1 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
2 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4 * config/arm/t-aprofile: New file.
5 * config.gcc: Handle --with-multilib-list option.
7 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
9 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
12 2013-10-15 Richard Biener <rguenther@suse.de>
14 * tree-loop-distribution.c (build_empty_rdg): Inline into
16 (rdg_flag_vertex): Inline into single user.
17 (rdg_flag_vertex_and_dependent): Likewise.
18 (build_rdg_partition_for_vertex): Remove processed bitmap.
19 (rdg_build_partitions): Simplify.
21 2013-10-15 Richard Biener <rguenther@suse.de>
23 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
24 Restructure forwarding through conversions and copies to
25 avoid performing copy-propagation the wrong way. Adjust
26 recursion invocations.
27 (forward_propagate_addr_expr): Add argument stating if we
28 are recursing from a single-use.
29 (ssa_forward_propagate_and_combine): Adjust.
31 2013-10-14 David Malcolm <dmalcolm@redhat.com>
33 * dumpfile.h (gcc::dump_manager): New class, to hold state
34 relating to dumpfile management.
35 (get_dump_file_name): Remove in favor of method of dump_manager.
36 (dump_initialized_p): Likewise.
37 (dump_start): Likewise.
38 (dump_finish): Likewise.
39 (dump_switch_p): Likewise.
40 (dump_register): Likewise.
41 (get_dump_file_info): Likewise.
42 * context.c (gcc::context::context): Construct the dump_manager
44 * context.h (gcc::context::get_dumps): New.
45 (gcc::context::m_dumps): New.
46 * coverage.c (coverage_init): Port to dump_manager API.
47 * dumpfile.c (extra_dump_files): Convert to field of
49 (extra_dump_files_in_use): Likewise.
50 (extra_dump_files_alloced): Likewise.
51 (gcc::dump_manager::dump_manager): New.
52 (dump_register): Convert to...
53 (gcc::dump_manager::dump_register): ...method, replacing
54 function-static next_dump with m_next_dump field.
55 (get_dump_file_info): Convert to...
56 (gcc::dump_manager::get_dump_file_info): ...method.
57 (get_dump_file_name): Convert to...
58 (gcc::dump_manager::get_dump_file_name): ...method.
59 (dump_start): Convert to...
60 (gcc::dump_manager::dump_start): ...method.
61 (dump_finish): Convert to...
62 (gcc::dump_manager::dump_finish): ...method.
63 (dump_begin): Replace body with...
64 (gcc::dump_manager::dump_begin): ...new method.
65 (dump_phase_enabled_p): Convert to...
66 (gcc::dump_manager::dump_phase_enabled_p): ...method.
67 (dump_phase_enabled_p): Convert to...
68 (gcc::dump_manager::dump_phase_enabled_p): ...method.
69 (dump_initialized_p): Convert to...
70 (gcc::dump_manager::dump_initialized_p): ...method.
71 (dump_flag_name): Replace body with...
72 (gcc::dump_manager::dump_flag_name): ...new method.
73 (dump_enable_all): Convert to...
74 (gcc::dump_manager::dump_enable_all): ...new method.
75 (opt_info_enable_passes): Convert to...
76 (gcc::dump_manager::opt_info_enable_passes): ...new method.
77 (dump_switch_p_1): Convert to...
78 (gcc::dump_manager::dump_switch_p_1): ...new method.
79 (dump_switch_p): Convert to...
80 (gcc::dump_manager::dump_switch_p): ...new method.
81 (opt_info_switch_p): Port to dump_manager API.
82 (enable_rtl_dump_file): Likewise.
83 * opts-global.c (handle_common_deferred_options): Port to new
85 * passes.c (pass_manager::finish_optimization_passes): Likewise.
86 (pass_manager::register_one_dump_file): Likewise.
87 (pass_manager::register_pass): Likewise.
88 (pass_init_dump_file): Likewise.
89 (pass_fini_dump_file): Likewise.
90 * statistics.c (statistics_early_init): Likewise.
92 2013-10-14 Richard Biener <rguenther@suse.de>
94 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
95 iterative_hash_canonical_type, gimple_canonical_type_hash,
96 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
97 gimple_register_canonical_type, print_gimple_types_stats,
98 free_gimple_type_tables): Move to lto/lto.c
99 (gt-gimple.h): Do not include.
100 * gimple.h (gimple_register_canonical_type,
101 print_gimple_types_stats, free_gimple_type_tables): Remove.
102 * Makefile.in (GTFILES): Remove gimple.c.
104 2013-10-14 Travis Snoozy <quandary@remstate.com>
107 * config/msp430/msp430.c (msp430_option_override): Correct thinko
108 scanning for msp430x targets.
110 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
113 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
114 (registers_ok_for_ldd_peep): Move around.
115 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
116 * config/sparc/sparc.md (widening peepholes): Use it.
118 2013-10-14 Richard Biener <rguenther@suse.de>
122 * gimple.c (iterative_hash_canonical_type): Make sure to
123 record the hash into the correct hashtable slot.
125 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
127 PR rtl-optimization/58662
128 * combine.c (try_combine): Take into account death nodes on I2 when
129 splitting a PARALLEL of two independent SETs. Fix dump message.
131 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
134 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
135 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
136 * config/sh/t-sh (sh_treg_combine.o): New entry.
137 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
138 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
139 (register_sh_passes): New function. Register sh_treg_combine pass.
140 (sh_option_override): Invoke it.
141 (sh_canonicalize_comparison): Handle op0_preserve_value.
142 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
143 opportunities. Canonicalize branch condition.
144 (nott): Allow only if pseudos can be created for non-SH2A.
146 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
149 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
150 (ix86_expand_movmem): Replace copy_addr_to_reg with
151 ix86_copy_addr_to_reg.
152 (ix86_expand_setmem): Likewise.
154 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
156 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
157 provided by ix86_fp_compare_mode instead of CCFPUmode.
159 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
161 * config/aarch64/arm_neon.h
162 (vtbx<1,3>_<psu>8): Fix register constriants.
164 2013-10-11 Jeff Law <law@redhat.com>
166 PR tree-optimization/58640
167 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
168 threading paths that cross over two loop entry points.
170 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
172 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
173 handle vector float as well.
174 (*vsx_le_perm_load_v4si): Likewise.
175 (*vsx_le_perm_store_v2di): Likewise.
176 (*vsx_le_perm_store_v4si): Likewise.
178 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
180 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
181 directly to circumvent subtract from splat{31} workaround.
182 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
184 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
185 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
186 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
187 (altivec_vperm_<mode>): Convert to define_insn_and_split to
188 separate big and little endian logic.
189 (*altivec_vperm_<mode>_internal): New define_insn.
190 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
191 separate big and little endian logic.
192 (*altivec_vperm_<mode>_uns_internal): New define_insn.
193 (vec_permv16qi): Add little endian logic.
195 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
197 * doc/extend.texi (returns_nonnull): Remove arguments.
199 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
200 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
201 Sergey Lega <sergey.s.lega@intel.com>
202 Anna Tikhonova <anna.tikhonova@intel.com>
203 Ilya Tocar <ilya.tocar@intel.com>
204 Andrey Turetskiy <andrey.turetskiy@intel.com>
205 Ilya Verbin <ilya.verbin@intel.com>
206 Kirill Yukhin <kirill.yukhin@intel.com>
207 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
209 * config/i386/sse.md (VI48F_256_512): New.
210 (avx2_permvar<mode>): Change to ...
211 (<avx2_avx512f>_permvar<mode>): This.
213 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
214 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
215 Sergey Lega <sergey.s.lega@intel.com>
216 Anna Tikhonova <anna.tikhonova@intel.com>
217 Ilya Tocar <ilya.tocar@intel.com>
218 Andrey Turetskiy <andrey.turetskiy@intel.com>
219 Ilya Verbin <ilya.verbin@intel.com>
220 Kirill Yukhin <kirill.yukhin@intel.com>
221 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
223 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
224 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
225 * config/i386/sse.md (VI4_AVX): New.
226 (sf2simodelower): Ditto.
227 (sse2_cvtps2dq): Change to ...
228 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
230 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
231 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
232 Sergey Lega <sergey.s.lega@intel.com>
233 Anna Tikhonova <anna.tikhonova@intel.com>
234 Ilya Tocar <ilya.tocar@intel.com>
235 Andrey Turetskiy <andrey.turetskiy@intel.com>
236 Ilya Verbin <ilya.verbin@intel.com>
237 Kirill Yukhin <kirill.yukhin@intel.com>
238 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
240 * config/i386/sse.md (V_512): New.
242 (vcond<V_512:mode><VF_512:mode>): Ditto.
243 (vcond<V_512:mode><VI_512:mode>): Ditto.
244 (vcondu<V_512:mode><VI_512:mode>): Ditto.
246 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
247 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
248 Sergey Lega <sergey.s.lega@intel.com>
249 Anna Tikhonova <anna.tikhonova@intel.com>
250 Ilya Tocar <ilya.tocar@intel.com>
251 Andrey Turetskiy <andrey.turetskiy@intel.com>
252 Ilya Verbin <ilya.verbin@intel.com>
253 Kirill Yukhin <kirill.yukhin@intel.com>
254 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
256 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
257 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
259 (fma<mode>4): Removed condition.
261 (fnma<mode>4): Ditto.
262 (fnms<mode>4): Ditto.
263 (fma4i_fmadd_<mode>): Ditto.
264 (*fma_fmadd_<mode>): Ditto.
265 (*fma_fmsub_<mode>): Ditto.
266 (*fma_fnmadd_<mode>): Ditto.
267 (*fma_fnmsub_<mode>): Ditto.
268 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
269 (*fma_fmaddsub_<mode>): Ditto.
270 (*fma_fmsubadd_<mode>): Ditto.
271 (*fmai_fmadd_<mode>): Ditto.
272 (*fmai_fmsub_<mode>): Ditto.
273 (*fmai_fnmadd_<mode>): Ditto.
274 (*fmai_fnmsub_<mode>): Ditto.
276 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
277 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
278 Sergey Lega <sergey.s.lega@intel.com>
279 Anna Tikhonova <anna.tikhonova@intel.com>
280 Ilya Tocar <ilya.tocar@intel.com>
281 Andrey Turetskiy <andrey.turetskiy@intel.com>
282 Ilya Verbin <ilya.verbin@intel.com>
283 Kirill Yukhin <kirill.yukhin@intel.com>
284 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
286 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
287 (VI124_256): Changed to ...
288 (VI124_256_48_512): This.
289 (ssepackmode): Extended with wider modes.
290 (<code><mode>3): Changed iterator.
291 (*avx2_<code><mode>3): Ditto.
292 (vec_pack_trunc_<mode>): Ditto.
294 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
295 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
296 Sergey Lega <sergey.s.lega@intel.com>
297 Anna Tikhonova <anna.tikhonova@intel.com>
298 Ilya Tocar <ilya.tocar@intel.com>
299 Andrey Turetskiy <andrey.turetskiy@intel.com>
300 Ilya Verbin <ilya.verbin@intel.com>
301 Kirill Yukhin <kirill.yukhin@intel.com>
302 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
304 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
305 (VI8F_256_512): Ditto.
306 (abs<mode>2): Changed iterator.
307 (avx2_perm<mode>): Changed to ...
308 (<avx2_avx512f>_perm<mode>): This.
309 (avx2_perm<mode>_1): Changed to ...
310 (<avx2_avx512f>_perm<mode>_1): This.
312 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
313 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
314 Sergey Lega <sergey.s.lega@intel.com>
315 Anna Tikhonova <anna.tikhonova@intel.com>
316 Ilya Tocar <ilya.tocar@intel.com>
317 Andrey Turetskiy <andrey.turetskiy@intel.com>
318 Ilya Verbin <ilya.verbin@intel.com>
319 Kirill Yukhin <kirill.yukhin@intel.com>
320 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
322 * config/i386/sse.md (VI48_AVX512F): New.
323 (VI48_AVX2): Changed to ...
324 (VI48_AVX2_48_AVX512F): This.
325 (avx2_ashrv<mode>): Changed to ...
326 (<avx2_avx512f>_ashrv<mode>): This.
327 (avx2_<shift_insn>v<mode>): Changed to ...
328 (<avx2_avx512f>_<shift_insn>v<mode>): This.
330 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
331 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
332 Sergey Lega <sergey.s.lega@intel.com>
333 Anna Tikhonova <anna.tikhonova@intel.com>
334 Ilya Tocar <ilya.tocar@intel.com>
335 Andrey Turetskiy <andrey.turetskiy@intel.com>
336 Ilya Verbin <ilya.verbin@intel.com>
337 Kirill Yukhin <kirill.yukhin@intel.com>
338 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
340 * config/i386/sse.md (VI4_AVX512F): New.
341 (VI8_AVX2_AVX512F): Ditto.
342 (mul<mode>3): Extended with wider modes.
343 (*<sse4_1_avx2>_mul<mode>3): Ditto.
345 (vec_widen_<s>mult_odd_<mode>): Ditto.
347 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
348 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
349 Sergey Lega <sergey.s.lega@intel.com>
350 Anna Tikhonova <anna.tikhonova@intel.com>
351 Ilya Tocar <ilya.tocar@intel.com>
352 Andrey Turetskiy <andrey.turetskiy@intel.com>
353 Ilya Verbin <ilya.verbin@intel.com>
354 Kirill Yukhin <kirill.yukhin@intel.com>
355 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
357 * config/i386/sse.md (VI2_AVX512F): New.
358 (VI124_AVX512F): Ditto.
359 (sseunpackmode): Extended with wider modes.
360 (sseunpackfltmode): Ditto.
361 (vec_unpacks_float_hi_<mode>): Ditto.
362 (vec_unpacks_float_lo_<mode>): Ditto.
363 (vec_unpacku_float_hi_<mode>): Ditto.
364 (vec_unpacku_float_lo_<mode>): Ditto.
365 (vec_unpacks_lo_<mode>): Ditto.
366 (vec_unpacks_hi_<mode>): Ditto.
367 (vec_unpacku_lo_<mode>): Ditto.
368 (vec_unpacku_hi_<mode>): Ditto.
370 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
371 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
372 Sergey Lega <sergey.s.lega@intel.com>
373 Anna Tikhonova <anna.tikhonova@intel.com>
374 Ilya Tocar <ilya.tocar@intel.com>
375 Andrey Turetskiy <andrey.turetskiy@intel.com>
376 Ilya Verbin <ilya.verbin@intel.com>
377 Kirill Yukhin <kirill.yukhin@intel.com>
378 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
380 * config/i386/i386.md (multdiv): New.
381 (multdiv_mnemonic): Ditto.
382 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
383 (<sse>_vm<multdiv_mnemonic><mode>3): This.
384 (<sse>_vmdiv<mode>3): Removed.
386 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
387 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
388 Sergey Lega <sergey.s.lega@intel.com>
389 Anna Tikhonova <anna.tikhonova@intel.com>
390 Ilya Tocar <ilya.tocar@intel.com>
391 Andrey Turetskiy <andrey.turetskiy@intel.com>
392 Ilya Verbin <ilya.verbin@intel.com>
393 Kirill Yukhin <kirill.yukhin@intel.com>
394 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
396 * config/i386/sse.md (V): Extended with wider modes.
398 (ssehalfvecmode): Ditto.
400 (ssepackfltmode): Ditto.
401 (avx_vec_concat<mode>): Ditto.
402 (V_256_512): New iterator.
403 (VF2_512_256): Ditto.
404 (si2dfmode): New attribute.
405 (si2dfmodelower): Ditto.
407 (concat_tg_mode): Ditto.
408 (floatv4siv4df2): Changed to ...
409 (float<si2dfmodelower><mode>2): This.
410 (avx_cvtps2pd256): Changed to ...
411 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
412 (vec_pack_trunc_v4df): Changed to ...
413 (vec_pack_trunc_<mode>): This.
414 (avx_vpermil<mode>): Changed to ...
415 (<sse2_avx_avx512f>_vpermil<mode>): This.
416 (<fixsuffix>fix_truncv8dfv8si2): New.
417 (vec_pack_sfix_trunc_v8df): Ditto.
418 (avx512f_rndscale<mode>): Ditto.
419 (avx512f_roundpd512): Ditto.
420 (vec_pack_ufix_trunc_<mode>): Updated iterator.
422 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
423 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
424 Sergey Lega <sergey.s.lega@intel.com>
425 Anna Tikhonova <anna.tikhonova@intel.com>
426 Ilya Tocar <ilya.tocar@intel.com>
427 Andrey Turetskiy <andrey.turetskiy@intel.com>
428 Ilya Verbin <ilya.verbin@intel.com>
429 Kirill Yukhin <kirill.yukhin@intel.com>
430 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
432 * config/i386/i386.md (any_fix): New iterator.
433 (fixsuffix): New attribute.
434 * config/i386/sse.md (VF1): Extened with wider modes.
438 (sseintvecmodelower): Ditto.
439 (ssescalarmode): Ditto.
440 (ssescalarnum): Ditto.
443 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
444 (<sse>_rcp<mode>2): Change iterator.
445 (rsqrt<mode>2): Ditto.
446 (<sse>_rsqrt<mode>2): Ditto.
447 (avx2_vec_dup<mode>): Ditto.
448 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
449 (round<mode>2_sfix): Ditto.
450 (avx2_pbroadcast<mode>): Ditto.
451 (*andnot<mode>3): Handle XI mode.
452 (*<code><mode>3): Ditto.
453 (AVXTOSSEMODE): Removed.
454 (avx_vpermil<mode>): Changed to ...
455 (<sse2_avx_avx512f>_vpermil<mode>): This.
457 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
458 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
459 Sergey Lega <sergey.s.lega@intel.com>
460 Anna Tikhonova <anna.tikhonova@intel.com>
461 Ilya Tocar <ilya.tocar@intel.com>
462 Andrey Turetskiy <andrey.turetskiy@intel.com>
463 Ilya Verbin <ilya.verbin@intel.com>
464 Kirill Yukhin <kirill.yukhin@intel.com>
465 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
467 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
469 (<sse>_ucomi): Ditto.
470 (sse_cvtss2siq_2): Ditto.
471 (sse2_cvtsd2si): Ditto.
472 (sse2_cvtsd2siq): Ditto.
473 (sse2_cvttsd2si): Ditto.
474 (sse2_cvttsd2siq): Ditto.
475 (<shift_insn><mode>3): Ditto.
476 (sse2_cvtsi2sdq): Update constraint and prefix.
477 (sse_cvtsi2ss): Update prefix.
478 (sse_cvtsi2ssq): Ditto.
480 2013-10-11 Jakub Jelinek <jakub@redhat.com>
482 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
483 ignore internal calls.
485 2013-10-11 Richard Biener <rguenther@suse.de>
487 * tree-pretty-print.c (dump_generic_node): Allow to dump
488 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
489 (OVF) if TREE_OVERFLOW is set.
491 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
493 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
495 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
496 * gimple.h: Likewise.
498 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
500 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
503 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
505 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
506 * configure: Regenerate.
508 2013-10-11 Jakub Jelinek <jakub@redhat.com>
510 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
511 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
512 formatting fixes, use pp_colon instead of pp_character (..., ':'),
513 similarly pp_right_paren.
514 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
515 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
516 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
517 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
519 (omp_declare_simd_clauses_equal,
520 omp_remove_redundant_declare_simd_attrs): New functions.
521 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
522 (walk_tree_1): Handle new OpenMP 4.0 clauses.
523 * tree.h (OMP_LOOP_CHECK): Define.
524 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
525 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
526 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
527 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
528 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
529 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
530 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
531 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
532 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
533 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
534 OMP_CLAUSE_SIMDLEN_EXPR): Define.
535 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
536 (omp_remove_redundant_declare_simd_attrs): New prototype.
537 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
538 GIMPLE_OMP_TEAMS): New codes.
539 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
540 * omp-low.c (struct omp_context): Add cancel_label and cancellable
542 (target_nesting_level): New variable.
543 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
544 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
545 for collapse > 1 static schedule unless ordered.
546 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
547 (determine_parallel_type): Adjust get_ws_args_for caller.
548 (install_var_field): Handle mask & 4 for double indirection.
549 (scan_sharing_clauses): Ignore shared clause on teams construct.
550 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
551 (create_omp_child_function): If inside target or declare target
552 constructs, set "omp declare target" attribute on the child function.
553 (find_combined_for): New function.
554 (scan_omp_parallel): Handle combined loops.
555 (scan_omp_target, scan_omp_teams): New functions.
556 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
557 restrictions and set ctx->cancellable for cancellable constructs.
558 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
559 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
560 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
561 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
562 (omp_clause_aligned_alignment): New function.
563 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
565 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
566 on teams constructs. Handle user defined reductions and new
568 (lower_reduction_clauses): Don't set placeholder to address of ref
569 if it has already the right type.
570 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
571 (expand_parallel_call): Use the new non-_start suffixed builtins,
572 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
573 and GOMP_parallel_end after the call.
574 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
575 (expand_omp_for_init_counts): Handle combined loops.
576 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
578 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
579 end of cancellable loops.
580 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
581 Likewise. Handle collapse > 1 loops.
582 (expand_omp_simd): Handle combined loops.
583 (expand_omp_for): Add inner_stmt argument, adjust callers of
584 expand_omp_for* functions, use expand_omp_for_static*chunk even
585 for collapse > 1 unless ordered.
586 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
587 of cancellable sections.
588 (expand_omp_single): Remove need_barrier variable, just rely on
589 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
590 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
591 (expand_omp_atomic_load, expand_omp_atomic_store,
592 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
593 (expand_omp_target): New function.
594 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
595 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
596 (build_omp_regions_1): Immediately close region for
597 GF_OMP_TARGET_KIND_UPDATE.
598 (maybe_add_implicit_barrier_cancel): New function.
599 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
601 (lower_omp_single): Likewise. Add clobber after the barrier.
602 (lower_omp_taskgroup): New function.
603 (lower_omp_for): Handle combined loops. Adjust
604 lower_rec_input_clauses caller. Handle cancellation.
605 (lower_depend_clauses): New function.
606 (lower_omp_taskreg): Lower depend clauses. Adjust
607 lower_rec_input_clauses caller. Add clobber after the call. Handle
609 (lower_omp_target, lower_omp_teams): New functions.
610 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
611 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
612 and GOMP_cancellation_point calls.
613 (lower_omp): Fold stmts inside of target region.
614 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
615 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
616 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
617 (BT_FN_VOID_OMPFN_PTR_UINT,
618 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
619 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
620 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
621 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
622 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
623 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
624 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
625 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
626 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
627 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
628 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
629 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
630 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
631 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
632 BUILT_IN_GOMP_PARALLEL_END.
633 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
634 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
635 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
636 GF_OMP_FOR_KIND_DISTRIBUTE.
637 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
638 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
639 (dump_gimple_omp_return): Print lhs if it has any.
640 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
641 gimple_omp_atomic_seq_cst_p.
642 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
643 and GIMPLE_OMP_TEAMS.
644 * langhooks.c (lhd_omp_mappable_type): New function.
645 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
646 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
648 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
649 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
650 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
652 (struct gimplify_omp_ctx): Add combined_loop field.
653 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
654 on stmts inside of target region.
655 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
656 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
657 ORT_TARGET and ORT_TARGET_DATA.
658 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
660 (omp_notice_threadprivate_variable): Complain about threadprivate
661 variables in target region.
662 (omp_notice_variable): Complain about vars with non-mappable type
663 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
664 (omp_check_private): Ignore ORT_TARGET* regions.
665 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
666 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
667 (find_combined_omp_for): New function.
668 (gimplify_omp_for): Handle gimplification of combined loops.
669 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
671 (gimplify_omp_target_update): New function.
672 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
673 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
674 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
675 (gimplify_body): If fndecl has "omp declare target" attribute, add
676 implicit ORT_TARGET context around it.
677 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
678 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
679 * tree-nested.c (convert_nonlocal_reference_stmt,
680 convert_local_reference_stmt, convert_gimple_call): Handle
681 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
682 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
683 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
684 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
685 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
686 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
687 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
688 BUILT_IN_GOMP_LOOP_END_CANCEL,
689 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
690 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
691 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
692 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
693 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
694 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
695 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
696 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
697 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
698 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
699 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
700 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
701 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
702 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
703 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
704 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
705 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
706 gimple_build_omp_teams): New functions.
707 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
708 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
709 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
710 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
711 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
712 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
713 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
714 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
715 GF_OMP_ATOMIC_SEQ_CST): New.
716 (gimple_build_omp_taskgroup, gimple_build_omp_target,
717 gimple_build_omp_teams): New prototypes.
718 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
719 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
720 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
721 GIMPLE_OMP_SINGLE as end of range.
722 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
723 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
724 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
725 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
726 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
727 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
728 gimple_omp_target_kind, gimple_omp_target_set_kind,
729 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
730 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
731 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
732 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
733 gimple_omp_teams_set_clauses): New inlines.
734 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
735 and GIMPLE_OMP_TASKGROUP.
736 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
737 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
738 enum omp_clause_proc_bind_kind): New.
739 (union omp_clause_subcode): Add depend_kind, map_kind and
740 proc_bind_kind fields.
741 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
742 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
743 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
744 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
745 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
747 2013-10-10 Teresa Johnson <tejohnson@google.com>
749 * predict.c (tree_estimate_probability): Add new parameter
750 for estimate_bb_frequencies.
751 (estimate_bb_frequencies): Add new parameter to force estimation.
752 (rebuild_frequencies): When max frequency in function is small,
753 recompute counts from frequencies.
754 * predict.h (estimate_bb_frequencies): New parameter.
756 2013-10-10 David Malcolm <dmalcolm@redhat.com>
758 * ipa-inline.c (ipa_inline): Fix leak of "order" when
759 optimizations are disabled.
761 2013-10-10 David Malcolm <dmalcolm@redhat.com>
763 * coverage.c (coverage_finish): Fix leak of da_file_name.
765 2013-10-10 Jan Hubicka <jh@suse.cz>
767 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
768 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
771 2013-10-10 Jakub Jelinek <jakub@redhat.com>
774 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
775 if any labels are in FALLTHRU_BB, use a special label emitted
776 immediately after the asm goto insn rather than label_rtx
778 (expand_asm_stmt): Adjust caller.
779 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
780 edge if the last insn in predecessor is a jump with single successor,
781 but it isn't simplejump_p.
783 2013-10-10 Richard Biener <rguenther@suse.de>
785 PR tree-optimization/58656
786 * tree-ssa-pre.c (phi_translate): Do not cache failed
789 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
791 * gimplify.c: Include expr.h and tm_p.h for targets with special
792 va-arg padding requirements.
794 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
796 * tree-flow.h: Move some prototypes to gimple.h.
797 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
798 * gimple.h: Relocate some prototypes from tree-flow.h
799 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
801 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
802 (build_va_arg_indirect_ref): Relocate and make static.
803 (std_gimplify_va_arg_expr): Relocate here.
804 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
805 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
807 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
809 * doc/md.texi: Document the mnemonic attribute.
811 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
814 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
815 (set_attr_alternative x ...) when searching for user defined
818 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
820 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
822 2013-10-09 Easwaran Raman <eraman@google.com>
824 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
825 * cfgexpand.c (defer_stack_allocation): ...use here
826 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
828 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
830 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
831 (operand_equal_for_value_replacement): New function, extracted from
832 value_replacement and enhanced to catch more cases.
833 (value_replacement): Use operand_equal_for_value_replacement.
835 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
837 * loop-doloop.c (doloop_modify, doloop_optimize): Use
838 get_max_loop_iterations.
840 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
842 * config/arm/aarch-common.c (arm_early_load_addr_dep):
843 Place comment above function.
845 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
847 * tree-flow.h: Remove all remaining prototypes, enums and structs that
848 are not related to tree-cfg.c.
849 * tree-ssa-address.h: New file. Relocate prototypes.
850 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
851 (addr_for_mem_ref): New. Combine call to get_address_description and
852 return addr_for_mem_ref.
853 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
854 * tree-ssa-live.h: Adjust prototypes.
855 * passes.c: Include tree-ssa-live.h.
856 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
857 * graphite.c (graphite_transform_loops): Make static.
858 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
859 make_pass_graphite, pass_data_graphite_transforms,
860 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
861 * ipa-pure-const.c (warn_function_noreturn): Make static.
862 (execute_warn_function_noreturn, gate_warn_function_noreturn,
863 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
864 Relocate from tree-cfg.c
865 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
867 (execute_warn_function_noreturn, gate_warn_function_noreturn,
868 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
869 Move to ipa-pure-const.c.
870 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
871 Relocate from tree-optimize.c.
872 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
873 make_pass_fixup_cfg): Move to tree-cfg.c.
874 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
875 Relocate some prototypes.
876 * tree-data-ref.h (tree_check_data_deps) Add prototype.
877 * tree-dump.c (dump_function_to_file): Remove prototype.
878 Add tree-flow.h to the include file.
879 * tree-dump.h: Remove prototype.
880 * tree-parloops.h: New File. Add prototypes.
881 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
882 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
883 from tree-ssa-loop.c.
884 * tree-predcom.c (run_tree_predictive_commoning,
885 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
886 Relocate here from tree-ssa-loop.c.
887 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
888 ssa_name_values.release ().
889 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
890 (ssa_name_values): Relocate from tree-flow.h.
891 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
892 * tree-ssa-loop.c (run_tree_predictive_commoning,
893 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
894 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
895 make_pass_graphite, pass_data_graphite_transforms,
896 make_pass_graphite_transforms, gate_tree_parallelize_loops,
897 tree_parallelize_loops, pass_data_parallelize_loops,
898 make_pass_parallelize_loops): Move to other files.
899 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
901 * tree.h: Remove prototypes from tree-address.c.
903 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
905 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
906 (struct int_tree_map): Move to tree-hasher.h
907 (SCALE, LABEL, PERCENT): Move to gimple.h
908 * tree-flow-inline.h: Delete. Move functions to other files.
909 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
910 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
911 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
912 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
913 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
914 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
915 * tree-hasher.h: Don't include tree-flow.h.
916 (struct int_tree_map): Relocate from tree-flow.h.
917 * tree-sra.c (contains_view_convert_expr_p): Relocate from
918 tree-flow-inline.h and make static.
919 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
921 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
922 tree-flow-inline.h and make static.
923 * tree.h (is_global_var, may_be_aliased): Relocate from
925 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
926 * value-prof.c: No longer include tree-flow-inline.h.
927 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
929 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
931 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
932 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
933 (enum move_pos): Move to tree-ssa-loop-im.h
934 * cfgloop.h: Move some prototypes.
935 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
936 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
937 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
938 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
939 (loop_containing_stmt): Relocate from tree-flow-inline.h.
940 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
941 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
942 (enum move_pos): Relocate here.
943 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
945 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
946 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
947 make_pass_lim): Relocate here from tree-ssa-loop.c.
948 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
950 (loop_edge_to_cancel, unloop_loops): Make static.
951 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
952 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
953 (tree_complete_unroll, gate_tree_complete_unroll,
954 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
955 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
956 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
957 * tree-ssa-loop-ivopts.c: Remove local prototypes.
958 (stmt_invariant_in_loop_p): Remove unused function.
959 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
960 * tree-ssa-loop-manip.h: New file. Add prototypes.
961 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
962 (gcov_type_to_double_int): Move to cfgloop.h.
963 (double_int_cmp, bound_index,
964 estimate_numbers_of_iterations_loop): Make static.
965 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
966 (max_loop_iterations): Factor out get_max_loop_iterations.
967 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
969 * tree-ssa-loop-niter.h: New file. Add prototypes.
970 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
971 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
972 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
973 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
974 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
975 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
976 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
977 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
978 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
979 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
980 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
981 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
982 pass_data_complete_unroll, make_pass_complete_unroll,
983 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
984 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
985 tree-ssa-loop-ivcanon.c.
986 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
987 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
988 tree-ssa-loop-prefetch.c.
989 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
991 (get_lsm_tmp_name): Relocate and add suffix parameter.
992 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
993 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
994 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
995 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
996 (get_estimated_loop_iterations): Factor out accessor from
997 estimated_loop_iterations in tree-ssa-loop-niter.c.
998 (get_max_loop_iterations): Factor out accessor from
999 _max_loop_iterations in tree-ssa-niter.c.
1000 * loop-unroll.c (decide_unroll_constant_iterations,
1001 decide_unroll_runtime_iterations, decide_peel_simple,
1002 decide_unroll_stupid): Use new get_* accessors.
1004 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
1006 PR tree-optimization/20318
1007 * doc/extend.texi (returns_nonnull): New function attribute.
1008 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
1010 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
1011 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
1013 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
1016 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
1017 false if both components are bitfields.
1019 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1021 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
1022 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
1023 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
1024 * config/aarch64/aarch64.h
1025 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
1026 * config/aarch64/aarch64-simd-builtins.def
1027 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
1029 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1031 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
1033 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1035 * config/aarch64/arm_neon.h (vdiv_f64): Added.
1037 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1039 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
1040 (vneg_f64): New intrinsic.
1041 (vneg_s8): Asm replaced with C.
1042 (vneg_s16): Likewise.
1043 (vneg_s32): Likewise.
1044 (vneg_s64): New intrinsic.
1045 (vnegq_f32): Asm replaced with C.
1046 (vnegq_f64): Likewise.
1047 (vnegq_s8): Likewise.
1048 (vnegq_s16): Likewise.
1049 (vnegq_s32): Likewise.
1050 (vnegq_s64): Likewise.
1052 2013-10-09 Renlin Li <Renlin.Li@arm.com>
1054 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
1056 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1058 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
1059 packed stack special handling.
1060 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
1061 back to fixed stack slots for FPRs saved due to stdarg.
1063 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1065 * config/s390/s390.c (s390_frame_info): Restructure function.
1067 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1069 * config/s390/s390.c (struct s390_frame_layout): New field
1071 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
1072 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
1073 regs_ever_clobbered to char*.
1074 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
1075 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
1076 (s390_register_info): Enable FPR save slots. Move/Copy some
1077 functionality into ...
1078 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
1079 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
1081 (s390_frame_info): Do gpr slot allocation here now. stdarg does
1082 not imply a stack frame.
1083 (s390_init_frame_layout): Remove variable clobbered_regs.
1084 (s390_update_register_info): Remove function.
1085 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
1086 cannot be used for register renaming.
1087 (s390_hard_regno_scratch_ok): New function.
1088 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
1089 (s390_initial_elimination_offset): Change offset calculation of
1090 the return address pointer.
1091 (save_gprs): Deal with only r6 being saved from the call-saved regs.
1092 (restore_gprs): Set frame related flag.
1093 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
1094 (s390_emit_prologue): Call s390_register_info instead of
1095 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
1096 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
1097 (s390_optimize_prologue): Call s390_optimize_register_info.
1098 Try to remove also FPR slot save/restore INSNs. Remove frame
1099 related flags from restore INSNs.
1101 2013-10-08 DJ Delorie <dj@redhat.com>
1103 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
1106 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
1107 avoid conflict with the MI use of %c.
1108 * config/rl78/rl78-real.md: change %c to %C throughout.
1109 * config/rl78/rl78-virt.md: Likewise.
1111 2013-10-08 Jan Hubicka <jh@suse.cz>
1113 * config/i386/i386.c (ix86_option_override_internal): Switch
1114 to SSE math for -ffast-math when target ISA supports SSE2.
1116 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
1118 * tree-flow.h: Remove some prototypes.
1119 * tree.h: Remove some protypes, add a couple.
1120 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
1121 using_eh_for_cleanups_p): Add interface routines for front ends.
1122 * tree-eh.h: New file. Add protoptyes.
1123 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
1124 (add_stmt_to_eh_lp_fn): Make static.
1125 (lower_try_finally): Use new using_eh_for_cleanups_p.
1126 * emit-rtl.c: Include tree-eh.h.
1127 * gimple.h: Include tree-eh.h.
1129 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
1131 PR tree-optimization/58480
1132 * tree-vrp.c (infer_nonnull_range): New function.
1133 (infer_value_range): Call infer_nonnull_range.
1135 2013-10-08 Dehao Chen <dehao@google.com>
1137 PR tree-optimization/58619
1138 * tree-inline.c (copy_phis_for_bb): Combine location data
1141 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1144 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
1145 RTX_FRAME_RELATED_P on INSN.
1147 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1149 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
1150 (altivec_expand_vec_perm_const): Call it.
1152 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1154 * config/rs6000/vector.md (mov<mode>): Emit permuted move
1155 sequences for LE VSX loads and stores at expand time.
1156 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
1158 * config/rs6000/rs6000.c (rs6000_const_vec): New.
1159 (rs6000_gen_le_vsx_permute): New.
1160 (rs6000_gen_le_vsx_load): New.
1161 (rs6000_gen_le_vsx_store): New.
1162 (rs6000_gen_le_vsx_move): New.
1163 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
1164 (*vsx_le_perm_load_v4si): New.
1165 (*vsx_le_perm_load_v8hi): New.
1166 (*vsx_le_perm_load_v16qi): New.
1167 (*vsx_le_perm_store_v2di): New.
1168 (*vsx_le_perm_store_v4si): New.
1169 (*vsx_le_perm_store_v8hi): New.
1170 (*vsx_le_perm_store_v16qi): New.
1171 (*vsx_xxpermdi2_le_<mode>): New.
1172 (*vsx_xxpermdi4_le_<mode>): New.
1173 (*vsx_xxpermdi8_le_V8HI): New.
1174 (*vsx_xxpermdi16_le_V16QI): New.
1175 (*vsx_lxvd2x2_le_<mode>): New.
1176 (*vsx_lxvd2x4_le_<mode>): New.
1177 (*vsx_lxvd2x8_le_V8HI): New.
1178 (*vsx_lxvd2x16_le_V16QI): New.
1179 (*vsx_stxvd2x2_le_<mode>): New.
1180 (*vsx_stxvd2x4_le_<mode>): New.
1181 (*vsx_stxvd2x8_le_V8HI): New.
1182 (*vsx_stxvd2x16_le_V16QI): New.
1184 2013-10-07 Renlin Li <Renlin.Li@arm.com>
1186 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
1188 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1190 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
1191 loop to work also for 31bit ABI.
1192 Save the stack pointer for frame_size > 0.
1194 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1196 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
1197 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
1198 constraint letters from expanders.
1199 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
1200 retry count to general_operand.
1201 ("tabort"): Give operand 0 a mode.
1202 ("tabort_1"): Add mode and constraint letter for operand 0.
1203 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
1205 2013-10-04 Jeff Law <law@redhat.com>
1207 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
1209 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
1211 (thread_across_edge): Here. Call it.
1213 2013-10-04 Cary Coutant <ccoutant@google.com>
1215 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
1216 discarding a location list expression (or a piece of one).
1218 2013-10-03 Jan Hubicka <jh@suse.cz>
1220 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
1221 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
1222 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
1224 2013-10-03 Jan Hubicka <jh@suse.cz>
1226 * config/i386/i386.c (ix86_option_override_internal): Do not enable
1227 accumulate-outgoing-args when producing unwind info.
1229 2013-10-03 Wei Mi <wmi@google.com>
1231 * lra-constraints.c (insert_move_for_subreg): New function
1232 extracted from simplify_operand_subreg.
1233 (simplify_operand_subreg): Add reload for paradoxical subreg.
1235 2013-10-03 Rong Xu <xur@google.com>
1237 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
1238 the candidate of builtin_expect such that we should fix the
1239 size/time estimation.
1240 (estimate_function_body_sizes): Do the acutally size/time fix-up
1243 2013-10-03 Rong Xu <xur@google.com>
1245 * predict.c (tree_predict_by_opcode): Get the probability
1246 for builtin_expect from param builtin_expect_probability.
1247 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
1248 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
1249 * doc/invoke.texi: Add documentation for builtin-expect-probability.
1251 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
1254 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
1255 * calls.c (alloca_call_p): Use get_callee_fndecl.
1256 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
1257 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
1259 (vrp_visit_stmt): Remove duplicated code.
1261 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
1263 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
1264 ceildf2, btruncdf2, instead of vsx_* name.
1266 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
1267 iterators to only do V2DF and V4SF here. Move the DF code to
1268 rs6000.md where it is combined with SF mode. Replace <VSv> with
1269 just 'v' since only vector operations are handled with these insns
1270 after moving the DF support to rs6000.md.
1271 (vsx_sub<mode>3): Likewise.
1272 (vsx_mul<mode>3): Likewise.
1273 (vsx_div<mode>3): Likewise.
1274 (vsx_fre<mode>2): Likewise.
1275 (vsx_neg<mode>2): Likewise.
1276 (vsx_abs<mode>2): Likewise.
1277 (vsx_nabs<mode>2): Likewise.
1278 (vsx_smax<mode>3): Likewise.
1279 (vsx_smin<mode>3): Likewise.
1280 (vsx_sqrt<mode>2): Likewise.
1281 (vsx_rsqrte<mode>2): Likewise.
1282 (vsx_fms<mode>4): Likewise.
1283 (vsx_nfma<mode>4): Likewise.
1284 (vsx_copysign<mode>3): Likewise.
1285 (vsx_btrunc<mode>2): Likewise.
1286 (vsx_floor<mode>2): Likewise.
1287 (vsx_ceil<mode>2): Likewise.
1288 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
1289 (vsx_sminsf3): Likewise.
1290 (vsx_fmadf4): Likewise.
1291 (vsx_fmsdf4): Likewise.
1292 (vsx_nfmadf4): Likewise.
1293 (vsx_nfmsdf4): Likewise.
1294 (vsx_cmpdf_internal1): Likewise.
1296 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
1297 simpler to select whether a target has SPE or traditional floating
1298 point support in iterators.
1299 (TARGET_DF_SPE): Likewise.
1300 (TARGET_SF_FPR): Likewise.
1301 (TARGET_DF_FPR): Likewise.
1302 (TARGET_SF_INSN): Macros to say whether floating point support
1303 exists for a given operation for expanders.
1304 (TARGET_DF_INSN): Likewise.
1306 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
1307 combining of SF/DF mode operations, using both traditional and VSX
1315 (abs<mode>2): Combine SF/DF modes using traditional floating point
1316 instructions. Add support for using the upper DF registers with
1317 VSX support, and SF registers with power8-vector support. Update
1318 expanders for operations supported by both the SPE and traditional
1319 floating point units.
1320 (abs<mode>2_fpr): Likewise.
1321 (nabs<mode>2): Likewise.
1322 (nabs<mode>2_fpr): Likewise.
1323 (neg<mode>2): Likewise.
1324 (neg<mode>2_fpr): Likewise.
1325 (add<mode>3): Likewise.
1326 (add<mode>3_fpr): Likewise.
1327 (sub<mode>3): Likewise.
1328 (sub<mode>3_fpr): Likewise.
1329 (mul<mode>3): Likewise.
1330 (mul<mode>3_fpr): Likewise.
1331 (div<mode>3): Likewise.
1332 (div<mode>3_fpr): Likewise.
1333 (sqrt<mode>3): Likewise.
1334 (sqrt<mode>3_fpr): Likewise.
1335 (fre<Fs>): Likewise.
1336 (rsqrt<mode>2): Likewise.
1337 (cmp<mode>_fpr): Likewise.
1338 (smax<mode>3): Likewise.
1339 (smin<mode>3): Likewise.
1340 (smax<mode>3_vsx): Likewise.
1341 (smin<mode>3_vsx): Likewise.
1342 (negsf2): Delete SF operations that are merged with DF.
1349 (fmasf4_fpr): Likewise.
1350 (fmssf4_fpr): Likewise.
1351 (nfmasf4_fpr): Likewise.
1352 (nfmssf4_fpr): Likewise.
1353 (sqrtsf2): Likewise.
1354 (rsqrtsf_internal1): Likewise.
1355 (smaxsf3): Likewise.
1356 (sminsf3): Likewise.
1357 (cmpsf_internal1): Likewise.
1358 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
1359 (negdf2): Delete DF operations that are merged with SF.
1361 (nabsdf2): Likewise.
1367 (rsqrtdf_internal1): Likewise.
1368 (fmadf4_fpr): Likewise.
1369 (fmsdf4_fpr): Likewise.
1370 (nfmadf4_fpr): Likewise.
1371 (nfmsdf4_fpr): Likewise.
1372 (sqrtdf2): Likewise.
1373 (smaxdf3): Likewise.
1374 (smindf3): Likewise.
1375 (cmpdf_internal1): Likewise.
1376 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
1377 (btrunc<mode>2): Delete separate expander, and combine with the
1378 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
1379 (btrunc<mode>2_fpr): Likewise.
1380 (ceil<mode>2): Likewise.
1381 (ceil<mode>2_fpr): Likewise.
1382 (floor<mode>2): Likewise.
1383 (floor<mode>2_fpr): Likewise.
1384 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
1385 Add support for using the upper registers with VSX and
1386 power8-vector. Move insns to be closer to the define_expands. On
1387 VSX systems, prefer the traditional form of FMA over the VSX
1388 version, since the traditional form allows the target not to
1389 overlap with the inputs.
1390 (fms<mode>4_fpr): Likewise.
1391 (nfma<mode>4_fpr): Likewise.
1392 (nfms<mode>4_fpr): Likewise.
1394 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1395 Richard Earnshaw <richard.earnshaw@arm.com>
1397 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
1398 (struct mult_cost_table): Likewise.
1399 (struct mem_cost_table): Likewise.
1400 (struct fp_cost_table): Likewise.
1401 (struct vector_cost_table): Likewise.
1402 (cpu_cost_table): Likewise.
1403 * config/arm/arm.opt (mold-rts-costs): New option.
1404 (mnew-generic-costs): Likewise.
1405 * config/arm/arm.c (generic_extra_costs): New table.
1406 (cortexa15_extra_costs): Likewise.
1407 (arm_slowmul_tune): Use NULL as new costs.
1408 (arm_fastmul_tune): Likewise.
1409 (arm_strongarm_tune): Likewise.
1410 (arm_xscale_tune): Likewise.
1411 (arm_9e_tune): Likewise.
1412 (arm_v6t2_tune): Likewise.
1413 (arm_cortex_a5_tune): Likewise.
1414 (arm_cortex_a9_tune): Likewise.
1415 (arm_v6m_tune): Likewise.
1416 (arm_fa726te_tune): Likewise.
1417 (arm_cortex_a15_tune): Use cortex15_extra_costs.
1418 (arm_cortex_tune): Use generict_extra_costs.
1419 (shifter_op_p): New function.
1420 (arm_unspec_cost): Likewise.
1421 (LIBCALL_COST): Define.
1422 (arm_new_rtx_costs): New function.
1423 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
1424 table is available. Use old costs otherwise unless mnew-generic-costs
1426 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
1427 (cpu_cost_table): Declare.
1429 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
1432 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
1433 (*subs_mul_imm_<mode>)
1434 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
1435 (*sub_<shift>_<mode>)
1436 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
1437 Remove k constraint.
1439 2013-10-03 Ian Bolton <ian.bolton@arm.com>
1441 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
1443 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
1445 2013-10-02 Teresa Johnson <tejohnson@google.com>
1447 * predict.c (probably_never_executed): New function.
1448 (probably_never_executed_bb_p): Invoke probably_never_executed.
1449 (probably_never_executed_edge_p): Ditto.
1450 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
1451 Treat profile insanities conservatively.
1453 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
1455 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
1457 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
1459 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
1460 use it. Use smaller increase for scratch. Don't increase reject
1461 for early clobber scratch.
1462 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
1463 setting eliminated regs except setting fp from hfp.
1464 (lra_eliminate): Check lra_insn_recog_data on NULL.
1466 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1469 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
1470 setting -mvsx-timode by default until the underlying problem is fixed.
1471 (RS6000_CPU, power7 defaults): Likewise.
1473 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
1475 * config/x-linux (host-linux.o): Remove header dependencies.
1476 Use $(COMPILE) and $(POSTCOMPILE).
1477 * config/t-linux-android (linux-android.o): Ditto.
1479 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
1481 * Makefile.in (expmed.o-warn): Remove.
1483 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1485 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
1486 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
1488 2013-10-02 Teresa Johnson <tejohnson@google.com>
1490 * dojump.c (do_jump_1): Divide probability between
1491 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
1493 2013-10-02 Tom Tromey <tromey@redhat.com>
1495 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
1497 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1499 * tree-flow.h: Remove some prototypes.
1500 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
1501 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
1502 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
1503 mark_virtual_phi_result_for_renaming): Relocate here.
1504 * tree-into-ssa.h: Add prototypes.
1505 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
1506 single_pred_before_succ_order.
1507 (blocks_in_phiopt_order): Rename and move to cfganal.c.
1508 (nonfreeing_call_p) Move to gimple.c.
1509 * cfganal.c (single_pred_before_succ_order): Move and renamed from
1511 * basic-block.h (single_pred_before_succ_order): Add prototype.
1512 * gimple.c (nonfreeing_call_p): Relocate here.
1513 * gimple.h: Add prototype.
1514 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
1515 * tree-ssa-dom.h: New file. Relocate prototypes here.
1516 * tree-ssa.h: Include tree-ssa-dom.h.
1518 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
1520 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
1521 Use $(COMPILE) and $(POSTCOMPILE).
1523 * config/alpha/x-alpha (driver-alpha.o): Ditto.
1525 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1527 * tree-flow.h: Remove some prototypes.
1528 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
1529 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
1530 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
1531 propagate_tree_value*): Move from here to...
1532 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
1533 propagate_tree_value*): Relocate here.
1534 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
1535 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
1536 * gimple-fold.c: Remove gimple-fold.h from include list.
1537 * tree-vrp.c: Remove gimple-fold.h from include list.
1538 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
1539 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
1540 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
1541 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
1542 * sese.c: Add tree-ssa-propagate.h to include list.
1544 2013-10-02 Richard Biener <rguenther@suse.de>
1546 * tree-loop-distribution.c: Include tree-vectorizer.h for
1548 (enum partition_kind): Remove PKIND_REDUCTION.
1549 (struct partition_s): Remove has_writes member, add reduction_p member.
1550 (partition_alloc): Adjust.
1551 (partition_builtin_p): Likewise.
1552 (partition_has_writes): Remove.
1553 (partition_reduction_p): New function.
1554 (partition_merge_into): Likewise.
1555 (generate_code_for_partition): Commonize builtin partition
1557 (rdg_cannot_recompute_vertex_p): Remove.
1558 (already_processed_vertex_p): Likewise.
1559 (rdg_flag_vertex): Do not set has_writes.
1560 (classify_partition): Adjust.
1561 (rdg_build_partitions): Do not set has_writes, treat all
1562 partitions as useful.
1563 (distribute_loop): Record number of library calls generated. Adjust.
1564 (tree_loop_distribution): Report number of loops and library
1565 calls generated as opt-info.
1567 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1569 * tree-flow.h: Include new .h files. Move prototypes.
1570 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
1571 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
1572 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
1573 * tree-pretty-print.h: Add prototypes from tree-flow.h.
1574 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
1575 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
1576 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
1577 * tree.h (get_ref_base_and_extent): Move prototype out.
1578 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
1580 * gimple-low.h: New File. Add prototypes from tree-flow.h.
1581 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
1582 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
1583 * tree-scalar-evolution.c: Include tree.h.
1584 * sese.c: Include tree.h.
1585 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
1586 * dwarf2out.c: Include tree-dfa.h.
1587 * tree-chrec.c: Include tree.h.
1588 * tree-data-ref.c: Include tree.h.
1590 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
1592 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
1595 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1597 * config/t-sol2 (sol2-c.o): Remove header dependencies.
1598 Use $(COMPILE) and $(POSTCOMPILE).
1599 (sol2-cxx.o): Likewise.
1600 (sol2-stubs.o): Likewise.
1602 * config/x-solaris (host-solaris.o): Likewise.
1604 * config/sparc/t-sparc (sparc.o): Remove.
1605 (sparc-c.o): Remove header dependencies.
1606 Use $(COMPILE) and $(POSTCOMPILE).
1607 * config/sparc/x-sparc: Likewise.
1609 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
1611 * config/arc/arc-opts.h: Add 2013 to Copyright years.
1612 * config/arc/arc700.md: Likewise.
1613 * config/arc/arc-modes.def: Likewise.
1614 * config/arc/arc-simd.h: Likewise.
1615 * config/arc/t-arc-uClibc: Likewise.
1616 * config/arc/t-arc-newlib: Likewise.
1618 2013-10-02 Renlin Li <renlin.li@arm.com>
1620 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1622 (aarch64_expand_epilogue): Likewise.
1624 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1625 Yufeng Zhang <yufeng.zhang@arm.com>
1627 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
1629 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
1630 'base_in' represent a conversion and legal_cast_p_1 holds; set
1631 'base_in' with the returned value from get_unwidened.
1633 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1635 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
1636 plus_constant is not used.
1638 2013-10-01 Wei Mi <wmi@google.com>
1640 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
1641 * config/i386/i386.md: Add define_peephole2 to
1642 break partial reg stall for cvtss2sd/cvtsd2ss.
1644 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1646 * config/arc/arc.c (pass_arc_ifcvt::clone):
1647 Update for ctxt_ -> m_ctxt change.
1649 2013-10-01 Jeff Law <law@redhat.com>
1651 * tree-ssa-threadupdate.c (struct redirection_data): Delete
1652 outgoing_edge and intermediate_edge fields. Instead store the path.
1653 (redirection_data::hash): Hash on the last edge's destination index.
1654 (redirection_data::equal): Check the entire thread path.
1655 (lookup_redirectio_data): Corresponding changes.
1656 (create_edge_and_update_destination_phis): Likewise.
1657 (thread_single_edge): Likewise.
1659 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1660 Diego Novillo <dnovillo@google.com>
1662 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
1663 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
1664 (vld32wh_insn, vld32wl_insn): Delete commented-out old
1665 versions of these patterns.
1667 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
1668 (__builtin_arc_aligned): Likewise.
1670 * config/arc/arc.md: Expand adc_0 comment stating the intended
1671 purpose and why it isn't ready.
1672 Replace commented out call_value_via_label_mixed with a
1673 plain comment about bl_s.
1675 * config/arc/arc.c (stdio.h): Don't include directly.
1676 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
1677 Remove if (1) condition.
1678 (arc_encode_section_info): Fix comment.
1680 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1682 * config/arc/arc.c (arc_conditional_register_usage):
1683 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
1684 Also set reg_alloc_order for DMA config regs.
1686 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1687 Jeremy Bennett <jeremy.bennett@embecosm.com>
1689 * doc/install.texi (--with-cpu): Mention ARC.
1690 (arc-*-elf32): New paragraph.
1691 (arc-linux-uclibc): Likewise.
1692 * doc/md.texi (Machine Constraints): Add ARC part.
1693 * doc/invoke.texi: (menu): Add ARC Options.
1694 (Machine Dependent Options) <ARC Options>: Add synopsis.
1695 (node ARC Options): Add.
1696 * doc/extend.texi (long_call / short_call attribute): Add ARC.
1697 (ARC Built-in Functions): New section defining
1698 generic ARC built-in functions.
1699 (ARC SIMD Built-in Functions): New section defining SIMD specific
1701 (Declaring Attributes of Functions): Extended
1702 description of short_call and long_call attributes for ARC and
1703 added index entries.
1705 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
1706 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
1707 Joern Rennecke <joern.rennecke@embecosm.com>
1708 Muhammad Khurram Riaz <khurram.riaz@arc.com>
1709 Brendan Kehoe <brendan@zen.org>
1710 Michael Eager <eager@eagercon.com>
1711 Simon Cook <simon.cook@embecosm.com>
1712 Jeremy Bennett <jeremy.bennett@embecosm.com>
1714 * config/arc, common/config/arc: New directories.
1716 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1717 Brendan Kehoe <brendan@zen.org>
1718 Simon Cook <simon.cook@embecosm.com>
1720 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
1722 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
1724 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
1725 * tree-ssa-coalesce.h: New. Move prototype to here.
1726 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
1727 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
1728 (gimple_can_coalesce_p): Move to...
1729 * gimple.c (gimple_can_coalesce_p): Here.
1731 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
1733 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
1734 (dump_decl_set): Move to gimple.c.
1735 * gimple.h: Don't include tree-ssa-operands.h.
1736 (dump_decl_set): Add prototype.
1737 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
1738 Move to gimple-ssa.h.
1739 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
1740 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
1741 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
1742 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
1743 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
1745 (dump_decl_set): Relocate here.
1746 * gimple-ssa.h: New file.
1747 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
1748 Relocate from gimple.h.
1749 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
1750 * tree-ssa-operands.c (swap_ssa_operands): Rename from
1751 swap_tree_operands and remove non-ssa path.
1752 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
1753 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
1755 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
1756 vect_is_simple_reduction_1): Use swap_ssa_operands.
1757 * tree-flow.h: Move various prototypes to tree-phinodes.h.
1758 (enum need_phi_state): Move to tree-into-ssa.c.
1759 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
1760 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
1761 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
1762 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
1763 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
1764 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
1765 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
1766 num_imm_uses): Move to ssa-iterators.h.
1767 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
1768 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
1770 (op_iter_done, op_iter_next_def, op_iter_next_tree,
1771 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
1772 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
1773 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
1774 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
1775 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
1776 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
1777 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
1778 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
1779 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
1780 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
1781 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
1782 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
1783 (set_phi_nodes): Move to tree-phinodes.h.
1784 * tree-ssa-operands.h (enum ssa_op_iter_type,
1785 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
1786 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
1787 (dump_decl_set): Remove prototype.
1788 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
1789 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
1790 (set_phi_nodes): Relocate from tree-flow-inline.h.
1791 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
1793 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
1794 include list. Temporarily add gimple.h to include list.
1795 * ssa-iterators.h: New file.
1796 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
1797 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
1798 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
1799 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
1800 Relocate from tree-ssa-operands.h.
1801 (delink_imm_use, link_imm_use_to_list, link_imm_use,
1802 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
1803 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
1804 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
1805 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
1806 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
1807 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
1808 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
1809 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
1810 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
1811 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
1812 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
1813 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
1814 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
1815 Relocate from tree-flow-inline.h.
1816 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
1818 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
1821 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
1822 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
1823 Insert '\t' to output template.
1824 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
1825 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
1826 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
1827 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
1829 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
1831 * doc/install.texi (Host/target specific installation notes for GCC):
1832 Put @anchor before @heading.
1833 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
1834 Use @email for email addresses.
1836 2013-10-01 Jeff Law <law@redhat.com>
1838 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
1839 a vec. Only delete the path if we create one without successfully
1840 registering a jump thread.
1841 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
1843 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
1844 (paths): New vector of jump threading paths.
1845 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
1846 (THREAD_PATH): New accessor macro for the entire thread path.
1847 (lookup_redirection_data): Get intermediate and final outgoing edge
1848 from the thread path.
1849 (create_edge_and_update_destination_phis): Copy the threading path.
1850 (ssa_fix_duplicate_block_edges): Get edges and block types from the
1851 jump threading path.
1852 (ssa_redirect_edges): Get edges and block types from the jump threading
1853 path. Free the path vector.
1854 (thread_block): Get edges from the jump threading path. Look at the
1855 entire path to see if we thread to a loop exit. If we cancel a jump
1856 thread request, then free the path vector.
1857 (thread_single_edge): Get edges and block types from the jump threading
1858 path. Free the path vector.
1859 (thread_through_loop_header): Get edges and block types from the jump
1860 threading path. Free the path vector.
1861 (mark_threaded_blocks): Iterate over the vector of paths and store
1862 the path on the appropriate edge. Get edges and block types from the
1863 jump threading path.
1864 (mark_threaded_blocks): Get edges and block types from the jump
1865 threading path. Free the path vector.
1866 (thread_through_all_blocks): Use the vector of paths rather than
1867 a vector of 3-edge sets.
1868 (register_jump_thread): Accept pointer to a path vector rather
1869 than the path vector itself. Store the path vector for later use.
1872 2013-10-01 Jakub Jelinek <jakub@redhat.com>
1873 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1876 * config/s390/s390.c (s390_split_branches): Modify check for table
1878 (s390_chunkify_start): Rearrange table jump insn check in order to
1879 deal with compare and branch insns correctly.
1881 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
1885 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
1886 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
1887 define_insn_and_split.
1888 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
1889 (shiftsi3_compare): New pattern.
1891 * config/arm/unspecs.md (UNSPEC_RRX): New.
1893 2013-10-01 Alan Modra <amodra@gmail.com>
1895 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
1896 casing inout operands.
1898 2013-10-01 Richard Biener <rguenther@suse.de>
1900 PR tree-optimization/58553
1901 * tree-loop-distribution.c (struct partition_s): Add niter member.
1902 (classify_partition): Populate niter member for the partition
1903 and properly identify whether the relevant store happens before
1904 or after the loop exit.
1905 (generate_memset_builtin): Use niter member from the partition.
1906 (generate_memcpy_builtin): Likewise.
1908 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1910 * vec.h (vec_prefix, vec): Prefix member names with "m_".
1911 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
1913 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1915 * basic-block.h (edge_list): Prefix member names with "m_".
1916 * context.h (context): Likewise.
1917 * domwalk.h (dom_walker): Likewise.
1918 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
1919 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1920 * hash-table.h (hash_table): Likewise.
1921 * machmode.h (bit_field_mode_iterator): Likewise.
1922 * pass_manager.h (pass_list): Likewise.
1923 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
1924 * tree-pass.h (pass_data): Likewise.
1925 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
1926 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
1927 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1928 * asan.c (pass_data_asan): Update accordingly.
1929 * cfganal.c (control_dependences::find_control_dependence): Likewise.
1930 (control_dependences::control_dependences): Likewise.
1931 (control_dependences::~control_dependences): Likewise.
1932 (control_dependences::~control_dependences): Likewise.
1933 (control_dependences::get_edges_dependent_on): Likewise.
1934 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
1935 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
1936 * context.c (gcc::context::context): Likewise.
1937 * cprop.c (pass_rtl_cprop::clone): Likewise.
1938 * domwalk.c (dom_walker::walk): Likewise.
1939 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
1940 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
1941 * mode-switching.c (pass_mode_switching::clone): Likewise.
1942 * passes.c (opt_pass::opt_pass): Likewise.
1943 (pass_manager::pass_manager): Likewise.
1944 * predict.c (pass_strip_predict_hints::clone): Likewise.
1945 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
1946 (pass_split_all_insns::clone): Likewise.
1947 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
1949 (bit_field_mode_iterator::next_mode): Likewise.
1950 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
1951 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
1952 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
1953 * tree-complex.c (pass_lower_complex::clone): Likewise.
1954 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
1955 * tree-object-size.c (pass_object_sizes::clone): Likewise.
1956 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
1957 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
1958 (pass_fold_builtins::clone): Likewise.
1959 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
1960 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
1961 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
1962 (pass_cd_dce::clone): Likewise.
1963 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
1964 (pass_phi_only_cprop::clone): Likewise.
1965 * tree-ssa-dse.c (pass_dse::clone): Likewise.
1966 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
1967 * tree-ssa-loop.c (pass_lim::clone): Likewise.
1968 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
1969 * tree-ssa-pre.c (pass_fre::clone): Likewise.
1970 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
1971 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
1972 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
1973 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
1974 * tree-vrp.c (pass_vrp::clone): Likewise.
1975 * tsan.c (pass_tsan::clone): Likewise.
1977 2013-09-30 Jakub Jelinek <jakub@redhat.com>
1980 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
1981 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
1984 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
1985 optimization, punt if sign_bit_p looked through any zero extension.
1987 2013-09-30 Teresa Johnson <tejohnson@google.com>
1989 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
1990 Update redirected out edge count in joiner case.
1991 (ssa_redirect_edges): Common the joiner and non-joiner cases
1992 so that joiner case gets profile updates.
1994 2013-09-30 Richard Biener <rguenther@suse.de>
1996 PR tree-optimization/58554
1997 * tree-loop-distribution.c (classify_partition): Require
1998 unconditionally executed stores for memcpy and memset recognition.
1999 (tree_loop_distribution): Calculate dominance info.
2001 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2003 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
2004 (NO_PROFILE_COUNTERS): Likewise.
2005 (PROFILE_HOOK): Likewise.
2006 (FUNCTION_PROFILER): Likewise.
2007 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
2009 2013-09-30 Iain Sandoe <iain@codesourcery.com>
2011 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
2012 calls and defines in TARGET_MACHO conditional.
2013 (load_macho_picbase_di): Likewise.
2014 (reload_macho_picbase): Likewise.
2015 (reload_macho_picbase_si): Likewise.
2016 (reload_macho_picbase_di): Likewise.
2017 (nonlocal_goto_receiver): Likewise.
2019 2013-09-30 Nick Clifton <nickc@redhat.com>
2021 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
2022 that use the MSP430X ISA.
2023 (msp430_option_override): Scan -mmcu command line option for any
2024 MCU name that supports the MSP430X ISA.
2025 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
2026 -mmcu options which enable the MSP430X ISA.
2028 2013-09-30 Richard Biener <rguenther@suse.de>
2031 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
2032 before looking for setjmp-like calls.
2034 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2037 * config/darwin-protos.h (machopic_get_function_picbase): New.
2038 * config/darwin.c (machopic_get_function_picbase): New.
2039 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
2040 label for a new func. (load_macho_picbase_di): Likewise.
2041 (reload_macho_picbase): New expand.
2042 (reload_macho_picbase_si): New insn.
2043 (reload_macho_picbase_di): New insn.
2044 (nonlocal_goto_receiver): New define and split.
2045 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
2046 (unspecv enum): Add UNSPECV_NLGR.
2048 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2050 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
2051 that altivec registers are correctly sized on Darwin.
2053 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2055 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
2056 darwin-driver.o): Use COMPILE and POSTCOMPILE.
2057 * config/x-darwin (host-darwin.o): Likewise.
2058 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
2059 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
2060 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
2062 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
2064 * doc/invoke.texi: Fix usage of @tie{} command.
2066 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
2068 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
2070 2013-09-28 Jan Hubicka <jh@suse.cz>
2072 * config/i386/x86-tune.def: Add documentation for each of the options;
2075 2013-09-28 Jan Hubicka <jh@suse.cz>
2077 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
2079 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
2080 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
2081 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
2082 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
2083 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
2085 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
2087 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
2088 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
2089 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
2090 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
2091 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
2092 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
2093 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
2094 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
2095 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
2096 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
2097 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
2098 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
2099 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
2100 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
2101 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
2102 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
2103 gimple.h, godump.c, graphite-clast-to-gimple.c,
2104 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
2105 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
2106 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
2107 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2108 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
2109 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
2110 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
2111 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
2112 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
2113 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
2114 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
2115 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
2116 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
2117 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
2118 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
2119 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
2120 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
2121 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
2122 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
2123 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
2124 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
2125 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
2126 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
2127 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
2128 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2129 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2130 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
2131 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
2132 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
2133 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2134 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2135 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
2136 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
2137 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2138 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
2139 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
2140 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
2141 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
2142 whitespace before "(".
2144 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
2146 * expr.h (extract_bit_field): Remove packedp parameter.
2147 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
2148 from forward declaration.
2149 (store_split_bit_field): Remove packedp arg from calls to
2150 extract_fixed_bit_field.
2151 (extract_bit_field_1): Remove packedp parameter and packedp
2152 argument from recursive calls and calls to extract_fixed_bit_field.
2153 (extract_bit_field): Remove packedp parameter and corresponding
2154 arg to extract_bit_field_1.
2155 (extract_fixed_bit_field): Remove packedp parameter. Remove code
2157 (extract_split_bit_field): Remove packedp arg from call to
2158 extract_fixed_bit_field.
2159 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
2160 (copy_blkmode_from_reg): Likewise.
2161 (copy_blkmode_to_reg): Likewise.
2162 (read_complex_part): Likewise.
2163 (store_field): Likewise.
2164 (expand_expr_real_1): Likewise.
2165 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
2166 to extract_bit_field.
2167 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
2168 call to extract_bit_field.
2169 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
2170 call to extract_bit_field.
2171 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
2172 and special packedp behavior from -fstrict-volatile-bitfields
2175 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2177 * lra-eliminations.c (init_elim_table): Guard value_p.
2179 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2181 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
2182 DFmode, DImode, and SFmode in the upper VSX registers based on the
2183 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
2184 if -mpower8-vector. Combine -mvsx-timode handling with the rest
2185 of the VSX register handling.
2187 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
2189 (zero_extendsidi2_lfiwzx): Add support for loading into the
2190 Altivec registers with -mpower8-vector. Use wu/wv constraints to
2191 only do VSX memory options on Altivec registers.
2192 (extendsidi2_lfiwax): Likewise.
2193 (extendsfdf2_fpr): Likewise.
2194 (mov<mode>_hardfloat, SF/SD modes): Likewise.
2195 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
2196 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
2197 (movdi_internal64): Likewise.
2199 2013-09-27 Xinliang David Li <davidxl@google.com>
2201 * opts.c (finish_options): Adjust parameters
2202 according to vect cost model.
2203 (common_handle_option): Set dynamic vect cost
2205 targhooks.c (default_add_stmt_cost): Compute stmt cost
2207 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
2208 Use helper function.
2209 * tree-vectorizer.h (unlimited_cost_model): New function.
2210 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
2211 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
2213 (vect_enhance_data_refs_alignment): Ditto.
2214 * flag-types.h: New enum.
2215 * common/config/i386/i386-common.c (ix86_option_init_struct):
2216 No need to initialize vect_cost_model flag.
2217 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
2220 2013-09-27 Diego Novillo <dnovillo@google.com>
2222 * gimple.h (enum ssa_mode): Remove.
2224 2013-09-27 Paulo Matos <pmatos@broadcom.com>
2227 * gcc.dg/pr58463.c: New test.
2229 2013-09-27 Paulo Matos <pmatos@broadcom.com>
2231 * cfgloop.h (number_of_loops): Fix typo in check for null.
2233 2013-09-27 Jakub Jelinek <jakub@redhat.com>
2236 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
2237 are children of outermost saved_cfun's loop, and set it up to
2238 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
2239 if loop != loop0 and SESE region contains bbs that belong to loop0.
2241 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
2243 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
2244 (binary_scale_code_p, get_base_term, get_index_term): New functions.
2245 (set_address_segment, set_address_base, set_address_index)
2246 (set_address_disp): Accept the argument unconditionally.
2247 (baseness): Remove must_be_base_p and must_be_index_p checks.
2248 (decompose_normal_address): Classify as much as possible in the
2251 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
2253 * cse.c (count_reg_usage): Handle INT_LIST.
2254 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
2255 * reginfo.c (reg_scan_mark_refs): Likewise.
2256 * reload1.c (eliminate_regs_1): Likewise.
2258 2013-09-27 Iain Sandoe <iain@codesourcery.com>
2261 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
2264 2013-09-27 Richard Biener <rguenther@suse.de>
2266 PR tree-optimization/58459
2267 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
2268 restriction not propagating into loops.
2270 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
2272 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
2273 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
2274 * doc/tree-ssa.texi (Walking use-def chains): Delete.
2276 2013-09-26 Richard Biener <rguenther@suse.de>
2278 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
2280 2013-09-26 Richard Biener <rguenther@suse.de>
2282 * alias.h (component_uses_parent_alias_set): Rename to ...
2283 (component_uses_parent_alias_set_from): ... this.
2284 * alias.c (component_uses_parent_alias_set): Rename to ...
2285 (component_uses_parent_alias_set_from): ... this and return
2287 (reference_alias_ptr_type_1): Use the result from
2288 component_uses_parent_alias_set_from instead of stripping
2289 components one at a time.
2290 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
2292 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
2294 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
2295 Move prototypes to...
2296 * tree-ssa-ter.h: New File. Move prototypes here.
2297 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
2298 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
2299 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
2300 from is_replaceable_p.
2301 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
2302 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
2303 (process_replaceable): Use ter_is_replaceable_p.
2304 (find_replaceable_in_bb): Use ter_is_replaceable_p.
2305 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
2306 newly refactored ssa_is_replaceable_p.
2307 * cfgexpand.c: Include tree-outof-ssa.h.
2308 * ssaexpand.h: Delete.
2310 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
2312 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
2313 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
2315 (create_gimple_tmp): Delete.
2316 (get_expr_type, build_assign, build_type_cast): Move to...
2317 * gimple-builder.c: New File.
2318 (get_expr_type): Relocate from gimple.c.
2319 (build_assign, build_type_cast): Change to only create ssanames.
2320 * gimple.h: Move prototypes to...
2321 * gimple-builder.h: New File. Here.
2322 * tree-ssa.h: And here.
2323 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
2324 count_uses_and_derefs): Relocate from gimple.c.
2325 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
2326 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
2327 * tree-ssa-math-opts (execute_cse_reciprocals): Use
2328 gimple_replace_ssa_lhs.
2329 * asan.c: Include gimple-builder.h.
2330 * Makefile.in: Add gimple-builder.o.
2332 2013-09-26 Richard Biener <rguenther@suse.de>
2334 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
2335 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
2336 (loe_visit_block): Use gcc_checking_assert.
2337 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
2338 gimple_assign_ssa_name_copy_p.
2339 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
2342 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
2344 * config/rs6000/t-rs6000 (rs6000.o): Remove.
2345 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
2347 2013-09-26 Richard Biener <rguenther@suse.de>
2349 PR tree-optimization/58539
2350 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
2351 the fact that debug statements are not taking part in loop-closed
2354 2013-09-26 Nick Clifton <nickc@redhat.com>
2356 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
2357 time warning message.
2358 (msp430_print_operand_raw): Delete unused letter parameter.
2359 (TARGET_PRINT_OPERAND_ADDRESS): Define.
2360 (msp430_print_operand_address): New function.
2361 (msp430_print_operand): Move address printing code from here to
2363 * config/msp430/msp430.md (movsipsi2): Add comment in generated
2365 (zero_extendpsisi2): Likewise.
2366 (extendpsisi2): New pattern.
2367 (andneghi3): New pattern.
2369 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
2371 * config/aarch64/aarch64.opt (mlra): New option.
2372 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
2373 (TARGET_LRA_P): Define.
2375 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
2377 * expr.c (expand_assignment): Remove obsolete comment.
2379 2013-09-25 Jeff Law <law@redhat.com>
2381 * tree-flow.h (thread_through_all_blocks): Prototype moved into
2382 tree-ssa-threadupdate.h.
2383 (register_jump_thread): Similarly.
2384 * tree-ssa-threadupdate.h: New header file.
2385 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
2386 * tree-vrp.c: Likewise.
2387 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
2388 (thread_around_empty_blocks): Change type of path vector argument to
2389 an edge,type pair from just an edge. Initialize both elements when
2390 appending to a jump threading path. Tweak references to elements
2392 (thread_across_edge): Similarly. Release memory for the elements
2394 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
2395 (dump_jump_thread_path): New function broken out from
2396 register_jump_thread.
2397 (register_jump_thread): Use dump_jump_thread_path. Change type of
2398 path vector entries. Search the path for NULL edges and dump
2399 the path if one is found. Tweak the conversion of path to 3-edge
2400 form to use the block copy type information embedded in the path.
2402 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
2404 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
2406 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
2407 Vladimir Makarov <vmakarov@redhat.com>
2409 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
2410 from the least significant bit.
2411 (strip_address_mutations): Add bitfield operations handling.
2412 (must_be_index_p): Add shifting and rotate operations handling.
2413 (set_address_base): Use must_be_base_p predicate.
2414 (set_address_index): Use must_be_index_p predicate.
2416 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
2417 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2418 Sergey Lega <sergey.s.lega@intel.com>
2419 Anna Tikhonova <anna.tikhonova@intel.com>
2420 Ilya Tocar <ilya.tocar@intel.com>
2421 Andrey Turetskiy <andrey.turetskiy@intel.com>
2422 Ilya Verbin <ilya.verbin@intel.com>
2423 Kirill Yukhin <kirill.yukhin@intel.com>
2424 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2426 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
2428 (ix86_expand_vector_move_misalign): Support new unaligned load and
2429 stores and use new names.
2430 (CODE_FOR_sse2_storedqu): Rename to ...
2431 (CODE_FOR_sse2_storedquv16qi): ... this.
2432 (CODE_FOR_sse2_loaddqu): Rename to ...
2433 (CODE_FOR_sse2_loaddquv16qi): ... this.
2434 (CODE_FOR_avx_loaddqu256): Rename to ...
2435 (CODE_FOR_avx_loaddquv32qi): ... this.
2436 (CODE_FOR_avx_storedqu256): Rename to ...
2437 (CODE_FOR_avx_storedquv32qi): ... this.
2438 * config/i386/i386.md (fpint_logic): New.
2439 * config/i386/sse.md (VMOVE): Extend for AVX512.
2443 (VI_UNALIGNED_LOADSTORE): Ditto.
2444 (sse2_avx_avx512f): Ditto.
2445 (sse2_avx2): Extend for AVX512.
2446 (sse4_1_avx2): Ditto.
2447 (avx2_avx512f): New.
2448 (sse): Extend for AVX512.
2451 (avxsizesuffix): Ditto.
2452 (sseintvecmode): Ditto.
2454 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
2455 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
2456 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
2457 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
2458 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
2459 (<sse2_avx_avx512f>_storedqu<mode): ... this.
2460 (<sse>_movnt<mode>): Replace constraint "x" with "v".
2461 (STORENT_MODE): Extend for AVX512.
2462 (*absneg<mode>2): Replace constraint "x" with "v".
2463 (*mul<mode>3): Ditto.
2464 (*ieee_smin<mode>3): Ditto.
2465 (*ieee_smax<mode>3): Ditto.
2466 (avx_cmp<mode>3): Replace VF with VF_128_256.
2467 (*<sse>_maskcmp<mode>3_comm): Ditto.
2468 (<sse>_maskcmp<mode>3): Ditto.
2469 (<sse>_andnot<mode>3): Extend for AVX512.
2470 (<code><mode>3, anylogic): Replace VF with VF_128_256.
2471 (<code><mode>3, fpint_logic): New.
2472 (*<code><mode>3): Extend for AVX512.
2473 (avx512flogicsuff): New.
2474 (avx512f_<logic><mode>): Ditto.
2475 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
2477 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
2478 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
2479 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
2480 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
2481 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
2482 (xop_vpermil2<mode>3): Ditto.
2483 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
2484 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
2485 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
2486 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
2488 2013-09-25 Tom Tromey <tromey@redhat.com>
2490 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
2491 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
2492 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
2493 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
2494 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
2495 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
2496 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
2497 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
2498 (GRAPHITE_HTAB_H): Remove.
2500 2013-09-25 Tom Tromey <tromey@redhat.com>
2502 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
2504 2013-09-25 Tom Tromey <tromey@redhat.com>
2506 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
2508 2013-09-25 Tom Tromey <tromey@redhat.com>
2510 * config/i386/t-i386 (i386.o): Remove.
2511 (i386-c.o): Use COMPILE and POSTCOMPILE.
2513 2013-09-25 Tom Tromey <tromey@redhat.com>
2515 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
2517 2013-09-25 Tom Tromey <tromey@redhat.com>
2519 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
2520 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
2521 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
2522 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
2523 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
2524 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
2525 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
2526 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
2527 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
2528 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
2529 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
2530 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
2531 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
2532 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
2533 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
2534 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
2535 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
2536 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
2537 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
2538 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
2539 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
2540 (tree-ssa-pre.o, tree-ssa-sccvn.o)
2541 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
2542 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
2543 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
2544 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
2545 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
2546 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
2547 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
2548 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
2549 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
2550 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
2551 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
2552 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
2553 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
2554 (graphite-clast-to-gimple.o, graphite-dependences.o)
2555 (graphite-interchange.o, graphite-poly.o)
2556 (graphite-scop-detection.o, graphite-sese-to-poly.o)
2557 (graphite-optimize-isl.o, tree-vect-loop.o)
2558 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
2559 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
2560 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
2561 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
2562 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
2563 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
2564 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
2565 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
2566 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
2567 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
2568 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
2569 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
2570 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
2571 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
2572 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
2573 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
2574 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
2575 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
2576 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
2577 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
2578 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
2579 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
2580 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
2581 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
2582 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
2583 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
2584 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
2585 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
2586 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
2587 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
2588 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
2589 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
2590 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
2591 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
2592 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
2593 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
2594 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
2595 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
2596 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
2597 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
2598 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
2599 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
2600 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
2601 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
2602 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
2603 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
2604 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
2605 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
2606 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
2607 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
2608 (gcov-dump.o): Remove.
2609 (default-c.o): Use COMPILE and POSTCOMPILE.
2610 (CFLAGS-gcc.o): New variable.
2611 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
2613 2013-09-25 Tom Tromey <tromey@redhat.com>
2615 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
2616 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
2617 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
2618 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
2619 (c-family/c-pragma.o, c-family/c-pretty-print.o)
2620 (c-family/c-semantics.o, c-family/c-ada-spec.o)
2621 (c-family/array-notation-common.o, c-family/stub-objc.o)
2622 (c-family/c-ubsan.o): Remove.
2624 2013-09-25 Tom Tromey <tromey@redhat.com>
2626 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
2628 2013-09-25 Tom Tromey <tromey@redhat.com>
2630 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
2631 to add -DENABLE_SHARED_LIBGCC.
2632 (gcc.o): Don't use subshell.
2634 2013-09-25 Tom Tromey <tromey@redhat.com>
2636 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
2637 * configure.ac: Don't invoke AM_PROG_CC_C_O.
2638 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
2639 * configure, config.in: Rebuild.
2641 2013-09-25 Tom Tromey <tromey@redhat.com>
2643 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
2644 (COMPILE, POSTCOMPILE): New variables.
2645 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
2646 (DEPFILES): New variable.
2647 Include ".Po" files.
2648 * configure.ac: Add checks for dependency checking.
2649 * configure, aclocal.m4: Regenerate.
2651 2013-09-25 Tom Tromey <tromey@redhat.com>
2653 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
2654 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
2656 2013-09-25 Tom Tromey <tromey@redhat.com>
2658 * Makefile.in (generated_files): Add options.h,
2659 target-hooks-def.h, insn-opinit.h,
2660 common/common-target-hooks-def.h, pass-instances.def,
2661 c-family/c-target-hooks-def.h.
2663 2013-09-25 Jeff Law <law@redhat.com>
2665 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
2666 than foo[foo.length () - 1] to access last member in a vec.
2667 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
2669 2013-09-25 Richard Biener <rguenther@suse.de>
2672 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
2674 2013-09-25 Jan Hubicka <jh@suse.cz>
2676 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
2679 2013-09-25 Marek Polacek <polacek@redhat.com>
2682 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
2683 when determining the type name.
2685 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
2687 * config/sh/sh.md: Fix formatting.
2689 2013-09-24 Xinliang David Li <davidxl@google.com>
2691 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
2692 max peel iterations parameter.
2693 * param.def: New parameter.
2694 * doc/invoke.texi: Document New parameter.
2696 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
2698 * gimple-pretty-print.c: Various whitespace tweaks.
2699 * tree-core.h: Likewise.
2700 * tree-pretty-print.c: Likewise.
2701 * tree-ssa-alias.c: Likewise.
2702 * tree-ssa-copy.c: Likewise.
2703 * tree-ssanames.c: Likewise.
2704 * tree-ssanames.h: Likewise.
2705 * tree-vrp.c: Likewise.
2707 2013-09-24 Alan Modra <amodra@gmail.com>
2711 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
2712 for output operands that disallow regs. Don't use EXPAND_WRITE on
2715 2013-09-24 Richard Biener <rguenther@suse.de>
2718 * tree.c (reference_alias_ptr_type): Move ...
