1 2019-01-10 Tamar Christina <tamar.christina@arm.com>
3 * config/aarch64/aarch64-builtins.c
4 (aarch64_init_builtins): Move aarch64_init_fcmla_laneq_builtins...
5 (aarch64_init_simd_builtins): ...Here
7 2019-01-10 Vladimir Makarov <vmakarov@redhat.com>
9 PR rtl-optimization/87305
11 (setup_live_pseudos_and_spill_after_risky_transforms): Check
12 allocation for big endian pseudos used as paradoxical subregs and
13 spill them if it is wrong.
14 * lra-constraints.c (lra_constraints): Add a comment.
16 2019-01-10 Richard Biener <rguenther@suse.de>
18 PR tree-optimization/88792
19 * tree-ssa-pre.c (get_representative_for): Do not return a
22 2019-01-10 Jakub Jelinek <jakub@redhat.com>
26 * function.c (assign_stack_local_1): Revert the 2018-11-21 changes.
27 (assign_parm_setup_block): Do the argument slot realignment here
30 2019-01-10 Stefan Agner <stefan@agner.ch>
33 * config/arm/arm.c (arm_option_override_internal): Force
34 opts->x_inline_asm_unified to true only if TARGET_THUMB2_P.
36 2019-01-10 Jakub Jelinek <jakub@redhat.com>
39 * attribs.c (handle_dll_attribute): Clear TREE_STATIC after setting
42 2019-01-10 Tamar Christina <tamar.christina@arm.com>
44 * config/arm/arm-builtins.c
45 (enum arm_type_qualifiers): Add qualifier_lane_pair_index.
46 (MAC_LANE_PAIR_QUALIFIERS): New.
47 (arm_expand_builtin_args): Use it.
48 (arm_expand_builtin_1): Likewise.
49 * config/arm/arm-protos.h (neon_vcmla_lane_prepare_operands): New.
50 * config/arm/arm.c (neon_vcmla_lane_prepare_operands): New.
51 * config/arm/arm-c.c (arm_cpu_builtins): Add __ARM_FEATURE_COMPLEX.
52 * config/arm/arm_neon.h:
53 (vcadd_rot90_f16): New.
54 (vcaddq_rot90_f16): New.
55 (vcadd_rot270_f16): New.
56 (vcaddq_rot270_f16): New.
59 (vcmla_lane_f16): New.
60 (vcmla_laneq_f16): New.
61 (vcmlaq_lane_f16): New.
62 (vcmlaq_laneq_f16): New.
63 (vcmla_rot90_f16): New.
64 (vcmlaq_rot90_f16): New.
65 (vcmla_rot90_lane_f16): New.
66 (vcmla_rot90_laneq_f16): New.
67 (vcmlaq_rot90_lane_f16): New.
68 (vcmlaq_rot90_laneq_f16): New.
69 (vcmla_rot180_f16): New.
70 (vcmlaq_rot180_f16): New.
71 (vcmla_rot180_lane_f16): New.
72 (vcmla_rot180_laneq_f16): New.
73 (vcmlaq_rot180_lane_f16): New.
74 (vcmlaq_rot180_laneq_f16): New.
75 (vcmla_rot270_f16): New.
76 (vcmlaq_rot270_f16): New.
77 (vcmla_rot270_lane_f16): New.
78 (vcmla_rot270_laneq_f16): New.
79 (vcmlaq_rot270_lane_f16): New.
80 (vcmlaq_rot270_laneq_f16): New.
81 (vcadd_rot90_f32): New.
82 (vcaddq_rot90_f32): New.
83 (vcadd_rot270_f32): New.
84 (vcaddq_rot270_f32): New.
87 (vcmla_lane_f32): New.
88 (vcmla_laneq_f32): New.
89 (vcmlaq_lane_f32): New.
90 (vcmlaq_laneq_f32): New.
91 (vcmla_rot90_f32): New.
92 (vcmlaq_rot90_f32): New.
93 (vcmla_rot90_lane_f32): New.
94 (vcmla_rot90_laneq_f32): New.
95 (vcmlaq_rot90_lane_f32): New.
96 (vcmlaq_rot90_laneq_f32): New.
97 (vcmla_rot180_f32): New.
98 (vcmlaq_rot180_f32): New.
99 (vcmla_rot180_lane_f32): New.
100 (vcmla_rot180_laneq_f32): New.
101 (vcmlaq_rot180_lane_f32): New.
102 (vcmlaq_rot180_laneq_f32): New.
103 (vcmla_rot270_f32): New.
104 (vcmlaq_rot270_f32): New.
105 (vcmla_rot270_lane_f32): New.
106 (vcmla_rot270_laneq_f32): New.
107 (vcmlaq_rot270_lane_f32): New.
108 (vcmlaq_rot270_laneq_f32): New.
109 * config/arm/arm_neon_builtins.def (vcadd90, vcadd270, vcmla0, vcmla90,
110 vcmla180, vcmla270, vcmla_lane0, vcmla_lane90, vcmla_lane180, vcmla_lane270,
111 vcmla_laneq0, vcmla_laneq90, vcmla_laneq180, vcmla_laneq270,
112 vcmlaq_lane0, vcmlaq_lane90, vcmlaq_lane180, vcmlaq_lane270): New.
113 * config/arm/neon.md (neon_vcmla_lane<rot><mode>,
114 neon_vcmla_laneq<rot><mode>, neon_vcmlaq_lane<rot><mode>): New.
115 * config/arm/arm.c (arm_arch8_3, arm_arch8_4): New.
116 * config/arm/arm.h (TARGET_COMPLEX, arm_arch8_3, arm_arch8_4): New.
117 (arm_option_reconfigure_globals): Use them.
118 * config/arm/iterators.md (VDF, VQ_HSF): New.
120 (VF_constraint, rot, rotsplit1, rotsplit2): Add V4HF and V8HF.
121 * config/arm/neon.md (neon_vcadd<rot><mode>, neon_vcmla<rot><mode>): New.
122 * config/arm/unspecs.md (UNSPEC_VCADD90, UNSPEC_VCADD270,
123 UNSPEC_VCMLA, UNSPEC_VCMLA90, UNSPEC_VCMLA180, UNSPEC_VCMLA270): New.
125 2019-01-10 Tamar Christina <tamar.christina@arm.com>
127 * config/aarch64/aarch64-builtins.c (enum aarch64_type_qualifiers): Add qualifier_lane_pair_index.
128 (emit-rtl.h): Include.
129 (TYPES_QUADOP_LANE_PAIR): New.
130 (aarch64_simd_expand_args): Use it.
131 (aarch64_simd_expand_builtin): Likewise.
132 (AARCH64_SIMD_FCMLA_LANEQ_BUILTINS, aarch64_fcmla_laneq_builtin_datum): New.
133 (FCMLA_LANEQ_BUILTIN, AARCH64_SIMD_FCMLA_LANEQ_BUILTIN_BASE,
134 AARCH64_SIMD_FCMLA_LANEQ_BUILTINS, aarch64_fcmla_lane_builtin_data,
135 aarch64_init_fcmla_laneq_builtins, aarch64_expand_fcmla_builtin): New.
136 (aarch64_init_builtins): Add aarch64_init_fcmla_laneq_builtins.
