 Richard Biener <rguenther@suse.de>
2268 PR middle-end/59630
2269 * gimple.h (is_gimple_builtin_call): Remove.
2270 (gimple_builtin_call_types_compatible_p): New.
2271 (gimple_call_builtin_p): New overload.
2272 * gimple.c (is_gimple_builtin_call): Remove.
2273 (validate_call): Rename to ...
2274 (gimple_builtin_call_types_compatible_p): ... this and export. Also
2275 check return types.
2276 (validate_type): New static function.
2277 (gimple_call_builtin_p): New overload and adjust.
2278 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
2279 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
2280 (gimple_fold_stmt_to_constant_1): Likewise.
2281 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
2283 2014-01-08 Richard Biener <rguenther@suse.de>
2285 PR middle-end/59471
2286 * gimplify.c (gimplify_expr): Gimplify register-register type
2287 VIEW_CONVERT_EXPRs to separate stmts.
2289 2014-01-07 Jeff Law <law@redhat.com>
2291 PR middle-end/53623
2292 * ree.c (combine_set_extension): Handle case where source
2293 and destination registers in an extension insn are different.
2294 (combine_reaching_defs): Allow source and destination registers
2295 in extension to be different under limited circumstances.
2296 (add_removable_extension): Remove restriction that the
2297 source and destination registers in the extension are the same.
2298 (find_and_remove_re): Emit a copy from the extension's
2299 destination to its source after the defining insn if
2300 the source and destination registers are different.
2302 PR middle-end/59285
2303 * ifcvt.c (merge_if_block): If we are merging a block with more than
2304 one successor with a block with no successors, remove any BARRIER
2305 after the second block.
2307 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
2309 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
2311 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
2313 PR target/59652
2314 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
2315 for 14-bit register offsets when INT14_OK_STRICT is false.
2317 2014-01-07 Roland Stigge <stigge@antcom.de>
2318 Michael Meissner <meissner@linux.vnet.ibm.com>
2320 PR 57386/target
2321 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
2322 Only check TFmode for SPE constants. Don't check TImode or TDmode.
2324 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
2326 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
2327 -mcpu.
2329 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
2331 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
2332 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
2333 rtx is const0_rtx or not.
2335 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
2337 PR target/58115
2338 * target-globals.c (save_target_globals): Remove this_fn_optab
2339 handling.
2340 * toplev.c: Include optabs.h.
2341 (target_reinit): Temporarily restore the global options if another
2342 set of options are in force.
2344 2014-01-07 Jakub Jelinek <jakub@redhat.com>
2346 PR rtl-optimization/58668
2347 * cfgcleanup.c (flow_find_cross_jump): Don't count
2348 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
2349 to determine what is counted.
2350 (flow_find_head_matching_sequence): Use active_insn_p to determine
2351 what is counted.
2352 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
2353 counting change.
2354 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
2355 determine what is counted.
2357 PR tree-optimization/59643
2358 * tree-predcom.c (split_data_refs_to_components): If one dr is
2359 read and one write, determine_offset fails and the write isn't
2360 in the bad component, just put the read into the bad component.
2362 2014-01-07 Mike Stump <mikestump@comcast.net>
2363 Jakub Jelinek <jakub@redhat.com>
2365 PR pch/59436
2366 * tree-core.h (struct tree_optimization_option): Change optabs
2367 type from unsigned char * to void *.
2368 * optabs.c (init_tree_optimization_optabs): Adjust
2369 TREE_OPTIMIZATION_OPTABS initialization.
2371 2014-01-06 Jakub Jelinek <jakub@redhat.com>
2373 PR target/59644
2374 * config/i386/i386.h (struct machine_function): Add
2375 no_drap_save_restore field.
2376 * config/i386/i386.c (ix86_save_reg): Use
2377 !cfun->machine->no_drap_save_restore instead of
2378 crtl->stack_realign_needed.
2379 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
2380 this function clears frame_pointer_needed. Set
2381 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
2382 and DRAP reg is needed.
2384 2014-01-06 Marek Polacek <polacek@redhat.com>
2386 PR c/57773
2387 * doc/implement-c.texi: Mention that other integer types are
2388 permitted as bit-field types in strictly conforming mode.
2390 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
2392 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
2393 is newly allocated.
2395 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
2397 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
2399 2014-01-06 Martin Jambor <mjambor@suse.cz>
2401 PR ipa/59008
2402 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
2403 to int.
2404 * ipa-prop.c (ipa_print_node_params): Fix indentation.
2406 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
2408 PR debug/59350
2409 PR debug/59510
2410 * var-tracking.c (add_stores): Preserve the value of the source even if
2411 we don't record the store.
2413 2014-01-06 Terry Guo <terry.guo@arm.com>
2415 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
2417 2014-01-05 Iain Sandoe <iain@codesourcery.com>
2419 PR bootstrap/59541
2420 * config/darwin.c (darwin_function_section): Adjust return values to
2421 correspond to optimisation changes made in r206070.
2423 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
2425 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
2426 from prefetch_block tune setting.
2427 (nocona_cost): Correct size of prefetch block to 64.
2429 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
2431 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
2432 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
2433 used to save the static chain register in the computation of the offset
2434 from which the FP registers need to be restored.