2719 * alias.c (reference_alias_ptr_type): ... here and implement
2720 in terms of the new reference_alias_ptr_type_1.
2721 (ref_all_alias_ptr_type_p): New helper.
2722 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
2723 use ref_all_alias_ptr_type_p.
2724 (get_deref_alias_set): Add flag_strict_aliasing check here.
2725 (reference_alias_ptr_type_1): New function, split out from ...
2726 (get_alias_set): ... here.
2727 (alias_ptr_types_compatible_p): New function.
2728 * alias.h (reference_alias_ptr_type): Declare.
2729 (alias_ptr_types_compatible_p): Likewise.
2730 * tree.h (reference_alias_ptr_type): Remove.
2731 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
2732 to compare MEM_REF alias types.
2734 2013-09-24 Richard Biener <rguenther@suse.de>
2736 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
2738 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2740 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
2741 reload helper function arrays into a single array reg_addr.
2742 (reload_fpr_gpr): Likewise.
2743 (reload_gpr_vsx): Likewise.
2744 (reload_vsx_gpr): Likewise.
2745 (struct rs6000_reg_addr): Likewise.
2746 (reg_addr): Likewise.
2747 (rs6000_debug_reg_global): Change rs6000_vector_reload,
2748 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
2749 (rs6000_init_hard_regno_mode_ok): Likewise.
2750 (rs6000_secondary_reload_direct_move): Likewise.
2751 (rs6000_secondary_reload): Likewise.
2753 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
2754 constraints: wu, ww, and wy. Repurpose wv constraint added during
2755 power8 changes. Put wg constraint in alphabetical order.
2757 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
2758 for future work to add ISA 2.07 VSX single precision support.
2759 (-mvsx-scalar-double): Change default from -1 to 1, update
2760 documentation comment.
2761 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
2762 (-mupper-regs-df): New debug switch to control whether DF values
2763 can go in the traditional Altivec registers.
2764 (-mupper-regs-sf): New debug switch to control whether SF values
2765 can go in the traditional Altivec registers.
2767 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
2769 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
2770 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
2771 Add new constraints, wu/ww/wy. Repurpose wv constraint.
2772 (rs6000_debug_legitimate_address_p): Print if we are running
2773 before, during, or after reload.
2774 (rs6000_secondary_reload): Add a comment.
2775 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
2777 * config/rs6000/constraints.md (wa constraint): Sort w<x>
2778 constraints. Update documentation string.
2779 (wd constraint): Likewise.
2780 (wf constraint): Likewise.
2781 (wg constraint): Likewise.
2782 (wn constraint): Likewise.
2783 (ws constraint): Likewise.
2784 (wt constraint): Likewise.
2785 (wx constraint): Likewise.
2786 (wz constraint): Likewise.
2787 (wu constraint): New constraint for ISA 2.07 SFmode scalar
2789 (ww constraint): Likewise.
2790 (wy constraint): Likewise.
2791 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
2792 the previous submissions.
2793 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
2795 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
2797 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
2798 (REG_BR_PROB): Say that the probability is stored in an int_list.
2799 * reg-notes.def: Update commentary to mention INT_LIST.
2800 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
2801 (INT_LIST): New rtx.
2802 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
2803 * rtlanal.c (int_reg_note_p): New function.
2804 (alloc_reg_note): Assert that the note does not have an int argument.
2805 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
2806 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
2807 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
2808 rather than an INSN_LIST. Handle INT_LIST.
2809 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
2810 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
2811 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
2812 Manipulate them as ints rather than rtxes.
2813 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
2814 * regmove.c (copy_src_to_dest): Likewise.
2815 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
2817 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
2818 into the cases that need it.
2819 * config/arm/arm.c (arm_unwind_emit): Likewise.
2821 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
2822 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
2823 * loop-doloop.c (add_test, doloop_modify): Likewise.
2824 * loop-unswitch.c (compare_and_jump_seq): Likewise.
2825 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
2826 * predict.c (combine_predictions_for_insn): Likewise.
2827 * print-rtl.c (print_rtx): Handle INT_LIST.
2828 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
2829 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
2830 * config/arm/arm.c (emit_unlikely_jump): Likewise.
2831 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
2832 (ix86_split_fp_branch, predict_jump): Likewise.
2833 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
2834 * config/sh/sh.c (expand_cbranchsi4): Likewise.
2835 * config/spu/spu.c (ea_load_store_inline): Likewise.
2837 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
2838 value of a REG_BR_PROB note.
2839 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
2840 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
2841 * emit-rtl.c (try_split): Likewise.
2842 * predict.c (br_prob_note_reliable_p): Likewise.
2843 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
2844 * reorg.c (mostly_true_jump): Likewise.
2845 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
2846 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
2847 * config/i386/i386.c (ix86_print_operand): Likewise.
2848 * config/ia64/ia64.c (ia64_print_operand): Likewise.
2849 * config/mmix/mmix.c (mmix_print_operand): Likewise.
2850 * config/rs6000/rs6000.c (output_cbranch): Likewise.
2851 * config/s390/s390.c (s390_expand_tbegin): Likewise.
2852 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
2853 * config/sparc/sparc.c (output_cbranch): Likewise.
2854 * config/spu/spu.c (get_branch_target): Likewise.
2855 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
2856 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
2858 2013-09-23 Jan Hubicka <jh@suse.cz>
2860 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
2862 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
2863 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
2865 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
2866 (gimple_fold_call): Dump inconsistent devirtualizations; add
2867 sanity check for type based devirtualizations.
2868 * ipa-prop.c: Include ipa-utils.h
2869 (ipa_intraprocedural_devirtualization): Add sanity check.
2870 (try_make_edge_direct_virtual_call): Likewise.
2872 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
2874 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
2875 assignment statements.
2877 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
2879 * gimple-pretty-print.c (dump_ssaname_info): New function.
2880 (dump_gimple_phi): Call it.
2881 (pp_gimple_stmt_1): Likewise.
2882 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
2883 (range_info_def): Declare.
2884 * tree-pretty-print.c (pp_double_int): New function.
2885 (dump_generic_node): Call it.
2886 * tree-pretty-print.h (pp_double_int): Declare.
2887 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
2888 * tree-ssanames.h (range_info_def): New structure.
2889 (value_range_type): Move definition here.
2890 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
2892 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
2894 (set_range_info): New function.
2895 (get_range_info): Likewise.
2896 (duplicate_ssa_name_range_info): Likewise.
2897 (duplicate_ssa_name_fn): Check pointer type and call
2898 duplicate_ssa_name_range_info.
2899 * tree-ssa-copy.c (fini_copy_prop): Likewise.
2900 * tree-vrp.c (value_range_type): Remove definition, now in
2902 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
2903 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
2904 (SSA_NAME_RANGE_INFO): New macro.
2906 2013-09-23 Richard Biener <rguenther@suse.de>
2908 PR tree-optimization/58464
2909 * tree-ssa-pre.c (phi_trans_lookup): Remove.
2910 (phi_trans_add): Change to add conditionally on being not present.
2911 (phi_translate_1): Remove recursion detection here.
2912 (phi_translate): Pre-seed the cache with NULL to catch
2913 recursion here in a more generic way.
2914 (bitmap_find_leader): Adjust comment.
2915 (get_representative_for): Dump value-numbers.
2916 (create_expression_by_pieces): Likewise.
2917 (insert_into_preds_of_block): Likewise.
2919 2013-09-23 Christian Bruel <christian.bruel@st.com>
2922 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
2923 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
2925 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
2928 2013-09-20 Renlin Li <renlin.li@arm.com>
2930 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2932 (aarch64_expand_epilogue): Likewise.
2933 (aarch64_legitimize_reload_address): Likewise.
2935 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
2937 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
2938 NULL_TREE before pushing them onto the vector. Likewise for labels.
2940 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
2942 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
2943 comparison operators when -fno-trapping-math is in effect.
2944 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
2945 comparison operators in TFmode and assert that unsupported operators
2947 (ia64_print_operand): Likewise.
2949 2013-09-21 Jan Hubicka <jh@suse.cz>
2951 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
2952 (sse_typeless_stores): Enable for core
2953 (sse_load0_by_pxor): Likewise.
2954 (four_jump_limit): Disable for core.
2955 (pad_returns): Likewise.
2956 (avoid_vector_decode): Likewise.
2957 (fuse_cmp_and_branch): Enable for cores.
2958 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
2960 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2963 * config/pa/pa.c (pa_option_override): Disable auto increment and
2964 decrement instructions until reload is completed.
2966 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
2967 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
2968 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
2970 2013-09-20 DJ Delorie <dj@redhat.com>
2971 Nick Clifton <nickc@redhat.com>
2973 * config/rl78/rl78.c: Various whitespace and comment tweaks.
2974 (need_to_save): Save bank 0 on interrupts.
2975 (characterize_address): Strip far address wrappers.
2976 (rl78_as_legitimate_address): Likewise.
2977 (transcode_memory_rtx): Likewise.
2978 (rl78_peep_movhi_p): Disable this peephole after devirt.
2979 (rl78_propogate_register_origins): Forget all origins when a
2981 * config/rl78/rl78-virt.md: Various whitespace tweaks.
2982 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
2984 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
2985 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
2986 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
2988 * config/rl78/constraints.md: Various whitespace and paren tweaks.
2990 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2992 * config/pa/pa.md: In "scc" insn patterns, change output template to
2993 handle const0_rtx in reg_or_0_operand operands.
2995 2013-09-20 Martin Husemann <martin@NetBSD.org>
2998 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
2999 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
3001 2013-09-20 Richard Biener <rguenther@suse.de>
3004 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
3005 remembering SSA name version and block index.
3006 (new_scev_info_str): Adjust.
3007 (hash_scev_info): Likewise. Also hash the block index.
3008 (eq_scev_info): Adjust.
3009 (find_var_scev_info): Likewise.
3010 (struct instantiate_cache_entry): Remove.
3011 (struct instantiate_cache_type): Use a htab to map name, block
3013 (instantiate_cache_type::~instantiate_cache_type): Adjust.
3014 (get_instantiated_value_entry): Likewise.
3015 (hash_idx_scev_info, eq_idx_scev_info): New functions.
3016 (instantiate_scev_name): Adjust.
3018 2013-09-20 Jeff Law <law@redhat.com>
3020 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
3022 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
3024 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
3025 Call aarch64_simd_expand_args to update op[argc].
3027 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3029 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
3032 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3034 * gengtype.c (file_rules): Added rule for *.cc files.
3035 (get_output_file_with_visibility): Give fatal message when no
3038 2013-09-20 Renlin Li <renlin.li@arm.com>
3040 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
3041 (aarch64_expand_epilogue): Likewise.
3042 (aarch64_legitimize_reload_address): Likewise.
3044 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
3047 * expr.c (expand_assignment): Remove misalignp code path.
3049 2013-09-20 Marek Polacek <polacek@redhat.com>
3052 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
3053 TYPE_PRECISION. Add asserts.
3055 2013-09-20 Richard Biener <rguenther@suse.de>
3057 PR tree-optimization/58453
3058 * tree-loop-distribution.c (distribute_loop): Apply the cost
3059 model for -ftree-loop-distribute-patterns, too.
3061 2013-09-20 Richard Biener <rguenther@suse.de>
3064 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
3065 make type comparison less strict.
3067 2013-09-20 Alan Modra <amodra@gmail.com>
3069 * configure: Regenerate.
3070 * aclocal.m4: Regenerate.
3072 2013-09-20 Marek Polacek <polacek@redhat.com>
3075 * doc/extend.texi: Document that attribute used is meant to be used
3076 on variables with static storage duration.
3078 2013-09-19 Jakub Jelinek <jakub@redhat.com>
3080 PR tree-optimization/58472
3081 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
3082 simd_lane_access set inv_p = false.
3083 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
3084 the simduid magic VAR_DECL.
3086 2013-09-19 Jan Hubicka <jh@suse.cz>
3088 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
3090 2013-09-17 Jeff Law <law@redhat.com>
3092 * tree-ssa-dom.c (record_temporary_equivalences): New function
3093 split out of dom_opt_dom_walker::after_dom_children.
3094 (dom_opt_dom_walker::thread_across_edge): Move common code
3095 in here from dom_opt_dom_walker::after_dom_children.
3096 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
3098 2013-09-19 Jan Hubicka <jh@suse.cz>
3100 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
3101 (TARGET_GENERIC): Use PROCESOR_GENERIC
3102 (enum processor_type): Unify generic32 and 64.
3103 * i386.md (cpu): Likewise.
3104 * x86-tune.def (use_leave): Enable for generic32.
3105 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
3106 * athlon.md: Change generic64 to generic in all occurences.
3107 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
3108 (ix86_target_macros_internal): Likewise.
3109 * driver-i386.c (host_detect_local_cpu): Likewise.
3110 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
3112 (generic_memcpy, generic_memset, generic_cost): This one.
3113 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
3114 (m_GENERIC32, m_GENERIC64): Remove.
3115 (m_GENERIC): Turn into one flag.
3116 (processor_target): Unify generic tunnings.
3117 (ix86_option_override_internal): Replace generic32/64 by generic.
3118 (ix86_issue_rate): Likewise.
3119 (ix86_adjust_cost): Likewise.
3121 2013-09-19 Jan Hubicka <jh@suse.cz>
3123 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
3124 of speculative flag.
3126 2013-09-19 Jakub Jelinek <jakub@redhat.com>
3128 * omp-low.c (expand_omp_sections): Always pass len - 1 to
3129 GOMP_sections_start, even if !exit_reachable.
3131 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
3133 * lra-constraints.c (need_for_all_save_p): Use macro
3134 HARD_REGNO_CALL_PART_CLOBBERED.
3135 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
3136 set up pseudo conflict hard regs.
3138 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
3141 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
3144 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
3146 PR rtl-optimization/58438
3147 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
3148 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
3149 from previous subpasses.
3151 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
3153 * arm.c (arm_get_frame_offsets): Validate architecture supports
3154 LDRD/STRD before accepting the tuning preference.
3155 (arm_expand_prologue): Likewise.
3156 (arm_expand_epilogue): Likewise.
3158 2013-09-18 Richard Biener <rguenther@suse.de>
3160 PR tree-optimization/58417
3161 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
3162 have chrecs with symbols defined in the loop as operands.
3163 (chrec_fold_multiply): Likewise.
3164 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
3165 parameters before folding binary operations.
3166 (struct instantiate_cache_entry_hasher): Remove.
3167 (struct instantiate_cache_type): Use a pointer-map.
3168 (instantiate_cache_type::instantiate_cache_type): New function.
3169 (instantiate_cache_type::get): Likewise.
3170 (instantiate_cache_type::set): Likewise.
3171 (instantiate_cache_type::~instantiate_cache_type): Adjust.
3172 (get_instantiated_value_entry): Likewise.
3173 (global_cache): New global.
3174 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
3175 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
3176 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
3177 (instantiate_scev_name): Adjust.
3178 (instantiate_scev): Construct global instead of local cache.
3179 (resolve_mixers): Likewise.
3181 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
3182 Paolo Carlini <paolo.carlini@oracle.com>
3185 * doc/implement-cxx.texi: Fix references to the C++ standards.
3187 2013-09-18 Jakub Jelinek <jakub@redhat.com>
3189 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
3190 * tree-vect-data-refs.c (vect_analyze_data_refs): For
3191 simd_lane_access drs, update also DR_ALIGNED_TO.
3193 2013-09-18 Marek Polacek <polacek@redhat.com>
3196 * doc/extend.texi: Document no_sanitize_undefined attribute.
3197 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
3198 no_sanitize_undefined attribute.
3200 2013-09-18 Nick Clifton <nickc@redhat.com>
3202 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
3203 (ASM_DECLARE_FUNCTION_NAME): Define.
3205 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
3207 * compare-elim.c (find_comparison_dom_walker): New class
3208 (find_comparisons_in_bb): Rename to
3209 find_comparison_dom_walker::before_dom_children
3210 (find_comparisons): Adjust
3211 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
3213 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
3214 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
3215 (init_walk_dominator_tree): Remove declaration.
3216 (fini_walk_dominator_tree): Remove declaration.
3217 * fwprop.c (single_def_use_dom_walker): New class
3218 (single_def_use_enter_block): Convert to
3219 single_def_use_dom_walker::before_dom_children.
3220 (single_def_use_leave_block): Convert to
3221 single_def_use_dom_walker::after_dom_children.
3222 (build_single_def_use_links): Adjust.
3223 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
3225 (find_candidates_in_block): Convert to
3226 find_candidates_dom_walker::before_dom_children.
3227 (execute_strength_reduction): Adjust.
3228 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
3229 (sese_dom_walker): New class.
3230 (sese_dom_walker::sese_dom_walker): New constructor.
3231 (sese_dom_walker::~sese_dom_walker): New destructor.
3232 (build_sese_conditions_before): Convert to
3233 sese_dom_walker::before_dom_children.
3234 (build_sese_conditions_after): Convert to
3235 sese_dom_walker::after_dom_children.
3236 (build_poly_scop): Adjust
3237 * tree-into-ssa.c (rewrite_dom_walker): New class
3238 (rewrite_enter_block): Convert to
3239 rewrite_dom_walker::before_dom_children.
3240 (rewrite_leave_block): Convert to
3241 rewrite_dom_walker::after_dom_children.
3242 (rewrite_update_dom_walker): New class.
3243 (rewrite_update_enter_block): Convert to
3244 rewrite_update_dom_walker::before_dom_children.
3245 (rewrite_update_leave_block): Convert to
3246 rewrite_update_dom_walker::after_dom_children.
3247 (rewrite_blocks, rewrite_into_ssa): Adjust.
3248 (mark_def_dom_walker): New class.
3249 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
3250 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
3251 (mark_def_sites_blocks): Convert to
3252 mark_def_dom_walker::before_dom_children.
3253 (mark_def_site_blocks): Remove.
3254 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
3255 (tree_ssa_dominator_optimize): Adjust.
3256 (dom_thread_across_edge): Convert to method
3257 dom_opt_dom_walker::thread_across_edge.
3258 (dom_opt_enter_block): Convert to member function
3259 dom_opt_dom_walker::before_dom_children.
3260 (dom_opt_leave_block): Convert to member function
3261 dom_opt_dom_walker::after_dom_children.
3262 * tree-ssa-dse.c (dse_dom_walker): New class.
3263 (dse_enter_block): Convert to member function
3264 dse_dom_walker::before_dom_children.
3265 (tree_ssa_dse): Adjust.
3266 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
3267 (determine_invariantness_stmt): Convert to method
3268 invariantness_dom_walker::before_dom_children.
3269 (determine_invariantness): Remove
3270 (move_computations_dom_walker): New class.
3271 (move_computations_stmt): Convert to method
3272 move_computations_dom_walker::before_dom_children.
3273 (move_computations, tree_ssa_lim): Adjust.
3274 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
3275 (nt_init_block): Make method
3276 notrappping_dom_walker::before_dom_children.
3277 (nt_fini_block): Make
3278 method nontrapping_dom_walker::after_dom_children.
3279 (get_non_trapping): Adjust.
3280 * tree-ssa-pre.c (eliminate_dom_walker): New class.
3281 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
3282 (eliminate_leave_block): Make method.
3283 eliminate_dom_walker::after_dom_children.
3285 * tree-ssa-strlen.c (strlen_dom_walker): New class.
3286 (strlen_enter_block): Make method
3287 strlen_dom_walker::before_dom_children.
3288 (strlen_leave_block): Make
3289 method strlen_dom_walker::after_dom_children.
3290 (tree_ssa_strlen): Adjust.
3291 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
3292 (tree_ssa_uncprop): Adjust.
3293 (uncprop_leave_block): Make method
3294 uncprop_dom_walker::after_dom_children.
3295 (uncprop_leave_block): Make method
3296 uncprop_dom_walker::before_dom_children.
3298 2013-09-18 Bin Cheng <bin.cheng@arm.com>
3300 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
3301 branch in same basic block. Check both src and dest of the move insn.
3303 2013-09-17 Nick Clifton <nickc@redhat.com>
3305 * config/rl78/rl78-real.md (bf): New pattern.
3307 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
3308 (rl78_print_operand): Do not put a # before a %B.
3309 * config/rl78/rl78.opt: Tweak doc strings.
3311 2013-09-17 DJ Delorie <dj@redhat.com>
3313 * config/rl78/constraints.md (Wcv): Allow up to $r31.
3314 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
3315 (rl78_option_override): Likewise, if -mallregs.
3316 (is_virtual_register): Likewise.
3317 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
3318 (REGNO_OK_FOR_BASE_P): Likewise.
3319 * config/rl78/rl78.opt (-mallregs): New.
3321 2013-09-17 Nick Clifton <nickc@redhat.com>
3323 * config/rl78/rl78.c (need_to_save): Change return type to bool.
3324 For interrupt functions: save all call clobbered registers if the
3325 interrupt handler is not a leaf function.
3326 (rl78_expand_prologue): Always recompute the frame information.
3327 For interrupt functions: only select bank 0 if one of the bank 0
3328 registers is going to be psuhed.
3330 2013-09-17 DJ Delorie <dj@redhat.com>
3332 * config/rl78/constraints.md: For each W* constraint, rename to C*
3333 and create a W* constraint that checks for an optional ES: prefix
3335 * config/rl78/rl78.md (UNS_ES_ADDR): New.
3336 (es_addr): New. Used to wrap far addresses.
3337 * config/rl78/rl78-protos.h (rl78_es_addr): New.
3338 (rl78_es_base): New.
3339 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
3340 wrapped far addresses.
3341 (rl78_print_operand_1): Unwrap far addresses before processing.
3342 (rl78_lo16): Wrap far addresses in unspecs.
3343 (rl78_es_addr): New.
3344 (rl78_es_base): New.
3345 (insn_ok_now): Check for not-yet-wrapped far addresses.
3346 (transcode_memory_rtx): Properly re-wrap far addresses.
3348 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
3350 * config/sparc/t-rtems: Add leon3 multilibs.
3352 2013-09-17 Cong Hou <congh@google.com>
3354 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
3355 when checking the dot production pattern. The type of rhs operand
3356 of multiply is now checked correctly.
3358 2013-09-17 Jeff Law <law@redhat.com>
3360 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3361 edge implied equivalences into successor phis.
3362 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
3363 here from tree-ssa-threadedge.c.
3364 (mark_threaded_blocks): When threading through a joiner, if both
3365 successors of the joiner's clone reach the same block, verify the
3366 PHI arguments are equal. If not, cancel the jump threading request.
3367 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
3368 tree-ssa-threadupdate.c
3369 (thread_across_edge): Don't check PHI argument equality when
3370 threading through joiner block here.
3372 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
3374 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
3375 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
3376 (warn_uninit, warn_uninitialized_vars,
3377 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
3378 Move to tree-ssa-uninit.c.
3379 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
3380 (has_undefined_value_p): New. Pass dependant parts of
3381 ssa_undefined_value_p.
3382 (uninit_undefined_value_p): Use has_undefined_value_p.
3383 (warn_uninit, warn_uninitialized_vars,
3384 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
3385 Move from tree-ssa.c.
3386 * tree-ssa.h: Adjust prototypes.
3388 2013-09-17 Jan Hubicka <jh@suse.cz>
3391 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
3392 * ipa-inline.c (can_inline_edge_p): Do not downgrade
3393 FUNCTION_NOT_OPTIMIZED.
3394 * ipa-inline-analysis.c (compute_inline_parameters): Function
3395 not optimized is not inlinable unless it is alwaysinline.
3396 (inline_analyze_function): Force calls in not optimized
3397 function not inlinable.
3399 2013-09-17 Jan Hubicka <jh@suse.cz>
3402 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
3404 * ipa.c (function_and_variable_visibility): Likewise.
3405 * ipa-profile.c (ipa_profile): Likewise.
3407 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
3410 * cgraph.c (cgraph_function_body_availability): Check for ifunc
3411 attribute, and don't inline the resolver in this case.
3413 2013-09-17 Teresa Johnson <tejohnson@google.com>
3415 * coverage.c (get_coverage_counts): Add missing newline.
3417 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3419 PR tree-optimization/58088
3420 * fold-const.c (mask_with_trailing_zeros): New function.
3421 (fold_binary_loc): Make sure we don't recurse infinitely
3422 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
3423 Use mask_with_trailing_zeros where appropriate.
3425 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
3427 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
3428 of 'prev' var to get better distance estimation.
3430 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
3432 * tree-inline.h (struct copy_body_data): Add transform_parameter.
3433 * tree-inline.c (is_parameter_of): New predicate.
3434 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
3435 a parameter has been remapped.
3436 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
3437 (optimize_inline_calls): Initialize transform_parameter.
3438 (copy_gimple_seq_and_replace_locals): Likewise.
3439 (tree_function_versioning): Likewise.
3440 (maybe_inline_call_in_expr): Likewise.
3442 2013-09-17 Nick Clifton <nickc@redhat.com>
3444 * config/msp430/msp430-protos.h: Add prototypes for new functions.
3445 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
3447 (is_attr_func): New function.
3448 (msp430_is_interrupt_func): New function.
3449 (is_naked_func): New function.
3450 (is_reentrant_func): New function.
3451 (is_critical_func): New function.
3452 (msp430_start_function): Add annotations for function attributes.
3453 (msp430_attr): New function.
3454 (msp430_attribute_table): New.
3455 (msp430_function_section): New function.
3456 (TARGET_ASM_FUNCTION_SECTION): Define.
3457 (msp430_builtin): New enum.
3458 (msp430_init_builtins): New function.
3459 (msp430_builtin_devl): New function.
3460 (msp430_expand_builtin): New function.
3461 (TARGET_INIT_BUILTINS): Define.
3462 (TARGET_EXPAND_BUILTINS): Define.
3463 (TARGET_BUILTIN_DECL): Define.
3464 (msp430_expand_prologue): Add support for naked, interrupt,
3465 critical and reentrant functions.
3466 (msp430_expand_epilogue): Likewise.
3467 (msp430_print_operand): Handle 'O' character.
3468 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
3470 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
3471 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
3472 (pushm): Use a 'n' rather than an 'i' constraint.
3473 (msp_return): Add generation of the interrupt return instruction.
3474 (disable_interrupts): New pattern.
3475 (enable_interrupts): New pattern.
3476 (push_intr_state): New pattern.
3477 (pop_intr_state): New pattern.
3478 (bic_SR): New pattern.
3479 (bis_SR): New pattern.
3480 * doc/extend.texi: Document MSP430 function attributes and builtin
3483 2013-09-17 Richard Biener <rguenther@suse.de>
3485 PR tree-optimization/58432
3486 * tree-loop-distribution.c (tree_loop_distribution): Also
3487 scan PHIs for outside loop uses and seed a partition from them.
3489 2013-09-17 Bin Cheng <bin.cheng@arm.com>
3491 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
3492 (restructure_reference): Call backtrace_base_for_ref.
3494 2013-09-17 Alan Modra <amodra@gmail.com>
3497 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
3500 2013-09-16 DJ Delorie <dj@redhat.com>
3502 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
3503 vregs location for RL78/G10.
3504 (rl78_expand_prologue): Avoid SEL on G10.
3505 (rl78_expand_epilogue): Likewise.
3506 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
3507 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
3508 __RL78_G10__ when appropriate.
3509 (ASM_SPEC): Pass -mg10 along to the assembler.
3510 * config/rl78/rl78.md (sel_rb): Disable for G10.
3511 * config/rl78/rl78.opt: Add -mg10 option.
3512 * config/rl78/t-rl78: Add -mg10 multilib.
3514 2013-09-16 Xinliang David Li <davidxl@google.com>
3516 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
3517 * omp-low.c (omp_max_vf): Ditto.
3518 (expand_omp_simd): Ditto.
3519 * tree-vectorizer.c (vectorize_loops): Ditto.
3520 (gate_vect_slp): Ditto.
3521 (gate_increase_alignment): Ditto.
3522 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
3523 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
3524 (gate_tree_vectorize): Name change.
3525 (tree_vectorize): Ditto.
3526 (pass_vectorize::gate): Call new function.
3527 (pass_vectorize::execute): Ditto.
3528 * opts.c: O3 default setting change.
3529 (finish_options): Check new flag.
3530 * doc/invoke.texi: Document new flags.
3531 * common.opt: New flags.
3533 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
3535 * doc/tm.texi.in (Cond Exec Macros): Remove node.
3536 (Condition Code): Don't reference it.
3537 * doc/tm.texi: Regenerate.
3539 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
3542 * lra-constraints.c (undo_optional_reloads): Consider all optional
3543 reload even if it did not get a hard reg.
3545 2013-09-16 Teresa Johnson <tejohnson@google.com>
3547 * dumpfile.c (dump_loc): Remove newline emission.
3548 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
3549 emission to dump_printf_loc calls where missing.
3550 (vect_mark_for_runtime_alias_test): Ditto.
3551 (vect_analyze_data_ref_dependence): Ditto.
3552 (vect_analyze_data_ref_dependences): Ditto.
3553 (vect_slp_analyze_data_ref_dependence): Ditto.
3554 (vect_slp_analyze_data_ref_dependences): Ditto.
3555 (vect_compute_data_ref_alignment): Ditto.
3556 (vect_update_misalignment_for_peel): Ditto.
3557 (vect_verify_datarefs_alignment): Ditto.
3558 (vector_alignment_reachable_p): Ditto.
3559 (vect_get_data_access_cost): Ditto.
3560 (vect_enhance_data_refs_alignment): Ditto.
3561 (vect_find_same_alignment_drs): Ditto.
3562 (vect_analyze_data_refs_alignment): Ditto.
3563 (vect_analyze_group_access): Ditto.
3564 (vect_analyze_data_ref_access): Ditto.
3565 (vect_analyze_data_ref_accesses): Ditto.
3566 (vect_prune_runtime_alias_test_list): Ditto.
3567 (vect_analyze_data_refs): Ditto.
3568 (vect_create_addr_base_for_vector_ref): Ditto.
3569 (vect_create_data_ref_ptr): Ditto.
3570 (vect_grouped_store_supported): Ditto.
3571 (vect_grouped_load_supported): Ditto.
3572 * value-prof.c (check_counter): Ditto.
3573 (check_ic_target): Ditto.
3574 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
3575 (vect_recog_widen_mult_pattern): Ditto.
3576 (vect_recog_widen_sum_pattern): Ditto.
3577 (vect_recog_over_widening_pattern): Ditto.
3578 (vect_recog_widen_shift_pattern): Ditto.
3579 (vect_recog_rotate_pattern): Ditto.
3580 (vect_recog_vector_vector_shift_pattern): Ditto.
3581 (vect_recog_divmod_pattern): Ditto.
3582 (vect_recog_mixed_size_cond_pattern): Ditto.
3583 (vect_recog_bool_pattern): Ditto.
3584 (vect_pattern_recog_1): Ditto.
3585 (vect_pattern_recog): Ditto.
3586 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
3587 (vect_is_simple_iv_evolution): Ditto.
3588 (vect_analyze_scalar_cycles_1): Ditto.
3589 (vect_get_loop_niters): Ditto.
3590 (vect_analyze_loop_1): Ditto.
3591 (vect_analyze_loop_form): Ditto.
3592 (vect_analyze_loop_operations): Ditto.
3593 (vect_analyze_loop_2): Ditto.
3594 (vect_analyze_loop): Ditto.
3595 (report_vect_op): Ditto.
3596 (vect_is_slp_reduction): Ditto.
3597 (vect_is_simple_reduction_1): Ditto.
3598 (vect_get_known_peeling_cost): Ditto.
3599 (vect_estimate_min_profitable_iters): Ditto.
3600 (vect_model_reduction_cost): Ditto.
3601 (vect_model_induction_cost): Ditto.
3602 (get_initial_def_for_induction): Ditto.
3603 (vect_create_epilog_for_reduction): Ditto.
3604 (vectorizable_reduction): Ditto.
3605 (vectorizable_induction): Ditto.
3606 (vectorizable_live_operation): Ditto.
3607 (vect_loop_kill_debug_uses): Ditto.
3608 (vect_transform_loop): Ditto.
3609 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
3610 (vect_stmt_relevant_p): Ditto.
3611 (process_use): Ditto.
3612 (vect_mark_stmts_to_be_vectorized): Ditto.
3613 (vect_model_simple_cost): Ditto.
3614 (vect_model_promotion_demotion_cost): Ditto.
3615 (vect_model_store_cost): Ditto.
3616 (vect_get_store_cost): Ditto.
3617 (vect_model_load_cost): Ditto.
3618 (vect_get_load_cost): Ditto.
3619 (vect_init_vector_1): Ditto.
3620 (vect_get_vec_def_for_operand): Ditto.
3621 (vect_finish_stmt_generation): Ditto.
3622 (vectorizable_call): Ditto.
3623 (vectorizable_conversion): Ditto.
3624 (vectorizable_assignment): Ditto.
3625 (vectorizable_shift): Ditto.
3626 (vectorizable_operation): Ditto.
3627 (vectorizable_store): Ditto.
3628 (vectorizable_load): Ditto.
3629 (vectorizable_condition): Ditto.
3630 (vect_analyze_stmt): Ditto.
3631 (vect_transform_stmt): Ditto.
3632 (vect_is_simple_use): Ditto.
3633 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
3634 (vect_can_advance_ivs_p): Ditto.
3635 (vect_update_ivs_after_vectorizer): Ditto.
3636 (vect_do_peeling_for_loop_bound): Ditto.
3637 (vect_gen_niters_for_prolog_loop): Ditto.
3638 (vect_update_inits_of_drs): Ditto.
3639 (vect_create_cond_for_alias_checks): Ditto.
3640 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
3641 (vect_build_slp_tree_1): Ditto.
3642 (vect_supported_load_permutation_p): Ditto.
3643 (vect_analyze_slp_instance): Ditto.
3644 (vect_analyze_slp): Ditto.
3645 (vect_make_slp_decision): Ditto.
3646 (vect_detect_hybrid_slp): Ditto.
3647 (vect_bb_vectorization_profitable_p): Ditto.
3648 (vect_slp_analyze_bb_1): Ditto.
3649 (vect_update_slp_costs_according_to_vf): Ditto.
3650 (vect_get_mask_element): Ditto.
3651 (vect_transform_slp_perm_load): Ditto.
3652 (vect_schedule_slp_instance): Ditto.
3653 (vect_schedule_slp): Ditto.
3654 (vect_slp_transform_bb): Ditto.
3655 * profile.c (read_profile_edge_counts): Ditto.
3656 (compute_branch_probabilities): Ditto.
3657 * coverage.c (get_coverage_counts): Ditto.
3659 2013-09-16 Diego Novillo <dnovillo@google.com>
3661 * tree-core.h: Add missing comment lines from refactoring of tree.h.
3663 2013-09-16 Jan Hubicka <jh@suse.cz>
3665 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
3666 abstract functions; for static functions check the presence of body.
3668 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3670 * config/aarch64/aarch64-simd-builtins.def (fma): New.
3671 * config/aarch64/aarch64-simd.md
3672 (aarch64_mla_elt<mode>): New.
3673 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3674 (aarch64_mls_elt<mode>): Likewise.
3675 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3676 (aarch64_fma4_elt<mode>): Likewise.
3677 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3678 (aarch64_fma4_elt_to_128v2df): Likewise.
3679 (aarch64_fma4_elt_to_64df): Likewise.
3680 (fnma<mode>4): Likewise.
3681 (aarch64_fnma4_elt<mode>): Likewise.
3682 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
3683 (aarch64_fnma4_elt_to_128v2df): Likewise.
3684 (aarch64_fnma4_elt_to_64df): Likewise.
3685 * config/aarch64/iterators.md (VDQSF): New.
3686 * config/aarch64/arm_neon.h
3687 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
3688 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
3690 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3692 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
3693 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3694 (aarch64_mul3_elt_to_128df): Likewise.
3695 (aarch64_mul3_elt_to_64v2df): Likewise.
3696 * config/aarch64/iterators.md (VEL): Also handle DFmode.
3698 (VMUL_CHANGE_NLANES) Likewise.
3701 * config/aarch64/arm_neon.h
3702 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
3704 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3706 * config/aarch64/arm_neon.h
3707 (vcvtx_high_f32_f64): Fix parameters.
3709 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3710 Uros Bizjak <ubizjak@gmail.com>
3712 * config/alpha.c: Include tree-ssanames.h.
3714 2013-09-16 Richard Biener <rguenther@suse.de>
3716 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
3717 (dot_rdg_1): Handle control_dd.
3718 (create_edge_for_control_dependence): New function.
3719 (create_rdg_edges): Add control dependences if asked for.
3720 (build_rdg): Likewise.
3721 (generate_loops_for_partition): If there are not necessary
3722 control stmts remove all their dependencies.
3723 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
3724 (distribute_loop): Pass on control dependences.
3725 (tree_loop_distribution): Compute control dependences and remove
3726 restriction on number of loop nodes.
3728 2013-09-16 Jakub Jelinek <jakub@redhat.com>
3730 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
3733 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
3735 * cse.c (try_const_anchors): Punt on CC modes.
3737 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3739 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
3741 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
3744 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
3747 2013-09-14 Iain Sandoe <iain@codesourcery.com>
3750 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
3752 (darwin_objc1_section): Likewise.
3753 (darwin_file_end): Emit Image Info section when required.
3755 2013-09-14 Jan Hubicka <jh@suse.cz>
3757 * tree-into-ssa.c (gate_into_ssa): New.
3758 (pass_data_build_ssa): Use it.
3759 * cgraph.h (expand_thunk): Update prototype.
3760 * cgraphunit.c (analyze_function): Expand thunks early.
3761 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
3762 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
3763 set lowered flag; do not add new function.
3764 (assemble_thunks_and_aliases): Update.
3765 * tree-ssa.c (gate_init_datastructures): New gate.
3766 (pass_data_init_datastructures): Use it.
3768 2013-09-14 Iain Sandoe <iain@codesourcery.com>
3771 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
3772 xmm register set described in the psABI.
3774 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
3776 * dwarf2out.c (should_emit_struct_debug): Add check
3777 for type_decl variable is not NULL.
3779 2013-09-13 Jacek Caban <jacek@codeweavers.com>
3781 * config.gcc: Use new winnt-c.c target hooks
3782 * config/t-winnt: New file
3783 * config/winnt-c.c: New file
3784 * doc/tm.texi.in: Document new hook
3785 * doc/tm.texi: Regenerated
3787 2013-09-13 Jan Hubicka <jh@suse.cz>
3790 * ipa-inline.c (check_callers): New function.
3791 (check_caller_edge): Remove.
3792 (want_inline_function_to_all_callers_p): Also permit alises that are
3794 (inline_to_all_callers): Terminate the walk when devirtualization
3795 introduce new calls.
3797 2013-09-13 Jan Hubicka <jh@suse.cz>
3799 * ipa-inline-analysis.c (struct growth_data): Add node.
3800 (do_estimate_growth_1): Fix detection of recursion.
3802 2013-09-13 Jakub Jelinek <jakub@redhat.com>
3804 PR tree-optimization/58392
3805 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
3806 to avoid shadowing of outer loop variable. If
3807 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
3808 replace_by_duplicate_decl simduid of loops that have it set and
3809 set dest_cfun->has_simduid_loops and/or
3810 dest_cfun->has_force_vect_loops.
3811 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
3812 instead of maybe_lookup_decl.
3813 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
3814 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
3815 call. Copy over force_vect and copy and remap simduid. Set
3816 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
3817 (copy_cfg_body): Remove blocks_to_copy argument. Use
3818 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
3819 caller. Don't set cfun->has_simduid_loops and/or
3820 cfun->has_force_vect_loops here.
3821 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
3823 (expand_call_inline, tree_function_versioning): Adjust copy_body
3826 2013-09-13 Martin Jambor <mjambor@suse.cz>
3829 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
3830 the assert if the edge was a speculative one.
3832 2013-09-13 Richard Biener <rguenther@suse.de>
3834 * tree-data-ref.h (known_dependences_p): Move here ...
3835 * tree-loop-distribution.c (known_dependences_p): ... from here.
3836 (dump_rdg_component, debug_rdg_component): Remove.
3838 (generate_loops_for_partition): Use gimple_uid instead of
3839 relying on matching stmt visit order.
3840 (rdg_build_partitions): Take starting stmt vector.
3841 (ldist_gen): Merge into ...
3842 (distribute_loop): ... this function. Do not compute starting
3844 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
3846 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3848 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
3849 Set type attribute correctly. Set predicable_short_it attribute.
3850 (cmpsi_shiftsi): Remove %? from output template.
3852 2013-09-13 Richard Biener <rguenther@suse.de>
3854 * tree-loop-distribution.c (struct rdg_component,
3855 rdg_defs_used_in_other_loops_p, free_rdg_components,
3856 rdg_build_components): Remove.
3857 (stmts_from_loop): Do not record virtual PHIs.
3858 (generate_loops_for_partition): Skip virtual PHIs.
3859 (build_rdg_partition_for_component): Rename to ...
3860 (build_rdg_partition_for_vertex): ... this and adjust.
3861 (rdg_build_partitions): Take a vector of starting vertices
3862 instead of components. Remove unnecessary leftover handling.
3863 (ldist_gen): Do not build components or record other stores.
3864 (distribute_loop): Do not distribute loops containing stmts
3867 2013-09-13 Christian Bruel <christian.bruel@st.com>
3870 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
3872 2013-09-13 Kai Tietz <ktietz@redhat.com>
3874 * config.gcc: Separate cases for mingw and cygwin targets,
3875 and add 64-bit cygwin target case.
3877 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
3878 dll-export inline-functions.
3879 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
3882 2013-09-13 Jeff Law <law@redhat.com>
3886 2013-09-06 Jeff Law <law@redhat.com>
3888 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3889 edge implied equivalences into successor phis.
3891 2013-09-12 DJ Delorie <dj@redhat.com>
3893 * config/rl78/rl78-virt.md: Change from | to \; for asm line
3896 2013-09-12 Brooks Moses <bmoses@google.com>
3899 * Makefile.in: Do not install driver binaries in $(target)/bin.
3901 2013-09-12 DJ Delorie <dj@redhat.com>
3903 * config/rl78/rl78.opt (mrelax): New.
3904 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
3905 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
3907 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
3909 (rl78_expand_epilogue): Likewise.
3911 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
3914 * lra-eliminations.c (remove_reg_equal_offset_note): New.
3915 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
3916 pointer elimination with using remove_reg_equal_offset_note.
3918 2013-09-12 DJ Delorie <dj@redhat.com>
3920 * config/msp430/: New port.
3921 * config.gcc (msp430): Added.
3922 * doc/invoke.texi: Document MSP430 options.
3923 * doc/install.texi: Document msp430-elf
3924 * doc/md.texi: Document msp430-elf
3925 * doc/contrib.texi: Document msp430-elf
3927 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
3930 2013-09-12 Martin Jambor <mjambor@suse.cz>
3933 * ipa-prop.c (remove_described_reference): Give up if the edge in the
3934 reference descriptor is NULL.
3935 (ipa_edge_removal_hook): If owning a reference descriptor, set its
3938 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3940 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
3941 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
3942 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
3943 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
3944 make_temp_ssa_name): move to tree-ssanames.h
3945 * tree-ssa-alias.h: Move prototype.
3946 * tree-ssa.h: Include tree-ssanames.h.
3947 * tree-ssanames.c (FREE_SSANAMES): Move to here.
3948 * tree-ssanames.h: New. Move items from tree-flow*.h
3949 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
3951 2013-09-12 Richard Biener <rguenther@suse.de>
3953 PR tree-optimization/58404
3954 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
3955 propagate non-invariant addresses into dereferences wrapped
3956 in component references.
3958 2013-09-12 Richard Biener <rguenther@suse.de>
3960 PR tree-optimization/58402
3961 * passes.def: Move pass_late_warn_uninitialized later.
3963 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3965 * tree-ssa.h: New. Move content from tree-flow.h and
3967 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
3968 Move prototypes belonging to tree-ssa.c.
3969 * tree-flow-inline.h (redirect_edge_var_map_def,
3970 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
3972 * gimple.h: Adjust prototypes.
3973 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
3975 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
3976 * tree.h: Move prototype to tree-ssa.h.
3977 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
3978 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
3979 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
3980 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
3981 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
3982 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
3983 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
3984 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
3985 graphite-blocking.c, graphite-clast-to-gimple.c,
3986 graphite-dependences.c, graphite-interchange.c,
3987 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
3988 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
3989 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
3990 ipa-reference.c, ipa-split.c, ipa-utils.c,
3991 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
3992 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
3993 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
3994 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3995 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
3996 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
3997 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
3998 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
3999 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
4000 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
4001 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
4002 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4003 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
4004 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
4005 value-prof.c, var-tracking.c,
4006 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
4008 2013-09-12 Richard Biener <rguenther@suse.de>
4010 PR tree-optimization/58396
4011 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
4012 (build_rdg): Take a loop-nest parameter, fix memleaks.
4013 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
4015 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4017 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
4020 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
4022 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
4023 The FMA instruction names should have a 'v' prefix.
4025 2013-09-12 Richard Biener <rguenther@suse.de>
4027 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
4028 (dot_rdg): Use popen instead of system in optional code.
4029 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
4030 (already_processed_vertex_p): Adjust.
4031 (has_anti_or_output_dependence, predecessor_has_mem_write,
4032 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
4033 rdg_flag_uses): Remove.
4034 (rdg_flag_vertex): Simplify.
4035 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
4037 (build_rdg_partition_for_component): Process the first vertex
4038 of a component only.
4039 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
4041 2013-09-12 Alan Modra <amodra@gmail.com>
4043 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
4045 2013-09-11 DJ Delorie <dj@redhat.com>
4046 Nick Clifton <nickc@redhat.com>
4048 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
4049 (rl78_stack_based_mem): New.
4050 * config/rl78/constraints.md (Iv08): New.
4059 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
4061 (movsi): Change from expand to insn-and-split.
4062 (ashrsi3): Clobber AX.
4066 * config/rl78/rl78.md (CC_REG): Fix.
4067 (addsi3): Allow memory and immediate operands.
4068 (addsi3_internal): Split into...
4069 (addsi3_internal_virt): ...new, and ...
4070 (addsi3_internal_real): ...new.
4072 (subsi3_internal_virt): New.
4073 (subsi3_internal_real): New.
4074 (mulsi3): Add memory operand.
4075 (mulsi3_rl78): Likewise.
4076 (mulsi3_g13): Likewise.
4077 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
4078 (cbranchqi4_real): Add more constraint options.
4079 (cbranchhi4_real): Expand pattern.
4080 (cbranchhi4_real_signed): New.
4081 (cbranchhi4_real_inverted): New.
4082 (cbranchsi4_real_lt): New.
4083 (cbranchsi4_real_ge): New.
4084 (cbranchsi4_real_signed): New.
4085 (cbranchsi4_real): New.
4087 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
4089 (lshrsi3_virt): Likewise.
4090 (ashlsi3_virt): Likewise.
4091 (cbranchqi4_virt_signed): New.
4092 (cbranchhi4_virt_signed): New.
4093 (cbranchsi4_virt): New.
4094 * config/rl78/rl78.c: Whitespace fixes throughout.
4095 (move_elim_pass): New.
4096 (pass_data_rl78_move_elim): New.
4097 (pass_rl78_move_elim): New.
4098 (make_pass_rl78_move_elim): New.
4099 (rl78_devirt_info): Run devirt earlier.
4100 (rl78_move_elim_info): New.
4101 (rl78_asm_file_start): Register it.
4102 (rl78_split_movsi): New.
4103 (rl78_as_legitimate_address): Allow virtual base registers when
4105 (rl78_addr_space_convert): Remove spurious debug stuff.
4106 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
4107 (rl78_print_operand): More cases for not printing '#'.
4108 (rl78_expand_compare): Remove most of the logic.
4109 (content_memory): New.
4110 (clear_content_memory): New.
4111 (get_content_index): New.
4112 (get_content_name): New.
4113 (display_content_memory): New.
4114 (update_content): New.
4115 (record_content): New.
4116 (already_contains): New.
4117 (insn_ok_now): Re-recog insns with virtual registers.
4118 (add_postponed_content_update): New.
4119 (process_postponed_content_update): New.
4120 (gen_and_emit_move): New.
4121 (transcode_memory_rtx): Record new location content.
4122 Use gen_and_emit_move.
4123 (force_into_acc): New.
4124 (move_to_acc): Use gen_and_emit_move.
4125 (move_from_acc): Likewise.
4126 (move_acc_to_reg): Likewise.
4127 (move_to_x): Likewise.
4128 (move_to_hl): Likewise.
4129 (move_to_de): Likewise.
4130 (rl78_alloc_physical_registers_op1): Record location content.
4131 (has_constraint): New.
4132 (rl78_alloc_physical_registers_op2): Record location content.
4134 (rl78_alloc_physical_registers_ro1): Likewise.
4135 (rl78_alloc_physical_registers_cmp): Likewise.
4136 (rl78_alloc_physical_registers_umul): Likewise.
4137 (rl78_alloc_address_registers_macax): New.
4138 (rl78_alloc_physical_registers): Initialize and set location
4139 content memory as needed.
4140 (rl78_reorg): Make sure split2 is called.
4141 (rl78_rtx_costs): New.
4143 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
4145 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
4146 for (not (neg ...)) and (neg (not ...)) cases.
4148 2013-09-11 Richard Biener <rguenther@suse.de>
4151 * passes.def: Split critical edges before late uninit warning passes.
4152 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
4154 2013-09-11 Jakub Jelinek <jakub@redhat.com>
4156 PR tree-optimization/58385
4157 * fold-const.c (build_range_check): If both low and high are NULL,
4158 use omit_one_operand_loc to preserve exp side-effects.
4160 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4162 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
4164 2013-09-11 Richard Biener <rguenther@suse.de>
4166 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
4167 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
4168 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
4169 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
4170 stmts_from_loop, known_dependences_p, build_empty_rdg,
4171 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
4172 * tree-loop-distribution.c: ... here.
4173 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
4174 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
4175 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
4176 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
4177 * tree-loop-distribution.c: ... here.
4178 * tree-loop-distribution.c: Include gimple-pretty-print.h.
4179 (struct partition_s): Add loops member.
4180 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
4181 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
4182 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
4184 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4185 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4186 Sergey Lega <sergey.s.lega@intel.com>
4187 Anna Tikhonova <anna.tikhonova@intel.com>
4188 Ilya Tocar <ilya.tocar@intel.com>
4189 Andrey Turetskiy <andrey.turetskiy@intel.com>
4190 Ilya Verbin <ilya.verbin@intel.com>
4191 Kirill Yukhin <kirill.yukhin@intel.com>
4192 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4194 * config/i386/constraints.md (k): New.
4196 * config/i386/i386.c (const regclass_map): Add new mask registers.
4197 (dbx_register_map): Ditto.
4198 (dbx64_register_map): Ditto.
4199 (svr4_dbx_register_map): Ditto.
4200 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
4202 (ix86_preferred_reload_class): Disable constants for mask registers.
4203 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
4204 (ix86_hard_regno_mode_ok): Support new mask registers.
4205 (x86_order_regs_for_local_alloc): Ditto.
4206 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
4207 (FIXED_REGISTERS): Add new mask registers.
4208 (CALL_USED_REGISTERS): Ditto.
4209 (REG_ALLOC_ORDER): Ditto.
4210 (VALID_MASK_REG_MODE): New.
4211 (FIRST_MASK_REG): Ditto.
4212 (LAST_MASK_REG): Ditto.
4213 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
4214 (MAYBE_MASK_CLASS_P): New.
4215 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
4216 (REG_CLASS_CONTENTS): Ditto.
4217 (MASK_REGNO_P): New.
4218 (ANY_MASK_REG_P): Ditto.
4219 (HI_REGISTER_NAMES): Add new mask registers.
4220 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
4221 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
4223 (attribute "type"): Add mskmov, msklog.
4224 (attribute "length_immediate"): Support them.
4225 (attribute "memory"): Ditto.
4226 (attribute "prefix_0f"): Ditto.
4227 (*movhi_internal): Support new mask registers.
4228 (*movqi_internal): Ditto.
4229 (define_split): Split out clobber pattern is a logic
4230 insn on mask registers.
4231 (*k<logic><mode>): New.
4232 (*andhi_1): Extend to support mask regs.
4233 (*andqi_1): Extend to support mask regs.
4235 (define_split): Split and-not to and and not if operands
4237 (*<code><mode>_1): Separate HI mode to new pattern...
4238 (*<code>hi_1): This.
4239 (*<code>qi_1): Extend to support mask regs.
4241 (kortestzhi): Ditto.
4242 (kortestchi): Ditto.
4244 (*one_cmpl<mode>2_1): Remove HImode and handle it...
4245 (*one_cmplhi2_1): ...Here, now with mask registers support.
4246 (*one_cmplqi2_1): Support new mask registers.
4247 (HI/QImode arithmetics splitter): Don't split if mask registers
4249 (HI/QImode not splitter): Ditto.
4250 * config/i386/predicated.md (mask_reg_operand): New.
4251 (general_reg_operand): Ditto.
4253 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4255 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
4256 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
4258 2013-09-10 Jeff Law <law@redhat.com>
4260 PR tree-optimization/58380
4261 * tree-ssa-threadupdate.c (thread_block): Recognize another case
4262 of threading through a buried loop header.
4264 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
4265 return value for single successor case.
4267 2013-09-10 Jan Hubicka <jh@suse.cz>
4269 * ipa-devirt.c (ipa_devirt): Enable with LTO.
4271 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
4274 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
4275 support conditional execution.
4276 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
4278 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
4280 * lra.c (lra): Clear lra_optional_reload_pseudos before every
4282 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
4283 Check destination too to check move insn.
4284 (undo_optional_reloads): Add check that the original peudo did not
4285 changed its allocation and the optional reload was inherited on last
4286 inheritance pass. Break loop after deciding to keep optional reload.
4287 (lra_undo_inheritance): Add check that inherited pseudo still in
4290 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
4292 * config/aarch64/aarch64.md (generic_sched): New.
4293 * config/aarch64/aarch64-generic.md (load): Make conditional
4294 on generic_sched attribute.
4295 (nonload): Likewise.
4297 2013-09-10 Jan Hubicka <jh@suse.cz>
4299 * lto-cgraph.c: Include ipa-utils.h.
4300 (compute_ltrans_boundary): Also add possible targets into the boundary.
4302 2013-09-10 Jan Hubicka <jh@suse.cz>
4304 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
4305 VAR_DECL of vtable rather than full expression.
4307 2013-09-10 Jan Hubicka <jh@suse.cz>
4308 Paolo Carlini <paolo.carlini@oracle.com>
4310 * cgraphunit.c (analyze_functions): Save input_location, set it
4311 to UNKNOWN_LOCATION and restore it at the end.
4313 2013-09-10 Martin Jambor <mjambor@suse.cz>
4315 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
4316 represented by a thunk.
4318 2013-09-10 Jeff Law <law@redhat.com>
4320 PR tree-optimization/58343
4321 * tree-ssa-threadupdate.c (thread_block): Identify and disable
4322 jump threading requests through loop headers buried in the middle
4323 of a jump threading path.
4325 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
4326 in return value/type.
4328 2013-09-10 Jakub Jelinek <jakub@redhat.com>
4330 PR rtl-optimization/58365
4331 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
4332 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
4335 2013-09-10 Richard Biener <rguenther@suse.de>
4337 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
4338 * tree-data-ref.c (create_rdg_vertices): Collect all data
4339 references, signal failure to the caller, use data-ref API.
4340 (build_rdg): Compute data references only once. Maintain lifetime
4341 of data references and data dependences from within RDG.
4342 (free_rdg): Free dependence relations.
4343 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
4344 inventing extra dependences.
4345 (distribute_loop): Update for RDG API changes.
4347 2013-09-10 Kai Tietz <ktietz@redhat.com>
4349 * doc/invoke.texi (fms-extensions): Document changed
4350 behavior for ms-abi targets.
4351 * config/i386/i386.c (ix86_option_override_internal):
4352 Set default value of option -fms-extension for ms-abi targets.
4354 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4356 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
4358 2013-09-10 Alan Modra <amodra@gmail.com>
4361 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
4363 2013-09-10 Alan Modra <amodra@gmail.com>
4365 * config/rs6000/predicates.md (add_cint_operand): New.
4366 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
4367 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
4368 using add_cint_operand.
4369 (largetoc_high_plus_aix): Likewise.
4371 2013-09-09 Jakub Jelinek <jakub@redhat.com>
4373 PR tree-optimization/58364
4374 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
4375 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
4376 the current range can't be an unconditional true or false.
4378 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
4380 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
4382 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
4384 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
4386 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
4389 * doc/invoke.texi (-Wdelete-incomplete): Document it.
4391 2013-09-09 Ian Bolton <ian.bolton@arm.com>
4393 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
4394 NO_REGS for immediate that can't be moved directly into FP_REGS.
4396 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4398 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
4399 comparison with negated operand.
4400 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
4403 2013-09-09 Richard Biener <rguenther@suse.de>
4406 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
4407 of a subloop record all its block as affecting loop-closed SSA form.
4409 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4411 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
4412 of an rtx/bitpos pair.
4413 (store_fixed_bit_field): Update accordingly.
4415 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4417 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
4419 * builtins.c (expand_errno_check): Likewise.
4420 * dwarf2cfi.c (init_return_column_size): Likewise.
4421 * except.c (sjlj_mark_call_sites): Likewise.
4422 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
4423 * lra-constraints.c (emit_inc): Likewise.
4424 * ree.c (combine_set_extension): Likewise.
4425 * regmove.c (fixup_match_2): Likewise.
4426 * reload1.c (inc_for_reload): Likewise.
4428 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4430 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
4431 (force_to_mode, simplify_shift_const_1, simplify_comparison):
4432 Use gen_int_mode with the mode of the associated simplify_* call.
4433 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
4434 * expmed.c (expand_shift_1): Likewise.
4435 * function.c (instantiate_virtual_regs_in_insn): Likewise.
4436 * loop-iv.c (iv_number_of_iterations): Likewise.
4437 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
4438 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
4440 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4442 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
4443 of the associated expand_* call.
4444 (asan_emit_stack_protection): Likewise.
4445 * builtins.c (round_trampoline_addr): Likewise.
4446 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
4447 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
4448 (emit_store_flag): Likewise.
4449 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
4450 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
4452 * function.c (instantiate_virtual_regs_in_insn): Likewise.
4453 * ifcvt.c (noce_try_store_flag_constants): Likewise.
4454 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
4455 * modulo-sched.c (generate_prolog_epilog): Likewise.
4456 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
4457 (expand_ctz, expand_ffs, expand_unop): Likewise.
4459 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4461 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
4462 of the associated gen_rtx_* call.
4463 * caller-save.c (init_caller_save): Likewise.
4464 * combine.c (find_split_point, make_extraction): Likewise.
4465 (make_compound_operation): Likewise.
4466 * dwarf2out.c (mem_loc_descriptor): Likewise.
4467 * explow.c (plus_constant, probe_stack_range): Likewise.
4468 * expmed.c (expand_mult_const): Likewise.
4469 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
4470 * reload1.c (init_reload): Likewise.
4471 * valtrack.c (cleanup_auto_inc_dec): Likewise.
4472 * var-tracking.c (adjust_mems): Likewise.
4473 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
4474 rather than gen_rtx_fmt_ee.
4476 2013-09-09 Jan Hubicka <jh@suse.cz>
4479 * value-prof.c (gimple_ic): Copy also abnormal edges.
4481 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4483 * asan.c (asan_shadow_cst): Use gen_int_mode.
4485 2013-09-08 Jan Hubicka <jh@suse.cz>
4487 * ipa-profile.c: Add toplevel comment.
4488 (ipa_propagate_frequency_1): Be more conservative when profile is read.
4489 (contains_hot_call_p): New function.
4490 (ipa_propagate_frequency): Set frequencies based on counts when
4492 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
4493 profile; do not tamper with profile after inlining if it is read.
4495 2013-09-08 Jan Hubicka <jh@suse.cz>
4497 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
4500 2013-09-08 Jan Hubicka <jh@suse.cz>
4502 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
4505 2013-09-08 Jeff Law <law@redhat.com>
4508 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
4511 2013-09-08 Andi Kleen <ak@linux.intel.com>
4513 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
4515 2013-09-08 Jan Hubicka <jh@suse.cz>
4517 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
4519 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
4521 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
4522 for non-debug insns.
4523 * lra.c (new_insn_reg): Take the containing insn as a parameter.
4524 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
4525 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
4528 2013-09-08 Jan Hubicka <jh@suse.cz>
4530 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
4531 targets and devirtualize to BUILT_IN_UNREACHABLE.
4532 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
4533 * ipa.c (walk_polymorphic_call_targets): New function.
4534 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
4535 functions; use the new timevar.
4536 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
4537 was removed from the program.
4538 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
4539 not consider it in the walk when its vtable is dead.
4540 (possible_polymorphic_call_targets_1): Pass anonymous flag to
4542 (devirt_variable_node_removal_hook): New function.
4543 (possible_polymorphic_call_targets): Also register
4544 devirt_variable_node_removal_hook.
4545 (ipa_devirt): Do not do non-speculative devirtualization.
4546 (gate_ipa_devirt): One execute if devirtualizing speculatively.
4548 2013-09-08 Jan Hubicka <jh@suse.cz>
4550 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
4551 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
4552 varpool_remove_variable_insertion_hook): Declare.
4553 * varpool.c (varpool_node_hook_list): New structure.
4554 (first_varpool_node_removal_hook,
4555 first_varpool_variable_insertion_hook): New variables.
4556 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
4557 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
4558 varpool_remove_variable_insertion_hook,
4559 varpool_call_variable_insertion_hooks): New functions.
4560 (varpool_remove_node): Use it.
4562 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
4565 * diagnostic.c (diagnostic_build_prefix): When s.file is
4566 "<built-in>" don't output line and column numbers.
4568 2013-09-06 Jan Hubicka <jh@suse.cz>
4570 * cgraphunit.c (expand_thunk): Get body before touching arguments.
4571 * lto-streamer-out.c: Stream thunks, too.
4572 * lto-streamer-in.c (input_function): Pop cfun here
4573 (lto_read_body): Instead of here.
4575 2013-09-06 Caroline Tice <cmtice@google.com>
4577 * doc/install.texi: Add documentation for the --enable-vtable-verify
4578 and the --disable-libvtv configure options.
4580 2013-09-06 Jeff Law <law@redhat.com>
4582 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4583 edge implied equivalences into successor phis.
4585 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
4587 * resource.c (mark_referenced_resources): Handle COND_EXEC.
4589 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
4591 * resource.c (mark_target_live_regs): Compute resources taking
4592 into account if a call is predicated or not.
4594 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
4596 * toplev.c (output_stack_usage): Be prepared for suffixes created by
4597 the compiler in the function names.
4599 2013-09-06 Jan Hubicka <jh@suse.cz>
4602 * ipa-inline.c (has_caller_p): New function.
4603 (want_inline_function_to_all_callers_p): Use it.
4604 (sum_callers, inline_to_all_callers): Break out from ...
4605 (ipa_inline): ... here.
4607 2013-09-06 Jan Hubicka <jh@suse.cz>
4609 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
4610 only when AVX is enabled.
4612 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4614 * config/aarch64/aarch64.md
4615 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
4616 is fpsimd_<load/store>2.
4617 (load_pair<mode>): Likewise.
4618 (store_pair<mode>): Likewise.
4620 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4622 * config/arm/types.md (type): Add "mrs" type.
4623 * config/aarch64/aarch64.md
4624 (aarch64_load_tp_hard): Make type "mrs".
4626 (load_tp_hard): Make type "mrs".
4627 * config/arm/cortex-a15.md: Update with new attributes.
4628 * config/arm/cortex-a5.md: Update with new attributes.
4629 * config/arm/cortex-a53.md: Update with new attributes.
4630 * config/arm/cortex-a7.md: Update with new attributes.
4631 * config/arm/cortex-a8.md: Update with new attributes.
4632 * config/arm/cortex-a9.md: Update with new attributes.
4633 * config/arm/cortex-m4.md: Update with new attributes.
4634 * config/arm/cortex-r4.md: Update with new attributes.
4635 * config/arm/fa526.md: Update with new attributes.
4636 * config/arm/fa606te.md: Update with new attributes.
4637 * config/arm/fa626te.md: Update with new attributes.
4638 * config/arm/fa726te.md: Update with new attributes.
4640 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4642 * config/aarch64/aarch64.md
4643 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
4644 (*movtf_aarch64): Likewise.
4646 (thumb1_movdi_insn): Use "multiple" for type where more than one
4647 instruction is used for a move.
4648 (*arm32_movhf): Likewise.
4649 (*thumb_movdf_insn): Likewise.
4651 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4653 * config/arm/types.md (type): Rename fcpys to fmov.
4655 (*arm_movsi_vfp): Rename type fcpys as fmov.
4656 (*thumb2_movsi_vfp): Likewise
4657 (*movhf_vfp_neon): Likewise
4658 (*movhf_vfp): Likewise
4659 (*movsf_vfp): Likewise
4660 (*thumb2_movsf_vfp): Likewise
4661 (*movsfcc_vfp): Likewise
4662 (*thumb2_movsfcc_vfp): Likewise
4663 * config/aarch64/aarch64-simd.md
4664 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
4665 * config/aarch64/aarch64.md
4666 (*movsi_aarch64): Replace type mov_reg with fmovs.
4667 (*movdi_aarch64): Likewise
4668 (*movsf_aarch64): Likewise
4669 (*movdf_aarch64): Likewise
4671 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
4672 * config/arm/iwmmxt.md
4673 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
4674 * config/arm/arm1020e.md: Update with new attributes.
4675 * config/arm/cortex-a15-neon.md: Update with new attributes.
4676 * config/arm/cortex-a5.md: Update with new attributes.
4677 * config/arm/cortex-a53.md: Update with new attributes.
4678 * config/arm/cortex-a7.md: Update with new attributes.
4679 * config/arm/cortex-a8-neon.md: Update with new attributes.
4680 * config/arm/cortex-a9.md: Update with new attributes.
4681 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4682 * config/arm/cortex-r4f.md: Update with new attributes.
4683 * config/arm/marvell-pj4.md: Update with new attributes.
4684 * config/arm/vfp11.md: Update with new attributes.
4686 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4688 * config/aarch64/aarch64.md
4689 (*madd<mode>): Fix type attribute.
4690 (*maddsi_uxtw): Likewise.
4691 (*msub<mode>): Likewise.
4692 (*msubsi_uxtw): Likewise.
4693 (<su_optab>maddsidi4): Likewise.
4694 (<su_optab>msubsidi4): Likewise.
4696 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4698 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
4699 * config/arm/arm.md (core_cycles): Remove fdiv.
4700 * config/arm/vfp.md:
4701 (*sqrtsf2_vfp): Update for attribute changes.
4702 (*sqrtdf2_vfp): Likewise.
4703 * config/aarch64/aarch64.md:
4704 (sqrt<mode>2): Update for attribute changes.
4705 * config/arm/arm1020e.md: Update with new attributes.
4706 * config/arm/cortex-a15-neon.md: Update with new attributes.
4707 * config/arm/cortex-a5.md: Update with new attributes.
4708 * config/arm/cortex-a53.md: Update with new attributes.
4709 * config/arm/cortex-a7.md: Update with new attributes.
4710 * config/arm/cortex-a8-neon.md: Update with new attributes.
4711 * config/arm/cortex-a9.md: Update with new attributes.
4712 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4713 * config/arm/cortex-r4f.md: Update with new attributes.
4714 * config/arm/marvell-pj4.md: Update with new attributes.
4715 * config/arm/vfp11.md: Update with new attributes.
4717 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4719 * config/arm/types.md
4720 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
4721 * config/aarch64/aarch64.md
4722 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
4724 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
4725 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
4726 (float<GPI:mode><GPF:mode>2): Likewise.
4728 (*truncsisf2_vfp): Update with new attributes.
4729 (*truncsidf2_vfp): Likewise.
4730 (fixuns_truncsfsi2): Likewise.
4731 (fixuns_truncdfsi2): Likewise.
4732 (*floatsisf2_vfp): Likewise.
4733 (*floatsidf2_vfp): Likewise.
4734 (floatunssisf2): Likewise.
4735 (floatunssidf2): Likewise.
4736 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
4737 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
4738 * config/arm/arm1020e.md: Update with new attributes.
4739 * config/arm/cortex-a15-neon.md: Update with new attributes.
4740 * config/arm/cortex-a5.md: Update with new attributes.
4741 * config/arm/cortex-a53.md: Update with new attributes.
4742 * config/arm/cortex-a7.md: Update with new attributes.
4743 * config/arm/cortex-a8-neon.md: Update with new attributes.
4744 * config/arm/cortex-a9.md: Update with new attributes.
4745 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4746 * config/arm/cortex-r4f.md: Update with new attributes.
4747 * config/arm/marvell-pj4.md: Update with new attributes.
4748 * config/arm/vfp11.md: Update with new attributes.
4750 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4752 * config/aarch64/arm_neon.h
4753 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
4754 (vqtbx<1,2,3,4><q>_s8): Likewise.
4756 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4758 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
4759 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
4760 (add<mode>3): Add type attribute.
4761 (add<mode>3): Likewise.
4762 (usadd<mode>3): Likewise.
4763 (ssadd<mode>3): Likewise.
4764 (sub<mode>3): Likewise.
4765 (sub<mode>3): Likewise.
4766 (ussub<mode>3): Likewise.
4767 (sssub<mode>3): Likewise.
4768 (ssmulsa3): Likewise.
4769 (usmulusa3): Likewise.
4770 (arm_usatsihi): Likewise.
4772 (*movdi_vfp): Add types for all instructions.
4773 (*movdi_vfp_cortexa8): Likewise.
4774 (*movhf_vfp_neon): Likewise.
4775 (*movhf_vfp): Likewise.
4776 (*movdf_vfp): Likewise.
4777 (*thumb2_movdf_vfp): Likewise.
4778 (*thumb2_movdfcc_vfp): Likewise.
4779 * config/arm/arm.md: Add type attribute to all insn patterns.
4780 (*thumb1_adddi3): Add type attribute.
4781 (*arm_adddi3): Likewise.
4782 (*adddi_sesidi_di): Likewise.
4783 (*adddi_zesidi_di): Likewise.
4784 (*thumb1_addsi3): Likewise.
4785 (addsi3_compare0): Likewise.
4786 (*addsi3_compare0_scratch): Likewise.
4787 (*compare_negsi_si): Likewise.
4788 (cmpsi2_addneg): Likewise.
4789 (*addsi3_carryin_<optab>): Likewise.
4790 (*addsi3_carryin_alt2_<optab>): Likewise.
4791 (*addsi3_carryin_clobercc_<optab>): Likewise.
4792 (*subsi3_carryin): Likewise.
4793 (*subsi3_carryin_const): Likewise.
4794 (*subsi3_carryin_compare): Likewise.
4795 (*subsi3_carryin_compare_const): Likewise.
4796 (*arm_subdi3): Likewise.
4797 (*thumb_subdi3): Likewise.
4798 (*subdi_di_zesidi): Likewise.
4799 (*subdi_di_sesidi): Likewise.
4800 (*subdi_zesidi_di): Likewise.
4801 (*subdi_sesidi_di): Likewise.
4802 (*subdi_zesidi_ze): Likewise.
4803 (thumb1_subsi3_insn): Likewise.
4804 (*arm_subsi3_insn): Likewise.
4805 (*anddi3_insn): Likewise.
4806 (*anddi_zesidi_di): Likewise.
4807 (*anddi_sesdi_di): Likewise.
4808 (*ne_zeroextracts): Likewise.
4809 (*ne_zeroextracts): Likewise.
4810 (*ite_ne_zeroextr): Likewise.
4811 (*ite_ne_zeroextr): Likewise.
4812 (*anddi_notdi_di): Likewise.
4813 (*anddi_notzesidi): Likewise.
4814 (*anddi_notsesidi): Likewise.
4815 (andsi_notsi_si): Likewise.
4816 (thumb1_bicsi3): Likewise.
4817 (*iordi3_insn): Likewise.
4818 (*iordi_zesidi_di): Likewise.
4819 (*iordi_sesidi_di): Likewise.
4820 (*thumb1_iorsi3_insn): Likewise.
4821 (*xordi3_insn): Likewise.
4822 (*xordi_zesidi_di): Likewise.
4823 (*xordi_sesidi_di): Likewise.
4824 (*arm_xorsi3): Likewise.
4825 (*andsi_iorsi3_no): Likewise.
4826 (*smax_0): Likewise.
4827 (*smax_m1): Likewise.
4828 (*arm_smax_insn): Likewise.
4829 (*smin_0): Likewise.
4830 (*arm_smin_insn): Likewise.
4831 (*arm_umaxsi3): Likewise.
4832 (*arm_uminsi3): Likewise.
4833 (*minmax_arithsi): Likewise.
4834 (*minmax_arithsi_): Likewise.
4835 (*satsi_<SAT:code>): Likewise.
4836 (arm_ashldi3_1bit): Likewise.
4837 (arm_ashrdi3_1bit): Likewise.
4838 (arm_lshrdi3_1bit): Likewise.
4839 (*arm_negdi2): Likewise.
4840 (*thumb1_negdi2): Likewise.
4841 (*arm_negsi2): Likewise.
4842 (*thumb1_negsi2): Likewise.
4843 (*negdi_extendsid): Likewise.
4844 (*negdi_zero_extend): Likewise.
4845 (*arm_abssi2): Likewise.
4846 (*thumb1_abssi2): Likewise.
4847 (*arm_neg_abssi2): Likewise.
4848 (*thumb1_neg_abss): Likewise.
4849 (one_cmpldi2): Likewise.
4850 (extend<mode>di2): Likewise.
4851 (*compareqi_eq0): Likewise.
4852 (*arm_extendhisi2addsi): Likewise.
4853 (*arm_movdi): Likewise.
4854 (*thumb1_movdi_insn): Likewise.
4855 (*arm_movt): Likewise.
4856 (*thumb1_movsi_insn): Likewise.
4857 (pic_add_dot_plus_four): Likewise.
4858 (pic_add_dot_plus_eight): Likewise.
4859 (tls_load_dot_plus_eight): Likewise.
4860 (*thumb1_movhi_insn): Likewise.
4861 (*thumb1_movsf_insn): Likewise.
4862 (*movdf_soft_insn): Likewise.
4863 (*thumb_movdf_insn): Likewise.
4864 (cbranchsi4_insn): Likewise.
4865 (cbranchsi4_scratch): Likewise.
4866 (*negated_cbranchsi4): Likewise.
4867 (*tbit_cbranch): Likewise.
4868 (*tlobits_cbranch): Likewise.
4869 (*tstsi3_cbranch): Likewise.
4870 (*cbranchne_decr1): Likewise.
4871 (*addsi3_cbranch): Likewise.
4872 (*addsi3_cbranch_scratch): Likewise.
4873 (*arm_cmpdi_insn): Likewise.
4874 (*arm_cmpdi_unsig): Likewise.
4875 (*arm_cmpdi_zero): Likewise.
4876 (*thumb_cmpdi_zero): Likewise.
4877 (*deleted_compare): Likewise.
4878 (*mov_scc): Likewise.
4879 (*mov_negscc): Likewise.
4880 (*mov_notscc): Likewise.
4881 (*cstoresi_eq0_thumb1_insn): Likewise.
4882 (cstoresi_nltu_thumb1): Likewise.
4883 (cstoresi_ltu_thu): Likewise.
4884 (thumb1_addsi3_addgeu): Likewise.
4885 (*arm_jump): Likewise.
4886 (*thumb_jump): Likewise.
4887 (*check_arch2): Likewise.
4888 (arm_casesi_internal): Likewise.
4889 (thumb1_casesi_dispatch): Likewise.
4890 (*arm_indirect_jump): Likewise.
4891 (*thumb1_indirect_jump): Likewise.
4893 (*and_scc): Likewise.
4894 (*ior_scc): Likewise.
4895 (*compare_scc): Likewise.
4896 (*cond_move): Likewise.
4897 (*cond_arith): Likewise.
4898 (*cond_sub): Likewise.
4899 (*cmp_ite0): Likewise.
4900 (*cmp_ite1): Likewise.
4901 (*cmp_and): Likewise.
4902 (*cmp_ior): Likewise.
4903 (*ior_scc_scc): Likewise.
4904 (*ior_scc_scc_cmp): Likewise.
4905 (*and_scc_scc): Likewise.
4906 (*and_scc_scc_cmp): Likewise.
4907 (*and_scc_scc_nod): Likewise.
4908 (*negscc): Likewise.
4909 (movcond_addsi): Likewise.
4910 (movcond): Likewise.
4911 (*ifcompare_plus_move): Likewise.
4912 (*if_plus_move): Likewise.
4913 (*ifcompare_move_plus): Likewise.
4914 (*if_move_plus): Likewise.
4915 (*ifcompare_arith_arith): Likewise.
4916 (*if_arith_arith): Likewise.
4917 (*ifcompare_arith_move): Likewise.
4918 (*if_arith_move): Likewise.
4919 (*ifcompare_move_arith): Likewise.
4920 (*if_move_arith): Likewise.
4921 (*ifcompare_move_not): Likewise.
4922 (*if_move_not): Likewise.
4923 (*ifcompare_not_move): Likewise.
4924 (*if_not_move): Likewise.
4925 (*ifcompare_shift_move): Likewise.
4926 (*if_shift_move): Likewise.
4927 (*ifcompare_move_shift): Likewise.
4928 (*if_move_shift): Likewise.
4929 (*ifcompare_shift_shift): Likewise.
4930 (*ifcompare_not_arith): Likewise.
4931 (*ifcompare_arith_not): Likewise.
4932 (*if_arith_not): Likewise.
4933 (*ifcompare_neg_move): Likewise.
4934 (*if_neg_move): Likewise.
4935 (*ifcompare_move_neg): Likewise.
4936 (*if_move_neg): Likewise.
4937 (prologue_thumb1_interwork): Likewise.
4938 (*cond_move_not): Likewise.
4939 (*sign_extract_onebit): Likewise.
4940 (*not_signextract_onebit): Likewise.
4941 (stack_tie): Likewise.
4942 (align_4): Likewise.
4943 (align_8): Likewise.
4944 (consttable_end): Likewise.
4945 (consttable_1): Likewise.
4946 (consttable_2): Likewise.
4947 (consttable_4): Likewise.
4948 (consttable_8): Likewise.
4949 (consttable_16): Likewise.
4950 (*thumb1_tablejump): Likewise.
4951 (prefetch): Likewise.
4952 (force_register_use): Likewise.
4953 (thumb_eh_return): Likewise.
4954 (load_tp_hard): Likewise.
4955 (load_tp_soft): Likewise.
4956 (tlscall): Likewise.
4957 (*arm_movtas_ze): Likewise.
4958 (*arm_rev): Likewise.
4959 (*arm_revsh): Likewise.
4960 (*arm_rev16): Likewise.
4961 * config/arm/thumb2.md
4962 (*thumb2_smaxsi3): Likewise.
4963 (*thumb2_sminsi3): Likewise.
4964 (*thumb32_umaxsi3): Likewise.
4965 (*thumb2_uminsi3): Likewise.
4966 (*thumb2_negdi2): Likewise.
4967 (*thumb2_abssi2): Likewise.
4968 (*thumb2_neg_abss): Likewise.
4969 (*thumb2_movsi_insn): Likewise.
4970 (tls_load_dot_plus_four): Likewise.
4971 (*thumb2_movhi_insn): Likewise.
4972 (*thumb2_mov_scc): Likewise.
4973 (*thumb2_mov_negs): Likewise.
4974 (*thumb2_mov_negs): Likewise.
4975 (*thumb2_mov_nots): Likewise.
4976 (*thumb2_mov_nots): Likewise.
4977 (*thumb2_movsicc_): Likewise.
4978 (*thumb2_movsfcc_soft_insn): Likewise.
4979 (*thumb2_indirect_jump): Likewise.
4980 (*thumb2_and_scc): Likewise.
4981 (*thumb2_ior_scc): Likewise.
4982 (*thumb2_ior_scc_strict_it): Likewise.
4983 (*thumb2_cond_move): Likewise.
4984 (*thumb2_cond_arith): Likewise.
4985 (*thumb2_cond_ari): Likewise.
4986 (*thumb2_cond_sub): Likewise.
4987 (*thumb2_negscc): Likewise.
4988 (*thumb2_movcond): Likewise.
4989 (thumb2_casesi_internal): Likewise.
4990 (thumb2_casesi_internal_pic): Likewise.
4991 (*thumb2_alusi3_short): Likewise.
4992 (*thumb2_mov<mode>_shortim): Likewise.
4993 (*thumb2_addsi_short): Likewise.
4994 (*thumb2_subsi_short): Likewise.
4995 (thumb2_addsi3_compare0): Likewise.
4996 (*thumb2_cbz): Likewise.
4997 (*thumb2_cbnz): Likewise.
4998 (*thumb2_one_cmplsi2_short): Likewise.
4999 (*thumb2_negsi2_short): Likewise.
5000 (*orsi_notsi_si): Likewise.
5001 * config/arm/arm1020e.md: Update with new attributes.
5002 * config/arm/arm1026ejs.md: Update with new attributes.
5003 * config/arm/arm1136jfs.md: Update with new attributes.
5004 * config/arm/arm926ejs.md: Update with new attributes.
5005 * config/arm/cortex-a15.md: Update with new attributes.
5006 * config/arm/cortex-a5.md: Update with new attributes.
5007 * config/arm/cortex-a53.md: Update with new attributes.
5008 * config/arm/cortex-a7.md: Update with new attributes.
5009 * config/arm/cortex-a8.md: Update with new attributes.
5010 * config/arm/cortex-a9.md: Update with new attributes.
5011 * config/arm/cortex-m4.md: Update with new attributes.
5012 * config/arm/cortex-r4.md: Update with new attributes.
5013 * config/arm/fa526.md: Update with new attributes.
5014 * config/arm/fa606te.md: Update with new attributes.
5015 * config/arm/fa626te.md: Update with new attributes.
5016 * config/arm/fa726te.md: Update with new attributes.
5018 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5020 * config/aarch64/aarch64-simd.md
5021 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
5022 <vwx> iterator to ensure correct register choice.
5023 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
5024 (aarch64_sqdmull_n<mode>): Likewise.
5025 (aarch64_sqdmull2_n<mode>_internal): Likewise.
5026 * config/aarch64/arm_neon.h
5027 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
5028 (vml<as><q>_n_<su>16): Likewise.
5029 (vml<as>l_high_lane<q>_<su>16): Likewise.
5030 (vml<as>l_high_n_<su>16): Likewise.
5031 (vml<as>l_lane<q>_<su>16): Likewise.
5032 (vml<as>l_n_<su>16): Likewise.
5033 (vmul<q>_lane<q>_<su>16): Likewise.
5034 (vmul<q>_n_<su>16): Likewise.
5035 (vmull_lane<q>_<su>16): Likewise.
5036 (vmull_n_<su>16): Likewise.
5037 (vmull_high_lane<q>_<su>16): Likewise.
5038 (vmull_high_n_<su>16): Likewise.
5039 (vqrdmulh<q>_n_s16): Likewise.