137 (aarch64_expand_buildin): Add AARCH64_SIMD_BUILTIN_FCMLA_LANEQ0_V2SF,
138 AARCH64_SIMD_BUILTIN_FCMLA_LANEQ90_V2SF, AARCH64_SIMD_BUILTIN_FCMLA_LANEQ180_V2SF,
139 AARCH64_SIMD_BUILTIN_FCMLA_LANEQ2700_V2SF, AARCH64_SIMD_BUILTIN_FCMLA_LANEQ0_V4HF,
140 AARCH64_SIMD_BUILTIN_FCMLA_LANEQ90_V4HF, AARCH64_SIMD_BUILTIN_FCMLA_LANEQ180_V4HF,
141 AARCH64_SIMD_BUILTIN_FCMLA_LANEQ270_V4HF.
142 * config/aarch64/iterators.md (FCMLA_maybe_lane): New.
143 * config/aarch64/aarch64-c.c (aarch64_update_cpp_builtins): Add __ARM_FEATURE_COMPLEX.
144 * config/aarch64/aarch64-simd-builtins.def (fcadd90, fcadd270, fcmla0, fcmla90,
145 fcmla180, fcmla270, fcmla_lane0, fcmla_lane90, fcmla_lane180, fcmla_lane270,
146 fcmla_laneq0, fcmla_laneq90, fcmla_laneq180, fcmla_laneq270,
147 fcmlaq_lane0, fcmlaq_lane90, fcmlaq_lane180, fcmlaq_lane270): New.
148 * config/aarch64/aarch64-simd.md (aarch64_fcmla_lane<rot><mode>,
149 aarch64_fcmla_laneq<rot>v4hf, aarch64_fcmlaq_lane<rot><mode>,aarch64_fcadd<rot><mode>,
150 aarch64_fcmla<rot><mode>): New.
151 * config/aarch64/arm_neon.h:
152 (vcadd_rot90_f16): New.
153 (vcaddq_rot90_f16): New.
154 (vcadd_rot270_f16): New.
155 (vcaddq_rot270_f16): New.
158 (vcmla_lane_f16): New.
159 (vcmla_laneq_f16): New.
160 (vcmlaq_lane_f16): New.
161 (vcmlaq_rot90_lane_f16): New.
162 (vcmla_rot90_laneq_f16): New.
163 (vcmla_rot90_lane_f16): New.
164 (vcmlaq_rot90_f16): New.
165 (vcmla_rot90_f16): New.
166 (vcmlaq_laneq_f16): New.
167 (vcmla_rot180_laneq_f16): New.
168 (vcmla_rot180_lane_f16): New.
169 (vcmlaq_rot180_f16): New.
170 (vcmla_rot180_f16): New.
171 (vcmlaq_rot90_laneq_f16): New.
172 (vcmlaq_rot270_laneq_f16): New.
173 (vcmlaq_rot270_lane_f16): New.
174 (vcmla_rot270_laneq_f16): New.
175 (vcmlaq_rot270_f16): New.
176 (vcmla_rot270_f16): New.
177 (vcmlaq_rot180_laneq_f16): New.
178 (vcmlaq_rot180_lane_f16): New.
179 (vcmla_rot270_lane_f16): New.
180 (vcadd_rot90_f32): New.
181 (vcaddq_rot90_f32): New.
182 (vcaddq_rot90_f64): New.
183 (vcadd_rot270_f32): New.
184 (vcaddq_rot270_f32): New.
185 (vcaddq_rot270_f64): New.
189 (vcmla_lane_f32): New.
190 (vcmla_laneq_f32): New.
191 (vcmlaq_lane_f32): New.
192 (vcmlaq_laneq_f32): New.
193 (vcmla_rot90_f32): New.
194 (vcmlaq_rot90_f32): New.
195 (vcmlaq_rot90_f64): New.
196 (vcmla_rot90_lane_f32): New.
197 (vcmla_rot90_laneq_f32): New.
198 (vcmlaq_rot90_lane_f32): New.
199 (vcmlaq_rot90_laneq_f32): New.
200 (vcmla_rot180_f32): New.
201 (vcmlaq_rot180_f32): New.
202 (vcmlaq_rot180_f64): New.
203 (vcmla_rot180_lane_f32): New.
204 (vcmla_rot180_laneq_f32): New.
205 (vcmlaq_rot180_lane_f32): New.
206 (vcmlaq_rot180_laneq_f32): New.
207 (vcmla_rot270_f32): New.
208 (vcmlaq_rot270_f32): New.
209 (vcmlaq_rot270_f64): New.
210 (vcmla_rot270_lane_f32): New.
211 (vcmla_rot270_laneq_f32): New.
212 (vcmlaq_rot270_lane_f32): New.
213 (vcmlaq_rot270_laneq_f32): New.
214 * config/aarch64/aarch64.h (TARGET_COMPLEX): New.
215 * config/aarch64/iterators.md (UNSPEC_FCADD90, UNSPEC_FCADD270,
216 UNSPEC_FCMLA, UNSPEC_FCMLA90, UNSPEC_FCMLA180, UNSPEC_FCMLA270): New.
219 * config/arm/types.md (neon_fcadd, neon_fcmla): New.
221 2019-01-09 Sandra Loosemore <sandra@codesourcery.com>
225 * config/pa/pa.c: Change "can not" to "cannot".
226 * gimple-ssa-evrp-analyze.c: Likewise.
227 * ipa-icf.c: Likewise.
228 * ipa-polymorphic-call.c: Likewise.
229 * ipa-pure-const.c: Likewise.
230 * lra-constraints.c: Likewise.
231 * lra-remat.c: Likewise.
232 * reload1.c: Likewise.
234 * tree-ssa-uninit.c: Likewise.
236 2019-01-09 Sandra Loosemore <sandra@codesourcery.com>
240 * Makefile.in: Mechanically replace "can not" with "cannot".
242 * builtins.c: Likewise.
244 * cgraph.c: Likewise.
245 * cgraph.h: Likewise.
246 * cgraphclones.c: Likewise.
247 * cgraphunit.c: Likewise.
248 * combine-stack-adj.c: Likewise.
249 * combine.c: Likewise.
250 * common/config/i386/i386-common.c: Likewise.
251 * config/aarch64/aarch64.c: Likewise.
252 * config/alpha/sync.md: Likewise.
253 * config/arc/arc.c: Likewise.
254 * config/arc/predicates.md: Likewise.
255 * config/arm/arm-c.c: Likewise.
256 * config/arm/arm.c: Likewise.
257 * config/arm/arm.h: Likewise.
258 * config/arm/arm.md: Likewise.
259 * config/arm/cortex-r4f.md: Likewise.
260 * config/csky/csky.c: Likewise.
261 * config/csky/csky.h: Likewise.
262 * config/darwin-f.c: Likewise.
263 * config/epiphany/epiphany.md: Likewise.
264 * config/i386/i386.c: Likewise.
265 * config/i386/sol2.h: Likewise.
266 * config/m68k/m68k.c: Likewise.
267 * config/mcore/mcore.h: Likewise.
268 * config/microblaze/microblaze.md: Likewise.
269 * config/mips/20kc.md: Likewise.
270 * config/mips/sb1.md: Likewise.
271 * config/nds32/nds32.c: Likewise.
272 * config/nds32/predicates.md: Likewise.
273 * config/pa/pa.c: Likewise.
274 * config/rs6000/e300c2c3.md: Likewise.
275 * config/rs6000/rs6000.c: Likewise.
276 * config/s390/s390.h: Likewise.
277 * config/sh/sh.c: Likewise.
278 * config/sh/sh.md: Likewise.