2436 2014-01-04 Jakub Jelinek <jakub@redhat.com>
2438 PR tree-optimization/59519
2439 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
2440 ICE if get_current_def (current_new_name) is already non-NULL, as long
2441 as it is a phi result of some other phi in *new_exit_bb that has
2442 the same argument.
2444 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
2445 or vmovdqu* for misaligned_operand.
2446 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
2447 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
2448 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
2449 aligned_mem for AVX512F masked aligned load and store builtins and for
2450 non-temporal moves.
2452 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
2454 PR tree-optimization/59651
2455 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
2456 Address range for negative step should be added by TYPE_SIZE_UNIT.
2458 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
2460 * config/m68k/m68k.c (handle_move_double): Handle pushes with
2461 overlapping registers also for registers other than the stack pointer.
2463 2014-01-03 Marek Polacek <polacek@redhat.com>
2465 PR other/59661
2466 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
2467 __builtin_FILE.
2469 2014-01-03 Jakub Jelinek <jakub@redhat.com>
2471 PR target/59625
2472 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
2473 asm goto as jump.
2475 * config/i386/i386.md (MODE_SIZE): New mode attribute.
2476 (push splitter): Use <P:MODE_SIZE> instead of
2477 GET_MODE_SIZE (<P:MODE>mode).
2478 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
2479 (mov -1, reg peephole2): Likewise.
2480 * config/i386/sse.md (*mov<mode>_internal,
2481 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
2482 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
2483 *<code><mode>3, *andnot<mode>3<mask_name>,
2484 <mask_codefor><code><mode>3<mask_name>): Likewise.
2485 * config/i386/subst.md (mask_mode512bit_condition,
2486 sd_mask_mode512bit_condition): Likewise.
2488 2014-01-02 Xinliang David Li <davidxl@google.com>
2490 PR tree-optimization/59303
2491 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
2492 (dump_predicates): Better output format.
2493 (pred_equal_p): New function.
2494 (is_neq_relop_p): Ditto.
2495 (is_neq_zero_form_p): Ditto.
2496 (pred_expr_equal_p): Ditto.
2497 (pred_neg_p): Ditto.
2498 (simplify_pred): Ditto.
2499 (simplify_preds_2): Ditto.
2500 (simplify_preds_3): Ditto.
2501 (simplify_preds_4): Ditto.
2502 (simplify_preds): Ditto.
2503 (push_pred): Ditto.
2504 (push_to_worklist): Ditto.
2505 (get_pred_info_from_cmp): Ditto.
2506 (is_degenerated_phi): Ditto.
2507 (normalize_one_pred_1): Ditto.
2508 (normalize_one_pred): Ditto.
2509 (normalize_one_pred_chain): Ditto.
2510 (normalize_preds): Ditto.
2511 (normalize_cond_1): Remove function.
2512 (normalize_cond): Ditto.
2513 (is_gcond_subset_of): Ditto.
2514 (is_subset_of_any): Ditto.
2515 (is_or_set_subset_of): Ditto.
2516 (is_and_set_subset_of): Ditto.
2517 (is_norm_cond_subset_of): Ditto.
2518 (pred_chain_length_cmp): Ditto.
2519 (convert_control_dep_chain_into_preds): Type change.
2520 (find_predicates): Ditto.
2521 (find_def_preds): Ditto.
2522 (destroy_predicates_vecs): Ditto.
2523 (find_matching_predicates_in_rest_chains): Ditto.
2524 (use_pred_not_overlap_with_undef_path_pred): Ditto.
2525 (is_pred_expr_subset): Ditto.
2526 (is_pred_chain_subset_of): Ditto.
2527 (is_included_in): Ditto.
2528 (is_superset_of): Ditto.
2530 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2532 Update copyright years.
2534 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
2536 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
2537 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
2538 config/arc/arc.md, config/arc/arc.opt,
2539 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
2540 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
2541 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
2542 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
2543 config/linux-protos.h, config/linux.c, config/winnt-c.c,
2544 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
2545 vtable-verify.c, vtable-verify.h: Use the standard form for the
2546 copyright notice.
2548 2014-01-02 Tobias Burnus <burnus@net-b.de>
2550 * gcc.c (process_command): Update copyright notice dates.
2551 * gcov-dump.c: Ditto.
2552 * gcov.c: Ditto.
2553 * doc/cpp.texi: Bump @copying's copyright year.
2554 * doc/cppinternals.texi: Ditto.
2555 * doc/gcc.texi: Ditto.
2556 * doc/gccint.texi: Ditto.
2557 * doc/gcov.texi: Ditto.
2558 * doc/install.texi: Ditto.
2559 * doc/invoke.texi: Ditto.
2561 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2563 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
2565 2014-01-01 Jakub Jelinek <jakub@redhat.com>
2567 * config/i386/sse.md (*mov<mode>_internal): Guard
2568 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
2570 PR rtl-optimization/59647
2571 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
2572 new_rtx into UNSIGNED_FLOAT rtxes.
2573 \f
2574 Copyright (C) 2014 Free Software Foundation, Inc.
2576 Copying and distribution of this file, with or without modification,
2577 are permitted in any medium without royalty provided the copyright
2578 notice and this notice are preserved.