5041 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
5043 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
5044 have the correct lane parameter.
5046 2013-09-06 Richard Biener <rguenther@suse.de>
5048 * cfganal.c (control_dependences::~control_dependences):
5049 Properly free all of the vector.
5051 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
5054 * config/i386/i386.c (ix86_conditional_register_usage):
5055 Proper initialize extended SSE registers.
5057 2013-09-06 Jan Hubicka <jh@suse.cz>
5059 PR tree-optimization/58311
5060 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
5062 2013-09-06 Jan Hubicka <jh@suse.cz>
5064 * Makefile.in (tree-sra.o): Update dependencies.
5065 * tree-sra.c: Include ipa-utils.h
5066 (scan_function): Use recursive_call_p.
5067 (has_caller_p): New function.
5068 (cgraph_for_node_and_aliases): Count also callers of aliases.
5070 2013-09-06 Jan Hubicka <jh@suse.cz>
5073 * cgraph.h (symtab_semantically_equivalent_p): Declare.
5074 * tree-tailcall.c: Include ipa-utils.h.
5075 (find_tail_calls): Use it.
5076 * ipa-pure-const.c (check_call): Likewise.
5077 * ipa-utils.c (recursive_call_p): New function.
5078 * ipa-utils.h (recursive_call_p): Dclare.
5079 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
5080 (symtab_semantically_equivalent_p): New function.
5081 * Makefile.in (tree-tailcall.o): Update dependencies.
5083 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5085 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
5088 2013-09-06 Richard Biener <rguenther@suse.de>
5090 * lto-streamer.h (lto_global_var_decls): Remove.
5091 * Makefile.in (OBJS): Remove lto-symtab.o.
5092 (lto-symtab.o): Remove.
5093 (GTFILES): Remove lto-symtab.c
5094 * lto-symtab.c: Move to lto/
5096 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5098 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
5099 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
5100 (UNSPEC_FPINT_RINT): New constant definitions.
5101 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
5102 definition with 2 attributes.
5103 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
5104 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
5107 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5109 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
5110 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
5111 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
5112 attribute to "z196_alone".
5113 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
5116 2013-09-06 Richard Biener <rguenther@suse.de>
5118 * basic-block.h (class control_dependences): New.
5119 * tree-ssa-dce.c (control_dependence_map): Remove.
5121 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
5122 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
5123 find_pdom, find_control_dependence, find_all_control_dependences):
5125 (mark_control_dependent_edges_necessary,
5126 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
5127 tree_dce_done, perform_tree_ssa_dce): Adjust.
5128 * cfganal.c (set_control_dependence_map_bit,
5129 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
5130 find_all_control_dependences): Move from tree-ssa-dce.c and
5131 implement as methods of control_dependences class.
5132 (control_dependences::control_dependences): New.
5133 (control_dependences::~control_dependences): Likewise.
5134 (control_dependences::get_edges_dependent_on): Likewise.
5135 (control_dependences::get_edge): Likewise.
5137 2013-09-04 Jan Hubicka <jh@suse.cz>
5139 * tree.c (types_same_for_odr): Drop overactive check.
5140 * ipa-devirt.c (hash_type_name): Likewise.
5142 2013-09-04 Jan Hubicka <jh@suse.cz>
5144 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
5145 (analyze_functions): ... here.
5147 2013-09-04 Jan Hubicka <jh@suse.cz>
5150 * cgraphunit.c (analyze_functions): Clear AUX fields
5151 after processing; initialize assembler name has.
5153 2013-09-05 Jeff Law <law@redhat.com>
5155 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
5156 from thread_around_empty_block. Record threading path into PATH.
5157 Recurse if threading through the initial block is successful.
5158 (thread_across_edge): Corresponding changes to slightly simplify.
5160 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
5162 * config/aarch64/aarch64.md
5163 (type): Remove frecpe, frecps, frecpx.
5164 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
5165 fix to be a TARGET_SIMD instruction.
5166 (aarch64_frecps): Remove.
5167 * config/aarch64/aarch64-simd.md
5168 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
5169 (aarch64_frecps<mode>): Handle all float/vector of float modes.
5171 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
5172 Sofiane Naci <sofiane.naci@arm.com>
5174 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
5175 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
5176 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
5177 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
5178 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
5179 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
5180 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
5181 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
5182 into "clz, "rbit". Rename "shift" to "shift_imm".
5183 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
5184 changes. Update for attribute changes all occurrences of arlo_* and
5186 * config/arm/arm-fixed.md: Update for attribute changes
5187 all occurrences of arlo_* types.
5188 * config/arm/thumb2.md: Update for attribute changes all occurrences
5190 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
5191 (cortexa7_older_only): Likewise.
5192 (cortexa7_younger): Likewise.
5193 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5194 (1020alu_shift_op): Likewise.
5195 (1020alu_shift_reg_op): Likewise.
5196 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5197 (alu_shift_op): Likewise.
5198 (alu_shift_reg_op): Likewise.
5199 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5200 (11_alu_shift_op): Likewise.
5201 (11_alu_shift_reg_op): Likewise.
5202 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5203 (9_alu_shift_reg_op): Likewise.
5204 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
5206 (cortex_a15_alu_shift): Likewise.
5207 (cortex_a15_alu_shift_reg): Likewise.
5208 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
5210 (cortex_a5_alu_shift): Likewise.
5211 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
5213 (cortex_a53_alu_shift): Likewise.
5214 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
5216 (cortex_a7_alu_reg): Likewise.
5217 (cortex_a7_alu_shift): Likewise.
5218 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
5220 (cortex_a8_alu_shift): Likewise.
5221 (cortex_a8_alu_shift_reg): Likewise.
5222 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5223 (cortex_a9_dp_shift): Likewise.
5224 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
5226 * config/arm/cortex-r4.md
5227 (cortex_r4_alu): Update for attribute changes.
5228 (cortex_r4_mov): Likewise.
5229 (cortex_r4_alu_shift_reg): Likewise.
5230 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5231 (526_alu_shift_op): Likewise.
5232 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5233 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5234 (626te_alu_shift_op): Likewise.
5235 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
5236 (726te_alu_shift_op): Likewise.
5237 (726te_alu_shift_reg_op): Likewise.
5238 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5239 (mp626_alu_shift_op): Likewise.
5240 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
5241 (pj4_alu_conds): Likewise.
5242 (pj4_shift): Likewise.
5243 (pj4_shift_conds): Likewise.
5244 (pj4_alu_shift): Likewise.
5245 (pj4_alu_shift_conds): Likewise.
5246 * config/aarch64/aarch64.md: Update for attribute change
5247 all occurrences of arlo_* and shift* types.
5249 2013-09-05 Mike Stump <mikestump@comcast.net>
5251 * tree.h: Move documentation for tree_function_decl to tree-core.h
5252 with the declaration.
5254 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
5257 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
5258 looking for widest mode.
5260 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
5262 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
5264 2013-09-05 Richard Biener <rguenther@suse.de>
5266 PR tree-optimization/58137
5267 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
5268 Do not create vectors of pointers.
5269 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
5270 types for the components of the vector initializer.
5271 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
5272 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
5274 2013-09-05 Martin Jambor <mjambor@suse.cz>
5276 * ipa-prop.c (remove_described_reference): Accept missing references,
5277 return false if that hppens, otherwise return true.
5278 (cgraph_node_for_jfunc): New function.
5279 (try_decrement_rdesc_refcount): Likewise.
5280 (try_make_edge_direct_simple_call): Use them.
5281 (ipa_edge_removal_hook): Remove references from rdescs.
5282 (ipa_edge_duplication_hook): Clone rdescs and their references
5283 when the new edge has the same caller as the old one.
5284 * cgraph.c (cgraph_resolve_speculation): Remove speculative
5285 reference before removing any edges.
5287 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
5289 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
5291 * thumb2.md (thumb2_storewb_parisi): New pattern.
5293 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
5295 * config/aarch64/aarch64-option-extensions.def: Add
5296 AARCH64_OPT_EXTENSION of 'crc'.
5297 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
5298 (AARCH64_ISA_CRC): Ditto.
5299 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
5300 description of the CRC extension.
5302 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
5304 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
5305 * config/rs6000/linux.h: Ditto.
5306 * alpha/linux.h: Ditto.
5307 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
5308 no_c99_libc_has_function.
5309 * config/c6x/uclinux-elf.h: Ditto.
5310 * config/lm32/uclinux-elf.h: Ditto.
5311 * config/m68k/uclinux.h: Ditto.
5312 * config/moxie/uclinux.h: Ditto.
5313 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
5314 (crisv32-*-linux*, cris-*-linux*): Ditto.
5315 * config/bfin/bfin.c: Include "tm_p.h".
5317 2013-09-05 Richard Biener <rguenther@suse.de>
5319 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
5320 check for a definition without a basic-block.
5322 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
5323 Sofiane Naci <sofiane.naci@arm.com>
5325 * config/aarch64/aarch64.md
5326 (*movti_aarch64): Rename r_2_f and f_2_r.
5327 (*movsf_aarch64): Likewise.
5328 (*movdf_aarch64): Likewise.
5329 (*movtf_aarch64): Likewise.
5330 (aarch64_movdi_<mode>low): Likewise.
5331 (aarch64_movdi_<mode>high): Likewise.
5332 (aarch64_mov<mode>high_di): Likewise.
5333 (aarch64_mov<mode>low_di): Likewise.
5334 (aarch64_movtilow_tilow): Likewise.
5335 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
5336 values to config/arm/types.md
5337 (attribute "conds"): Update for attribute change.
5338 (anddi3_insn): Likewise.
5339 (iordi3_insn): Likewise.
5340 (xordi3_insn): Likewise.
5341 (one_cmpldi2): Likewise.
5342 * config/arm/types.md (type): Add Neon types.
5343 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
5344 use "type" attribute.
5345 (movmisalign<mode>_neon_store): Likewise.
5346 (movmisalign<mode>_neon_load): Likewise.
5347 (vec_set<mode>_internal): Likewise.
5348 (vec_setv2di_internal): Likewise.
5349 (vec_extract<mode>): Likewise.
5350 (vec_extractv2di): Likewise.
5351 (add<mode>3_neon): Likewise.
5352 (adddi3_neon): Likewise.
5353 (sub<mode>3_neon): Likewise.
5354 (subdi3_neon): Likewise.
5355 (mul<mode>3_neon): Likewise.
5356 (mul<mode>3add<mode>_neon): Likewise.
5357 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
5358 (fma<VCVTF:mode>4)): Likewise.
5359 (fma<VCVTF:mode>4_intrinsic): Likewise.
5360 (fmsub<VCVTF:mode>4)): Likewise.
5361 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
5362 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
5363 (ior<mode>3): Likewise.
5364 (and<mode>3): Likewise.
5365 (anddi3_neon): Likewise.
5366 (orn<mode>3_neon): Likewise.
5367 (orndi3_neon): Likewise.
5368 (bic<mode>3_neon): Likewise.
5369 (bicdi3_neon): Likewise.
5370 (xor<mode>3): Likewise.
5371 (one_cmpl<mode>2): Likewise.
5372 (abs<mode>2): Likewise.
5373 (neg<mode>2): Likewise.
5374 (umin<mode>3_neon): Likewise.
5375 (umax<mode>3_neon): Likewise.
5376 (smin<mode>3_neon): Likewise.
5377 (smax<mode>3_neon): Likewise.
5378 (vashl<mode>3): Likewise.
5379 (vashr<mode>3_imm): Likewise.
5380 (vlshr<mode>3_imm): Likewise.
5381 (ashl<mode>3_signed): Likewise.
5382 (ashl<mode>3_unsigned): Likewise.
5383 (neon_load_count): Likewise.
5384 (ashldi3_neon_noclobber): Likewise.
5385 (signed_shift_di3_neon): Likewise.
5386 (unsigned_shift_di3_neon): Likewise.
5387 (ashrdi3_neon_imm_noclobber): Likewise.
5388 (lshrdi3_neon_imm_noclobber): Likewise.
5389 (widen_ssum<mode>3): Likewise.
5390 (widen_usum<mode>3): Likewise.
5391 (quad_halves_<code>v4si): Likewise.
5392 (quad_halves_<code>v4sf): Likewise.
5393 (quad_halves_<code>v8hi): Likewise.
5394 (quad_halves_<code>v16qi): Likewise.
5395 (reduc_splus_v2di): Likewise.
5396 (neon_vpadd_internal<mode>): Likewise.
5397 (neon_vpsmin<mode>): Likewise.
5398 (neon_vpsmax<mode>): Likewise.
5399 (neon_vpumin<mode>): Likewise.
5400 (neon_vpumax<mode>): Likewise.
5401 (ss_add<mode>_neon): Likewise.
5402 (us_add<mode>_neon): Likewise.
5403 (ss_sub<mode>_neon): Likewise.
5404 (us_sub<mode>_neon): Likewise.
5405 (neon_vadd<mode>_unspec): Likewise.
5406 (neon_vaddl<mode>): Likewise.
5407 (neon_vaddw<mode>): Likewise.
5408 (neon_vhadd<mode>): Likewise.
5409 (neon_vqadd<mode>): Likewise.
5410 (neon_vaddhn<mode>): Likewise.
5411 (neon_vmul<mode>): Likewise.
5412 (neon_vmla<mode>): Likewise.
5413 (neon_vmlal<mode>): Likewise.
5414 (neon_vmls<mode>): Likewise.
5415 (neon_vmlsl<mode>): Likewise.
5416 (neon_vqdmulh<mode>): Likewise.
5417 (neon_vqdmlal<mode>): Likewise.
5418 (neon_vqdmlsl<mode>): Likewise.
5419 (neon_vmull<mode>): Likewise.
5420 (neon_vqdmull<mode>): Likewise.
5421 (neon_vsub<mode>_unspec): Likewise.
5422 (neon_vsubl<mode>): Likewise.
5423 (neon_vsubw<mode>): Likewise.
5424 (neon_vqsub<mode>): Likewise.
5425 (neon_vhsub<mode>): Likewise.
5426 (neon_vsubhn<mode>): Likewise.
5427 (neon_vceq<mode>): Likewise.
5428 (neon_vcge<mode>): Likewise.
5429 (neon_vcgeu<mode>): Likewise.
5430 (neon_vcgt<mode>): Likewise.
5431 (neon_vcgtu<mode>): Likewise.
5432 (neon_vcle<mode>): Likewise.
5433 (neon_vclt<mode>): Likewise.
5434 (neon_vcage<mode>): Likewise.
5435 (neon_vcagt<mode>): Likewise.
5436 (neon_vtst<mode>): Likewise.
5437 (neon_vabd<mode>): Likewise.
5438 (neon_vabdl<mode>): Likewise.
5439 (neon_vaba<mode>): Likewise.
5440 (neon_vabal<mode>): Likewise.
5441 (neon_vmax<mode>): Likewise.
5442 (neon_vmin<mode>): Likewise.
5443 (neon_vpaddl<mode>): Likewise.
5444 (neon_vpadal<mode>): Likewise.
5445 (neon_vpmax<mode>): Likewise.
5446 (neon_vpmin<mode>): Likewise.
5447 (neon_vrecps<mode>): Likewise.
5448 (neon_vrsqrts<mode>): Likewise.
5449 (neon_vqabs<mode>): Likewise.
5450 (neon_vqneg<mode>): Likewise.
5451 (neon_vcls<mode>): Likewise.
5452 (clz<mode>2): Likewise.
5453 (popcount<mode>2): Likewise.
5454 (neon_vrecpe): Likewise.
5455 (neon_vrsqrte): Likewise.
5456 (neon_vget_lane<mode>_sext_internal): Likewise.
5457 (neon_vget_lane<mode>_zext_internal): Likewise.
5458 (neon_vdup_n<mode>): Likewise.
5459 (neon_vdup_nv2di): Likewise.
5460 (neon_vdpu_lane<mode>_internal): Likewise.
5461 (neon_vswp<mode>): Likewise.
5462 (float<mode><V_cvtto>2): Likewise.
5463 (floatuns<mode><V_cvtto>2): Likewise.
5464 (fix_trunc<mode><V_cvtto>)2): Likewise
5465 (fixuns_trunc<mode><V_cvtto)2): Likewise.
5466 (neon_vcvt<mode>): Likewise.
5467 (neon_vcvtv4sfv4hf): Likewise.
5468 (neon_vcvtv4hfv4sf): Likewise.
5469 (neon_vcvt_n<mode>): Likewise.
5470 (neon_vmovn<mode>): Likewise.
5471 (neon_vqmovn<mode>): Likewise.
5472 (neon_vqmovun<mode>): Likewise.
5473 (neon_vmovl<mode>): Likewise.
5474 (neon_vmul_lane<mode>): Likewise.
5475 (neon_vmull_lane<mode>): Likewise.
5476 (neon_vqdmull_lane<mode>): Likewise.
5477 (neon_vqdmulh_lane<mode>): Likewise.
5478 (neon_vmla_lane<mode>): Likewise.
5479 (neon_vmlal_lane<mode>): Likewise.
5480 (neon_vqdmlal_lane<mode>): Likewise.
5481 (neon_vmls_lane<mode>): Likewise.
5482 (neon_vmlsl_lane<mode>): Likewise.
5483 (neon_vqdmlsl_lane<mode>): Likewise.
5484 (neon_vext<mode>): Likewise.
5485 (neon_vrev64<mode>): Likewise.
5486 (neon_vrev32<mode>): Likewise.
5487 (neon_vrev16<mode>): Likewise.
5488 (neon_vbsl<mode>_internal): Likewise.
5489 (neon_vshl<mode>): Likewise.
5490 (neon_vqshl<mode>): Likewise.
5491 (neon_vshr_n<mode>): Likewise.
5492 (neon_vshrn_n<mode>): Likewise.
5493 (neon_vqshrn_n<mode>): Likewise.
5494 (neon_vqshrun_n<mode>): Likewise.
5495 (neon_vshl_n<mode>): Likewise.
5496 (neon_vqshl_n<mode>): Likewise.
5497 (neon_vqshlu_n<mode>): Likewise.
5498 (neon_vshll_n<mode>): Likewise.
5499 (neon_vsra_n<mode>): Likewise.
5500 (neon_vsri_n<mode>): Likewise.
5501 (neon_vsli_n<mode>): Likewise.
5502 (neon_vtbl1v8qi): Likewise.
5503 (neon_vtbl2v8qi): Likewise.
5504 (neon_vtbl3v8qi): Likewise.
5505 (neon_vtbl4v8qi): Likewise.
5506 (neon_vtbx1v8qi): Likewise.
5507 (neon_vtbx2v8qi): Likewise.
5508 (neon_vtbx3v8qi): Likewise.
5509 (neon_vtbx4v8qi): Likewise.
5510 (neon_vtrn<mode>_internal): Likewise.
5511 (neon_vzip<mode>_internal): Likewise.
5512 (neon_vuzp<mode>_internal): Likewise.
5513 (neon_vld1<mode>): Likewise.
5514 (neon_vld1_lane<mode>): Likewise.
5515 (neon_vld1_dup<mode>): Likewise.
5516 (neon_vld1_dupv2di): Likewise.
5517 (neon_vst1<mode>): Likewise.
5518 (neon_vst1_lane<mode>): Likewise.
5519 (neon_vld2<mode>): Likewise.
5520 (neon_vld2_lane<mode>): Likewise.
5521 (neon_vld2_dup<mode>): Likewise.
5522 (neon_vst2<mode>): Likewise.
5523 (neon_vst2_lane<mode>): Likewise.
5524 (neon_vld3<mode>): Likewise.
5525 (neon_vld3qa<mode>): Likewise.
5526 (neon_vld3qb<mode>): Likewise.
5527 (neon_vld3_lane<mode>): Likewise.
5528 (neon_vld3_dup<mode>): Likewise.
5529 (neon_vst3<mode>): Likewise.
5530 (neon_vst3qa<mode>): Likewise.
5531 (neon_vst3qb<mode>): Likewise.
5532 (neon_vst3_lane<mode>): Likewise.
5533 (neon_vld4<mode>): Likewise.
5534 (neon_vld4qa<mode>): Likewise.
5535 (neon_vld4qb<mode>): Likewise.
5536 (neon_vld4_lane<mode>): Likewise.
5537 (neon_vld4_dup<mode>): Likewise.
5538 (neon_vst4<mode>): Likewise.
5539 (neon_vst4qa<mode>): Likewise.
5540 (neon_vst4qb<mode>): Likewise.
5541 (neon_vst4_lane<mode>): Likewise.
5542 (neon_vec_unpack<US>_lo_<mode>): Likewise.
5543 (neon_vec_unpack<US>_hi_<mode>): Likewise.
5544 (neon_vec_<US>mult_lo_<mode>): Likewise.
5545 (neon_vec_<US>mult_hi_<mode>): Likewise.
5546 (neon_vec_<US>shiftl_<mode>): Likewise.
5547 (neon_unpack<US>_<mode>): Likewise.
5548 (neon_vec_<US>mult_<mode>): Likewise.
5549 (vec_pack_trunc_<mode>): Likewise.
5550 (neon_vec_pack_trunk_<mode>): Likewise.
5551 (neon_vabd<mode>_2): Likewise.
5552 (neon_vabd<mode>_3): Likewise.
5553 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5554 (thumb2_movsi_vfp): Likewise.
5555 (movdi_vfp): Likewise.
5556 (movdi_vfp_cortexa8): Likewise.
5557 (movhf_vfp_neon): Likewise.
5558 (movhf_vfp): Likewiwse.
5559 (movsf_vfp): Likewiwse.
5560 (thumb2_movsf_vfp): Likewiwse.
5561 (movdf_vfp): Likewise.
5562 (thumb2_movdf_vfp): Likewise.
5563 (movsfcc_vfp): Likewise.
5564 (thumb2_movsfcc_vfp): Likewise.
5565 (movdfcc_vfp): Likewise.
5566 (thumb2_movdfcc_vfp): Likewise.
5567 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
5568 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
5569 (v10_v2c): Likewise.
5570 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
5572 (cortex_a15_neon_int_2): Likewise.
5573 (cortex_a15_neon_int_3): Likewise.
5574 (cortex_a15_neon_int_4): Likewise.
5575 (cortex_a15_neon_int_5): Likewise.
5576 (cortex_a15_neon_vqneg_vqabs): Likewise.
5577 (cortex_a15_neon_vmov): Likewise.
5578 (cortex_a15_neon_vaba): Likewise.
5579 (cortex_a15_neon_vaba_qqq): Likewise.
5580 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5581 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5582 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
5583 scalar_64_32_long_scalar): Likewise.
5584 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5585 (cortex_a15_neon_mla_qqq_8_16): Likewise.
5586 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5587 lotype_qdd_64_32_long): Likewise.
5588 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5589 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5590 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
5591 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5592 (cortex_a15_neon_shift_1): Likewise.
5593 (cortex_a15_neon_shift_2): Likewise.
5594 (cortex_a15_neon_shift_3): Likewise.
5595 (cortex_a15_neon_vshl_ddd): Likewise.
5596 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5597 (cortex_a15_neon_vsra_vrsra): Likewise.
5598 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
5599 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
5600 (cortex_a15_neon_fp_vmul_ddd): Likewise.
5601 (cortex_a15_neon_fp_vmul_qqd): Likewise.
5602 (cortex_a15_neon_fp_vmla_ddd): Likewise.
5603 (cortex_a15_neon_fp_vmla_qqq): Likewise.
5604 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5605 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5606 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5607 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5608 (cortex_a15_neon_bp_simple): Likewise.
5609 (cortex_a15_neon_bp_2cycle): Likewise.
5610 (cortex_a15_neon_bp_3cycle): Likewise.
5611 (cortex_a15_neon_vld1_1_2_regs): Likewise.
5612 (cortex_a15_neon_vld1_3_4_regs): Likewise.
5613 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5614 (cortex_a15_neon_vld2_4_regs): Likewise.
5615 (cortex_a15_neon_vld3_vld4): Likewise.
5616 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5617 (cortex_a15_neon_vst1_3_4_regs): Likewise.
5618 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
5619 (cortex_a15_neon_vst3_vst4): Likewise.
5620 (cortex_a15_neon_vld1_vld2_lane): Likewise.
5621 (cortex_a15_neon_vld3_vld4_lane" 10
5622 (cortex_a15_neon_vst1_vst2_lane): Likewise.
5623 (cortex_a15_neon_vst3_vst4_lane): Likewise.
5624 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
5625 (cortex_a15_neon_ldm_2): Likewise.0
5626 (cortex_a15_neon_stm_2): Likewise.
5627 (cortex_a15_neon_mcr): Likewise.
5628 (cortex_a15_neon_mcr_2_mcrr): Likewise.
5629 (cortex_a15_neon_mrc): Likewise.
5630 (cortex_a15_neon_mrrc): Likewise.
5631 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5633 (cortex_a15_alu_shift): Likewise.
5634 (cortex_a15_alu_shift_reg): Likewise.
5635 (cortex_a15_mult32): Likewise.
5636 (cortex_a15_mult64): Likewise.
5637 (cortex_a15_block): Likewise.
5638 (cortex_a15_branch): Likewise.
5639 (cortex_a15_load1): Likewise.
5640 (cortex_a15_load3): Likewise.
5641 (cortex_a15_store1): Likewise.
5642 (cortex_a15_store3): Likewise.
5643 (cortex_a15_call): Likewise.
5644 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
5645 (cortex_a5_f2r): Likewise.
5646 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
5648 (cortex_a53_f2r): Likewise.
5649 * config/arm/cortex-a7.md
5650 (cortex_a7_branch): Update for attribute change.
5651 (cortex_a7_call): Likewise.
5652 (cortex_a7_alu_imm): Likewise.
5653 (cortex_a7_alu_reg): Likewise.
5654 (cortex_a7_alu_shift): Likewise.
5655 (cortex_a7_mul): Likewise.
5656 (cortex_a7_load1): Likewise.
5657 (cortex_a7_store1): Likewise.
5658 (cortex_a7_load2): Likewise.
5659 (cortex_a7_store2): Likewise.
5660 (cortex_a7_load3): Likewise.
5661 (cortex_a7_store3): Likewise.
5662 (cortex_a7_load4): Likewise.
5663 (cortex_a7_store4): Likewise.
5664 (cortex_a7_fpalu): Likewise.
5665 (cortex_a7_fconst): Likewise.
5666 (cortex_a7_fpmuls): Likewise.
5667 (cortex_a7_neon_mul): Likewise.
5668 (cortex_a7_fpmacs): Likewise.
5669 (cortex_a7_neon_mla: Likewise.
5670 (cortex_a7_fpmuld: Likewise.
5671 (cortex_a7_fpmacd: Likewise.
5672 (cortex_a7_fpfmad: Likewise.
5673 (cortex_a7_fdivs: Likewise.
5674 (cortex_a7_fdivd: Likewise.
5675 (cortex_a7_r2f: Likewise.
5676 (cortex_a7_f2r: Likewise.
5677 (cortex_a7_f_flags: Likewise.
5678 (cortex_a7_f_loads: Likewise.
5679 (cortex_a7_f_loadd: Likewise.
5680 (cortex_a7_f_stores: Likewise.
5681 (cortex_a7_f_stored: Likewise.
5682 (cortex_a7_neon): Likewise.
5683 * config/arm/cortex-a8-neon.md
5684 (cortex_a8_neon_mrc): Update for attribute change.
5685 (cortex_a8_neon_mrrc): Likewise.
5686 (cortex_a8_neon_int_1): Likewise.
5687 (cortex_a8_neon_int_2): Likewise.
5688 (cortex_a8_neon_int_3): Likewise.
5689 (cortex_a8_neon_int_4): Likewise.
5690 (cortex_a8_neon_int_5): Likewise.
5691 (cortex_a8_neon_vqneg_vqabs): Likewise.
5692 (cortex_a8_neon_vmov): Likewise.
5693 (cortex_a8_neon_vaba): Likewise.
5694 (cortex_a8_neon_vaba_qqq): Likewise.
5695 (cortex_a8_neon_vsma): Likewise.
5696 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5697 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5698 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
5699 long_scalar): Likewise.
5700 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5701 (cortex_a8_neon_mla_qqq_8_16): Likewise.
5702 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5703 long_scalar_qdd_64_32_long): Likewise.
5704 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5705 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5706 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
5707 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5708 (cortex_a8_neon_shift_1): Likewise.
5709 (cortex_a8_neon_shift_2): Likewise.
5710 (cortex_a8_neon_shift_3): Likewise.
5711 (cortex_a8_neon_vshl_ddd): Likewise.
5712 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5713 (cortex_a8_neon_vsra_vrsra): Likewise.
5714 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
5715 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
5716 (cortex_a8_neon_fp_vsum): Likewise.
5717 (cortex_a8_neon_fp_vmul_ddd): Likewise.
5718 (cortex_a8_neon_fp_vmul_qqd): Likewise.
5719 (cortex_a8_neon_fp_vmla_ddd): Likewise.
5720 (cortex_a8_neon_fp_vmla_qqq): Likewise.
5721 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
5722 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
5723 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5724 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5725 (cortex_a8_neon_bp_simple): Likewise.
5726 (cortex_a8_neon_bp_2cycle): Likewise.
5727 (cortex_a8_neon_bp_3cycle): Likewise.
5728 (cortex_a8_neon_ldr): Likewise.
5729 (cortex_a8_neon_str): Likewise.
5730 (cortex_a8_neon_vld1_1_2_regs): Likewise.
5731 (cortex_a8_neon_vld1_3_4_regs): Likewise.
5732 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5733 (cortex_a8_neon_vld2_4_regs): Likewise.
5734 (cortex_a8_neon_vld3_vld4): Likewise.
5735 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5736 (cortex_a8_neon_vst1_3_4_regs): Likewise.
5737 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
5738 (cortex_a8_neon_vst3_vst4): Likewise.
5739 (cortex_a8_neon_vld1_vld2_lane): Likewise.
5740 (cortex_a8_neon_vld3_vld4_lane): Likewise.
5741 (cortex_a8_neon_vst1_vst2_lane): Likewise.
5742 (cortex_a8_neon_vst3_vst4_lane): Likewise.
5743 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
5744 (cortex_a8_neon_mcr): Likewise.
5745 (cortex_a8_neon_mcr_2_mcrr): Likewise.
5746 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5747 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
5749 (ca9_neon_mrrc): Likewise.
5750 (cortex_a9_neon_int_1): Likewise.
5751 (cortex_a9_neon_int_2): Likewise.
5752 (cortex_a9_neon_int_3): Likewise.
5753 (cortex_a9_neon_int_4): Likewise.
5754 (cortex_a9_neon_int_5): Likewise.
5755 (cortex_a9_neon_vqneg_vqabs): Likewise.
5756 (cortex_a9_neon_vmov): Likewise.
5757 (cortex_a9_neon_vaba): Likewise.
5758 (cortex_a9_neon_vaba_qqq): Likewise.
5759 (cortex_a9_neon_vsma): Likewise.
5760 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5761 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5762 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
5763 long_scalar): Likewise.
5764 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5765 (cortex_a9_neon_mla_qqq_8_16): Likewise.
5766 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5767 long_scalar_qdd_64_32_long): Likewise.
5768 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5769 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5770 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
5771 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5772 (cortex_a9_neon_shift_1): Likewise.
5773 (cortex_a9_neon_shift_2): Likewise.
5774 (cortex_a9_neon_shift_3): Likewise.
5775 (cortex_a9_neon_vshl_ddd): Likewise.
5776 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5777 (cortex_a9_neon_vsra_vrsra): Likewise.
5778 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
5779 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
5780 (cortex_a9_neon_fp_vsum): Likewise.
5781 (cortex_a9_neon_fp_vmul_ddd): Likewise.
5782 (cortex_a9_neon_fp_vmul_qqd): Likewise.
5783 (cortex_a9_neon_fp_vmla_ddd): Likewise.
5784 (cortex_a9_neon_fp_vmla_qqq): Likewise.
5785 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
5786 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
5787 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5788 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5789 (cortex_a9_neon_bp_simple): Likewise.
5790 (cortex_a9_neon_bp_2cycle): Likewise.
5791 (cortex_a9_neon_bp_3cycle): Likewise.
5792 (cortex_a9_neon_ldr): Likewise.
5793 (cortex_a9_neon_str): Likewise.
5794 (cortex_a9_neon_vld1_1_2_regs): Likewise.
5795 (cortex_a9_neon_vld1_3_4_regs): Likewise.
5796 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5797 (cortex_a9_neon_vld2_4_regs): Likewise.
5798 (cortex_a9_neon_vld3_vld4): Likewise.
5799 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5800 (cortex_a9_neon_vst1_3_4_regs): Likewise.
5801 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
5802 (cortex_a9_neon_vst3_vst4): Likewise.
5803 (cortex_a9_neon_vld1_vld2_lane): Likewise.
5804 (cortex_a9_neon_vld3_vld4_lane): Likewise.
5805 (cortex_a9_neon_vst1_vst2_lane): Likewise.
5806 (cortex_a9_neon_vst3_vst4_lane): Likewise.
5807 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5808 (cortex_a9_neon_mcr): Likewise.
5809 (cortex_a9_neon_mcr_2_mcrr): Likewise.
5810 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
5811 (cortex_a9_fps): Likewise.
5812 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
5814 (cortex_m4_fmuls): Likewise.
5815 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
5817 (cortex_r4_mrc): Likewise.
5818 * config/arm/iterators.md: Update comment referring to neon_type.
5819 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
5820 (iwmmxt_movsi_insn): Likewise.
5821 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
5823 (pj4_core_to_vfp): Likewise.
5824 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
5826 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
5827 (vfp_fstore): Likewise.
5828 * doc/md.texi: Change references to neon_type to refer to type.
5830 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5832 * tree.h (DECL_BUILT_IN): Fix typo in comment.
5834 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
5836 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
5839 2013-09-04 Easwaran Raman <eraman@google.com>
5842 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
5843 (build_and_add_sum): Use it.
5844 (appears_later_in_bb): Simplify code.
5846 2013-09-04 Teresa Johnson <tejohnson@google.com>
5848 * dumpfile.c (dump_finish): Don't close stderr/stdout.
5850 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
5852 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
5854 2013-09-04 Jan Hubicka <jh@suse.cz>
5856 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
5857 * ipa-devirt.c: Include diganostic.h
5858 (odr_type_d): Add types and types_set.
5859 (hash_type_name): Work for types with vtables during LTO.
5860 (odr_hasher::remove): Fix comment; destroy types_set.
5861 (add_type_duplicate): New function,
5862 (get_odr_type): Use it.
5863 (dump_type_inheritance_graph): Dump type duplicates.
5864 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
5866 * tree.c (types_same_for_odr): Give exact answers on types with
5869 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5871 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
5872 explaining their differences.
5874 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
5876 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
5878 2013-09-03 Jeff Law <law@redhat.com>
5880 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
5881 when not threading through a joiner block. Pass joiner/no joiner
5882 state to register_jump_thread.
5883 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
5884 state from argument rather than implying on path length.
5885 Dump the entire jump thread path into debugging dump.
5886 * tree-flow.h (register_jump_thread): Update prototype.
5888 2013-08-29 Xinliang David Li <davidxl@google.com>
5890 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5891 Remove a trivial gcc_assert.
5893 2013-08-29 Xinliang David Li <davidxl@google.com>
5895 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
5896 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
5897 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5898 Delay base decl alignment adjustment.
5899 * tree-vectorizer.c (vect_destroy_datarefs): New function.
5900 * tree-vectorizer.h: New data structure.
5901 (set_dr_misalignment): New function.
5902 (dr_misalignment): Ditto.
5903 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
5904 (vectorizable_load): Ditto.
5905 (ensure_base_align): New function.
5906 (vectorize_loops): Add dbg_cnt support.
5907 (execute_vect_slp): Ditto.
5908 * dbgcnt.def: New debug counter.
5909 * Makefile: New dependency.
5911 2013-09-03 Meador Inge <meadori@codesourcery.com>
5915 2013-08-30 Meador Inge <meadori@codesourcery.com>
5917 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
5919 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
5921 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
5922 function descriptor.
5924 2013-09-03 Richard Biener <rguenther@suse.de>
5926 * tree-affine.c (add_elt_to_tree): Fix association issue,
5927 avoid useless converts and make sure to always return a
5928 properly typed result.
5930 2013-09-03 Richard Biener <rguenther@suse.de>
5933 * fold-const.c (negate_expr_p): Fix division case.
5934 (negate_expr): Likewise.
5936 2013-09-03 Richard Biener <rguenther@suse.de>
5939 * tree-streamer-out.c: Include tm.h.
5940 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
5942 2013-09-03 Jan Hubicka <jh@suse.cz>
5944 * tree-profile.c (tree_profiling): Cleanup CFG when done.
5946 2013-09-03 Alan Modra <amodra@gmail.com>
5948 * config.gcc (powerpc*-*-linux*): Add support for little-endian
5949 multilibs to big-endian target and vice versa.
5950 * config/rs6000/t-linux64: Use := assignment on all vars.
5951 (MULTILIB_EXTRA_OPTS): Remove fPIC.
5952 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
5953 * config/rs6000/t-linux64le: New file.
5954 * config/rs6000/t-linux64bele: New file.
5955 * config/rs6000/t-linux64lebe: New file.
5957 2013-09-02 Jan Hubicka <jh@suse.cz>
5959 * ipa-inline-transform.c (inline_transform): Do not
5960 optimize_inline_calls when not optimizing.
5962 2013-09-02 Jan Hubicka <jh@suse.cz>
5964 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
5965 duplicated nodes for assembler names.
5966 * symtab.c (symtab_unregister_node): Do not attempt to unlink
5967 hard registers from assembler name hash.
5969 2013-09-02 Jan Hubicka <jh@suse.cz>
5971 * ipa-split.c (execute_split_functions): Split externally visible
5972 functions called once.
5974 2013-09-02 Martin Jambor <mjambor@suse.cz>
5977 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
5978 linked list. When finding the correct duplicate, also consider also
5979 the caller in additon to its inlined_to node.
5981 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
5983 * config/aarch64/aarch64-simd-builtins.def
5984 (dup_lane_scalar): Remove.
5985 * config/aarch64/aarch64-simd.md
5986 (aarch64_simd_dup): Add 'w->w' alternative.
5987 (aarch64_dup_lane<mode>): Allow for VALL.
5988 (aarch64_dup_lane_scalar<mode>): Remove.
5989 (aarch64_dup_lane_<vswap_width_name><mode>): New.
5990 (aarch64_get_lane_signed<mode>): Add w->w altenative.
5991 (aarch64_get_lane_unsigned<mode>): Likewise.
5992 (aarch64_get_lane<mode>): Likewise.
5993 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
5994 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
5995 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
5996 (VCON): Change container of V2SF.
5997 (vswap_width_name): Likewise.
5998 * config/aarch64/arm_neon.h
5999 (__aarch64_vdup_lane_any): New.
6000 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6001 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
6002 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6004 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
6007 * expr.c (emit_move_complex): Do not move complex FP values as parts if
6008 the source or the destination is a single hard register.
6010 2013-09-02 Richard Biener <rguenther@suse.de>
6013 * tree-scalar-evolution.c (instantiate_scev_name): Allow
6016 2013-09-02 Richard Biener <rguenther@suse.de>
6018 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
6019 arithmetic to sizetype.
6021 2013-09-02 Bin Cheng <bin.cheng@arm.com>
6023 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
6024 Find auto-increment use both before and after candidate.
6026 2013-09-02 Marek Polacek <polacek@redhat.com>
6028 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
6030 2013-09-01 Jan Hubicka <jh@suse.cz>
6032 * Makefile.in: Add ipa-profile.o
6033 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
6034 * cgraph.c (struct cgraph_propagate_frequency_data,
6035 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
6036 ipa-profile.c; replace cgraph_ by ipa_ prefix.
6037 * cgraph.h (cgraph_propagate_frequency): Remove.
6038 * ipa-inline-analysis.c: Include ipa-utils.h;
6039 drop duplicated cfgloop.h.
6040 (inline_update_callee_summaries): Update.
6041 * ipa-profile.c: New file.
6042 * ipa-utils.h (ipa_propagate_frequency): Declare.
6043 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
6044 data-streamer.h, value-prof.h.
6045 (symtab_remove_unreachable_nodes): Update profile.
6046 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
6047 account_time_size, cmp_counts, dump_histogram,
6048 ipa_profile_generate_summary, ipa_profile_write_summary,
6049 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
6050 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
6051 Move to ipa-profile.c.
6053 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
6055 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
6057 2013-09-01 Jan Hubicka <jh@suse.cz>
6059 * common.opt (fdevirtualize-speculatively): New function.
6060 * invoke.texi (fdevirtualize-speculatively): Document.
6061 * ipa-devirt.c: Include ipa-inline.h
6062 (likely_target_p): New function.
6063 (ipa_devirt): New function.
6064 (gate_ipa_devirt): New function.
6065 (pass_data_ipa_devirt): New static var.
6066 (pass_ipa_devirt): Likewise.
6067 (make_pass_ipa_devirt): New function.
6068 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
6069 (common_handle_option): Disable devirtualization when
6070 value range profiling is available.
6071 * passes.def (pass_ipa_devirt): Add.
6072 * timever.def (TV_IPA_DEVIRT): New timevar.
6073 * tree-pass.h (make_pass_ipa_devirt):
6075 2013-09-01 Iain Sandoe <iain@codesourcery.com>
6077 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
6078 include sanitize(undefined).
6080 2013-08-31 Diego Novillo <dnovillo@google.com>
6082 * Makefile.in (TREE_CORE_H): Define.
6084 (GTFILES): Add tree-core.h.
6085 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
6087 * tree-core.h: New file.
6088 Move all data structures, enum, typedefs, global
6089 declarations and constants from ...
6092 2013-08-31 Jan Hubicka <jh@suse.cz>
6094 * bulitins.c (expand_builtin): Do not early exit for gcov
6095 instrumented functions.
6097 2013-08-31 Marek Polacek <polacek@redhat.com>
6099 * ubsan.c: Include tm_p.h.
6101 2013-08-31 Jan Hubicka <jh@suse.cz>
6103 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
6106 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
6107 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
6108 * tree-cfg.c (verify_gimple_label): ... here.
6109 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
6110 (ipa_merge_profiles): New function.
6111 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
6112 (lto_input_function_body): Likewise.
6113 * ipa-utils.h (ipa_merge_profiles): Declare.
6114 * lto-streamer.h (lto_input_function_body): Update prototype.
6115 (emit_label_in_global_context_p): Remove.
6116 * lto-symtab.c: Include ipa-utils.h
6117 (lto_cgraph_replace_node): Use ipa_merge_profiles.
6119 2013-08-31 Jan Hubicka <jh@suse.cz>
6121 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
6123 2013-08-31 Jan Hubicka <jh@suse.cz>
6125 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
6127 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
6129 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
6130 "cmp" RTX before signed_comparison_operator check to account
6133 2013-08-30 Jan Hubicka <jh@suse.cz>
6135 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
6136 (detect_type_change_1): Rename to ...
6137 (detect_type_change): ... this one; early return on non-polymorphic
6139 (detect_type_change_ssa): Add comp_type parameter; update
6140 use of detect_type_change.
6141 (compute_complex_assign_jump_func): Add param_type parameter;
6142 update use of detect_type_change_ssa.
6143 (compute_complex_ancestor_jump_func): Likewise.
6144 (ipa_get_callee_param_type): New function.
6145 (ipa_compute_jump_functions_for_edge): Compute parameter type;
6146 update calls to the jump function computation functions.
6148 2013-08-30 Teresa Johnson <tejohnson@google.com>
6149 Steven Bosscher <steven@gcc.gnu.org>
6151 * cfgrtl.c (fixup_new_cold_bb): New routine.
6152 (commit_edge_insertions): Invoke fixup_partitions.
6153 (find_partition_fixes): New routine.
6154 (fixup_partitions): Ditto.
6155 (verify_hot_cold_block_grouping): Update comments.
6156 (rtl_verify_edges): Invoke find_partition_fixes.
6157 (rtl_verify_bb_pointers): Update comments.
6158 (rtl_verify_bb_layout): Ditto.
6159 * basic-block.h (probably_never_executed_edge_p): Declare.
6160 (fixup_partitions): Ditto.
6161 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
6162 * bb-reorder.c (sanitize_hot_paths): New function.
6163 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
6165 * predict.c (probably_never_executed_edge_p): New routine.
6166 * cfg.c (check_bb_profile): Add partition insanity warnings.
6168 2013-08-30 Meador Inge <meadori@codesourcery.com>
6170 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
6172 2013-08-30 Marek Polacek <polacek@redhat.com>
6174 * Makefile.in (ubsan.o): Add.
6175 (c-family/c-ubsan.o): Add.
6176 (builtins.o): Add ubsan.h dependency.
6177 * ubsan.h: New file.
6178 * ubsan.c: New file.
6179 * common.opt: Add -fsanitize=undefined option.
6180 (flag_sanitize): Add variable.
6181 (fsanitize=): Add option. Add Driver.
6182 (fsanitize=thread): Remove option.
6183 (fsanitize=address): Likewise.
6184 (static-libubsan): New option.
6185 * doc/invoke.texi: Document the new flag and -static-libubsan.
6186 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
6187 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
6188 * builtin-attrs.def (ATTR_COLD): Define.
6189 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
6190 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
6191 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
6192 * flag-types.h (sanitize_code): New enum.
6193 * opts.c (common_handle_option): Parse command line arguments
6194 of -fsanitize=. Add -fsanitize=unreachable option.
6195 * varasm.c (get_variable_section): Adjust.
6196 (assemble_noswitch_variable): Likewise.
6197 (assemble_variable): Likewise.
6198 (output_constant_def_contents): Likewise.
6199 (categorize_decl_for_section): Likewise.
6200 (place_block_symbol): Likewise.
6201 (output_object_block): Likewise.
6202 * builtins.def: Likewise.
6203 * toplev.c (compile_file): Likewise.
6204 (process_options): Likewise.
6205 * cppbuiltin.c: Likewise.
6206 * tsan.c (tsan_pass): Likewise.
6207 (tsan_gate): Likewise.
6208 (tsan_gate_O0): Likewise.
6209 * cfgexpand.c (partition_stack_vars): Likewise.
6210 (expand_stack_vars): Likewise.
6211 (defer_stack_allocation): Likewise.
6212 (expand_used_vars): Likewise.
6213 * cfgcleanup.c (old_insns_match_p): Likewise.
6214 * asan.c (asan_finish_file): Likewise.
6215 (asan_instrument): Likewise.
6216 (gate_asan): Likewise.
6217 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
6218 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
6219 (asan_global_struct): Use pointer_sized_int_node instead
6220 calling build_nonstandard_integer_type.
6221 (initialize_sanitizer_builtins): Likewise.
6222 (asan_finish_file): Likewise.
6223 * gcc.c: Document %{%:function(args):X}.
6224 (static_spec_functions): Add sanitize.
6225 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
6226 store funcval != NULL there.
6227 (do_spec_1): Adjust handle_spec_function caller.
6228 (handle_braces): Allow %:function(args) as condition.
6229 (sanitize_spec_function): New function.
6230 (ADD_STATIC_LIBUBSAN_LIBS): Define.
6231 (LIBUBSAN_SPEC): Likewise.
6232 (LIBUBSAN_EARLY_SPEC): Likewise.
6233 (SANITIZER_SPEC): Handle libubsan.
6234 (SANITIZER_EARLY_SPEC): Likewise.
6235 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
6236 instead of fsanitize=address.
6237 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
6238 instead of fsanitize=address*.
6239 * builtins.c: Include ubsan.h.
6240 (fold_builtin_0): Instrument __builtin_unreachable.
6241 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
6242 instead of flag_asan.
6243 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
6244 (pointer_sized_int_node): Define.
6245 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
6247 2013-08-30 Mike Stump <mikestump@comcast.net>
6249 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
6252 2013-08-29 Jan Hubicka <jh@suse.cz>
6254 * cgraph.c (cgraph_function_body_availability): Handle weakref
6256 * passes.def: Remove pass_fixup_cfg.
6257 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
6258 track when we need to remove functions.
6259 (gate_ipa_inline): Execute inlining always; add comment why.
6260 (pass_data_ipa_inline): Remove TODO_remove_functions.
6261 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
6262 do not produce summaries.
6263 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
6264 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
6265 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
6266 constant pool and vtable.
6268 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
6270 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
6271 New arm_neon.h's internal macros to specify 64-bit constants.
6272 Avoid using stdint.h's macros.
6274 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
6276 * recog.c (verify_changes): Verify that changes[i].old is non-zero
6277 before applying REG_P.
6279 2013-08-30 Jakub Jelinek <jakub@redhat.com>
6281 PR tree-optimization/58277
6282 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
6283 after seeing too many stmts with vdef in between dombb and current
6284 bb, invalidate everything.
6286 2013-08-30 Richard Biener <rguenther@suse.de>
6288 * fold-const.c (fold_single_bit_test): Fix overflow test.
6290 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
6292 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
6293 and which can live in a register, always retrieve the value on entry.
6294 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
6295 passed by invisible reference specially.
6296 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
6297 (vt_add_function_parameter): Correctly deal with a parameter passed by
6298 invisible reference.
6300 2013-08-30 Jan Hubicka <jh@suse.cz>
6302 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
6304 2013-08-30 Richard Biener <rguenther@suse.de>
6306 PR tree-optimization/58228
6307 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
6308 allow invariant loads in nested loop vectorization.
6310 2013-08-30 Richard Biener <rguenther@suse.de>
6312 PR tree-optimization/58223
6313 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
6314 (has_anti_or_output_dependence): ... this and adjust to also
6315 look for output dependences.
6316 (mark_nodes_having_upstream_mem_writes): Adjust.
6317 (rdg_flag_uses): Likewise.
6319 2013-08-30 Richard Biener <rguenther@suse.de>
6321 PR tree-optimization/58010
6322 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
6323 assert that we have a loop-closed PHI.
6325 2013-08-29 Jan Hubicka <jh@suse.cz>
6327 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
6328 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
6329 * lto-section-in.c (lto_free_function_in_decl_state): New function.
6330 (lto_free_function_in_decl_state_for_node): New function.
6332 2013-08-29 Xinliang David Li <davidxl@google.com>
6334 * loop-unroll.c (report_unroll_peel): Minor message change.
6335 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
6336 Emit alignment peeling message with default -fopt-info.
6337 (vect_loop_versioning): Emit loop version info message.
6338 * tree-vectorizer.c (vectorize_loops): Minor message change.
6339 (execute_vect_slp): Ditto.
6341 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
6343 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
6344 of the clone from the DECL_NAME of the original function.
6346 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
6348 * passes.c (register_pass): Add overload.
6349 * tree-pass.h (register_pass): Forward declare it. Add comment.
6351 2013-08-29 Jan Hubicka <jh@suse.cz>
6353 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
6354 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
6355 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
6356 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
6357 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
6358 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
6359 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
6360 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
6362 2013-08-29 Teresa Johnson <tejohnson@google.com>
6364 * dumpfile.c (dump_loc): Output column number.
6365 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
6366 * doc/invoke.texi: Document optall -fopt-info flag.
6367 * profile.c (read_profile_edge_counts): Use new dump framework.
6368 (compute_branch_probabilities): Ditto.
6369 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
6370 when pass not in any opt group.
6371 * pass_manager.h (pass_manager::get_pass_profile): New method.
6372 * value-prof.c (check_counter): Use new dump framework.
6373 (check_ic_target): Ditto.
6374 * coverage.c (get_coverage_counts): Ditto.
6375 (coverage_init): Setup new dump framework.
6377 2013-08-29 Richard Biener <rguenther@suse.de>
6379 PR tree-optimization/58246
6380 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
6381 handle the dominance check inside a basic-block.
6383 2013-08-29 Richard Biener <rguenther@suse.de>
6386 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
6387 of default defs that appear in abnormal PHI nodes.
6389 2013-08-29 Richard Biener <rguenther@suse.de>
6391 PR tree-optimization/57685
6392 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
6393 single-use operands to avoid exponential complexity.
6395 2013-08-28 Dehao Chen <dehao@google.com>
6397 * ipa-inline.c (edge_badness): Fix integer underflow.
6399 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
6401 * gtm-builtins.def (_ITM_free): Declare leaf.
6403 2013-08-28 Jakub Jelinek <jakub@redhat.com>
6406 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
6407 (*tls_local_dynamic_base_64_largepic): Likewise.
6408 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
6409 Remove predicate from call operand.
6410 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
6411 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
6413 2013-08-28 Jeff Law <law@redhat.com>
6415 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
6416 checks for the number of predecessors and successors allowed.
6417 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
6418 which require copying a joiner block if there is a request which
6419 is a subpath that requires no joiner block copying.
6421 2013-08-28 Jan Hubicka <jh@suse.cz>
6423 * lto-streamer-out.c (DFS_write_tree_body): Drop
6424 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
6425 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
6426 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
6427 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
6428 Do not read DECL_ERROR_ISSUED.
6429 (unpack_ts_decl_with_vis_value_fields): Do not read
6431 (lto_input_ts_binfo_tree_pointers): Do not read
6432 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
6433 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
6434 write DECL_ERROR_ISSUED..
6435 (pack_ts_decl_with_vis_value_fields): Do not write
6437 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
6438 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
6439 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
6440 * tree.h (tree_decl_common): Update comment.
6441 (DECL_ERROR_ISSUED): Remove.
6443 2013-08-28 Jakub Jelinek <jakub@redhat.com>
6446 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
6448 2013-08-28 Jan Hubicka <jh@suse.cz>
6450 * builtins.def (free): Declare leaf.
6452 2013-08-27 David Malcolm <dmalcolm@redhat.com>
6455 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
6456 * configure: Regenerate.
6458 2013-08-27 Martin Jambor <mjambor@suse.cz>
6460 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
6461 (ipa_ancestor_jf_data): Likewise.
6462 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
6463 (ipa_get_jf_pass_through_type_preserved): New function.
6464 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
6465 (ipa_get_jf_ancestor_type_preserved): New function.
6466 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
6467 (ipa_get_jf_ancestor_result): Likewise.
6468 (propagate_vals_accross_pass_through): Use
6469 ipa_get_jf_pass_through_result to do all the value mappings.
6470 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
6471 type_preserved flag.
6472 (ipa_set_jf_cst_copy): New function.
6473 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
6474 (ipa_set_jf_arith_pass_through): Likewise.
6475 (ipa_set_ancestor_jf): Likewise.
6476 (compute_complex_assign_jump_func): Set type_preserved instead of
6478 (ipa_compute_jump_functions_for_edge): Likewise.
6479 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
6480 (update_jump_functions_after_inlining): Update type_preserved.
6481 Explicitely create jump functions when combining one with pass_through.
6482 (ipa_write_jump_function): Stream the type_preserved flags.
6483 (ipa_read_jump_function): Likewise.
6485 2013-08-27 Jakub Jelinek <jakub@redhat.com>
6486 Aldy Hernandez <aldyh@redhat.com>
6488 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
6489 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
6490 * function.h (struct function): Add has_force_vect_loops and
6492 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
6493 * gimple.c (gimple_build_omp_critical): Add KIND argument and
6495 * gimple.def: Update CLAUSES comments.
6496 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
6497 (gimple_build_omp_for): Add argument to prototype.
6498 (gimple_omp_for_kind): New.
6499 (gimple_omp_for_set_kind): New.
6500 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
6501 GOVD_DATA_SHARE_CLASS.
6502 (enum omp_region_type): Add ORT_SIMD.
6503 (gimple_add_tmp_var): Handle ORT_SIMD.
6504 (gimplify_var_or_parm_decl): Same.
6505 (is_gimple_stmt): Same.
6506 (omp_firstprivatize_variable): Same.
6507 (omp_add_variable): Only use splay_tree_insert if lookup failed.
6508 (omp_notice_variable): Handle ORT_SIMD.
6509 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
6510 (omp_check_private): Handle ORT_SIMD.
6511 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
6513 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
6514 Handle OMP_CLAUSE_LASTPRIVATE.
6515 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
6517 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
6518 (gimplify_expr): Handle OMP_SIMD.
6519 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
6520 (expand_GOMP_SIMD_VF): New.
6521 (expand_GOMP_SIMD_LAST_LANE): New.
6522 * internal-fn.def (GOMP_SIMD_LANE): New.
6523 (GOMP_SIMD_VF): New.
6524 (GOMP_SIMD_LAST_LANE): New.
6525 * omp-low.c: Include target.h.
6526 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
6528 (check_omp_nesting_restrictions): Same.
6530 (lower_rec_simd_input_clauses): New.
6531 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
6533 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
6534 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
6535 (expand_omp_build_assign): New.
6536 (expand_omp_for_init_counts): New.
6537 (expand_omp_for_init_vars): New.
6538 (extract_omp_for_update_vars): New.
6539 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
6540 and rewrite accordingly.
6541 (expand_omp_simd): New.
6542 (expand_omp_for): Use expand_omp_simd.
6543 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
6544 (lower_omp_for): Do not lower the body.
6545 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
6547 * tree-flow.h (find_omp_clause): Remove prototype.
6548 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
6549 forcing vectorization of the loop, or if flag_tree_vectorize.
6550 (gate_tree_if_conversion): Similarly.
6551 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
6552 gimple_build_omp_for.
6553 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
6554 * tree-parloops (create_parallel_loop): Pass kind argument to
6555 gimple_build_omp_for.
6556 * tree-pretty-print.c (dump_omp_clause): Add cases for
6557 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
6558 OMP_CLAUSE__SIMDUID_.
6559 (dump_generic_node): Handle OMP_SIMD.
6560 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
6561 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
6562 unroll OMP_SIMD loops here.
6563 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
6564 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
6566 (vect_analyze_data_refs): Handle simd loops.
6567 * tree-vect-loop.c (vectorizable_live_operation): Handle
6569 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
6570 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
6571 (vectorizable_load): Same.
6572 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
6573 (struct simduid_to_vf): New.
6574 (simduid_to_vf::hash): New.
6575 (simduid_to-vf::equal): New.
6576 (struct simd_array_to_simduid): New.
6577 (simd_array_to_simduid::hash): New.
6578 (simd_array_to_simduid::equal): New.
6579 (adjust_simduid_builtins): New.
6580 (struct note_simd_array_uses_struct): New.
6581 (note_simd_array_uses_cb): New.
6582 (note_simd_array_uses): New.
6583 (vectorize_loops): Handle simd hints and adjust simd builtins
6585 * tree-vectorizer.h (struct _stmt_vec_info): Add
6586 simd_lane_access_p field.
6587 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
6588 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
6589 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
6590 (omp_clause_code_name): Same.
6591 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
6592 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
6593 * tree.def (OMP_SIMD): New entry.
6594 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
6595 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
6596 (OMP_CLAUSE_DECL): Adjust range for new clauses.
6597 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
6598 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
6599 (OMP_CLAUSE_LINEAR_STEP): New.
6600 (OMP_CLAUSE_SAFELEN_EXPR): New.
6601 (OMP_CLAUSE__SIMDUID__DECL): New.
6602 (find_omp_clause): New prototype.
6604 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
6606 * config/i386/driver-i386.c (host_detect_local_cpu): Update
6607 Haswell processor detection.
6609 2013-08-27 Christian Widmer <shadow@umbrox.de>
6612 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
6613 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
6614 AVX2 capable processors.
6616 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
6618 * config/aarch64/arm_neon.h: Replace all inline asm implementations
6619 of vget_low_* with implementations in terms of other intrinsics.
6621 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
6624 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
6625 values to -1, 0 and 1.
6627 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
6629 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
6630 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
6631 (<optab><mode>3_insn): Remove.
6632 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
6633 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
6634 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
6635 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
6636 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
6637 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
6638 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
6639 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
6640 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
6641 (ror<mode>3_insn): Likewise.
6642 * config/aarch64/predicates.md (aarch64_simd_register): New.
6644 2013-08-27 Richard Biener <rguenther@suse.de>
6646 PR tree-optimization/57521
6647 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
6648 one edge is non-critical.
6649 (find_phi_replacement_condition): Make sure to use a non-critical
6650 edge. Cleanup and remove old bug workarounds.
6651 (bb_postdominates_preds): Remove.
6652 (if_convertible_loop_p_1): Do not compute post-dominators.
6653 (combine_blocks): Do not free post-dominators.
6654 (main_tree_if_conversion): Likewise.
6655 (pass_data_if_conversion): Add TODO_verify_ssa.
6657 2013-08-27 DJ Delorie <dj@redhat.com>
6659 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
6661 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
6663 * function.c (assign_parm_find_data_types): Set passed_mode and
6664 nominal_mode to the TYPE_MODE of nominal_type for the built
6665 pointer type in case of the struct-pass-by-reference.
6667 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6669 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
6670 (UINT16_TYPE): Change default to "unsigned int".
6672 * config/avr/avr.opt (mfract-convert-truncate): New option.
6673 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
6674 is set, round negative fractional integers according to n1169
6675 when converting to integer types.
6677 2013-08-26 Jan Hubicka <jh@suse.cz>
6679 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
6680 methods can not be called indirectly when their address is not taken.
6682 2013-08-26 Jan Hubicka <jh@suse.cz>
6684 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
6687 2013-08-26 Jan Hubicka <jh@suse.cz>
6689 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
6692 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6694 * reload.c (find_valid_class): Allow classes that do not include
6695 FIRST_PSEUDO_REGISTER - 1.
6697 2013-08-26 Jan Hubicka <jh@suse.cz>
6699 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
6700 fix edge count/frequency when speculation failed; fix type check
6701 for the direct call.
6703 2013-08-26 Jan Hubicka <jh@suse.cz>
6705 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
6707 2013-08-26 Jan Hubicka <jh@suse.cz>
6709 * ipa-inline-transform.c (inline_transform): Be ready for basic block
6710 to be changed by edge redirection.
6712 2013-08-26 Jan Hubicka <jh@suse.cz>
6714 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
6715 formating; add sanity check.
6716 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
6717 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
6718 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
6719 (resolve_noninline_speculation): Update callee keys, too.
6721 2013-08-26 Jan Hubicka <jh@suse.cz>
6723 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
6724 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
6726 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6728 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
6731 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
6733 * config/i386/i386.c (ix86_debug_options): Remove prototype.
6734 (x86_64_elf_select_section): Ditto.
6735 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
6737 (ix86_pass_by_reference): Ditto.
6738 (output_set_got): Ditto.
6739 (ix86_unary_operator_ok): Ditto.
6740 (ix86_expand_builtin): Ditto.
6742 2013-08-23 Jan Hubicka <jh@suse.cz>
6744 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
6746 2013-08-23 Jan Hubicka <jh@suse.cz>
6748 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
6749 (tree_decl_with_vis): Add FINAL field.
6751 2013-08-23 Jeff Law <law@redhat.com>
6753 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
6754 the debugging dump when the expression is fully redundant.
6756 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
6758 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
6759 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
6760 * pretty-print.c (pp_formatted_text_data): Likewise.
6761 (pp_write_text_to_stream): Likewise.
6762 (pp_write_text_as_dot_label_to_stream): Likewise.
6763 (pp_append_r): Likewise.
6764 (pp_format): Likewise.
6765 (pp_flush): Likewise.
6766 (pp_clear_output_area): Likewise.
6767 (pp_append_text): Likewise.
6768 (pp_formatted_text): Likewise.
6769 (pp_remaining_character_count_for_line): Likewise.
6770 (pp_newline): Likewise.
6771 (pp_character): Likewise.
6772 (output_buffer::~output_buffer): Define.
6773 (pretty_printer::~pretty_printer): Destruct output buffer.
6774 * pretty-print.h (output_buffer::~output_buffer): Declare.
6775 (pretty_printer::~pretty_printer): Declare virtual.
6777 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
6780 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
6781 HOST_WIDE_INT_M1U): New macros.
6782 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
6783 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
6784 unsigned -1 for lshift.
6785 * cse.c (cse_insn): Likewise.
6786 * double-int.c (rshift_double, lshift_double): Likewise.
6787 * builtins.c (fold_builtin_bitop): Likewise.
6788 * combine.c (force_to_mode): Likewise.
6789 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
6790 * simplify-rtx.c (simplify_const_unary_operation,
6791 simplify_const_binary_operation): Likewise.
6792 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
6793 check_va_list_escapes): Likewise.
6794 * rtlanal.c (nonzero_bits1): Likewise.
6795 * expmed.c (expand_smod_pow2): Likewise.
6796 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
6798 2013-08-23 Jan Hubicka <jh@suse.cz>
6800 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
6801 as having address taken.
6803 2013-08-23 Jan Hubicka <jh@suse.cz>
6805 * ipa-utils.h (method_class_type): Declare.
6806 * ipa-devirt.c (method_class_type): Export.
6808 * cgraphunit.c (analyze_functions): Do basic devirtualization;
6809 do not walk base classes of anonymous types.
6811 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
6813 PR rtl-optimization/58220
6815 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
6816 handle SEQUENCE insns properly.
6818 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
6820 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
6822 (pp_newline_and_indent): Likewise.
6823 (pp_separate_with): Likewise.
6824 * pretty-print.c (pp_newline_and_flush): Define.
6825 (pp_newline_and_indent): Likewise.
6826 (pp_separate_with): Likewise.
6828 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6831 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
6832 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
6834 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
6836 * config/i386/predicates.md (ext_sse_reg_operand): New.
6837 * config/i386/i386.md (*movti_internal): Use
6838 predicate to determine if EVEX is needed.
6839 (*movsi_internal): Ditto.
6840 (*movdf_internal): Ditto.
6841 (*movsf_internal): Ditto.
6842 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
6844 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6846 PR tree-optimization/58209
6847 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
6848 (find_tail_calls): Give up for pointer result types if m is non-NULL.
6849 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
6850 emit POINTER_PLUS_EXPR.
6851 (create_tailcall_accumulator): For pointer result type accumulate in
6854 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
6856 * configure.ac: Add backslashes missing from the last change.
6857 * configure: Regenerate.
6859 2013-08-22 Jan Hubicka <jh@susue.cz>
6861 * ipa.c (function_and_variable_visibility): First remember function
6862 was global and then make it local.
6864 2013-08-22 Julian Brown <julian@codesourcery.com>
6866 * configure.ac: Add aarch64 to list of arches which use "nop" in
6868 * configure: Regenerate.
6870 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6872 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
6873 gnu_libc_has_function.
6874 * config/s390/tpf.h: Likewise.
6876 2013-08-22 Jan Hubicka <jh@susue.cz>
6878 * timevar.c (validate_phases): Add cast.
6880 2013-08-22 Jan Hubicka <jh@susue.cz>
6882 * timevar.c (validate_phases): Use size_t for memory.
6883 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
6885 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
6887 * pretty-print.h (output_buffer::output_buffer): Declare.
6888 (pretty_printer::pretty_printer): Likewise.
6889 (pp_construct): Remove.
6890 * pretty-print.c (output_buffer::output_buffer): Define.
6891 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
6892 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
6893 (print_gimple_expr): Likewise.
6894 (print_gimple_seq): Likewise.
6895 (gimple_dump_bb): Likewise.
6896 * sched-vis.c (dump_value_slim): Likewise.
6897 (dump_insn_slim): Likewise.
6898 (dump_rtl_slim): Likewise.
6899 (str_pattern_slim): Likewise.
6900 * tree-mudflap.c (mf_varname_tree): Likewise.
6901 * graph.c (print_graph_cfg): Likewise.
6902 (start_graph_dump): Likewise.
6903 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
6905 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
6906 pretty printer initialization.
6907 * coretypes.h (diagnostic_context): Remove superflous type alias
6909 (pretty_printer): Likewise. Declare directly as a class.
6910 (pretty_print_info): Remove declaration as class.
6911 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
6912 and pp_clear_output_area.
6913 (asan_add_global): Likewise.
6915 2013-08-22 Jan Hubicka <jh@suse.cz>
6917 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
6918 * ipa-utils.h (update_type_inheritance_graph): Declare.
6919 (possible_polymorphic_call_target_p): Declare.
6920 (possible_polymorphic_call_target_p): New.
6921 * ipa-devirt.c: Update toplevel comments.
6922 (cached_polymorphic_call_targets): Move up.
6923 (odr_type_d): Move ID down.
6924 (polymorphic_type_binfo_p): Update comment.
6925 (odr_hasher::remove): Likewise;
6926 (get_odr_type): Set anonymous_namespace.
6927 (dump_odr_type): Dump it.
6928 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
6929 (maybe_record_node): Record node in cached_polymorphic_call_targets.
6930 (record_binfo): Add comment.
6931 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
6933 (devirt_node_removal_hook): Do not iCE when cache is freed.
6934 (possible_polymorphic_call_target_p): New predicate.
6935 (update_type_inheritance_graph): New function.
6937 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
6938 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6939 Sergey Lega <sergey.s.lega@intel.com>
6940 Anna Tikhonova <anna.tikhonova@intel.com>
6941 Ilya Tocar <ilya.tocar@intel.com>
6942 Andrey Turetskiy <andrey.turetskiy@intel.com>
6943 Ilya Verbin <ilya.verbin@intel.com>
6944 Kirill Yukhin <kirill.yukhin@intel.com>
6945 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6947 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
6948 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
6949 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
6950 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
6951 (OPTION_MASK_ISA_AVX2_UNSET): Update.
6952 (OPTION_MASK_ISA_AVX512F_UNSET): New.
6953 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
6954 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
6955 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
6956 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
6957 OPT_mavx512pf, OPT_mavx512er cases.
6958 * config/i386/constraints.md (v): New constraint.
6959 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
6960 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
6961 (bit_AVX512CD): New.
6962 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6963 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
6964 * config/i386/i386-c.c (ix86_target_macros_internal):
6965 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
6967 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
6968 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
6969 * config/i386/i386.c (regclass_map, dbx_register_map)
6970 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
6971 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
6972 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
6974 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
6975 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
6976 -mavx512pf options. Fix formatting.
6977 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
6978 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
6979 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
6980 -mavx512cd, -mavx512pf options.
6981 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
6982 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
6983 (ix86_preferred_output_reload_class): Replace SSE_REGS with
6985 (ix86_hard_regno_mode_ok): Support 512-bit registers.
6986 (ix86_set_reg_reg_cost): Ditto.
6987 (x86_order_regs_for_local_alloc): Ditto.
6988 (MAX_VECT_LEN): Extend to 64-byte.
6989 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
6990 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
6991 (TARGET_AVX512ER, TARGET_AVX512CD): New.
6992 (BIGGEST_ALIGNMENT): Extend to 512-bits.
6993 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
6994 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
6995 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
6996 (SSE_REG_MODE_P): Support new modes.
6997 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
6998 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
6999 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
7000 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
7001 (REG_CLASS_CONTENTS): Add new registers.
7002 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
7003 (EXT_REX_SSE_REGNO_P): New.
7004 (HI_REGISTER_NAMES): Add new registers.
7005 * config/i386/i386.md: Define constants for new registers.
7006 (mode): Add new 512-bit modes.
7007 (prefix): Support evex prefix.
7008 (isa): Support avx512f, noavx512f, fma_avx512f.
7009 (ssemodesuffix): Add new 512-bit modes.
7011 (*movxi_internal_avx512f): Ditto.
7012 (*movdi_internal): Replace constraint "x" with the new constraint "v".
7014 (*movsi_internal): Likewise.
7015 (*movdf_internal): Likewise.
7016 (*movsf_internal): Likewise.
7017 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
7018 (<code><mode>3): Likewise.
7019 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
7021 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
7022 with the new constraint "v".
7023 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
7025 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
7026 with the new constraint "v".
7027 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
7028 (<sse2>_storedqu<avxsizesuffix>): Likewise.
7029 (*<plusminus_insn><mode>3): Likewise.
7030 (<sse>_vm<plusminus_insn><mode>3): Likewise.
7031 (*mul<mode>3): Likewise.
7032 (<sse>_vmmul<mode>3): Likewise.
7033 (<sse>_div<mode>3): Likewise.
7034 (<sse>_vmdiv<mode>3): Likewise.
7035 (<sse>_sqrt<mode>2): Likewise.
7036 (<sse>_vmsqrt<mode>2): Likewise.
7037 (*<code><mode>3_finite): Likewise.
7038 (*<code><mode>3) <smaxmin>: Likewise.
7039 (<sse>_vm<code><mode>3): Likewise.
7040 (*<code><mode>3) <any_logic>: Likewise.
7041 (*fma_fmadd_<mode>): Likewise.
7042 (*fma_fmsub_<mode>): Likewise.
7043 (*fma_fnmadd_<mode>): Likewise.
7044 (*fma_fnmsub_<mode>): Likewise.
7045 (*fma_fmaddsub_<mode>): Likewise.
7046 (*fma_fmsubadd_<mode>): Likewise.
7047 (*fmai_fmadd_<mode>): Likewise.
7048 (*fmai_fmsub_<mode>): Likewise.
7049 (*fmai_fnmadd_<mode>): Likewise.
7050 (*fmai_fnmsub_<mode>): Likewise.
7051 (sse_cvtsi2ss): Likewise.
7052 (sse_cvtsi2ssq): Likewise.
7053 (sse_cvtss2si): Likewise.
7054 (sse_cvtss2si_2): Likewise.
7055 (sse_cvtss2siq): Likewise.
7056 (sse_cvtss2siq_2): Likewise.
7057 (sse_cvttss2si): Likewise.
7058 (sse_cvtss2siq_2): Likewise.
7059 (float<sseintvecmodelower><mode>2): Likewise.
7060 (sse2_cvtsd2si_2): Likewise.
7061 (sse2_cvtsd2siq_2): Likewise.
7062 (*<plusminus_insn><mode>3): Likewise.
7063 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
7064 (*<sse4_1_avx2>_mul<mode>3): Likewise.
7065 (ashr<mode>3): Likewise.
7066 (<shift_insn><mode>3): Likewise.
7067 (avx2_<code><mode>3): Likewise.
7068 (*avx2_<code><mode>3): Likewise.
7069 (*andnot<mode>3): Likewise.
7070 (*<code><mode>3) <any_logic>: Likewise.
7071 (abs<mode>2): Likewise.
7072 (avx2_permvar<mode>): Likewise.
7073 (avx2_perm<mode>_1): Likewise.
7074 (*avx_vpermilp<mode>): Likewise.
7075 (avx_vpermilvar<mode>3): Likewise.
7076 (avx2_ashrv<mode>): Likewise.
7077 (avx2_<shift_insn>v<mode>): Likewise.
7078 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
7080 * doc/rtl.texi: Document XImode.
7082 2013-08-21 Jeff Law <law@redhat.com>
7084 * tree-flow.h (register_jump_thread): Pass vector of edges
7085 instead of each important edge.
7086 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
7087 thread path into a vector and pass that to register_jump_thread.
7088 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
7089 passed in edge vector to the current 3-edge form.
7092 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
7094 * dce.c (fini_dce): Call df_analyze again just in case
7095 delete_unmarked_insns removed anything.
7097 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
7099 * reload.h (struct reg_equivs): Rename to ..
7100 (struct reg_equivs_s): .. this.
7102 2013-08-20 Martin Liska <marxin.liska@gmail.com>
7104 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
7106 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7108 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
7110 2013-08-21 Jeff Law <law@redhat.com>
7112 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
7113 simplify assignments too. If the RHS collapses to a singleton
7114 range, then return the value for the range.
7116 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
7118 * config/i386/sse.md (V16): Rename to...
7120 (mov<mode>): Update iterator name.
7121 (*mov<mode>_internal): Ditto.
7122 (push<mode>1): Ditto.
7123 (movmisalign<mode>): Ditto.
7125 2013-08-20 Jan Hubicka <jh@suse.cz>
7128 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
7129 entry for direct edges.
7130 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
7132 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7134 Revert my last two changes, r201865 and r201864:
7137 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7139 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
7140 instances can own GC refs.
7142 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
7143 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
7144 (gcc::context::gt_pch_nx): Likewise.
7145 (gcc::context::gt_pch_nx): Likewise.
7146 * ggc.h (gt_ggc_mx <T>): New.
7147 (gt_pch_nx_with_op <T>): New.
7148 (gt_pch_nx <T>): New.
7149 * passes.c (opt_pass::gt_ggc_mx): New.
7150 (opt_pass::gt_pch_nx): New.
7151 (opt_pass::gt_pch_nx_with_op): New.
7152 (pass_manager::gt_ggc_mx): New.
7153 (pass_manager::gt_pch_nx): New.
7154 (pass_manager::gt_pch_nx_with_op): New.
7155 (pass_manager::operator new): Use
7156 ggc_internal_cleared_alloc_stat rather than xcalloc.
7157 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
7158 (pass_manager::gt_ggc_mx): New.
7159 (pass_manager::gt_pch_nx): New.
7160 (pass_manager::gt_pch_nx_with_op): New.
7161 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
7162 (opt_pass::operator new): New.
7163 (opt_pass::gt_ggc_mx): New.
7164 (opt_pass::gt_pch_nx): New.
7165 (opt_pass::gt_pch_nx_with_op): New.
7168 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7170 * Makefile.in (GTFILES): Add context.h.
7171 * context.c (gcc::context::operator new): New.
7172 (gcc::context::gt_ggc_mx): New.
7173 (gcc::context::gt_pch_nx): New.
7174 (gcc::context::gt_pch_nx): New.
7175 * context.h (gcc::context): Add GTY((user)) marking.
7176 (gcc::context::operator new): New.
7177 (gcc::context::gt_ggc_mx): New.
7178 (gcc::context::gt_pch_nx): New.
7179 (gcc::context::gt_pch_nx): New.
7180 (g): Add GTY marking.
7181 (gt_ggc_mx (gcc::context *)): New.
7182 (gt_pch_nx (gcc::context *)): New.
7183 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
7184 void *cookie)): New.
7185 * gengtype.c (open_base_files) <ifiles>: Add context.h.
7187 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
7189 * dce.c (fini_dce): Call df_analyze again just in case
7190 delete_unmarked_insns removed anything.
7192 2013-08-20 Teresa Johnson <tejohnson@google.com>
7194 PR rtl-optimizations/57451
7195 * final.c (reemit_insn_block_notes): Prevent lexical blocks
7196 from crossing split section boundaries.
7198 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7200 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
7201 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
7202 with MULTLIB_DEFAULTS.
7204 2013-08-20 Nick Clifton <nickc@redhat.com>
7206 * target.def (narrow_volatile_bitfield): Note that the default
7207 value is false, not !TARGET_STRICT_ALIGN.
7208 * doc/tm.texi: Regenerate.
7210 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
7212 Fix LIB_SPEC for systems without libpthread.
7214 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
7215 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
7217 * config/i386/linux-common.h: Likewise.
7218 * config/mips/linux-common.h: Likewise.
7220 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
7222 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
7225 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7227 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
7228 instances can own GC refs.
7230 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
7231 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
7232 (gcc::context::gt_pch_nx): Likewise.
7233 (gcc::context::gt_pch_nx): Likewise.
7234 * ggc.h (gt_ggc_mx <T>): New.
7235 (gt_pch_nx_with_op <T>): New.
7236 (gt_pch_nx <T>): New.
7237 * passes.c (opt_pass::gt_ggc_mx): New.
7238 (opt_pass::gt_pch_nx): New.
7239 (opt_pass::gt_pch_nx_with_op): New.
7240 (pass_manager::gt_ggc_mx): New.
7241 (pass_manager::gt_pch_nx): New.
7242 (pass_manager::gt_pch_nx_with_op): New.
7243 (pass_manager::operator new): Use
7244 ggc_internal_cleared_alloc_stat rather than xcalloc.
7245 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
7246 (pass_manager::gt_ggc_mx): New.
7247 (pass_manager::gt_pch_nx): New.
7248 (pass_manager::gt_pch_nx_with_op): New.
7249 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
7250 (opt_pass::operator new): New.
7251 (opt_pass::gt_ggc_mx): New.
7252 (opt_pass::gt_pch_nx): New.
7253 (opt_pass::gt_pch_nx_with_op): New.
7255 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7257 * Makefile.in (GTFILES): Add context.h.
7258 * context.c (gcc::context::operator new): New.
7259 (gcc::context::gt_ggc_mx): New.
7260 (gcc::context::gt_pch_nx): New.
7261 (gcc::context::gt_pch_nx): New.
7262 * context.h (gcc::context): Add GTY((user)) marking.
7263 (gcc::context::operator new): New.
7264 (gcc::context::gt_ggc_mx): New.
7265 (gcc::context::gt_pch_nx): New.
7266 (gcc::context::gt_pch_nx): New.
7267 (g): Add GTY marking.
7268 (gt_ggc_mx (gcc::context *)): New.
7269 (gt_pch_nx (gcc::context *)): New.
7270 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
7271 void *cookie)): New.
7272 * gengtype.c (open_base_files) <ifiles>: Add context.h.
7274 2013-08-20 Alan Modra <amodra@gmail.com>
7277 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
7278 (rs6000_emit_epilogue): Likewise.
7280 2013-08-19 Dehao Chen <dehao@google.com>
7282 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
7284 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
7285 Jakub Jelinek <jakub@redhat.com>
7287 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
7288 (BUILT_IN_FABSD64): Likewise.
7289 (BUILT_IN_FABSD128): Likewise.
7290 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
7291 (fold_builtin_1): Likewise.
7292 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
7293 destination and source operands.
7294 (*abstd2_fpr): Likewise.
7295 (*nabstd2_fpr): Likewise.
7297 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
7299 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
7302 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
7304 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
7307 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
7309 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
7310 * builtins.c (default_libc_has_function): New.
7311 (gnu_libc_has_function): Ditto.
7312 (no_c99_libc_has_function): Ditto.
7313 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
7314 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
7315 (fold_builtin_sincos): Likewise.
7316 (fold_builtin_cexp): Likewise.
7317 * builtins.def (DEF_C94_BUILTIN): Likewise.
7318 (DEF_C99_BUILTIN): Likewise.
7319 (DEF_C99_C90RES_BUILTIN): Likewise.
7320 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
7321 definitions to using this define.
7322 * config/darwin-protos.h (darwin_libc_has_function): New.
7323 * config/darwin.c (darwin_libc_has_function): Ditto.
7324 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
7325 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
7326 * config/darwin.h: Ditto.
7327 * config/elfos.h: Ditto.
7328 * config/freebsd.h: Ditto.
7329 * config/i386/cygming.h: Ditto.
7330 * config/i386/djgpp.h: Ditto.
7331 * config/i386/i386-interix.h: Ditto.
7332 * config/microblaze/microblaze.h: Ditto.
7333 * config/mmix/mmix.h: Ditto.
7334 * config/gnu-user.h: Ditto.
7335 * config/ia64/hpux.h: Ditto.
7336 * config/pa/pa-hpux.h: Ditto.
7337 * config/pdp11/pdp11.h: Ditto.
7338 * config/picochip/picochip.h: Ditto.
7339 * config/linux.h: Ditto.
7340 * config/netbsd.h: Ditto.
7341 * config/openbsd.h: Ditto.
7342 * config/rs6000/aix43.h: Ditto.
7343 * config/rs6000/aix51.h: Ditto.
7344 * config/rs6000/aix52.h: Ditto.
7345 * config/rs6000/aix53.h: Ditto.
7346 * config/rs6000/aix61.h: Ditto.
7347 * config/rs6000/darwin.h: Ditto.
7348 * config/rs6000/linux.h: Ditto.
7349 * config/rs6000/linux64.h: Ditto.