279 * config/spu/vmx2spu.h: Likewise.
281 * dbxout.c: Likewise.
282 * df-scan.c: Likewise.
283 * doc/cfg.texi: Likewise.
284 * doc/extend.texi: Likewise.
285 * doc/fragments.texi: Likewise.
286 * doc/gty.texi: Likewise.
287 * doc/invoke.texi: Likewise.
288 * doc/lto.texi: Likewise.
289 * doc/md.texi: Likewise.
290 * doc/objc.texi: Likewise.
291 * doc/rtl.texi: Likewise.
292 * doc/tm.texi: Likewise.
294 * emit-rtl.c: Likewise.
295 * emit-rtl.h: Likewise.
296 * except.c: Likewise.
297 * expmed.c: Likewise.
299 * fold-const.c: Likewise.
300 * genautomata.c: Likewise.
301 * gimple-fold.c: Likewise.
302 * hard-reg-set.h: Likewise.
304 * ipa-comdats.c: Likewise.
305 * ipa-cp.c: Likewise.
306 * ipa-devirt.c: Likewise.
307 * ipa-fnsummary.c: Likewise.
308 * ipa-icf.c: Likewise.
309 * ipa-inline-transform.c: Likewise.
310 * ipa-inline.c: Likewise.
311 * ipa-polymorphic-call.c: Likewise.
312 * ipa-profile.c: Likewise.
313 * ipa-prop.c: Likewise.
314 * ipa-pure-const.c: Likewise.
315 * ipa-reference.c: Likewise.
316 * ipa-split.c: Likewise.
317 * ipa-visibility.c: Likewise.
319 * ira-build.c: Likewise.
320 * ira-color.c: Likewise.
321 * ira-conflicts.c: Likewise.
322 * ira-costs.c: Likewise.
323 * ira-int.h: Likewise.
324 * ira-lives.c: Likewise.
327 * loop-invariant.c: Likewise.
328 * loop-unroll.c: Likewise.
329 * lower-subreg.c: Likewise.
330 * lra-assigns.c: Likewise.
331 * lra-constraints.c: Likewise.
332 * lra-eliminations.c: Likewise.
333 * lra-lives.c: Likewise.
334 * lra-remat.c: Likewise.
335 * lra-spills.c: Likewise.
337 * lto-cgraph.c: Likewise.
338 * lto-streamer-out.c: Likewise.
339 * postreload-gcse.c: Likewise.
340 * predict.c: Likewise.
341 * profile-count.h: Likewise.
342 * profile.c: Likewise.
345 * reload.c: Likewise.
346 * reload1.c: Likewise.
348 * resource.c: Likewise.
351 * rtlanal.c: Likewise.
352 * sched-deps.c: Likewise.
353 * sched-ebb.c: Likewise.
354 * sched-rgn.c: Likewise.
355 * sel-sched-ir.c: Likewise.
356 * sel-sched.c: Likewise.
357 * shrink-wrap.c: Likewise.
358 * simplify-rtx.c: Likewise.
359 * symtab.c: Likewise.
360 * target.def: Likewise.
361 * toplev.c: Likewise.
362 * tree-call-cdce.c: Likewise.
363 * tree-cfg.c: Likewise.
364 * tree-complex.c: Likewise.
365 * tree-core.h: Likewise.
366 * tree-eh.c: Likewise.
367 * tree-inline.c: Likewise.
368 * tree-loop-distribution.c: Likewise.
369 * tree-nrv.c: Likewise.
370 * tree-profile.c: Likewise.
371 * tree-sra.c: Likewise.
372 * tree-ssa-alias.c: Likewise.
373 * tree-ssa-dce.c: Likewise.
374 * tree-ssa-dom.c: Likewise.
375 * tree-ssa-forwprop.c: Likewise.
376 * tree-ssa-loop-im.c: Likewise.
377 * tree-ssa-loop-ivcanon.c: Likewise.
378 * tree-ssa-loop-ivopts.c: Likewise.
379 * tree-ssa-loop-niter.c: Likewise.
380 * tree-ssa-phionlycprop.c: Likewise.
381 * tree-ssa-phiopt.c: Likewise.
382 * tree-ssa-propagate.c: Likewise.
383 * tree-ssa-threadedge.c: Likewise.
384 * tree-ssa-threadupdate.c: Likewise.
385 * tree-ssa-uninit.c: Likewise.
386 * tree-ssanames.c: Likewise.
387 * tree-streamer-out.c: Likewise.
390 * vr-values.c: Likewise.
392 2019-01-09 Uroš Bizjak <ubizjak@gmail.com>
394 * config/i386/i386-protos.h (ix86_expand_xorsign): New prototype.
395 (ix86_split_xorsign): Ditto.
396 * config/i386/i386.c (ix86_expand_xorsign): New function.
397 (ix86_split_xorsign): Ditto.
398 * config/i386/i386.md (UNSPEC_XORSIGN): New unspec.
399 (xorsign<mode>3): New expander.
400 (xorsign<mode>3_1): New insn_and_split pattern.
401 * config/i386/sse.md (xorsign<mode>3): New expander.
403 2019-01-09 Eric Botcazou <ebotcazou@adacore.com>
405 * config/sparc/sparc.md (*tablejump_sp32): Merge into...
406 (*tablejump_sp64): Likewise.
407 (*tablejump<P:mode>): ...this.
408 (*call_address_sp32): Merge into...
409 (*call_address_sp64): Likewise.
410 (*call_address<P:mode>): ...this.
411 (*call_symbolic_sp32): Merge into...
412 (*call_symbolic_sp64): Likewise.
413 (*call_symbolic<P:mode>): ...this.
414 (call_value): Remove constraint and add predicate.
415 (*call_value_address_sp32): Merge into...
416 (*call_value_address_sp64): Likewise.
417 (*call_value_address<P:mode>): ...this.
418 (*call_value_symbolic_sp32): Merge into...
419 (*call_value_symbolic_sp64): Likewise.
420 (*call_value_symbolic<P:mode>): ...this.
421 (*sibcall_symbolic_sp32): Merge into...
422 (*sibcall_symbolic_sp64): Likewise.
423 (*sibcall_symbolic<P:mode>): ...this.
424 (sibcall_value): Remove constraint and add predicate.
425 (*sibcall_value_symbolic_sp32): Merge into...
426 (*sibcall_value_symbolic_sp64): Likewise.
427 (*sibcall_value_symbolic<P:mode>): ...this.
428 (window_save): Minor tweak.
429 (*branch_sp32): Merge into...
430 (*branch_sp64): Likewise.
431 (*branch<P:mode>): ...this.
433 2019-01-09 Eric Botcazou <ebotcazou@adacore.com>
434 James Clarke <jrtc27@jrtc27.com>
437 * config/sparc/sparc.c (sparc_legitimize_tls_address): Only use Pmode
438 consistently in TLS address generation and adjust code to the renaming
439 of patterns. Mark calls to __tls_get_addr as const.
440 * config/sparc/sparc.md (tgd_hi22): Turn into...
441 (tgd_hi22<P:mode>): ...this and use Pmode throughout.
442 (tgd_lo10): Turn into...
443 (tgd_lo10<P:mode>): ...this and use Pmode throughout.
444 (tgd_add32): Merge into...
445 (tgd_add64): Likewise.
446 (tgd_add<P:mode>): ...this and use Pmode throughout.
447 (tldm_hi22): Turn into...
448 (tldm_hi22<P:mode>): ...this and use Pmode throughout.