7350 * config/s390/tpf.h: Ditto.
7351 * config/sol2-10.h: Ditto.
7352 * config/sol2.h: Ditto.
7353 * config/vms/vms.h: Ditto.
7354 * config/vxworks.h: Ditto.
7355 * config/linux-android.c (linux_android_libc_has_function):
7356 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
7357 * config/linux-protos.h (linux_android_libc_has_function):
7359 * config/i386/i386.c (ix86_libc_has_function): New.
7360 * config/i386/i386-protos.h
7361 (ix86_libc_has_function): New declaration.
7362 * config/i386/i386.md
7363 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
7364 ("isinf<mode>2): Likewise.
7365 * convert.c (convert_to_integer): Using new target hook
7366 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
7367 TARGET_C99_FUNCTIONS.
7368 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
7369 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
7370 * coretypes.h (function_class): New enum for different
7371 classes of functions.
7372 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
7373 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
7374 (TARGET_HAS_SINCOS): Likewise.
7375 (TARGET_LIBC_HAS_FUNCTION): New.
7376 * doc/tm.texi: Regenerated.
7377 * targhooks.h (default_libc_has_function): New declaration.
7378 (no_c99_libc_has_function): Ditto.
7379 (gnu_libc_has_function): Ditto.
7380 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
7381 and TARGET_HAS_SINCOS.
7383 2013-08-18 Jan Hubicka <jh@suse.cz>
7385 * Makeifle-in (ipa-devirt.o): New.
7386 (GTFILES): Add ipa-utils.h and ipa-devirt.c
7387 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
7388 (analyze_functions): Look into possible targets of polymorphic call.
7389 * dumpfile.c (dump_files): Add type-inheritance dump.
7390 * dumpfile.h (TDI_inheritance): New.
7391 * ipa-devirt.c: New file.
7392 * ipa-utils.h (odr_type_d): Forward declare.
7393 (odr_type): New type.
7394 (build_type_inheritance_graph): Declare.
7395 (possible_polymorphic_call_targets): Declare and introduce inline
7396 variant when only edge is pased.
7397 (dump_possible_polymorphic_call_targets): Likewise.
7398 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
7399 * tree.c (type_in_anonymous_namespace_p): Break out from ...
7400 (types_same_for_odr): ... here.
7401 * tree.h (type_in_anonymous_namespace_p): Declare.
7403 2013-08-18 Jakub Jelinek <jakub@redhat.com>
7405 PR tree-optimization/58006
7406 * tree-parloops.c (take_address_of): Don't ICE if get_name
7408 (eliminate_local_variables_stmt): Remove clobber stmts.
7410 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
7412 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
7413 error message is issued for an alias to undefined symbol.
7415 2013-08-18 Jan Hubicka <jh@suse.cz>
7417 * cgraph.c (cgraph_create_indirect_edge): Discover
7418 polymorphic calls and record basic info into indirect_info.
7419 * gimple-fold.c (gimple_fold_call): When doing BINFO based
7420 devirtualization, ignore objc function calls.
7421 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
7422 call with no parm index info.
7423 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
7424 * tree.c (virtual_method_call_p): New function.
7425 * tree.h (virtual_method_call_p): Declare.
7427 2013-08-16 Jan Hubicka <jh@suse.cz>
7430 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
7432 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
7434 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
7437 2013-08-16 David Malcolm <dmalcolm@redhat.com>
7439 * gengtype.c (type_for_name): Add special-case support for
7440 locating types within the "gcc::" namespace.
7441 (open_base_files): Emit a "using namespace gcc" directive.
7443 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7446 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
7447 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
7449 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
7450 array instead of each individual operand as a separate argument.
7451 (emit_fusion_gpr_load): Likewise.
7452 (expand_fusion_gpr_load): Add new function declaration.
7454 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
7455 signature to have the operands passed as an array, instead of as
7456 separate arguments. Allow ZERO_EXTEND to be in the memory
7457 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
7458 depend on the register live/dead flags when peepholes are run.
7459 (expand_fusion_gpr_load): New function to be called from the
7460 peephole2 pass, to change the register that addis sets to be the
7462 (emit_fusion_gpr_load): Change the calling signature to have the
7463 operands passed as an array, instead of as separate arguments.
7464 Allow ZERO_EXTEND to be in the memory address, and also
7465 SIGN_EXTEND if -mpower8-fusion-sign.
7467 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
7469 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
7470 adjust the register addis loads up in the peephole2 pass. Do not
7471 depend on the register live/dead state when the peephole pass is done.
7473 2013-08-16 David Malcolm <dmalcolm@redhat.com>
7475 * gengtype.c (create_user_defined_type): Ensure that the kind
7476 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
7477 declaration is seen before the GTY((user)) marking.
7479 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
7482 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
7484 2013-08-16 Jan Hubicka <jh@suse.cz>
7486 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
7487 arugment expected_type.
7488 (gimple_fold_call): Use it.
7489 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
7490 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
7491 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
7492 (try_make_edge_direct_virtual_call): Likewise.
7493 * tree.c (obj_type_ref_class): New.
7494 * tree.h (obj_type_ref_class): Use it.
7496 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
7498 * sched-vis.c (rtl_slim_pp_initialized): Remove.
7499 (rtl_slim_pp): Likewise.
7500 (init_rtl_slim_pretty_print): Likewise.
7501 (dump_value_slim): Don't call it. Use local pretty printer.
7502 (dump_insn_slim): Likewise.
7503 (dump_rtl_slim): Likewise.
7504 (str_pattern_slim): Likewise.
7505 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
7508 2013-08-16 Jakub Jelinek <jakub@redhat.com>
7510 PR tree-optimization/58164
7511 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
7512 walk gimple_goto_dest of GIMPLE_GOTO.
7514 PR tree-optimization/58165
7515 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
7516 bi_call must be the last stmt in a bb, don't split_block, instead
7517 use fallthru edge from it and give up if there is none.
7518 Release conds vector when returning early.
7520 2013-08-14 Xinliang David Li <davidxl@google.com>
7522 * config/i386/i386.c (ix86_option_override_internal):
7523 Remove unused variable and field.
7525 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7528 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
7529 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
7530 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
7531 and Linux, correct BLKmode alignment when 128-bit alignment is
7532 required and compatibility flag is not set.
7533 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
7534 for zero-size arguments when compatibility flag is not set.
7536 2013-08-14 Jakub Jelinek <jakub@redhat.com>
7538 PR tree-optimization/58145
7539 * tree-sra.c (build_ref_for_offset): If prev_base has
7540 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
7542 2013-08-14 Xinliang David Li <davidxl@google.com>
7544 * config/i386/i386.c (ix86_option_override_internal):
7545 Fix uninitialized variable error.
7547 2013-08-14 Xinliang David Li <davidxl@google.com>
7549 * config/i386/i386.opt: Define two new options.
7550 * config/i386/x86-tune.def: Add arch selector field in macros.
7551 * config/i386/i386.h: Adjust macro definition.
7552 * config/i386/i386.c (ix86_option_override_internal):
7554 (parse_mtune_ctrl_str): New function.
7555 (set_ix86_tune_features): New function.
7556 (ix86_function_specific_restore): Call the new helper function.
7558 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
7560 PR rtl-optimization/57662
7561 * sel-sched.c (code_motion_process_successors): When the current insn
7562 is removed after the recursive traversal, break from the loop.
7563 Add comments and debug printouts.
7565 2013-08-14 Jakub Jelinek <jakub@redhat.com>
7566 Alexandre Oliva <aoliva@redhat.com>
7569 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
7570 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
7571 there also UNSPEC_PLTOFF.
7573 2013-08-14 Marek Polacek <polacek@redhat.com>
7575 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
7576 NUM_CONDITIONS bit positions.
7578 2013-08-13 Cary Coutant <ccoutant@google.com>
7580 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
7581 (attr_checksum): Hash vector contents instead of pointer.
7582 (attr_checksum_ordered): Likewise.
7584 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
7586 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
7587 when Pmode != word_mode. Add length_address attribute.
7588 (sse3_monitor_<mode>): Merge from sse3_monitor and
7589 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
7590 Pmode != word_mode. Update insn length attribute.
7591 * config/i386/i386.c (ix86_option_override_internal): Update
7592 ix86_gen_monitor selection for merged sse3_monitor insn.
7594 2013-08-13 Julian Brown <julian@codesourcery.com>
7596 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
7597 perform invalid legitimization on greater-than-word-size modes for
7600 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
7602 * ira.c (setup_class_translate_array): Use aclass instead of cl
7603 for classes not fully covered by allocno classes.
7605 2013-08-13 Jakub Jelinek <jakub@redhat.com>
7607 PR tree-optimization/57661
7608 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
7609 * tree-inline.c (tree_function_versioning): Initialize it.
7610 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
7611 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
7612 that is not being copied.
7615 * asan.c (instrument_strlen_call): Fix typo in comment.
7616 Use char * type even for the lhs of POINTER_PLUS_EXPR.
7618 2013-08-13 Steve Ellcey <sellcey@mips.com>
7620 * config/mips/mips.md (prefetch): Use lw instead of ld on
7621 loongson in 32bit mode.
7623 2013-08-13 Nick Clifton <nickc@redhat.com>
7625 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
7627 2013-08-13 Jan Hubicka <jh@suse.cz>
7629 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
7630 introduced edge; fix typo in sanity check.
7631 (cgraph_resolve_speculation): Export; improve diagnostic.
7632 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
7633 speculation at type mismatch.
7634 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
7635 (cgraph_resolve_speculation): Declare.
7636 (symtab_can_be_discarded): New function.
7637 * value-prof.c (gimple_ic_transform): Remove actual transform code.
7638 * ipa-inline-transform.c (speculation_removed): New global var.
7639 (clone_inlined_nodes): See if speculation can be removed.
7640 (inline_call): If speculations was removed, we growths may not match.
7641 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
7642 (speculation_useful_p): New function.
7643 (resolve_noninline_speculation): New function.
7644 (inline_small_functions): Resolve useless speculations.
7645 * ipa-inline.h (speculation_useful_p): Declare
7646 * ipa.c (can_replace_by_local_alias): Simplify.
7647 (ipa_profile): Produce speculative calls in non-lto, too;
7648 add simple cost model; produce local aliases.
7650 2013-08-13 David Malcolm <dmalcolm@redhat.com>
7652 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
7655 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
7657 * config/i386/i386.c (ix86_function_versions): Use error + inform.
7659 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
7661 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
7662 iterator instead of X87MODEF.
7664 2013-08-12 Perez Read <netfirewall@gmail.com>
7667 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
7668 operand 0 for intel asm alternative.
7669 (*movabs<mode>_2): Ditto for operand 1.
7671 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
7673 * config/aarch64/arm_none.h
7674 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
7676 2013-08-12 Nick Clifton <nickc@redhat.com>
7678 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
7680 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
7682 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
7683 expand for QI/HImode operand to produce more effictive code for
7684 unsigned char(short) --> float(double) conversion.
7686 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
7688 * doc/invoke.texi: Mention that -ftls-model does not force the final
7691 2013-08-12 Marek Polacek <polacek@redhat.com>
7692 Marc Glisse <marc.glisse@inria.fr>
7694 PR tree-optimization/57980
7695 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
7696 when creating -1 constant.
7698 2013-08-10 Jan Hubicka <jh@suse.cz>
7700 Workaround binutils PR14342.
7701 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
7702 (gimple_init_edge_profiler): Likewise.
7703 (gimple_gen_ic_func_profiler): Likewise.
7705 2013-08-09 Jan Hubicka <jh@suse.cz>
7707 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
7709 2013-08-09 Xinliang David Li <davidxl@google.com>
7711 * config/i386/stringop.def: New file.
7712 * config/i386/stringop.opt: New file.
7713 * config/i386/i386-opts.h: Include stringopt.def.
7714 * config/i386/i386.opt: Include stringopt.opt.
7715 * config/i386/i386.c (ix86_option_override_internal):
7716 Override default size based stringop inline strategies with options.
7717 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
7720 2013-08-09 Jan Hubicka <jh@suse.cz>
7722 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
7724 2013-08-09 Jan Hubicka <jh@suse.cz>
7726 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
7728 (dump_cgraph_node): Dump profile-id.
7729 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
7730 and common_target_probability.
7731 * lto-cgraph.c (lto_output_edge): Stream common targets.
7732 (lto_output_node): Stream profile ids.
7733 (input_node): Stream profile ids.
7734 (input_edge): Stream common targets.
7735 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
7736 * ipa.c: Include value-prof.h
7737 (ipa_profile_generate_summary): Turn indirect call statement histograms
7738 into common targets.
7739 (ipa_profile): Turn common targets into speculative edges.
7741 2013-08-09 Jan Hubicka <jh@suse.cz>
7743 * cgraph.h (cgraph_node): Add profile_id.
7744 * value-prof.c (cgraph_node_map): Turn into pointer_map.
7745 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
7746 (del_node_map): Update.
7747 (find_func_by_funcdef_no): Replace by ...
7748 (find_func_by_profile_id): ... this one.
7749 (gimple_ic_transform): Do not remove useful histograms when
7750 speculation is not done; dump info when indirect call removal
7752 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
7753 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
7754 (__gcov_indirect_call_profiler_v2): .. this one.
7755 * profile.h (init_node_map): Update.
7756 * coverage.c (coverage_compute_profile_id): New function.
7757 * coverage.h (coverage_compute_profile_id): Declare.
7758 * tree-profile.c (init_ic_make_global_vars): Make
7759 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
7760 (gimple_init_edge_profiler): Update prototype of
7761 __gcov_indirect_call_profiler.
7762 (gimple_gen_ic_func_profiler): Simplify.
7763 (tree_profiling): Use init_node_map
7765 2013-08-09 Jan Hubicka <jh@suse.cz>
7767 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
7768 non-speculative refs.
7769 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
7770 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
7771 (cgraph_set_call_stmt): Likewise.
7772 (cgraph_create_edge_1): Fix release checking compilatoin;
7774 (cgraph_free_edge): Free indirect info.
7775 (cgraph_turn_edge_to_speculative): New function.
7776 (cgraph_speculative_call_info): New function.
7777 (cgraph_make_edge_direct): Return direct edge; handle speculation.
7778 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
7779 (dump_cgraph_node): Dump speculation.
7780 (verify_edge_count_and_frequency): Accept speculative edges.
7781 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
7782 (verify_cgraph_node): Handle speculation.
7783 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
7784 (cgraph_set_call_stmt): Update prototype.
7785 (cgraph_make_edge_direct): Update prototype.
7786 (cgraph_speculative_call_info): Declare.
7787 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
7788 to change; update call of ipa_find_references.
7789 * ipa-ref.c (ipa_record_reference): Fix return value; clear
7790 lto_stmt_uid and speculative flags.
7791 (ipa_dump_references): Dump speculation.
7792 (ipa_clone_references): Clone speculative flag.
7793 (ipa_clone_referring): Likewise.
7794 (ipa_clone_ref): New function.
7795 (ipa_find_reference): Look into lto_stmt_uids
7796 (ipa_clear_stmts_in_references): Do not clear speculative calls.
7797 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
7798 (ipa_find_reference): Update declaration.
7799 (ipa_clone_ref): Declare.
7800 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
7801 stream speculative flag.
7802 (lto_output_ref): Stream statements uids and speculation.
7803 (input_ref): Likewise.
7804 (input_edge): Stream speuclation.
7805 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
7806 (cgraph_set_call_stmt_including_clones): Handle speculation.
7807 * ipa-inline.c (heap_edge_removal_hook): New function.
7808 (inline_small_functions): Register it.
7809 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
7810 also initialize refs.
7811 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
7813 (try_make_edge_direct_simple_call): Likewise.
7814 (try_make_edge_direct_simple_call): Likewise.
7815 (update_indirect_edges_after_inlining): Likewise.
7816 (remove_described_reference): Look proper lto_stmt_uid.
7817 (propagate_controlled_uses): Likewise.
7818 (propagate_controlled_uses): Liekwise.
7819 * tree-inline.c (copy_bb): Copy speculative edges.
7820 (redirect_all_calls): New function.
7821 (copy_cfg_body): Do redirection after loop info is updated.
7822 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
7824 2013-08-09 Jan Hubicka <jh@suse.cz>
7826 * lto-streamer-out.c (output_function): Renumber PHIs.
7827 * lto-streamer-in.c (input_function): Likewise.
7829 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
7831 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
7832 (get_lane_unsigned): Likewise.
7833 (dup_lane_scalar): Likewise.
7834 (get_lane): enable for VALL.
7835 * config/aarch64/aarch64-simd.md
7836 (aarch64_dup_lane_scalar<mode>): Remove.
7837 (aarch64_get_lane_signed<mode>): Likewise.
7838 (aarch64_get_lane_unsigned<mode>): Likewise.
7839 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
7840 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
7841 (aarch64_get_lane<mode>): Enable for all vector modes.
7842 (aarch64_get_lanedi): Remove misleading constraints.
7843 * config/aarch64/arm_neon.h
7844 (__aarch64_vget_lane_any): Define.
7845 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
7846 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
7847 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
7848 * config/aarch64/iterators.md (VDQQH): New.
7852 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
7854 * configure.ac: Add GAS check for LEON instructions on SPARC.
7855 * configure: Regenerate.
7856 * config.in: Likewise.
7857 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
7859 * config/sparc/sparc.opt (LEON, LEON3): New masks.
7860 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
7862 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
7863 (AS_LEON_FLAG): New macro.
7864 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
7865 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
7866 Deal with LEON and LEON3 for the memory model.
7867 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
7868 (atomic_compare_and_swap<mode>_1): Likewise.
7869 (*atomic_compare_and_swap<mode>_1): Likewise.
7871 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7873 * config/arm/neon.md (vcond): Fix floating-point vector
7874 comparisons against 0.
7876 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
7878 * lra-constraints.c (emit_spill_move): Remove assert.
7879 (process_alt_operands): Add more debugging
7880 output. Increase reject for spilling into memory. Decrease
7881 reject for reloading scratch.
7882 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
7884 2013-08-08 Steve Ellcey <sellcey@mips.com>
7886 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
7887 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
7888 micromips incompatible. Add nan2008.
7889 (MULTILIB_DIRNAMES): Add nan2008.
7890 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7891 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
7892 and micromips incompatible. Add nan2008.
7893 (MULTILIB_DIRNAMES): Add nan2008.
7894 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7896 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
7898 PR rtl-optimization/58079
7899 * combine.c (combine_simplify_rtx): Avoid using SUBST if
7900 simplify_comparison has widened a comparison with an integer.
7902 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7904 * config/arm/neon.md (movmisalign<mode>): Disable when we
7905 don't allow unaligned accesses.
7906 (*movmisalign<mode>_neon_store): Likewise.
7907 (*movmisalign<mode>_neon_load): Likewise.
7908 (*movmisalign<mode>_neon_store): Likewise.
7909 (*movmisalign<mode>_neon_load): Likewise.
7911 2013-08-08 Jan Hubicka <jh@suse.cz>
7913 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
7914 (make_pass_rebuild_cgraph_edges): Also clear references.
7915 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
7916 * ipa-inline-transform.c (inline_transform): Remove all references
7918 * cgraphunit.c (expand_function): Remove all references after
7920 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
7921 (ipa_find_reference): Rewrite to iterator.
7922 (remove_stmt_references): Likewise.
7923 (ipa_clear_stmts_in_references): New function.
7924 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
7925 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
7927 * ipa-split.c (split_function): Remove references in split function.
7929 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7932 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
7933 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
7935 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7938 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
7939 suggested mode for the assignment isn't compatible with the
7942 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7945 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
7947 2013-08-07 Xinliang David Li <davidxl@google.com>
7949 * config/i386/i386.opt: New option -mtune-ctrl=.
7950 * config/i386/x86-tune.def: New file.
7951 * config/i386/i386.h: include x86-tune.def.
7952 * config/i386/i386.c (ix86_option_override_internal):
7953 Parsing -mtune-ctrl= option and set tune features.
7955 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
7958 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
7959 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
7962 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7964 * rtl.h (update_alignments): Declare.
7965 * final.c (grow_label_align): New function extracted from...
7966 (shorten_branches): ...here. Call it.
7967 (update_alignments): New function.
7968 * reorg.c (sibling_labels): New variable.
7969 (get_label_before): Add SIBLING parameter. If it is non-zero, push
7970 the new label along with it onto the sibling_labels vector.
7971 (fill_simple_delay_slots): Adjust call to get_label_before.
7972 (fill_slots_from_thread): Likewise.
7973 (relax_delay_slots): Likewise.
7974 (make_return_insns): Likewise.
7975 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
7977 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7979 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
7980 document its semantics.
7981 (diagnostic_report_diagnostic): Adjust accordingly.
7983 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7985 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
7986 (sparc_option_override): ...and port to new C++ pass API.
7987 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
7989 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
7991 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
7993 2013-08-06 Caroline Tice <cmtice@google.com>
7995 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
7996 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
7997 * tree-pass.h: Add pass_vtable_verify.
7998 * varasm.c (assemble_variable): Add code to properly set the comdat
7999 section and name for the .vtable_map_vars section.
8000 (assemble_vtyv_preinit_initializer): New function.
8001 (default_sectin_type_flags): Make sure .vtable_map_vars section has
8003 * output.h: Add function decl for assemble_vtv_preinit_initializer.
8004 * vtable-verify.c: New file.
8005 * vtable-verify.h: New file.
8006 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
8007 initialiation levels.
8008 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
8009 * passes.def: Insert pass_vtable_verify.
8010 * aclocal.m4: Reorder includes.
8011 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
8012 -fvtv-counts options.
8013 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
8014 as appropriate, if -fvtable-verify=... is used.
8015 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
8016 -fvtable-verify=... is used.
8017 * Makefile.in (OBJS): Add vtable-verify.o to list.
8018 (vtable-verify.o): Add new build rule.
8019 (GTFILES): Add vtable-verify.c to list.
8020 * common.opt (fvtable-verify=): New flag.
8021 (vtv_priority): Values for fvtable-verify= flag.
8022 (fvtv-counts): New flag.
8023 (fvtv-debug): New flag.
8024 * tree.h (save_vtable_map_decl): New extern function decl.
8026 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8028 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
8029 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
8030 (pass_data_rl78_devirt): ...new pass_data instance and...
8031 (make_pass_rl78_devirt): ...new function.
8032 (rl78_asm_file_start): Port pass registration to new C++ API.
8034 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8036 * coretypes.h (rtl_opt_pass): Add.
8037 (gcc::context): Add.
8038 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
8039 (epiphany_init): Port to new C++ pass API.
8040 (epiphany_optimize_mode_switching): Likewise.
8041 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
8042 (pass_manager::get_pass_mode_switching): New.
8043 (pass_manager::get_pass_peephole2): New.
8044 * mode-switching.c (pass_mode_switching): Add clone method.
8045 * recog.c (pass_peephole2): Add clone method.
8046 (pass_split_all_insns): Add clone method.
8048 2013-08-06 David Malcolm <dmalcolm@redhat.com>
8050 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
8051 (mips_option_override): ...here, porting to new C++ API for passes.
8053 2013-08-06 Jan Hubicka <jh@suse.cz>
8055 * cgraph.c (cgraph_get_body): New function based on lto.c
8057 * cgraph.h (cgraph_get_body): Declare.
8058 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
8060 * cgraphunit.c (expand_function): Get body prior expanding.
8061 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
8062 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
8064 * passes.c (do_per_function_toporder): Get body.
8065 * tree-inline.c (expand_call_inline): Get body prior inlining it.
8066 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
8068 2013-08-06 Martin Jambor <mjambor@suse.cz>
8071 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
8072 is not re-finalized. Rename second parameter to no_collect.
8074 2013-08-06 Martin Jambor <mjambor@suse.cz>
8077 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
8078 MEM_REF has proper alignment information.
8080 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
8083 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
8085 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
8086 rtx (*) (rtx, ...) with insn_gen_fn.
8087 * genoutput.c (output_insn_data): Cast gen_? function pointers to
8088 insn_gen_fn::stored_funcptr. Add initializer braces.
8090 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8092 Rewrite how instances of passes are cloned to remove assumptions
8093 about their sizes (thus allowing pass subclasses to have
8094 additional data fields, albeit non-GC-managed ones at this point).
8096 * passes.c (make_pass_instance): Now that passes have clone
8097 methods, rewrite this function to eliminate XNEW and memcpy
8098 calls that used hardcoded sizes. Since this function no longer
8099 creates pass instances, rename it to...
8100 (add_pass_instance): ...this. Document the old way that passes were
8101 numbered and flagged, and rework this function to continue using it.
8102 (next_pass_1): Add an initial_pass argument for use by
8104 (position_pass): When adding multiple instances of a pass, use
8105 the pass's clone method, rather than relying on the XNEW/memcpy
8106 within the former make_pass_instance (now add_pass_instance).
8107 (pass_manager::pass_manager): When invoking next_pass_1, also supply
8108 the initial instance of the current pass within the pass manager.
8110 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8112 This is the automated part of the conversion of passes from C
8113 structs to C++ classes.
8115 Patch autogenerated by refactor_passes.py from
8116 https://github.com/davidmalcolm/gcc-refactoring-scripts
8117 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
8119 * asan.c (pass_asan): Convert from a global struct to a subclass of
8120 gimple_opt_pass along with...
8121 (pass_data_asan): ...new pass_data instance and...
8122 (make_pass_asan): ...new function.
8123 (pass_asan_O0): Convert from a global struct to a subclass of
8124 gimple_opt_pass along with...
8125 (pass_data_asan_O0): ...new pass_data instance and...
8126 (make_pass_asan_O0): ...new function.
8127 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
8128 subclass of rtl_opt_pass along with...
8129 (pass_data_inc_dec): ...new pass_data instance and...
8130 (make_pass_inc_dec): ...new function.
8131 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
8132 a subclass of rtl_opt_pass along with...
8133 (pass_data_reorder_blocks): ...new pass_data instance and...
8134 (make_pass_reorder_blocks): ...new function.
8135 (pass_duplicate_computed_gotos): Convert from a global struct to a
8136 subclass of rtl_opt_pass along with...
8137 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
8138 (make_pass_duplicate_computed_gotos): ...new function.
8139 (pass_partition_blocks): Convert from a global struct to a subclass of
8140 rtl_opt_pass along with...
8141 (pass_data_partition_blocks): ...new pass_data instance and...
8142 (make_pass_partition_blocks): ...new function.
8143 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
8144 struct to a subclass of rtl_opt_pass along with...
8145 (pass_data_branch_target_load_optimize1): ...new pass_data instance
8147 (make_pass_branch_target_load_optimize1): ...new function.
8148 (pass_branch_target_load_optimize2): Convert from a global struct to a
8149 subclass of rtl_opt_pass along with...
8150 (pass_data_branch_target_load_optimize2): ...new pass_data instance
8152 (make_pass_branch_target_load_optimize2): ...new function.
8153 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
8154 of rtl_opt_pass along with...
8155 (pass_data_jump): ...new pass_data instance and...
8156 (make_pass_jump): ...new function.
8157 (pass_jump2): Convert from a global struct to a subclass of
8158 rtl_opt_pass along with...
8159 (pass_data_jump2): ...new pass_data instance and...
8160 (make_pass_jump2): ...new function.
8161 * cfgexpand.c (pass_expand): Convert from a global struct to a
8162 subclass of rtl_opt_pass along with...
8163 (pass_data_expand): ...new pass_data instance and...
8164 (make_pass_expand): ...new function.
8165 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
8166 of rtl_opt_pass along with...
8167 (pass_data_free_cfg): ...new pass_data instance and...
8168 (make_pass_free_cfg): ...new function.
8169 (pass_into_cfg_layout_mode): Convert from a global struct to a
8170 subclass of rtl_opt_pass along with...
8171 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
8172 (make_pass_into_cfg_layout_mode): ...new function.
8173 (pass_outof_cfg_layout_mode): Convert from a global struct to a
8174 subclass of rtl_opt_pass along with...
8175 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
8176 (make_pass_outof_cfg_layout_mode): ...new function.
8177 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
8178 struct to a subclass of gimple_opt_pass along with...
8179 (pass_data_build_cgraph_edges): ...new pass_data instance and...
8180 (make_pass_build_cgraph_edges): ...new function.
8181 (pass_rebuild_cgraph_edges): Convert from a global struct to a
8182 subclass of gimple_opt_pass along with...
8183 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
8184 (make_pass_rebuild_cgraph_edges): ...new function.
8185 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
8186 subclass of gimple_opt_pass along with...
8187 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
8189 (make_pass_remove_cgraph_callee_edges): ...new function.
8190 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
8191 struct to a subclass of rtl_opt_pass along with...
8192 (pass_data_stack_adjustments): ...new pass_data instance and...
8193 (make_pass_stack_adjustments): ...new function.
8194 * combine.c (pass_combine): Convert from a global struct to a subclass
8195 of rtl_opt_pass along with...
8196 (pass_data_combine): ...new pass_data instance and...
8197 (make_pass_combine): ...new function.
8198 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
8199 global struct to a subclass of rtl_opt_pass along with...
8200 (pass_data_compare_elim_after_reload): ...new pass_data instance
8202 (make_pass_compare_elim_after_reload): ...new function.
8203 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
8204 of rtl_opt_pass along with...
8205 (pass_data_rtl_cprop): ...new pass_data instance and...
8206 (make_pass_rtl_cprop): ...new function.
8207 * cse.c (pass_cse): Convert from a global struct to a subclass of
8208 rtl_opt_pass along with...
8209 (pass_data_cse): ...new pass_data instance and...
8210 (make_pass_cse): ...new function.
8211 (pass_cse2): Convert from a global struct to a subclass of
8212 rtl_opt_pass along with...
8213 (pass_data_cse2): ...new pass_data instance and...
8214 (make_pass_cse2): ...new function.
8215 (pass_cse_after_global_opts): Convert from a global struct to a
8216 subclass of rtl_opt_pass along with...
8217 (pass_data_cse_after_global_opts): ...new pass_data instance and...
8218 (make_pass_cse_after_global_opts): ...new function.
8219 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
8220 of rtl_opt_pass along with...
8221 (pass_data_ud_rtl_dce): ...new pass_data instance and...
8222 (make_pass_ud_rtl_dce): ...new function.
8223 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
8224 rtl_opt_pass along with...
8225 (pass_data_fast_rtl_dce): ...new pass_data instance and...
8226 (make_pass_fast_rtl_dce): ...new function.
8227 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
8228 a subclass of rtl_opt_pass along with...
8229 (pass_data_df_initialize_opt): ...new pass_data instance and...
8230 (make_pass_df_initialize_opt): ...new function.
8231 (pass_df_initialize_no_opt): Convert from a global struct to a
8232 subclass of rtl_opt_pass along with...
8233 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
8234 (make_pass_df_initialize_no_opt): ...new function.
8235 (pass_df_finish): Convert from a global struct to a subclass of
8236 rtl_opt_pass along with...
8237 (pass_data_df_finish): ...new pass_data instance and...
8238 (make_pass_df_finish): ...new function.
8239 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
8240 rtl_opt_pass along with...
8241 (pass_data_rtl_dse1): ...new pass_data instance and...
8242 (make_pass_rtl_dse1): ...new function.
8243 (pass_rtl_dse2): Convert from a global struct to a subclass of
8244 rtl_opt_pass along with...
8245 (pass_data_rtl_dse2): ...new pass_data instance and...
8246 (make_pass_rtl_dse2): ...new function.
8247 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
8248 subclass of rtl_opt_pass along with...
8249 (pass_data_dwarf2_frame): ...new pass_data instance and...
8250 (make_pass_dwarf2_frame): ...new function.
8251 * except.c (pass_set_nothrow_function_flags): Convert from a global
8252 struct to a subclass of rtl_opt_pass along with...
8253 (pass_data_set_nothrow_function_flags): ...new pass_data instance
8255 (make_pass_set_nothrow_function_flags): ...new function.
8256 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
8257 subclass of rtl_opt_pass along with...
8258 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
8260 (make_pass_convert_to_eh_region_ranges): ...new function.
8261 * final.c (pass_compute_alignments): Convert from a global struct to a
8262 subclass of rtl_opt_pass along with...
8263 (pass_data_compute_alignments): ...new pass_data instance and...
8264 (make_pass_compute_alignments): ...new function.
8265 (pass_final): Convert from a global struct to a subclass of
8266 rtl_opt_pass along with...
8267 (pass_data_final): ...new pass_data instance and...
8268 (make_pass_final): ...new function.
8269 (pass_shorten_branches): Convert from a global struct to a subclass of
8270 rtl_opt_pass along with...
8271 (pass_data_shorten_branches): ...new pass_data instance and...
8272 (make_pass_shorten_branches): ...new function.
8273 (pass_clean_state): Convert from a global struct to a subclass of
8274 rtl_opt_pass along with...
8275 (pass_data_clean_state): ...new pass_data instance and...
8276 (make_pass_clean_state): ...new function.
8277 * function.c (pass_instantiate_virtual_regs): Convert from a global
8278 struct to a subclass of rtl_opt_pass along with...
8279 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
8280 (make_pass_instantiate_virtual_regs): ...new function.
8281 (pass_leaf_regs): Convert from a global struct to a subclass of
8282 rtl_opt_pass along with...
8283 (pass_data_leaf_regs): ...new pass_data instance and...
8284 (make_pass_leaf_regs): ...new function.
8285 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
8286 subclass of rtl_opt_pass along with...
8287 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
8289 (make_pass_thread_prologue_and_epilogue): ...new function.
8290 (pass_match_asm_constraints): Convert from a global struct to a
8291 subclass of rtl_opt_pass along with...
8292 (pass_data_match_asm_constraints): ...new pass_data instance and...
8293 (make_pass_match_asm_constraints): ...new function.
8294 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
8295 subclass of rtl_opt_pass along with...
8296 (pass_data_rtl_fwprop): ...new pass_data instance and...
8297 (make_pass_rtl_fwprop): ...new function.
8298 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
8299 rtl_opt_pass along with...
8300 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
8301 (make_pass_rtl_fwprop_addr): ...new function.
8302 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
8303 rtl_opt_pass along with...
8304 (pass_data_rtl_pre): ...new pass_data instance and...
8305 (make_pass_rtl_pre): ...new function.
8306 (pass_rtl_hoist): Convert from a global struct to a subclass of
8307 rtl_opt_pass along with...
8308 (pass_data_rtl_hoist): ...new pass_data instance and...
8309 (make_pass_rtl_hoist): ...new function.
8310 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
8311 subclass of gimple_opt_pass along with...
8312 (pass_data_lower_cf): ...new pass_data instance and...
8313 (make_pass_lower_cf): ...new function.
8314 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
8315 from a global struct to a subclass of gimple_opt_pass along with...
8316 (pass_data_strength_reduction): ...new pass_data instance and...
8317 (make_pass_strength_reduction): ...new function.
8318 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
8319 of rtl_opt_pass along with...
8320 (pass_data_rtl_ifcvt): ...new pass_data instance and...
8321 (make_pass_rtl_ifcvt): ...new function.
8322 (pass_if_after_combine): Convert from a global struct to a subclass of
8323 rtl_opt_pass along with...
8324 (pass_data_if_after_combine): ...new pass_data instance and...
8325 (make_pass_if_after_combine): ...new function.
8326 (pass_if_after_reload): Convert from a global struct to a subclass of
8327 rtl_opt_pass along with...
8328 (pass_data_if_after_reload): ...new pass_data instance and...
8329 (make_pass_if_after_reload): ...new function.
8330 * init-regs.c (pass_initialize_regs): Convert from a global struct to
8331 a subclass of rtl_opt_pass along with...
8332 (pass_data_initialize_regs): ...new pass_data instance and...
8333 (make_pass_initialize_regs): ...new function.
8334 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
8335 of ipa_opt_pass_d along with...
8336 (pass_data_ipa_cp): ...new pass_data instance and...
8337 (make_pass_ipa_cp): ...new function.
8338 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
8339 global struct to a subclass of gimple_opt_pass along with...
8340 (pass_data_inline_parameters): ...new pass_data instance and...
8341 (make_pass_inline_parameters): ...new function.
8342 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
8343 subclass of gimple_opt_pass along with...
8344 (pass_data_early_inline): ...new pass_data instance and...
8345 (make_pass_early_inline): ...new function.
8346 (pass_ipa_inline): Convert from a global struct to a subclass of
8347 ipa_opt_pass_d along with...
8348 (pass_data_ipa_inline): ...new pass_data instance and...
8349 (make_pass_ipa_inline): ...new function.
8350 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
8351 struct to a subclass of gimple_opt_pass along with...
8352 (pass_data_local_pure_const): ...new pass_data instance and...
8353 (make_pass_local_pure_const): ...new function.
8354 (pass_ipa_pure_const): Convert from a global struct to a subclass of
8355 ipa_opt_pass_d along with...
8356 (pass_data_ipa_pure_const): ...new pass_data instance and...
8357 (make_pass_ipa_pure_const): ...new function.
8358 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
8359 to a subclass of ipa_opt_pass_d along with...
8360 (pass_data_ipa_reference): ...new pass_data instance and...
8361 (make_pass_ipa_reference): ...new function.
8362 * ipa-split.c (pass_split_functions): Convert from a global struct to
8363 a subclass of gimple_opt_pass along with...
8364 (pass_data_split_functions): ...new pass_data instance and...
8365 (make_pass_split_functions): ...new function.
8366 (pass_feedback_split_functions): Convert from a global struct to a
8367 subclass of gimple_opt_pass along with...
8368 (pass_data_feedback_split_functions): ...new pass_data instance and...
8369 (make_pass_feedback_split_functions): ...new function.
8370 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
8371 global struct to a subclass of simple_ipa_opt_pass along with...
8372 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
8374 (make_pass_ipa_function_and_variable_visibility): ...new function.
8375 (pass_ipa_free_inline_summary): Convert from a global struct to a
8376 subclass of simple_ipa_opt_pass along with...
8377 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
8378 (make_pass_ipa_free_inline_summary): ...new function.
8379 (pass_ipa_whole_program_visibility): Convert from a global struct to a
8380 subclass of ipa_opt_pass_d along with...
8381 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
8383 (make_pass_ipa_whole_program_visibility): ...new function.
8384 (pass_ipa_profile): Convert from a global struct to a subclass of
8385 ipa_opt_pass_d along with...
8386 (pass_data_ipa_profile): ...new pass_data instance and...
8387 (make_pass_ipa_profile): ...new function.
8388 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
8389 ipa_opt_pass_d along with...
8390 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
8391 (make_pass_ipa_cdtor_merge): ...new function.
8392 * ira.c (pass_ira): Convert from a global struct to a subclass of
8393 rtl_opt_pass along with...
8394 (pass_data_ira): ...new pass_data instance and...
8395 (make_pass_ira): ...new function.
8396 (pass_reload): Convert from a global struct to a subclass of
8397 rtl_opt_pass along with...
8398 (pass_data_reload): ...new pass_data instance and...
8399 (make_pass_reload): ...new function.
8400 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
8401 subclass of rtl_opt_pass along with...
8402 (pass_data_cleanup_barriers): ...new pass_data instance and...
8403 (make_pass_cleanup_barriers): ...new function.
8404 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
8405 of rtl_opt_pass along with...
8406 (pass_data_loop2): ...new pass_data instance and...
8407 (make_pass_loop2): ...new function.
8408 (pass_rtl_loop_init): Convert from a global struct to a subclass of
8409 rtl_opt_pass along with...
8410 (pass_data_rtl_loop_init): ...new pass_data instance and...
8411 (make_pass_rtl_loop_init): ...new function.
8412 (pass_rtl_loop_done): Convert from a global struct to a subclass of
8413 rtl_opt_pass along with...
8414 (pass_data_rtl_loop_done): ...new pass_data instance and...
8415 (make_pass_rtl_loop_done): ...new function.
8416 (pass_rtl_move_loop_invariants): Convert from a global struct to a
8417 subclass of rtl_opt_pass along with...
8418 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
8419 (make_pass_rtl_move_loop_invariants): ...new function.
8420 (pass_rtl_unswitch): Convert from a global struct to a subclass of
8421 rtl_opt_pass along with...
8422 (pass_data_rtl_unswitch): ...new pass_data instance and...
8423 (make_pass_rtl_unswitch): ...new function.
8424 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
8425 subclass of rtl_opt_pass along with...
8426 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
8428 (make_pass_rtl_unroll_and_peel_loops): ...new function.
8429 (pass_rtl_doloop): Convert from a global struct to a subclass of
8430 rtl_opt_pass along with...
8431 (pass_data_rtl_doloop): ...new pass_data instance and...
8432 (make_pass_rtl_doloop): ...new function.
8433 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
8434 a subclass of rtl_opt_pass along with...
8435 (pass_data_lower_subreg): ...new pass_data instance and...
8436 (make_pass_lower_subreg): ...new function.
8437 (pass_lower_subreg2): Convert from a global struct to a subclass of
8438 rtl_opt_pass along with...
8439 (pass_data_lower_subreg2): ...new pass_data instance and...
8440 (make_pass_lower_subreg2): ...new function.
8441 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
8442 struct to a subclass of ipa_opt_pass_d along with...
8443 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
8444 (make_pass_ipa_lto_gimple_out): ...new function.
8445 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
8446 of ipa_opt_pass_d along with...
8447 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
8448 (make_pass_ipa_lto_finish_out): ...new function.
8449 * mode-switching.c (pass_mode_switching): Convert from a global struct
8450 to a subclass of rtl_opt_pass along with...
8451 (pass_data_mode_switching): ...new pass_data instance and...
8452 (make_pass_mode_switching): ...new function.
8453 * modulo-sched.c (pass_sms): Convert from a global struct to a
8454 subclass of rtl_opt_pass along with...
8455 (pass_data_sms): ...new pass_data instance and...
8456 (make_pass_sms): ...new function.
8457 * omp-low.c (pass_expand_omp): Convert from a global struct to a
8458 subclass of gimple_opt_pass along with...
8459 (pass_data_expand_omp): ...new pass_data instance and...
8460 (make_pass_expand_omp): ...new function.
8461 (pass_lower_omp): Convert from a global struct to a subclass of
8462 gimple_opt_pass along with...
8463 (pass_data_lower_omp): ...new pass_data instance and...
8464 (make_pass_lower_omp): ...new function.
8465 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
8466 of gimple_opt_pass along with...
8467 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
8468 (make_pass_diagnose_omp_blocks): ...new function.
8469 * passes.c (pass_early_local_passes): Convert from a global struct to
8470 a subclass of simple_ipa_opt_pass along with...
8471 (pass_data_early_local_passes): ...new pass_data instance and...
8472 (make_pass_early_local_passes): ...new function.
8473 (pass_all_early_optimizations): Convert from a global struct to a
8474 subclass of gimple_opt_pass along with...
8475 (pass_data_all_early_optimizations): ...new pass_data instance and...
8476 (make_pass_all_early_optimizations): ...new function.
8477 (pass_all_optimizations): Convert from a global struct to a subclass
8478 of gimple_opt_pass along with...
8479 (pass_data_all_optimizations): ...new pass_data instance and...
8480 (make_pass_all_optimizations): ...new function.
8481 (pass_all_optimizations_g): Convert from a global struct to a subclass
8482 of gimple_opt_pass along with...
8483 (pass_data_all_optimizations_g): ...new pass_data instance and...
8484 (make_pass_all_optimizations_g): ...new function.
8485 (pass_rest_of_compilation): Convert from a global struct to a subclass
8486 of rtl_opt_pass along with...
8487 (pass_data_rest_of_compilation): ...new pass_data instance and...
8488 (make_pass_rest_of_compilation): ...new function.
8489 (pass_postreload): Convert from a global struct to a subclass of
8490 rtl_opt_pass along with...
8491 (pass_data_postreload): ...new pass_data instance and...
8492 (make_pass_postreload): ...new function.
8493 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
8494 subclass of rtl_opt_pass along with...
8495 (pass_data_gcse2): ...new pass_data instance and...
8496 (make_pass_gcse2): ...new function.
8497 * postreload.c (pass_postreload_cse): Convert from a global struct to
8498 a subclass of rtl_opt_pass along with...
8499 (pass_data_postreload_cse): ...new pass_data instance and...
8500 (make_pass_postreload_cse): ...new function.
8501 * predict.c (pass_profile): Convert from a global struct to a subclass
8502 of gimple_opt_pass along with...
8503 (pass_data_profile): ...new pass_data instance and...
8504 (make_pass_profile): ...new function.
8505 (pass_strip_predict_hints): Convert from a global struct to a subclass
8506 of gimple_opt_pass along with...
8507 (pass_data_strip_predict_hints): ...new pass_data instance and...
8508 (make_pass_strip_predict_hints): ...new function.
8509 * recog.c (pass_peephole2): Convert from a global struct to a subclass
8510 of rtl_opt_pass along with...
8511 (pass_data_peephole2): ...new pass_data instance and...
8512 (make_pass_peephole2): ...new function.
8513 (pass_split_all_insns): Convert from a global struct to a subclass of
8514 rtl_opt_pass along with...
8515 (pass_data_split_all_insns): ...new pass_data instance and...
8516 (make_pass_split_all_insns): ...new function.
8517 (pass_split_after_reload): Convert from a global struct to a subclass
8518 of rtl_opt_pass along with...
8519 (pass_data_split_after_reload): ...new pass_data instance and...
8520 (make_pass_split_after_reload): ...new function.
8521 (pass_split_before_regstack): Convert from a global struct to a
8522 subclass of rtl_opt_pass along with...
8523 (pass_data_split_before_regstack): ...new pass_data instance and...
8524 (make_pass_split_before_regstack): ...new function.
8525 (pass_split_before_sched2): Convert from a global struct to a subclass
8526 of rtl_opt_pass along with...
8527 (pass_data_split_before_sched2): ...new pass_data instance and...
8528 (make_pass_split_before_sched2): ...new function.
8529 (pass_split_for_shorten_branches): Convert from a global struct to a
8530 subclass of rtl_opt_pass along with...
8531 (pass_data_split_for_shorten_branches): ...new pass_data instance
8533 (make_pass_split_for_shorten_branches): ...new function.
8534 * ree.c (pass_ree): Convert from a global struct to a subclass of
8535 rtl_opt_pass along with...
8536 (pass_data_ree): ...new pass_data instance and...
8537 (make_pass_ree): ...new function.
8538 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
8539 subclass of rtl_opt_pass along with...
8540 (pass_data_stack_regs): ...new pass_data instance and...
8541 (make_pass_stack_regs): ...new function.
8542 (pass_stack_regs_run): Convert from a global struct to a subclass of
8543 rtl_opt_pass along with...
8544 (pass_data_stack_regs_run): ...new pass_data instance and...
8545 (make_pass_stack_regs_run): ...new function.
8546 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
8547 subclass of rtl_opt_pass along with...
8548 (pass_data_cprop_hardreg): ...new pass_data instance and...
8549 (make_pass_cprop_hardreg): ...new function.
8550 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
8551 subclass of rtl_opt_pass along with...
8552 (pass_data_reginfo_init): ...new pass_data instance and...
8553 (make_pass_reginfo_init): ...new function.
8554 * regmove.c (pass_regmove): Convert from a global struct to a subclass
8555 of rtl_opt_pass along with...
8556 (pass_data_regmove): ...new pass_data instance and...
8557 (make_pass_regmove): ...new function.
8558 * regrename.c (pass_regrename): Convert from a global struct to a
8559 subclass of rtl_opt_pass along with...
8560 (pass_data_regrename): ...new pass_data instance and...
8561 (make_pass_regrename): ...new function.
8562 * reorg.c (pass_delay_slots): Convert from a global struct to a
8563 subclass of rtl_opt_pass along with...
8564 (pass_data_delay_slots): ...new pass_data instance and...
8565 (make_pass_delay_slots): ...new function.
8566 (pass_machine_reorg): Convert from a global struct to a subclass of
8567 rtl_opt_pass along with...
8568 (pass_data_machine_reorg): ...new pass_data instance and...
8569 (make_pass_machine_reorg): ...new function.
8570 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
8571 of rtl_opt_pass along with...
8572 (pass_data_sched): ...new pass_data instance and...
8573 (make_pass_sched): ...new function.
8574 (pass_sched2): Convert from a global struct to a subclass of
8575 rtl_opt_pass along with...
8576 (pass_data_sched2): ...new pass_data instance and...
8577 (make_pass_sched2): ...new function.
8578 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
8579 to a subclass of rtl_opt_pass along with...
8580 (pass_data_stack_ptr_mod): ...new pass_data instance and...
8581 (make_pass_stack_ptr_mod): ...new function.
8582 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
8583 to a subclass of rtl_opt_pass along with...
8584 (pass_data_rtl_store_motion): ...new pass_data instance and...
8585 (make_pass_rtl_store_motion): ...new function.
8586 * tracer.c (pass_tracer): Convert from a global struct to a subclass
8587 of gimple_opt_pass along with...
8588 (pass_data_tracer): ...new pass_data instance and...
8589 (make_pass_tracer): ...new function.
8590 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
8591 to a subclass of gimple_opt_pass along with...
8592 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
8593 (make_pass_diagnose_tm_blocks): ...new function.
8594 (pass_lower_tm): Convert from a global struct to a subclass of
8595 gimple_opt_pass along with...
8596 (pass_data_lower_tm): ...new pass_data instance and...
8597 (make_pass_lower_tm): ...new function.
8598 (pass_tm_init): Convert from a global struct to a subclass of
8599 gimple_opt_pass along with...
8600 (pass_data_tm_init): ...new pass_data instance and...
8601 (make_pass_tm_init): ...new function.
8602 (pass_tm_mark): Convert from a global struct to a subclass of
8603 gimple_opt_pass along with...
8604 (pass_data_tm_mark): ...new pass_data instance and...
8605 (make_pass_tm_mark): ...new function.
8606 (pass_tm_edges): Convert from a global struct to a subclass of
8607 gimple_opt_pass along with...
8608 (pass_data_tm_edges): ...new pass_data instance and...
8609 (make_pass_tm_edges): ...new function.
8610 (pass_tm_memopt): Convert from a global struct to a subclass of
8611 gimple_opt_pass along with...
8612 (pass_data_tm_memopt): ...new pass_data instance and...
8613 (make_pass_tm_memopt): ...new function.
8614 (pass_ipa_tm): Convert from a global struct to a subclass of
8615 simple_ipa_opt_pass along with...
8616 (pass_data_ipa_tm): ...new pass_data instance and...
8617 (make_pass_ipa_tm): ...new function.
8618 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
8619 subclass of gimple_opt_pass along with...
8620 (pass_data_call_cdce): ...new pass_data instance and...
8621 (make_pass_call_cdce): ...new function.
8622 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
8623 subclass of gimple_opt_pass along with...
8624 (pass_data_build_cfg): ...new pass_data instance and...
8625 (make_pass_build_cfg): ...new function.
8626 (pass_split_crit_edges): Convert from a global struct to a subclass of
8627 gimple_opt_pass along with...
8628 (pass_data_split_crit_edges): ...new pass_data instance and...
8629 (make_pass_split_crit_edges): ...new function.
8630 (pass_warn_function_return): Convert from a global struct to a
8631 subclass of gimple_opt_pass along with...
8632 (pass_data_warn_function_return): ...new pass_data instance and...
8633 (make_pass_warn_function_return): ...new function.
8634 (pass_warn_function_noreturn): Convert from a global struct to a
8635 subclass of gimple_opt_pass along with...
8636 (pass_data_warn_function_noreturn): ...new pass_data instance and...
8637 (make_pass_warn_function_noreturn): ...new function.
8638 (pass_warn_unused_result): Convert from a global struct to a subclass
8639 of gimple_opt_pass along with...
8640 (pass_data_warn_unused_result): ...new pass_data instance and...
8641 (make_pass_warn_unused_result): ...new function.
8642 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
8643 a subclass of gimple_opt_pass along with...
8644 (pass_data_merge_phi): ...new pass_data instance and...
8645 (make_pass_merge_phi): ...new function.
8646 * tree-complex.c (pass_lower_complex): Convert from a global struct to
8647 a subclass of gimple_opt_pass along with...
8648 (pass_data_lower_complex): ...new pass_data instance and...
8649 (make_pass_lower_complex): ...new function.
8650 (pass_lower_complex_O0): Convert from a global struct to a subclass of
8651 gimple_opt_pass along with...
8652 (pass_data_lower_complex_O0): ...new pass_data instance and...
8653 (make_pass_lower_complex_O0): ...new function.
8654 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
8655 subclass of gimple_opt_pass along with...
8656 (pass_data_lower_eh): ...new pass_data instance and...
8657 (make_pass_lower_eh): ...new function.
8658 (pass_refactor_eh): Convert from a global struct to a subclass of
8659 gimple_opt_pass along with...
8660 (pass_data_refactor_eh): ...new pass_data instance and...
8661 (make_pass_refactor_eh): ...new function.
8662 (pass_lower_resx): Convert from a global struct to a subclass of
8663 gimple_opt_pass along with...
8664 (pass_data_lower_resx): ...new pass_data instance and...
8665 (make_pass_lower_resx): ...new function.
8666 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
8667 of gimple_opt_pass along with...
8668 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
8669 (make_pass_lower_eh_dispatch): ...new function.
8670 (pass_cleanup_eh): Convert from a global struct to a subclass of
8671 gimple_opt_pass along with...
8672 (pass_data_cleanup_eh): ...new pass_data instance and...
8673 (make_pass_cleanup_eh): ...new function.
8674 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
8675 to a subclass of simple_ipa_opt_pass along with...
8676 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
8677 (make_pass_ipa_lower_emutls): ...new function.
8678 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
8679 a subclass of gimple_opt_pass along with...
8680 (pass_data_if_conversion): ...new pass_data instance and...
8681 (make_pass_if_conversion): ...new function.
8682 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
8683 subclass of gimple_opt_pass along with...
8684 (pass_data_build_ssa): ...new pass_data instance and...
8685 (make_pass_build_ssa): ...new function.
8686 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
8687 global struct to a subclass of gimple_opt_pass along with...
8688 (pass_data_loop_distribution): ...new pass_data instance and...
8689 (make_pass_loop_distribution): ...new function.
8690 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
8691 subclass of gimple_opt_pass along with...
8692 (pass_data_mudflap_1): ...new pass_data instance and...
8693 (make_pass_mudflap_1): ...new function.
8694 (pass_mudflap_2): Convert from a global struct to a subclass of
8695 gimple_opt_pass along with...
8696 (pass_data_mudflap_2): ...new pass_data instance and...
8697 (make_pass_mudflap_2): ...new function.
8698 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
8699 subclass of gimple_opt_pass along with...
8700 (pass_data_mudflap_1): ...new pass_data instance and...
8701 (make_pass_mudflap_1): ...new function.
8702 (pass_mudflap_2): Convert from a global struct to a subclass of
8703 gimple_opt_pass along with...
8704 (pass_data_mudflap_2): ...new pass_data instance and...
8705 (make_pass_mudflap_2): ...new function.
8706 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
8707 gimple_opt_pass along with...
8708 (pass_data_nrv): ...new pass_data instance and...
8709 (make_pass_nrv): ...new function.
8710 (pass_return_slot): Convert from a global struct to a subclass of
8711 gimple_opt_pass along with...
8712 (pass_data_return_slot): ...new pass_data instance and...
8713 (make_pass_return_slot): ...new function.
8714 * tree-object-size.c (pass_object_sizes): Convert from a global struct
8715 to a subclass of gimple_opt_pass along with...
8716 (pass_data_object_sizes): ...new pass_data instance and...
8717 (make_pass_object_sizes): ...new function.
8718 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
8719 global struct to a subclass of gimple_opt_pass along with...
8720 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
8722 (make_pass_cleanup_cfg_post_optimizing): ...new function.
8723 (pass_fixup_cfg): Convert from a global struct to a subclass of
8724 gimple_opt_pass along with...
8725 (pass_data_fixup_cfg): ...new pass_data instance and...
8726 (make_pass_fixup_cfg): ...new function.
8727 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
8728 (make_pass_mudflap_1): ...new function.
8729 (pass_mudflap_2): Replace declaration with that of...
8730 (make_pass_mudflap_2): ...new function.
8731 (pass_asan): Replace declaration with that of...
8732 (make_pass_asan): ...new function.
8733 (pass_asan_O0): Replace declaration with that of...
8734 (make_pass_asan_O0): ...new function.
8735 (pass_tsan): Replace declaration with that of...
8736 (make_pass_tsan): ...new function.
8737 (pass_tsan_O0): Replace declaration with that of...
8738 (make_pass_tsan_O0): ...new function.
8739 (pass_lower_cf): Replace declaration with that of...
8740 (make_pass_lower_cf): ...new function.
8741 (pass_refactor_eh): Replace declaration with that of...
8742 (make_pass_refactor_eh): ...new function.
8743 (pass_lower_eh): Replace declaration with that of...
8744 (make_pass_lower_eh): ...new function.
8745 (pass_lower_eh_dispatch): Replace declaration with that of...
8746 (make_pass_lower_eh_dispatch): ...new function.
8747 (pass_lower_resx): Replace declaration with that of...
8748 (make_pass_lower_resx): ...new function.
8749 (pass_build_cfg): Replace declaration with that of...
8750 (make_pass_build_cfg): ...new function.
8751 (pass_early_tree_profile): Replace declaration with that of...
8752 (make_pass_early_tree_profile): ...new function.
8753 (pass_cleanup_eh): Replace declaration with that of...
8754 (make_pass_cleanup_eh): ...new function.
8755 (pass_sra): Replace declaration with that of...
8756 (make_pass_sra): ...new function.
8757 (pass_sra_early): Replace declaration with that of...
8758 (make_pass_sra_early): ...new function.
8759 (pass_early_ipa_sra): Replace declaration with that of...
8760 (make_pass_early_ipa_sra): ...new function.
8761 (pass_tail_recursion): Replace declaration with that of...
8762 (make_pass_tail_recursion): ...new function.
8763 (pass_tail_calls): Replace declaration with that of...
8764 (make_pass_tail_calls): ...new function.
8765 (pass_tree_loop): Replace declaration with that of...
8766 (make_pass_tree_loop): ...new function.
8767 (pass_tree_loop_init): Replace declaration with that of...
8768 (make_pass_tree_loop_init): ...new function.
8769 (pass_lim): Replace declaration with that of...
8770 (make_pass_lim): ...new function.
8771 (pass_tree_unswitch): Replace declaration with that of...
8772 (make_pass_tree_unswitch): ...new function.
8773 (pass_predcom): Replace declaration with that of...
8774 (make_pass_predcom): ...new function.
8775 (pass_iv_canon): Replace declaration with that of...
8776 (make_pass_iv_canon): ...new function.
8777 (pass_scev_cprop): Replace declaration with that of...
8778 (make_pass_scev_cprop): ...new function.
8779 (pass_empty_loop): Replace declaration with that of...
8780 (make_pass_empty_loop): ...new function.
8781 (pass_record_bounds): Replace declaration with that of...
8782 (make_pass_record_bounds): ...new function.
8783 (pass_graphite): Replace declaration with that of...
8784 (make_pass_graphite): ...new function.
8785 (pass_graphite_transforms): Replace declaration with that of...
8786 (make_pass_graphite_transforms): ...new function.
8787 (pass_if_conversion): Replace declaration with that of...
8788 (make_pass_if_conversion): ...new function.
8789 (pass_loop_distribution): Replace declaration with that of...
8790 (make_pass_loop_distribution): ...new function.
8791 (pass_vectorize): Replace declaration with that of...
8792 (make_pass_vectorize): ...new function.
8793 (pass_slp_vectorize): Replace declaration with that of...
8794 (make_pass_slp_vectorize): ...new function.
8795 (pass_complete_unroll): Replace declaration with that of...
8796 (make_pass_complete_unroll): ...new function.
8797 (pass_complete_unrolli): Replace declaration with that of...
8798 (make_pass_complete_unrolli): ...new function.
8799 (pass_parallelize_loops): Replace declaration with that of...
8800 (make_pass_parallelize_loops): ...new function.
8801 (pass_loop_prefetch): Replace declaration with that of...
8802 (make_pass_loop_prefetch): ...new function.
8803 (pass_iv_optimize): Replace declaration with that of...
8804 (make_pass_iv_optimize): ...new function.
8805 (pass_tree_loop_done): Replace declaration with that of...
8806 (make_pass_tree_loop_done): ...new function.
8807 (pass_ch): Replace declaration with that of...
8808 (make_pass_ch): ...new function.
8809 (pass_ccp): Replace declaration with that of...
8810 (make_pass_ccp): ...new function.
8811 (pass_phi_only_cprop): Replace declaration with that of...
8812 (make_pass_phi_only_cprop): ...new function.
8813 (pass_build_ssa): Replace declaration with that of...
8814 (make_pass_build_ssa): ...new function.
8815 (pass_build_alias): Replace declaration with that of...
8816 (make_pass_build_alias): ...new function.
8817 (pass_build_ealias): Replace declaration with that of...
8818 (make_pass_build_ealias): ...new function.
8819 (pass_dominator): Replace declaration with that of...
8820 (make_pass_dominator): ...new function.
8821 (pass_dce): Replace declaration with that of...
8822 (make_pass_dce): ...new function.
8823 (pass_dce_loop): Replace declaration with that of...
8824 (make_pass_dce_loop): ...new function.
8825 (pass_cd_dce): Replace declaration with that of...
8826 (make_pass_cd_dce): ...new function.
8827 (pass_call_cdce): Replace declaration with that of...
8828 (make_pass_call_cdce): ...new function.
8829 (pass_merge_phi): Replace declaration with that of...
8830 (make_pass_merge_phi): ...new function.
8831 (pass_split_crit_edges): Replace declaration with that of...
8832 (make_pass_split_crit_edges): ...new function.
8833 (pass_pre): Replace declaration with that of...
8834 (make_pass_pre): ...new function.
8835 (pass_profile): Replace declaration with that of...
8836 (make_pass_profile): ...new function.
8837 (pass_strip_predict_hints): Replace declaration with that of...
8838 (make_pass_strip_predict_hints): ...new function.
8839 (pass_lower_complex_O0): Replace declaration with that of...
8840 (make_pass_lower_complex_O0): ...new function.
8841 (pass_lower_complex): Replace declaration with that of...
8842 (make_pass_lower_complex): ...new function.
8843 (pass_lower_vector): Replace declaration with that of...
8844 (make_pass_lower_vector): ...new function.
8845 (pass_lower_vector_ssa): Replace declaration with that of...
8846 (make_pass_lower_vector_ssa): ...new function.
8847 (pass_lower_omp): Replace declaration with that of...
8848 (make_pass_lower_omp): ...new function.
8849 (pass_diagnose_omp_blocks): Replace declaration with that of...
8850 (make_pass_diagnose_omp_blocks): ...new function.
8851 (pass_expand_omp): Replace declaration with that of...
8852 (make_pass_expand_omp): ...new function.
8853 (pass_expand_omp_ssa): Replace declaration with that of...
8854 (make_pass_expand_omp_ssa): ...new function.
8855 (pass_object_sizes): Replace declaration with that of...
8856 (make_pass_object_sizes): ...new function.
8857 (pass_strlen): Replace declaration with that of...
8858 (make_pass_strlen): ...new function.
8859 (pass_fold_builtins): Replace declaration with that of...
8860 (make_pass_fold_builtins): ...new function.
8861 (pass_stdarg): Replace declaration with that of...
8862 (make_pass_stdarg): ...new function.
8863 (pass_early_warn_uninitialized): Replace declaration with that of...
8864 (make_pass_early_warn_uninitialized): ...new function.
8865 (pass_late_warn_uninitialized): Replace declaration with that of...
8866 (make_pass_late_warn_uninitialized): ...new function.
8867 (pass_cse_reciprocals): Replace declaration with that of...
8868 (make_pass_cse_reciprocals): ...new function.
8869 (pass_cse_sincos): Replace declaration with that of...
8870 (make_pass_cse_sincos): ...new function.
8871 (pass_optimize_bswap): Replace declaration with that of...
8872 (make_pass_optimize_bswap): ...new function.
8873 (pass_optimize_widening_mul): Replace declaration with that of...
8874 (make_pass_optimize_widening_mul): ...new function.
8875 (pass_warn_function_return): Replace declaration with that of...
8876 (make_pass_warn_function_return): ...new function.
8877 (pass_warn_function_noreturn): Replace declaration with that of...
8878 (make_pass_warn_function_noreturn): ...new function.
8879 (pass_cselim): Replace declaration with that of...
8880 (make_pass_cselim): ...new function.
8881 (pass_phiopt): Replace declaration with that of...
8882 (make_pass_phiopt): ...new function.
8883 (pass_forwprop): Replace declaration with that of...
8884 (make_pass_forwprop): ...new function.
8885 (pass_phiprop): Replace declaration with that of...
8886 (make_pass_phiprop): ...new function.
8887 (pass_tree_ifcombine): Replace declaration with that of...
8888 (make_pass_tree_ifcombine): ...new function.
8889 (pass_dse): Replace declaration with that of...
8890 (make_pass_dse): ...new function.
8891 (pass_nrv): Replace declaration with that of...
8892 (make_pass_nrv): ...new function.
8893 (pass_rename_ssa_copies): Replace declaration with that of...
8894 (make_pass_rename_ssa_copies): ...new function.
8895 (pass_sink_code): Replace declaration with that of...
8896 (make_pass_sink_code): ...new function.
8897 (pass_fre): Replace declaration with that of...
8898 (make_pass_fre): ...new function.
8899 (pass_check_data_deps): Replace declaration with that of...
8900 (make_pass_check_data_deps): ...new function.
8901 (pass_copy_prop): Replace declaration with that of...
8902 (make_pass_copy_prop): ...new function.
8903 (pass_vrp): Replace declaration with that of...
8904 (make_pass_vrp): ...new function.
8905 (pass_uncprop): Replace declaration with that of...
8906 (make_pass_uncprop): ...new function.
8907 (pass_return_slot): Replace declaration with that of...
8908 (make_pass_return_slot): ...new function.
8909 (pass_reassoc): Replace declaration with that of...
8910 (make_pass_reassoc): ...new function.
8911 (pass_rebuild_cgraph_edges): Replace declaration with that of...
8912 (make_pass_rebuild_cgraph_edges): ...new function.
8913 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
8914 (make_pass_remove_cgraph_callee_edges): ...new function.
8915 (pass_build_cgraph_edges): Replace declaration with that of...
8916 (make_pass_build_cgraph_edges): ...new function.
8917 (pass_local_pure_const): Replace declaration with that of...
8918 (make_pass_local_pure_const): ...new function.
8919 (pass_tracer): Replace declaration with that of...
8920 (make_pass_tracer): ...new function.
8921 (pass_warn_unused_result): Replace declaration with that of...
8922 (make_pass_warn_unused_result): ...new function.
8923 (pass_diagnose_tm_blocks): Replace declaration with that of...
8924 (make_pass_diagnose_tm_blocks): ...new function.
8925 (pass_lower_tm): Replace declaration with that of...
8926 (make_pass_lower_tm): ...new function.
8927 (pass_tm_init): Replace declaration with that of...
8928 (make_pass_tm_init): ...new function.
8929 (pass_tm_mark): Replace declaration with that of...
8930 (make_pass_tm_mark): ...new function.
8931 (pass_tm_memopt): Replace declaration with that of...
8932 (make_pass_tm_memopt): ...new function.
8933 (pass_tm_edges): Replace declaration with that of...
8934 (make_pass_tm_edges): ...new function.
8935 (pass_split_functions): Replace declaration with that of...
8936 (make_pass_split_functions): ...new function.
8937 (pass_feedback_split_functions): Replace declaration with that of...
8938 (make_pass_feedback_split_functions): ...new function.
8939 (pass_strength_reduction): Replace declaration with that of...
8940 (make_pass_strength_reduction): ...new function.
8941 (pass_ipa_lower_emutls): Replace declaration with that of...
8942 (make_pass_ipa_lower_emutls): ...new function.
8943 (pass_ipa_function_and_variable_visibility): Replace declaration with
8945 (make_pass_ipa_function_and_variable_visibility): ...new function.
8946 (pass_ipa_tree_profile): Replace declaration with that of...
8947 (make_pass_ipa_tree_profile): ...new function.
8948 (pass_early_local_passes): Replace declaration with that of...
8949 (make_pass_early_local_passes): ...new function.
8950 (pass_ipa_whole_program_visibility): Replace declaration with that
8952 (make_pass_ipa_whole_program_visibility): ...new function.
8953 (pass_ipa_lto_gimple_out): Replace declaration with that of...
8954 (make_pass_ipa_lto_gimple_out): ...new function.
8955 (pass_ipa_increase_alignment): Replace declaration with that of...
8956 (make_pass_ipa_increase_alignment): ...new function.
8957 (pass_ipa_inline): Replace declaration with that of...
8958 (make_pass_ipa_inline): ...new function.
8959 (pass_ipa_free_lang_data): Replace declaration with that of...
8960 (make_pass_ipa_free_lang_data): ...new function.
8961 (pass_ipa_free_inline_summary): Replace declaration with that of...
8962 (make_pass_ipa_free_inline_summary): ...new function.
8963 (pass_ipa_cp): Replace declaration with that of...
8964 (make_pass_ipa_cp): ...new function.
8965 (pass_ipa_reference): Replace declaration with that of...
8966 (make_pass_ipa_reference): ...new function.
8967 (pass_ipa_pure_const): Replace declaration with that of...
8968 (make_pass_ipa_pure_const): ...new function.
8969 (pass_ipa_pta): Replace declaration with that of...
8970 (make_pass_ipa_pta): ...new function.
8971 (pass_ipa_lto_finish_out): Replace declaration with that of...
8972 (make_pass_ipa_lto_finish_out): ...new function.
8973 (pass_ipa_tm): Replace declaration with that of...
8974 (make_pass_ipa_tm): ...new function.
8975 (pass_ipa_profile): Replace declaration with that of...
8976 (make_pass_ipa_profile): ...new function.
8977 (pass_ipa_cdtor_merge): Replace declaration with that of...
8978 (make_pass_ipa_cdtor_merge): ...new function.
8979 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
8981 (make_pass_cleanup_cfg_post_optimizing): ...new function.
8982 (pass_init_datastructures): Replace declaration with that of...
8983 (make_pass_init_datastructures): ...new function.
8984 (pass_fixup_cfg): Replace declaration with that of...
8985 (make_pass_fixup_cfg): ...new function.
8986 (pass_expand): Replace declaration with that of...
8987 (make_pass_expand): ...new function.
8988 (pass_instantiate_virtual_regs): Replace declaration with that of...
8989 (make_pass_instantiate_virtual_regs): ...new function.
8990 (pass_rtl_fwprop): Replace declaration with that of...
8991 (make_pass_rtl_fwprop): ...new function.
8992 (pass_rtl_fwprop_addr): Replace declaration with that of...
8993 (make_pass_rtl_fwprop_addr): ...new function.
8994 (pass_jump): Replace declaration with that of...
8995 (make_pass_jump): ...new function.
8996 (pass_jump2): Replace declaration with that of...
8997 (make_pass_jump2): ...new function.
8998 (pass_lower_subreg): Replace declaration with that of...
8999 (make_pass_lower_subreg): ...new function.
9000 (pass_cse): Replace declaration with that of...
9001 (make_pass_cse): ...new function.
9002 (pass_fast_rtl_dce): Replace declaration with that of...
9003 (make_pass_fast_rtl_dce): ...new function.
9004 (pass_ud_rtl_dce): Replace declaration with that of...
9005 (make_pass_ud_rtl_dce): ...new function.
9006 (pass_rtl_dce): Replace declaration with that of...
9007 (make_pass_rtl_dce): ...new function.
9008 (pass_rtl_dse1): Replace declaration with that of...
9009 (make_pass_rtl_dse1): ...new function.
9010 (pass_rtl_dse2): Replace declaration with that of...
9011 (make_pass_rtl_dse2): ...new function.
9012 (pass_rtl_dse3): Replace declaration with that of...
9013 (make_pass_rtl_dse3): ...new function.
9014 (pass_rtl_cprop): Replace declaration with that of...
9015 (make_pass_rtl_cprop): ...new function.
9016 (pass_rtl_pre): Replace declaration with that of...
9017 (make_pass_rtl_pre): ...new function.
9018 (pass_rtl_hoist): Replace declaration with that of...
9019 (make_pass_rtl_hoist): ...new function.
9020 (pass_rtl_store_motion): Replace declaration with that of...
9021 (make_pass_rtl_store_motion): ...new function.
9022 (pass_cse_after_global_opts): Replace declaration with that of...
9023 (make_pass_cse_after_global_opts): ...new function.
9024 (pass_rtl_ifcvt): Replace declaration with that of...
9025 (make_pass_rtl_ifcvt): ...new function.
9026 (pass_into_cfg_layout_mode): Replace declaration with that of...
9027 (make_pass_into_cfg_layout_mode): ...new function.
9028 (pass_outof_cfg_layout_mode): Replace declaration with that of...
9029 (make_pass_outof_cfg_layout_mode): ...new function.
9030 (pass_loop2): Replace declaration with that of...
9031 (make_pass_loop2): ...new function.
9032 (pass_rtl_loop_init): Replace declaration with that of...
9033 (make_pass_rtl_loop_init): ...new function.
9034 (pass_rtl_move_loop_invariants): Replace declaration with that of...
9035 (make_pass_rtl_move_loop_invariants): ...new function.
9036 (pass_rtl_unswitch): Replace declaration with that of...
9037 (make_pass_rtl_unswitch): ...new function.
9038 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
9039 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9040 (pass_rtl_doloop): Replace declaration with that of...
9041 (make_pass_rtl_doloop): ...new function.
9042 (pass_rtl_loop_done): Replace declaration with that of...
9043 (make_pass_rtl_loop_done): ...new function.
9044 (pass_web): Replace declaration with that of...
9045 (make_pass_web): ...new function.
9046 (pass_cse2): Replace declaration with that of...
9047 (make_pass_cse2): ...new function.
9048 (pass_df_initialize_opt): Replace declaration with that of...
9049 (make_pass_df_initialize_opt): ...new function.
9050 (pass_df_initialize_no_opt): Replace declaration with that of...
9051 (make_pass_df_initialize_no_opt): ...new function.
9052 (pass_reginfo_init): Replace declaration with that of...
9053 (make_pass_reginfo_init): ...new function.
9054 (pass_inc_dec): Replace declaration with that of...
9055 (make_pass_inc_dec): ...new function.
9056 (pass_stack_ptr_mod): Replace declaration with that of...
9057 (make_pass_stack_ptr_mod): ...new function.
9058 (pass_initialize_regs): Replace declaration with that of...
9059 (make_pass_initialize_regs): ...new function.
9060 (pass_combine): Replace declaration with that of...
9061 (make_pass_combine): ...new function.
9062 (pass_if_after_combine): Replace declaration with that of...
9063 (make_pass_if_after_combine): ...new function.
9064 (pass_ree): Replace declaration with that of...
9065 (make_pass_ree): ...new function.
9066 (pass_partition_blocks): Replace declaration with that of...
9067 (make_pass_partition_blocks): ...new function.
9068 (pass_match_asm_constraints): Replace declaration with that of...
9069 (make_pass_match_asm_constraints): ...new function.
9070 (pass_regmove): Replace declaration with that of...
9071 (make_pass_regmove): ...new function.
9072 (pass_split_all_insns): Replace declaration with that of...
9073 (make_pass_split_all_insns): ...new function.
9074 (pass_fast_rtl_byte_dce): Replace declaration with that of...
9075 (make_pass_fast_rtl_byte_dce): ...new function.
9076 (pass_lower_subreg2): Replace declaration with that of...
9077 (make_pass_lower_subreg2): ...new function.
9078 (pass_mode_switching): Replace declaration with that of...
9079 (make_pass_mode_switching): ...new function.
9080 (pass_sms): Replace declaration with that of...
9081 (make_pass_sms): ...new function.
9082 (pass_sched): Replace declaration with that of...
9083 (make_pass_sched): ...new function.
9084 (pass_ira): Replace declaration with that of...
9085 (make_pass_ira): ...new function.
9086 (pass_reload): Replace declaration with that of...
9087 (make_pass_reload): ...new function.
9088 (pass_clean_state): Replace declaration with that of...
9089 (make_pass_clean_state): ...new function.
9090 (pass_branch_prob): Replace declaration with that of...
9091 (make_pass_branch_prob): ...new function.
9092 (pass_value_profile_transformations): Replace declaration with that
9094 (make_pass_value_profile_transformations): ...new function.
9095 (pass_postreload_cse): Replace declaration with that of...
9096 (make_pass_postreload_cse): ...new function.
9097 (pass_gcse2): Replace declaration with that of...
9098 (make_pass_gcse2): ...new function.
9099 (pass_split_after_reload): Replace declaration with that of...
9100 (make_pass_split_after_reload): ...new function.
9101 (pass_branch_target_load_optimize1): Replace declaration with that
9103 (make_pass_branch_target_load_optimize1): ...new function.
9104 (pass_thread_prologue_and_epilogue): Replace declaration with that
9106 (make_pass_thread_prologue_and_epilogue): ...new function.
9107 (pass_stack_adjustments): Replace declaration with that of...
9108 (make_pass_stack_adjustments): ...new function.
9109 (pass_peephole2): Replace declaration with that of...
9110 (make_pass_peephole2): ...new function.
9111 (pass_if_after_reload): Replace declaration with that of...
9112 (make_pass_if_after_reload): ...new function.
9113 (pass_regrename): Replace declaration with that of...
9114 (make_pass_regrename): ...new function.
9115 (pass_cprop_hardreg): Replace declaration with that of...
9116 (make_pass_cprop_hardreg): ...new function.
9117 (pass_reorder_blocks): Replace declaration with that of...
9118 (make_pass_reorder_blocks): ...new function.
9119 (pass_branch_target_load_optimize2): Replace declaration with that
9121 (make_pass_branch_target_load_optimize2): ...new function.
9122 (pass_leaf_regs): Replace declaration with that of...
9123 (make_pass_leaf_regs): ...new function.
9124 (pass_split_before_sched2): Replace declaration with that of...
9125 (make_pass_split_before_sched2): ...new function.
9126 (pass_compare_elim_after_reload): Replace declaration with that of...
9127 (make_pass_compare_elim_after_reload): ...new function.
9128 (pass_sched2): Replace declaration with that of...
9129 (make_pass_sched2): ...new function.
9130 (pass_stack_regs): Replace declaration with that of...
9131 (make_pass_stack_regs): ...new function.
9132 (pass_stack_regs_run): Replace declaration with that of...
9133 (make_pass_stack_regs_run): ...new function.
9134 (pass_df_finish): Replace declaration with that of...
9135 (make_pass_df_finish): ...new function.
9136 (pass_compute_alignments): Replace declaration with that of...
9137 (make_pass_compute_alignments): ...new function.
9138 (pass_duplicate_computed_gotos): Replace declaration with that of...
9139 (make_pass_duplicate_computed_gotos): ...new function.
9140 (pass_variable_tracking): Replace declaration with that of...
9141 (make_pass_variable_tracking): ...new function.
9142 (pass_free_cfg): Replace declaration with that of...
9143 (make_pass_free_cfg): ...new function.
9144 (pass_machine_reorg): Replace declaration with that of...
9145 (make_pass_machine_reorg): ...new function.
9146 (pass_cleanup_barriers): Replace declaration with that of...
9147 (make_pass_cleanup_barriers): ...new function.
9148 (pass_delay_slots): Replace declaration with that of...
9149 (make_pass_delay_slots): ...new function.
9150 (pass_split_for_shorten_branches): Replace declaration with that of...
9151 (make_pass_split_for_shorten_branches): ...new function.
9152 (pass_split_before_regstack): Replace declaration with that of...
9153 (make_pass_split_before_regstack): ...new function.
9154 (pass_convert_to_eh_region_ranges): Replace declaration with that
9156 (make_pass_convert_to_eh_region_ranges): ...new function.
9157 (pass_shorten_branches): Replace declaration with that of...
9158 (make_pass_shorten_branches): ...new function.
9159 (pass_set_nothrow_function_flags): Replace declaration with that of...
9160 (make_pass_set_nothrow_function_flags): ...new function.
9161 (pass_dwarf2_frame): Replace declaration with that of...
9162 (make_pass_dwarf2_frame): ...new function.
9163 (pass_final): Replace declaration with that of...
9164 (make_pass_final): ...new function.
9165 (pass_rtl_seqabstr): Replace declaration with that of...
9166 (make_pass_rtl_seqabstr): ...new function.
9167 (pass_release_ssa_names): Replace declaration with that of...
9168 (make_pass_release_ssa_names): ...new function.
9169 (pass_early_inline): Replace declaration with that of...
9170 (make_pass_early_inline): ...new function.
9171 (pass_inline_parameters): Replace declaration with that of...
9172 (make_pass_inline_parameters): ...new function.
9173 (pass_update_address_taken): Replace declaration with that of...
9174 (make_pass_update_address_taken): ...new function.
9175 (pass_convert_switch): Replace declaration with that of...
9176 (make_pass_convert_switch): ...new function.
9177 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
9178 to a subclass of simple_ipa_opt_pass along with...
9179 (pass_data_ipa_tree_profile): ...new pass_data instance and...
9180 (make_pass_ipa_tree_profile): ...new function.
9181 * tree-sra.c (pass_sra_early): Convert from a global struct to a
9182 subclass of gimple_opt_pass along with...
9183 (pass_data_sra_early): ...new pass_data instance and...
9184 (make_pass_sra_early): ...new function.
9185 (pass_sra): Convert from a global struct to a subclass of
9186 gimple_opt_pass along with...
9187 (pass_data_sra): ...new pass_data instance and...
9188 (make_pass_sra): ...new function.
9189 (pass_early_ipa_sra): Convert from a global struct to a subclass of
9190 gimple_opt_pass along with...
9191 (pass_data_early_ipa_sra): ...new pass_data instance and...
9192 (make_pass_early_ipa_sra): ...new function.
9193 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
9194 subclass of gimple_opt_pass along with...
9195 (pass_data_ccp): ...new pass_data instance and...
9196 (make_pass_ccp): ...new function.
9197 (pass_fold_builtins): Convert from a global struct to a subclass of
9198 gimple_opt_pass along with...
9199 (pass_data_fold_builtins): ...new pass_data instance and...
9200 (make_pass_fold_builtins): ...new function.
9201 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
9202 subclass of gimple_opt_pass along with...
9203 (pass_data_copy_prop): ...new pass_data instance and...
9204 (make_pass_copy_prop): ...new function.
9205 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
9206 global struct to a subclass of gimple_opt_pass along with...
9207 (pass_data_rename_ssa_copies): ...new pass_data instance and...
9208 (make_pass_rename_ssa_copies): ...new function.
9209 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
9210 subclass of gimple_opt_pass along with...
9211 (pass_data_dce): ...new pass_data instance and...
9212 (make_pass_dce): ...new function.
9213 (pass_dce_loop): Convert from a global struct to a subclass of
9214 gimple_opt_pass along with...
9215 (pass_data_dce_loop): ...new pass_data instance and...
9216 (make_pass_dce_loop): ...new function.
9217 (pass_cd_dce): Convert from a global struct to a subclass of
9218 gimple_opt_pass along with...
9219 (pass_data_cd_dce): ...new pass_data instance and...
9220 (make_pass_cd_dce): ...new function.