449 (tldm_lo10): Turn into...
450 (tldm_lo10<P:mode>): ...this and use Pmode throughout.
451 (tldm_add32): Merge into...
452 (tldm_add64): Likewise.
453 (tldm_add<P:mode>): ...this and use Pmode throughout.
454 (tldm_call32): Merge into...
455 (tldm_call64): Likewise.
456 (tldm_call<P:mode>): ...this and use Pmode throughout.
457 (tldo_hix22): Turn into...
458 (tldo_hix22<P:mode>): ...this and use Pmode throughout.
459 (tldo_lox10): Turn into...
460 (tldo_lox10<P:mode>): ...this and use Pmode throughout.
461 (tldo_add32): Merge into...
462 (tldo_add64): Likewise.
463 (tldo_add<P:mode>): ...this and use Pmode throughout.
464 (tie_hi22): Turn into...
465 (tie_hi22<P:mode>): ...this and use Pmode throughout.
466 (tie_lo10): Turn into...
467 (tie_lo10<P:mode>): ...this and use Pmode throughout.
468 (tie_ld64): Use DImode throughout.
469 (tie_add32): Merge into...
470 (tie_add64): Likewise.
471 (tie_add<P:mode>): ...this and use Pmode throughout.
472 (tle_hix22_sp32): Merge into...
473 (tle_hix22_sp64): Likewise.
474 (tle_hix22<P:mode>): ...this and use Pmode throughout.
475 (tle_lox22_sp32): Merge into...
476 (tle_lox22_sp64): Likewise.
477 (tle_lox22<P:mode>): ...this and use Pmode throughout.
478 (*tldo_ldub_sp32): Merge into...
479 (*tldo_ldub_sp64): Likewise.
480 (*tldo_ldub<P:mode>): ...this and use Pmode throughout.
481 (*tldo_ldub1_sp32): Merge into...
482 (*tldo_ldub1_sp64): Likewise.
483 (*tldo_ldub1<P:mode>): ...this and use Pmode throughout.
484 (*tldo_ldub2_sp32): Merge into...
485 (*tldo_ldub2_sp64): Likewise.
486 (*tldo_ldub2<P:mode>): ...this and use Pmode throughout.
487 (*tldo_ldsb1_sp32): Merge into...
488 (*tldo_ldsb1_sp64): Likewise.
489 (*tldo_ldsb1<P:mode>): ...this and use Pmode throughout.
490 (*tldo_ldsb2_sp32): Merge into...
491 (*tldo_ldsb2_sp64): Likewise.
492 (*tldo_ldsb2<P:mode>): ...this and use Pmode throughout.
493 (*tldo_ldub3_sp64): Use DImode throughout.
494 (*tldo_ldsb3_sp64): Likewise.
495 (*tldo_lduh_sp32): Merge into...
496 (*tldo_lduh_sp64): Likewise.
497 (*tldo_lduh<P:mode>): ...this and use Pmode throughout.
498 (*tldo_lduh1_sp32): Merge into...
499 (*tldo_lduh1_sp64): Likewise.
500 (*tldo_lduh1<P:mode>): ...this and use Pmode throughout.
501 (*tldo_ldsh1_sp32): Merge into...
502 (*tldo_ldsh1_sp64): Likewise.
503 (*tldo_ldsh1<P:mode>): ...this and use Pmode throughout.
504 (*tldo_lduh2_sp64): Use DImode throughout.
505 (*tldo_ldsh2_sp64): Likewise.
506 (*tldo_lduw_sp32): Merge into...
507 (*tldo_lduw_sp64): Likewise.
508 (*tldo_lduw<P:mode>): ...this and use Pmode throughout.
509 (*tldo_lduw1_sp64): Use DImode throughout.
510 (*tldo_ldsw1_sp64): Likewise.
511 (*tldo_ldx_sp64): Likewise.
512 (*tldo_stb_sp32): Merge into...
513 (*tldo_stb_sp64): Likewise.
514 (*tldo_stb<P:mode>): ...this and use Pmode throughout.
515 (*tldo_sth_sp32): Merge into...
516 (*tldo_sth_sp64): Likewise.
517 (*tldo_sth<P:mode>): ...this and use Pmode throughout.
518 (*tldo_stw_sp32): Merge into...
519 (*tldo_stw_sp64): Likewise.
520 (*tldo_stw<P:mode>): ...this and use Pmode throughout.
521 (*tldo_stx_sp64): Use DImode throughout.
523 2018-01-09 Sudakshina Das <sudi.das@arm.com>
525 * config/aarch64/aarch64.c (aarch64_override_options): Add case to
526 check configure option to set BTI and Return Address Signing.
527 * configure.ac: Add --enable-standard-branch-protection and
528 --disable-standard-branch-protection.
529 * configure: Regenerated.
530 * doc/install.texi: Document the same.
532 2018-01-09 Sudakshina Das <sudi.das@arm.com>
533 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
535 * config.gcc (aarch64*-*-*): Add aarch64-bti-insert.o.
536 * config/aarch64/aarch64.h: Update comment for TRAMPOLINE_SIZE.
537 * config/aarch64/aarch64.c (aarch64_asm_trampoline_template): Update
539 * config/aarch64/aarch64-bti-insert.c: New file.
540 * config/aarch64/aarch64-passes.def (INSERT_PASS_BEFORE): Insert bti
542 * config/aarch64/aarch64-protos.h (make_pass_insert_bti): Declare the
544 * config/aarch64/aarch64.md (unspecv): Add UNSPECV_BTI_NOARG,
545 UNSPECV_BTI_C, UNSPECV_BTI_J and UNSPECV_BTI_JC.
546 (bti_noarg, bti_j, bti_c, bti_jc): New define_insns.
547 * config/aarch64/t-aarch64: Add rule for aarch64-bti-insert.o.
549 2018-01-09 Sudakshina Das <sudi.das@arm.com>
551 * config/aarch64/aarch64-protos.h (aarch64_bti_enabled): Declare.
552 * config/aarch64/aarch64.c (aarch64_handle_no_branch_protection):
553 Disable bti for -mbranch-protection=none.
554 (aarch64_handle_standard_branch_protection): Enable bti for
555 -mbranch-protection=standard.
556 (aarch64_handle_bti_protection): Enable bti for "bti" in the string to
558 (aarch64_bti_enabled): Check if bti is enabled.
559 * config/aarch64/aarch64.opt: Declare target variable.
560 * doc/invoke.texi: Add bti to the -mbranch-protection documentation.
562 2018-01-09 Sudakshina Das <sudi.das@arm.com>
564 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use new
565 epilogue/prologue scratch registers EP0_REGNUM and EP1_REGNUM.
566 (aarch64_expand_epilogue): Likewise.
567 (aarch64_output_mi_thunk): Likewise
568 * config/aarch64/aarch64.h (REG_CLASS_CONTENTS): Change
569 TAILCALL_ADDR_REGS to x16 and x17.
570 * config/aarch64/aarch64.md: Define EP0_REGNUM and EP1_REGNUM.
572 2018-01-09 Sudakshina Das <sudi.das@arm.com>
574 * config/aarch64/aarch64-option-extensions.def: Define
575 AARCH64_OPT_EXTENSION for memtag, rng, sb, ssbs and predres.
576 * gcc/config/aarch64/aarch64.h (AARCH64_FL_RNG): New.
577 (AARCH64_FL_MEMTAG, ARCH64_FL_SB, AARCH64_FL_SSBS): New.
578 (AARCH64_FL_PREDRES): New.
579 (AARCH64_FL_FOR_ARCH8_5): Add AARCH64_FL_SB, AARCH64_FL_SSBS and
580 AARCH64_FL_PREDRES by default.
581 * gcc/doc/invoke.texi: Document rng, memtag, sb, ssbs and predres.
583 2018-01-09 Sudakshina Das <sudi.das@arm.com>
585 * config/aarch64/aarch64-arches.def: Define AARCH64_ARCH for
587 * gcc/config/aarch64/aarch64.h (AARCH64_FL_V8_5): New.
588 (AARCH64_FL_FOR_ARCH8_5, AARCH64_ISA_V8_5): New.
589 * gcc/doc/invoke.texi: Document ARMv8.5-A.
591 2019-01-09 Alejandro Martinez <alejandro.martinezvicente@arm.com>
593 * config/aarch64/aarch64-sve.md (copysign<mode>3): New define_expand.
594 (xorsign<mode>3): Likewise.
596 2019-01-09 Jelinek <jakub@redhat.com>
599 * tree.c (initializer_each_zero_or_onep) <case VECTOR_CST>: Use
600 vector_cst_elt instead of VECTOR_CST_ENCODED_ELT.
602 PR rtl-optimization/88331
603 * function.c (assign_stack_local_1): Don't set dynamic_align_addr if
604 not currently_expanding_to_rtl.
606 2019-01-09 Eric Botcazou <ebotcazou@adacore.com>
608 * doc/invoke.texi (-Os): Remove trailing spaces.
609 (-finline-functions): Remove reference to -O2.
611 2019-01-08 Jakub Jelinek <jakub@redhat.com>
613 PR rtl-optimization/79593
614 * config/i386/i386.md (reg = mem; mem = reg): New define_peephole2.
616 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Delegitimize
617 UNSPEC_FUSION_GPR to its argument. Formatting fixes.
619 2019-01-08 Eric Botcazou <ebotcazou@adacore.com>
622 * config/sparc/sparc.c (function_arg_slotno): Set *PPREGNO & *PPADDING
626 * config/sparc/sparc.c (sparc_delegitimize_address): Deal with naked
627 UNSPECs and UNSPEC_MOVE_GOTDATA specifically.
629 2019-01-08 H.J. Lu <hongjiu.lu@intel.com>
632 * config/i386/i386.c (ix86_avx_u128_mode_exit): Call
633 ix86_avx_u128_mode_entry.
635 2019-01-08 Martin Liska <mliska@suse.cz>
637 PR tree-optimization/88753
638 * tree-switch-conversion.c (switch_conversion::build_one_array):
639 Come up with local variable constructor. Convert first to
640 type of constructor values.
642 2019-01-08 Richard Biener <rguenther@suse.de>
644 PR tree-optimization/86554
645 * tree-ssa-sccvn.c (eliminate_dom_walker, rpo_elim,
646 rpo_avail): Move earlier.
647 (visit_nary_op): When value-numbering to expressions
648 with different overflow behavior make sure there's an
649 available expression on the path.
651 2019-01-08 Sam Tebbs <sam.tebbs@arm.com>
653 * config/aarch64/aarch64.c (BRANCH_PROTECT_STR_MAX,
654 aarch64_parse_branch_protection,
655 struct aarch64_branch_protect_type,
656 aarch64_handle_no_branch_protection,
657 aarch64_handle_standard_branch_protection,
658 aarch64_validate_mbranch_protection,
659 aarch64_handle_pac_ret_protection,
660 aarch64_handle_attr_branch_protection,
661 accepted_branch_protection_string,
662 aarch64_pac_ret_subtypes,
663 aarch64_branch_protect_types,
664 aarch64_handle_pac_ret_leaf): Define.
665 (aarch64_override_options_after_change_1, aarch64_override_options):
666 Add check for accepted_branch_protection_string.
667 (aarch64_option_save): Save accepted_branch_protection_string.
668 (aarch64_option_restore): Save accepted_branch_protection_string.
669 * config/aarch64/aarch64.c (aarch64_attributes): Add branch-protection.
670 * config/aarch64/aarch64.opt: Add mbranch-protection. Deprecate
671 msign-return-address.
672 * doc/invoke.texi: Add mbranch-protection.
674 2019-01-08 Alan Modra <amodra@gmail.com>
677 * genattrtab.c (max_attr_value, min_attr_value, or_attr_value):
678 Delete "unknownp" parameter. Adjust callers. Handle
679 CONST_INT, PLUS, MINUS, and MULT.
680 (attr_value_aligned): Renamed from or_attr_value.
681 (min_attr_value): Return INT_MIN for unhandled rtl case..
682 (min_fn): ..and translate to INT_MAX here.
683 (write_length_unit_log): Modify to cope without "unknown".
684 (write_attr_value): Handle IF_THEN_ELSE.
686 2019-01-07 Richard Sandiford <richard.sandiford@arm.com>
688 * tree-vect-stmts.c (vectorizable_store): Don't use the dataref_offset
689 optimization for masked stores.
691 2019-01-07 Richard Sandiford <richard.sandiford@arm.com>
694 * tree-vect-loop.c (get_initial_defs_for_reduction): Pass the
695 output vector directly to duplicate_and_interleave instead of
696 going through a temporary. Postpone insertion of ctor_seq to
699 2019-01-07 Richard Earnshaw <rearnsha@arm.com>
702 * config/aarch64/aarch64.c (aarch64_expand_subvti): New parameter
703 unsigned_p. Handle signed and unsigned overflow correction as
705 * config/aarch64/aarch64-protos.h (aarch64_expand_subvti): Update
707 * config/aarch64/aarch64.md (addv<mode>4): Use aarch64_plus_operand
709 (add<mode>3_compareV_imm): Make this callable for expanding.
710 (subv<GPI:mode>4): Use register_operand for operand 1. Use
711 aarch64_plus_operand for operand 2.
712 (subv<GPI:mode>_insn): New insn pattern.
713 (subv<GPI:mode>_imm): Likewise.
714 (negv<GPI:mode>3): New expand pattern.
715 (negv<GPI:mode>_insn): New insn pattern.
716 (negv<GPI:mode>_cmp_only): Likewise.
717 (cmpv<GPI:mode>_insn): Likewise.
718 (subvti4): Use register_operand for operand 1. Update call to
719 aarch64_expand_subvti.
720 (usubvti4): Likewise.
721 (negvti3): New expand pattern.
722 (negdi_carryout): New insn pattern.
723 (negvdi_carryinV): New insn pattern.
724 (sub<mode3>_compare1_imm): Delete named insn pattern, make anonymous
725 version the named version.
726 (peepholes to convert to sub<mode3>_compare1_imm): Adjust order of
728 (usub<GPI:mode>3_carryinC, usub<GPI:mode>3_carryinC_z1): New insn
730 (usub<GPI:mode>3_carryinC_z2, usub<GPI:mode>3_carryinC): New insn
732 (sub<mode>3_carryinCV, sub<mode>3_carryinCV_z1_z2): Delete.
733 (sub<mode>3_carryinCV_z1, sub<mode>3_carryinCV_z2): Delete.
734 (sub<mode>3_carryinCV): Delete.
735 (sub<GPI:mode>3_carryinV): New expand pattern.
736 sub<mode>3_carryinV, sub<mode>3_carryinV_z2): New insn patterns.