9221 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
9222 subclass of gimple_opt_pass along with...
9223 (pass_data_dominator): ...new pass_data instance and...
9224 (make_pass_dominator): ...new function.
9225 (pass_phi_only_cprop): Convert from a global struct to a subclass of
9226 gimple_opt_pass along with...
9227 (pass_data_phi_only_cprop): ...new pass_data instance and...
9228 (make_pass_phi_only_cprop): ...new function.
9229 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
9230 subclass of gimple_opt_pass along with...
9231 (pass_data_dse): ...new pass_data instance and...
9232 (make_pass_dse): ...new function.
9233 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
9234 a subclass of gimple_opt_pass along with...
9235 (pass_data_forwprop): ...new pass_data instance and...
9236 (make_pass_forwprop): ...new function.
9237 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
9238 struct to a subclass of gimple_opt_pass along with...
9239 (pass_data_tree_ifcombine): ...new pass_data instance and...
9240 (make_pass_tree_ifcombine): ...new function.
9241 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
9242 subclass of gimple_opt_pass along with...
9243 (pass_data_ch): ...new pass_data instance and...
9244 (make_pass_ch): ...new function.
9245 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
9246 subclass of gimple_opt_pass along with...
9247 (pass_data_tree_loop): ...new pass_data instance and...
9248 (make_pass_tree_loop): ...new function.
9249 (pass_tree_loop_init): Convert from a global struct to a subclass of
9250 gimple_opt_pass along with...
9251 (pass_data_tree_loop_init): ...new pass_data instance and...
9252 (make_pass_tree_loop_init): ...new function.
9253 (pass_lim): Convert from a global struct to a subclass of
9254 gimple_opt_pass along with...
9255 (pass_data_lim): ...new pass_data instance and...
9256 (make_pass_lim): ...new function.
9257 (pass_tree_unswitch): Convert from a global struct to a subclass of
9258 gimple_opt_pass along with...
9259 (pass_data_tree_unswitch): ...new pass_data instance and...
9260 (make_pass_tree_unswitch): ...new function.
9261 (pass_predcom): Convert from a global struct to a subclass of
9262 gimple_opt_pass along with...
9263 (pass_data_predcom): ...new pass_data instance and...
9264 (make_pass_predcom): ...new function.
9265 (pass_vectorize): Convert from a global struct to a subclass of
9266 gimple_opt_pass along with...
9267 (pass_data_vectorize): ...new pass_data instance and...
9268 (make_pass_vectorize): ...new function.
9269 (pass_graphite): Convert from a global struct to a subclass of
9270 gimple_opt_pass along with...
9271 (pass_data_graphite): ...new pass_data instance and...
9272 (make_pass_graphite): ...new function.
9273 (pass_graphite_transforms): Convert from a global struct to a subclass
9274 of gimple_opt_pass along with...
9275 (pass_data_graphite_transforms): ...new pass_data instance and...
9276 (make_pass_graphite_transforms): ...new function.
9277 (pass_check_data_deps): Convert from a global struct to a subclass of
9278 gimple_opt_pass along with...
9279 (pass_data_check_data_deps): ...new pass_data instance and...
9280 (make_pass_check_data_deps): ...new function.
9281 (pass_iv_canon): Convert from a global struct to a subclass of
9282 gimple_opt_pass along with...
9283 (pass_data_iv_canon): ...new pass_data instance and...
9284 (make_pass_iv_canon): ...new function.
9285 (pass_scev_cprop): Convert from a global struct to a subclass of
9286 gimple_opt_pass along with...
9287 (pass_data_scev_cprop): ...new pass_data instance and...
9288 (make_pass_scev_cprop): ...new function.
9289 (pass_record_bounds): Convert from a global struct to a subclass of
9290 gimple_opt_pass along with...
9291 (pass_data_record_bounds): ...new pass_data instance and...
9292 (make_pass_record_bounds): ...new function.
9293 (pass_complete_unroll): Convert from a global struct to a subclass of
9294 gimple_opt_pass along with...
9295 (pass_data_complete_unroll): ...new pass_data instance and...
9296 (make_pass_complete_unroll): ...new function.
9297 (pass_complete_unrolli): Convert from a global struct to a subclass of
9298 gimple_opt_pass along with...
9299 (pass_data_complete_unrolli): ...new pass_data instance and...
9300 (make_pass_complete_unrolli): ...new function.
9301 (pass_parallelize_loops): Convert from a global struct to a subclass
9302 of gimple_opt_pass along with...
9303 (pass_data_parallelize_loops): ...new pass_data instance and...
9304 (make_pass_parallelize_loops): ...new function.
9305 (pass_loop_prefetch): Convert from a global struct to a subclass of
9306 gimple_opt_pass along with...
9307 (pass_data_loop_prefetch): ...new pass_data instance and...
9308 (make_pass_loop_prefetch): ...new function.
9309 (pass_iv_optimize): Convert from a global struct to a subclass of
9310 gimple_opt_pass along with...
9311 (pass_data_iv_optimize): ...new pass_data instance and...
9312 (make_pass_iv_optimize): ...new function.
9313 (pass_tree_loop_done): Convert from a global struct to a subclass of
9314 gimple_opt_pass along with...
9315 (pass_data_tree_loop_done): ...new pass_data instance and...
9316 (make_pass_tree_loop_done): ...new function.
9317 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
9318 struct to a subclass of gimple_opt_pass along with...
9319 (pass_data_cse_reciprocals): ...new pass_data instance and...
9320 (make_pass_cse_reciprocals): ...new function.
9321 (pass_cse_sincos): Convert from a global struct to a subclass of
9322 gimple_opt_pass along with...
9323 (pass_data_cse_sincos): ...new pass_data instance and...
9324 (make_pass_cse_sincos): ...new function.
9325 (pass_optimize_bswap): Convert from a global struct to a subclass of
9326 gimple_opt_pass along with...
9327 (pass_data_optimize_bswap): ...new pass_data instance and...
9328 (make_pass_optimize_bswap): ...new function.
9329 (pass_optimize_widening_mul): Convert from a global struct to a
9330 subclass of gimple_opt_pass along with...
9331 (pass_data_optimize_widening_mul): ...new pass_data instance and...
9332 (make_pass_optimize_widening_mul): ...new function.
9333 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
9334 subclass of gimple_opt_pass along with...
9335 (pass_data_phiopt): ...new pass_data instance and...
9336 (make_pass_phiopt): ...new function.
9337 (pass_cselim): Convert from a global struct to a subclass of
9338 gimple_opt_pass along with...
9339 (pass_data_cselim): ...new pass_data instance and...
9340 (make_pass_cselim): ...new function.
9341 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
9342 subclass of gimple_opt_pass along with...
9343 (pass_data_phiprop): ...new pass_data instance and...
9344 (make_pass_phiprop): ...new function.
9345 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
9346 subclass of gimple_opt_pass along with...
9347 (pass_data_pre): ...new pass_data instance and...
9348 (make_pass_pre): ...new function.
9349 (pass_fre): Convert from a global struct to a subclass of
9350 gimple_opt_pass along with...
9351 (pass_data_fre): ...new pass_data instance and...
9352 (make_pass_fre): ...new function.
9353 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
9354 subclass of gimple_opt_pass along with...
9355 (pass_data_reassoc): ...new pass_data instance and...
9356 (make_pass_reassoc): ...new function.
9357 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
9358 subclass of gimple_opt_pass along with...
9359 (pass_data_sink_code): ...new pass_data instance and...
9360 (make_pass_sink_code): ...new function.
9361 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
9362 subclass of gimple_opt_pass along with...
9363 (pass_data_strlen): ...new pass_data instance and...
9364 (make_pass_strlen): ...new function.
9365 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
9366 struct to a subclass of gimple_opt_pass along with...
9367 (pass_data_build_alias): ...new pass_data instance and...
9368 (make_pass_build_alias): ...new function.
9369 (pass_build_ealias): Convert from a global struct to a subclass of
9370 gimple_opt_pass along with...
9371 (pass_data_build_ealias): ...new pass_data instance and...
9372 (make_pass_build_ealias): ...new function.
9373 (pass_ipa_pta): Convert from a global struct to a subclass of
9374 simple_ipa_opt_pass along with...
9375 (pass_data_ipa_pta): ...new pass_data instance and...
9376 (make_pass_ipa_pta): ...new function.
9377 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
9378 subclass of gimple_opt_pass along with...
9379 (pass_data_uncprop): ...new pass_data instance and...
9380 (make_pass_uncprop): ...new function.
9381 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
9382 global struct to a subclass of gimple_opt_pass along with...
9383 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
9384 (make_pass_late_warn_uninitialized): ...new function.
9385 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
9386 to a subclass of gimple_opt_pass along with...
9387 (pass_data_init_datastructures): ...new pass_data instance and...
9388 (make_pass_init_datastructures): ...new function.
9389 (pass_early_warn_uninitialized): Convert from a global struct to a
9390 subclass of gimple_opt_pass along with...
9391 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
9392 (make_pass_early_warn_uninitialized): ...new function.
9393 (pass_update_address_taken): Convert from a global struct to a
9394 subclass of gimple_opt_pass along with...
9395 (pass_data_update_address_taken): ...new pass_data instance and...
9396 (make_pass_update_address_taken): ...new function.
9397 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
9398 struct to a subclass of gimple_opt_pass along with...
9399 (pass_data_release_ssa_names): ...new pass_data instance and...
9400 (make_pass_release_ssa_names): ...new function.
9401 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
9402 subclass of gimple_opt_pass along with...
9403 (pass_data_stdarg): ...new pass_data instance and...
9404 (make_pass_stdarg): ...new function.
9405 * tree-switch-conversion.c (pass_convert_switch): Convert from a
9406 global struct to a subclass of gimple_opt_pass along with...
9407 (pass_data_convert_switch): ...new pass_data instance and...
9408 (make_pass_convert_switch): ...new function.
9409 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
9410 to a subclass of gimple_opt_pass along with...
9411 (pass_data_tail_recursion): ...new pass_data instance and...
9412 (make_pass_tail_recursion): ...new function.
9413 (pass_tail_calls): Convert from a global struct to a subclass of
9414 gimple_opt_pass along with...
9415 (pass_data_tail_calls): ...new pass_data instance and...
9416 (make_pass_tail_calls): ...new function.
9417 * tree-vect-generic.c (pass_lower_vector): Convert from a global
9418 struct to a subclass of gimple_opt_pass along with...
9419 (pass_data_lower_vector): ...new pass_data instance and...
9420 (make_pass_lower_vector): ...new function.
9421 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
9422 gimple_opt_pass along with...
9423 (pass_data_lower_vector_ssa): ...new pass_data instance and...
9424 (make_pass_lower_vector_ssa): ...new function.
9425 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
9426 to a subclass of gimple_opt_pass along with...
9427 (pass_data_slp_vectorize): ...new pass_data instance and...
9428 (make_pass_slp_vectorize): ...new function.
9429 (pass_ipa_increase_alignment): Convert from a global struct to a
9430 subclass of simple_ipa_opt_pass along with...
9431 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
9432 (make_pass_ipa_increase_alignment): ...new function.
9433 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
9434 gimple_opt_pass along with...
9435 (pass_data_vrp): ...new pass_data instance and...
9436 (make_pass_vrp): ...new function.
9437 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
9438 subclass of simple_ipa_opt_pass along with...
9439 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
9440 (make_pass_ipa_free_lang_data): ...new function.
9441 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
9442 gimple_opt_pass along with...
9443 (pass_data_tsan): ...new pass_data instance and...
9444 (make_pass_tsan): ...new function.
9445 (pass_tsan_O0): Convert from a global struct to a subclass of
9446 gimple_opt_pass along with...
9447 (pass_data_tsan_O0): ...new pass_data instance and...
9448 (make_pass_tsan_O0): ...new function.
9449 * var-tracking.c (pass_variable_tracking): Convert from a global
9450 struct to a subclass of rtl_opt_pass along with...
9451 (pass_data_variable_tracking): ...new pass_data instance and...
9452 (make_pass_variable_tracking): ...new function.
9453 * web.c (pass_web): Convert from a global struct to a subclass of
9454 rtl_opt_pass along with...
9455 (pass_data_web): ...new pass_data instance and...
9456 (make_pass_web): ...new function.
9457 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
9458 declaration with that of...
9459 (make_pass_mode_switch_use): ...new function.
9460 (pass_resolve_sw_modes): Replace declaration with that of...
9461 (make_pass_resolve_sw_modes): ...new function.
9462 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
9463 from a global struct to a subclass of rtl_opt_pass along with...
9464 (pass_data_mode_switch_use): ...new pass_data instance and...
9465 (make_pass_mode_switch_use): ...new function.
9466 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
9467 from a global struct to a subclass of rtl_opt_pass along with...
9468 (pass_data_resolve_sw_modes): ...new pass_data instance and...
9469 (make_pass_resolve_sw_modes): ...new function.
9470 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
9471 struct to a subclass of rtl_opt_pass along with...
9472 (pass_data_insert_vzeroupper): ...new pass_data instance and...
9473 (make_pass_insert_vzeroupper): ...new function.
9474 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
9475 global struct to a subclass of rtl_opt_pass along with...
9476 (pass_data_work_around_errata): ...new pass_data instance and...
9477 (make_pass_work_around_errata): ...new function.
9478 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
9479 struct to a subclass of rtl_opt_pass along with...
9480 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
9481 (make_pass_mips_machine_reorg2): ...new function.
9483 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9485 * passes.c (pass_manager::operator new): New.
9487 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9489 Handwritten part of conversion of passes to C++ classes.
9491 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
9492 (toplev.o): Add dep on PASS_MANAGER_H.
9493 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
9494 of early local pases to reflect this moving from a global to a
9495 member of gcc::pass_manager.
9496 (cgraph_add_new_function): Likewise.
9497 * lto-cgraph.c (lto_output_node): Update for conversion of
9498 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
9499 * passes.c (opt_pass::clone): New.
9500 (opt_pass::gate): New.
9501 (opt_pass::execute): New.
9502 (opt_pass::opt_pass): New.
9503 (pass_manager::execute_early_local_passes): New.
9504 (pass_manager::execute_pass_mode_switching): new.
9505 (finish_optimization_passes): Convert to...
9506 (pass_manager::finish_optimization_passes): ...this.
9507 (finish_optimization_passes): Update for conversion of passes to
9509 (register_dump_files_1): Use has_gate since we cannot portably
9510 check a vtable entry against NULL.
9511 (dump_one_pass): Likewise.
9512 (ipa_write_summaries_2): Likewise.
9513 (ipa_write_optimization_summaries_1): Likewise.
9514 (ipa_read_summaries_1): Likewise.
9515 (ipa_read_optimization_summaries_1): Likewise.
9516 (execute_ipa_stmt_fixups): Likewise.
9517 (pass_manager::pass_manager): Rewrite pass-creation, invoking
9518 pass-creation functions rather than wiring up globals, and
9519 storing the results in fields of pass_manager generated using
9521 (pass_manager::dump_profile_report): Update for conversion of
9522 passes to C++ classes.
9523 (pass_manager::execute_ipa_summary_passes): Likewise.
9524 (execute_one_ipa_transform_pass): Likewise.
9525 (execute_one_pass): Use has_gate and has_execute since we cannot
9526 portably check a vtable entry against NULL.
9527 * pass_manager.h (pass_manager::finish_optimization_passes): New.
9528 (pass_manager): Use pass-instances.def to add fields for the
9529 various pass instances.
9530 * toplev.c (finalize): Update for move of
9531 finish_optimization_passes to a method of gcc::pass_manager.
9532 * toplev.h (finish_optimization_passes): Move to method of class
9534 * tree-pass.h (struct pass_data): New.
9535 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
9536 (opt_pass::gate): Convert to virtual function.
9537 (opt_pass::~opt_pass): New.
9538 (opt_pass::clone): New.
9539 (opt_pass::execute): Convert to virtual function.
9540 (opt_pass::opt_pass): New.
9541 (opt_pass::ctxt_): new.
9542 (gimple_opt_pass): Convert to subclass of opt_pass.
9543 (gimple_opt_pass::gimple_opt_pass): New.
9544 (rtl_opt_pass): Convert to subclass of opt_pass.
9545 (rtl_opt_pass::rtl_opt_pass): New.
9546 (ipa_opt_pass_d): Convert to subclass of opt_pass.
9547 (ipa_opt_pass_d::ipa_opt_pass_d): New.
9548 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
9549 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
9550 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
9551 invocation of pass_mode_switching to reflect this moving from a
9552 global to a member of gcc::pass_manager.
9553 (ix86_option_override): Rework how pass_insert_vzeroupper is
9554 added to the pass_manager to reflect autogenerated changes.
9555 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
9557 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
9559 PR rtl-optimization/57708
9560 * recog.c (peep2_find_free_register): Validate all regs in a
9563 2013-08-05 Jan Hubicka <jh@suse.cz>
9566 * cgraph.c (verify_cgraph_node): Accept local flags from other
9568 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
9569 (function_and_variable_visibility): Likewise.
9570 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
9572 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9574 * graph.c (init_graph_slim_pretty_print): Remove.
9575 (print_graph_cfg): Do not call it. Use local pretty printer.
9576 (start_graph_dump): Likewise.
9578 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9580 * gimple-pretty-print.c (buffer): Remove.
9581 (initialized): Likewise.
9582 (maybe_init_pretty_print): Likewise.
9583 (print_gimple_stmt): Do not call it. Use non-static local
9584 pretty_printer variable.
9585 (print_gimple_expr): Likewise.
9586 (print_gimple_seq): Likewise.
9587 (gimple_dump_bb): Likewise.
9589 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9591 * asan.c (asan_pp): Remove.
9592 (asan_pp_initialized): Likewise.
9593 (asan_pp_initialize): Likewise.
9594 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
9595 (asan_emit_stack_protection): Tidy. Use local pretty printer.
9596 (asan_add_global): Likewise.
9598 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
9600 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
9601 * diagnostic.h (diagnostic_flush_buffer): Adjust.
9602 * pretty-print.c (pp_formatted_text_data): Likewise.
9603 (pp_indent): Rename from pp_base_indent.
9604 (pp_format): Rename from pp_base_format.
9605 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
9606 (pp_format_verbatim): Rename from pp_base_format_verbatim.
9607 (pp_flush): Rename from pp_base_flush.
9608 (pp_set_line_maximum_length): Rename from
9609 pp_base_set_line_maximum_length.
9610 (pp_clear_output_area): Rename from pp_base_clear_output_area.
9611 (pp_set_prefix): Rename from pp_base_set_prefix.
9612 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
9613 (pp_emit_prefix): Rename from pp_base_emit_prefix.
9614 (pp_append_text): Rename from pp_base_append_text.
9615 (pp_formatted_text): Rename from pp_base_formatted_text.
9616 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
9617 (pp_remaining_character_count_for_line): Rename from
9618 pp_base_remaining_character_count_for_line.
9619 (pp_newline): Rename from pp_base_newline.
9620 (pp_character): Rename from pp_base_character.
9621 (pp_string): Rename from pp_base_string.
9622 (pp_maybe_space): Rename from pp_base_maybe_space.
9623 * asan.c (asan_pp_string): Adjust.
9624 (asan_emit_stack_protection): Likewise.
9625 (asan_add_global): Likewise.
9626 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
9627 * tree-mudflap.c (mf_varname_tree): Likewise.
9628 * tree-pretty-print.c (pp_tree_identifier): Rename from
9629 pp_base_tree_identifier.
9630 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
9631 Declare as function.
9633 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
9635 * pretty-print.h (pp_bar_bar): New.
9636 (pp_ampersand_ampersand): Likewise.
9637 (pp_less_equal): Likewise.
9638 (pp_greater_equal): Likewise.
9639 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
9640 printer functions instead of pp_string or operators and punctuators.
9641 (dump_gimple_call): Likewise.
9642 (dump_gimple_omp_for): Likewise.
9643 (dump_gimple_transaction): Likewise.
9644 (dump_gimple_phi): Likewise.
9645 (pp_gimple_stmt_1): Likewise.
9646 * sched-vis.c (print_insn): Likewise.
9647 * tree-mudflap.c (mf_varname_tree): Likewise.
9648 * tree-pretty-print.c (dump_block_node): Likewise.
9649 (dump_generic_node): Likewise.
9651 2013-08-02 Jan Hubicka <jh@suse.cz>
9653 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
9655 * lto-streamer-out.c (tree_is_indexable): Results decls and
9656 parm decls are not indexable.
9657 (DFS_write_tree_body): Do not follow args and results.
9658 (hash_tree): Likewise.
9659 (output_functions): Rearrange so struct function is needed
9660 only when real body is output; be able to also ouptut abstract
9661 functions; output DECL_ARGUMENTS and DECL_RESULT.
9662 (lto_output): When not in WPA, ale store abstract functions.
9663 (write_symbol): Do not care about RESULT_DECL.
9664 (output_symbol_p): Handle correctly sbtract decls.
9665 * lto-streamer-in.c (input_function): Rearrange so struct
9666 function can be NULL at entry; allow streaming of
9667 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
9668 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
9669 sanity check during LTO.
9670 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
9671 RESULT_DECl and DECL_ARGUMENTS.
9672 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
9675 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
9677 * pretty-print.h (pp_underscore): New.
9679 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
9680 printer functions instead of pp_character.
9681 (dump_binary_rhs): Likewise.
9682 (dump_ternary_rhs): Likewise.
9683 (dump_gimple_call_args): Likewise.
9684 (pp_points_to_solution): Likewise.
9685 (dump_gimple_call): Likewise.
9686 (dump_gimple_switch): Likewise.
9687 (dump_gimple_cond): Likewise.
9688 (dump_gimple_bind): Likewise.
9689 (dump_gimple_try): Likewise.
9690 (dump_gimple_omp_for): Likewise.
9691 (dump_gimple_omp_continue): Likewise.
9692 (dump_gimple_omp_single): Likewise.
9693 (dump_gimple_omp_sections): Likewise.
9694 (dump_gimple_omp_block): Likewise.
9695 (dump_gimple_omp_critical): Likewise.
9696 (dump_gimple_transaction): Likewise.
9697 (dump_gimple_asm): Likewise.
9698 (dump_gimple_phi): Likewise.
9699 (dump_gimple_omp_parallel): Likewise.
9700 (dump_gimple_omp_task): Likewise.
9701 (dump_gimple_omp_atomic_load): Likewise.
9702 (dump_gimple_omp_atomic_store): Likewise.
9703 (dump_gimple_mem_ops): Likewise.
9704 (pp_gimple_stmt_1): Likewise.
9705 (pp_cfg_jump): Likewise.
9706 (dump_implicit_edges): Likewise.
9707 (gimple_dump_bb_for_graph): Likewise.
9708 * graph.c (draw_cfg_node): Likewise.
9709 * langhooks.c (lhd_print_error_function): Likewise.
9710 * sched-vis.c (print_exp): Likewise.
9711 (print_value): Likewise.
9712 (print_pattern): Likewise.
9713 (print_insn): Likewise.
9714 (rtl_dump_bb_for_graph): Likewise.
9715 * tree-pretty-print.c (dump_function_declaration): Likewise.
9716 (dump_array_domain): Likewise.
9717 (dump_omp_clause): Likewise.
9718 (dump_location): Likewise.
9719 (dump_generic_node): Likewise.
9720 (print_struct_decl): Likewise.
9721 * diagnostic.c (diagnostic_show_locus): Use pp_space.
9723 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9725 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
9726 candidate table when replacing a candidate statement.
9727 (replace_rhs_if_not_dup): Likewise.
9728 (replace_one_candidate): Likewise.
9730 2013-08-02 Jan Hubicka <jh@suse.cz>
9731 Martin Liska <marxin.liska@gmail.com>
9733 * cgraphunit.c (add_new_function): Fix logic when adding from
9735 (assemble_thunk): Rename to ...
9736 (expand_thunk); .. this one; export; get it working with
9737 general functions; make produced gimple valid.
9738 * cgraph.h (expand_thunk): Declare.
9740 2013-08-02 Jan Hubicka <jh@suse.cz>
9742 * ipa-cp.c (gather_context_independent_values): Use
9743 ipa_get_param_move_cost.
9744 (get_replacement_map): Remove PARAM; move parameter folding
9746 (create_specialized_node): Update.
9747 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
9748 assert that we have gimple body; update move_cost.
9749 (count_formal_params): Assert that we have gimple body.
9750 (ipa_dump_param): New function.
9751 (ipa_alloc_node_params): Break out from ...
9752 (ipa_initialize_node_params): ... here.
9753 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
9754 (ipa_write_node_info): Stream move costs.
9755 (ipa_read_node_info): Read move costs.
9756 (ipa_update_after_lto_read): Do not recompute node params.
9757 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
9758 (ipa_get_param): Check we are not in WPA.
9759 (ipa_get_param_move_cost): New.
9760 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
9761 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
9762 parm numbers to be present.
9764 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
9766 PR rtl-optimization/58048
9767 * lra-constraints.c (process_alt_operands): Don't check asm
9768 operand on register.
9770 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
9772 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
9773 the implied StoreLoad barrier for atomic operations if before.
9775 2013-08-02 Jan Hubicka <jh@suse.cz>
9776 Martin Liska <marxin.liska@gmail.com>
9778 * cgraph.c (cgraph_function_body_availability): Do not check
9780 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
9781 symtab_node_availability): Declare.
9782 * ipa.c (can_replace_by_local_alias): New.
9783 (function_and_variable_visibility): Use it.
9784 * symtab.c (symtab_for_node_and_aliases,
9785 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
9787 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
9789 PR rtl-optimization/57963
9790 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
9791 (lra_constraints): Use them.
9793 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
9795 * config/arm/types.md (define_attr "type"): Add "load_acq"
9797 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
9799 (cortex_a53_store1): Likewise.
9801 2013-08-01 Jan Hubicka <jh@suse.cz>
9803 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
9804 partitions are not needed.
9806 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
9808 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
9809 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
9810 MAYBE_NON_Q_CLASS_P where appropriate.
9812 2013-08-01 Jan Hubicka <jh@suse.cz>
9814 * cgraph.h (release_function_body): Declare.
9815 * tree.c (free_lang_data_in_decl): Free, parameters and return values
9816 of unused delcarations.
9818 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9820 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
9821 RTL form when subtracting a constant.
9823 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9825 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
9826 Generate canonical plus rtx with negated immediate instead of minus
9828 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
9830 2013-08-01 Jan Hubicka <jh@suse.cz>
9832 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
9833 (cgraph_release_function_body): Likewise.
9834 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
9835 * cgraph.h (cgrpah_node): Rename abstract_and_needed
9836 to used_as_abstract_origin.
9837 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
9838 symbols used as abstract origins.
9839 * cgraphunit.c (analyze_functions): Update.
9840 * ipa.c (symtab_remove_unreachable_nodes): Recompute
9841 used_as_abstract_origin.
9842 * tree-inline.c (tree_function_versioning): Update
9843 used_as_abstract_origin; be ready for DECL_RESULT and
9844 DECL_ARGUMENTS to be NULL.
9846 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
9847 for abstract functions.
9848 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
9851 2013-08-01 Jan Hubicka <jh@suse.cz>
9853 * profile.c (compute_value_histograms): Fix thinko.
9855 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9857 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
9858 aarch-common-protos.h to extra_headers.
9859 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9860 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
9861 * config/aarch64/t-aarch64 (aarch-common.o): Define.
9863 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9865 * config/aarch64/aarch64.md (define_attr "type"): Delete.
9866 Include "../arm/types.md". Define "type" attribute for all patterns.
9867 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
9870 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
9872 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
9873 to support power8 load fusion.
9874 (fusion_gpr_mem_load): Likewise.
9876 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
9878 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
9879 declarations for power8 load fusion.
9880 (emit_fusion_gpr_load): Likewise.
9882 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
9883 tuning for power8, turn on fusion mode by default. Turn on sign
9884 extending fusion mode if normal fusion mode is on, and we are at
9886 (fusion_gpr_load_p): New function, return true if we can fuse an
9887 addis instruction with a dependent load to a GPR.
9888 (emit_fusion_gpr_load): Emit the instructions for power8 load
9891 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
9892 (VSX load fusion peepholes): New peepholes to fuse together an
9893 addi instruction with a VSX load instruction.
9895 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
9896 peepholes to fuse an addis instruction with a load to a GPR base
9897 register. If we are supporting sign extending fusions, convert
9898 sign extending loads to zero extending loads and add an explicit
9901 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9903 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
9904 aarch-common-protos.h to extra_headers.
9905 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9906 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
9907 (arm_early_store_addr_dep): Likewise.
9908 (arm_no_early_alu_shift_dep): Likewise.
9909 (arm_no_early_alu_shift_value_dep): Likewise.
9910 (arm_no_early_mul_dep): Likewise.
9911 (arm_no_early_store_addr_dep): Likewise.
9912 (arm_mac_accumulator_is_mul_result): Likewise.
9913 (arm_mac_accumulator_is_result): Likewise.
9914 * config/arm/aarch-common.c: ... here. New file.
9915 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
9917 (arm_early_store_addr_dep): Likewise.
9918 (arm_no_early_alu_shift_dep): Likewise.
9919 (arm_no_early_alu_shift_value_dep): Likewise.
9920 (arm_no_early_mul_dep): Likewise.
9921 (arm_no_early_store_addr_dep): Likewise.
9922 (arm_mac_accumulator_is_mul_result): Likewise.
9923 (arm_mac_accumulator_is_result): Likewise.
9924 * config/arm/aarch-common-protos.h: ... here. New file.
9925 * config/arm/t-arm (aarch-common.o): Define.
9927 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9929 * config/arm/arm.md: Include new file "types.md".
9930 (define_attr "type"): Move from here to ...
9931 (define_attr "mul32"): Likewise.
9932 (define_attr "mul64"): Likewise.
9933 * config/arm/types.md: ... here. New file.
9935 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
9937 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
9938 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
9940 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9942 * gen-pass-instances.awk: Fix offset of substr().
9944 2013-07-31 David Malcolm <dmalcolm@redhat.com>
9946 * Makefile.in (pass-instances.def): New.
9947 (passes.o): Replace dependency on passes.def with one on
9950 * gen-pass-instances.awk: New.
9952 * passes.c (pass_manager::pass_manager): Use pass-instances.def
9953 rather than passes.def, updating local definition of NEXT_PASS
9954 macro to add an extra NUM parameter (currently unused).
9956 2013-07-30 David Malcolm <dmalcolm@redhat.com>
9958 * Makefile.in (PASS_MANAGER_H): New.
9959 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
9960 (passes.o): Likewise.
9961 (statistics.o): Likewise.
9962 (cgraphunit.o): Likewise.
9963 (context.o): Depend on PASS_MANAGER_H.
9965 * pass_manager.h: New.
9967 * cgraphunit.c (cgraph_add_new_function): Update for moves
9968 of globals to fields of pass_manager.
9969 (analyze_function): Likewise.
9970 (expand_function): Likewise.
9971 (ipa_passes): Likewise.
9972 (compile): Likewise.
9974 * context.c (context::context): New.
9975 * context.h (context::context): New.
9976 (context::get_passes): New.
9977 (context::passes_): New.
9979 * lto-cgraph.c (input_node): Update for moves of globals to
9980 fields of pass_manager.
9982 * passes.c (all_passes): Remove, in favor of a field of the
9983 same name within the new class pass_manager.
9984 (all_small_ipa_passes): Likewise.
9985 (all_lowering_passes): Likewise.
9986 (all_regular_ipa_passes): Likewise.
9987 (all_late_ipa_passes): Likewise.
9988 (all_lto_gen_passes): Likewise.
9989 (passes_by_id): Likewise.
9990 (passes_by_id_size): Likewise.
9991 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
9992 the new class pass_manager.
9993 (set_pass_for_id): Convert to...
9994 (pass_manager::set_pass_for_id): ...method.
9995 (get_pass_for_id): Convert to...
9996 (pass_manager::get_pass_for_id): ...method.
9997 (register_one_dump_file): Move body of implementation into...
9998 (pass_manager::register_one_dump_file): ...here.
9999 (register_dump_files_1): Convert to...
10000 (pass_manager::register_dump_files_1): ...method.
10001 (register_dump_files): Convert to...
10002 (pass_manager::register_dump_files): ...method.
10003 (create_pass_tab): Update for moves of globals to fields of
10005 (dump_passes): Move body of implementation into...
10006 (pass_manager::dump_passes): ...here.
10007 (register_pass): Move body of implementation into...
10008 (pass_manager::register_pass): ...here.
10009 (init_optimization_passes): Convert into...
10010 (pass_manager::pass_manager): ...constructor for new
10011 pass_manager class, and initialize the pass_lists array.
10012 (check_profile_consistency): Update for moves of globals to
10013 fields of pass_manager.
10014 (dump_profile_report): Move body of implementation into...
10015 (pass_manager::dump_profile_report): ...here.
10016 (ipa_write_summaries_1): Update for moves of pass lists from
10017 being globals to fields of pass_manager.
10018 (ipa_write_optimization_summaries): Likewise.
10019 (ipa_read_summaries): Likewise.
10020 (ipa_read_optimization_summaries): Likewise.
10021 (execute_all_ipa_stmt_fixups): Likewise.
10023 * statistics.c (statistics_fini): Update for moves of globals to
10024 fields of pass_manager.
10026 * toplev.c (general_init): Replace call to
10027 init_optimization_passes with construction of the pass_manager
10030 * tree-pass.h (all_passes): Remove, in favor of a field of the
10031 same name within the new class pass_manager.
10032 (all_small_ipa_passes): Likewise.
10033 (all_lowering_passes): Likewise.
10034 (all_regular_ipa_passes): Likewise.
10035 (all_lto_gen_passes): Likewise.
10036 (all_late_ipa_passes): Likewise.
10037 (passes_by_id): Likewise.
10038 (passes_by_id_size): Likewise.
10039 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
10040 the new class pass_manager.
10041 (get_pass_for_id): Remove.
10043 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
10045 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
10048 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
10050 * arm.md (mulhi3): New expand pattern.
10052 2013-07-30 Jan Hubicka <jh@suse.cz>
10053 Martin Liska <marxin.liska@gmail.com>
10055 * profile.c (compute_value_histograms): Do not ICE when
10056 there is mismatch only on some counters.
10058 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10060 PR rtl-optimization/57637
10061 * function.c (move_insn_for_shrink_wrap): Also check the
10062 GEN set of the LIVE problem for the liveness analysis
10063 if it exists, otherwise give up.
10065 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10067 PR tree-optimization/57993
10068 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
10069 replaced statement in the candidate table.
10070 (phi_add_costs): Return infinite cost when the hidden basis does
10071 not dominate all phis on which the candidate is dependent.
10072 (replace_one_candidate): Record replaced statement in the
10075 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
10077 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
10078 (ashlv2si3): New expander.
10079 (*ashlv2si3_i): New define_insn_and_split.
10080 * predicates.md (float_operation): Allow patterns with three
10081 basic sub-patterns.
10083 PR rtl-optimization/58021
10084 * mode-switching.c (create_pre_exit): Always split off preceding
10085 insns if we are not at the basic block head.
10087 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
10089 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
10090 (UCLIBC_DYNAMIC_LINKER): New macro.
10091 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
10093 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
10094 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
10096 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
10097 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
10098 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
10099 for SF and DF modes. Use ieee_quad_format for TF mode.
10100 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
10101 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
10102 (mips_option_override): Handle `-mnan=legacy'.
10103 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
10104 `-mabs=2008' and `-mnan=2008'.
10105 (OPTION_DEFAULT_SPECS): Add "nan" default.
10106 (ASM_SPEC): Handle `-mnan='.
10107 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
10108 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
10109 comment accordingly.
10110 (neg<mode>2): Likewise.
10111 * config/mips/mips.opt (mabs, mnan): New options.
10112 * doc/install.texi (Configuration): Document `--with-nan=' option.
10113 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
10115 (MIPS Options): Document them.
10116 * config.gcc <mips*-*-*>: Handle `--with-nan='.
10117 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
10118 * configure: Regenerate.
10119 * config.in: Regenerate.
10121 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
10123 * config/i386/i386.md (float post-reload splitters): Do not check
10124 for subregs of SSE registers.
10126 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
10127 H.J. Lu <hongjiu.lu@intel.com>
10131 * config/i386/i386.md (post-reload splitter
10132 to avoid partial SSE reg dependency stalls): New pattern.
10134 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
10136 * config/s390/s390.md ("movcc"): Swap load and store instructions.
10138 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
10140 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
10141 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
10143 2013-07-26 Cary Coutant <ccoutant@google.com>
10145 * dwarf2out.c (die_checksum_ordered): Don't include template
10146 instantiations in signature.
10147 (is_template_parameter): New function.
10148 (is_template_instantiation): New function.
10149 (generate_skeleton_bottom_up): Don't include template instantiations
10151 (generate_skeleton): Likewise.
10152 (break_out_comdat_types): Move recursive call to break out nested
10154 (prune_unused_types_mark_generic_parms_dies): Call
10155 is_template_parameter.
10157 2013-07-26 Ian Bolton <ian.bolton@arm.com>
10159 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
10160 uses vector registers.
10161 * config/aarch64/iterators.md: Add attributes rtn and vas.
10163 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10164 Richard Earnshaw <richard.earnshaw@arm.com>
10166 * combine.c (simplify_comparison): Re-canonicalize operands
10168 * config/arm/arm.md (movcond_addsi): New splitter.
10170 2013-07-25 Sterling Augustine <saugustine@google.com>
10172 * dwarf2out.c (size_of_pubnames): Move code to...
10173 (include_pubname_in_output): ...here. New.
10174 (want_pubnames): Rearrange.
10175 (output_pubnames): Call include_pubname_in_output. Move assertion.
10177 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
10179 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
10181 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
10184 * doc/extend.texi: Remove obsolete builtins. Fix
10185 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
10187 2013-07-25 Jan Hubicka <jh@suse.cz>
10189 * cgraph.c (release_function_body): Break out from ...
10190 (cgraph_release_function_body): ... this one; also release DECL_RESULT
10191 and DECL_ARGUMENTS.
10192 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
10193 old_tree in the map.
10194 (create_specialized_node): Update.
10195 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
10197 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
10198 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
10199 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
10200 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
10203 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10205 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
10206 addsi3_carryin_alt2_<optab>): Correct output template.
10208 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10210 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
10211 Adjust for arm_restrict_it.
10212 Remove trailing whitespace.
10214 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
10216 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
10217 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
10219 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
10221 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
10223 PR rtl-optimization/57960
10224 * lra-constraints.c (process_alt_operands): Use the right mode
10225 when checking strict_low.
10227 2013-07-25 Jan Hubicka <jh@suse.cz>
10229 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
10230 * cgraph.c (cgraph_remove_node): Do not release function body
10231 when in cgraph streaming.
10232 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
10233 in other partitions are not considered reachable; fix handling of
10236 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10238 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
10240 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10245 * config/arm/arm.md ("*sibcall_insn): Replace use of
10246 Ss with US. Adjust output for v5 and v4t.
10247 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
10249 * config/arm/constraints.md ("Ss"): Rename to US.
10251 2013-07-25 Terry Guo <terry.guo@arm.com>
10253 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
10254 shift_add/shift_sub0/shift_sub1 RTXs.
10256 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
10257 Anton Blanchard <anton@au1.ibm.com>
10259 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
10260 (altivec_vpks<VI_char>ss): Likewise.
10261 (altivec_vpks<VI_char>us): Likewise.
10262 (altivec_vpku<VI_char>us): Likewise.
10263 (altivec_vpku<VI_char>um): Likewise.
10265 2013-07-24 David Malcolm <dmalcolm@redhat.com>
10267 Introduce context class.
10269 * Makefile.in (CONTEXT_H): New.
10270 (OBJS): Add context.o.
10271 (toplev.o): Add CONTEXT_H to dependencies.
10274 * toplev.c (general_init): Create the singleton gcc::context instance.
10280 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
10282 PR rtl-optimization/57968
10283 * mode-switching.c (create_pre_exit): Allow instructions that
10284 don't set a return register to need a non-exit mode.
10286 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10287 Anton Blanchard <anton@au1.ibm.com>
10289 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
10290 operands to vperm for little endian.
10291 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
10292 of lvsl to create the control mask for a vperm for little endian.
10294 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10295 Anton Blanchard <anton@au1.ibm.com>
10297 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
10298 two operands for little-endian.
10300 2013-07-23 Steve Ellcey <sellcey@mips.com>
10302 * config/mips/mips.c (mips_case_values_threshold): New.
10303 (TARGET_CASE_VALUES_THRESHOLD): Define.
10305 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10306 Anton Blanchard <anton@au1.ibm.com>
10308 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
10309 selection of field for vector splat in little endian mode.
10311 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10313 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
10314 expanders to rs6000.md.
10315 (ior<mode>3): Likewise.
10316 (and<mode>3): Likewise.
10317 (one_cmpl<mode>2): Likewise.
10318 (nor<mode>3): Likewise.
10319 (andc<mode>3): Likewise.
10320 (eqv<mode>3): Likewise.
10321 (nand<mode>3): Likewise.
10322 (orc<mode>3): Likewise.
10324 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
10327 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
10328 to split multi-word logical operations.
10329 (rs6000_split_logical_di): Likewise.
10330 (rs6000_split_logical): Likewise.
10332 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
10333 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
10334 and allow TImode operations in 32-bit.
10335 (vsx_and<mode>3_64bit): Likewise.
10336 (vsx_ior<mode>3_32bit): Likewise.
10337 (vsx_ior<mode>3_64bit): Likewise.
10338 (vsx_xor<mode>3_32bit): Likewise.
10339 (vsx_xor<mode>3_64bit): Likewise.
10340 (vsx_one_cmpl<mode>2_32bit): Likewise.
10341 (vsx_one_cmpl<mode>2_64bit): Likewise.
10342 (vsx_nor<mode>3_32bit): Likewise.
10343 (vsx_nor<mode>3_64bit): Likewise.
10344 (vsx_andc<mode>3_32bit): Likewise.
10345 (vsx_andc<mode>3_64bit): Likewise.
10346 (vsx_eqv<mode>3_32bit): Likewise.
10347 (vsx_eqv<mode>3_64bit): Likewise.
10348 (vsx_nand<mode>3_32bit): Likewise.
10349 (vsx_nand<mode>3_64bit): Likewise.
10350 (vsx_orc<mode>3_32bit): Likewise.
10351 (vsx_orc<mode>3_64bit): Likewise.
10353 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
10354 logical types in GPRs.
10356 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
10357 logical insns to rs6000.md, and allow TImode operations in
10359 (altivec_ior<mode>3): Likewise.
10360 (altivec_xor<mode>3): Likewise.
10361 (altivec_one_cmpl<mode>2): Likewise.
10362 (altivec_nor<mode>3): Likewise.
10363 (altivec_andc<mode>3): Likewise.
10365 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
10366 attributes for moving the 128-bit logical operations into
10368 (BOOL_REGS_OUTPUT): Likewise.
10369 (BOOL_REGS_OP1): Likewise.
10370 (BOOL_REGS_OP2): Likewise.
10371 (BOOL_REGS_UNARY): Likewise.
10372 (BOOL_REGS_AND_CR0): Likewise.
10373 (one_cmpl<mode>2): Add support for DI logical operations on
10374 32-bit, splitting the operations to 32-bit.
10375 (anddi3): Likewise.
10376 (iordi3): Likewise.
10377 (xordi3): Likewise.
10378 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
10379 changes to combine the 32/64-bit code, allow logical operations on
10380 TI mode in 32-bit, and to use similar match_operator patterns like
10381 scalar mode uses. Combine the Altivec and VSX code for logical
10382 operations, and move it here.
10383 (ior<mode>3, 128-bit types): Likewise.
10384 (xor<mode>3, 128-bit types): Likewise.
10385 (one_cmpl<mode>3, 128-bit types): Likewise.
10386 (nor<mode>3, 128-bit types): Likewise.
10387 (andc<mode>3, 128-bit types): Likewise.
10388 (eqv<mode>3, 128-bit types): Likewise.
10389 (nand<mode>3, 128-bit types): Likewise.
10390 (orc<mode>3, 128-bit types): Likewise.
10391 (and<mode>3_internal): Likewise.
10392 (bool<mode>3_internal): Likewise.
10393 (boolc<mode>3_internal1): Likewise.
10394 (boolc<mode>3_internal2): Likewise.
10395 (boolcc<mode>3_internal1): Likewise.
10396 (boolcc<mode>3_internal2): Likewise.
10397 (eqv<mode>3_internal1): Likewise.
10398 (eqv<mode>3_internal2): Likewise.
10399 (one_cmpl1<mode>3_internal): Likewise.
10401 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
10403 * config/microblaze/microblaze.c (microblaze_expand_prologue):
10404 Rename flag_stack_usage to flag_stack_usage_info.
10406 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
10408 * config/microblaze/sync.md: New file.
10409 * config/microblaze/microblaze.md: Include sync.md
10410 * config/microblaze/microblaze.c: Add print_operand 'y'.
10411 * config/microblaze/constraints.md: Add memory_contraint
10412 'Q' which is a single register.
10414 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
10416 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
10418 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
10420 * reload.c (find_reloads): Exit loop once we find this operand
10421 cannot be reloaded somehow for this alternative.
10423 * reload.c (find_reloads): Exit loop once we find a hard register.
10425 * rtlanal.c (computed_jump_p): Exit loop once we find label
10428 * i386.c (ix86_pad_returns): Exit loop after setting replace.
10430 * cfgloopmanip.c (remove_path): Exit loop after setting
10433 * gensupport.c (subst_dup): Avoid loop if code is not
10434 MATCH_DUP nor MATCH_OP_DUP.
10436 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
10438 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
10440 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10442 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
10443 true for SP_REGNUM if mode == ptr_mode.
10444 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
10445 with value R0_REGNUM + 31.
10447 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10449 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
10450 pad pointer-typed argument downward.
10452 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10454 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
10455 and __ILP32__ when the ILP32 model is in use.
10457 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10459 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
10460 (aarch64_load_symref_appropriately): In the case of
10461 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
10462 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
10463 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
10464 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
10465 if 'mode' doesn't equal to 'ptr_mode'.
10466 (aarch64_output_mi_thunk): Add an assertion on the alignment of
10467 'vcall_offset'; change to call aarch64_emit_move differently depending
10468 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
10469 to calculate the upper bound of 'vcall_offset'.
10470 (aarch64_cannot_force_const_mem): Change to also return true if
10472 (aarch64_legitimize_reload_address): In the case of large
10473 displacements, add new local variable 'xmode' and an assertion
10474 based on it; change to use 'xmode' to generate the new rtx and
10476 (aarch64_asm_trampoline_template): Change to generate the template
10477 differently depending on TARGET_ILP32 or not; change to use
10478 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
10479 (aarch64_trampoline_size): Removed.
10480 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
10481 and replace immediate literals with it. Change to use 'ptr_mode'
10482 instead of 'DImode' and call convert_memory_address if the mode
10483 of 'fnaddr' doesn't equal to 'ptr_mode'.
10484 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
10486 (aarch64_elf_asm_destructor): Likewise.
10487 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
10488 on TARGET_ILP32 instead of aarch64_trampoline_size.
10489 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
10490 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
10491 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
10492 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
10493 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
10494 (storewb_pair<GPI:mode>_<P:mode>): ... this.
10495 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
10496 depending on the value of 'mode'.
10497 (add_losym_<mode>): New.
10498 (ldr_got_small_<mode>): New, based on ldr_got_small.
10499 (ldr_got_small): Remove.
10500 (ldr_got_small_sidi): New.
10501 * config/aarch64/iterators.md (P): New.
10502 (PTR): Change to 'ptr_mode' in the condition.
10504 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10506 * config.gcc (aarch64*-*-*): Support --with-abi.
10507 (aarch64*-*-elf): Support --with-multilib-list.
10508 (aarch64*-*-linux*): Likewise.
10509 (supported_defaults): Add abi to aarch64*-*-*.
10510 * configure.ac: Mention AArch64 for --with-multilib-list.
10511 * configure: Re-generated.
10512 * config/aarch64/biarchilp32.h: New file.
10513 * config/aarch64/biarchlp64.h: New file.
10514 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
10516 (MULTILIB_DEFAULTS): Ditto.
10517 (DRIVER_SELF_SPECS): Ditto.
10518 (ASM_SPEC): Update to also substitute -mabi.
10519 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
10520 file whose name depends on -mabi= and -mbig-endian.
10521 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
10523 (POINTER_SIZE): New define.
10524 (POINTERS_EXTEND_UNSIGNED): Ditto.
10525 (enum aarch64_abi_type): New enumeration tag.
10526 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
10527 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
10528 (TARGET_ILP32): New define.
10529 * config/aarch64/aarch64.opt (mabi): New.
10530 (aarch64_abi): New.
10531 (ilp32, lp64): New values for -mabi.
10532 * config/aarch64/t-aarch64 (comma): New define.
10533 (MULTILIB_OPTIONS): Ditto.
10534 (MULTILIB_DIRNAMES): Ditto.
10535 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
10536 * doc/invoke.texi: Document -mabi for AArch64.
10538 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
10540 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
10542 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10543 Anton Blanchard <anton@au1.ibm.com>
10545 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
10546 endianness when selecting field to splat.
10548 2013-07-22 Eric Christopher <echristo@gmail.com>
10550 * dwarf2out.c (die_odr_checksum): New function to use
10551 CHECKSUM_ macros and ULEB128 for DIE tag.
10552 (generate_type_signature): Use.
10554 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
10556 * config.gcc (sparc*-*-*): Accept leon3 processor.
10557 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
10558 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
10559 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
10560 * config/sparc/sparc.opt (enum processor_type): Add leon3.
10561 (mfix-ut699): Adjust comment.
10562 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
10563 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
10564 (CPP_CPU_SPEC): Likewise.
10565 (ASM_CPU_SPEC): Likewise.
10566 * config/sparc/sparc.c (leon3_cost): New constant.
10567 (sparc_option_override): Add leon3 support.
10568 (mem_ref): New function.
10569 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
10570 (sparc_do_work_around_errata): Look into the instruction in the delay
10571 slot and adjust accordingly. Add fix for the data cache nullify issues
10572 of the UT699. Change insertion position for the NOP.
10573 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
10574 (leon3_load): New reservation.
10575 (leon_store): Bump latency to 2.
10576 (grfpu): New automaton.
10577 (grfpu_alu): New unit.
10578 (grfpu_ds): Likewise.
10579 (leon_fp_alu): Adjust.
10580 (leon_fp_mult): Delete.
10581 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
10582 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
10583 * config/sparc/sparc.md (cpu): Add leon3.
10584 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
10585 (swapsi): Likewise.
10586 (atomic_test_and_set): Likewise.
10587 (ldstub): Likewise.
10589 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
10591 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
10592 default for R5900 targets.
10593 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
10594 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
10595 * config/mips/mips.c (mips_option_override): Report an error for
10596 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
10597 for -march=r5900 -mhard-float.
10599 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
10601 * df-problems.c (can_move_insns_across): Exit loop once we
10602 find a non-fixed, non-global register.
10604 * ipa-pure-const.c (propagate_nothrow): Exit loop after
10607 * omega.c (omega_eliminate_red): Break after setting red_found.
10608 (omega_problem_has_red_equations): Similarly after setting found.
10609 (omega_query_variable): Similarly after setting coupled.
10611 2013-07-22 Marek Polacek <polacek@redhat.com>
10613 * gimplify.c: Don't include gimple.h twice.
10615 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10617 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
10618 instead of TARGET_THUMB1.
10619 (Pz): New constraint.
10620 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
10622 (compare_negsi_si): Likewise.
10623 (compare_addsi2_op0): Likewise.
10624 (compare_addsi2_op1): Likewise.
10625 (addsi3_carryin_<optab>): Likewise.
10626 (addsi3_carryin_alt2_<optab>): Likewise.
10627 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
10628 for arm_restrict_it.
10629 (subsi3_carryin): Likewise.
10630 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
10631 (minmax_arithsi): Disable for arm_restrict_it.
10632 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
10633 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
10634 (satsi_<SAT:code>_shift): Likewise.
10635 (arm_shiftsi3): Add alternative for 16-bit encoding.
10636 (arm32_movhf): Disable for arm_restrict_it.
10637 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
10638 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
10640 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
10642 * config/arm/arm.md (attribute "insn"): Delete.
10643 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
10644 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
10645 (not_shiftsi): Update for attribute change.
10646 (not_shiftsi_compare0): Likewise.
10647 (not_shiftsi_compare0_scratch): Likewise.
10648 (arm_one_cmplsi2): Likewise.
10649 (thumb1_one_cmplsi2): Likewise.
10650 (notsi_compare0): Likewise.
10651 (notsi_compare0_scratch): Likewise.
10652 (thumb1_movdi_insn): Likewise.
10653 (arm_movsi_insn): Likewise.
10654 (movhi_insn_arch4): Likewise.
10655 (movhi_bytes): Likewise.
10656 (arm_movqi_insn): Likewise.
10657 (thumb1_movqi_insn): Likewise.
10658 (arm32_movhf): Likewise.
10659 (thumb1_movhf): Likewise.
10660 (arm_movsf_soft_insn): Likewise.
10661 (thumb1_movsf_insn): Likewise.
10662 (thumb_movdf_insn): Likewise.
10663 (movsicc_insn): Likewise.
10664 (movsfcc_soft_insn): Likewise.
10665 (and_scc): Likewise.
10666 (cond_move): Likewise.
10667 (if_move_not): Likewise.
10668 (if_not_move): Likewise.
10669 (if_shift_move): Likewise.
10670 (if_move_shift): Likewise.
10671 (if_shift_shift): Likewise.
10672 (if_not_arith): Likewise.
10673 (if_arith_not): Likewise.
10674 (cond_move_not): Likewise.
10675 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
10676 (neon_mov<mode>): Likewise.
10677 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
10678 (thumb2_movsi_vfp): Likewise.
10679 (movsf_vfp): Likewise.
10680 (thumb2_movsf_vfp): Likewise.
10681 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
10683 (cortexa7_older_only): Likewise.
10684 (cortexa7_younger): Likewise.
10685 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
10686 (1020alu_shift_op): Likewise.
10687 (1020alu_shift_reg_op): Likewise.
10688 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
10689 (alu_shift_op): Likewise.
10690 (alu_shift_reg_op): Likewise.
10691 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
10692 (11_alu_shift_op): Likewise.
10693 (11_alu_shift_reg_op): Likewise.
10694 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
10695 (9_alu_shift_reg_op): Likewise.
10696 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10698 (cortex_a15_alu_shift): Likewise.
10699 (cortex_a15_alu_shift_reg): Likewise.
10700 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
10701 (cortex_a5_alu_shift): Likewise.
10702 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
10704 (cortex_a53_alu_shift): Likewise.
10705 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
10707 (cortex_a7_alu_reg): Likewise.
10708 (cortex_a7_alu_shift): Likewise.
10709 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10710 (cortex_a8_alu_shift): Likewise.
10711 (cortex_a8_alu_shift_reg): Likewise.
10712 (cortex_a8_mov): Likewise.
10713 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10714 (cortex_a9_dp_shift): Likewise.
10715 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
10716 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
10717 (cortex_r4_mov): Likewise.
10718 (cortex_r4_alu_shift): Likewise.
10719 (cortex_r4_alu_shift_reg): Likewise.
10720 * config/arm/fa526.md (526_alu_op): Update for attribute change.
10721 (526_alu_shift_op): Likewise.
10722 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
10723 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
10724 (626te_alu_shift_op): Likewise.
10725 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
10726 (726te_alu_op): Likewise.
10727 (726te_alu_shift_op): Likewise.
10728 (726te_alu_shift_reg_op): Likewise.
10729 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
10730 (mp626_alu_shift_op): Likewise.
10731 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
10732 (pj4_alu_e1_conds): Likewise.
10733 (pj4_alu): Likewise.
10734 (pj4_alu_conds): Likewise.
10735 (pj4_shift): Likewise.
10736 (pj4_shift_conds): Likewise.
10737 (pj4_alu_shift): Likewise.
10738 (pj4_alu_shift_conds): Likewise.
10740 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10742 * config/arm/predicates.md (shiftable_operator_strict_it):
10744 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
10745 Disable cond_exec version for arm_restrict_it.
10746 (thumb2_smaxsi3): Convert to generate cond_exec.
10747 (thumb2_sminsi3): Likewise.
10748 (thumb32_umaxsi3): Likewise.
10749 (thumb2_uminsi3): Likewise.
10750 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
10751 (thumb2_neg_abssi2): Likewise.
10752 (thumb2_mov_scc): Add alternative for 16-bit encoding.
10753 (thumb2_movsicc_insn): Adjust alternatives.
10754 (thumb2_mov_negscc): Disable for arm_restrict_it.
10755 (thumb2_mov_negscc_strict_it): New pattern.
10756 (thumb2_mov_notscc_strict_it): New pattern.
10757 (thumb2_mov_notscc): Disable for arm_restrict_it.
10758 (thumb2_ior_scc): Likewise.
10759 (thumb2_ior_scc_strict_it): New pattern.
10760 (thumb2_cond_move): Adjust for arm_restrict_it.
10761 (thumb2_cond_arith): Disable for arm_restrict_it.
10762 (thumb2_cond_arith_strict_it): New pattern.
10763 (thumb2_cond_sub): Adjust for arm_restrict_it.
10764 (thumb2_movcond): Likewise.
10765 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
10766 (thumb2_zero_extendhisi2_v6): Likewise.
10767 (thumb2_zero_extendqisi2_v6): Likewise.
10768 (orsi_notsi_si): Likewise.
10769 (orsi_not_shiftsi_si): Likewise.
10771 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
10773 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
10774 instruction sequence is 1 byte shorter.
10776 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
10778 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
10779 it is not needed after split.
10781 2013-07-20 Iain Sandoe <iain@codesourcery.com>
10784 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
10785 second label for nonlocal goto receivers. Don't output pic base labels
10786 unless we're producing PIC; mark that action unreachable().
10787 (ix86_save_reg): If the function contains a nonlocal label, save the
10789 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
10790 * config/darwin.c (emitted_pic_label_num): New GTY.
10791 (update_pic_label_number_if_needed): New.
10792 (machopic_output_function_base_name): Adjust for nonlocal receiver
10794 (machopic_should_output_picbase_label): New.
10795 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
10796 (nonlocal_goto_receiver): New insn and split.
10798 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
10800 * config/aarch64/aarch64-builtins.c
10801 (aarch64_fold_builtin): Fold abs in all modes.
10802 * config/aarch64/aarch64-simd-builtins.def
10803 (abs): Enable for all modes.
10804 * config/aarch64/arm_neon.h
10805 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
10806 (vabs_f64): Add missing intrinsic.
10808 2013-07-19 Ian Bolton <ian.bolton@arm.com>
10810 * config/aarch64/arm_neon.h (vabs_s64): New function
10812 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
10815 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
10816 * config/avr/avr.md (adjust_len): Add `round'.
10817 * config/avr/avr-protos.h (avr_out_round): New prototype.
10818 (avr_out_plus): Add `out_label' argument.
10819 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
10820 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
10821 Handle the case where `insn' is just a pattern.
10822 (avr_out_bitop): Handle the case where `insn' is just a pattern.
10823 (avr_out_round): New function.
10824 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
10826 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
10828 * config/microblaze/microblaze.c (microblaze_expand_prologue):
10829 Add check for flag_stack_usage to handle -fstack-usage support
10831 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
10833 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
10834 interaction for new Power8 flags and VSX.
10836 2013-07-18 Sriraman Tallam <tmsriram@google.com>
10838 PR middle-end/57698
10839 * tree-inline.c (expand_call_inline): Emit errors during
10840 early_inlining only if optimization is not turned on.
10842 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10846 * passes.c (init_optimization_passes): Move the construction of
10847 the pass hierarchy into a new passes.def file.
10849 * Makefile.in (passes.o): Add dependency on passes.def.
10851 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10853 * passes.c (init_optimization_passes): Introduce macros for
10854 constructing the tree of passes (INSERT_PASSES_AFTER,
10855 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
10856 TERMINATE_PASS_LIST).
10858 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
10859 Wei Mi <wmi@google.com>
10861 PR rtl-optimization/57878
10862 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
10864 (reload_pseudo_compare_func): Check nregs first for reload
10867 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10869 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
10871 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
10873 * read-rtl.c (validate_const_int): Once an invalid character is
10874 seen, quit the loop.
10876 * gengtype.c (write_roots): Similarly once we find the "deletable"
10877 or "if_marked" option.
10879 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10881 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
10882 "xtab" and "sat". Move value "clz" from here to ...
10883 (attriubte "type"): ... here.
10884 (satsi_<SAT:code>): Delete "insn" attribute.
10885 (satsi_<SAT:code>_shift): Likewise.
10886 (arm_zero_extendqisi2addsi): Likewise.
10887 (arm_extendqisi2addsi): Likewise.
10888 (clzsi2): Update for attribute changes.
10889 (rbitsi2): Likewise.
10890 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
10892 (arm_usatsihi): Likewise.
10893 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10895 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10897 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
10898 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
10899 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
10900 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
10901 in alphabetical order.
10902 (attribute "core_cycles"): Update for attribute changes.
10903 (arm_addsi3): Likewise.
10904 (addsi3_compare0): Likewise.
10905 (addsi3_compare0_scratch): Likewise.
10906 (addsi3_compare_op1): Likewise.
10907 (addsi3_compare_op2): Likewise.
10908 (compare_addsi2_op0): Likewise.
10909 (compare_addsi2_op1): Likewise.
10910 (addsi3_carryin_shift_<optab>): Likewise.
10911 (subsi3_carryin_shift): Likewise.
10912 (rsbsi3_carryin_shift): Likewise.
10913 (arm_subsi3_insn): Likewise.
10914 (subsi3_compare0): Likewise.
10915 (subsi3_compare): Likewise.
10916 (arm_andsi3_insn): Likewise.
10917 (thumb1_andsi3_insn): Likewise.
10918 (andsi3_compare0): Likewise.
10919 (andsi3_compare0_scratch): Likewise.
10920 (zeroextractsi_compare0_scratch
10921 (andsi_not_shiftsi_si): Likewise.
10922 (iorsi3_insn): Likewise.
10923 (iorsi3_compare0): Likewise.
10924 (iorsi3_compare0_scratch): Likewise.
10925 (arm_xorsi3): Likewise.
10926 (thumb1_xorsi3_insn): Likewise.
10927 (xorsi3_compare0): Likewise.
10928 (xorsi3_compare0_scratch): Likewise.
10929 (satsi_<SAT:code>_shift): Likewise.
10931 (arm_shiftsi3): Likewise.
10932 (shiftsi3_compare0): Likewise.
10933 (not_shiftsi): Likewise.
10934 (not_shiftsi_compare0): Likewise.
10935 (not_shiftsi_compare0_scratch): Likewise.
10936 (arm_one_cmplsi2): Likewise.
10937 (thumb_one_complsi2): Likewise.
10938 (notsi_compare0): Likewise.
10939 (notsi_compare0_scratch): Likewise.
10940 (thumb1_zero_extendhisi2): Likewise.
10941 (arm_zero_extendhisi2): Likewise.
10942 (arm_zero_extendhisi2_v6): Likewise.
10943 (arm_zero_extendhisi2addsi): Likewise.
10944 (thumb1_zero_extendqisi2): Likewise.
10945 (thumb1_zero_extendqisi2_v6): Likewise.
10946 (arm_zero_extendqisi2): Likewise.
10947 (arm_zero_extendqisi2_v6): Likewise.
10948 (arm_zero_extendqisi2addsi): Likewise.
10949 (thumb1_extendhisi2): Likewise.
10950 (arm_extendhisi2): Likewise.
10951 (arm_extendhisi2_v6): Likewise.
10952 (arm_extendqisi): Likewise.
10953 (arm_extendqisi_v6): Likewise.
10954 (arm_extendqisi2addsi): Likewise.
10955 (thumb1_extendqisi2): Likewise.
10956 (thumb1_movdi_insn): Likewise.
10957 (arm_movsi_insn): Likewise.
10958 (movsi_compare0): Likewise.
10959 (movhi_insn_arch4): Likewise.
10960 (movhi_bytes): Likewise.
10961 (arm_movqi_insn): Likewise.
10962 (thumb1_movqi_insn): Likewise.
10963 (arm32_movhf): Likewise.
10964 (thumb1_movhf): Likewise.
10965 (arm_movsf_soft_insn): Likewise.
10966 (thumb1_movsf_insn): Likewise.
10967 (movdf_soft_insn): Likewise.
10968 (thumb_movdf_insn): Likewise.
10969 (arm_cmpsi_insn): Likewise.
10970 (cmpsi_shiftsi): Likewise.
10971 (cmpsi_shiftsi_swp): Likewise.
10972 (arm_cmpsi_negshiftsi_si): Likewise.
10973 (movsicc_insn): Likewise.
10974 (movsfcc_soft_insn): Likewise.
10975 (arith_shiftsi): Likewise.
10976 (arith_shiftsi_compare0
10977 (arith_shiftsi_compare0_scratch
10978 (sub_shiftsi): Likewise.
10979 (sub_shiftsi_compare0
10980 (sub_shiftsi_compare0_scratch
10981 (and_scc): Likewise.
10982 (cond_move): Likewise.
10983 (if_plus_move): Likewise.
10984 (if_move_plus): Likewise.
10985 (if_move_not): Likewise.
10986 (if_not_move): Likewise.
10987 (if_shift_move): Likewise.
10988 (if_move_shift): Likewise.
10989 (if_shift_shift): Likewise.
10990 (if_not_arith): Likewise.
10991 (if_arith_not): Likewise.
10992 (cond_move_not): Likewise.
10993 (thumb1_ashlsi3): Set type attribute.
10994 (thumb1_ashrsi3): Likewise.
10995 (thumb1_lshrsi3): Likewise.
10996 (thumb1_rotrsi3): Likewise.
10997 (shiftsi3_compare0_scratch): Likewise.
10998 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
10999 (neon_mov<mode>): Likewise.
11000 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
11002 (thumb2_movsi_insn): Likewise.
11003 (thumb2_cmpsi_neg_shiftsi): Likewise.
11004 (thumb2_extendqisi_v6): Likewise.
11005 (thumb2_zero_extendhisi2_v6): Likewise.
11006 (thumb2_zero_extendqisi2_v6): Likewise.
11007 (thumb2_shiftsi3_short): Likewise.
11008 (thumb2_addsi3_compare0_scratch): Likewise.
11009 (orsi_not_shiftsi_si): Likewise.
11010 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
11011 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
11013 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11014 (1020alu_shift_op): Likewise.
11015 (1020alu_shift_reg_op): Likewise.
11016 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11017 (alu_shift_op): Likewise.
11018 (alu_shift_reg_op): Likewise.
11019 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11020 (11_alu_shift_op): Likewise.
11021 (11_alu_shift_reg_op): Likewise.
11022 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11023 (9_alu_shift_reg_op): Likewise.
11024 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11026 (cortex_a15_alu_shift): Likewise.
11027 (cortex_a15_alu_shift_reg): Likewise.
11028 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
11030 (cortex_a5_alu_shift): Likewise.
11031 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
11033 (cortex_a53_alu_shift): Likewise.
11034 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11036 (cortex_a7_alu_reg): Likewise.
11037 (cortex_a7_alu_shift): Likewise.
11038 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
11040 (cortex_a8_alu_shift): Likewise.
11041 (cortex_a8_alu_shift_reg): Likewise.
11042 (cortex_a8_mov): Likewise.
11043 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11044 (cortex_a9_dp_shift): Likewise.
11045 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
11047 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
11049 (cortex_r4_mov): Likewise.
11050 (cortex_r4_alu_shift): Likewise.
11051 (cortex_r4_alu_shift_reg): Likewise.
11052 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11053 (526_alu_shift_op): Likewise.
11054 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11055 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11056 (626te_alu_shift_op): Likewise.
11057 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
11058 (726te_alu_op): Likewise.
11059 (726te_alu_shift_op): Likewise.
11060 (726te_alu_shift_reg_op): Likewise.
11061 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11062 (mp626_alu_shift_op): Likewise.
11063 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
11064 (pj4_alu_e1_conds): Likewise.
11065 (pj4_alu): Likewise.
11066 (pj4_alu_conds): Likewise.
11067 (pj4_shift): Likewise.
11068 (pj4_shift_conds): Likewise.
11069 (pj4_alu_shift): Likewise.
11070 (pj4_alu_shift_conds): Likewise.
11071 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11073 (cortexa7_older_only): Likewise.
11074 (cortexa7_younger): Likewise.
11076 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11078 * ipa-pure-const.c (generate_summary): Rename to...
11079 (pure_const_generate_summary): ... this.
11081 2013-07-17 Iain Sandoe <iain@codesourcery.com>
11083 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
11085 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
11088 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
11091 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11093 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
11094 enabled without -march=zEC12.
11095 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
11098 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11100 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
11101 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
11102 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
11103 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
11104 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
11105 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
11106 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
11108 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
11109 and ISA_HAS_NMADD3_NMSUB3.
11110 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
11111 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
11112 (nmsub4<mode>, nmsub3<mode>): Likewise.
11113 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
11115 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11117 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
11118 TARGET_MIPS5400 checking.
11120 2013-07-16 Jakub Jelinek <jakub@redhat.com>
11121 Peter Bergner <bergner@vnet.ibm.com>
11123 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
11124 registers in the comment.
11125 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
11126 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
11127 rather than FIRST_PSEUDO_REGISTERS.
11129 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
11131 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
11132 enable extra ISA flags with TARGET_HTM.
11134 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11136 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
11139 2013-07-15 Cong Hou <congh@google.com>
11141 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
11142 in compare function for sorting.
11144 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
11146 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
11147 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
11148 * config/rs6000/rs6000.opt: Add -mhtm option.
11149 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
11150 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
11151 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11152 __HTM__ if the HTM instructions are available.
11153 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
11154 htm_spr_reg_operand): New define_predicates.
11155 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
11156 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
11158 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
11159 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
11160 HTM builtin functions.
11161 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
11162 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
11163 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
11164 (rs6000_builtin_mask_calculate): Likewise.
11165 (rs6000_option_override_internal): Likewise.
11166 (bdesc_htm): Add new HTM builtin support.
11167 (htm_spr_num): New function.
11168 (htm_spr_regno): Likewise.
11169 (rs6000_htm_spr_icode): Likewise.
11170 (htm_expand_builtin): Likewise.
11171 (htm_init_builtins): Likewise.
11172 (rs6000_expand_builtin): Add support for HTM builtin functions.
11173 (rs6000_init_builtins): Likewise.
11174 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
11176 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
11177 (TARGET_HTM, MASK_HTM): Define macros.
11178 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
11179 (FIXED_REGISTERS): Likewise.
11180 (CALL_USED_REGISTERS): Likewise.
11181 (CALL_REALLY_USED_REGISTERS): Likewise.
11182 (REG_ALLOC_ORDER): Likewise.
11183 (enum reg_class): Likewise.
11184 (REG_CLASS_NAMES): Likewise.
11185 (REG_CLASS_CONTENTS): Likewise.
11186 (REGISTER_NAMES): Likewise.
11187 (ADDITIONAL_REGISTER_NAMES): Likewise.
11188 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
11189 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
11190 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
11191 * config/rs6000/htm.md: New file.
11192 * config/rs6000/htmintrin.h: New file.
11193 * config/rs6000/htmxlintrin.h: New file.
11195 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
11197 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
11198 Define SYMBOL_TINY_GOT, update comment.
11199 * config/aarch64/aarch64.c
11200 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
11201 (aarch64_expand_mov_immediate): Likewise.
11202 (aarch64_print_operand): Likewise.
11203 (aarch64_classify_symbol): Likewise.
11204 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
11205 (ldr_got_tiny): Define.
11207 2013-07-13 Tobias Grosser <tobias@grosser.es>
11209 PR tree-optimization/54094
11210 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
11211 scheduling dimension for the parallelism check from the polyhedral
11212 information in the AST.
11213 * graphite-dependences.c (carries_deps): Do not assume the schedule is
11216 2013-07-13 Jason Merrill <jason@redhat.com>
11218 * print-tree.c (debug_vec_tree): Use debug_raw.
11219 (debug_raw (vec<tree, va_gc> &)): New.
11220 (debug_raw (vec<tree, va_gc> *)): New.
11221 * tree.h: Declare them.
11223 2013-07-13 Bin Cheng <bin.cheng@arm.com>
11225 * ifcvt.c (ifcvt_after_combine): New static variable.
11226 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
11228 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
11229 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
11230 rest_of_handle_if_after_reload): Pass new argument for if_convert.
11232 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
11234 * config/mips/mips.c (mips_expand_call): Remove empty statement.
11236 2013-07-12 Michael Matz <matz@suse.de>
11238 PR middle-end/55771
11239 * convert.c (convert_to_real): Reject non-float inner types.
11241 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
11243 * config/aarch64/aarch64-protos.h
11244 (aarch64_simd_immediate_valid_for_move): Remove.
11245 * config/aarch64/aarch64.c (simd_immediate_info): New member.
11246 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
11248 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
11250 2013-07-11 Steve Ellcey <sellcey@mips.com>
11252 * config/mips/mips.c (mips_conditional_register_usage): Do not
11253 use t[0-7] registers in MIPS16 mode when optimizing for size.
11255 2013-07-11 Sriraman Tallam <tmsriram@google.com>
11257 * config/i386/i386.c (dispatch_function_versions): Fix array
11258 indexing of function_version_info to match actual_versions.
11260 2013-07-11 Teresa Johnson <tejohnson@google.com>
11262 * vec.h (struct va_gc): Move release out-of-line.
11263 (va_gc::release): Call ggc_free on released vec.
11265 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11267 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
11268 Require GOT register as additional operand in UNSPEC.
11269 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
11270 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
11271 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
11272 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
11273 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
11274 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
11275 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
11277 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
11280 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
11281 name seen by assembler/linker rather if available.
11283 2013-07-11 Andreas Schwab <schwab@suse.de>
11285 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
11287 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
11289 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
11291 2013-07-10 Joseph Myers <joseph@codesourcery.com>
11293 * doc/tm.texi.in: Move hook documentation to ....
11294 * target.def: ... here.
11296 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
11297 text on @hook line.
11298 * doc/tm.texi: Regenerate.
11300 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
11303 * doc/invoke.texi: Document -Wconditionally-supported.
11305 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
11308 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
11311 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
11314 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
11315 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
11316 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
11317 Remove duplicate devices.
11318 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
11319 * config/avr/t-multilib: Regenerate.
11320 * config/avr/avr-tables.opt: Regenerate.
11321 * doc/avr-mmcu.texi: Regenerate.
11323 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
11326 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
11328 2013-07-10 Graham Stott <graham.stott@btinternet.com>
11330 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
11331 the cost of MULT when optimizing for size.
11333 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11335 * config/cr16/cr16-protos.h: Don't include target.h.
11337 2013-07-09 Joseph Myers <joseph@codesourcery.com>
11339 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
11340 adjust register size for TDmode and TFmode for VSX registers.
11342 2013-07-08 Kai Tietz <ktietz@redhat.com>
11345 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
11346 hook_bool_const_tree_true.
11348 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11350 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
11351 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
11352 * config/s390/s390.md: Define FPR*_REGNUM constants.
11353 Fix FPR2_REGNUM constant (18 -> 17).
11354 ("*trunc<BFP:mode><DFP_ALL:mode>2")
11355 ("*trunc<DFP_ALL:mode><BFP:mode>2")
11356 ("trunc<BFP:mode><DFP_ALL:mode>2")
11357 ("trunc<DFP_ALL:mode><BFP:mode>2")
11358 ("*extend<BFP:mode><DFP_ALL:mode>2")
11359 ("*extend<DFP_ALL:mode><BFP:mode>2")
11360 ("extend<BFP:mode><DFP_ALL:mode>2")
11361 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
11364 2013-07-08 Graham Stott <graham.stott@btinternet.com>
11366 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
11368 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11370 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
11371 and cfun_fpr_bit_p to cfun_fpr_save_p.
11372 (s390_frame_area, s390_register_info, s390_frame_info)
11373 (s390_emit_prologue, s390_emit_epilogue)
11374 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
11376 * config/s390/s390.h: Define *_REGNUM macros for floating point
11379 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
11381 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
11383 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
11385 PR rtl-optimization/57786
11386 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
11387 and break out of the loop when it is set to false.
11389 2013-07-08 Jakub Jelinek <jakub@redhat.com>
11392 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
11393 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
11394 (const_int 63)) 0)).
11395 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
11396 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
11397 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
11399 PR rtl-optimization/57829
11400 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
11401 mask bits outside of mode are just sign-extension from mode to HWI.
11403 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
11405 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
11406 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
11407 adjust_address instead of change_address to keep info about alignment.
11408 (emit_strmov): Remove.
11409 (emit_memmov): New function.
11410 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
11411 (expand_movmem_epilogue): Likewise and return updated rtx for
11413 (expand_constant_movmem_prologue): Likewise and return updated rtx for
11414 destination and source.
11415 (decide_alignment): Refactor, handle vector_loop.
11416 (ix86_expand_movmem): Likewise.
11417 (ix86_expand_setmem): Likewise.
11418 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
11420 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
11422 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
11423 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
11425 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
11427 * config/i386/sse.md (sse_movlhps): Change alternative 3
11428 of operand 2 to "m".
11430 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
11433 * config/i386/sse.md (iptr): New mode attribute.
11434 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
11435 (<sse>_vm<plusminus_insn><mode>3): Ditto.
11436 (<sse>_vmmul<mode>3): Ditto.
11437 (<sse>_vmdiv<mode>3): Ditto.
11438 (sse_vmrcpv4sf2): Ditto.
11439 (<sse>_vmsqrt<mode>2): Ditto.
11440 (sse_vmrsqrtv4sf2): Ditto.
11441 (<sse>_vm<code><mode>3): Ditto.
11442 (avx_vmcmp<mode>3): Ditto.
11443 (<sse>_vmmaskcmp<mode>3): Ditto.
11444 (<sse>_comi): Ditto.
11445 (<sse>_ucomi): Ditto.
11446 (*xop_vmfrcz_<mode>): Ditto.
11447 (*fmai_fmadd_<mode>): Ditto.
11448 (*fmai_fmsub_<mode>): Ditto.
11449 (*fmai_fnmadd_<mode>): Ditto.
11450 (*fmai_fnmsub_<mode>): Ditto.
11451 (*fma4i_vmfmadd_<mode>): Ditto.
11452 (*fma4i_vmfmsub_<mode>): Ditto.
11453 (*fma4i_vmfnmadd_<mode>): Ditto.
11454 (*fma4i_vmfnmsub_<mode>): Ditto.
11455 (*xop_vmfrcz_<mode>): Ditto.
11456 (sse_cvtps2pi): Ditto.
11457 (sse_cvttps2pi): Ditto.
11458 (sse_cvtss2si): Ditto.
11459 (sse_cvtss2si_2): Ditto.
11460 (sse_cvtss2siq_2): Ditto.
11461 (sse_cvttss2si): Ditto.
11462 (sse_cvttss2siq): Ditto.
11463 (sse_cvtsd2si): Ditto.
11464 (sse_cvtsd2si_2): Ditto.
11465 (sse_cvtsd2siq_2): Ditto.
11466 (sse_cvttsd2si): Ditto.
11467 (sse_cvttsd2siq): Ditto.
11468 (sse_cvtsd2ss): Ditto.
11469 (sse_cvtss2sd): Ditto.
11470 (avx2_pbroadcast<mode>): Ditto.
11471 (avx2_pbroadcast<mode>_1): Ditto.
11472 (*avx_vperm_broadcast_v4sf): Ditto.
11474 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
11475 (sse_movlhps): Ditto.
11476 (sse_storehps): Ditto.
11477 (sse_loadhps): Ditto.
11478 (sse_storelps): Ditto.
11479 (sse_loadlps): Ditto.
11480 (*vec_concatv4sf): Ditto.
11481 (*vec_interleave_highv2df): Ditto.
11482 (*vec_interleave_lowv2df): Ditto.
11483 (*vec_extractv2df_1_sse): Ditto.
11484 (*vec_extractv2df_0_sse): Ditto.
11485 (sse2_storelpd): Ditto.
11486 (sse2_loadlpd): Ditto.
11487 (sse2_movsd): Ditto.
11488 (*vec_concatv4si): Ditto.
11489 (vec_concatv2di): Ditto.
11491 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
11492 for Intel asm dialect.
11493 (mmx_punpcklwd): Ditto.
11494 (mmx_punpckldq): Ditto.
11496 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
11497 for intel assembler dialect.
11499 2013-07-06 Jakub Jelinek <jakub@redhat.com>
11502 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
11503 for BUILT_IN_C{LZ,LRSB}*.
11504 * tree.h (CASE_INT_FN): Add FN##IMAX case.
11505 * tree-vrp.c (extract_range_basic): Handle
11506 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
11507 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
11508 fall thru to code calling set_value*.
11509 * builtins.c (expand_builtin): Remove *IMAX cases.
11510 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
11511 if width is bigger than 2*HWI.
11513 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
11515 PR rtl-optimization/55342
11516 * lra-int.h (lra_subreg_reload_pseudos): New.
11517 * lra.c: Add undoing optional reloads to the block diagram.
11518 (lra_subreg_reload_pseudos): New.
11519 (lra_optional_reload_pseudos): Change comments.
11520 (lra): Init and clear lra_subreg_reload_pseudos. Clear
11521 lra_optional_reload_pseudos after undo transformations.
11522 * lra-assigns.c (pseudo_prefix_title): New.
11523 (lra_setup_reg_renumber): Use it.
11524 (spill_for): Ditto. Check subreg reload pseudos too.
11525 (assign_by_spills): Consider subreg reload pseudos too.
11526 * lra-constraints.c (simplify_operand_subreg): Use
11527 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
11528 (curr_insn_transform): Recognize and do optional reloads.
11529 (undo_optional_reloads): New.
11530 (lra_undo_inheritance): Call undo_optional_reloads.
11532 2013-07-05 Thomas Quinot <quinot@adacore.com>
11534 * tree-complex.c (expand_complex_operations_1): Fix typo.
11536 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
11538 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
11539 (tune_params): New member 'const vec_costs'.
11540 * config/aarch64/aarch64.c (generic_vector_cost): New.
11541 (generic_tunings): New member 'generic_vector_cost'.
11542 (aarch64_builtin_vectorization_cost): New.
11543 (aarch64_add_stmt_cost): New.
11544 (TARGET_VECTORIZE_ADD_STMT_COST): New.
11545 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
11547 2013-07-03 Jakub Jelinek <jakub@redhat.com>
11550 * config/i386/predicates.md (vsib_address_operand): Disallow
11551 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
11553 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
11555 PR middle-end/55030
11556 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
11557 expand_builtin_setjmp_receiver.
11558 (expand_label): Adjust, call expand_builtin_setjmp_receiver
11559 with NULL for the label parameter.
11560 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
11561 the frame-pointer. Adjust comments.
11562 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
11563 only if LABEL is non-NULL.
11565 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
11567 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
11568 (ARM_ABI_AAPCS64): Ditto.
11570 (ARM_DEFAULT_ABI): Ditto.
11572 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
11574 * config/aarch64/aarch64-builtins.c
11575 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
11576 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
11578 * config/aarch64/aarch64-simd.md
11579 (aarch64_ld1<VALL:mode>): New.