738 2019-01-07 Richard Biener <rguenther@suse.de>
740 * tree-ssa-uncprop.c (ssa_equip_hash_traits): Remove in favor
741 of tree_operand_hash.
743 2019-01-07 Richard Sandiford <richard.sandiford@arm.com>
745 PR tree-optimization/88598
746 * tree.h (single_nonzero_element): Declare.
747 * tree.c (single_nonzero_element): New function.
748 * match.pd: Fold certain reductions of X & CST to X[I] & CST[I]
749 if I is the only nonzero element of CST.
751 2019-01-07 Richard Sandiford <richard.sandiford@arm.com>
753 PR tree-optimization/88598
754 * tree.h (initializer_each_zero_or_onep): Declare.
755 * tree.c (initializer_each_zero_or_onep): New function.
756 (signed_or_unsigned_type_for): Handle float types too.
757 (unsigned_type_for, signed_type_for): Update comments accordingly.
758 * match.pd: Fold x * { 0 or 1, 0 or 1, ...} to
759 x & { 0 or -1, 0 or -1, ... }.
761 2019-01-07 Jonathan Wakely <jwakely@redhat.com>
763 * doc/install.texi: Replace references to x86_64-unknown-linux-gnu
764 with x86_64-pc-linux-gnu.
766 2019-01-07 Tom de Vries <tdevries@suse.de>
769 * config/nvptx/nvptx.c (has_vector_partitionable_routine_calls_p): New
771 (nvptx_goacc_validate_dims): Force vl32 if calling vector-partitionable
774 2019-01-07 Jakub Jelinek <jakub@redhat.com>
776 * config/i386/sse.md (vec_extract<mode><ssehalfvecmodelower>): Use
777 V_256_512 iterator instead of V_512 and TARGET_AVX instead of
778 TARGET_AVX512F as condition.
781 * dwarf2out.c (const_ok_for_output_1): Remove redundant call to
782 const_not_ok_for_debug_p target hook.
783 (mem_loc_descriptor) <case UNSPEC>: Only call const_ok_for_output_1
784 on UNSPEC and subexpressions thereof if all subexpressions of the
785 UNSPEC are CONSTANT_P.
787 PR tree-optimization/88676
788 * tree-ssa-phiopt.c (two_value_replacement): New function.
789 (tree_ssa_phiopt_worker): Call it.
792 * cfgexpand.c (expand_stack_vars): Only align prev_offset to
793 ASAN_MIN_RED_ZONE_SIZE, not to maximum of that and alignb.
796 * tree-vect-generic.c: Include insn-config.h and recog.h.
797 (expand_vector_piecewise): Add defaulted ret_type argument,
798 if non-NULL, use that in preference to type for the result type.
799 (expand_vector_parallel): Formatting fix.
800 (do_vec_conversion, do_vec_narrowing_conversion,
801 expand_vector_conversion): New functions.
802 (expand_vector_operations_1): Call expand_vector_conversion
803 for VEC_CONVERT ifn calls.
804 * internal-fn.def (VEC_CONVERT): New internal function.
805 * internal-fn.c (expand_VEC_CONVERT): New function.
806 * fold-const-call.c (fold_const_vec_convert): New function.
807 (fold_const_call): Use it for CFN_VEC_CONVERT.
808 * doc/extend.texi (__builtin_convertvector): Document.
810 2019-01-07 Tom de Vries <tdevries@suse.de>
812 * config/nvptx/nvptx-protos.h (nvptx_output_red_partition): Declare.
813 * config/nvptx/nvptx.c (vector_red_size, vector_red_align,
814 vector_red_partition, vector_red_sym): New global variables.
815 (nvptx_option_override): Initialize vector_red_sym.
816 (nvptx_declare_function_name): Restore red_partition register.
817 (nvptx_file_end): Emit code to declare the vector reduction variables.
818 (nvptx_output_red_partition): New function.
819 (nvptx_expand_shared_addr): Add vector argument. Use it to handle
820 large vector reductions.
821 (enum nvptx_builtins): Add NVPTX_BUILTIN_VECTOR_ADDR.
822 (nvptx_init_builtins): Add VECTOR_ADDR.
823 (nvptx_expand_builtin): Update call to nvptx_expand_shared_addr.
824 Handle nvptx_expand_shared_addr.
825 (nvptx_get_shared_red_addr): Add vector argument and handle large
827 (nvptx_goacc_reduction_setup): Add offload_attrs argument and handle
829 (nvptx_goacc_reduction_init): Likewise.
830 (nvptx_goacc_reduction_fini): Likewise.
831 (nvptx_goacc_reduction_teardown): Likewise.
832 (nvptx_goacc_reduction): Update calls to nvptx_goacc_reduction_{setup,
834 (nvptx_init_axis_predicate): Initialize vector_red_partition.
835 (nvptx_set_current_function): Init vector_red_partition.
836 * config/nvptx/nvptx.md (UNSPECV_RED_PART): New unspecv.
837 (nvptx_red_partition): New insn.
838 * config/nvptx/nvptx.h (struct machine_function): Add red_partition.
840 2019-01-07 Tom de Vries <tdevries@suse.de>
843 * config/nvptx/nvptx.c (nvptx_process_pars): Don't emit barriers for
846 2019-01-07 Tom de Vries <tdevries@suse.de>
848 * config/nvptx/nvptx.c (oacc_bcast_partition): Declare.
849 (nvptx_option_override): Init oacc_bcast_partition.
850 (nvptx_init_oacc_workers): New function.
851 (nvptx_declare_function_name): Call nvptx_init_oacc_workers.
852 (nvptx_needs_shared_bcast): New function.
853 (nvptx_find_par): Generalize to enable vectors to use shared-memory
855 (nvptx_shared_propagate): Initialize vector bcast partition and
856 synchronization state.
857 (nvptx_single): Generalize to enable vectors to use shared-memory
859 (nvptx_process_pars): Likewise.
860 (nvptx_set_current_function): Initialize oacc_broadcast_partition.
861 * config/nvptx/nvptx.h (struct machine_function): Add
862 bcast_partition and sync_bar members.
864 2019-01-07 Tom de Vries <tdevries@suse.de>
866 * config/nvptx/nvptx.c (nvptx_welformed_vector_length_p)
867 (nvptx_apply_dim_limits): New function.
868 (nvptx_goacc_validate_dims_1): Allow PTX_MAX_VECTOR_LENGTH larger than
871 2019-01-07 Tom de Vries <tdevries@suse.de>
873 * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): Move warnings to
876 2019-01-07 Tom de Vries <tdevries@suse.de>
878 * config/nvptx/nvptx.c (PTX_VECTOR_LENGTH): Remove.
879 (PTX_DEFAULT_VECTOR_LENGTH, PTX_MAX_VECTOR_LENGTH): Define.
880 (nvptx_goacc_validate_dims_1, nvptx_dim_limit)
881 (nvptx_goacc_reduction_fini): Use PTX_DEFAULT_VECTOR_LENGTH,
882 PTX_MAX_VECTOR_LENGTH and PTX_WARP_SIZE instead of PTX_VECTOR_LENGTH.
884 2019-01-07 Tom de Vries <tdevries@suse.de>
886 * config/nvptx/nvptx.c (nvptx_goacc_validate_dims): Add asserts.
888 2019-01-07 Tom de Vries <tdevries@suse.de>
890 * omp-offload.c (oacc_get_min_dim): New function.
891 * omp-offload.h (oacc_get_min_dim): Declare.
893 2018-12-26 Mateusz B <mateuszb@poczta.onet.pl>
896 * config/i386/i386.c (function_value_ms_64): Return small sturct in
897 AX_REG and float/double in FIRST_SSE_REG for 4 or 8 byte modes.
899 2019-01-05 Jan Hubicka <hubicka@ucw.cz>
903 2017-05-22 Jan Hubicka <hubicka@ucw.cz>
905 * ipa-inline.c (edge_badness): Use inlined_time instead of
906 inline_summaries->get.
908 2019-01-05 Jan Hubicka <hubicka@ucw.cz>
910 * opts.c (enable_fdo_optimizations): Enable
911 version-loops-for-strides, loop-interchange, unrol-and-jam
912 and tree-loop-distribution.
913 * invoke.texi: Document newly enabled options.
915 2019-01-05 Jan Hubicka <hubicka@ucw.cz>
917 * doc/invoke.texi (max-inline-insns-small): New parameters.
918 * ipa-inline.c (want_early_inline_function_p): simplify.
919 (want_inline_small_function_p): Fix pasto from previous patch;
920 use max-inline-insns-small bound.
921 * params.def (max-inline-insns-small): New param.
922 * ipa-fnsummary.c (analyze_function_body): Initialize time/size
925 2019-01-05 Jan Hubicka <hubicka@ucw.cz>
927 * doc/invoke.texi: Document max-inline-insns-size,
928 uninlined-function-insns, uninlined-function-time,
929 uninlined-thunk-insns and uninlined-thunk-time.
930 * params.def: Add max-inline-insns-size,
931 uninlined-function-insns, uninlined-function-time,
932 uninlined-thunk-insns and uninlined-thunk-time.
933 * ipa-fnsummary.c (compute_fn_summary, analyze_function_body): Use
935 * ipa-inline.c (can_inline_edge_by_limits_p,
936 want_inline_small_function_p): Use new parameters.
938 2019-01-05 Jan Hubicka <hubicka@ucw.cz>
940 * ipa-fnsummary.c (analyze_function_body): Fix accounting of time.
942 2019-01-05 Jakub Jelinek <jakub@redhat.com>
946 * expr.c (expand_assignment): For calls returning VLA structures
947 if to_rtx is not a MEM, force it into a stack temporary.
950 * dwarf2out.c (const_ok_for_output_1): Reject MINUS that contains
951 SYMBOL_REF, CODE_LABEL or UNSPEC in subexpressions of second argument.
952 Reject PLUS that contains SYMBOL_REF, CODE_LABEL or UNSPEC in
953 subexpressions of both operands.
954 (mem_loc_descriptor): Handle UNSPEC if target hook acks it and all the
955 subrtxes are CONSTANT_P.
956 * config/i386/i386.c (ix86_const_not_ok_for_debug_p): Revert
959 2019-01-04 Jan Hubicka <hubicka@ucw.cz>
961 * params.def (hot-bb-count-ws-permille): Set to 990.
963 2019-01-04 Martin Sebor <msebor@redhat.com>
966 * attribs.c (decls_mismatched_attributes): Avoid warning for attribute
969 2019-01-04 Martin Sebor <msebor@redhat.com>
972 * doc/extend.texi (attribute alloc_align, alloc_size): Update.
974 2019-01-04 Jakub Jelinek <jakub@redhat.com>
976 * gdbinit.in: Turn off pagination for the skip commands, restore
977 it to previous state afterwards.
979 2019-01-04 Sam Tebbs <sam.tebbs@arm.com>
982 * gcc.target/aarch64/combine_bfxil.c: Change scan-assembler-times bfxil
985 2019-01-04 Jakub Jelinek <jakub@redhat.com>
988 * config/i386/i386.c (ix86_expand_divmod_libfunc): Use mode instead
989 of GET_MODE (opN) as modes of the libcall arguments.
991 2019-01-04 Jan Beulich <jbeulich@suse.com>
994 (<avx512>_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>,
995 <avx512>_cmp<mode>3<mask_scalar_merge_name>,
996 <avx512>_ucmp<mode>3<mask_scalar_merge_name>,
997 <avx512>_ucmp<mode>3<mask_scalar_merge_name>,
998 avx512f_vmcmp<mode>3<round_saeonly_name>,
999 avx512f_vmcmp<mode>3_mask<round_saeonly_name>,
1000 avx512f_maskcmp<mode>3,
1001 <avx512>_cvt<ssemodesuffix>2mask<mode>,
1002 <avx512>_cvt<ssemodesuffix>2mask<mode>,
1003 *<avx512>_cvtmask2<ssemodesuffix><mode>,
1004 *<avx512>_cvtmask2<ssemodesuffix><mode>,
1005 <avx512>_eq<mode>3<mask_scalar_merge_name>_1,
1006 <avx512>_eq<mode>3<mask_scalar_merge_name>_1,
1007 <avx512>_gt<mode>3<mask_scalar_merge_name>,
1008 <avx512>_gt<mode>3<mask_scalar_merge_name>,
1009 <avx512>_testm<mode>3<mask_scalar_merge_name>,
1010 <avx512>_testnm<mode>3<mask_scalar_merge_name>,
1011 *<avx512>_testm<VI1248_AVX512VLBW:mode>3_zext,
1012 *<avx512>_testm<VI1248_AVX512VLBW:mode>3_zext_mask,
1013 *<avx512>_testnm<VI1248_AVX512VLBW:mode>3_zext,
1014 *<avx512>_testnm<VI1248_AVX512VLBW:mode>3_zext_mask,
1015 avx512cd_maskb_vec_dup<mode>,
1016 avx512cd_maskw_vec_dup<mode>,
1017 avx512dq_fpclass<mode><mask_scalar_merge_name>,
1018 avx512dq_vmfpclass<mode>,
1019 avx512vl_vpshufbitqmb<mode><mask_scalar_merge_name>): Use =k
1022 2019-01-03 Martin Sebor <msebor@redhat.com>
1024 PR tree-optimization/88659
1025 * calls.c (maybe_warn_nonstring_arg): Avoid assuming maxlen is set.
1027 2019-01-03 Aaron Sawdey <acsawdey@linux.ibm.com>
1029 * config/rs6000/rs6000-string.c (expand_block_move): Don't use
1030 unaligned vsx and avoid lxvd2x/stxvd2x.
1031 (gen_lvx_v4si_move): New function.
1033 2019-01-03 Tom de Vries <tdevries@suse.de>
1035 * config/nvptx/nvptx.c (MACH_VECTOR_LENGTH, MACH_MAX_WORKERS): Define.
1036 (init_axis_dim, nvptx_mach_max_workers, nvptx_mach_vector_length): New
1038 * config/nvptx/nvptx.h (struct machine_function): Add axis_dims.
1040 2019-01-03 Tom de Vries <tdevries@suse.de>
1042 * config/nvptx/nvptx.c (struct offload_attrs): New.
1043 (populate_offload_attrs): New function. Factor mask extraction out of
1044 nvptx_reorg. Add extraction of dimensions.
1045 (nvptx_reorg): Use populate_offload_attrs.
1047 2019-01-03 Tom de Vries <tdevries@suse.de>
1049 * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): Add early-out
1050 cases for oacc_min_dims_p and routine_p. Add asserts for
1051 oacc_default_dims_p and offload_region_p.