11580 (aarch64_st1<VALL:mode>): Likewise.
11581 * config/aarch64/arm_neon.h
11582 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
11584 2013-07-02 Sriraman Tallam <tmsriram@google.com>
11586 * config/i386/i386.c (gate_insert_vzeroupper): Check if
11588 (ix86_option_override_internal):Turn on all -mavx target flags by
11589 default as they are dependent on AVX anyway.
11591 2013-07-02 Cary Coutant <ccoutant@google.com>
11593 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
11594 deterministic hash.
11595 (loc_checksum_ordered): Likewise.
11596 (hash_loc_operands): Remove inline keyword.
11598 2013-07-02 Jakub Jelinek <jakub@redhat.com>
11600 PR tree-optimization/57741
11601 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
11602 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
11603 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
11604 Allow REAL_CST step_exprs if flag_associative_math.
11605 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
11607 2013-07-02 Ian Bolton <ian.bolton@arm.com>
11609 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
11611 2013-07-02 Ian Bolton <ian.bolton@arm.com>
11613 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
11615 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11617 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
11619 (iorsi3_insn): Likewise.
11620 (arm_xorsi3): Likewise.
11622 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
11624 * arm.md (attribute "wtype"): Delete. Move attribute values from here
11626 (attribute "type"): ... here, and prefix with "wmmx_".
11627 (attribute "core_cycles"): Update for attribute changes.
11628 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
11629 (tbcstv4hi): Likewise.
11630 (tbcstv2si): Likewise.
11631 (iwmmxt_iordi3): Likewise.
11632 (iwmmxt_xordi3): Likewise.
11633 (iwmmxt_anddi3): Likewise.
11634 (iwmmxt_nanddi3): Likewise.
11635 (iwmmxt_arm_movdi): Likewise.
11636 (iwmmxt_movsi_insn): Likewise.
11637 (mov<mode>_internal): Likewise.
11638 (and<mode>3_iwmmxt): Likewise.
11639 (ior<mode>3_iwmmxt): Likewise.
11640 (xor<mode>3_iwmmxt): Likewise.
11641 (add<mode>3_iwmmxt): Likewise.
11642 (ssaddv8qi3): Likewise.
11643 (ssaddv4hi3): Likewise.
11644 (ssaddv2si3): Likewise.
11645 (usaddv8qi3): Likewise.
11646 (usaddv4hi3): Likewise.
11647 (usaddv2si3): Likewise.
11648 (sub<mode>3_iwmmxt): Likewise.
11649 (sssubv8qi3): Likewise.
11650 (sssubv4hi3): Likewise.
11651 (sssubv2si3): Likewise.
11652 (ussubv8qi3): Likewise.
11653 (ussubv4hi3): Likewise.
11654 (ussubv2si3): Likewise.
11655 (mulv4hi3_iwmmxt): Likewise.
11656 (smulv4hi3_highpart): Likewise.
11657 (umulv4hi3_highpart): Likewise.
11658 (iwmmxt_wmacs): Likewise.
11659 (iwmmxt_wmacsz): Likewise.
11660 (iwmmxt_wmacu): Likewise.
11661 (iwmmxt_wmacuz): Likewise.
11662 (iwmmxt_clrdi): Likewise.
11663 (iwmmxt_clrv8qi): Likewise.
11664 (iwmmxt_clr4hi): Likewise.
11665 (iwmmxt_clr2si): Likewise.
11666 (iwmmxt_uavgrndv8qi3): Likewise.
11667 (iwmmxt_uavgrndv4hi3): Likewise.
11668 (iwmmxt_uavgv8qi3): Likewise.
11669 (iwmmxt_uavgv4hi3): Likewise.
11670 (iwmmxt_tinsrb): Likewise.
11671 (iwmmxt_tinsrh): Likewise.
11672 (iwmmxt_tinsrw): Likewise.
11673 (iwmmxt_textrmub): Likewise.
11674 (iwmmxt_textrmsb): Likewise.
11675 (iwmmxt_textrmuh): Likewise.
11676 (iwmmxt_textrmsh): Likewise.
11677 (iwmmxt_textrmw): Likewise.
11678 (iwmxxt_wshufh): Likewise.
11679 (eqv8qi3): Likewise.
11680 (eqv4hi3): Likewise.
11681 (eqv2si3): Likewise.
11682 (gtuv8qi3): Likewise.
11683 (gtuv4hi3): Likewise.
11684 (gtuv2si3): Likewise.
11685 (gtv8qi3): Likewise.
11686 (gtv4hi3): Likewise.
11687 (gtv2si3): Likewise.
11688 (smax<mode>3_iwmmxt): Likewise.
11689 (umax<mode>3_iwmmxt): Likewise.
11690 (smin<mode>3_iwmmxt): Likewise.
11691 (umin<mode>3_iwmmxt): Likewise.
11692 (iwmmxt_wpackhss): Likewise.
11693 (iwmmxt_wpackwss): Likewise.
11694 (iwmmxt_wpackdss): Likewise.
11695 (iwmmxt_wpackhus): Likewise.
11696 (iwmmxt_wpackwus): Likewise.
11697 (iwmmxt_wpackdus): Likewise.
11698 (iwmmxt_wunpckihb): Likewise.
11699 (iwmmxt_wunpckihh): Likewise.
11700 (iwmmxt_wunpckihw): Likewise.
11701 (iwmmxt_wunpckilb): Likewise.
11702 (iwmmxt_wunpckilh): Likewise.
11703 (iwmmxt_wunpckilw): Likewise.
11704 (iwmmxt_wunpckehub): Likewise.
11705 (iwmmxt_wunpckehuh): Likewise.
11706 (iwmmxt_wunpckehuw): Likewise.
11707 (iwmmxt_wunpckehsb): Likewise.
11708 (iwmmxt_wunpckehsh): Likewise.
11709 (iwmmxt_wunpckehsw): Likewise.
11710 (iwmmxt_wunpckelub): Likewise.
11711 (iwmmxt_wunpckeluh): Likewise.
11712 (iwmmxt_wunpckeluw): Likewise.
11713 (iwmmxt_wunpckelsb): Likewise.
11714 (iwmmxt_wunpckelsh): Likewise.
11715 (iwmmxt_wunpckelsw): Likewise.
11716 (ror<mode>3): Likewise.
11717 (ashr<mode>3_iwmmxt): Likewise.
11718 (lshr<mode>3_iwmmxt): Likewise.
11719 (ashl<mode>3_iwmmxt): Likewise.
11720 (ror<mode>3_di): Likewise.
11721 (ashr<mode>3_di): Likewise.
11722 (lshr<mode>3_di): Likewise.
11723 (ashl<mode>3_di): Likewise.
11724 (iwmmxt_wmadds): Likewise.
11725 (iwmmxt_wmaddu): Likewise.
11726 (iwmmxt_tmia): Likewise.
11727 (iwmmxt_tmiaph): Likewise.
11728 (iwmmxt_tmiabb): Likewise.
11729 (iwmmxt_tmiatb): Likewise.
11730 (iwmmxt_tmiabt): Likewise.
11731 (iwmmxt_tmiatt): Likewise.
11732 (iwmmxt_tmovmskb): Likewise.
11733 (iwmmxt_tmovmskh): Likewise.
11734 (iwmmxt_tmovmskw): Likewise.
11735 (iwmmxt_waccb): Likewise.
11736 (iwmmxt_wacch): Likewise.
11737 (iwmmxt_waccw): Likewise.
11738 (iwmmxt_waligni): Likewise.
11739 (iwmmxt_walignr): Likewise.
11740 (iwmmxt_walignr0): Likewise.
11741 (iwmmxt_walignr1): Likewise.
11742 (iwmmxt_walignr2): Likewise.
11743 (iwmmxt_walignr3): Likewise.
11744 (iwmmxt_wsadb): Likewise.
11745 (iwmmxt_wsadh): Likewise.
11746 (iwmmxt_wsadbz): Likewise.
11747 (iwmmxt_wsadhz): Likewise.
11748 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
11749 (iwmmxt_wabsdiffb): Likewise.
11750 (iwmmxt_wabsdiffh): Likewise.
11751 (iwmmxt_wabsdiffw): Likewise.
11752 (iwmmxt_waddsubhx): Likewise
11753 (iwmmxt_wsubaddhx): Likewise.
11754 (addc<mode>3): Likewise.
11755 (iwmmxt_avg4): Likewise.
11756 (iwmmxt_avg4r): Likewise.
11757 (iwmmxt_wmaddsx): Likewise.
11758 (iwmmxt_wmaddux): Likewise.
11759 (iwmmxt_wmaddsn): Likewise.
11760 (iwmmxt_wmaddun): Likewise.
11761 (iwmmxt_wmulwsm): Likewise.
11762 (iwmmxt_wmulwum): Likewise.
11763 (iwmmxt_wmulsmr): Likewise.
11764 (iwmmxt_wmulumr): Likewise.
11765 (iwmmxt_wmulwsmr): Likewise.
11766 (iwmmxt_wmulwumr): Likewise.
11767 (iwmmxt_wmulwl): Likewise.
11768 (iwmmxt_wqmulm): Likewise.
11769 (iwmmxt_wqmulwm): Likewise.
11770 (iwmmxt_wqmulmr): Likewise.
11771 (iwmmxt_wqmulwmr): Likewise.
11772 (iwmmxt_waddbhusm): Likewise.
11773 (iwmmxt_waddbhusl): Likewise.
11774 (iwmmxt_wqmiabb): Likewise.
11775 (iwmmxt_wqmiabt): Likewise.
11776 (iwmmxt_wqmiatb): Likewise.
11777 (iwmmxt_wqmiatt): Likewise.
11778 (iwmmxt_wqmiabbn): Likewise.
11779 (iwmmxt_wqmiabtn): Likewise.
11780 (iwmmxt_wqmiatbn): Likewise.
11781 (iwmmxt_wqmiattn): Likewise.
11782 (iwmmxt_wmiabb): Likewise.
11783 (iwmmxt_wmiabt): Likewise.
11784 (iwmmxt_wmiatb): Likewise.
11785 (iwmmxt_wmiatt): Likewise.
11786 (iwmmxt_wmiabbn): Likewise.
11787 (iwmmxt_wmiabtn): Likewise.
11788 (iwmmxt_wmiatbn): Likewise.
11789 (iwmmxt_wmiattn): Likewise.
11790 (iwmmxt_wmiawbb): Likewise.
11791 (iwmmxt_wmiawbt): Likewise.
11792 (iwmmxt_wmiawtb): Likewise.
11793 (iwmmxt_wmiawtt): Likewise.
11794 (iwmmxt_wmiawbbn): Likewise.
11795 (iwmmxt_wmiawbtn): Likewise.
11796 (iwmmxt_wmiawtbn): Likewise.
11797 (iwmmxt_wmiawttn): Likewise.
11798 (iwmmxt_wmerge): Likewise.
11799 (iwmmxt_tandc<mode>3): Likewise.
11800 (iwmmxt_torc<mode>3): Likewise.
11801 (iwmmxt_torvsc<mode>3): Likewise.
11802 (iwmmxt_textrc<mode>3): Likewise.
11803 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
11804 (wmmxt_pack): Likewise.
11805 (wmmxt_mult_c1): Likewise.
11806 (wmmxt_mult_c2): Likewise.
11807 (wmmxt_alu_c1): Likewise.
11808 (wmmxt_alu_c2): Likewise.
11809 (wmmxt_alu_c3): Likewise.
11810 (wmmxt_transfer_c1): Likewise.
11811 (wmmxt_transfer_c2): Likewise.
11812 (wmmxt_transfer_c3): Likewise.
11813 (marvell_f_iwmmxt_wstr): Likewise.
11814 (marvell_f_iwmmxt_wldr): Likewise.
11816 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
11818 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
11820 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11823 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11824 * lra-constraints.c (need_for_split_p): Check call used hard regs
11825 living through calls.
11827 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
11828 call used regs for call insn.
11830 2013-06-28 Jakub Jelinek <jakub@redhat.com>
11833 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
11834 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
11836 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11838 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
11840 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11842 * lra-constraints.c (need_for_split_p): Check call used hard regs
11843 living through calls.
11845 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
11848 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
11849 to tie with any other modes. Eliminate Altivec vector mode tests,
11850 since these are a subset of ALTIVEC or VSX vector modes. Simplify
11851 code, to return 0 if testing MODE2 for a condition, if we've
11852 already tested MODE1 for the same condition.
11854 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11856 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
11859 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11861 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
11862 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
11864 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11866 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
11868 (aarch64_symbolic_constant_p): Remove.
11869 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
11870 static. Fix line length and white space.
11871 (aarch64_symbolic_constant_p): Remove.
11872 * config/aarch64/predicates.md (aarch64_valid_symref):
11873 Use aarch64_classify_symbol_expression.
11875 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11877 * config/arm/constraints.md (Ts): New constraint.
11878 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
11880 (compare_scc): Use "Ts" constraint for operand 0.
11881 (ior_scc_scc): Likewise.
11882 (and_scc_scc): Likewise.
11883 (and_scc_scc_nodom): Likewise.
11884 (ior_scc_scc_cmp): Likewise for operand 7.
11885 (and_scc_scc_cmp): Likewise.
11886 * config/arm/thumb2.md (thumb2_movsi_insn):
11887 Add alternatives for 16-bit encodings.
11888 (thumb2_movhi_insn): Likewise.
11889 (thumb2_movsicc_insn): Likewise.
11890 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
11891 (thumb2_negscc): Use "Ts" constraint.
11892 Move mvn instruction outside cond_exec block.
11893 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
11894 for 16-bit encodings.
11896 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11898 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
11900 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
11901 (mulsi3subsi): Likewise.
11902 (mulsidi3adddi): Likewise.
11903 (mulsidi3_v6): Likewise.
11904 (umulsidi3_v6): Likewise.
11905 (umulsidi3adddi_v6): Likewise.
11906 (smulsi3_highpart_v6): Likewise.
11907 (umulsi3_highpart_v6): Likewise.
11908 (mulhisi3tb): Likewise.
11909 (mulhisi3bt): Likewise.
11910 (mulhisi3tt): Likewise.
11911 (maddhisi4): Likewise.
11912 (maddhisi4tb): Likewise.
11913 (maddhisi4tt): Likewise.
11914 (maddhidi4): Likewise.
11915 (maddhidi4tb): Likewise.
11916 (maddhidi4tt): Likewise.
11917 (zeroextractsi_compare0_scratch): Likewise.
11918 (insv_zero): Likewise.
11919 (insv_t2): Likewise.
11920 (anddi_notzesidi_di): Likewise.
11921 (anddi_notsesidi_di): Likewise.
11922 (andsi_notsi_si): Likewise.
11923 (iordi_zesidi_di): Likewise.
11924 (xordi_zesidi_di): Likewise.
11925 (andsi_iorsi3_notsi): Likewise.
11926 (smax_0): Likewise.
11927 (smax_m1): Likewise.
11928 (smin_0): Likewise.
11929 (not_shiftsi): Likewise.
11930 (unaligned_loadsi): Likewise.
11931 (unaligned_loadhis): Likewise.
11932 (unaligned_loadhiu): Likewise.
11933 (unaligned_storesi): Likewise.
11934 (unaligned_storehi): Likewise.
11935 (extv_reg): Likewise.
11936 (extzv_t2): Likewise.
11937 (divsi3): Likewise.
11938 (udivsi3): Likewise.
11939 (arm_zero_extendhisi2addsi): Likewise.
11940 (arm_zero_extendqisi2addsi): Likewise.
11941 (compareqi_eq0): Likewise.
11942 (arm_extendhisi2_v6): Likewise.
11943 (arm_extendqisi2addsi): Likewise.
11944 (arm_movt): Likewise.
11945 (thumb2_ldrd): Likewise.
11946 (thumb2_ldrd_base): Likewise.
11947 (thumb2_ldrd_base_neg): Likewise.
11948 (thumb2_strd): Likewise.
11949 (thumb2_strd_base): Likewise.
11950 (thumb2_strd_base_neg): Likewise.
11951 (arm_negsi2): Add alternative for 16-bit encoding.
11952 (arm_one_cmplsi2): Likewise.
11954 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11956 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
11957 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
11958 (movdfcc): Likewise.
11959 * config/arm/vfp.md (*thumb2_movsf_vfp):
11960 Disable predication for arm_restrict_it.
11961 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
11962 (*thumb2_movdfcc_vfp): Likewise.
11963 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
11964 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
11965 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
11966 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
11967 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
11968 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
11969 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
11970 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
11971 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
11972 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
11973 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
11974 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
11975 Disable predication for arm_restrict_it.
11977 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
11979 * config/i386/bmiintrin.h (_bextr_u32): New.
11980 (_bextr_u64): Ditto.
11982 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
11984 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
11985 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
11986 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
11987 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
11988 * config/mips/n32-elf.h: ...this new file.
11990 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
11993 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
11994 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
11996 2013-06-27 Catherine Moore <clm@codesourcery.com>
11998 * config/mips/mips-tables.opt: Regenerate.
11999 * config/mips/mips-cpus.def: Add m14ke and m14kec.
12000 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
12001 * doc/invoke.texi: Add -m14kc.
12003 2013-06-27 Jakub Jelinek <jakub@redhat.com>
12006 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
12007 constraints of operand 1 and 2.
12010 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
12011 to match RTL canonicalization. Swap predicates and
12012 constraints of operand 1 and 2.
12014 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
12016 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
12017 Process OP_INOUT regs for splitting too.
12019 2013-06-27 Jakub Jelinek <jakub@redhat.com>
12021 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
12022 decl before the loop, initialize to NULL.
12023 (vectorizable_load): Initialize ptr_incr to NULL.
12025 2013-06-27 Martin Jambor <mjambor@suse.cz>
12028 * ipa-ref.h (ipa_maybe_record_reference): Declare.
12029 * ipa-ref.c (ipa_maybe_record_reference): New function.
12030 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
12031 * ipa-cp.c (create_specialized_node): Record potential references from
12033 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
12035 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12037 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
12038 parameter 'mode' of type 'enum machine_mode mode'; change to pass
12039 'mode' to force_reg.
12040 (aarch64_add_offset): Update calls to aarch64_force_temporary.
12041 (aarch64_expand_mov_immediate): Likewise.
12043 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12045 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
12046 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
12048 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12050 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
12051 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12052 (struct machine_function): Add tbegin_p.
12053 (s390_canonicalize_comparison): Fold CC mode compares to
12054 conditional jump if possible.
12055 (s390_emit_jump): Return the emitted jump.
12056 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
12057 Handle CCRAWmode compares.
12058 (s390_option_override): Default to -mhtm if available.
12059 (s390_reg_clobbered_rtx): Handle floating point regs as well.
12060 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
12061 FPRs instead of df_regs_ever_live_p.
12062 (s390_optimize_nonescaping_tx): New function.
12063 (s390_init_frame_layout): Extend clobbered_regs array to cover
12065 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
12066 (s390_expand_tbegin): New function.
12067 (enum s390_builtin): New enum definition.
12068 (code_for_builtin): New array definition.
12069 (s390_init_builtins): New function.
12070 (s390_expand_builtin): New function.
12071 (TARGET_INIT_BUILTINS): Define.
12072 (TARGET_EXPAND_BUILTIN): Define.
12073 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
12074 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
12075 (s390_alc_comparison): Likewise.
12076 * config/s390/s390-modes.def: Add CCRAWmode.
12077 * config/s390/s390.h (processor_flags): Add PF_TX.
12078 (TARGET_CPU_HTM): Define macro.
12079 (TARGET_HTM): Define macro.
12080 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
12081 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
12082 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12083 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
12084 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
12085 (TBEGIN_MASK, TBEGINC_MASK): New constants.
12086 ("*cc_to_int"): Move up.
12087 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
12088 constants other than 0.
12089 ("*ccraw_to_int"): New insn and splitter definition.
12090 ("tbegin", "tbegin_nofloat", "tbegin_retry")
12091 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
12092 ("tx_assist"): New expander.
12093 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
12094 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
12095 * config/s390/s390.opt: Add -mhtm option.
12096 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
12097 * config/s390/htmxlintrin.h: New file.
12098 * config/s390/htmintrin.h: New file.
12099 * config/s390/s390intrin.h: New file.
12100 * doc/extend.texi: Document htm builtins.
12101 * config.gcc: Add the new header files to extra_headers.
12103 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
12105 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
12106 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
12108 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12109 Pat Haugen <pthaugen@us.ibm.com>
12110 Peter Bergner <bergner@vnet.ibm.com>
12112 * config/rs6000/power8.md: New.
12113 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
12114 setting for power8 entry.
12115 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
12116 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
12117 test for Power4/Power5 only.
12118 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
12120 (force_new_group): Adjust comment.
12121 * config/rs6000/rs6000.md: Include power8.md.
12123 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
12125 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
12126 * config/arm/arm-protos.h (arm_max_conditional_execute): New
12128 (tune_params): Update comment.
12129 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
12130 (arm_max_conditional_execute): New function.
12131 (thumb2_final_prescan_insn): Use max_insn_skipped and
12132 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
12134 2013-06-25 Jakub Jelinek <jakub@redhat.com>
12136 PR tree-optimization/57705
12137 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
12138 SSA_NAME step, provided that it is not defined inside the loop.
12139 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
12140 (get_initial_def_for_induction): Handle SSA_NAME IV step.
12142 2013-06-25 Martin Jambor <mjambor@suse.cz>
12144 PR middle-end/57670
12145 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
12146 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
12148 (ipa_note_param_call): Initialize member_ptr flag.
12149 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
12150 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
12151 (ipa_write_indirect_edge_info): Stream member_ptr flag.
12152 (ipa_read_indirect_edge_info): Likewise.
12154 2013-06-25 Richard Biener <rguenther@suse.de>
12156 PR middle-end/56977
12157 * passes.c (init_optimization_passes): Move pass_fold_builtins
12158 and pass_dce earlier with -Og.
12160 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
12162 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
12163 <BIT_FIELD_REF>: Remove trailing TAB.
12164 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
12167 2013-06-24 Martin Jambor <mjambor@suse.cz>
12169 PR tree-optimization/57358
12170 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
12171 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
12172 (ipa_analyze_params_uses): Generate pessimistic info when true.
12174 2013-06-24 Martin Jambor <mjambor@suse.cz>
12176 PR tree-optimization/57539
12177 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
12178 global.inlined_to of the new node to it. All callers changed.
12179 * ipa-inline-transform.c (clone_inlined_nodes): New variable
12180 inlining_into, pass it to cgraph_clone_node.
12181 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
12182 ipa_free_edge_args_substructures.
12183 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
12184 rdesc linked list. Do not assert rdesc edges have inlined caller.
12185 Assert we have found an rdesc in the rdesc list.
12187 2013-06-24 Richard Biener <rguenther@suse.de>
12189 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
12190 (pointer_set_lookup): Declare.
12191 (class pointer_map): New template class implementing a
12192 generic pointer to T map.
12193 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
12194 pointer_map<T>::contains, pointer_map<T>::insert,
12195 pointer_map<T>::traverse): New functions.
12196 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
12197 (pointer_set_lookup): New function.
12198 (pointer_set_contains): Use pointer_set_lookup.
12199 (pointer_set_insert): Likewise.
12200 (insert_aux): Remove.
12201 (struct pointer_map_t): Embed a pointer_set_t.
12202 (pointer_map_create): Adjust.
12203 (pointer_map_destroy): Likewise.
12204 (pointer_map_contains): Likewise.
12205 (pointer_map_insert): Likewise.
12206 (pointer_map_traverse): Likewise.
12207 * tree-streamer.h (struct streamer_tree_cache_d): Use a
12208 pointer_map<unsigned> instead of a pointer_map_t.
12209 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
12210 (streamer_tree_cache_lookup): Likewise.
12211 (streamer_tree_cache_create): Likewise.
12212 (streamer_tree_cache_delete): Likewise.
12213 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
12214 pointer_map<unsigned> instead of a pointer_map_t.
12215 (lto_init_tree_ref_encoder): Adjust.
12216 (lto_destroy_tree_ref_encoder): Likewise.
12217 * lto-section-out.c (lto_output_decl_index): Likewise.
12218 (lto_record_function_out_decl_state): Likewise.
12219 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
12221 2013-06-24 Richard Biener <rguenther@suse.de>
12223 PR tree-optimization/57488
12224 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
12226 2013-06-24 Alan Modra <amodra@gmail.com>
12228 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
12229 (gen_easy_altivec_constant): Likewise.
12230 * config/rs6000/predicates.md (easy_vector_constant_add_self,
12231 easy_vector_constant_msb): Likewise.
12233 2013-06-23 Jakub Jelinek <jakub@redhat.com>
12236 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
12237 add missing return true.
12239 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
12242 * config/sh/predicates.md (general_extend_operand): Invoke
12243 general_movsrc_operand for memory operands.
12244 (general_movsrc_operand): Allow reg+reg addressing, do not use
12245 general_operand for memory operands.
12247 2013-06-23 Sriraman Tallam <tmsriram@google.com>
12249 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
12250 when current target options does not apply.
12251 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
12252 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
12253 * config/i386/bmiintrin.h: Pass appropriate target
12254 attributes to header.
12255 * config/i386/mmintrin.h: Ditto.
12256 * config/i386/nmmintrin.h: Ditto.
12257 * config/i386/avx2intrin.h: Ditto.
12258 * config/i386/fxsrintrin.h: Ditto.
12259 * config/i386/tbmintrin.h: Ditto.
12260 * config/i386/xsaveintrin.h: Ditto.
12261 * config/i386/f16cintrin.h: Ditto.
12262 * config/i386/xtestintrin.h: Ditto.
12263 * config/i386/xsaveoptintrin.h: Ditto.
12264 * config/i386/bmi2intrin.h: Ditto.
12265 * config/i386/lzcntintrin.h: Ditto.
12266 * config/i386/smmintrin.h: Ditto.
12267 * config/i386/wmmintrin.h: Ditto.
12268 * config/i386/x86intrin.h: Remove all header include guards.
12269 * config/i386/prfchwintrin.h: Ditto.
12270 * config/i386/pmmintrin.h: Ditto.
12271 * config/i386/tmmintrin.h: Ditto.
12272 * config/i386/xmmintrin.h: Ditto.
12273 * config/i386/popcntintrin.h: Ditto.
12274 * config/i386/rdseedintrin.h: Ditto.
12275 * config/i386/ammintrin.h: Ditto.
12276 * config/i386/emmintrin.h: Ditto.
12277 * config/i386/immintrin.h: Remove all header include guards.
12278 * config/i386/fma4intrin.h: Ditto.
12279 * config/i386/lwpintrin.h: Ditto.
12280 * config/i386/xopintrin.h: Ditto.
12281 * config/i386/ia32intrin.h: Ditto.
12282 * config/i386/avxintrin.h: Ditto.
12283 * config/i386/rtmintrin.h: Ditto.
12284 * config/i386/fmaintrin.h: Ditto.
12285 * config/i386/mm3dnow.h: Ditto.
12287 2013-06-22 Sriraman Tallam <tmsriram@google.com>
12289 * common/config/i386/i386-common.c: Handle LZCNT.
12291 2013-06-22 Andi Kleen <ak@linux.intel.com>
12293 * doc/extend.texi: Use __atomic_store_n instead of
12294 __atomic_store in HLE example.
12296 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
12298 * config/sh/sh.c: Remove <cstdlib> workaround.
12300 2013-06-21 Andi Kleen <ak@linux.intel.com>
12302 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
12304 2013-06-21 Andi Kleen <ak@linux.intel.com>
12306 * doc/extend.texi: Document that __atomic_clear and
12307 __atomic_test_and_set should only be used with bool.
12309 2013-06-20 Jan Hubicka <jh@suse.cz>
12311 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
12312 types_same_for_odr.
12313 * tree.c (decls_same_for_odr): New function.
12314 (same_for_edr): New function.
12315 (types_same_for_odr): New function.
12316 (get_binfo_at_offset): Use it.
12317 * tree.h (types_same_for_odr): Declare.
12319 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
12320 Jason Merrill <jason@redhat.com>
12322 * system.h: Include <cstdlib> as well as <stdlib.h>.
12324 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
12327 * config/i386/i386.c (construct_container): Report error if
12328 long double is used with disabled x87 float returns.
12330 2013-06-20 Jan Hubicka <jh@suse.cz>
12332 * lto-cgraph.c (input_symtab): Do not set cgraph state.
12334 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
12336 PR rtl-optimization/57425
12337 PR rtl-optimization/57569
12338 * alias.c (write_dependence_p): Remove parameters mem_mode and
12339 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
12340 Changed all callers.
12341 (canon_anti_dependence): Get comments and semantics in sync.
12342 Add parameter mem_canonicalized. Changed all callers.
12343 * rtl.h (canon_anti_dependence): Update prototype.
12345 2013-06-20 Richard Biener <rguenther@suse.de>
12347 * data-streamer-in.c (streamer_read_uhwi): Optimize single
12348 byte case, inline streamer_read_uchar and defer section
12351 2013-06-20 Richard Biener <rguenther@suse.de>
12353 PR tree-optimization/57584
12354 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
12355 SSA names into the expanded expression that take part in
12356 abnormal coalescing.
12358 2013-06-19 Sharad Singhai <singhai@google.com>
12360 * gcov.c (print_usage): Handle new option.
12361 (process_args): Ditto.
12362 (get_gcov_intermediate_filename): New function.
12363 (output_intermediate_file): New function.
12364 (output_gcov_file): New function
12365 (generate_results): Handle new option.
12366 (release_function): Relase demangled name.
12367 (read_graph_file): Handle demangled name.
12368 (output_lines): Ditto.
12369 * doc/gcov.texi: Document gcov intermediate format.
12371 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
12374 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
12375 (lra_emit_add): Use the functions. Add comment about Y as an
12378 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
12381 * collect2.c (collect_atexit): New.
12382 (collect_exit): Delete.
12383 (main): Register collect_atexit with atexit.
12384 (collect_wait): Change collect_exit to exit.
12386 * collect2.h (collect_exit): Delete.
12387 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
12389 2013-06-19 Wei Mi <wmi@google.com>
12391 PR rtl-optimization/57518
12392 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
12393 if regno is used in paradoxical subreg.
12394 (update_equiv_regs): Check pdx_subregs[regno] before
12395 set a reg to be equivalent with a mem.
12397 2013-06-19 Matthias Klose <doko@ubuntu.com>
12400 * file-find.h (find_a_file): Add a mode parameter.
12401 * file-find.c (find_a_file): Likewise.
12402 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
12403 with X_OK for the executables.
12404 * collect2.c (main): Call find_a_file with X_OK.
12406 2013-06-19 Steve Ellcey <sellcey@mips.com>
12409 * config/mips/mips.md (casesi_internal_mips16_<mode>):
12410 Use NEXT_INSN instead of next_real_insn.
12412 2013-06-19 Jan Hubicka <jh@suse.cz>
12414 * cgraph.h (const_value_known_p): Replace by ...
12415 (ctor_for_folding): .. this one.
12416 * cgraphunit.c (process_function_and_variable_attributes): Use it.
12417 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
12418 * expr.c (expand_expr_real_1): Likewise.
12419 (string_constant): Likewise.
12420 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
12421 * ipa.c (process_references): Likewise.
12422 (symtab_remove_unreachable_nodes): Likewise.
12423 * ipa-inline-analysis.c (param_change_prob): Likewise.
12424 * gimple-fold.c (canonicalize_constructor_val): Likewise.
12425 (get_base_constructor): Likwise.
12426 * varpool.c (varpool_remove_node): Likewise.
12427 (varpool_remove_initializer): LIkewise.
12428 (dump_varpool_node): LIkwise.
12429 (const_value_known_p): Rewrite to ...
12430 (ctor_for_folding): ... this one.
12432 2013-06-19 Jakub Jelinek <jakub@redhat.com>
12435 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
12436 PERSONALITY in $PATH derived prefixes.
12438 2013-06-19 Jeff Law <law@redhat.com>
12440 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
12443 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
12444 (simplify_bitwise_binary): Use it to simpify certain binary ops on
12447 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
12449 * config/arm/vfp.md: Move VFP instruction classification documentation
12451 * config/arm/arm.md: ... here. Update instruction classification
12454 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
12456 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
12457 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
12458 pattern. Use more efficient sequences on ARMv5 and Thumb2.
12460 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
12463 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
12464 with NEXT_INSN. Use tablejump_p to check for jump table data
12467 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
12470 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
12471 that now in C++ the value is correct per the C++ standards.
12473 2013-06-19 Richard Biener <rguenther@suse.de>
12475 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
12476 for global context.
12478 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12481 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12484 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
12485 with next_active_insn.
12487 2013-06-18 Sriraman Tallam <tmsriram@google.com>
12489 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
12490 functions are inlined during failures to flag an error.
12491 * tree-inline.c (expand_call_inline): Allow the error to be flagged
12492 in early inline pass.
12494 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
12496 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
12499 2013-06-18 Julian Brown <julian@codesourcery.com>
12501 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
12502 Permit virtual register pre-reload if !strict.
12503 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
12505 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
12507 * config/arm/neon.md (movmisalign<mode>): Use
12508 neon_perm_struct_or_reg_operand instead of
12509 neon_struct_or_register_operand.
12510 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
12511 neon_permissive_struct_operand instead of neon_struct_operand.
12512 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
12513 neon_vector_mem_operand.
12514 * config/arm/predicates.md (neon_struct_operand): Adjust call to
12515 neon_vector_mem_operand.
12516 (neon_permissive_struct_operand): New.
12517 (neon_struct_or_register_operand): Rename to...
12518 (neon_perm_struct_or_reg_operand): This. Adjust call to
12519 neon_vector_mem_operand.
12521 2013-06-18 Richard Biener <rguenther@suse.de>
12523 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
12524 * lto-streamer.h: Include pointer-set.h.
12525 (struct lto_decl_slot): Remove.
12526 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
12527 Remove next_index entry.
12528 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
12529 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
12530 (lto_init_tree_ref_encoder): Adjust.
12531 (lto_destroy_tree_ref_encoder): Likewise.
12532 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
12533 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
12534 (lto_output_decl_index): Adjust.
12535 (lto_new_out_decl_state): Likewise.
12536 (lto_record_function_out_decl_state): Likewise.
12537 * lto-streamer-out.c (copy_function): Likewise.
12539 2013-06-18 Richard Biener <rguenther@suse.de>
12541 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
12542 * cgraphunit.c: Include cfgloop.h.
12543 (init_lowered_empty_function): Initialize the loop tree.
12544 (assemble_thunk): Insert new BBs into loops.
12546 2013-06-18 Richard Biener <rguenther@suse.de>
12548 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
12549 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
12550 the map from cache entry to cache index optional.
12551 (streamer_tree_cache_replace_tree): Adjust accordingly.
12552 (streamer_tree_cache_append): Likewise.
12553 (streamer_tree_cache_delete): Likewise.
12554 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
12555 streamer cache map from cache entry to cache index.
12556 * lto-streamer-out.c (create_output_block): Adjust.
12558 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
12560 * config/arm/arm.md (attribute "insn"): Move multiplication and
12561 division attributes to...
12562 (attribute "type"): ... here. Remove mult.
12563 (attribute "mul32"): New attribute.
12564 (attribute "mul64"): Add umaal.
12565 (*arm_mulsi3): Update attributes.
12566 (*arm_mulsi3_v6): Likewise.
12567 (*thumb_mulsi3): Likewise.
12568 (*thumb_mulsi3_v6): Likewise.
12569 (*mulsi3_compare0): Likewise.
12570 (*mulsi3_compare0_v6): Likewise.
12571 (*mulsi_compare0_scratch): Likewise.
12572 (*mulsi_compare0_scratch_v6): Likewise.
12573 (*mulsi3addsi): Likewise.
12574 (*mulsi3addsi_v6): Likewise.
12575 (*mulsi3addsi_compare0): Likewise.
12576 (*mulsi3addsi_compare0_v6): Likewise.
12577 (*mulsi3addsi_compare0_scratch): Likewise.
12578 (*mulsi3addsi_compare0_scratch_v6): Likewise.
12579 (*mulsi3subsi): Likewise.
12580 (*mulsidi3adddi): Likewise.
12581 (*mulsi3addsi_v6): Likewise.
12582 (*mulsidi3adddi_v6): Likewise.
12583 (*mulsidi3_nov6): Likewise.
12584 (*mulsidi3_v6): Likewise.
12585 (*umulsidi3_nov6): Likewise.
12586 (*umulsidi3_v6): Likewise.
12587 (*umulsidi3adddi): Likewise.
12588 (*umulsidi3adddi_v6): Likewise.
12589 (*smulsi3_highpart_nov6): Likewise.
12590 (*smulsi3_highpart_v6): Likewise.
12591 (*umulsi3_highpart_nov6): Likewise.
12592 (*umulsi3_highpart_v6): Likewise.
12593 (mulhisi3): Likewise.
12594 (*mulhisi3tb): Likewise.
12595 (*mulhisi3bt): Likewise.
12596 (*mulhisi3tt): Likewise.
12597 (maddhisi4): Likewise.
12598 (*maddhisi4tb): Likewise.
12599 (*maddhisi4tt): Likewise.
12600 (maddhidi4): Likewise.
12601 (*maddhidi4tb): Likewise.
12602 (*maddhidi4tt): Likewise.
12603 (divsi3): Likewise.
12604 (udivsi3): Likewise.
12605 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
12606 (thumb2_mulsi_short_compare0): Likewise.
12607 (thumb2_mulsi_short_compare0_scratch): Likewise.
12608 * config/arm/arm1020e.md (1020mult1): Update attribute change.
12609 (1020mult2): Likewise.
12610 (1020mult3): Likewise.
12611 (1020mult4): Likewise.
12612 (1020mult5): Likewise.
12613 (1020mult6): Likewise.
12614 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
12616 (cortex_a15_mult64): Likewise.
12617 (cortex_a15_sdiv): Likewise.
12618 (cortex_a15_udiv): Likewise.
12619 * config/arm/arm1026ejs.md (mult1): Update attribute change.
12625 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
12626 (pj4_ir_div): Likewise.
12627 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
12628 (11_mult2): Likewise.
12629 (11_mult3): Likewise.
12630 (11_mult4): Likewise.
12631 (11_mult5): Likewise.
12632 (11_mult6): Likewise.
12633 (11_mult7): Likewise.
12634 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
12635 (cortex_a8_mla): Likewise.
12636 (cortex_a8_mull): Likewise.
12637 (cortex_a8_smulwy): Likewise.
12638 (cortex_a8_smlald): Likewise.
12639 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
12640 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
12641 (cortex_r4_mul_3): Likewise.
12642 (cortex_r4_mla_4): Likewise.
12643 (cortex_r4_mla_3): Likewise.
12644 (cortex_r4_smlald): Likewise.
12645 (cortex_r4_mull): Likewise.
12646 (cortex_r4_sdiv): Likewise.
12647 (cortex_r4_udiv): Likewise.
12648 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
12649 (cortex_a7_idiv): Likewise.
12650 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
12651 (9_mult2): Likewise.
12652 (9_mult3): Likewise.
12653 (9_mult4): Likewise.
12654 (9_mult5): Likewise.
12655 (9_mult6): Likewise.
12656 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
12657 (cortex_a53_sdiv): Likewise.
12658 (cortex_a53_udiv): Likewise.
12659 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
12660 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
12661 (mp626_mult2): Likewise.
12662 (mp626_mult3): Likewise.
12663 (mp626_mult4): Likewise.
12664 * config/arm/fa526.md (526_mult1): Update attribute change.
12665 (526_mult2): Likewise.
12666 * config/arm/arm-generic.md (mult): Update attribute change.
12667 (mult_ldsched_strongarm): Likewise.
12668 (mult_ldsched): Likewise.
12669 (multi_cycle): Likewise.
12670 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
12671 * config/arm/fa606te.md (606te_mult1): Update attribute change.
12672 (606te_mult2): Likewise.
12673 (606te_mult3): Likewise.
12674 (606te_mult4): Likewise.
12675 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
12676 (cortex_a9_mac16): Likewise.
12677 (cortex_a9_multiply): Likewise.
12678 (cortex_a9_mac): Likewise.
12679 (cortex_a9_multiply_long): Likewise.
12680 * config/arm/fa626te.md (626te_mult1): Update attribute change.
12681 (626te_mult2): Likewise.
12682 (626te_mult3): Likewise.
12683 (626te_mult4): Likewise.
12685 2013-06-18 Richard Biener <rguenther@suse.de>
12688 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
12690 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12693 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
12694 with next_active_insn.
12696 2013-06-18 Alan Modra <amodra@gmail.com>
12698 * config/rs6000/rs6000.h (enum data_align): New.
12699 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
12700 (DATA_ABI_ALIGNMENT): Define.
12701 (CONSTANT_ALIGNMENT): Correct comment.
12702 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
12703 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
12705 2013-06-17 David Malcolm <dmalcolm@redhat.com>
12707 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
12708 ATTRIBUTE_UNUSED marking.
12710 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
12712 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
12713 alternative and update.
12714 (aarch64_dup_lanedi): Delete.
12715 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
12716 * config/aarch64/aarch64-simd-builtins.def: Update.
12718 2013-06-17 Richard Biener <rguenther@suse.de>
12720 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
12721 (lto_input_scc): Declare.
12722 (lto_input_tree_1): Likewise.
12723 (struct lto_stats_d): Add num_tree_bodies_output and
12724 num_pickle_refs_output.
12725 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
12726 (lto_read_tree_1): Split out from ...
12727 (lto_read_tree): ... this.
12728 (lto_input_scc): New function.
12729 (lto_input_tree_1): Split out from ...
12730 (lto_input_tree): ... this. Handle LTO_tree_scc.
12731 (lto_data_in_create): Create the streamer cache without hashes.
12732 * lto-streamer-out.c (create_output_block): Create the streamer
12733 cache with hashes when not doing WPA.
12734 (lto_write_tree_1): Split out from ...
12735 (lto_write_tree): ... this.
12736 (get_symbol_initial_value): New function.
12737 (lto_output_tree_1): Split out from ...
12738 (lto_output_tree): ... this. Write trees as series of SCCs
12739 using a DFS walk via DFS_write_tree.
12740 (struct sccs, struct scc_entry): New types.
12741 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
12742 (DFS_write_tree_body): New function.
12743 (DFS_write_tree): Likewise.
12744 (hash_tree): Likewise.
12745 (scc_entry_compare): Likewise.
12746 (hash_scc): Likewise.
12747 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
12748 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
12749 TREE_CHAIN as regular reference.
12750 (streamer_read_integer_cst): Remove.
12751 (streamer_get_pickled_tree): Adjust.
12752 * tree-streamer-out.c (streamer_write_chain): Disable streaming
12753 of DECL_EXTERNALs in BLOCK_VARS for now.
12754 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
12756 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
12757 Add hash value argument and record that if hashes are recorded
12759 (streamer_tree_cache_insert_1): Adjust.
12760 (streamer_tree_cache_insert): Likewise.
12761 (streamer_tree_cache_insert_at): Rename to ...
12762 (streamer_tree_cache_replace_tree): ... this and adjust.
12763 (streamer_tree_cache_append): Adjust.
12764 (record_common_node): Likewise.
12765 (streamer_tree_cache_create): Add argument whether to
12766 record hash values together with trees.
12767 (streamer_tree_cache_delete): Adjust.
12768 * tree-streamer.h (struct streamer_tree_cache_d): Add
12770 (streamer_read_integer_cst): Remove.
12771 (streamer_tree_cache_insert): Adjust.
12772 (streamer_tree_cache_append): Likewise.
12773 (streamer_tree_cache_insert_at): Rename to ...
12774 (streamer_tree_cache_replace_tree): ... this and adjust.
12775 (streamer_tree_cache_create): Add argument whether to record hashes.
12776 (streamer_tree_cache_get): Rename to ...
12777 (streamer_tree_cache_get_tree): ... this.
12778 (streamer_tree_cache_get_hash): New function.
12779 * tree.c (cache_integer_cst): New function.
12780 * tree.h (cache_integer_cst): Declare.
12781 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
12782 * lto-symtab.c (lto_varpool_replace_node): Only release
12783 DECL_INITIAL of non-prevailing decls.
12784 * varpool.c (varpool_remove_initializer): Do not release
12785 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
12787 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12789 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
12790 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
12791 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
12792 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
12793 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
12794 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
12795 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
12796 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
12797 instead of TARGET_64BIT.
12798 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
12799 Require ISA_HAS_<D>DIV.
12801 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
12803 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
12804 (mips*-*-linux*): Move default with_llsc setting to where other
12806 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
12808 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
12809 Likewise. Remove default with_tune setting. Move default float
12810 setting to its own block. Handle with_llsc in the same block as above.
12812 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
12814 PR rtl-optimization/57425
12815 PR rtl-optimization/57569
12816 * alias.c (write_dependence_p): Add new parameters mem_mode,
12817 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
12818 Changed all callers.
12819 (canon_anti_dependence): New function.
12820 * cse.c (check_dependence): Use canon_anti_dependence.
12821 * cselib.c (cselib_invalidate_mem): Likewise.
12822 * rtl.h (canon_anti_dependence): Declare.
12824 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12826 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
12827 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
12828 ".set mips3" for 64-bit targets.
12830 2013-06-15 Dehao Chen <dehao@google.com>
12832 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
12833 * gimple-low.c (gimple_check_call_matching_types): Likewise.
12834 (gimple_check_call_args): Likewise.
12835 * value-prof.c (check_ic_target): Likewise.
12836 * ipa-inline.c (early_inliner): Likewise.
12837 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
12838 * cgraph.c (cgraph_create_edge_1): Likewise.
12839 (cgraph_make_edge_direct): Likewise.
12841 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
12844 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
12845 rs6000_output_move_128bit to handle emitting quad memory
12846 operations. Set attribute length to 8 bytes.
12848 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
12850 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
12852 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
12853 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
12854 (aarch64_<su>mlsl<mode>): Likewise.
12856 2013-06-14 Mike Stump <mikestump@comcast.net>
12858 * Makefile.in (TARGET_H): Add insn-codes.h.
12860 2013-06-14 Alan Modra <amodra@gmail.com>
12862 PR middle-end/57134
12863 PR middle-end/57586
12864 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
12865 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
12866 bitfield expansion when EXPAND_MEMORY.
12867 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
12869 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
12871 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
12872 test for clearing quad memory on 32-bit later.
12874 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12876 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
12877 (fold_negate_expr): Likewise.
12878 (fold_real_zero_addition_p): Handle vectors.
12879 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
12881 2013-06-14 Alan Modra <amodra@gmail.com>
12883 * varasm.c (force_const_mem): Revert 2013-06-07 change.
12885 2013-06-13 Jan Hubicka <jh@suse.cz>
12887 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12888 Local comdats are not externally visible.
12889 * symtab.c (dump_symtab_base): Dump externally visible.
12890 (verify_symtab_base): Verify back links in the symtab hash.
12892 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12894 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
12895 CONVERT_EXPR as equal nodes.
12897 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12899 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
12901 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12903 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
12904 Generalize to complex and vector.
12905 * tree.c (build_all_ones_cst): New function.
12906 * tree.h (build_all_ones_cst): Declare it.
12908 2013-06-13 Alan Modra <amodra@gmail.com>
12910 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
12911 * config/rs6000/rs6000.md (signbittf2): New insn.
12912 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
12913 (abstf2_internal, cmptf_internal2): Likewise.
12914 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
12916 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12917 Pat Haugen <pthaugen@us.ibm.com>
12918 Peter Bergner <bergner@vnet.ibm.com>
12920 * config/rs6000/rs6000.c (emit_load_locked): Add support for
12921 power8 byte, half-word, and quad-word atomic instructions.
12922 (emit_store_conditional): Likewise.
12923 (rs6000_expand_atomic_compare_and_swap): Likewise.
12924 (rs6000_expand_atomic_op): Likewise.
12926 * config/rs6000/sync.md (larx): Add new modes for power8.
12928 (AINT): New mode iterator to include TImode as well as normal
12929 integer modes on power8.
12930 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
12931 that VSX registers are not considered. Use AINT mode iterator
12932 instead of INT1 to allow inclusion of quad word atomic operations
12934 (load_locked<mode>): Likewise.
12935 (store_conditional<mode>): Likewise.
12936 (atomic_compare_and_swap<mode>): Likewise.
12937 (atomic_exchange<mode>): Likewise.
12938 (atomic_nand<mode>): Likewise.
12939 (atomic_fetch_<fetchop_name><mode>): Likewise.
12940 (atomic_nand_fetch<mode>): Likewise.
12941 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
12943 (ATOMIC): On power8, add QImode, HImode modes.
12944 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
12945 modes that promote to SImode.
12946 (load_lockedti): Convert TImode arguments to PTImode, so that we
12947 get a guaranteed even/odd register pair.
12948 (load_lockedpti): Likewise.
12949 (store_conditionalti): Likewise.
12950 (store_conditionalpti): Likewise.
12952 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
12953 atomic load/store instructions.
12956 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
12958 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
12960 (insn_count): New attribute, with most cases extracted from...
12961 (length): ...here. Redefine most cases in terms of insn_count.
12962 (single_insn): Delete.
12963 (can_delay): Use insn_count to check for single instructions.
12964 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
12965 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
12966 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
12967 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
12968 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
12969 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
12970 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
12971 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
12972 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
12973 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
12974 rather than "length".
12975 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
12976 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
12977 Use "insn_count" rather than "length".
12978 * config/mips/mips-dsp.md
12979 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
12980 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
12983 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
12985 PR tree-optimization/57361
12986 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
12988 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
12990 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
12992 (aarch64_simd_combine<mode>): New instruction expansion.
12993 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
12994 function prototype.
12995 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
12996 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
12998 2013-06-12 Jan Hubicka <jh@suse.cz>
13000 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
13001 decl has when in streaming stage.
13002 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
13003 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
13005 2013-06-12 Roland Stigge <stigge@antcom.de>
13008 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
13010 2013-06-12 Jakub Jelinek <jakub@redhat.com>
13012 PR tree-optimization/57537
13013 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
13014 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
13016 2013-06-12 Richard Biener <rguenther@suse.de>
13018 * data-streamer.h (streamer_write_char_stream): CSE
13019 obs->current_pointer.
13020 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
13021 streamer_write_char_stream manually and optimize the resulting loop.
13022 (streamer_write_hwi_stream): Likewise.
13024 2013-06-12 Jan Hubicka <jh@suse.cz>
13026 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
13027 * cgraph.h (varpool_create_empty_node): Declare.
13028 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
13030 * symtab.c (symtab_unregister_node): Be lax about missin entries
13032 (symtab_get_node): Update comment.
13033 * varpool.c (varpool_create_empty_node): Break out from ...
13034 (varpool_node_for_decl): ... here.
13035 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
13037 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
13039 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
13040 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
13041 part. Use straight-line flow at the end.
13042 <COMPONENT_REF>: Remove superfluous else.
13043 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
13045 2013-06-12 Jakub Jelinek <jakub@redhat.com>
13048 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
13049 target hook even for !TREE_PUBLIC decls. If no resolution info
13050 is available, return false for common and external decls.
13052 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13054 * config/rl78/constraints.md (U): New constraint.
13055 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
13058 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
13061 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
13062 to allow returning address to AT_PLATFORM name.
13064 2013-06-11 Jan Hubicka <jh@suse.cz>
13066 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
13067 * cgraph.h (symtab_node_base): Add weakref flag.
13068 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
13069 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
13070 (output_weakrefs): Use weakref flag.
13071 * fold-const.c (simple_operand_p): Handle WEAK.
13072 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
13073 * ipa.c (varpool_externally_visible_p): Drop weakref.
13074 (function_and_variable_visibility): Update comment; fix weakref
13075 sanity checks; do not clear DECL_WEAK on them.
13076 * lto-cgraph.c (lto_output_node): update.
13077 (lto_output_varpool_node): Update.
13078 (input_overwrite_node): Update.
13079 (input_node): Update.
13080 (input_varpool_node): Update.
13081 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
13082 (lto_symtab_merge_symbols): Add sanity check.
13083 (lto_symtab_prevailing_decl): Do not special case weakrefs.
13084 * passes.c (rest_of_decl_compilation): Set static flag, too.
13085 * symtab.c (dump_symtab_base): Dump weakref.
13086 (verify_symtab_base): Sanity check weakrefs.
13087 (symtab_make_decl_local): Remove duplicated code.
13088 (symtab_alias_ultimate_target): Simplify.
13089 * varpool.c (varpool_create_variable_alias): Set weakref flag.
13091 2013-06-11 Tom de Vries <tom@codesourcery.com>
13093 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
13094 sequence_vect == NULL.
13096 2013-06-11 DJ Delorie <dj@redhat.com>
13098 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
13099 (rl78_unwind_word_mode): New.
13101 2013-06-11 David Malcolm <dmalcolm@redhat.com>
13103 * final.c (debug_prefix_maps): Make static.
13105 2013-06-11 David Malcolm <dmalcolm@redhat.com>
13107 * function.c (initial_trampoline): Remove stray copy.
13109 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
13111 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
13113 2013-06-11 Martin Jambor <mjambor@suse.cz>
13115 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
13116 within bounds at the beginning of the function.
13118 2013-06-11 Alan Modra <amodra@gmail.com>
13120 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
13122 (get_named_section): Don't NULL !DECL_P decl.
13124 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
13126 * doc/invoke.texi (core-avx2): Document.
13128 (atom): Updated with MOVBE.
13130 2013-06-11 Richard Biener <rguenther@suse.de>
13132 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
13134 2013-06-11 Anton Blanchard <anton@samba.org>
13136 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
13137 correct shift value in little-endian mode.
13139 2013-06-11 Jakub Jelinek <jakub@redhat.com>
13142 * varasm.c (get_variable_align): Move #endif to the right place.
13144 2013-06-10 Cary Coutant <ccoutant@google.com>
13146 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
13147 for hash so that hash table traversal order is deterministic.
13149 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
13150 Pat Haugen <pthaugen@us.ibm.com>
13151 Peter Bergner <bergner@vnet.ibm.com>
13153 * config/rs6000/vector.md (GPR move splitter): Do not split moves
13154 of vectors in GPRS if they are direct moves or quad word load or
13157 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
13159 (direct_move_p): Likewise.
13160 (quad_load_store_p): Likewise.
13162 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
13163 classes into bins based on the physical register type.
13164 (reg_class_to_reg_type): Likewise.
13165 (IS_STD_REG_TYPE): Likewise.
13166 (IS_FP_VECT_REG_TYPE): Likewise.
13167 (reload_fpr_gpr): Arrays to determine what insn to use if we can
13168 use direct move instructions.
13169 (reload_gpr_vsx): Likewise.
13170 (reload_vsx_gpr): Likewise.
13171 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
13172 information that is a simplification of register classes. Also
13173 precalculate direct move reload helpers.
13174 (direct_move_p): New function to return true if the operation can
13175 be done as a direct move instruciton.
13176 (quad_load_store_p): New function to return true if the operation
13177 is a quad memory operation.
13178 (rs6000_legitimize_address): If quad memory, only allow register
13179 indirect for TImode addresses.
13180 (rs6000_legitimate_address_p): Likewise.
13181 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
13182 (rs6000_reload_register_type): Likewise.
13183 (register_to_reg_type): Return register type.
13184 (rs6000_secondary_reload_simple_move): New helper function for
13185 secondary reload and secondary memory needed to identify anything
13186 that is a simple move, and does not need reloading.
13187 (rs6000_secondary_reload_direct_move): New helper function for
13188 secondary reload to identify cases that can be done with several
13189 instructions via the direct move instructions.
13190 (rs6000_secondary_reload_move): New helper function for secondary
13191 reload to identify moves between register types that can be done.
13192 (rs6000_secondary_reload): Add support for quad memory operations
13193 and for direct move.
13194 (rs6000_secondary_memory_needed): Likewise.
13195 (rs6000_debug_secondary_memory_needed): Change argument names.
13196 (rs6000_output_move_128bit): New function to return the move to
13197 use for 128-bit moves, including knowing about the various
13198 limitations of quad memory operations.
13200 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
13201 memory operations. call rs6000_output_move_128bit for the actual
13202 instruciton(s) to generate.
13203 (vsx_movti_64bit): Likewise.
13205 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
13206 (UNSPEC_P8V_MTVSRWZ): Likewise.
13207 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
13208 (UNSPEC_P8V_MTVSRD): Likewise.
13209 (UNSPEC_P8V_XXPERMDI): Likewise.
13210 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
13211 (UNSPEC_FUSION_GPR): Likewise.
13212 (FMOVE128_GPR): New iterator for direct move.
13213 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
13214 (f32_sv): Likewise.
13215 (f32_dm): Likewise.
13216 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
13217 loads and direct move instructions.
13218 (zero_extendsidi2_lfiwzx): Likewise.
13219 (extendsidi2_lfiwax): Likewise.
13220 (extendsidi2_nocell): Likewise.
13221 (floatsi<mode>2_lfiwax): Likewise.
13222 (lfiwax): Likewise.
13223 (floatunssi<mode>2_lfiwzx): Likewise.
13224 (lfiwzx): Likewise.
13225 (fix_trunc<mode>_stfiwx): Likewise.
13226 (fixuns_trunc<mode>_stfiwx): Likewise.
13227 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
13228 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
13229 (parity<mode>2_cmpb): Set length/type attr.
13230 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
13231 for 'mr.' to fast_compare.
13232 (bpermd_<mode>): Change type attr to popcnt.
13233 (p8_fmrgow_<mode>): New insns for power8 direct move support.
13234 (p8_mtvsrwz_1): Likewise.
13235 (p8_mtvsrwz_2): Likewise.
13236 (reload_fpr_from_gpr<mode>): Likewise.
13237 (p8_mtvsrd_1): Likewise.
13238 (p8_mtvsrd_2): Likewise.
13239 (p8_xxpermdi_<mode>): Likewise.
13240 (reload_vsx_from_gpr<mode>): Likewise.
13241 (reload_vsx_from_gprsf): Likewise.
13242 (p8_mfvsrd_3_<mode>): LIkewise.
13243 (reload_gpr_from_vsx<mode>): Likewise.
13244 (reload_gpr_from_vsxsf): Likewise.
13245 (p8_mfvsrd_4_disf): Likewise.
13246 (multi-word GPR splits): Do not split direct moves or quad memory
13249 2013-06-10 David Malcolm <dmalcolm@redhat.com>
13251 * tree-into-ssa.c (interesting_blocks): Make static.
13253 2013-06-10 Jakub Jelinek <jakub@redhat.com>
13256 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
13257 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
13258 Use DATA_ABI_ALIGNMENT for that case instead if defined.
13259 (get_variable_align): New function.
13260 (get_variable_section, emit_bss, emit_common,
13261 assemble_variable_contents, place_block_symbol): Use
13262 get_variable_align instead of DECL_ALIGN.
13263 (assemble_noswitch_variable): Add align argument, use it
13264 instead of DECL_ALIGN.
13265 (assemble_variable): Adjust caller. Use get_variable_align
13266 instead of DECL_ALIGN.
13267 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
13269 (DATA_ABI_ALIGNMENT): Define.
13270 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
13271 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
13272 opt is false, only return the psABI mandated alignment increase.
13273 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
13274 (DATA_ABI_ALIGNMENT): ... this.
13275 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
13276 (DATA_ABI_ALIGNMENT): ... this.
13277 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
13278 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
13279 (DATA_ABI_ALIGNMENT): ... this.
13280 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
13281 * doc/tm.texi: Regenerated.
13283 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
13285 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
13286 cmp_code to construct REG_EQUAL note.
13288 2013-06-09 Jakub Jelinek <jakub@redhat.com>
13291 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
13292 that operands[2] doesn't overlap with operands[0].
13294 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
13295 Jan Hubicka <jh@suse.cz>
13297 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
13298 hack to mark symbols as used.
13300 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
13302 PR rtl-optimization/57559
13303 * lra-constraints.c (process_alt_operands): Don't discourage
13304 memory with known offset for offsetable memory constraint.
13305 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
13307 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
13309 * varasm.c (struct oc_local_state): Reorder fields.
13310 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
13311 and adjust accordingly.
13312 (output_constructor): Reorder initialization code and adjust call to
13313 output_constructor_bitfield.
13315 2013-06-07 Jan Hubicka <jh@suse.cz>
13317 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
13319 2013-06-07 David Malcolm <dmalcolm@redhat.com>
13321 * tree-object-size.c (unknown): Make const.
13323 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13325 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
13326 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
13327 for last alternative in the cpu_facility attribute.
13329 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13332 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
13333 (xordi3): Change operand 2 constraint to arm_xordi_operand.
13334 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
13335 * config/arm/constraints.md (Dg): New constraint.
13336 * config/arm/neon.md (xordi3_neon): Remove.
13337 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
13338 * config/arm/predicates.md (arm_xordi_operand): New predicate.
13340 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13342 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
13343 Clean up alternatives.
13345 2013-06-07 Alan Modra <amodra@gmail.com>
13347 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
13350 2013-06-07 Alan Modra <amodra@gmail.com>
13352 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
13354 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13356 * config/arm/constraints.md (Df): New constraint.
13357 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
13358 Correct length attribute for last two alternatives.
13360 2013-06-07 Alan Modra <amodra@gmail.com>
13362 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
13363 override user -mfp-in-toc.
13364 (offsettable_ok_by_alignment): Consider just the current access
13365 rather than the whole object, unless BLKmode. Handle
13366 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
13367 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
13368 for -mcmodel=medium.
13369 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
13370 override user -mfp-in-toc or -msum-in-toc. Default to
13371 -mno-fp-in-toc for -mcmodel=medium.
13373 2013-06-06 DJ Delorie <dj@redhat.com>
13375 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
13376 TARGET_VALID_POINTER_MODE.
13378 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
13379 Pat Haugen <pthaugen@us.ibm.com>
13380 Peter Bergner <bergner@vnet.ibm.com>
13382 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
13383 Document new power8 builtins.
13385 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
13386 condition code register, to allow 128-bit logical operations to be
13387 done in the VSX or GPR registers.
13388 (nor<mode>3): Use the canonical form for nor.
13389 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
13390 vclz*, and vpopcnt* vector instructions.
13391 (nand<mode>3): Likewise.
13392 (orc<mode>3): Likewise.
13393 (clz<mode>2): LIkewise.
13394 (popcount<mode>2): Likewise.
13396 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
13397 that only the GPRs are recognized.
13399 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
13400 support for new power8 builtins.
13402 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
13404 (xscvdpspn): Likewise.
13410 (vpopcnt): Likewise.
13411 (vpopcntb): Likewise.
13412 (vpopcnth): Likewise.
13413 (vpopcntw): Likewise.
13414 (vpopcntd): Likewise.
13416 (vmrgew): Likewise.
13417 (vmrgow): Likewise.
13419 (eqv_v16qi3): Likewise.
13420 (eqv_v8hi3): Likewise.
13421 (eqv_v4si3): Likewise.
13422 (eqv_v2di3): Likewise.
13423 (eqv_v4sf3): Likewise.
13424 (eqv_v2df3): Likewise.
13426 (nand_v16qi3): Likewise.
13427 (nand_v8hi3): Likewise.
13428 (nand_v4si3): Likewise.
13429 (nand_v2di3): Likewise.
13430 (nand_v4sf3): Likewise.
13431 (nand_v2df3): Likewise.
13433 (orc_v16qi3): Likewise.
13434 (orc_v8hi3): Likewise.
13435 (orc_v4si3): Likewise.
13436 (orc_v2di3): Likewise.
13437 (orc_v4sf3): Likewise.
13438 (orc_v2df3): Likewise.
13440 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
13441 allow power8 quad mode in 64-bit.
13442 (rs6000_builtin_vectorized_function): Add support to vectorize
13443 ISA 2.07 count leading zeros, population count builtins.
13444 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
13445 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
13446 (builtin_function_type): Add vgbbd builtin function which takes an
13448 (altivec_expand_vec_perm_const): Add support for new power8 merge
13451 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
13452 that does not include TImdoe for use with 32-bit.
13453 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
13455 (UNSPEC_VSX_CVDPSPN): Likewise.
13456 (vsx_xscvdpspn): Likewise.
13457 (vsx_xscvspdpn): Likewise.
13458 (vsx_xscvdpspn_scalar): Likewise.
13459 (vsx_xscvspdpn_directmove): Likewise.
13460 (vsx_and<mode>3): Split logical operations into 32-bit and
13461 64-bit. Add support to do logical operations on TImode as well as
13462 VSX vector types. Allow logical operations to be done in either
13463 VSX registers or in general purpose registers in 64-bit mode. Add
13464 splitters if GPRs were used. For AND, add clobber of CCmode to
13465 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
13467 (vsx_and<mode>3_32bit): Likewise.
13468 (vsx_and<mode>3_64bit): Likewise.
13469 (vsx_ior<mode>3): Likewise.
13470 (vsx_ior<mode>3_32bit): Likewise.
13471 (vsx_ior<mode>3_64bit): Likewise.
13472 (vsx_xor<mode>3): Likewise.
13473 (vsx_xor<mode>3_32bit): Likewise.
13474 (vsx_xor<mode>3_64bit): Likewise.
13475 (vsx_one_cmpl<mode>2): Likewise.
13476 (vsx_one_cmpl<mode>2_32bit): Likewise.
13477 (vsx_one_cmpl<mode>2_64bit): Likewise.
13478 (vsx_nor<mode>3): Likewise.
13479 (vsx_nor<mode>3_32bit): Likewise.
13480 (vsx_nor<mode>3_64bit): Likewise.
13481 (vsx_andc<mode>3): Likewise.
13482 (vsx_andc<mode>3_32bit): Likewise.
13483 (vsx_andc<mode>3_64bit): Likewise.
13484 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
13485 and xxlorc instructions.
13486 (vsx_eqv<mode>3_64bit): Likewise.
13487 (vsx_nand<mode>3_32bit): Likewise.
13488 (vsx_nand<mode>3_64bit): Likewise.
13489 (vsx_orc<mode>3_32bit): Likewise.
13490 (vsx_orc<mode>3_64bit): Likewise.
13492 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
13494 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
13496 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
13497 (p8_vmrgow): Likewise.
13498 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
13499 GPRs to be split under VSX.
13500 (p8v_clz<mode>2): Add power8 count leading zero support.
13501 (p8v_popcount<mode>2): Add power8 population count support.
13502 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
13505 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
13508 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
13510 (vec_nand): Likewise.
13511 (vec_vclz): Likewise.
13512 (vec_vclzb): Likewise.
13513 (vec_vclzd): Likewise.
13514 (vec_vclzh): Likewise.
13515 (vec_vclzw): Likewise.
13516 (vec_vgbbd): Likewise.
13517 (vec_vmrgew): Likewise.
13518 (vec_vmrgow): Likewise.
13519 (vec_vpopcnt): Likewise.
13520 (vec_vpopcntb): Likewise.
13521 (vec_vpopcntd): Likewise.
13522 (vec_vpopcnth): Likewise.
13523 (vec_vpopcntw): Likewise.
13525 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13527 PR rtl-optimization/57468
13528 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
13531 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13533 PR rtl-optimization/57459
13534 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
13535 type when setting live regs.
13537 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13539 * config/s390/s390.opt (mlra): New option.
13540 * config/s390/s390.c (s390_decompose_address): Check displacement
13541 for all registers for LRA.
13542 (s390_secondary_reload): Don't used secondary reloads for LRA.
13543 (s390_lra_p): New function.
13544 (TARGET_LRA_P): Define.
13545 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
13546 of attribute cpu_facility to zarch for the last alternative.
13547 (*cmpmem_short): Ditto.
13549 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
13551 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
13552 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
13553 (arm_expand_prologue): Likewise.
13555 2013-06-06 Teresa Johnson <tejohnson@google.com>
13558 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
13559 as this is now done by redirect_edge_and_branch_force.
13560 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
13561 barriers, and fix interaction with splitting.
13562 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
13563 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
13564 reflect changes made in the routine.
13565 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
13566 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
13567 since this is called in cfglayout mode, and replace partition fixup
13568 with assert as that is now done by force_nonfallthru_and_redirect.
13569 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
13570 already be marked with region crossing note.
13571 (insert_section_boundary_note): Make non-static, gate on flag
13572 has_bb_partition, rewrite to also check for multiple partitions.
13573 (rest_of_handle_reorder_blocks): Remove call to
13574 insert_section_boundary_note, now done later during free_cfg.
13575 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
13576 * bb-reorder.h (insert_section_boundary_note): Declare.
13577 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
13578 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
13579 invoke insert_section_boundary_note.
13580 (try_redirect_by_replacing_jump): Remove unnecessary
13581 check for region crossing note.
13582 (fixup_partition_crossing): New function.
13583 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
13584 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
13585 in non-cfglayout mode.
13586 (force_nonfallthru_and_redirect): Fixup partition boundaries,
13587 remove old code that tried to do this. Emit barrier correctly
13588 when we are in cfglayout mode.
13589 (last_bb_in_partition): New function.
13590 (rtl_split_edge): Correctly fixup partition boundaries.
13591 (commit_one_edge_insertion): Remove old code that tried to
13592 fixup region crossing edge since this is now handled in
13593 split_block, and set up insertion point correctly since
13594 block may now end in a jump.
13595 (verify_hot_cold_block_grouping): Guard against checking when not in
13596 linearized RTL mode.
13597 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
13599 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
13600 rtl_verify_flow_info, so not called in cfglayout mode.
13601 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
13602 (fixup_reorder_chain): Remove old code that attempted to fixup region
13603 crossing note as this is now handled in force_nonfallthru_and_redirect.
13604 (duplicate_insn_chain): Don't duplicate switch section notes.
13605 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
13607 * basic-block.h (emit_barrier_after_bb): Declare.
13609 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13611 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
13612 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
13613 arm_usatsihi): Adjust alternatives for arm_restrict_it.
13615 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13617 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
13619 * config/arm/ldmstm.md: Regenerate.
13621 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13623 * config/arm/sync.md (atomic_loaddi_1):
13624 Disable predication for arm_restrict_it.
13625 (arm_load_exclusive<mode>): Likewise.
13626 (arm_load_exclusivesi): Likewise.
13627 (arm_load_exclusivedi): Likewise.
13628 (arm_load_acquire_exclusive<mode>): Likewise.
13629 (arm_load_acquire_exclusivesi): Likewise.
13630 (arm_load_acquire_exclusivedi): Likewise.
13631 (arm_store_exclusive<mode>): Likewise.
13632 (arm_store_exclusive<mode>): Likewise.
13633 (arm_store_release_exclusivedi): Likewise.
13634 (arm_store_release_exclusive<mode>): Likewise.
13636 2013-06-06 Richard Biener <rguenther@suse.de>
13638 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
13640 (lto_tag_is_tree_code_p): Adjust.
13641 (lto_tag_is_gimple_code_p): Likewise.
13642 (lto_gimple_code_to_tag): Likewise.
13643 (lto_tag_to_gimple_code): Likewise.
13644 (lto_tree_code_to_tag): Likewise.
13645 (lto_tag_to_tree_code): Likewise.
13646 * data-streamer.h (streamer_write_hwi_in_range): Use
13647 uhwi streaming to stream the normalized range.
13648 (streamer_read_hwi_in_range): Likewise.
13650 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13652 * config/arm/arm.md (enabled_for_depr_it): New attribute.
13653 (predicable_short_it): Likewise.
13654 (predicated): Likewise.
13655 (enabled): Handle above.
13656 (define_cond_exec): Set predicated attribute to yes.
13658 2013-06-05 Mike Stump <mikestump@comcast.net>
13660 * gdbinit.in (__FUNCTION__): Add.
13662 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
13664 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
13665 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
13667 2013-06-05 Jan Hubicka <jh@suse.cz>
13669 * varasm.c (mark_decl_referenced): Revert the removal until targets
13672 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
13674 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
13675 instead of mark_decl_referenced.
13677 2013-06-05 Jan Hubicka <jh@suse.cz>
13679 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
13680 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
13681 and symtab_used_from_object_file_p.
13682 (cgraph_make_node_local_1): Clear forced_by_abi.
13683 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
13684 * cgraph.h (symtab_node_base): Add forced_by_abi.
13685 (decide_is_variable_needed): Remove.
13686 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
13687 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
13688 (decide_is_symbol_needed): ... this one; handle symbols in general;
13689 always analyze virtuals; honnor forced_by_abi.
13690 (cgraph_finalize_function): Update.
13691 (varpool_finalize_decl): Update.
13692 (symbol_defined_and_needed): Remove.
13693 (analyze_functions): Update.
13694 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13695 output_refs, input_overwrite_node): Handle forced_by_abi.
13696 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
13697 (address_taken_from_non_vtable_p): ... this one.
13698 (comdat_can_be_unshared_p_1): New function.
13699 (cgraph_comdat_can_be_unshared_p): Rename to ...
13700 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
13701 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
13702 (function_and_variable_visibility): Clear forced_by_abi as needed.
13703 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
13704 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
13705 * symtab.c (dump_symtab_base): Dump forced_by_abi.
13706 * varpool.c (decide_is_variable_needed): Remove.
13708 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13710 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
13711 (arm_option_override): Override arm_restrict_it where appropriate.
13712 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
13713 * config/arm/arm.opt (mrestrict-it): New command-line option.
13714 * doc/invoke.texi: Document -mrestrict-it.
13716 2013-06-05 David Malcolm <dmalcolm@redhat.com>
13718 * tsan.c (tsan_atomic_table): Make const.
13720 2013-06-05 Richard Biener <rguenther@suse.de>
13722 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
13723 index associated with the tree we are supposed to replace.
13724 * tree-streamer-out.c (pack_ts_base_value_fields): Output
13725 TREE_ASM_WRITTEN as zero for everything but SSA names.
13727 2013-06-05 David Malcolm <dmalcolm@redhat.com>
13729 * tree-ssa-structalias.c (call_stmt_vars): Make static.
13731 2013-06-04 Jan Hubicka <jh@suse.cz>
13733 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
13734 (input_node, input_varpool_node): Handle correctly external same
13736 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
13737 nodes at ltrans stage.
13739 2013-06-04 Jan Hubicka <jh@suse.cz>
13741 * ipa-inline.c (update_caller_keys): Fix availability test.
13742 (update_callee_keys): Likewise.
13743 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
13744 to follow ELF standard.
13746 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
13748 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
13749 (mips64r5900el-*-elf*): New configurations.
13750 * config/mips/mips-cpus.def (r5900): New processor.
13751 * config/mips/mips-tables.opt: Regenerate.
13752 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
13753 (mips_issue_rate): Handle PROCESSOR_R5900.
13754 (mips_reorg_process_insns): Force reorder mode for the R5900.
13755 * config/mips/mips.h (TARGET_MIPS5900): Define.
13756 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
13758 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
13760 * config/mips/mips.md (processor): Add r5900.
13761 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
13763 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13765 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
13766 into function to generate MOVI instruction.
13767 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
13768 (aarch64_preferred_simd_mode): Turn into wrapper.
13769 (aarch64_output_scalar_simd_mov_immediate): New function.
13770 * config/aarch64/aarch64-protos.h: Add prototype for above.
13772 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13774 * config/aarch64/aarch64.c (simd_immediate_info): Remove
13775 element_char member.
13776 (sizetochar): Return signed char.
13777 (aarch64_simd_valid_immediate): Remove elchar and other
13778 unnecessary variables.
13779 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
13780 Calculate element_char as required.
13781 * config/aarch64/aarch64-protos.h: Update and move prototype
13782 for aarch64_output_simd_mov_immediate.
13783 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
13786 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13788 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
13789 information completed by aarch64_simd_valid_immediate.
13790 (aarch64_legitimate_constant_p): Update arguments.
13791 (aarch64_simd_valid_immediate): Work with struct rather than many
13793 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
13794 (aarch64_simd_make_constant): Update arguments.
13795 (aarch64_output_simd_mov_immediate): Work with struct rather than
13796 many pointers. Output immediate directly rather than as operand.
13797 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13799 * config/aarch64/constraints.md (Dn): Update arguments.
13801 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13803 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
13805 (aarch64_simd_immediate_valid_for_move): Remove.
13806 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
13807 (aarch64_simd_make_constant): Update call.
13808 (aarch64_output_simd_mov_immediate): Update call.
13809 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13811 * config/aarch64/constraints.md (Dn): Update call.
13813 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13815 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
13816 return type to bool for prototype.
13817 (aarch64_legitimate_constant_p): Check for true instead of not -1.
13818 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
13819 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
13821 2013-06-04 Catherine Moore <clm@codesourcery.com>
13823 * config/mips/mips.opt (meva): New.
13824 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
13825 (ASM_SPEC): Handle -meva.
13826 * doc/invoke.texi (meva): Document.
13828 2013-06-04 Alan Modra <amodra@gmail.com>
13830 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
13833 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13835 * rtl.def: Add extra fourth optional field to define_cond_exec.
13836 * gensupport.c (process_one_cond_exec): Process attributes from
13838 * doc/md.texi: Document fourth field in define_cond_exec.
13840 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
13842 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
13843 out the processing order as in store_bit_field_1.
13845 2013-06-04 Jan Hubicka <jh@suse.cz>
13847 PR middle-end/57500
13848 * cgraphunit.c (cgraph_process_same_body_aliases): Create
13849 non-VAR_DECL node if it does not exist yet.
13851 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
13853 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
13854 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
13855 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
13856 target_cpu_default setting.
13858 2013-06-03 Teresa Johnson <tejohnson@google.com>
13860 * dumpfile.c (opt_info_switch_p): Change -fopt-info
13861 default to -fopt-info=optimized instead of all.
13862 * doc/invoke.texi: Ditto.
13863 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
13864 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
13865 (execute_vect_slp): Emit BB vectorization success under
13866 MSG_OPTIMIZED_LOCATIONS.
13867 * tree-vect-slp.c (vect_slp_transform_bb): Change
13868 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
13869 * tree-vect-loop.c (vect_transform_loop): Ditto.
13871 2013-06-03 Jason Merrill <jason@redhat.com>
13874 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
13875 Use TARGET_EXPR for C++.
13877 2013-06-03 Jakub Jelinek <jakub@redhat.com>
13879 PR rtl-optimization/57268
13880 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
13881 if DEBUG_INSN_P (insn).
13884 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
13886 PR rtl-optimization/57268
13887 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13888 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13890 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13892 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
13893 (ix86_avoid_lea_for_addr): Likewise.
13894 (exact_dependency_1): Likewise.
13895 (ix86_adjust_cost): Likewise.
13896 (swap_top_of_ready_list): Fix formatting and !reload_completed check
13898 (do_reorder_for_imul): Fix typo, formatting and
13899 !reload_completed check removed.
13900 (ix86_sched_reorder): Fix typo and formatting.
13901 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
13904 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
13906 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
13908 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
13910 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
13911 <STRING_CST>: Likewise.
13912 <VECTOR_CST>: Likewise.
13914 2013-06-01 Janus Weil <janus@gcc.gnu.org>
13915 Mikael Morin <mikael@gcc.gnu.org>
13917 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
13918 * config.in: Regenerated.
13919 * configure: Regenerated.
13921 2013-06-01 Jan Hubicka <jh@suse.cz>
13923 PR middle-end/57366
13924 * cgraphunit.c (compile): When weakref is not supported,
13925 set up transparent aliases before final output pass.
13926 * varasm.c (assemble_alias): Do not try to do it here.
13928 2013-06-01 Jan Hubicka <jh@suse.cz>
13930 PR middle-end/57467
13931 * passes.c (for_per_function): Skip unanalyzed functions.
13933 2013-06-01 Jan Hubicka <jh@suse.cz>
13935 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
13936 (lto_symtab_merge_symbols_1): ... this one.
13937 (lto_symtab_merge_cgraph_nodes): Rename to ...
13938 (lto_symtab_merge_symbols): ... this one; simplify.
13939 * cgraph.c (same_body_aliases_done): Rename to ...
13940 (cpp_implicit_aliases_done): ... this one.
13941 (cgraph_create_function_alias): Update.
13942 (cgraph_same_body_alias): Update.
13943 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
13944 (verify_edge_corresponds_to_fndecl): Simplify.
13945 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
13946 (cgraph_node): Remove same_body_alias.
13947 (varpool_node): Remove alias_of and extra_name_alias.
13948 (same_body_aliases_done): Rename to ..
13949 (cpp_implicit_aliases_done): ... this one.
13950 (symtab_alias_ultimate_target): Add default parameter.
13951 (symtab_resolve_alias): New function.
13952 (fixup_same_cpp_alias_visibility): Declare.
13953 (cgraph_function_node): Add default parameter.
13954 (cgraph_node_asm_name): Likewise.
13955 (cgraph_function_or_thunk_node): Add default parameter; do
13956 not ICE when it is NULL.
13957 (varpool_variable_node): Likewise.
13958 * tree-emutls.c (create_emultls_var): Update.
13959 (ipa_lower_emutls): Update.
13960 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
13961 (cgraph_reset_node): Reset alias info.
13962 (cgraph_finalize_function): Update.
13963 (fixup_same_cpp_alias_visibility): Move to symtab.c.
13964 (analyze_function): Simplify.
13965 (cgraph_process_same_body_aliases): Simplify.
13966 (analyze_functions): Fixup same body aliases.
13967 (handle_alias_pairs): Simplify.
13968 (assemble_thunk): Update.
13969 (assemble_thunks_and_aliases): Update.
13970 (output_weakrefs): Rewrite.
13971 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
13972 (lto_output_varpool_node): Likewise.
13973 (compute_ltrans_boundary): Remve assert.
13974 (get_alias_symbol): New functoin.
13975 (input_node): Rewrite alias handling.
13976 (input_varpool_node): Likewise.
13977 * ipa-pure-const.c (propagate_pure_const): Fix formating.
13978 * ipa.c (process_references): Handle weakrefs correctly.
13979 (symtab_remove_unreachable_nodes): Likewise.
13980 * trans-mem.c (get_cg_data): Update.
13981 (ipa_tm_create_version_alias): Update.
13982 (ipa_tm_execute): Update.
13983 * symtab.c (dump_symtab_base): Dump aliases.
13984 (verify_symtab_base): Verify aliases.
13985 (symtab_node_availability): New function.
13986 (symtab_alias_ultimate_target): Simplify.
13987 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
13988 handle all the fixup cases.
13989 (symtab_resolve_alias): New function.
13990 * passes.c (ipa_write_summaries): Handle weakrefs.
13991 * varpool.c (varpool_analyze_node): Simplify.
13992 (assemble_aliases): Update.
13993 (varpool_create_variable_alias): Simplify.
13994 (varpool_extra_name_alias): Simplify.
13995 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
13996 (lto_symtab_merge_symbols): ... this one.
13998 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
14001 PR rtl-optimization/57268
14002 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14003 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14005 2013-06-01 Tobias Burnus <burnus@net-b.de>
14007 Partially reverted:
14008 2013-05-31 Tobias Burnus <burnus@net-b.de>
14010 PR middle-end/57073
14011 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
14014 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14016 PR rtl-optimization/57268
14017 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14018 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14020 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14022 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
14023 unordered comparison operators when -fno-trapping-math is in effect
14025 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
14026 and implement unordered comparison operators properly on the e500.
14028 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14030 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
14031 for constant scalar integers.
14032 (simplify_relational_operation_1): Likewise.