1053 2019-01-03 Tom de Vries <tdevries@suse.de>
1055 * config/nvptx/nvptx.c (nvptx_goacc_validate_dims_1): New function,
1057 (nvptx_goacc_validate_dims): ... here.
1059 2019-01-03 Jan Hubicka <hubicka@ucw.cz>
1061 PR tree-optimization/85574
1062 * tree-ssa-uncprop.c (struct equiv_hash_elt): Remove unused
1064 (struct ssa_equip_hash_traits): Declare.
1065 (val_ssa_equiv): Use custom hash traits using operand_equal_p.
1067 2019-01-03 Jakub Jelinek <jakub@redhat.com>
1070 * dwarf2out.c (modified_type_die): If type is equal to sizetype,
1071 change it to qualified_type.
1073 2019-01-03 Jan Hubicka <hubicka@ucw.cz>
1075 * ipa-utils.c (scale_ipa_profile_for_fn): Break out from ...
1076 (ipa_merge_profiles): ... here; do not ICE on thunks and aliases.
1078 2019-01-02 Martin Sebor <msebor@redhat.com>
1079 Jeff Law <law@redhat.com>
1081 * gimple-fold.c (strlen_range_kind): Remove SRK_LENRANGE_2.
1082 (get_range_strlen_tree): Update appropriately.
1084 * gimple-fold.h (get_range_strlen): Drop unused last argument.
1086 * gimple-fold.c (gimple_fold_builtin_strlen): Use set_strlen_range
1087 rather than set_range_info.
1088 * tree-ssa-strlen.c (set_strlen_range): Extracted from
1089 maybe_set_strlen_range. Handle potentially boundary crossing
1090 cases more conservatively.
1091 (maybe_set_strlen_range): Parts refactored into set_strlen_range.
1092 Call set_strlen_range.
1093 * tree-ssa-strlen.h (set_strlen_range): Add prototype.
1096 * gimple-fold.c (get_range_strlen): Update prototype to no longer
1097 need the flexp argument.
1098 (get_range_strlen_tree): Drop flexp argument. Drop flexp argument
1099 from calls to get_range_strlen. Update comments. Just update
1100 VAL for an unterminated const char array and let the reset of the
1101 code handle it normally. No longer try to set *flexp. Adjust
1103 (get_range_strlen): Update for the new get_range_strlen API.
1104 (get_maxval_strlen): Similarly.
1105 (gimple_fold_builtin_strlen): Handle update meaning of return value
1106 from get_range_strlen.
1107 * gimple-ssa-sprintf.c (get_string_length): Update for the new
1108 get_range_strlen API.
1110 2019-01-02 Jan Hubicka <hubicka@ucw.cz>
1113 * varpool.c (varpool_node::ctor_useable_for_folding_p): Also return
1114 false at WPA time when body was removed.
1116 2019-01-02 Martin Liska <mliska@suse.cz>
1118 PR tree-optimization/88650
1119 * predict.c (set_even_probabilities): Calculate probability
1120 remainer only when really used.
1122 2019-01-02 Richard Biener <rguenther@suse.de>
1125 * tree-data-ref.c (analyze_subscript_affine_affine): Use
1126 widest_ints when mangling max_stmt_execution results.
1128 2019-01-02 Richard Biener <rguenther@suse.de>
1130 PR tree-optimization/88621
1131 * tree-ssa-loop-im.c (gather_mem_refs_stmt): Fix pastos, avoid
1132 bitfields when canoncalizing.
1134 2019-01-02 Richard Biener <rguenther@suse.de>
1137 * config/i386/x86-tune-costs.h (intel_cost): Adjust
1138 cost of cheap SSE instruction.
1140 2019-01-02 Richard Biener <rguenther@suse.de>
1143 * ipa-icf.h (sem_item_optimizer::sort_congruence_split): Declare.
1144 * ipa-icf.c (sem_item_optimizer::sort_congruence_split): New
1146 (sem_item_optimizer::do_congruence_step_f): Sort the congruence
1147 set after UIDs before splitting them.
1149 2019-01-01 Martin Sebor <msebor@redhat.com>
1150 Jeff Law <law@redhat.com>
1152 * gimple-fold.c (get_range_strlen_tree): Record if the computed
1153 length is optimistic. If it is, then arrange to compute the
1154 conservative length as well.
1156 * gimple-fold.h (get_range_strlen): Update prototype.
1157 * builtins.c (check_access): Update call to get_range_strlen to use
1158 c_strlen_data pointer. Change various variable accesses to instead
1159 pull data from the c_strlen_data structure.
1160 (check_strncat_sizes, expand_builtin_strncat): Likewise.
1161 * calls.c (maybe_warn_nonstring_arg): Likewise.
1162 * tree-ssa-strlen.c (maybe_diag_stxncpy_trunc): Likewise. Reset
1163 minimum length if maximum lengh is unknown.
1164 * gimple-ssa-sprintf.c (get_string_length): Likewise. Drop code
1165 that used c_strlen, it's no longer needed. Restructure slightly.
1166 (format_string): Set unlikely range appropriately.
1167 * gimple-fold.c (get_range_strlen): Update comments. Fix minor
1169 (get_range_strlen): Accept c_strlen_data pointer for external
1170 call sites as well. Pass through to call to internal get_range_strlen.
1171 Adjust minlen, maxlen and maxbound as needed.
1172 (get_maxval_strlen): Update comments.
1173 (gimple_fold_builtin_strlen): Update call to get_range_strlen
1174 to use c_strlen_data pointer. Change variable accesses to instead
1175 use c_strlen_data data members.
1177 * gimple-fold.c (get_range_strlen): Update prototype.
1178 (get_range_strlen_tree): Update prototype. Drop minlen/maxlen
1179 local variables. Use pdata to return information to caller.
1180 Update calls to get_range_strlen. Update pdata->maxbound.
1181 (get_range_strlen -- static version): Similarly.
1182 (get_range_strlen -- extern version): Update for internal
1183 get_range_strlen API change. Convert to external data format.
1184 (get_maxval_strlen): Similarly.
1186 2019-01-01 Jan Hubicka <hubicka@ucw.cz>
1188 * coverage.c (get_coverage_counts): Use current_function_decl.
1189 * profile.c (read_thunk_profile): New function.
1190 (branch_prob): Add THUNK parameter.
1191 * tree-profile.c (tree_profiling): Handle thunks.
1192 * value-prof.c (init_node_map): Handle thunks.
1193 * value-prof.h (branch_prob): Upate prototype.
1194 (read_thunk_profile): Declare.
1196 2019-01-01 Jakub Jelinek <jakub@redhat.com>
1198 Update copyright years.
1200 * gcc.c (process_command): Update copyright notice dates.
1201 * gcov-dump.c (print_version): Ditto.
1202 * gcov.c (print_version): Ditto.
1203 * gcov-tool.c (print_version): Ditto.
1204 * gengtype.c (create_file): Ditto.
1205 * doc/cpp.texi: Bump @copying's copyright year.
1206 * doc/cppinternals.texi: Ditto.
1207 * doc/gcc.texi: Ditto.
1208 * doc/gccint.texi: Ditto.
1209 * doc/gcov.texi: Ditto.
1210 * doc/install.texi: Ditto.
1211 * doc/invoke.texi: Ditto.
1213 Copyright (C) 2019 Free Software Foundation, Inc.
1215 Copying and distribution of this file, with or without modification,
1216 are permitted in any medium without royalty provided the copyright
1217 notice and this notice are preserved.