14034 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
14036 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
14037 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
14040 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14041 Igor Zamyatin <igor.zamyatin@intel.com>
14043 Silvermont (SLM) architecture performance tuning.
14044 * config/i386/i386.h (enum ix86_tune_indices): Add
14045 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
14046 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
14048 * config/i386/i386.c (initial_ix86_tune_features)
14049 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
14050 (ix86_lea_outperforms): Handle Silvermont tuning.
14051 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
14053 (ix86_use_lea_for_mov): Likewise.
14054 (ix86_avoid_lea_for_addr): Likewise.
14055 (ix86_lea_for_add_ok): Likewise.
14056 (exact_dependency_1): New function.
14057 (exact_store_load_dependency): Likewise.
14058 (ix86_adjust_cost): Handle Silvermont tuning.
14059 (do_reoder_for_imul): Likewise.
14060 (swap_top_of_ready_list): New function.
14061 (ix86_sched_reorder): Changed to handle Silvermont tuning.
14063 * config/i386/i386.md (peepholes that split memory operand in fp
14066 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14068 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
14069 Remove un-necessary braces.
14071 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14073 * config/aarch64/aarch64.c (aarch64_classify_symbol):
14074 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
14076 2013-05-31 Tobias Burnus <burnus@net-b.de>
14078 PR middle-end/57073
14079 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
14081 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14084 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
14085 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
14086 * config/arm/neon.md (iordi3_neon): Remove.
14087 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
14088 * config/arm/predicates.md (imm_for_neon_logic_operand):
14089 Move to earlier in the file.
14090 (neon_logic_op2): Likewise.
14091 (arm_iordi_operand_neon): New predicate.
14093 2013-05-31 Richard Biener <rguenther@suse.de>
14095 PR tree-optimization/57478
14096 PR tree-optimization/57453
14097 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
14100 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14102 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
14103 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
14105 2013-05-30 Tobias Burnus <burnus@net-b.de>
14106 Thomas Koenig <tkoenig@gcc.gnu.org>
14108 PR middle-end/57073
14109 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
14110 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
14112 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
14114 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
14116 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
14118 * target.def (register_usage_leveling_p): New hook.
14119 * targhooks.c (default_register_usage_leveling_p): New.
14120 * targhooks.h (default_register_usage_leveling_p): New prototype.
14121 * lra-assigns.c (register_usage_leveling_p): Use the hook.
14122 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
14123 * doc/tm.texi: Update.
14124 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
14126 2013-05-30 Ian Bolton <ian.bolton@arm.com>
14128 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
14129 (*insv_reg<mode>): New define_insn.
14131 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
14133 PR rtl-optimization/57439
14134 * postreload.c (move2add_valid_value_p): Check that we have
14135 a zero subreg_regno_offset when accessing the register in
14136 the requested mode.
14138 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14139 Igor Zamyatin <igor.zamyatin@intel.com>
14141 Silvermont (SLM) architecture pipeline model, tuning and
14143 * config.gcc: Add slm config options and target.
14145 * config/i386/slm.md: New.
14147 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
14149 * config/i386/i386-c.c (ix86_target_macros_internal): New case
14151 (ix86_target_macros_internal): Likewise.
14153 * config/i386/i386.c (slm_cost): New cost.
14154 (m_SLM): New macro flag.
14155 (initial_ix86_tune_features): Set m_SLM.
14156 (x86_accumulate_outgoing_args): Likewise.
14157 (x86_arch_always_fancy_math_387): Likewise.
14158 (processor_target_table): Add slm cost.
14159 (cpu_names): Add slm cpu name.
14160 (x86_option_override_internal): Set SLM ISA.
14161 (ix86_issue_rate): New case PROCESSOR_SLM.
14162 (ia32_multipass_dfa_lookahead): Likewise.
14163 (fold_builtin_cpu): Add slm.
14165 * config/i386/i386.h (TARGET_SLM): New target macro.
14166 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
14167 (processor_type): Add PROCESSOR_SLM.
14169 * config/i386/i386.md (cpu): Add new value "slm".
14170 (slm.md): Include slm.md.
14172 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
14173 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14175 * config/arm/arm-protos.h: Add and update function protos.
14176 * config/arm/arm.c (use_simple_return_p): New added.
14177 (thumb2_expand_return): Check simple_return flag.
14178 * config/arm/arm.md: Add simple_return and conditional simple_return.
14179 * config/arm/iterators.md: Add iterator for return and simple_return.
14181 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14183 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
14184 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
14185 (arm_emit_vfp_multi_reg_pop): Likewise.
14186 (thumb2_emit_ldrd_pop): Likewise.
14187 (arm_expand_epilogue): Add misc REG_CFA notes.
14188 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
14190 2013-05-29 Lawrence Crowl <crowl@google.com>
14192 * config/arm/t-arm: Update for below.
14194 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
14195 Change type to hash_table. Update dependent calls and types.
14197 * config/i386/t-cygming: Update for below.
14199 * config/i386/t-interix: Update for below.
14201 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
14202 Change type to hash_table. Update dependent calls and types.
14203 (i386_find_on_wrapper_list::wrappers): Likewise.
14205 * config/ia64/t-ia64: Update for below.
14207 * config/ia64/ia64.c (bundle_state_table):
14208 Change type to hash_table. Update dependent calls and types.
14210 * config/mips/mips.c (mips_reorg_process_insns::htab):
14211 Change type to hash_table. Update dependent calls and types.
14213 * config/sol2.c (solaris_comdat_htab):
14214 Change type to hash_table. Update dependent calls and types.
14216 * config/t-sol2: Update for above.
14218 2013-05-29 Teresa Johnson <tejohnson@google.com>
14220 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
14221 functions are not yet marked as defined.
14223 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
14224 Pat Haugen <pthaugen@us.ibm.com>
14225 Peter Bergner <bergner@vnet.ibm.com>
14227 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
14231 (vrotl<mode>3): Likewise.
14232 (vashl<mode>3): Likewise.
14233 (vlshr<mode>3): Likewise.
14234 (vashr<mode>3): Likewise.
14236 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14237 support for power8 V2DI builtins.
14239 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
14240 power8 V2DI builtins.
14241 (vupkhsw): Likewise.
14242 (vupklsw): Likewise.
14243 (vaddudm): Likewise.
14244 (vminsd): Likewise.
14245 (vmaxsd): Likewise.
14246 (vminud): Likewise.
14247 (vmaxud): Likewise.
14248 (vpkudum): Likewise.
14249 (vpksdss): Likewise.
14250 (vpkudus): Likewise.
14251 (vpksdus): Likewise.
14256 (vsubudm): Likewise.
14257 (vcmpequd): Likewise.
14258 (vcmpgtsd): Likewise.
14259 (vcmpgtud): Likewise.
14260 (vcmpequd_p): Likewise.
14261 (vcmpgtsd_p): Likewise.
14262 (vcmpgtud_p): Likewise.
14263 (vupkhsw): Likewise.
14264 (vupklsw): Likewise.
14265 (vaddudm): Likewise.
14266 (vmaxsd): Likewise.
14267 (vmaxud): Likewise.
14268 (vminsd): Likewise.
14269 (vminud): Likewise.
14270 (vpksdss): Likewise.
14271 (vpksdus): Likewise.
14272 (vpkudum): Likewise.
14273 (vpkudus): Likewise.
14278 (vsubudm): Likewise.
14280 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
14281 support for power8 V2DI instructions.
14283 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
14284 power8 V2DI instructions. Combine pack and unpack insns to use an
14285 iterator for each mode. Check whether a particular mode supports
14286 Altivec instructions instead of just checking TARGET_ALTIVEC.
14287 (UNSPEC_VPKUWUM): Likewise.
14288 (UNSPEC_VPKSHSS): Likewise.
14289 (UNSPEC_VPKSWSS): Likewise.
14290 (UNSPEC_VPKUHUS): Likewise.
14291 (UNSPEC_VPKSHUS): Likewise.
14292 (UNSPEC_VPKUWUS): Likewise.
14293 (UNSPEC_VPKSWUS): Likewise.
14294 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
14295 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
14296 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
14297 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
14298 (UNSPEC_VUPKHSB): Likewise.
14299 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
14300 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
14301 (UNSPEC_VUPKHSH): Likewise.
14302 (UNSPEC_VUPKLSB): Likewise.
14303 (UNSPEC_VUPKLSH): Likewise.
14305 (VI_char): Likewise.
14306 (VI_scalar): Likewise.
14307 (VI_unit): Likewise.
14309 (VP_small): Likewise.
14310 (VP_small_lc): Likewise.
14311 (VU_char): Likewise.
14312 (add<mode>3): Likewise.
14313 (altivec_vaddcuw): Likewise.
14314 (altivec_vaddu<VI_char>s): Likewise.
14315 (altivec_vadds<VI_char>s): Likewise.
14316 (sub<mode>3): Likewise.
14317 (altivec_vsubcuw): Likewise.
14318 (altivec_vsubu<VI_char>s): Likewise.
14319 (altivec_vsubs<VI_char>s): Likewise.
14320 (altivec_vavgs<VI_char>): Likewise.
14321 (altivec_vcmpbfp): Likewise.
14322 (altivec_eq<mode>): Likewise.
14323 (altivec_gt<mode>): Likewise.
14324 (altivec_gtu<mode>): Likewise.
14325 (umax<mode>3): Likewise.
14326 (smax<mode>3): Likewise.
14327 (umin<mode>3): Likewise.
14328 (smin<mode>3): Likewise.
14329 (altivec_vpkuhum): Likewise.
14330 (altivec_vpkuwum): Likewise.
14331 (altivec_vpkshss): Likewise.
14332 (altivec_vpkswss): Likewise.
14333 (altivec_vpkuhus): Likewise.
14334 (altivec_vpkshus): Likewise.
14335 (altivec_vpkuwus): Likewise.
14336 (altivec_vpkswus): Likewise.
14337 (altivec_vpks<VI_char>ss): Likewise.
14338 (altivec_vpks<VI_char>us): Likewise.
14339 (altivec_vpku<VI_char>us): Likewise.
14340 (altivec_vpku<VI_char>um): Likewise.
14341 (altivec_vrl<VI_char>): Likewise.
14342 (altivec_vsl<VI_char>): Likewise.
14343 (altivec_vsr<VI_char>): Likewise.
14344 (altivec_vsra<VI_char>): Likewise.
14345 (altivec_vsldoi_<mode>): Likewise.
14346 (altivec_vupkhsb): Likewise.
14347 (altivec_vupkhs<VU_char>): Likewise.
14348 (altivec_vupkls<VU_char>): Likewise.
14349 (altivec_vupkhsh): Likewise.
14350 (altivec_vupklsb): Likewise.
14351 (altivec_vupklsh): Likewise.
14352 (altivec_vcmpequ<VI_char>_p): Likewise.
14353 (altivec_vcmpgts<VI_char>_p): Likewise.
14354 (altivec_vcmpgtu<VI_char>_p): Likewise.
14355 (abs<mode>2): Likewise.
14356 (vec_unpacks_hi_v16qi): Likewise.
14357 (vec_unpacks_hi_v8hi): Likewise.
14358 (vec_unpacks_lo_v16qi): Likewise.
14359 (vec_unpacks_hi_<VP_small_lc>): Likewise.
14360 (vec_unpacks_lo_v8hi): Likewise.
14361 (vec_unpacks_lo_<VP_small_lc>): Likewise.
14362 (vec_pack_trunc_v8h): Likewise.
14363 (vec_pack_trunc_v4si): Likewise.
14364 (vec_pack_trunc_<mode>): Likewise.
14366 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
14368 (vec_vmaxsd): Likewise.
14369 (vec_vmaxud): Likewise.
14370 (vec_vminsd): Likewise.
14371 (vec_vminud): Likewise.
14372 (vec_vpksdss): Likewise.
14373 (vec_vpksdus): Likewise.
14374 (vec_vpkudum): Likewise.
14375 (vec_vpkudus): Likewise.
14376 (vec_vrld): Likewise.
14377 (vec_vsld): Likewise.
14378 (vec_vsrad): Likewise.
14379 (vec_vsrd): Likewise.
14380 (vec_vsubudm): Likewise.
14381 (vec_vupkhsw): Likewise.
14382 (vec_vupklsw): Likewise.
14384 2013-05-29 Jan Hubicka <jh@suse.cz>
14386 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
14387 flags; reorder rest of fields in more consistent way.
14388 (varpool_node): Remove analyzed, finalized and alias.
14389 (cgraph_ndoe): Likewise.
14390 (symtab_alias_ultimate_target): New function.
14391 (cgraph_function_node): Move offline.
14392 (cgraph_reset_node): Declare.
14393 (cgraph_comdat_can_be_unshared_p): Remove.
14394 (varpool_remove_initializer): Declare.
14395 (varpool_first_defined_variable, varpool_next_defined_variable
14396 cgraph_first_defined_function, cgraph_next_defined_function): Update.
14397 (cgraph_function_with_gimple_body_p): Update.
14398 (varpool_all_refs_explicit_p): Update.
14399 (symtab_alias_target): New function.
14400 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
14401 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
14402 (cgraph_function_or_thunk_node): Simplify using
14403 symtab_alias_ultimate_target.
14404 (varpool_variable_node): Likewise.
14405 * cgraph.c (cgraph_create_function_alias): Update.
14406 (cgraph_add_thunk): Update.
14407 (cgraph_remove_node): Update.
14408 (dump_cgraph_node): Do not dump removed flags.
14409 (cgraph_function_body_availability): Update.
14410 (cgraph_propagate_frequency): Update.
14411 (verify_cgraph_node): Check sanity of local flag.
14412 (cgraph_function_node): Move here from cgraph.h; revamp for
14413 cgraph_function_or_thunk_node.
14414 * lto-symtab.c (lto_varpool_replace_node): Update.
14415 (lto_symtab_resolve_can_prevail_p): Update.
14416 (lto_symtab_merge_cgraph_nodes): Update.
14417 * ipa-cp.c (determine_versionability, initialize_node_lattices,
14418 propagate_constants_accross_call, devirtualization_time_bonus,
14419 ipcp_propagate_stage): Update.
14420 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
14421 * ipa-inline-transform.c (clone_inlined_nodes,
14422 preserve_function_body_p): Update.
14423 * ipa-reference.c (propagate): Update.
14424 (write_node_summary_p): Update.
14425 * toplev.c (wrapup_global_declaration_2): Update.
14426 * cgraphunit.c (cgraph_analyze_function): Rename to ...
14427 (analyze_function) ... this one.
14428 (cgraph_process_new_functions): Update.
14429 (cgraph_reset_node): Export.
14430 (cgraph_finalize_function): Update.
14431 (cgraph_add_new_function): Update.
14432 (process_function_and_variable_attributes): Update.
14433 (varpool_finalize_decl): Update.
14434 (symbol_finalized): Remove.
14435 (symbol_finalized_and_needed): Rename to ...
14436 (symbol_defined_and_needed): ... update.
14437 (cgraph_analyze_functions): Update.
14438 (handle_alias_pairs): Update.
14439 (mark_functions_to_output): Update.
14440 (assemble_thunk): Update.
14441 (output_in_order): Update.
14442 (output_weakrefs): Update.
14443 (finalize_compilation_unit): Update.
14444 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
14445 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
14446 input_node, input_varpool_node): Update.
14447 * dbxout.c (dbxout_expand_expr): Update.
14448 * cgraphclones.c (cgraph_clone_node): Update.
14449 (cgraph_copy_node_for_versioning): Update.
14450 (cgraph_materialize_clone): Update.
14451 (cgraph_materialize_all_clones): Update.
14452 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
14453 propagate_pure_const, propagate_nothrow): Update.
14454 * lto-streamer-out.c (lto_output, write_symbol): Update.
14455 * ipa-utils.c (ipa_reverse_postorder): Update.
14456 * ipa-inline.c (can_inline_edge_p): Update.
14457 (update_caller_keys, ipa_inline): Update.
14458 * dwarf2out.c (reference_to_unused,
14459 premark_types_used_by_global_vars_helper): Update.
14460 * tree-eh.c (tree_could_trap_p): Update.
14461 * ipa-split.c (consider_split, execute_split_functions): Update.
14462 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
14463 has_addr_references_p): Update; move ahead in file for better
14465 (process_references): Simplify.
14466 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
14467 bodies are removed.
14468 (cgraph_comdat_can_be_unshared_p): Make static.
14469 (cgraph_externally_visible_p): Update.
14470 (varpool_externally_visible_p): Update.
14471 (function_and_variable_visibility): Update.
14472 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
14473 ipa_tm_mark_force_output_node): Update.
14474 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
14475 estimate_edge_devirt_benefit, inline_generate_summary,
14476 inline_write_summary): Update.
14477 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
14478 * ipa-prop.c (ipa_compute_jump_functions): Update.
14479 (ipa_print_node_params, ipa_prop_read_section,
14480 ipa_update_after_lto_read, read_replacements_section): Update.
14481 * varasm.c (mark_decl_referenced): Update.
14482 (assemble_alias, dump_tm_clone_pairs): Update.
14483 * tree-inline.c (copy_bb): Update.
14484 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
14486 * symtab.c (dump_symtab_base): Print new flags.
14487 (verify_symtab_base): Verify new flags.
14488 (symtab_alias_ultimate_target): New function.
14489 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
14490 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
14492 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
14494 * i386.c (ix86_get_function_versions_dispatcher,
14495 ix86_generate_version_dispatcher_body): Update.
14496 (fold_builtin_cpu): Use varpool_add_new_variable.
14497 * varpool.c (varpool_remove_initializer): Break out from ...
14498 (varpool_remove_node): ... this one.
14499 (dump_varpool_node, varpool_node_for_asm,
14500 cgraph_variable_initializer_availability, varpool_analyze_node,
14501 varpool_assemble_decl, varpool_remove_unreferenced_decls,
14502 varpool_finalize_named_section_flags, varpool_create_variable_alias):
14505 2013-05-29 Jan Hubicka <jh@suse.cz>
14507 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
14509 2013-05-29 Easwaran Raman <eraman@google.com>
14511 PR tree-optimization/57442
14512 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
14513 when control exits the main loop.
14515 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
14517 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
14519 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
14520 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
14521 * rx/t-rx: Add rx100 under multi library matches option for nofpu
14524 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14526 PR tree-optimization/57441
14527 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
14528 Don't limit size of incr_vec to number of candidates.
14530 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
14532 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
14533 and mips16 directories.
14534 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
14535 (MULTILIB_DIRNAMES): Ditto.
14536 (MULTILIB_EXCEPTIONS): Add new exceptions.
14537 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
14538 (MULTILIB_DIRNAMES): Ditto.
14539 (MULTILIB_EXCEPTIONS): Add new exceptions.
14541 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14542 Marcus Shawcroft <marcus.shawcroft@arm.com>
14544 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
14545 SYMBOL_TINY_ABSOLUTE.
14546 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
14547 SYMBOL_TINY_ABSOLUTE.
14548 (aarch64_expand_mov_immediate): Likewise.
14549 (aarch64_classify_symbol): Likewise.
14550 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
14551 Permit SYMBOL_TINY_ABSOLUTE.
14552 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
14554 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14555 Marcus Shawcroft <marcus.shawcroft@arm.com>
14557 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
14558 Refactor if/switch. Replace gcc_assert with if.
14560 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14562 * config/i386/i386.c (initial_ix86_tune_features): Enable
14563 FP Reassociation for AMD bdver1 and bdver2.
14565 2013-05-29 Martin Jambor <mjambor@suse.cz>
14567 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
14568 and IMAGPART_EXPR do not occur within other handled_components.
14570 2013-05-29 Richard Biener <rguenther@suse.de>
14572 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
14573 access on whether the use is in the BB we currently try to
14575 (vect_bb_vectorization_profitable_p): Pass the BB we currently
14576 vectorize to vect_bb_slp_scalar_cost.
14578 2013-05-29 Richard Biener <rguenther@suse.de>
14580 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
14581 computing scalar cost offsetted by stmts that are kept live
14583 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
14584 for computation of scalar cost.
14586 2013-05-28 Steve Ellcey <sellcey@mips.com>
14588 * config/mips/mips-cpus.def (mips32r2): Change processor type.
14590 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
14592 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
14593 array notation built-in reduction functions.
14594 * doc/passes.texi (Passes): Added documentation about changes done
14596 * doc/invoke.texi (C Dialect Options): Added documentation about
14597 the -fcilkplus flag.
14598 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
14599 (BUILTINS_DEF): Depend on cilkplus.def.
14600 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
14601 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
14602 * cilkplus.def: New file.
14604 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
14606 PR rtl-optimization/57439
14607 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
14609 2013-05-28 Easwaran Raman <eraman@google.com>
14611 PR tree-optimization/57337
14612 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
14613 (find_insert_point): Correctly identify the insertion point
14614 when two statements with the same UID is compared.
14616 2013-05-28 Richard Biener <rguenther@suse.de>
14618 PR tree-optimization/56787
14619 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
14620 from the list of data references.
14621 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
14623 (vect_analyze_loop_operations): Likewise.
14624 (vect_transform_loop): Remove clobbers.
14626 2013-05-28 Martin Jambor <mjambor@suse.cz>
14628 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
14629 and REALPART_EXPRs have scalar type.
14631 2013-05-28 Richard Biener <rguenther@suse.de>
14633 PR tree-optimization/57411
14634 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
14636 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
14637 virtual operand propagation.
14639 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
14641 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
14642 destination register for bmasksi_vis.
14643 (vector_init_bshuffle): Likewise.
14644 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
14646 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
14648 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
14649 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
14650 mode if the instruction isn't available in the original mode.
14651 * config/sparc/sparc.opt (mfix-ut699): New option.
14652 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
14653 (divdf3): Turn into expander.
14654 (divdf3_nofix): New insn.
14655 (divdf3_fix): Likewise.
14656 (divsf3): Disable if -mfix-ut699.
14657 (sqrtdf2): Turn into expander.
14658 (sqrtdf2_nofix): New insn.
14659 (sqrtdf2_fix): Likewise.
14660 (sqrtsf2): Disable if -mfix-ut699.
14662 2013-05-27 Richard Biener <rguenther@suse.de>
14664 PR middle-end/57412
14665 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
14666 block for the new loop.
14668 2013-05-27 Richard Biener <rguenther@suse.de>
14670 PR tree-optimization/57343
14671 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
14672 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
14673 (number_of_iterations_cond): Do not build the folded tree.
14675 2013-05-27 Richard Biener <rguenther@suse.de>
14678 PR middle-end/57381
14679 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
14680 OEP_CONSTANT_ADDRESS_OF retained.
14682 PR tree-optimization/57417
14683 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
14684 for unchanged base.
14685 (set_ssa_val_to): Compare addresses using
14686 get_addr_base_and_unit_offset.
14688 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
14690 PR rtl-optimization/56833
14691 * postreload.c (move2add_record_mode): New function.
14692 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
14693 (move2add_use_add2_insn): Use move2add_record_sym_value.
14694 (move2add_use_add3_insn): Likewise.
14695 (reload_cse_move2add): Use move2add_valid_value_p and
14696 move2add_record_mode. Invalidate call-clobbered and REG_INC
14697 affected regs by setting reg_mode to VOIDmode.
14698 (move2add_note_store): Don't pretend the inside of a SUBREG is
14699 the actual destination. Invalidate single/leading registers by
14700 setting reg_mode to VOIDmode.
14701 Use move2add_record_sym_value, move2add_valid_value_p and
14702 move2add_record_mode.
14704 2013-05-27 Richard Biener <rguenther@suse.de>
14706 PR tree-optimization/57396
14707 * tree-affine.c (double_int_constant_multiple_p): Properly
14708 return false for val == 0 and div != 0.
14710 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14712 * config/mips/mips.h: Use #elif in preprocessor conditions.
14714 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14717 * config/mips/constraints.md (kl): New constraint.
14718 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
14719 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
14720 constraint for operand 0. Split after CSE for MIPS16. Emit a move
14721 from LO for MIPS16.
14722 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
14724 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14727 * config/mips/mips.c (mips_can_inline_p): New function.
14728 (TARGET_CAN_INLINE_P): Define.
14730 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
14732 * sched-int.h (ds_t, dw_t): Make unsigned int.
14733 Fix documentation that describes how all the ds_t bits are used.
14734 Reserve the last bit for delayed-branch scheduling.
14735 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
14736 (BITS_PER_DEP_WEAK): Fix definition and documentation.
14737 (gen_dep_weak_1): Remove prototype.
14738 * sched-deps.c (get_dep_weak_1): Make static.
14739 * target.def (speculate_insn, needs_block_p, gen_spec_check,
14740 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
14741 * doc/tm.texi: Regenerate.
14742 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
14744 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
14747 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
14749 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
14750 Sandra Loosemore <sandra@codesourcery.com>
14752 * config.gcc (powerpc-*): Allow native for with-cpu.
14754 2013-05-24 Jeff Law <law@redhat.com>
14756 PR tree-optimization/57124
14757 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
14758 conversion feeding a condition if the range has an overflow
14759 if -fstrict-overflow. Add warnings for when we do make the
14762 2013-05-24 Dehao Chen <dehao@google.com>
14764 * tree-cfg.c (locus_discrim_map): Fix the typo.
14765 (locus_discrim_hasher): Likewise.
14766 (locus_discrim_hasher::hash): Likewise.
14767 (locus_discrim_hasher::equal): Likewise.
14769 2013-05-24 Martin Jambor <mjambor@suse.cz>
14771 PR tree-optimization/57294
14772 * cgraph.h (ipa_record_stmt_references): Declare.
14773 * cgraphbuild.c (ipa_record_stmt_references): New function.
14774 (build_cgraph_edges): Use ipa_record_stmt_references.
14775 (rebuild_cgraph_edges): Likewise.
14776 (cgraph_rebuild_references): Likewise.
14777 * ipa-prop.c (ipa_modify_call_arguments): Discard references
14778 associated with the old statement and build references from the
14779 newly built statements.
14780 * ipa-ref.c (ipa_remove_stmt_references): New function.
14781 * ipa-ref.h (ipa_remove_stmt_references): Declare.
14783 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
14785 * lra-constraints.c (emit_spill_move): Use smaller mode for
14787 (check_and_process_move): Consider mem-reg moves for secondary
14789 (curr_insn_transform): Don't lose insns emitted before for
14790 secondary memory moves.
14791 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
14792 reg set up in the current insn.
14794 2013-05-24 Dehao Chen <dehao@google.com>
14796 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
14798 (locus_descrim_hasher::equal): Likewise.
14799 (build_gimple_cfg): New discriminator assignment algorithm.
14800 (make_edges): Likewise.
14801 (next_discriminator_for_locus): Likewise.
14802 (same_line_p): Likewise.
14803 (assign_discriminators): Likewise.
14804 (make_cond_expr_edges): Likewise.
14805 (make_gimple_switch_edges): Likewise.
14806 (make_goto_expr_edges): Likewise.
14807 (make_gimple_asm_edges): Likewise.
14809 2013-05-24 Ian Bolton <ian.bolton@arm.com>
14811 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
14812 X format specifier to only display bottom 16 bits.
14813 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
14814 immediate to match for operand 2, since it will be masked.
14816 2013-05-24 Richard Biener <rguenther@suse.de>
14818 PR tree-optimization/57287
14819 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
14820 all SSA names that occur in abnormal PHIs.
14822 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
14825 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
14826 that index is not negative.
14828 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
14830 PR rtl-optimization/55177
14831 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
14832 (simplify_byte_swapping_operation): New.
14833 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
14834 (simplify_relational_operation_1): Deal with BSWAP.
14836 2013-05-23 Richard Henderson <rth@redhat.com>
14839 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
14840 (ix86_reorg): Call it.
14842 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
14845 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
14846 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
14847 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
14849 2013-05-23 Christian Bruel <christian.bruel@st.com>
14852 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
14854 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14855 Marcus Shawcroft <marcus.shawcroft@arm.com>
14857 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
14858 * config/aarch64/constraints.md (Usa): Remove.
14859 * doc/md.texi (AArch64 Usa): Remove.
14861 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14862 Marcus Shawcroft <marcus.shawcroft@arm.com>
14864 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
14865 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
14866 * config/aarch64/predicates.md (aarch64_const_address): Remove.
14867 (aarch64_mov_operand): Use aarch64_mov_operand_p.
14869 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
14871 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
14872 instruction (AdvSIMD).
14873 * config/aarch64/aarch64-builtins.c
14874 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
14875 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
14877 2013-05-23 Martin Jambor <mjambor@suse.cz>
14879 PR middle-end/57347
14880 * tree.h (contains_bitfld_component_ref_p): Declare.
14881 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
14882 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
14884 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
14885 not access a bit-field. Assert all final offsets are byte-aligned.
14887 2013-05-23 Richard Biener <rguenther@suse.de>
14889 PR tree-optimization/57380
14890 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
14891 least one invariant or re-used load.
14892 * passes.c (init_optimization_passes): Move pass_phiprop before
14895 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
14897 * config/aarch64/aarch64-simd.md
14898 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
14900 2013-05-23 Richard Biener <rguenther@suse.de>
14902 PR middle-end/57381
14903 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
14904 OEP_CONSTANT_ADDRESS_OF retained.
14906 2013-05-23 Jakub Jelinek <jakub@redhat.com>
14908 PR middle-end/57344
14909 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
14910 don't lower unit. Handle unit not being always BITS_PER_WORD.
14912 2013-05-23 Richard Biener <rguenther@suse.de>
14914 PR rtl-optimization/57341
14915 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
14916 instead of true_dependence.
14918 2013-05-22 David Malcolm <dmalcolm@redhat.com>
14920 * bb-reorder.c (branch_threshold): Make const.
14921 (exec_threshold): Ditto.
14923 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14924 Pat Haugen <pthaugen@us.ibm.com>
14925 Peter Bergner <bergner@vnet.ibm.com>
14927 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
14928 documentation for the power8 crypto builtins.
14930 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
14932 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
14933 macros for defining power8 builtin functions.
14934 (BU_P8V_AV_2): Likewise.
14935 (BU_P8V_AV_P): Likewise.
14936 (BU_P8V_VSX_1): Likewise.
14937 (BU_P8V_OVERLOAD_1): Likewise.
14938 (BU_P8V_OVERLOAD_2): Likewise.
14939 (BU_CRYPTO_1): Likewise.
14940 (BU_CRYPTO_2): Likewise.
14941 (BU_CRYPTO_3): Likewise.
14942 (BU_CRYPTO_OVERLOAD_1): Likewise.
14943 (BU_CRYPTO_OVERLOAD_2): Likewise.
14944 (XSCVSPDP): Fix typo, point to the correct instruction.
14945 (VCIPHER): Add power8 crypto builtins.
14946 (VCIPHERLAST): Likewise.
14947 (VNCIPHER): Likewise.
14948 (VNCIPHERLAST): Likewise.
14949 (VPMSUMB): Likewise.
14950 (VPMSUMH): Likewise.
14951 (VPMSUMW): Likewise.
14952 (VPERMXOR_V2DI): Likewise.
14953 (VPERMXOR_V4SI: Likewise.
14954 (VPERMXOR_V8HI: Likewise.
14955 (VPERMXOR_V16QI: Likewise.
14956 (VSHASIGMAW): Likewise.
14957 (VSHASIGMAD): Likewise.
14958 (VPMSUM): Likewise.
14959 (VPERMXOR): Likewise.
14960 (VSHASIGMA): Likewise.
14962 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
14963 __CRYPTO__ if the crypto instructions are available.
14964 (altivec_overloaded_builtins): Add support for overloaded power8
14967 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
14968 support for power8 crypto builtins.
14969 (builtin_function_type): Likewise.
14970 (altivec_init_builtins): Add support for builtins that take vector
14971 long long (V2DI) arguments.
14973 * config/rs6000/crypto.md: New file, define power8 crypto
14976 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14977 Pat Haugen <pthaugen@us.ibm.com>
14978 Peter Bergner <bergner@vnet.ibm.com>
14980 * doc/invoke.texi (Option Summary): Add power8 options.
14981 (RS/6000 and PowerPC Options): Likewise.
14983 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
14984 constraints.md instead of rs6000.h. Reorder w* constraints. Add
14985 wm, wn, wr documentation.
14987 * config/rs6000/constraints.md (wm): New constraint for VSX
14988 registers if direct move instructions are enabled.
14989 (wn): New constraint for no registers.
14990 (wq): New constraint for quad word even GPR registers.
14991 (wr): New constraint if 64-bit instructions are enabled.
14992 (wv): New constraint if power8 vector instructions are enabled.
14993 (wQ): New constraint for quad word memory locations.
14995 * config/rs6000/predicates.md (const_0_to_15_operand): New
14996 constraint for 0..15 for crypto instructions.
14997 (gpc_reg_operand): If VSX allow registers in VSX registers as well
14998 as GPR and floating point registers.
14999 (int_reg_operand): New predicate to match only GPR registers.
15000 (base_reg_operand): New predicate to match base registers.
15001 (quad_int_reg_operand): New predicate to match even GPR registers
15002 for quad memory operations.
15003 (vsx_reg_or_cint_operand): New predicate to allow vector logical
15004 operations in both GPR and VSX registers.
15005 (quad_memory_operand): New predicate for quad memory operations.
15006 (reg_or_indexed_operand): New predicate for direct move support.
15008 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
15009 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
15010 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
15011 (POWERPC_MASKS): Add power8 options.
15012 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
15015 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
15016 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
15018 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
15019 (-mpower8-fusion): New power8 options.
15020 (-mpower8-fusion-sign): Likewise.
15021 (-mpower8-vector): Likewise.
15022 (-mcrypto): Likewise.
15023 (-mdirect-move): Likewise.
15024 (-mquad-memory): Likewise.
15026 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
15027 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
15029 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
15030 (rs6000_debug_vector_unit): Add p8_vector.
15031 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
15032 definitions. Also print fusion state.
15033 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
15034 (rs6000_builtin_mask_calculate): Add power8 builtin support.
15035 (rs6000_option_override_internal): Add support for power8.
15036 (rs6000_common_init_builtins): Add debugging for skipped builtins
15037 if -mdebug=builtin.
15038 (rs6000_adjust_cost): Add power8 support.
15039 (rs6000_issue_rate): Likewise.
15040 (insn_must_be_first_in_group): Likewise.
15041 (insn_must_be_last_in_group): Likewise.
15042 (force_new_group): Likewise.
15043 (rs6000_register_move_cost): Likewise.
15044 (rs6000_opt_masks): Likewise.
15046 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
15047 power8 capable assembler, default to power7 options.
15048 (TARGET_DIRECT_MOVE): Likewise.
15049 (TARGET_CRYPTO): Likewise.
15050 (TARGET_P8_VECTOR): Likewise.
15051 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
15052 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
15053 (VECTOR_MEM_P8_VECTOR_P): Likewise.
15054 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
15055 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
15056 (TARGET_XSCVDPSPN): Likewise.
15057 (TARGET_XSCVSPDPN): Likewsie.
15058 (TARGET_SYNC_HI_QI): Likewise.
15059 (TARGET_SYNC_TI): Likewise.
15060 (MASK_CRYPTO): Likewise.
15061 (MASK_DIRECT_MOVE): Likewise.
15062 (MASK_P8_FUSION): Likewise.
15063 (MASK_P8_VECTOR): Likewise.
15064 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
15065 temporary used by some of the direct move instructions to get two FP
15066 temporary registers does not force creation of a stack frame.
15067 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
15068 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
15069 that any VSX registers are tieable, even if they are also an
15070 Altivec vector mode.
15071 (r6000_reg_class_enum): Add wm, wr, wv constraints.
15072 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
15073 (RS6000_BTM_CRYPTO): Likewise.
15074 (RS6000_BTM_COMMON): Likewise.
15076 * config/rs6000/rs6000.md (cpu attribute): Add power8.
15077 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
15078 (enum rs6000_vector): Add power8 vector support.
15080 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15084 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
15085 (any_sibcall_could_use_r3): this and handle indirect calls.
15086 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
15088 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15090 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
15092 2013-05-22 Richard Biener <rguenther@suse.de>
15094 PR middle-end/57349
15095 * profile.c (branch_prob): Do not split blocks that are
15096 abnormally receiving from ECF_RETURNS_TWICE functions.
15098 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15100 * recog.c (offsettable_address_addr_space_p): Fix calculation of
15101 address mode. Move pointer mode initialization to the same place.
15103 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
15105 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
15106 while it has any effect.
15108 2013-05-21 Easwaran Raman <eraman@google.com>
15110 PR tree-optimization/57322
15111 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
15112 UID of the statement added to the BB to be 1.
15114 2013-05-21 Jakub Jelinek <jakub@redhat.com>
15116 PR tree-optimization/57331
15117 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
15118 of conversion from pointer type to integral type with integer.
15120 2013-05-21 Martin Jambor <mjambor@suse.cz>
15123 * ipa-prop.c (ipa_read_node_info): Process param_used and
15124 controlled_uses in the same order as when writing.
15126 2013-05-21 Magnus Granberg <baldrick@free.fr>
15129 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
15131 2013-05-21 Richard Biener <rguenther@suse.de>
15133 PR tree-optimization/57318
15134 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
15135 estimate stmts with side-effects as likely eliminated.
15137 2013-05-21 Richard Biener <rguenther@suse.de>
15139 PR tree-optimization/57330
15140 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
15141 preserve the call stmts fntype.
15143 2013-05-21 Richard Biener <rguenther@suse.de>
15145 PR tree-optimization/57303
15146 * tree-ssa-sink.c (statement_sink_location): Improve killing
15147 stmt detection and properly handle self-assignments.
15149 2013-05-21 Christian Bruel <christian.bruel@st.com>
15151 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
15152 spanning registers. LEAF_REG_REMAP is supported only for contiguous
15153 registers. Set register size out of the PARALLEL loop.
15155 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
15158 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
15159 (*fmasf4, *fmasf4_media): New insns.
15161 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
15163 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
15164 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
15165 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
15166 (mips_idiv_insns): Update the comments to say that the returned
15167 instruction counts are in units of BASE_INSN_LENGTH.
15168 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
15169 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
15170 using 2 rather than 4 as the length of indirect MIPS16 and
15171 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
15172 length of a NOP. Don't divide MIPS16 lengths by 2.
15173 (mips16_split_long_branches): Assume a branch is long if the
15174 length is greater than 4 rather than 8.
15175 * config/mips/mips.md (length): Give MIPS16 lengths directly,
15176 rather than multiplying them by 2. Multiply instruction counts
15177 by BASE_INSN_LENGTH rather than 4.
15178 (*jump_mips16, tls_get_tp_mips16_<mode>)
15179 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
15181 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
15183 * config/mips/mips.md (extended_mips16): Remove branch case.
15184 (length): Remove duplicated extended_mips16 test.
15186 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
15188 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
15190 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
15192 * recog.h (Recog_data): Rename to...
15193 (recog_data_d): ...this.
15194 (recog_data): Update accordingly.
15195 * recog.c (recog_data): Likewise.
15196 * reload.c (save_recog_data): Likewise.
15197 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
15198 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
15200 2013-05-17 Julian Brown <julian@codesourcery.com>
15202 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
15203 found in a REG_EQUAL note, invalidate it.
15205 2013-05-17 Easwaran Raman <eraman@google.com>
15207 * tree-ssa-reassoc.c (find_insert_point): New function.
15208 (insert_stmt_after): Likewise.
15209 (get_def_stmt): Likewise.
15210 (ensure_ops_are_available): Likewise.
15211 (not_dominated_by): Likewise.
15212 (rewrite_expr_tree): Do not move statements beyond what is
15213 necessary. Remove call to swap_ops_for_binary_stmt...
15214 (reassociate_bb): ... and move it here.
15215 (build_and_add_sum): Assign UIDs for new statements.
15216 (linearize_expr): Likewise.
15217 (do_reassoc): Renumber gimple statement UIDs.
15219 2013-05-17 Jan Hubicka <jh@suse.cz>
15221 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
15223 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
15224 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
15226 (output_weakrefs): Update.
15228 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
15229 Martin Jambor <mjambor@suse.cz>
15231 PR middle-end/57276
15232 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
15233 value that corresponds to the given aggval is found in values vector.
15235 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
15237 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
15238 sse, sse2, sse3, ssse3 and sse4a flags to options.
15240 2013-05-17 David Malcolm <dmalcolm@redhat.com>
15242 * gengtype-state.c: (s_expr_writer): New class, to handle
15243 prettifying of output layout of s-expressions.
15244 (state_writer): New class, to write out gtype.state.
15245 (state_written_type_count): Move this variable into member data of
15247 (s_expr_writer::s_expr_writer): New code: constructor for new class
15248 (state_writer::state_writer(): ditto
15249 (s_expr_writer::write_new_line): New function
15250 (s_expr_writer::write_any_indent): ditto
15251 (s_expr_writer::begin_s_expr): ditto
15252 (s_expr_writer::end_s_expr): ditto
15253 (write_state_fileloc): convert to method of state_writer...
15254 (state_writer:: write_state_fileloc): ...and use methods of
15255 s_expr_writer to write indentation into the gtype.state output file
15256 to visually represent the hierarchical structure of the list
15258 (write_state_fields): ditto, renaming to...
15259 (state_writer::write_state_fields)
15260 (write_state_a_string): ditto, renaming to...
15261 (state_writer::write_state_a_string)
15262 (write_state_string_option): ditto, renaming to...
15263 (state_writer::write_state_string_option)
15264 (write_state_type_option): ditto, renaming to...
15265 (state_writer::write_state_type_option)
15266 (write_state_nested_option): ditto, renaming to...
15267 (state_writer::write_state_nested_option)
15268 (write_state_option): ditto, renaming to...
15269 (state_writer::write_state_option)
15270 (write_state_options): ditto, renaming to...
15271 (state_writer::write_state_options)
15272 (write_state_lang_bitmap): ditto, renaming to...
15273 (state_writer::write_state_lang_bitmap)
15274 (write_state_version): ditto, renaming to...
15275 (state_writer::write_state_version)
15276 (write_state_scalar_type): ditto, renaming to...
15277 (state_writer::write_state_scalar_type)
15278 (write_state_string_type): ditto, renaming to...
15279 (state_writer::write_state_string_type)
15280 (write_state_undefined_type): ditto, renaming to...
15281 (state_writer::write_state_undefined_type)
15282 (write_state_struct_union_type): ditto, renaming to...
15283 (state_writer::write_state_struct_union_type)
15284 (write_state_struct_type): ditto, renaming to...
15285 (state_writer::write_state_struct_type)
15286 (write_state_user_struct_type): ditto, renaming to...
15287 (state_writer::write_state_user_struct_type)
15288 (write_state_lang_struct_type): ditto, renaming to...
15289 (state_writer::write_state_lang_struct_type)
15290 (write_state_param_struct_type): ditto, renaming to...
15291 (state_writer::write_state_param_struct_type)
15292 (write_state_pointer_type): ditto, renaming to...
15293 (state_writer::write_state_pointer_type)
15294 (write_state_array_type): ditto, renaming to...
15295 (state_writer::write_state_array_type)
15296 (write_state_gc_used): ditto, renaming to...
15297 (state_writer::write_state_gc_used)
15298 (write_state_common_type_content): ditto, renaming to...
15299 (state_writer::write_state_common_type_content)
15300 (write_state_type): ditto, renaming to...
15301 (state_writer::write_state_type)
15302 (write_state_pair_list): ditto, renaming to...
15303 (state_writer::write_state_pair_list)
15304 (write_state_pair): ditto, renaming to...
15305 (state_writer::write_state_pair)
15306 (write_state_typedefs): ditto, renaming to...
15307 (state_writer::write_state_typedefs)
15308 (write_state_structures): ditto, renaming to...
15309 (state_writer::write_state_structures)
15310 (write_state_param_structs): ditto, renaming to...
15311 (state_writer::write_state_param_structs)
15312 (write_state_variables): ditto, renaming to...
15313 (state_writer::write_state_variables)
15314 (write_state_srcdir): ditto, renaming to...
15315 (state_writer::write_state_srcdir)
15316 (write_state_files_list): ditto, renaming to...
15317 (state_writer::write_state_files_list)
15318 (write_state_languages): ditto, renaming to...
15319 (state_writer::write_state_languages)
15320 (write_state): create a state_writer instance and use it when
15321 writing out the state file
15323 2013-05-17 Mike Stump <mikestump@comcast.net>
15325 PR rtl-optimization/57304
15326 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
15327 accessing DF_REF_REAL_LOC.
15329 2013-05-17 Jakub Jelinek <jakub@redhat.com>
15331 PR rtl-optimization/57281
15332 PR rtl-optimization/57300
15333 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
15334 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
15335 what the other splitter did if the registers are dead.
15337 2013-05-17 Richard Biener <rguenther@suse.de>
15339 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
15342 2013-05-17 Jakub Jelinek <jakub@redhat.com>
15344 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
15347 2013-05-17 Marek Polacek <polacek@redhat.com>
15349 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
15350 length when doing non-zero store of storing '\0' to '\0'.
15352 2013-05-17 Jakub Jelinek <jakub@redhat.com>
15354 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
15355 vect_external_def oprnd1 with loop_vinfo, try to emit
15356 optional cast, negation and and stmts on the loop preheader
15357 edge instead of into the pattern def seq.
15359 PR tree-optimization/57051
15360 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
15361 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
15363 2013-05-16 Nick Clifton <nickc@redhat.com>
15365 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
15366 (rl78_is_naked_func): New function.
15367 (rl78_expand_prologue): Skip prologue generation for naked functions.
15368 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
15369 * doc/extend.texi (naked): Add RL78 to the list of processors
15370 that supports this attribute.
15372 2013-05-16 Jeff Law <law@redhat.com>
15374 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
15376 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
15378 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
15379 cache parameters using detect_caches_amd also for CYRIX,
15380 NSC and TM2 signatures.
15382 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
15383 Dzianis Kahanovich <mahatma@eu.by>
15387 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
15388 VIA/Centaur processors and determine their cache parameters
15389 using detect_caches_amd.
15391 2013-05-16 Teresa Johnson <tejohnson@google.com>
15393 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
15394 (rtl_verify_edges): New function.
15395 (rtl_verify_bb_insns): Ditto.
15396 (rtl_verify_bb_pointers): Ditto.
15397 (rtl_verify_bb_insn_chain): Ditto.
15398 (rtl_verify_fallthru): Ditto.
15399 (rtl_verify_bb_layout): Ditto.
15400 (rtl_verify_flow_info_1): Outline checks into new functions.
15401 (rtl_verify_flow_info): Ditto.
15403 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
15405 * cfghooks.c (copy_bbs): Add update_dominance argument.
15406 * cfghooks.h (copy_bbs): Update prototype.
15407 * tree-cfg.c (gimple_duplicate_sese_region):
15408 Add update_dominance argument.
15409 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
15410 * tree-ssa-loop-ch.c (copy_loop_headers): Update
15411 gimple_duplicate_sese_region call.
15412 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
15413 Update copy_bbs call.
15414 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
15415 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
15417 2013-05-16 Jakub Jelinek <jakub@redhat.com>
15419 * tree-vectorizer.h (NUM_PATTERNS): Increment.
15420 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
15421 vect_recog_rotate_pattern.
15422 (vect_recog_rotate_pattern): New function.
15424 2013-05-16 Jason Merrill <jason@redhat.com>
15426 * Makefile.in (LLINKER): New variable.
15427 (mostlyclean): Remove link mutex.
15428 * configure.ac: Handle --enable-link-mutex.
15429 * lock-and-run.sh: New script.
15431 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15434 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
15437 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15439 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
15441 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
15443 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
15444 * config/arm/arm.c (next_consecutive_mem): New function.
15445 (gen_movmem_ldrd_strd): Likewise.
15446 * config/arm/arm.md (movmemqi): Update condition and code.
15447 (unaligned_loaddi, unaligned_storedi): New patterns.
15449 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15451 * config.gcc: Obsolete *-*-solaris2.9*.
15452 * doc/install.texi (Specific, *-*-solaris2*): Document it.
15454 2013-05-16 Richard Biener <rguenther@suse.de>
15456 * passes.c (init_optimization_passes): Move pass_parallelize_loops
15457 earlier, after GRAPHITE transforms and IV canonicalization.
15459 2013-05-16 Jakub Jelinek <jakub@redhat.com>
15461 * omp-low.c (extract_omp_for_data): For collapsed loops,
15462 if at least one of the loops is known at compile time to
15463 iterate zero times, set count to 0.
15464 (expand_omp_regimplify_p): New function.
15465 (expand_omp_for_generic): For collapsed loops, if at least
15466 one of the loops isn't known to iterate at least once,
15467 add runtime check with setting count to 0.
15468 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
15469 For unsigned types if it isn't known at compile time that
15470 the loop will iterate at least once, add runtime check to bypass
15471 the whole loop if initial condition isn't true.
15473 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
15475 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
15477 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
15479 PR middle-end/57286
15480 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
15481 transformations to avoid an infinite loop.
15483 2013-05-16 Marek Polacek <polacek@redhat.com>
15485 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
15487 2013-05-15 Leif Ekblad <leif@rdos.net>
15489 * config/i386/i386.c (ix86_decompose_address): Use
15490 DEFAULT_TLS_SEG_REG to access TLS segment register.
15491 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
15492 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
15493 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
15495 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
15498 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
15499 sibling calls to functions that would normally be lazily bound,
15500 unless $gp is call-clobbered.
15502 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
15504 * config/i386/i386.c (ix86_option_override_internal): Update
15505 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
15506 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
15507 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
15508 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
15510 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
15512 2013-05-15 Andreas Schwab <schwab@suse.de>
15514 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
15515 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
15518 2013-05-15 Teresa Johnson <tejohnson@google.com>
15520 * loop-unroll.c (report_unroll_peel): Check decision before
15521 emitting unroll/peel message.
15523 2013-05-15 Teresa Johnson <tejohnson@google.com>
15525 * function.h (has_bb_partition): New rtl_data flag.
15526 (bb_reorder_complete): Ditto.
15527 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
15528 instead of flag_reorder_blocks_and_partition.
15529 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
15530 with some enhancements.
15531 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
15532 * bb-reorder.c (connect_traces): Check for has_bb_partition
15533 instead of flag_reorder_blocks_and_partition.
15534 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
15535 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
15536 verify_hot_cold_block_grouping.
15537 (partition_hot_cold_basic_blocks): Set has_bb_partition.
15539 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15542 * config/arm/predicates.md (call_insn_operand): New predicate.
15543 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
15544 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
15545 if insn is not a tail call.
15546 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
15548 * config/arm/arm.h (enum reg_class): New caller save register class.
15549 (REG_CLASS_NAMES): Likewise.
15550 (REG_CLASS_CONTENTS): Likewise.
15551 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
15554 2013-05-15 Richard Biener <rguenther@suse.de>
15556 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
15557 of MSG_OPTIMIZED_LOCATIONS.
15558 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
15559 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
15561 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
15562 of MSG_OPTIMIZED_LOCATIONS.
15563 (execute_vect_slp): Likewise.
15564 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
15565 (vect_create_cond_for_alias_checks): Likewise.
15566 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
15567 (vect_recog_widen_mult_pattern): Likewise.
15568 (vect_recog_widen_sum_pattern): Likewise.
15569 (vect_recog_over_widening_pattern): Likewise.
15570 (vect_recog_widen_shift_pattern): Likewise.
15571 (vect_recog_vector_vector_shift_pattern): Likewise.
15572 (vect_recog_divmod_pattern): Likewise.
15573 (vect_recog_mixed_size_cond_pattern): Likewise.
15574 (vect_recog_bool_pattern): Likewise.
15575 (vect_pattern_recog_1): Likewise.
15577 2013-05-15 Martin Jambor <mjambor@suse.cz>
15579 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
15580 non-functions to builtin_unreachable.
15581 * ipa-inline-transform.c (inline_call): Do not assert estimates were
15582 correct when new direct edges were discovered.
15584 2013-05-15 Martin Jambor <mjambor@suse.cz>
15586 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
15587 header, print symbol order instead of node uid, print more information
15588 about indirect edge targets.
15589 (ipa_make_edge_direct_to_target): Print symbol order instead of node
15591 (ipa_make_edge_direct_to_target): Likewise.
15592 (remove_described_reference): Likewise.
15593 (propagate_controlled_uses): Likewise.
15594 (ipa_print_node_params): Also print symbol order.
15595 (ipcp_transform_function): Print symbol order instead of node uids.
15596 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
15597 (cgraph_get_create_real_symbol_node): Likewise.
15598 * ipa-cp.c (print_lattice): Likewise.
15599 (print_all_lattices): Likewise.
15600 (determine_versionability): Likewise.
15601 (initialize_node_lattices): Likewise.
15602 (estimate_local_effects): Likewise.
15603 (update_profiling_info): Likewise.
15604 (create_specialized_node): Likewise.
15605 (perhaps_add_new_callers): Likewise.
15606 (decide_about_value): Likewise.
15607 (decide_whether_version_node): Likewise.
15608 (identify_dead_nodes): Likewise.
15609 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
15610 (dump_inline_summary): Likewise.
15611 (estimate_node_size_and_time): Likewise.
15612 (inline_analyze_function): Likewise.
15613 * ipa-inline.c (report_inline_failed_reason): Likewise.
15614 (want_early_inline_function_p): Likewise.
15615 (edge_badness): Likewise.
15616 (update_edge_key): Likewise.
15617 (inline_small_functions): Likewise. Add dumping of order to two other
15619 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
15620 instead of node uids.
15621 (propagate_pure_const): Likewise.
15622 (propagate_pure_const): Likewise.
15623 * ipa-utils.c (dump_cgraph_node_set): Likewise.
15624 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
15625 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
15627 * tree-pretty-print.c (dump_function_header): Likewise.
15628 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
15629 Print symbol order instead of node uids.
15631 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15633 * config/s390/s390.c (s390_register_move_cost): Don't impose the
15634 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
15636 2013-05-15 Richard Biener <rguenther@suse.de>
15638 PR tree-optimization/57275
15639 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
15640 return value for fail to do runtime alias checks for gather loads.
15642 2013-05-15 Jan Hubicka <jh@suse.cz>
15646 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
15647 weakrefs are not external.
15648 (lto_symtab_merge_decls): Fix thinko when dealing with
15649 non-lto_symtab decls.
15650 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
15651 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
15652 * varpool.c (dump_varpool_node): Dump more flags.
15654 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15656 * config/i386/i386.c (processor_alias_table): Add instruction
15657 FSGSBASE for AMD bdver3 architecture.
15659 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15661 * tree.c (warn_deprecated_use): Print file:line using locus color.
15662 * diagnostic.c (diagnostic_report_current_module): Print file:line
15663 and file:line:column using locus color.
15665 2013-05-14 Mike Stump <mikestump@comcast.net>
15667 * gdbinit.in: Add __null.
15669 2013-05-14 Mike Stump <mikestump@comcast.net>
15671 * recog.h: Rename struct recog_data to Recog_data.
15672 * recog.c: Likewise.
15673 * reload.c (can_reload_into): Likewise.
15674 * config/picochip/picochip.c: Likewise.
15676 2013-05-14 Mike Stump <mikestump@comcast.net>
15678 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
15680 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
15682 * resource.h (struct resources): Remove unch_memory member.
15683 (CLEAR_RESOURCE): Don't clear unch_memory.
15684 * resource.c (mark_referenced_resources): Don't set it.
15685 (mark_set_resources): Likewise.
15686 (mark_target_live_regs): Don't clear it.
15687 (init_resource_info): Likewise.
15688 * reorg.c (resource_conflicts_p): Don't compare it.
15689 (redundant_insn): Don't set it.
15691 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
15693 * emit-rtl.c (next_label): Remove unused function.
15694 (skip_consecutive_labels, link_cc0_insns): Move to ...
15695 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
15696 only place where these functions are used, and make them static.
15698 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
15700 * fold-const.c (fold_negate_expr): Handle vectors.
15701 (fold_truth_not_expr): Make it static.
15702 (fold_invert_truthvalue): New static function.
15703 (invert_truthvalue_loc): Handle vectors. Do not call
15704 fold_truth_not_expr directly.
15705 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
15706 <TRUTH_NOT_EXPR>: Do not cast to boolean.
15707 (fold_comparison): Handle vector constants.
15708 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
15709 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
15710 * tree.h (fold_truth_not_expr): Remove declaration.
15712 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
15714 * config/aarch64/aarch64-simd.md
15715 (aarch64_vcond_internal<mode>): Rename to...
15716 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
15717 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
15718 float modes. Clarify all iterator modes.
15719 (vcond<mode><mode>): Use new name for vcond expanders.
15720 (vcond<v_cmp_result><mode>): Likewise.
15721 (vcondu<mode><mode>: Likewise.
15722 * config/aarch64/iterators.md (VDQF_COND): New.
15724 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
15727 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
15728 variable for the shift amount. Check that we shift by non-negative
15731 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
15734 * config/arm/arm.h (EPILOGUE_USES): Only return true
15735 for LR_REGNUM after epilogue_completed.
15737 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
15739 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
15740 is error_mark_node.
15742 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15745 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
15746 and Solaris 11+/x86 with gld.
15747 * configure: Regenerate.
15749 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15751 * expmed.c (expand_shift_1): Canonicalize rotates by
15752 constant bitsize / 2 to bitsize - 1.
15753 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
15754 case ROTATERT>: Likewise.
15757 2013-05-10 Jakub Jelinek <jakub@redhat.com>
15759 * config/i386/i386.md (rotateinv): New code attr.
15760 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
15761 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
15762 roll $31, %eax, etc.
15764 2013-05-14 Richard Biener <rguenther@suse.de>
15766 PR middle-end/57235
15767 * tree-eh.c (sink_clobbers): Give up for successors with
15768 multiple predecessors and no virtual uses.
15770 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
15772 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
15773 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
15775 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15777 PR middle-end/57251
15778 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
15779 the case when both op0 and op1 have VOIDmode.
15781 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15783 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
15784 in multiply-accumulate mode.
15786 2013-05-13 Guozhi Wei <carrot@google.com>
15788 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
15790 2013-05-13 Kai Tietz <ktietz@redhat.com>
15793 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
15794 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
15795 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
15796 * config/i386/i386.c (ix86_option_override_internal): Likewise.
15797 (ix86_expand_prologue): Likewise.
15798 (ix86_expand_split_stack_prologue): Likewise.
15799 (legitimate_pic_address_disp_p): Likewise.
15800 (legitimize_pic_address): Likewise.
15801 (legitimize_tls_address): Likewise.
15802 (legitimize_pe_coff_symbol): Likewise.
15803 (output_pic_addr_const): Likewise.
15804 (construct_plt_address): Likewise.
15805 (ix86_expand_call): Likewise.
15806 (x86_output_mi_thunk): Likewise.
15807 (x86_function_profiler): Likewise.
15809 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
15811 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
15812 similar switch cases.
15813 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
15814 (aarch64_simd_mov_to_<mode>low): Delete.
15815 (aarch64_simd_mov_to_<mode>high): Delete.
15816 (move_lo_quad_<mode>): Add w<-r alternative.
15817 (aarch64_simd_move_hi_quad_<mode>): Likewise.
15818 (aarch64_simd_mov_from_*): Update type attribute.
15819 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
15822 2013-05-13 Jan Hubicka <jh@suse.cz>
15824 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
15825 * config/i386/i386.c (ix86_compute_frame_layout,
15826 ix86_expand_epilogue, emit_i387_cw_initialization,
15827 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
15828 ix86_local_alignment): Fix use of size/speed predicates.
15830 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15832 PR tree-optimization/45216
15833 PR tree-optimization/57157
15834 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
15835 the (-Y) & (B - 1) variant if OP is |.
15836 * expmed.c (expand_shift_1): For rotations by const0_rtx just
15837 return shifted. Use (-op1) & (prec - 1) as other_amount
15838 instead of prec - op1.
15840 2013-05-13 Martin Jambor <mjambor@suse.cz>
15842 PR middle-end/42371
15843 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
15844 (ipa_constant_data): New type.
15845 (ipa_jump_func): Use ipa_constant_data to hold information about
15846 constant jump functions.
15847 (ipa_get_jf_constant): Adjust to jump function type changes.
15848 (ipa_get_jf_constant_rdesc): New function.
15849 (ipa_param_descriptor): New field controlled_uses.
15850 (ipa_get_controlled_uses): New function.
15851 (ipa_set_controlled_uses): Likewise.
15852 * ipa-ref.h (ipa_find_reference): Declare.
15853 * ipa-prop.c (ipa_cst_ref_desc): New type.
15854 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
15856 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
15857 New parameter cs. Adjust all callers.
15858 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
15859 (remove_described_reference): New function.
15860 (jfunc_rdesc_usable): Likewise.
15861 (try_make_edge_direct_simple_call): Decrement controlled use count,
15862 attempt to remove reference if it hits zero.
15863 (combine_controlled_uses_counters): New function.
15864 (propagate_controlled_uses): Likewise.
15865 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
15866 (ipa_edge_duplication_hook): Duplicate reference descriptions.
15867 (ipa_print_node_params): Print described use counter.
15868 (ipa_write_jump_function): Adjust to jump function type changes.
15869 (ipa_read_jump_function): New parameter CS, pass it to
15870 ipa_set_jf_constant. Adjust caller.
15871 (ipa_write_node_info): Stream controlled use count
15872 (ipa_read_node_info): Likewise.
15873 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
15875 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
15876 count. Remove cloning-added reference if it reaches zero.
15877 * ipa-ref.c (ipa_find_reference): New function.
15879 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15881 * config/i386/i386.c (processor_target_table): Modified default
15882 alignment values for AMD BD and BT architectures.
15884 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
15886 * tree-vect-generic.c (uniform_vector_p): Move ...
15887 * tree.c (uniform_vector_p): ... here.
15888 * tree.h (uniform_vector_p): Declare it.
15889 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
15892 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15894 PR tree-optimization/57230
15895 * tree-ssa-strlen.c (handle_char_store): Record length for
15896 array store from STRING_CST.
15898 PR tree-optimization/57230
15899 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
15902 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
15904 * config/epiphany/epiphany.c (epiphany_init): Check size of
15905 NUM_MODES_FOR_MODE_SWITCHING.
15906 (epiphany_expand_prologue):
15907 Remove CONFIG_REGNUM initial value handling code.
15908 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
15909 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
15910 (emit_set_fp_mode, epiphany_mode_after): Likewise.
15911 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
15912 Don't return 1 for FP_MODE_NONE.
15913 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
15914 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
15915 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
15916 * config/epiphany/epiphany.md (save_config): New pattern.
15918 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
15920 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
15922 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
15924 * config/i386/i386.md (memory): Handle sseishft1.
15925 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
15926 (*vec_extractv2di_1): Ditto.
15928 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
15930 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
15933 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
15935 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
15936 Add mthumb/march=armv7-a multilib.
15937 Add mthumb/march=armv7-r multilib.
15938 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
15940 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
15942 * config/v850/t-rtems: Add more multilibs.
15944 2013-05-10 Richard Biener <rguenther@suse.de>
15946 PR tree-optimization/57214
15947 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
15948 not propagate from SSA names that occur in abnormal PHI nodes.
15950 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
15952 * stor-layout.c (element_precision): New function.
15953 * machmode.h (element_precision): Declare it.
15954 * tree.c (build_minus_one_cst): New function.
15955 (element_precision): Likewise.
15956 * tree.h (build_minus_one_cst): Declare new function.
15957 (element_precision): Likewise.
15958 * fold-const.c (operand_equal_p): Use element_precision.
15959 (fold_binary_loc): Handle vector types.
15960 * convert.c (convert_to_integer): Use element_precision.
15961 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
15964 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
15966 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
15967 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
15968 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
15969 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15970 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
15971 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
15972 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
15973 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15974 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
15975 (Uuw8): New constraints.
15976 (Usb4): Move into alphabetical order.
15977 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
15978 (sd8_operand, ub8_operand, uw8_operand): New predicates.
15979 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
15980 previously unnamed patterns.
15981 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
15982 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
15983 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
15984 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
15985 of set_attr_alternative/if_then_else. Use extended_mips16 instead
15986 of specific lengths.
15988 2013-05-10 Jakub Jelinek <jakub@redhat.com>
15990 * config/i386/i386.md (rotateinv): New code attr.
15991 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
15992 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
15993 roll $31, %eax, etc.
15995 PR tree-optimization/45216
15996 PR tree-optimization/57157
15997 * tree-ssa-forwprop.c (simplify_rotate): New function.
15998 (ssa_forward_propagate_and_combine): Call it.
16000 2013-05-10 Richard Biener <rguenther@suse.de>
16002 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
16003 disable peeling when we version for aliasing.
16004 (vector_alignment_reachable_p): Honor explicit user alignment.
16005 (vect_supportable_dr_alignment): Likewise.
16006 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
16007 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
16008 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
16009 then peeling to arrange for the cost-model check to come first.
16011 2013-05-10 Alan Modra <amodra@gmail.com>
16013 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
16014 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
16015 * configure: Regenerate.
16017 2013-05-10 Alan Modra <amodra@gmail.com>
16020 * varasm.c (default_elf_select_section): Move !DECL_P check..
16021 (get_named_section): ..to here before calling get_section_name.
16023 (default_section_type_flags): Add DECL_P check.
16024 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
16025 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
16027 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
16029 * config/epiphany/epiphany.c (epiphany_expand_prologue):
16030 When using gen_stack_adjust_str with a register offset, add a
16031 REG_FRAME_RELATED_EXPR note.
16033 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
16035 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
16036 (*vec_extractv4si_zext_mem): Ditto.
16037 (*vec_extractv2di): Add 0->x and x->x alternatives.
16038 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
16039 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
16041 2013-05-09 Jason Merrill <jason@redhat.com>
16043 N3639 C++1y VLA support
16044 * gimplify.c (gimplify_vla_decl): Don't touch an existing
16047 * tree.c (build_constructor_va): New.
16048 * tree.h: Declare it.
16050 2013-05-09 Martin Jambor <mjambor@suse.cz>
16053 * gimple-fold.c (canonicalize_constructor_val): Call
16054 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
16056 2013-05-09 Jan Hubicka <jh@suse.cz>
16057 Richard Biener <rguenther@suse.de>
16060 * symtab.c (symtab_make_decl_local): Do not add private names.
16062 2013-05-09 Jan Hubicka <jh@suse.cz>
16065 * symtab.c (insert_to_assembler_name_hash): Handle clones.
16066 (unlink_from_assembler_name_hash): Likewise.
16067 (symtab_prevail_in_asm_name_hash, symtab_register_node,
16068 symtab_unregister_node, symtab_initialize_asm_name_hash,
16069 change_decl_assembler_name): Update.
16071 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
16073 * config/aarch64/aarch64.md: New movtf split.
16074 (*movtf_aarch64): Update.
16075 (aarch64_movdi_tilow): Handle TF modes and rename to
16076 aarch64_movdi_<mode>low.
16077 (aarch64_movdi_tihigh): Handle TF modes and rename to
16078 aarch64_movdi_<mode>high
16079 (aarch64_movtihigh_di): Handle TF modes and rename to
16080 aarch64_mov<mode>high_di
16081 (aarch64_movtilow_di): Handle TF modes and rename to
16082 aarch64_mov<mode>low_di
16083 (aarch64_movtilow_tilow): Remove spurious whitespace.
16084 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
16086 (aarch64_print_operand): Update.
16088 2013-05-09 Alan Modra <amodra@gmail.com>
16090 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
16092 * configure: Regenerate.
16094 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
16096 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
16097 splitter preparation statements.
16098 * config/i386/sse.md (*vec_extract* splitters): Ditto.
16099 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
16102 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16104 * gimple-ssa-strength-reduction.c (count_candidates): Change
16105 return value to int.
16106 (analyze_candidates_and_replace): Change type of length to int.
16108 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
16110 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
16111 (*vec_extract<mode>): Use VI12_128 mode iterator.
16112 (*vec_extract<mode>_mem): Ditto.
16113 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
16116 2013-05-08 Diego Novillo <dnovillo@google.com>
16121 2012-08-17 Diego Novillo <dnovillo@google.com>
16124 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
16125 * config.in: Regenerate.
16126 * configure: Regenerate.
16127 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
16129 2013-05-08 Jan Hubicka <jh@suse.cz>
16132 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
16133 * cgraph.h (symtab_node_base): Add unique_name.
16134 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
16135 input_overwrite_node, input_varpool_node): Stream unique_name.
16136 * cgraphclones.c (cgraph_create_virtual_clone,
16137 cgraph_function_versioning): Set unique_name.
16138 * ipa.c (function_and_variable_visibility): Set unique_name.
16140 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16142 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
16143 (alloc_cand_and_find_basis): Restrict conditional candidate
16144 processing to CAND_MULTs.
16146 2013-05-08 Jan Hubicka <jh@suse.cz>
16149 lto-symtab.c (lto_symtab_symbol_p): New function.
16150 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
16151 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
16152 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
16153 Skip static symbols.
16155 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
16157 PR tree-optimization/57200
16158 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
16159 Only call inform if the preceding warning_at returns true.
16161 2013-05-07 Han Shen <shenhan@google.com>
16163 * cfgexpand.c (record_or_union_type_has_array_p): New function.
16164 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
16165 * common.opt (fstack-protector-strong): New option.
16166 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
16167 * doc/invoke.texi (Optimization Options): Document
16168 "-fstack-protector-strong".
16169 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
16171 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
16173 * config/mips/mips.c (mips_machine_reorg2): Return 0.
16175 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
16177 * ira.c (update_equiv_regs): Add insn having equiv memory even if
16178 it is not lhs of the insn.
16179 (setup_reg_equiv): Remove insn having equiv memory which it is not
16181 * lra-constraints.c (process_address): Try to improve generation
16182 code for address base + disp.
16183 (lra_constraints): Make correct the code for checking insn setting
16184 up backward equivalence. Remove insn only if it is in the init
16186 * lra-eliminations.c (update_reg_eliminate): Change return value.
16187 (lra_eliminate): Use the result.
16189 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
16191 * config/i386/sse.md (ssescalarnummask): New mode attribute.
16192 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
16193 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
16194 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
16195 register target operands.
16196 (*vec_extractv8hi_sse2): New pattern.
16197 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
16198 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
16199 (*vec_extract<mode>_mem): New insn and split pattern.
16201 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
16203 * config/arm/arm.c (arm_asan_shadow_offset): New function.
16204 (TARGET_ASAN_SHADOW_OFFSET): Define.
16205 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
16206 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
16208 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16210 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
16211 (incr_vec_index): Return -1 if increment not found.
16212 (create_add_on_incoming_edge): Assert if increment not found.
16213 (record_increment): Limit number of increments recorded.
16214 (all_phi_incrs_profitable): Return false if an increment not found.
16215 (replace_profitable_candidates): Don't process increments that were
16217 (analyze_candidates_and_replace): Limit size of incr_vec.
16219 2013-05-07 Richard Biener <rguenther@suse.de>
16221 * calls.c (special_function_p): setjmp-like functions are leaf.
16222 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
16223 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
16225 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
16227 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
16228 (aarch64_simd_mov<mode>): New expander.
16229 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
16230 (aarch64_simd_mov_to_<mode>high): Likewise.
16231 (aarch64_simd_mov_from_<mode>low): Likewise.
16232 (aarch64_simd_mov_from_<mode>high): Likewise.
16233 (aarch64_dup_lane<mode>): Update.
16234 (aarch64_dup_lanedi): New instruction pattern.
16235 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
16236 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
16238 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16240 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
16241 (replace_mult_candidate): Remove unnecessary argument; remove
16242 unnecessary parameter from call to introduce_cast_before_cand.
16243 (replace_unconditional_candidate): Remove unnecessary parameter
16244 from call to replace_mult_candidate.
16245 (replace_conditional_candidate): Likewise.
16246 (insert_initializers): Use make_temp_ssa_name.
16247 (introduce_cast_before_cand): Remove unnecessary argument; use
16248 make_temp_ssa_name.
16249 (replace_one_candidate): Remove unnecessary argument; remove
16250 unnecessary parameter from calls to introduce_cast_before_cand.
16251 (replace_profitable_candidates): Remove unnecessary parameters
16252 from calls to replace_one_candidate.
16254 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16256 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
16257 phi def as possibly hiding a basis for a CAND_ADD whose operands
16258 have been commuted in the analysis.
16259 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
16261 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16263 * config/aarch64/aarch64.md
16264 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
16265 shift value between 0-4.
16267 2013-05-07 Richard Biener <rguenther@suse.de>
16269 * double-int.h (rshift): New overload.
16270 * double-int.c (rshift): New function.
16271 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
16272 (create_reference_ops_from_ref): Remove.
16273 (vn_reference_insert): Use shared ops for constructing the
16274 reference and copy it.
16276 2013-05-07 Richard Biener <rguenther@suse.de>
16278 PR middle-end/57190
16279 * tree-eh.c (sink_clobbers): Properly propagate
16280 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
16282 2013-05-07 Jakub Jelinek <jakub@redhat.com>
16284 PR tree-optimization/57149
16285 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
16286 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
16287 collect_phi_def_edges, execute_late_warn_uninitialized): Use
16288 uninit_undefined_value_p instead of ssa_undefined_value_p.
16291 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
16292 for modifier == EXPAND_INITIALIZER.
16294 2013-05-07 Anton Blanchard <anton@samba.org>
16296 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
16297 for powerpc64 little endian.
16298 * configure: Regenerate.
16300 2013-05-06 Graham Stott <grahams@btinternet.com>
16302 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
16303 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
16304 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
16305 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
16307 2013-05-06 Graham Stott <grahams@btinternet.com>
16309 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
16310 codes which allow non-lvalues.
16312 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
16314 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
16315 components are all 1s.
16316 (integer_minus_onep): New function.
16317 * tree.h (integer_minus_onep): Declare it.
16318 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
16319 integer_minus_onep instead of integer_all_onesp.
16321 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
16324 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
16325 variations of these patterns.
16327 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
16329 * config/i386/i386.md (isa): Add x64_sse4 member.
16330 (enabled): Handle x64_sse4.
16331 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
16332 instruction for 64bit SSE4_1 targets. Update insn attributes.
16333 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
16334 instruction for SSE4_1 targets. Update insn attributes.
16335 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
16336 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
16337 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
16339 (*vec_extractv4si): Rename from *sse4_1_pextrd.
16340 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
16341 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
16343 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
16346 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
16348 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
16350 * final.c (do_assembler_dialects): Don't handle curly braces and
16351 vertical bar escaped by % as dialect delimiters.
16352 (output_asm_insn): Print curly braces and vertical bar if escaped
16353 by % and ASSEMBLER_DIALECT defined.
16354 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
16355 * doc/tm.texi: Regenerated.
16357 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
16359 * config/mips/mips.c: Include tree-pass.h.
16360 (mips_reorg): Split in pre- and post-dbr_schedule parts.
16361 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
16362 (pass_mips_machine_reorg2): New machine specific pass.
16363 (insert_pass_mips_machine_reorg2): New pass plugin definition.
16364 (mips_option_override): Register the new pass.
16365 * rtl.h (cleanup_barriers): Remove prototype.
16366 (dbr_schedule): Likewise.
16367 * jump.c (cleanup_barriers): Make static.
16368 * reorg.c (dbr_schedule): Likewise.
16370 2013-05-06 Richard Biener <rguenther@suse.de>
16372 PR tree-optimization/57185
16373 * tree-parloops.c (add_field_for_reduction): Handle anonymous
16374 SSA names properly.
16376 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
16379 * config/i386/i386.c (add_parameter_dependencies): Add dependence
16380 between "first_arg" and "insn", not "last" and "insn".
16382 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16384 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
16385 (find_candidates_in_block): Re-enable slsr_process_phi.
16386 (create_phi_basis): Fix double counting of candidate adjustment.
16388 2013-05-06 Richard Biener <rguenther@suse.de>
16390 PR middle-end/57147
16391 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
16392 the edge is also fallthru, preserve it and just clear the
16394 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
16395 also complex, preserve that and just clear the fallthru flag.
16396 * tree-inline.c (update_ssa_across_abnormal_edges): Also
16397 update virtual operands.
16399 2013-05-06 Alan Modra <amodra@gmail.com>
16401 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
16402 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
16403 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
16404 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
16405 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
16406 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
16408 2013-05-06 Alan Modra <amodra@gmail.com>
16410 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
16411 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
16412 (DEFAULT_ASM_ENDIAN): Define.
16413 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
16414 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
16415 Update -K PIC clause from sysv4.h.
16416 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
16417 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
16419 2013-05-06 Alan Modra <amodra@gmail.com>
16421 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
16422 twice for little-endian.
16423 (ashrdi3_no_power, ashrdi3): Support little-endian.
16425 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
16428 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
16429 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
16431 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
16432 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
16433 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
16436 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
16437 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16439 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
16440 * config/pa/pa.opt: Make mbig-switch a no-op.
16441 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
16442 (CASE_VECTOR_MODE): Always return SImode.
16443 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
16444 for the !TARGET_BIG_SWITCH case.
16445 * config/pa/pa-linux.h: Likewise.
16446 * config/pa/pa-openbsd.h: Likewise.
16447 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
16448 * config/pa/pa.md (short_jump): Remove define_insn.
16449 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
16450 (casesi0): Remove define_insn.
16451 (type): Remove btable_branch.
16452 (pa_combine_type): Likewise.
16453 (in_nullified_branch_delay): Likewise.
16454 (in_call_delay): Likewise.
16455 (define_delay): Likewise.
16456 (define_insn_reservation "Z3"): Likewise.
16457 (define_insn_reservation "Z4"): Likewise.
16458 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
16459 (pa_adjust_insn_length): Remove adjustment for btable branches.
16460 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
16463 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
16465 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
16466 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
16467 Add m->r,x alternatives.
16468 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
16469 splitters using SWI48x mode iterator.
16470 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
16471 TARGET_64BIT. Add m->x alternative.
16472 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
16473 Add o->x alternative. Enable for TARGET_SSE.
16474 (sse_storeq): Remove expander.
16475 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
16476 with memory input operand.
16477 (*vec_extractv2di_1 splitter): New.
16478 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
16479 * config/i386/i386.md (ssevecmodelower): New mode attribute.
16481 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
16483 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
16484 (INT_LOWPART): Delete.
16485 (extract_MB): Adjust.
16486 (extract_ME): Adjust.
16487 (print_operand): Adjust.
16489 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
16491 * config/rs6000/predicates.md (reg_or_add_cint_operand,
16492 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
16493 (reg_or_logical_cint_operand, easy_fp_constant,
16494 logical_const_operand): Delete "CONST_DOUBLE" case.
16495 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
16496 "HOST_BITS_PER_WIDE_INT == 64" test.
16497 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
16498 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
16499 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
16500 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
16502 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
16503 CONST_DOUBLE DImode/VOIDmode case.
16504 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
16505 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
16506 CONST_DOUBLE VOIDmode case.
16507 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
16508 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
16509 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
16510 Delete CONST_DOUBLE case.
16511 (splitters for mov FMOVE64 const_double): Delete
16512 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
16513 "HOST_BITS_PER_WIDE_INT >= 64" test.
16514 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
16516 (mov DI const_double): Delete.
16518 2013-05-04 Jakub Jelinek <jakub@redhat.com>
16520 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
16521 on op shows all bits zero in mode of a lowpart subreg, return zero.
16523 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
16526 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
16527 to save TFmode registers and DImode to save TImode registers for
16528 caller save operations.
16529 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
16530 mark being partially clobbered since they only use the first
16533 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
16534 and TDmode only use the upper 64-bits of each VSX register.
16536 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16538 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
16539 (find_candidates_in_block): Disable slsr_process_phi.
16541 2013-05-03 Guozhi Wei <carrot@google.com>
16543 * coverage.c (coverage_obj_init): Move the construction of gcov
16545 (build_init_ctor): ... here.
16547 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16549 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
16550 (slsr_cand_d): Redefine def_phi.
16551 (stride_status, phi_adjust_status, count_phis_status): New enums.
16552 (find_phi_def): New.
16553 (find_basis_for_base_expr): New.
16554 (find_basis_for_candidate): Handle hidden bases.
16555 (alloc_cand_and_find_basis): Handle phi candidates.
16556 (slsr_process_phi): New.
16557 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
16558 (create_mul_imm_cand): Likewise.
16559 (create_add_ssa_cand): Exclude phi base candidates.
16560 (create_add_imm_cand): Likewise.
16561 (slsr_process_cast): Likewise.
16562 (slsr_process_copy): Likewise.
16563 (find_candidates_in_block): Handle phi candidates.
16564 (dump_candidate): Likewise.
16565 (unconditional_cands): Delete.
16566 (unconditional_cands_with_known_stride_p): Delete.
16567 (phi_dependent_cand_p): New.
16568 (cand_increment): Handle phi-dependent candidates.
16569 (replace_dependent): Delete.
16570 (replace_mult_candidate): New.
16571 (replace_unconditional_candidate): New.
16572 (incr_vec_index): Move to avoid forward reference.
16573 (create_add_on_incoming_edge): New.
16574 (create_phi_basis): New.
16575 (replace_dependents): Delete.
16576 (replace_conditional_candidate): New.
16577 (phi_add_costs): New.
16578 (replace_uncond_cands_and_profitable_phis): New.
16579 (record_increment): Handle phi adjustments.
16580 (record_phi_increments): New.
16581 (record_increments): Handle phi adjustments.
16582 (phi_incr_cost): New.
16583 (lowest_cost_path): Handle phis.
16584 (total_savings): Likewise.
16585 (analyze_increments): Likewise.
16586 (ncd_with_phi): New.
16587 (ncd_of_cand_and_phis): New.
16588 (nearest_common_dominator_for_cands): Handle phi increments.
16589 (all_phi_incrs_profitable): New.
16590 (replace_profitable_candidates): Handle phi-dependent candidates.
16591 (analyze_candidates_and_replace): Likewise.
16593 2013-05-03 Teresa Johnson <tejohnson@google.com>
16596 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
16597 do not exceed REG_BR_PROB_BASE.
16599 2013-05-03 Jeff Law <law@redhat.com>
16601 PR tree-optimization/57144
16602 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
16603 operand of the condition will bit into the new type when eliminating
16604 a cast feeding a condition.
16606 2013-05-03 Jakub Jelinek <jakub@redhat.com>
16608 PR rtl-optimization/57130
16609 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
16610 of COMPARE as in_code to the recursive call if needed.
16612 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
16614 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
16615 (enabled): Handle new members.
16616 * config/i386/sse.md (*vec_concatv2si): Merge from
16617 *vec_concatv2si_sse2 and vec_concatv2si_sse.
16618 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
16620 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
16622 PR tree-optimization/57027
16623 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
16624 for fnms opportunity, check we got the prerequisite kind
16625 of tree / gimple before using accessor functions.
16627 2013-05-03 Richard Biener <rguenther@suse.de>
16629 * double-int.h (lshift): New overload without precision
16630 and arith argument.
16631 (operator *=, operator +=, operator -=): Move ...
16632 * double-int.c (operator *=, operator +=, operator -=): ... here
16633 and implement more efficiently.
16634 (mul_double_with_sign): Remove.
16635 (lshift_double): Adjust to take unsinged shift argument, push
16636 dispatching code to callers.
16637 (mul_double_wide_with_sign): Add early out for callers that
16638 are not interested in high parts or overflow.
16639 (lshift): New function.
16640 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
16641 dispatch code here.
16642 (lrotate, rrotate): Use logical shifts.
16643 * expr.c (get_inner_reference): Use lshift.
16644 * fixed-value.c (do_fixed_divide): Likewise.
16645 * tree-dfa.c (get_ref_base_and_extent): Likewise.
16646 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
16647 (indirect_refs_may_alias_p): Likewise.
16648 (stmt_kills_ref_p_1): Likewise.
16650 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
16652 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
16654 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
16656 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
16657 scalar form of FABD instruction.
16659 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
16661 * lra-constraints.c (process_alt_operands): Add checking alt
16662 number to choose the best alternative.
16664 2013-05-02 Richard Biener <rguenther@suse.de>
16666 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
16667 bitmap and its handling.
16668 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
16670 2013-05-02 Richard Biener <rguenther@suse.de>
16672 PR middle-end/57140
16673 * tree-inline.c (copy_loops): Properly handle removed loops.
16674 (copy_cfg_body): Mark destination loops for fixup if source
16675 loops needed fixup.
16677 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
16680 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
16681 generating simple_return for naked functions.
16683 2013-05-02 Martin Jambor <mjambor@suse.cz>
16685 PR middle-end/56988
16686 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
16687 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
16689 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
16690 ipa_agg_replacement_value structures.
16691 (known_aggs_to_agg_replacement_list): Likewise.
16692 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
16693 (read_agg_replacement_chain): Likewise.
16694 (ipcp_transform_function): Also check that by_ref flags match.
16696 2013-05-02 Richard Biener <rguenther@suse.de>
16698 * graphds.h (struct graph): Add obstack member.
16699 * graphds.c (new_graph): Initialize obstack and allocate
16701 (add_edge): Allocate edge from the obstack.
16702 (free_graph): Free the obstack instead of all edges and vertices.
16704 2013-05-02 Teresa Johnson <tejohnson@google.com>
16706 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
16708 * cfg.c (update_bb_profile_for_threading): Ditto.
16709 * tree-inline.c (copy_bb): Ditto.
16710 (copy_edges_for_bb): Ditto.
16711 (initialize_cfun): Ditto.
16712 (copy_cfg_body): Ditto.
16713 (expand_call_inline): Ditto.
16714 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
16715 (estimate_node_size_and_time): Ditto.
16716 (inline_merge_summary): Ditto.
16717 * cgraphclones.c (cgraph_clone_edge): Ditto.
16718 (cgraph_clone_node): Ditto.
16719 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16720 (compute_trg_info): Ditto.
16722 2013-05-02 Ian Bolton <ian.bolton@arm.com>
16724 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
16725 S reg when fp attribute set.
16726 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
16728 2013-05-02 Ian Bolton <ian.bolton@arm.com>
16730 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
16732 (*and_one_cmplsi3_compare0_uxtw): Likewise.
16733 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
16734 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
16736 2013-05-02 Richard Biener <rguenther@suse.de>
16738 * tree-scalar-evolution.c (scev_info_hasher): Remove.
16739 (struct instantiate_cache_entry): New type.
16740 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
16741 (struct instantiate_cache_type): New type.
16742 (set_instantiated_value, get_instantiated_value): Remove.
16743 (get_instantiated_value_entry): New function.
16744 (instantiate_scev_name): Use the new cache and adjust.
16745 (instantiate_scev_poly): Adjust.
16746 (instantiate_scev_binary): Likewise.
16747 (instantiate_array_ref): Likewise.
16748 (instantiate_scev_convert): Likewise.
16749 (instantiate_scev_not): Likewise.
16750 (instantiate_scev_3): Likewise.
16751 (instantiate_scev_2): Likewise.
16752 (instantiate_scev_r): Likewise.
16753 (instantiate_scev): Likewise.
16754 (resolve_mixers): Likewise.
16756 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
16759 * lra-constraints.c (best_small_class_operands_num): Remove.
16760 (process_alt_operands): Remove small_class_operands_num. Take
16761 small classes operands into losers and only if the operand is not
16762 matched. Modify debugging output.
16763 (curr_insn_transform): Remove best_small_class_operands_num.
16766 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16768 * config/aarch64/aarch64-builtins.c
16769 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
16770 * config/aarch64/aarch64-simd-builtins.def
16771 (reduc_splus_): Add new modes.
16772 (reduc_uplus_): New.
16773 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
16774 (reduc_uplus_v4sf): Likewise.
16775 (reduc_splus_v4sf): Likewise.
16776 (aarch64_addv<mode>): Likewise.
16777 (reduc_uplus_<mode>): Likewise.
16778 (reduc_splus_<mode>): Likewise.
16779 (aarch64_addvv2di): Likewise.
16780 (reduc_uplus_v2di): Likewise.
16781 (reduc_splus_v2di): Likewise.
16782 (aarch64_addvv2si): Likewise.
16783 (reduc_uplus_v2si): Likewise.
16784 (reduc_splus_v2si): Likewise.
16785 (reduc_<sur>plus_<mode>): New.
16786 (reduc_<sur>plus_v2di): Likewise.
16787 (reduc_<sur>plus_v2si): Likewise.
16788 (reduc_<sur>plus_v4sf): Likewise.
16789 (aarch64_addpv4sf): Likewise.
16790 * config/aarch64/arm_neon.h
16791 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
16792 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
16793 add UNSPEC_SADDV, UNSPEC_UADDV.
16795 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
16797 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16799 * config/aarch64/arm_neon.h
16800 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
16802 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16804 * config/aarch64/aarch64-builtins
16805 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
16807 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16809 * config/aarch64/aarch64-simd-builtins.def
16810 (reduc_smax_): New.
16811 (reduc_smin_): Likewise.
16812 (reduc_umax_): Likewise.
16813 (reduc_umin_): Likewise.
16814 (reduc_smax_nan_): Likewise.
16815 (reduc_smin_nan_): Likewise.
16818 (smax): Update for V2SF, V4SF and V2DF modes.
16821 (smin_nan): Likewise.
16822 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
16823 (<su><maxmin><mode>3): ...This, refactor.
16824 (s<maxmin><mode>3): New.
16825 (<maxmin_uns><mode>3): Likewise.
16826 (reduc_<maxmin_uns>_<mode>): Refactor.
16827 (reduc_<maxmin_uns>_v4sf): Likewise.
16828 (reduc_<maxmin_uns>_v2si): Likewise.
16829 (aarch64_<fmaxmin><mode>: Remove.
16830 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
16832 (vmin<q>_f<32,64>): Likewise.
16833 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16835 (su): Add mappings for smax, smin, umax, umin.
16837 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16838 (FMAXMIN): Rename as...
16839 (FMAXMIN_UNS): ...This.
16841 (fmaxminv): Likewise.
16842 (fmaxmin): Likewise.
16844 (maxmin_uns_op): Likewise.
16846 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16848 * config/aarch64/arm_neon.h
16849 (vac<ge, gt><sd>_f<32, 64>): Rename to...
16850 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
16851 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
16853 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16855 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
16856 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
16858 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16860 * config/aarch64/aarch64-simd.md
16861 (vcond<mode>_internal): Handle special cases for constant masks.
16862 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
16863 (vcondu<mode><mode>): Likewise.
16864 (vcond<v_cmp_result><mode>): New.
16866 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16868 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
16869 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
16870 * config/aarch64/aarch64-simd-builtins.def
16871 (cmeq): Update to BUILTIN_VALLDI.
16876 * config/aarch64/arm_neon.h
16877 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
16878 to builtins or C as appropriate.
16880 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16882 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
16884 (cmhi): Rename to...
16886 * config/aarch64/aarch64-simd.md
16887 (simd_mode): Add SF.
16888 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
16889 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
16890 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
16891 (cstore<mode>_neg): ...This.
16892 * config/aarch64/iterators.md
16894 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
16895 (COMPARISONS): New.
16896 (UCOMPARISONS): Likewise.
16897 (optab): Add missing comparisons.
16903 (VCMP_S): Likewise.
16904 (VCMP_U): Likewise.
16905 (V_cmp_result): Add DF, SF modes.
16906 (v_cmp_result): Likewise.
16908 (vmtype): Likewise.
16909 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
16911 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
16913 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
16914 define_insn to define_insn_and_split.
16915 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
16916 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
16917 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
16918 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
16919 (thumb2_negscc): Likewise.
16921 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16923 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
16925 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16927 * config/arm/thumb2.md: Remove trailing whitespaces.
16929 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16931 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
16932 Use gen_int_mode rather than GEN_INT.
16934 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
16936 * value-prof.c (stream_in_histogram_value): Remove the strayed
16939 2013-04-30 Richard Biener <rguenther@suse.de>
16941 PR middle-end/57122
16942 * cfghooks.c (split_edge): Properly check for the loop latch edge.
16944 2013-04-30 Richard Biener <rguenther@suse.de>
16946 PR middle-end/57107
16947 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
16949 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
16951 PR rtl-optimization/56957
16952 PR rtl-optimization/57105
16953 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
16954 variable. Use just INSN_UID for determining whether an insn
16955 should be only disconnected from the insn stream.
16956 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
16958 2013-04-30 Jakub Jelinek <jakub@redhat.com>
16960 PR tree-optimization/57104
16961 * tsan.c (instrument_expr): Don't instrument accesses to
16962 DECL_HARD_REGISTER VAR_DECLs.
16964 2013-04-30 Richard Biener <rguenther@suse.de>
16966 * function.h (loops_for_fn): New inline function.
16967 (set_loops_for_fn): Likewise.
16968 * cfgloop.h (place_new_loop): Add struct function parameter.
16969 (get_loop): Likewise.
16970 (get_loops): Likewise.
16971 (number_of_loops): Likewise.
16972 (fel_next): Adjust.
16973 (fel_init): Likewise.
16974 * cfg.c (get_loop_copy): Adjust.
16975 * cfgloop.c (flow_loops_dump): Likewise.
16976 (record_loop_exits): Likewise.
16977 (verify_loop_structure): Likewise.
16978 * cfgloopanal.c (mark_irreducible_loops): Likewise.
16979 (estimate_reg_pressure_cost): Likewise.
16980 (mark_loop_exit_edges): Likewise.
16981 * cfgloopmanip.c (place_new_loop): Likewise.
16982 (add_loop): Likewise.
16983 (duplicate_loop): Likewise.
16984 * graph.c (draw_cfg_nodes): Likewise.
16985 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
16986 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
16987 (extract_affine_chrec): Likewise.
16988 (build_scop_iteration_domain): Likewise.
16989 * graphite.c (graphite_initialize): Likewise.
16990 * ira-build.c (create_loop_tree_nodes): Likewise.
16991 (more_one_region_p): Likewise.
16992 (rebuild_regno_allocno_maps): Likewise.
16993 (mark_loops_for_removal): Likewise.
16994 (mark_all_loops_for_removal): Likewise.
16995 (remove_unnecessary_regions): Likewise.
16996 (ira_build): Likewise.
16997 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16998 * loop-init.c (fix_loop_structure): Likewise.
16999 (gate_rtl_move_loop_invariants): Likewise.
17000 (gate_rtl_unswitch): Likewise.
17001 (gate_rtl_unroll_and_peel_loops): Likewise.
17002 (rtl_doloop): Likewise.
17003 * lto-streamer-in.c (input_cfg): Likewise.
17004 * lto-streamer-out.c (output_cfg): Likewise.
17005 * modulo-sched.c (sms_schedule): Likewise.
17006 * predict.c (tree_estimate_probability): Likewise.
17007 (tree_estimate_probability_driver): Likewise.
17008 (estimate_loops): Likewise.
17009 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
17010 (move_sese_region_to_fn): Likewise.
17011 (debug_loop_num): Likewise.
17012 * tree-chrec.c (chrec_evaluate): Likewise.
17013 (hide_evolution_in_other_loops_than_loop): Likewise.
17014 (chrec_component_in_loop_num): Likewise.
17015 (reset_evolution_in_loop): Likewise.
17016 (evolution_function_is_invariant_rec_p): Likewise.
17017 * tree-if-conv.c (main_tree_if_conversion): Likewise.
17018 * tree-inline.c (copy_loops): Likewise.
17019 (copy_cfg_body): Likewise.
17020 (tree_function_versioning): Likewise.
17021 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
17022 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
17024 (add_to_evolution_1): Likewise.
17025 (scev_const_prop): Likewise.
17026 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
17027 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
17028 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
17029 (tree_ssa_lim_initialize): Likewise.
17030 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
17031 (verify_loop_closed_ssa): Likewise.
17032 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
17033 (tree_ssa_loop_im): Likewise.
17034 (tree_ssa_loop_unswitch): Likewise.
17035 (tree_vectorize): Likewise.
17036 (check_data_deps): Likewise.
17037 (tree_ssa_loop_ivcanon): Likewise.
17038 (tree_ssa_loop_bounds): Likewise.
17039 (tree_complete_unroll): Likewise.
17040 (tree_complete_unroll_inner): Likewise.
17041 (tree_parallelize_loops): Likewise.
17042 (tree_ssa_loop_prefetch): Likewise.
17043 (tree_ssa_loop_ivopts): Likewise.
17044 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
17045 * tree-vectorizer.c (vectorize_loops): Likewise.
17047 2013-04-29 Mike Frysinger <vapier@gentoo.org>
17049 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
17050 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
17051 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
17052 with EABI_LINK_SPEC.
17054 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17057 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
17060 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
17063 * lra-constraints.c (process_alt_operands): Discourage a bit more
17064 using memory for pseudos. Print cost dump for alternatives.
17065 Modify cost values for conflicts with early clobbers.
17066 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
17068 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17071 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
17073 2013-04-29 Ian Bolton <ian.bolton@arm.com>
17075 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
17076 from/to S register.
17077 (movdi_aarch64): Support LDR/STR from/to D register.
17079 2013-04-29 Ian Bolton <ian.bolton@arm.com>
17081 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
17082 or higher by default.
17084 2013-04-29 Richard Biener <rguenther@suse.de>
17086 PR middle-end/57075
17087 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
17088 even if not adding abnormal edges for calls that can make
17091 2013-04-29 Richard Biener <rguenther@suse.de>
17093 PR middle-end/57103
17094 * tree-cfg.c (move_stmt_op): Fix condition under which to update
17096 (move_stmt_r): Remove redundant checking.
17098 2013-04-29 Teresa Johnson <tejohnson@google.com>
17101 * basic-block.h (apply_scale): New function.
17102 (apply_probability): Use apply_scale.
17103 * gimple-streamer-in.c (input_bb): Ditto.
17104 * lto-streamer-in.c (input_cfg): Ditto.
17105 * lto-cgraph.c (merge_profile_summaries): Ditto.
17106 * tree-optimize.c (execute_fixup_cfg): Ditto.
17107 * tree-inline.c (copy_bb): Update comment to use apply_scale.
17108 (copy_edges_for_bb): Ditto.
17109 (copy_cfg_body): Ditto.
17111 2013-04-29 Tom de Vries <tom@codesourcery.com>
17113 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
17114 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
17115 (tail_merge_optimize): Handle current_loops == NULL.
17117 2013-04-26 Jeff Law <law@redhat.com>
17119 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
17120 (simplify_cond_using_ranges): Generalize code to simplify
17121 COND_EXPRs where one argument is a constant and the other
17122 is an SSA_NAME created by an integral type conversion.
17124 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17126 * config/arm/arm.md (store_minmaxsi): Use only when
17127 optimize_insn_for_size_p.
17129 2013-04-29 Christian Bruel <christian.bruel@st.com>
17132 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
17134 2013-04-29 Richard Biener <rguenther@suse.de>
17136 PR middle-end/57089
17137 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
17138 loop tree make sure to schedule a fixup for the child as well.
17139 (expand_omp_for_generic): Properly add loops.
17140 (expand_omp_for_static_nochunk): Likewise.
17141 (expand_omp_for_static_chunk): Likewise.
17142 (expand_omp_for): For the degenerate case fixup loops.
17143 (expand_omp_sections): Fix default bb placement in loops.
17144 (expand_omp_atomic_pipeline): Properly add loops.
17146 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17148 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
17150 2013-04-29 Tom de Vries <tom@codesourcery.com>
17152 * tree-ssa-tail-merge.c: Update header comment.
17154 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17156 * config/aarch64/arm_neon.h
17157 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
17158 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
17159 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
17160 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
17161 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
17162 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
17163 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
17164 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
17166 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17168 * config/aarch64/aarch64-simd.md
17169 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
17170 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
17171 fix_trunc, fixuns_trunc.
17172 (ftrunc<VDQF:mode>2): New.
17173 * config/aarch64/iterators.md (optab): Add fix, fixuns.
17174 (fix_trunc_optab): New.
17176 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17178 * config/aarch64/aarch64-builtins.c
17179 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
17180 iceilf, lround, iroundf.
17182 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17185 * config/i386/i386.h (enum ix86_tune_indices)
17186 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
17187 New, split from X86_TUNE_INTER_UNIT_MOVES.
17188 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
17189 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
17190 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
17191 (TARGET_INTER_UNIT_MOVES): Remove.
17192 * config/i386/i386.c (initial_ix86_tune_features): Update.
17193 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
17194 (ix86_expand_convert_uns_didf_sse): Use
17195 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
17196 (ix86_expand_vector_init_one_nonzero): Ditto.
17197 (ix86_expand_vector_init_interleave): Ditto.
17198 (inline_secondary_memory_needed): Return true for moves from SSE class
17199 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
17200 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
17201 * config/i386/constraints.md (Yi, Ym): Depend on
17202 TARGET_INTER_UNIT_MOVES_TO_VEC.
17203 (Yj, Yn): New constraints.
17204 * config/i386/i386.md (*movdi_internal): Change constraints of
17205 operand 1 from Yi to Yj and from Ym to Yn.
17206 (*movsi_internal): Ditto.
17207 (*movdf_internal): Ditto.
17208 (*movsf_internal): Ditto.
17209 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
17210 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
17211 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
17212 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
17213 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
17214 * config/i386/sse.md (movdi_to_sse): Ditto.
17215 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
17216 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
17217 TARGET_INTER_UNIT_MOVES.
17218 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
17219 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
17220 instead of TARGET_INTER_UNIT_MOVES.
17221 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
17222 operand 1 from Yi to Yj and from Ym to Yn.
17224 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17226 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
17227 (float_truncate_hi_): Likewise.
17228 (float_extend_lo_): Likewise.
17229 (float_truncate_lo_): Likewise.
17230 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
17231 (aarch64_float_extend_lo_v2df): Likewise.
17232 (vec_unpacks_hi_v4sf): Likewise.
17233 (aarch64_float_truncate_lo_v2sf): Likewise.
17234 (aarch64_float_truncate_hi_v4sf): Likewise.
17235 (vec_pack_trunc_v2df): Likewise.
17236 (vec_pack_trunc_df): Likewise.
17238 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17240 * config/aarch64/aarch64-builtins.c
17241 (aarch64_fold_builtin): Fold float conversions.
17242 * config/aarch64/aarch64-simd-builtins.def
17243 (floatv2si, floatv4si, floatv2di): New.
17244 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
17245 * config/aarch64/aarch64-simd.md
17246 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
17247 * config/aarch64/iterators.md (FLOATUORS): New.
17248 (optab): Add float, floatuns.
17249 (su_optab): Likewise.
17251 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17253 * config/aarch64/aarch64-builtins.c
17254 (aarch64_builtin_vectorized_function): Use new names for
17256 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
17257 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
17258 (fcvtzu): Split as...
17259 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
17260 (fcvtas): Split as...
17261 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
17262 (fcvtau): Split as...
17263 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
17264 (fcvtps): Split as...
17265 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
17266 (fcvtpu): Split as...
17267 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
17268 (fcvtms): Split as...
17269 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
17270 (fcvtmu): Split as...
17271 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
17272 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
17273 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
17274 (lfrintnusf, lfrintnudf): Likewise.
17275 * config/aarch64/aarch64-simd.md
17276 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
17278 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
17279 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
17280 (fcvt_pattern): Likewise.
17282 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17284 * config/aarch64/aarch64-simd.md
17285 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
17286 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
17288 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17290 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
17291 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
17292 (vrnd<a,m,n,p>_f32): Implement using builtins.
17293 (vrnd<i,x><q>_f<32, 64>): New.
17295 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17297 * config/aarch64/aarch64-builtins.c
17298 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
17299 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
17300 (frintz): Rename to...
17302 (frintp): Rename to...
17304 (frintm): Rename to...
17306 (frinti): Rename to...
17307 (nearbyint): ...this.
17308 (frintx): Rename to...
17310 (frinta): Rename to...
17312 * config/aarch64/aarch64-simd.md
17313 (aarch64_frint<frint_suffix><mode>): Delete.
17314 (<frint_pattern><mode>2): Convert to insn.
17315 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
17316 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
17317 (frint_pattern): Likewise.
17318 (frint_suffix): Likewise.
17320 2013-04-29 Richard Biener <rguenther@suse.de>
17322 PR tree-optimization/57081
17323 * loop-init.c: Include tree-flow.h.
17324 (loop_optimizer_finalize): Free number of iteration estimates.
17325 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
17327 2013-04-29 Jakub Jelinek <jakub@redhat.com>
17329 PR tree-optimization/57083
17330 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
17331 non-singleton shift count range, zero extend low_bound for uns case.
17333 * config/i386/predicates.md (general_vector_operand): New predicate.
17334 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
17335 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
17336 if they aren't nonimmediate operands. If their original values
17337 satisfy const_vector_equal_evenodd_p, don't shift them.
17338 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
17339 predicates. For the SSE4.1 case force operands[{1,2}] into registers
17340 if not nonimmediate_operand.
17341 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
17342 instead of register_operand.
17343 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
17345 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
17347 * stor-layout.c (finalize_size_functions): Allocate a structure and
17348 reset cfun before dumping the functions.
17350 2013-04-27 Jakub Jelinek <jakub@redhat.com>
17352 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
17355 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
17356 use xop_pmacsdqh if uns_p.
17357 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
17358 the immediate rotate count.
17360 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
17362 * rtl.h (struct rtx_def): Add comment for field jump.
17363 (LRA_SUBREG_P): New macro.
17364 * recog.c (register_operand): Check LRA_SUBREG_P.
17365 * lra.c (lra): Add note at the end of RTL code. Align non-empty
17367 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
17368 (lra_final_code_change): Skip subreg change for operators.
17369 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
17370 if there are no operand changes.
17371 * lra-constraints.c (curr_insn_set): New.
17372 (match_reload): Set LRA_SUBREG_P.
17373 (emit_spill_move): Ditto.
17374 (check_and_process_move): Use curr_insn_set. Process only single
17375 set insns. Don't initialize sec_mem_p and change_p.
17376 (simplify_operand_subreg): Use LRA_SUBREG_P.
17377 (reg_in_class_p): New function.
17378 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
17379 of #ifdef. Add code to remove cycling.
17380 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
17381 non-null disp. Reload inner instead of disp when base and index
17382 are null. Try to put lo_sum into register.
17383 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
17384 (check_and_process_move): Move code for move cost check to
17385 simple_move_p. Remove equiv_substitution.
17386 (simple_move_p): New function.
17387 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
17388 curr_insn_set. Call check_and_process_move only for single set
17389 insns. Use the new function. Move call of check_and_process_move
17390 after operand equiv substitution and address process.
17392 2013-04-26 Jakub Jelinek <jakub@redhat.com>
17395 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
17396 with nonlocal goto receivers or returns twice calls, ignore
17397 unininitialized values from abnormal edges to nl goto receiver
17398 or returns twice call.
17400 2013-04-26 Jakub Jelinek <jakub@redhat.com>
17402 PR tree-optimization/57051
17403 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
17404 and VEC_RSHIFT_EXPR if shift count is a multiple of element
17407 2013-04-26 Richard Biener <rguenther@suse.de>
17409 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
17410 (expand_omp_taskreg): Likewise. Mark loops for fixup.
17411 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
17412 (fixup_loop_arrays_after_move): New function.
17413 (move_sese_region_to_fn): Properly outline the loop tree parts
17414 of the SESE region.
17416 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
17418 * config/i386/i386.md (type, unit): Fix long lines.
17420 2013-04-26 Richard Biener <rguenther@suse.de>
17422 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
17423 (lto-streamer-out.o): Likewise.
17424 * cfgloop.c (init_loops_structure): Export, add struct function
17425 argument and adjust.
17426 (flow_loops_find): Adjust.
17427 * cfgloop.h (enum loop_estimation): Add EST_LAST.
17428 (init_loops_structure): Declare.
17429 * lto-streamer-in.c: Include cfgloop.h.
17430 (input_cfg): Input the loop tree.
17431 * lto-streamer-out.c: Include cfgloop.h.
17432 (output_cfg): Output the loop tree.
17433 (output_struct_function_base): Do not drop PROP_loops.
17435 2013-03-26 Richard Biener <rguenther@suse.de>
17437 * tree-cfg.c (execute_build_cfg): Build the loop tree.
17438 (pass_build_cfg): Provide PROP_loops.
17439 (move_sese_region_to_fn): Remove loops that are outlined into fn
17441 * tree-inline.c: Include cfgloop.h.
17442 (initialize_cfun): Do not drop PROP_loops.
17443 (copy_loops): New function.
17444 (copy_cfg_body): Copy loop structure.
17445 (tree_function_versioning): Initialize destination loop tree.
17446 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
17447 (pass_parallelize_loops): Do IL verification.
17448 * loop-init.c (loop_optimizer_init): Fixup loops if required.
17449 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
17450 the CFG make sure we fixup loops as well.
17451 * tree-ssa-tail-merge.c: Include cfgloop.h.
17452 (replace_block_by): When merging loop latches mark loops for fixup.
17453 * lto-streamer-out.c (output_struct_function_base): Drop
17454 PROP_loops for now.
17455 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
17456 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
17457 * ipa-split.c: Include cfgloop.h.
17458 (split_function): Add the new return block to the loop tree root.
17459 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
17460 whether we have removed the forwarder block.
17461 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
17462 * cfgloop.h (place_new_loop): Declare.
17463 * cfgloopmanip.c (place_new_loop): Export.
17464 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
17465 (tree-switch-conversion.o): Likewise.
17466 (tree-complex.o): Likewise.
17467 (tree-inline.o): Likewise.
17468 (tree-ssa-tailmerge.o): Likewise.
17469 (ipa-split.o): Likewise.
17470 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
17471 (tree-ssa-copy.o): Likewise.
17472 * tree-switch-conversion.c: Include cfgloop.h
17473 (process_switch): If we emit a bit-test cascade, schedule loops
17475 * tree-complex.c: Include cfgloop.h.
17476 (expand_complex_div_wide): Properly add new basic-blocks to loops.
17477 * asan.c: Include cfgloop.h.
17478 (create_cond_insert_point): Properly add new basic-blocks to
17479 loops, schedule loop fixup.
17480 * cfgloop.c (verify_loop_structure): Check that looks are not
17482 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
17484 (expand_omp_for_generic): Likewise.
17485 (expand_omp_sections): Likewise.
17486 (expand_omp_atomic_pipeline): Schedule loops for fixup.
17487 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
17488 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
17489 is initialized, not when loops are present.
17490 * tree-parloops.c (parallelize_loops): Remove checking here.
17491 * passes.c (init_optimization_passes): Schedule a copy-propagation
17492 pass before complete unrolling of inner loops.
17494 2013-04-26 Jakub Jelinek <jakub@redhat.com>
17496 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
17497 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
17498 (colorize_init): Add argument to _WIN32 version.
17499 * toplev.c: Include diagnostic-color.h.
17500 (process_options): Default to -fdiagnostics-color=auto if
17501 GCC_COLORS env var is in the environment.
17502 * common.opt (fdiagnostics-color=): Add Var and Init.
17503 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
17504 env var is in the environment, the default is auto rather than never.
17506 * diagnostic.h (file_name_as_prefix): Add context argument.
17507 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
17508 the string as locus.
17509 * langhooks.c (lhd_print_error_function): Adjust caller.
17511 2013-04-25 Lawrence Crowl <crowl@google.com>
17513 * var-tracking.c (shared_hash_def::htab):
17514 Change type to hash_table. Update dependent calls and types.
17516 2013-04-25 Lawrence Crowl <crowl@google.com>
17518 * Makefile.in: Update as needed below.
17520 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
17521 Move declaration to after the type's method definitons.
17523 * attribs.c (htab_t scoped_attributes::attribute_hash):
17524 Change type to hash_table. Update dependent calls and types.
17526 * bitmap.c (htab_t bitmap_desc_hash):
17527 Change type to hash_table. Update dependent calls and types.
17529 * cselib.c (htab_t cselib_hash_table):
17530 Change type to hash_table. Update dependent calls and types.
17532 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
17533 (hash_string_slot_node): Move implementation into lto-streamer.h
17534 struct string_slot_hasher.
17535 (eq_string_slot_node): Likewise.
17537 * data-streamer-out.c: Update output_block::string_hash_table
17538 dependent calls and types.
17540 * dwarf2cfi.c (htab_t trace_index):
17541 Change type to hash_table. Update dependent calls and types.
17543 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
17544 Change type to hash_table. Update dependent calls and types.
17545 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
17546 (htab_t optimize_external_refs::map): Likewise.
17547 (htab_t output_comp_unit::extern_map): Likewise.
17548 (htab_t output_comdat_type_unit::extern_map): Likewise.
17549 (htab_t output_macinfo::macinfo_htab): Likewise.
17550 (htab_t optimize_location_lists::htab): Likewise.
17551 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
17553 * except.c (htab_t ehspec_hash_type):
17554 Change type to hash_table. Update dependent calls and types.
17555 (assign_filter_values::ttypes): Likewise.
17556 (assign_filter_values::ehspec): Likewise.
17557 (sjlj_assign_call_site_values::ar_hash): Likewise.
17558 (convert_to_eh_region_ranges::ar_hash): Likewise.
17560 * gcse.c (htab_t pre_ldst_table):
17561 Change type to hash_table. Update dependent calls and types.
17563 * ggc-common.c (htab_t saving_htab):
17564 Change type to hash_table. Update dependent calls and types.
17565 (htab_t loc_hash): Likewise.
17566 (htab_t ptr_hash): Likewise.
17567 (call_count): Rename ggc_call_count.
17568 (call_alloc): Rename ggc_call_alloc.
17569 (loc_descriptor): Rename make_loc_descriptor.
17570 (add_statistics): Rename ggc_add_statistics.
17572 * ggc-common.c (saving_htab):
17573 Change type to hash_table. Update dependent calls and types.
17575 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
17576 (push_gimplify_context): Likewise.
17577 (pop_gimplify_context): Likewise.
17578 (struct gimple_temp_hash_elt): Added.
17579 (struct gimplify_hasher): Likewise.
17580 (struct gimplify_ctx.temp_htab):
17581 Change type to hash_table. Update dependent calls and types.
17583 * gimple-fold.c: Include gimplify-ctx.h.
17585 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
17586 Change type to hash_table. Update dependent calls and types.
17587 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
17588 avoid potential global name collision.
17590 * gimplify.c: Include gimplify-ctx.h.
17591 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
17592 (htab_t gimplify_ctx::temp_htab):
17593 Update dependent calls and types for new type hash_table.
17594 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
17595 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
17597 * gimplify-ctx.h: New.
17598 (struct gimple_temp_hash_elt): Move from gimplify.c.
17599 (class gimplify_hasher): New.
17600 (struct gimplify_ctx): Move from gimple.h.
17601 (htab_t gimplify_ctx::temp_htab):
17602 Change type to hash_table. Update dependent calls and types.
17604 * graphite-clast-to-gimple.c: Include graphite-htab.h.
17605 (htab_t ivs_params::newivs_index):
17606 Change type to hash_table. Update dependent calls and types.
17607 (htab_t ivs_params::params_index): Likewise.
17608 (htab_t print_generated_program::params_index): Likewise.
17609 (htab_t gloog::newivs_index): Likewise.
17610 (htab_t gloog::params_index): Likewise.
17612 * graphite.c: Include graphite-htab.h.
17613 4htab_t graphite_transform_loops::bb_pbb_mapping):
17614 Change type to hash_table. Update dependent calls and types.
17616 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
17617 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
17618 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
17620 * graphite-dependences.c: Include graphite-htab.h.
17621 (loop_is_parallel_p): Change hash table type of parameter.
17623 * graphite-htab.h: New.
17624 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
17625 (extern find_pbb_via_hash): Move from graphite-poly.h.
17626 (extern loop_is_parallel_p): Move from graphite-poly.h.
17627 (extern get_loop_body_pbbs): Move from graphite-poly.h.
17629 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
17630 (extern loop_is_parallel_p): Move to graphite-htab.h.
17631 (extern get_loop_body_pbbs): Move to graphite-htab.h.
17633 * haifa-sched.c (htab_t delay_htab):
17634 Change type to hash_table. Update dependent calls and types.
17635 (htab_t delay_htab_i2): Likewise.
17637 * ira-color.c (htab_t allocno_hard_regs_htab):
17638 Change type to hash_table. Update dependent calls and types.
17640 * ira-costs.c (htab_t cost_classes_htab):
17641 Change type to hash_table. Update dependent calls and types.
17643 * loop-invariant.c (htab_t merge_identical_invariants::eq):
17644 Change type to hash_table. Update dependent calls and types.
17646 * loop-iv.c (htab_t bivs):
17647 Change type to hash_table. Update dependent calls and types.
17649 * loop-unroll.c (htab_t opt_info::insns_to_split):
17650 Change type to hash_table. Update dependent calls and types.
17651 (htab_t opt_info::insns_with_var_to_expand): Likewise.
17653 * lto-streamer.h (struct string_slot): Move from data-streamer.h
17654 (struct string_slot_hasher): New.
17655 (htab_t output_block::string_hash_table):
17656 Change type to hash_table. Update dependent calls and types.
17658 * lto-streamer-in.c (freeing_string_slot_hasher): New.
17659 (htab_t file_name_hash_table):
17660 Change type to hash_table. Update dependent calls and types.
17662 * lto-streamer-out.c: Update output_block::string_hash_table dependent
17665 * lto-streamer.c (htab_t tree_htab):
17666 Change type to hash_table. Update dependent calls and types.
17668 * omp-low.c: Include gimplify-ctx.h.
17670 * passes.c (htab_t name_to_pass_map):
17671 Change type to hash_table. Update dependent calls and types.
17672 (pass_traverse): Rename to passes_pass_traverse.
17674 * plugin.c (htab_t event_tab):
17675 Change type to hash_table. Update dependent calls and types.
17677 * postreload-gcse.c (htab_t expr_table):
17678 Change type to hash_table. Update dependent calls and types.
17679 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
17681 * sese.c (debug_rename_map_1): Make extern.
17682 (htab_t copy_bb_and_scalar_dependences::rename_map):
17683 Change type to hash_table. Update dependent calls and types.
17685 * sese.h (extern debug_rename_map): Move to .c file.
17687 * store-motion.c (htab_t store_motion_mems_table):
17688 Change type to hash_table. Update dependent calls and types.
17690 * trans-mem.c (htab_t tm_new_mem_hash):
17691 Change type to hash_table. Update dependent calls and types.
17693 * tree-browser.c (htab_t TB_up_ht):
17694 Change type to hash_table. Update dependent calls and types.
17696 * tree-cfg.c (htab_t discriminator_per_locus):
17697 Change type to hash_table. Update dependent calls and types.
17699 * tree-complex.c: Include tree-hasher.h
17700 (htab_t complex_variable_components):
17701 Change type to hash_table. Update dependent calls and types.
17703 * tree-eh.c (htab_t finally_tree):
17704 Change type to hash_table. Update dependent calls and types.
17706 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
17707 struct int_tree_hasher.
17708 (extern int_tree_map_eq): Likewise.
17709 (uid_decl_map_hash): Removed.
17710 (extern decl_tree_map_eq): Likewise.
17712 * tree-hasher.h: New.
17713 (struct int_tree_hasher): New.
17714 (typedef int_tree_htab_type): New.
17716 * tree-inline.c: Include gimplify-ctx.h.
17718 * tree-mudflap.c: Include gimplify-ctx.h.
17720 * tree-parloops.c: Include tree-hasher.h.
17721 (htab_t eliminate_local_variables_stmt::decl_address):
17722 Change type to hash_table. Update dependent calls and types.
17723 (htab_t separate_decls_in_region::decl_copies): Likewise.
17725 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
17726 Change type to hash_table. Update dependent calls and types.
17728 * tree-sra.c (candidates):
17729 Change type to hash_table. Update dependent calls and types.
17731 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
17733 (int_tree_map_hash): Likewise.
17735 * tree-ssa-dom.c (htab_t avail_exprs):
17736 Change type to hash_table. Update dependent calls and types.
17738 * tree-ssa-live.c (var_map_base_init::tree_to_index):
17739 Change type to hash_table. Update dependent calls and types.
17741 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
17742 Change type to hash_table. Update dependent calls and types.
17744 * tree-ssa-phiopt.c (seen_ssa_names):
17745 Change type to hash_table. Update dependent calls and types.
17747 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
17748 Change type to hash_table. Update dependent calls and types.
17750 * tree-ssa-uncprop.c (equiv):
17751 Change type to hash_table. Update dependent calls and types.
17753 2013-04-25 Jakub Jelinek <jakub@redhat.com>
17755 PR rtl-optimization/57003
17756 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
17757 call note_stores with kill_clobbered_value callback again after
17758 killing regs_invalidated_by_call.
17760 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17762 * config/aarch64/aarch64-simd.md
17763 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
17764 (aarch64_simd_bsl<mode>): Likewise.
17765 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
17767 2013-04-25 Marek Polacek <polacek@redhat.com>
17769 PR tree-optimization/57066
17770 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
17772 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
17774 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
17776 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17778 * config/aarch64/aarch64-builtins.c
17779 (aarch64_fold_builtin): New.
17780 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
17781 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
17782 * config/aarch64/aarch64-simd-builtins.def (abs): New.
17783 * config/aarch64/arm_neon.h
17784 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
17786 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17787 Tejas Belagod <tejas.belagod@arm.com>
17789 * config/aarch64/aarch64-builtins.c
17790 (aarch64_gimple_fold_builtin): New.
17791 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
17792 * config/aarch64/aarch64-simd-builtins.def (addv): New.
17793 * config/aarch64/aarch64-simd.md (addpv4sf): New.
17794 (addvv4sf): Update.
17795 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
17797 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17799 * config/aarch64/aarch64.md
17800 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
17802 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17804 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
17805 (*ngcsi_uxtw): New pattern.
17807 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17808 Julian Brown <julian@codesourcery.com>
17810 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
17811 (TB_DREG): Add T_V4HF.
17812 (v4hf_UP): New macro.
17813 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17814 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17815 Handle initialisation of V4HF. Adjust initialisation of reinterpret
17817 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17818 (arm_vector_mode_supported_p): Handle V4HF.
17819 (arm_mangle_map): Handle V4HFmode.
17820 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
17821 * config/arm/arm_neon_builtins.def: Add entries for
17822 vcvtv4hfv4sf, vcvtv4sfv4hf.
17823 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
17824 (neon_vcvtv4hfv4sf): Likewise.
17825 * config/arm/neon-gen.ml: Handle half-precision floating point
17827 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
17828 * config/arm/arm_neon.h: Regenerate.
17829 * config/arm/neon.ml (type elts): Add F16.
17830 (type vectype): Add T_float16x4, T_floatHF.
17831 (type vecmode): Add V4HF.
17832 (type features): Add Requires_FP_bit feature.
17833 (elt_width): Handle F16.
17834 (elt_class): Likewise.
17835 (elt_of_class_width): Likewise.
17836 (mode_of_elt): Refactor.
17837 (type_for_elt): Handle F16, fix error messages.
17838 (vectype_size): Handle T_float16x4.
17839 (vcvt_sh): New function.
17840 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
17841 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
17842 (string_of_mode): Handle V4HF.
17843 * doc/arm-neon-intrinsics.texi: Regenerate.
17845 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17847 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
17848 format specifier in 'X' case.
17850 2013-04-25 Alan Modra <amodra@gmail.com>
17853 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
17854 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
17855 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
17856 Repeat for many other rotate/shift and mask patterns using subregs.
17857 Name lshiftrt insns.
17858 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
17859 on WORDS_BIG_ENDIAN.
17861 2013-04-25 Alan Modra <amodra@gmail.com>
17863 * config.gcc: Support little-endian powerpc-linux targets.
17864 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
17865 (LINK_OS_LINUX_SPEC): Define.
17866 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
17867 Preserve MASK_LITTLE_ENDIAN.
17868 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
17869 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
17870 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
17871 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
17872 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
17873 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
17874 Correct fp word order for little-endian. Don't shift toc entries
17875 smaller than a word for little-endian.
17876 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
17877 (bswapdi2 splits): Correct low-part subreg for little-endian.
17878 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
17879 low/high where such is correct only for be.
17880 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
17881 little-endian for -mcall-aixdesc.
17883 2013-04-25 Alan Modra <amodra@gmail.com>
17885 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
17886 replace_equiv_address_nv.
17888 2013-04-25 Alan Modra <amodra@gmail.com>
17890 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
17892 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17895 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17896 * rtl.h (struct rtx_def): ...
17898 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17900 PR rtl-optimizations/57046
17901 * lra-constraints (split_reg): Set up lra_risky_transformations_p
17902 for multi-reg splits.
17904 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
17906 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
17908 2013-04-24 Sterling Augustine <saugustine@google.com>
17910 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
17911 (comp_dir_string, debug_str_dwo_section): New.
17912 (DEBUG_STR_DWO_SECTION): Rename to ...
17913 (DEBUG_DWO_STR_SECTION): ... this.
17914 (DEBUG_NORM_STR_SECTION): Delete.
17915 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
17916 (DEBUG_STR_DWO_SECTION_FLAGS): New.
17917 (find_AT_string): Move most logic to ...
17918 (find_AT_string_in_table): ... here. New.
17919 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
17920 add_skeleton_AT_string. Delete logic.
17921 (output_skeleton_debug_sections): Remove call to
17922 add_top_level_skeleton_die_attrs.
17923 (add_comp_dir_attribute): Move logic to comp_dir_string.
17924 (dwarf2out_init): Initialize debug_str_dwo_section.
17925 (output_indirect_string): Call find_string_form.
17926 (output_indirect_strings): Rewrite.
17927 (prune_unused_types): Empty skeleton_debug_str_hash.
17928 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
17929 (dwarf2out_finish): Call output_indirect_strings.
17931 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17933 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
17935 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17937 * rtl.h (struct rtx_def): Add comment for field jump.
17938 (LRA_SUBREG_P): New macro.
17939 * recog.c (register_operand): Check LRA_SUBREG_P.
17940 * lra.c (lra): Add note at the end of RTL code. Align non-empty
17942 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
17943 (lra_final_code_change): Skip subreg change for operators.
17944 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
17945 if there are no operand changes.
17946 * lra-constraints.c (curr_insn_set): New.
17947 (match_reload): Set LRA_SUBREG_P.
17948 (emit_spill_move): Ditto.
17949 (check_and_process_move): Use curr_insn_set. Process only single
17950 set insns. Don't initialize sec_mem_p and change_p.
17951 (simplify_operand_subreg): Use LRA_SUBREG_P.
17952 (reg_in_class_p): New function.
17953 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
17954 of #ifdef. Add code to remove cycling.
17955 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
17956 non-null disp. Reload inner instead of disp when base and index
17957 are null. Try to put lo_sum into register.
17958 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
17959 (check_and_process_move): Move code for move cost check to
17960 simple_move_p. Remove equiv_substitution.
17961 (simple_move_p): New function.
17962 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
17963 curr_insn_set. Call check_and_process_move only for single set
17964 insns. Use the new function. Move call of check_and_process_move
17965 after operand equiv substitution and address process.
17967 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
17969 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
17970 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
17971 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
17973 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17975 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
17977 2013-04-24 Marek Polacek <polacek@redhat.com>
17979 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
17980 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
17981 (select_loops_exit_conditions): Likewise.
17982 (number_of_iterations_for_all_loops): Likewise.
17983 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
17984 (scev_analysis): Likewise.
17986 2013-04-02 Catherine Moore <clm@codesourcery.com>
17987 Chao-ying Fu <fu@mips.com>
17989 * config/mips/micromips.md (jraddiusp): New pattern.
17990 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
17991 instruction if possible.
17993 2013-04-24 Alan Modra <amodra@gmail.com>
17995 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
17997 2013-04-24 Julian Brown <julian@codesourcery.com>
17998 Chung-Lin Tang <cltang@codesourcery.com>
18000 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
18001 dependency behavior in enumeration type DIE generation. Add TODO note
18002 to comments about future DW_FORM_sdata/udata re-work of related code.
18004 2013-04-23 Lawrence Crowl <crowl@google.com>
18006 * Makefile.in: Update as needed below.
18008 * hash-table.h (class hash_table):
18009 Correct many methods with parameter types compare_type to the correct
18010 value_type. (Correct code was unlikely to notice the change.)
18011 (hash_table::elements_with_deleted) New.
18012 (class hashtable::iterator): New.
18013 (hashtable::begin()): New.
18014 (hashtable::end()): New.
18015 (FOR_EACH_HASH_TABLE_ELEMENT): New.
18017 * statistics.c (statistics_hashes):
18018 Change type to hash_table. Update dependent calls and types.
18020 * tree-into-ssa.c (var_infos):
18021 Change type to hash_table. Update dependent calls and types.
18023 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
18024 Change type to hash_table. Update dependent calls and types.
18026 * tree-ssa-loop-im.c (struct mem_ref.refs):
18027 Change type to hash_table. Update dependent calls and types.
18029 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
18030 Change type to hash_table. Update dependent calls and types.
18032 * tree-ssa-sccvn.c (vn_tables_s::nary):
18033 Change type to hash_table. Update dependent calls and types.
18034 (vn_tables_s::phis): Likewise.
18035 (vn_tables_s::references): Likewise.
18037 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
18038 (vn_reference_eq): Update parameter and return types.
18040 * tree-ssa-structalias.c (pointer_equiv_class_table):
18041 Change type to hash_table. Update dependent calls and types.
18042 (location_equiv_class_table): Likewise.
18044 * tree-vect-data-refs.c: Consequential changes for making
18045 peeling a hash_table.
18047 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
18048 (destroy_loop_vec_info): Dependent hash_table update.
18050 * tree-vectorizer.h (peeling_htab):
18051 Change type to hash_table. Update dependent calls and types.
18053 2013-04-23 Shiva Chen <shiva0217@gmail.com>
18055 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
18056 to check the register content is equal or not.
18057 * lra-constraints.c (match_reload): Use lra_assign_reg_val
18058 to assign register content record.
18059 * lra-eliminations.c (update_reg_eliminate): Use
18060 lra_update_reg_val_offset to update register content offset.
18061 * lra-int.h (struct lra_reg): Add offset member.
18062 (lra_reg_val_equal_p): New static inline function.
18063 (lra_update_reg_val_offset): New static inline function.
18064 (lra_assign_reg_val): New static inline function.
18065 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
18066 to assign register content record.
18067 (initialize_lra_reg_info_element): Initial offset to zero.
18069 2013-04-23 Catherine Moore <clm@codesourcery.com>
18071 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
18072 operands. Record compression.
18074 2013-04-23 Xinliang David Li <davidxl@google.com>
18076 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
18078 2013-04-23 Richard Biener <rguenther@suse.de>
18080 PR middle-end/57036
18081 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
18082 parameter, only add abnormal goto edges from the copied body
18083 if the call could perform abnormal gotos.
18084 (copy_cfg_body): Adjust.
18086 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
18088 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
18090 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
18092 * coretypes.h (gimple_stmt_iterator): Add struct to make
18095 2013-04-23 Richard Biener <rguenther@suse.de>
18097 PR tree-optimization/57026
18098 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
18099 from SSA names occuring in abnormal PHI nodes.
18101 2013-04-22 Andi Kleen <ak@linux.intel.com>
18103 * lto/lto.c (print_lto_report_1): Fix LTO report names.
18105 2013-04-22 Andi Kleen <ak@linux.intel.com>
18107 * lto/lto.c (print_lto_report_1): Declare early.
18108 (read_cgraph_and_symbols): Call print_lto_report_1 early.
18110 2013-04-22 Andi Kleen <ak@linux.intel.com>
18112 * common.opt (-flto-report-wpa): Add.
18113 * doc/invoke.texi (-flto-report-wpa): Add.
18114 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
18117 2013-04-22 Xinliang David Li <davidxl@google.com>
18119 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
18120 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
18121 * Makefile.in: New dependency
18123 David Daney <ddaney.cavm@gmail.com>
18125 * configure.ac (gcc_cv_as_micromips_support): Use the
18126 --fatal-warnings option.
18127 * configure: Regenerate.
18129 2013-04-22 Marek Polacek <polacek@redhat.com>
18132 * tsan.c (instrument_expr): Don't instrument expression
18133 in case its size is zero.
18135 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
18139 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18141 * config/alpha/alpha.c (TARGET_LRA_P): New define.
18143 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
18145 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
18146 (gimple_stmt_iterator): New typedef.
18147 * gimple.h (gimple_stmt_iterator): Rename to...
18148 (gimple_stmt_iterator_d): ... This.
18149 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
18150 trees be valid for GIMPLE and GENERIC.
18151 (TARGET_GIMPLE_FOLD_BUILTIN): New.
18152 * gimple-fold.c (gimple_fold_call): Call target hook
18153 gimple_fold_builtin.
18154 * hooks.c (hook_bool_gsiptr_false): New.
18155 * hooks.h (hook_bool_gsiptr_false): New.
18156 * target.def (fold_stmt): New.
18157 * doc/tm.texi: Regenerate.
18159 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
18162 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
18163 a set sp if no stack realignment.
18165 2013-04-22 Nick Clifton <nickc@redhat.com>
18167 * config.gcc (tilegx-linux): Extend extra_objs rather than
18169 (tilepro-linux): Likewise.
18171 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
18173 * config/aarch64/aarch64-builtins.c
18175 (CF0, CF1, CF2, CF3, CF4, CF10): New.
18176 (VAR<1-12>): Add MAP parameter.
18177 (BUILTIN_*): Likewise.
18178 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
18179 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
18180 (aarch64_ushl_n<mode>): Likewise.
18181 (aarch64_sshr_n<mode>): Likewise.
18182 (aarch64_ushr_n<mode>): Likewise.
18183 (aarch64_<maxmin><mode>): Likewise.
18184 (aarch64_sqrt<mode>): Likewise.
18185 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
18186 (vshr<q>_n_*): Likewise.
18188 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
18190 * config/aarch64/aarch64-builtins.c
18191 (aarch64_simd_builtin_type_mode): Handle SF types.
18193 (BUILTIN_GPF): Define.
18194 (aarch64_init_simd_builtins): Handle SF types.
18195 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
18196 (frecps): Likewise.
18197 (frecpx): Likewise.
18198 * config/aarch64/aarch64-simd.md
18199 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
18200 (aarch64_frecpe<mode>): New.
18201 (aarch64_frecps<mode>): Likewise.
18202 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
18203 (v8type): Add frecp<esx>.
18204 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
18205 (aarch64_frecps<mode>): Likewise.
18206 * config/aarch64/iterators.md (FRECP): New.
18207 (frecp_suffix): Likewise.
18208 * config/aarch64/arm_neon.h
18209 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
18211 2013-04-22 Christian Bruel <christian.bruel@st.com>
18214 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
18215 (REG_CLASS_NAMES): Idem.
18216 (REG_CLASS_CONTENTS): Idem.
18217 (REGCLASS_HAS_FP_REG): Idem.
18218 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
18219 (sh_conditional_register_usage): Idem.
18221 2013-04-21 Jeff Law <law@redhat.com>
18223 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
18224 (ssa_forward_propagate_and_combine): Use it.
18226 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
18228 * lra.c: Update the flow chart diagram.
18230 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
18232 PR rtl-optimization/56847
18233 * lra-constraints.c (process_alt_operands): Discourage alternative
18234 with non-matche doffsettable memory constraint fro memory with
18237 2013-04-19 Richard Biener <rguenther@suse.de>
18239 PR tree-optimization/56982
18240 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
18242 * gimplify.c (gimplify_call_expr): Notice special calls.
18243 (gimplify_modify_expr): Likewise.
18244 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
18245 abnormal control flow receivers.
18246 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
18247 in the same way as cfun->has_nonlocal_labels.
18248 (gimple_purge_dead_abnormal_call_edges): Likewise.
18249 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
18250 receivers start a basic-block.
18252 2013-04-19 Richard Biener <rguenther@suse.de>
18254 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
18256 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
18257 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
18258 (SLP_TREE_LOAD_PERMUTATION): Add.
18259 (vect_transform_slp_perm_load): Adjust prototype.
18260 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
18261 (vect_free_slp_instance): Likewise.
18262 (vect_create_new_slp_node): Likewise.
18263 (vect_supported_slp_permutation_p): Remove.
18264 (vect_slp_rearrange_stmts): Adjust.
18265 (vect_supported_load_permutation_p): Likewise. Inline
18266 vect_supported_slp_permutation_p here.
18267 (vect_analyze_slp_instance): Compute load permutations per
18268 slp node instead of per instance.
18269 (vect_get_slp_defs): Adjust.
18270 (vect_transform_slp_perm_load): Likewise.
18271 (vect_schedule_slp_instance): Remove redundant code.
18272 (vect_schedule_slp): Remove hack for PR56270, add it ...
18273 * tree-vect-stmts.c (vectorizable_load): ... here, do not
18274 CSE loads for SLP. Adjust.
18276 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
18278 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
18279 spelling in two comments.
18281 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
18284 * config/arm/arm.c (load_multiple_sequence): Require SP
18285 as base register for loads if SP is in the register list.
18287 2013-04-19 Martin Jambor <mjambor@suse.cz>
18289 PR tree-optimization/56718
18290 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
18291 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
18292 and made public. Adjusted all callers.
18293 (ipa_intraprocedural_devirtualization): New function.
18294 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
18295 (ipa_intraprocedural_devirtualization): Likewise.
18296 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
18298 2013-04-19 Richard Biener <rguenther@suse.de>
18300 PR tree-optimization/57000
18301 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
18303 2013-04-19 Terry Guo <terry.guo@arm.com>
18305 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
18307 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
18308 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
18309 (cortex_m4_fmacs): Use new reservations.
18310 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
18312 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
18314 PR rtl-optimization/56999
18315 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
18317 (lra_coalesce): Remove split_origin_bitmap and related code.
18318 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
18319 ranges if necessary.
18321 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
18323 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
18325 (ix86_expand_call): Remove clobbered_registers array and use
18326 x86_64_ms_sysv_extra_clobbered_registers instead.
18327 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
18329 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
18331 * config/i386/i386.md (*call_rex64_ms_sysv): Use
18332 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
18333 (*call_value_rex64_ms_sysv): Ditto.
18335 2013-04-18 Cary Coutant <ccoutant@google.com>
18337 * dwarf2out.c (output_pubnames): Check die_perennial_p of
18338 parent instead of die_mark.
18340 2013-04-18 Diego Novillo <dnovillo@google.com>
18342 * gimple.c (create_gimple_tmp): New.
18343 (get_expr_type): New.
18344 (build_assign): New.
18345 (build_type_cast): New.
18346 * gimple.h (enum ssa_mode): Define.
18347 (gimple_seq_set_location): New.
18348 * asan.c (build_check_stmt): Change some gimple_build_* calls
18349 to use build_assign and build_type_cast.
18351 2013-04-18 Richard Biener <rguenther@suse.de>
18353 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
18354 handle negative step. Remove redundant checks.
18355 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
18356 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
18357 for negative step and grouped loads fail to vectorize.
18359 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
18361 * emit-rtl.c (reset_insn_used_flags): New function.
18362 (reset_all_used_flags): Use it.
18363 (verify_insn_sharing): New function.
18364 (verify_rtl_sharing): Fix verification for SEQUENCEs.
18366 2013-04-18 Jakub Jelinek <jakub@redhat.com>
18368 PR tree-optimization/56984
18369 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
18370 and (x >> M) >= N don't register any assertion if N << M is the
18373 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
18375 * lower-subreg.c (resolve_simple_move): If called self-recursive,
18376 do not delete_insn insns that have not yet been emitted, only
18377 unlink them with remove_insn.
18378 * df-scan.c (df_insn_delete): Revert r197492.
18380 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
18382 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
18383 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
18385 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
18387 * config/arm/arm.md (movsicc_insn): Convert define_insn into
18388 define_insn_and_split.
18389 (and_scc,ior_scc,negscc): Likewise.
18390 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
18392 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
18394 * config/arm/arm.c (use_return_insn): Return 0 for targets that
18395 can benefit from using a sequence of LDRD instructions in epilogue
18396 instead of a single LDM instruction.
18398 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
18401 * doc/extend.texi: Fix typo.
18403 2013-04-17 Richard Biener <rguenther@suse.de>
18405 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
18406 (vect_build_slp_tree): ... here.
18407 (vect_build_slp_tree_1): Compute which stmts of the SLP group
18408 match. Remove special-casing of mismatched complex loads.
18409 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
18410 re-try the match with swapped commutative operands.
18411 (vect_supported_load_permutation_p): Remove special-casing of
18412 mismatched complex loads.
18413 (vect_analyze_slp_instance): Adjust.
18415 2013-04-17 Richard Biener <rguenther@suse.de>
18417 PR rtl-optimization/56921
18418 * cfgloop.h (struct loop): Add simple_loop_desc member.
18419 (struct niter_desc): Mark with GTY(()).
18420 (simple_loop_desc): Do not use aux field but simple_loop_desc.
18421 * loop-iv.c (get_simple_loop_desc): Likewise.
18422 (free_simple_loop_desc): Likewise.
18425 2013-04-16 Richard Biener <rguenther@suse.de>
18427 PR rtl-optimization/56921
18428 * loop-init.c (pass_rtl_move_loop_invariants): Add
18429 TODO_do_not_ggc_collect to todo_flags_finish.
18430 (pass_rtl_unswitch): Same.
18431 (pass_rtl_unroll_and_peel_loops): Same.
18432 (pass_rtl_doloop): Same.
18434 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
18436 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
18437 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
18438 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
18440 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
18441 * tree-streamer.c (record_common_node): Adjust reference in comment.
18443 2013-04-17 Terry Guo <terry.guo@arm.com>
18445 * config/arm/cortex-m4.md: Add a new bypass.
18447 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18449 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
18451 (*subs_<optab><mode>_multp2): New pattern.
18452 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
18453 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
18455 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18457 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
18458 (*subs_mul_imm_<mode>): New pattern.
18460 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
18463 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
18464 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
18465 (vsx_movti_32bit): Change j->wa to O->wa.
18467 2013-04-16 Richard Biener <rguenther@suse.de>
18469 PR rtl-optimization/56921
18470 * loop-init.c (pass_rtl_move_loop_invariants): Add
18471 TODO_do_not_ggc_collect to todo_flags_finish.
18472 (pass_rtl_unswitch): Same.
18473 (pass_rtl_unroll_and_peel_loops): Same.
18474 (pass_rtl_doloop): Same.
18476 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
18478 * config/arm/arm.c (emit_multi_reg_push): New declaration
18479 for an existing function.
18480 (arm_emit_strd_push): New function.
18481 (arm_expand_prologue): Used here.
18482 (arm_emit_ldrd_pop): New function.
18483 (arm_expand_epilogue): Used here.
18484 (arm_get_frame_offsets): Update condition.
18485 (arm_emit_multi_reg_pop): Add a special case for load of a single
18486 register with writeback.
18488 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
18490 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
18493 2013-04-16 Richard Biener <rguenther@suse.de>
18495 PR tree-optimization/56756
18496 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
18497 (first_mem_ref_loc): New.
18498 (execute_sm): Place the load temporarily before a previous
18499 access instead of in the latch edge to ensure its SSA dependencies
18500 are defined at points dominating the load.
18502 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
18504 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
18505 correct fix by moving header and footer insn to the footer of
18506 the merged basic block. Clear BB_END of the merged-away block.
18508 PR middle-end/43631
18509 * emit-rtl.c (make_note_raw): New function.
18510 (link_insn_into_chain): New static inline function.
18511 (add_insn): Use it.
18512 (add_insn_before, add_insn_after): Factor insn chain linking code...
18513 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
18514 using link_insn_into_chain.
18515 (note_outside_basic_block_p): New helper function for emit_note_after
18516 and emit_note_before.
18517 (emit_note_after): Use nobb variant of add_insn_after if the note
18518 should not be contained in a basic block.
18519 (emit_note_before): Use nobb variant of add_insn_before if the note
18520 should not be contained in a basic block.
18521 (emit_note_copy): Use make_note_raw.
18522 (emit_note): Likewise.
18523 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
18524 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
18525 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
18526 the moved barrier the tail of the basic block it follows.
18527 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
18529 2013-04-15 Jakub Jelinek <jakub@redhat.com>
18531 PR tree-optimization/56962
18532 * gimple-ssa-strength-reduction.c (record_increment): Only set
18533 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
18534 either rhs1 or rhs2 is equal to c->base_expr.
18536 2013-04-15 Richard Biener <rguenther@suse.de>
18538 PR tree-optimization/56933
18539 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
18541 (GROUP_READ_WRITE_DEPENDENCE): Remove.
18542 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
18543 * tree-vect-data-refs.c (vect_analyze_group_access): Move
18544 dependence check ...
18545 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
18547 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
18548 GROUP_READ_WRITE_DEPENDENCE.
18550 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18552 * emit-rtl.c (reset_all_used_flags): New function.
18553 (verify_rtl_sharing): Call reset_all_used_flags before and after
18554 performing the checks.
18556 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18558 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
18559 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
18560 * config/arm/constraints.md (De): New constraint.
18561 * config/arm/neon.md (anddi3_neon): Delete.
18562 (neon_vand<mode>): Expand to standard anddi3 pattern.
18563 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
18564 Move earlier in the file.
18565 (neon_inv_logic_op2): Likewise.
18566 (arm_anddi_operand_neon): New predicate.
18568 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18570 * configure.ac (gcc_cv_ld_as_needed): Set
18571 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
18572 Use -z ignore, -z record on *-*-solaris2*.
18573 (HAVE_LD_AS_NEEDED): Update comment.
18574 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
18575 * configure: Regenerate.
18576 * config.in: Regenerate.
18577 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
18578 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
18579 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
18580 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
18581 equivalents. Fix markup.
18582 * doc/tm.texi: Regenerate.
18584 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
18586 * config/i386/i386.opt: New option mstack-protector-guard=.
18587 * config/i386/i386-opts.h: Add enum stack_protector_guard.
18588 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
18589 TARGET_SSP_TLS_GUARD.
18590 * config/i386/i386.c (ix86_option_override_internal): Set
18591 ix86_stack_protector_guard.
18592 * config/i386/i386.md (stack_protect_set): Enable for
18593 TARGET_SSP_TLS_GUARD only.
18594 (stack_protect_set_<mode>): Ditto.
18595 (stack_protect_test): Ditto.
18596 (stack_protect_test_<mode>): Ditto.
18597 * doc/invoke.texi (i386 Option): Document.
18599 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
18602 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
18603 (S_MODES): Set H_MODE bit.
18604 (SF_MODES): Set only S_MODE and SF_MODE bits.
18605 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
18606 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
18607 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
18608 <MODE_FLOAT>: Likewise.
18610 2013-04-15 Joey Ye <joey.ye@arm.com>
18612 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
18614 2013-04-15 Joey Ye <joey.ye@arm.com>
18616 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
18618 (thumb_far_jump_used_p): Count instruction size and set
18621 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
18623 * reorg.c (fill_simple_delay_slots): Reindent block of code.
18624 * resource.c (mark_target_live_regs): Reformat conditional block.
18626 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
18628 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
18629 notes, they are emitted only just before final.
18630 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
18632 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
18634 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
18635 * cfgrtl.c (delete_insn): Call it here instead.
18636 * lra-spills.c (lra_final_code_change): Use delete_insn.
18637 * haifa-sched.c (sched_remove_insn): Likewise.
18638 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
18639 returning to the nop pool.
18640 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
18641 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
18643 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
18645 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
18646 * doc/tm.texi: Regenerated.
18648 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
18650 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
18653 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
18655 * df-core.c (df_find_def): Compare register numbers.
18656 (df_find_use): Likewise.
18658 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
18661 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
18662 lra_in_progress for return.
18664 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
18666 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
18667 define_insn into define_insn_and_split and emit movsicc patterns.
18669 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
18671 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
18673 2013-04-12 Richard Biener <rguenther@suse.de>
18675 * tree-pass.h (TODO_do_not_ggc_collect): New.
18676 * passes.c (execute_one_ipa_transform_pass): Honor
18677 TODO_do_not_ggc_collect.
18678 (execute_one_pass): Likewise.
18681 2013-04-10 Richard Biener <rguenther@suse.de>
18683 * passes.c (init_optimization_passes): Remove reload pass.
18684 * ira.c (do_reload): Merge into ...
18686 (rest_of_handle_reload): Remove.
18687 (pass_reload): Likewise.
18688 * config/i386/i386.c (ix86_option_override): Refer to ira instead
18689 of reload for vzeroupper pass placement.
18691 2013-04-12 Jakub Jelinek <jakub@redhat.com>
18693 PR tree-optimization/56918
18694 PR tree-optimization/56920
18695 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
18696 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
18697 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
18698 use wide_mul_with_sign method.
18700 2013-04-12 Richard Biener <rguenther@suse.de>
18702 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
18703 not be considered a gimple constant.
18705 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
18707 * fold-const.c (const_binop): Handle vector shifts by a scalar.
18708 (fold_binary_loc): Call const_binop also for mixed vector-scalar
18711 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
18712 Jakub Jelinek <jakub@redhat.com>
18714 * opts.c: Include diagnostic-color.h.
18715 (common_handle_option): Handle OPT_fdiagnostics_color_.
18716 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
18717 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
18718 (diagnostic-color.o): New.
18719 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
18720 (diagnostic_color_rule): New enum.
18721 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
18722 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
18723 the location string.
18724 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
18725 either NULL, or color kind.
18726 * diagnostic-color.c: New file.
18727 * diagnostic-color.h: New file.
18728 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
18730 * doc/invoke.texi (-fdiagnostics-color): Document.
18731 * pretty-print.h (pp_show_color): Define.
18732 (struct pretty_print_info): Add show_color field.
18733 * diagnostic.c: Include diagnostic-color.h.
18734 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
18735 macros. Colorize error:, warning: etc. strings and also the location
18737 (diagnostic_show_locus): Colorize the caret line.
18738 * pretty-print.c: Include diagnostic-color.h.
18739 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
18740 inside of %< %> quotes or quoted through q format modifier.
18742 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18744 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
18746 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18748 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
18749 code in CC_NZ mode.
18750 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
18753 2013-04-11 Marek Polacek <polacek@redhat.com>
18755 PR tree-optimization/48184
18756 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
18758 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
18760 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
18761 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
18762 (skip_simple_arithmetic): Tidy up.
18763 * tree.h (skip_simple_constant_arithmetic): Declare.
18765 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18767 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
18769 2013-04-11 Richard Biener <rguenther@suse.de>
18771 * tree-vect-loop.c (get_initial_def_for_induction): Properly
18772 generate vector constants.
18774 2013-04-11 Richard Biener <rguenther@suse.de>
18776 PR tree-optimization/56878
18777 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
18778 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
18780 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
18781 Prefer to align the DR with the most invariant base address.
18783 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18785 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
18788 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
18790 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
18791 floating-point vector comparisons against 0.
18793 2013-04-11 Jakub Jelinek <jakub@redhat.com>
18795 PR tree-optimization/56899
18796 * fold-const.c (extract_muldiv_1): Apply distributive law
18797 only if TYPE_OVERFLOW_WRAPS (ctype).
18799 2013-04-11 Bin Cheng <bin.cheng@arm.com>
18802 * ira-costs.c (scan_one_insn): Check whether the source rtx of
18803 loading has side effect.
18805 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
18807 * config/sparc/sparc.c: Include tree-pass.h.
18808 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
18809 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
18810 head of file. Change return type. Split off gate function.
18811 (sparc_gate_work_around_errata): New function.
18812 (pass_work_around_errata): New pass definition.
18813 (insert_pass_work_around_errata) New pass insert definition to
18814 insert pass_work_around_errata just after delayed-branch scheduling.
18815 (sparc_option_override): Insert the pass.
18816 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
18818 2013-04-10 David S. Miller <davem@davemloft.net>
18820 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
18821 or -mcpu=hypersparc.
18823 * target.def (cstore_mode): New hook.
18824 * target.h: Include insn-codes.h
18825 * targhooks.c: Likewise.
18826 (default_cstore_mode): New function.
18827 * targhooks.h: Declare it.
18828 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
18829 * doc/tm.texi: Rebuild.
18830 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
18831 target hook, rather than inspecting the insn_data.
18832 * config/sparc/sparc.c (sparc_cstore_mode): New function.
18833 (TARGET_CSTORE_MODE): Redefine.
18834 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
18836 * config/sparc/predicates.md (cstore_result_operand): New special
18838 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
18839 Use it for operand 0.
18840 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
18841 (*snesi_special): Likewise.
18842 (*snesi_zero): Likewise.
18843 (*seqsi_zero): Likewise.
18844 (*sltu_insn): Likewise.
18845 (*sgeu_insn): Likewise.
18846 (*seqdi_special): Make operand 0 and comparison operation be of
18848 (*snedi_special): Likewise.
18849 (*snedi_special_vis3): Likewise.
18850 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
18851 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
18852 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
18853 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
18854 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
18855 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
18856 (*sltu_extend_sp64): Likewise.
18857 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
18858 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
18859 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
18860 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
18861 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
18863 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
18865 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
18866 (aarch64_start_file): Use the new function.
18868 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18869 Jason Merrill <jason@redhat.com>
18871 * common.opt: Add -gdwarf.
18872 * opts.c (common_handle_option): Handle it.
18873 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
18875 2013-04-10 Richard Biener <rguenther@suse.de>
18877 * passes.c (execute_todo): Do not call ggc_collect conditional here.
18878 (execute_one_ipa_transform_pass): But unconditionally here.
18879 (execute_one_pass): And here.
18880 (init_optimization_passes): Remove reload pass.
18881 * tree-pass.h (TODO_ggc_collect): Remove.
18882 (pass_reload): Likewise.
18883 * ira.c (do_reload): Merge into ...
18885 (rest_of_handle_reload): Remove.
18886 (pass_reload): Likewise.
18887 * config/i386/i386.c (ix86_option_override): Refer to ira instead
18888 of reload for vzeroupper pass placement.
18889 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
18890 and todo_flags_finish of all passes.
18892 2013-04-10 Richard Biener <rguenther@suse.de>
18894 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
18895 first_const_oprnd field, rename first_def_type to first_op_type.
18896 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
18897 (vect_get_and_check_slp_defs): Always use the type of the
18898 operand. Allow mixed vect_external_def, vect_constant_def types.
18899 (vect_get_constant_vectors): Handle mixed vect_external_def,
18900 vect_constant_def types.
18902 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
18904 PR tree-optimization/55524
18905 * tree-ssa-math-opts.c
18906 (convert_mult_to_fma): Don't use an fms construct
18907 when we don't have an fms operation, but fnma, and it looks
18908 likely that we'll be able to use the latter.
18910 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
18912 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
18914 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
18915 inline fail caused by overwritable functions.
18917 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
18919 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
18920 unnecessary bits in the constant power of two case.
18922 2013-04-10 Richard Biener <rguenther@suse.de>
18924 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
18925 broken code swapping operands.
18926 (vect_build_slp_tree): Do not compute load permutations here.
18927 (vect_analyze_slp_instance): Compute load permutations here,
18928 after building the SLP tree.
18930 2013-04-09 Christian Bruel <christian.bruel@st.com>
18932 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
18933 of next/prev_real_insn.
18935 2013-04-09 Jan Hubicka <jh@suse.cz>
18937 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18938 Drop aliased parameter.
18939 (function_and_variable_visibility): Do not handle alias pairs.
18940 * cgraph.c (varpool_externally_visible_p): Update prototype.
18941 * varpool.c (varpool_add_new_variable): Update.
18943 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18945 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
18947 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
18949 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
18951 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
18953 2013-04-09 Marek Polacek <polacek@redhat.com>
18955 PR tree-optimization/48762
18956 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
18958 2013-04-09 Richard Biener <rguenther@suse.de>
18960 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
18962 (vect_build_slp_tree): Likewise.
18963 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
18964 calculating the cost of a SLP instance.
18965 (vect_analyze_slp_instance): Use it from here, after building
18968 2013-04-09 Jakub Jelinek <jakub@redhat.com>
18970 PR middle-end/56883
18971 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
18972 expand_omp_for_static_chunk): Use simple_p = true in
18973 force_gimple_operand_gsi calls when assigning to addressable decls.
18975 2013-04-09 Jeff Law <law@redhat.com>
18977 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
18978 when the boolean was created by converting a wider object which
18979 had a boolean range.
18981 2013-04-09 Richard Biener <rguenther@suse.de>
18983 * tree-vectorizer.h (slp_void_p): Remove.
18984 (slp_tree): Typedef before _slp_tree declaration.
18985 (struct _slp_tree): Use a vector of slp_tree as children.
18986 (vect_get_place_in_interleaving_chain): Remove.
18987 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
18989 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
18991 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
18992 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
18993 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
18994 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
18995 Use slp_node instead of slp_void_p and adjust.
18997 2013-04-09 Richard Biener <rguenther@suse.de>
18999 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
19000 work that is not necessary.
19002 2013-04-09 Jakub Jelinek <jakub@redhat.com>
19004 PR tree-optimization/56854
19005 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
19006 forward into clobber stmts if it would change MEM_REF lhs into
19009 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
19011 * tree.c (type_hash_lookup, type_hash_add): Make static.
19012 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
19014 2013-04-09 Richard Biener <rguenther@suse.de>
19016 * tree.h (unsave_expr_now): Remove.
19017 * tree-inline.c (mark_local_for_remap_r): Remove.
19018 (unsave_expr_1): Likewise.
19019 (unsave_r): Likewise.
19020 (unsave_expr_now): Likewise.
19021 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
19022 (propagate_tree_value): Likewise.
19024 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
19026 * doc/rtl.texi (sequence): Rewrite documentation to match the
19027 current use of SEQUENCE rtl objects.
19028 * rtl.def (SEQUENCE): Likewise.
19030 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
19031 Update documentation.
19032 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
19033 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
19035 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
19037 2013-04-08 Teresa Johnson <tejohnson@google.com>
19039 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
19040 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
19042 (estimate_edge_size_and_time): Add comment to suggest using rounding
19044 (estimate_node_size_and_time): Ditto.
19045 (remap_edge_change_prob): Use helper rounding divide methods.
19046 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
19047 (gimple_mod_pow2_value_transform): Ditto.
19048 (gimple_mod_subtract_transform): Ditto.
19049 (gimple_ic_transform): Ditto.
19050 (gimple_stringops_transform): Ditto.
19051 * stmt.c (conditional_probability): Ditto.
19052 (emit_case_dispatch_table): Ditto.
19053 * lto-cgraph.c (merge_profile_summaries): Ditto.
19054 * tree-optimize.c (execute_fixup_cfg): Ditto.
19055 * cfgcleanup.c (try_forward_edges): Ditto.
19056 * cfgloopmanip.c (scale_loop_profile): Ditto.
19058 (duplicate_loop_to_header_edge): Ditto.
19059 (lv_adjust_loop_entry_edge): Ditto.
19060 * tree-vect-loop.c (vect_transform_loop): Ditto.
19061 * profile.c (compute_branch_probabilities): Ditto.
19062 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
19063 * lto-streamer-in.c (input_cfg): Ditto.
19064 * gimple-streamer-in.c (input_bb): Ditto.
19065 * ipa-cp.c (update_profiling_info): Ditto.
19066 (update_specialized_profile): Ditto.
19067 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
19068 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
19070 * sched-rgn.c (compute_dom_prob_ps): Ditto.
19071 (compute_trg_info): Ditto.
19072 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
19073 (purge_dead_edges): Ditto.
19074 * loop-unswitch.c (unswitch_loop): Ditto.
19075 * cgraphclones.c (cgraph_clone_edge): Ditto.
19076 (cgraph_clone_node): Ditto.
19077 * tree-inline.c (copy_bb): Ditto.
19078 (copy_edges_for_bb): Ditto.
19079 (initialize_cfun): Ditto.
19080 (copy_cfg_body): Ditto.
19081 (expand_call_inline): Ditto.
19083 2013-04-08 Kai Tietz <ktietz@redhat.com>
19085 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
19086 TARGET_CYGWIN64 by TARGET_64BIT.
19088 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
19090 * config/epiphany/epiphany.md (GPR_1): New constant.
19091 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
19092 * config/epiphany/epiphany.c (gen_compare_reg):
19093 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
19094 is already in place.
19095 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
19096 Don't require being called during rtl expansion; If y operlaps r0,
19098 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
19099 (epiphany_expand_epilogue): Likewise.
19101 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
19102 Don't use CC_FPmode for ORDERED / UNORDERED.
19103 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
19105 * config/epiphany/constraints.md (CnL): New constraint.
19106 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
19107 * config/epiphany/predicates.md (add_operand): Allow 1024.
19109 * config/epiphany/epiphany.md (logical_op): New code iterator.
19110 (op_mnc): New code attribute.
19111 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
19112 (mov_f+1, mov_f+2): New peephole2 patterns.
19114 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
19115 (cstoresi4): Also allow re-use of zero result when doing a NE
19116 comparison to a non-zero operand.
19117 Use (clobber (scratch)) for first insn if the gpr output is not needed.
19119 * config/epiphany/epiphany.md (<insn_opname>v2si3):
19120 Use gen_addsi3_i / gen_subsi3_i.
19122 2013-04-08 Jakub Jelinek <jakub@redhat.com>
19126 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
19127 contain anything but clobbers, at most one __builtin_stack_restore,
19128 optionally debug stmts and final resx, and if it has at least one
19129 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
19130 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
19131 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
19132 which isn't defaut definition, remove them.
19133 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
19134 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
19135 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
19136 with MEM_REF LHS with SSA_NAME address.
19138 2013-04-08 Jeff Law <law@redhat.com>
19140 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
19142 2013-04-08 Richard Biener <rguenther@suse.de>
19144 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
19146 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
19147 determined vector type.
19148 (vect_analyze_data_refs): Likewise.
19149 (vect_get_new_vect_var): Adjust.
19150 (vect_create_destination_var): Preserve SSA name versions.
19151 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
19152 not dump anything here.
19154 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
19156 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
19157 Add member lr_slot_known.
19158 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
19160 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
19161 Remove code that sets lr_slot_offset according to what a previous
19162 version of epiphany_emit_save_restore used to do.
19163 (epiphany_emit_save_restore): When doing an lr save or restore,
19164 set/verify lr_slot_known and lr_slot_offset.
19166 2013-04-08 Xinyu Qi <xyqi@marvell.com>
19169 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
19172 2013-04-08 Richard Biener <rguenther@suse.de>
19174 * alias.c (find_base_term): Fix thinko in previous change.
19176 2013-04-08 Jakub Jelinek <jakub@redhat.com>
19178 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
19179 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
19180 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
19181 if possible to compute val.
19182 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
19183 For QImode integers don't require anything about precision. Use
19184 const_with_all_bytes_same to find out if the constant doesn't have
19185 repeated bytes in it.
19187 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19189 * config/s390/s390.c (s390_expand_insv): Only accept insertions
19192 2013-04-08 Marek Polacek <polacek@redhat.com>
19194 PR rtl-optimization/48182
19195 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
19198 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19201 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
19202 nuses, make sure we have a label.
19204 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19207 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
19208 (rs6000_emit_swdiv_low_precision): Remove.
19209 (rs6000_emit_swdiv): Rewrite to handle between one and four
19210 iterations of Newton-Raphson generally; modify required number of
19211 iterations for some cases.
19212 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
19214 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
19216 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
19217 set-but-unused variable.
19219 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
19220 basic blocks of released function bodies garbage-collectable.
19222 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
19223 (struct rtl_opt_pass): Add TODO_df_finish.
19225 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
19227 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19229 * config/arm/constraints.md (q): New constraint.
19230 * config/arm/ldrdstrd.md: New file.
19231 * config/arm/arm.md (ldrdstrd.md) New include.
19232 (arm_movdi): Use "q" instead of "r" constraint
19233 for double-word memory access.
19234 (movdf_soft_insn): Likewise.
19235 * config/arm/vfp.md (movdi_vfp): Likewise.
19236 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
19237 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
19238 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
19239 (mem_ok_for_ldrd_strd): Likewise.
19240 (output_move_double): Update assertion.
19242 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19244 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
19246 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19248 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
19249 define_insn_and_split.
19250 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
19252 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19254 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
19255 define_insn_and_split.
19256 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
19257 (shiftsi3_compare): New pattern.
19258 (rrx): New pattern.
19259 * config/arm/unspecs.md (UNSPEC_RRX): New.
19261 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19263 * config/arm/arm.md (negdi_extendsidi): New pattern.
19264 (negdi_zero_extendsidi): Likewise.
19266 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19268 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
19269 define_insn_and_split.
19270 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
19271 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
19273 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19275 * config/arm/arm.md (arm_subdi3): Convert define_insn into
19276 define_insn_and_split.
19277 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
19278 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
19280 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19282 * config/arm/arm.md (subsi3_carryin): New pattern.
19283 (subsi3_carryin_const): Likewise.
19284 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
19285 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
19287 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19289 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
19291 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19293 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
19294 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
19296 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19298 * config/arm/arm.c (arm_expand_builtin): Change fcode
19299 type to unsigned int.
19301 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19303 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
19305 2013-04-04 Ian Lance Taylor <iant@google.com>
19307 * doc/standards.texi (Standards): The Go frontend supports the Go 1
19310 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
19312 PR middle-end/56729
19313 * df-scan.c (df_insn_delete): Disable failing assert.
19315 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19317 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
19318 New function prototype.
19319 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
19320 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
19321 (arm_builtin_vectorized_function): New function.
19323 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19325 * config/arm/arm_neon_builtins.def: New file.
19326 * config/arm/arm.c (neon_builtin_data): Move contents to
19327 arm_neon_builtins.def.
19328 (enum arm_builtins): Include neon builtin definitions.
19329 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
19330 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
19332 2013-04-04 Marek Polacek <polacek@redhat.com>
19334 PR tree-optimization/48186
19335 * predict.c (maybe_hot_frequency_p): Return false if
19336 HOT_BB_FREQUENCY_FRACTION is 0.
19337 (cgraph_maybe_hot_edge_p): Likewise.
19339 2013-04-04 Richard Biener <rguenther@suse.de>
19341 PR tree-optimization/56826
19342 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
19345 2013-04-04 Richard Biener <rguenther@suse.de>
19347 PR tree-optimization/56213
19348 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
19349 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
19351 2013-04-04 Richard Biener <rguenther@suse.de>
19353 PR tree-optimization/56837
19354 * tree-loop-distribution.c (classify_partition): For non-zero
19355 values require that the value has the same precision as its
19356 mode to be useful as memset value.
19358 2013-04-03 Nick Clifton <nickc@redhat.com>
19360 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
19361 (fmssf4): Use fmsf.s on E3V5 architectures.
19362 (fnmasf4): Use fnmaf.s on E3V5 architectures.
19363 (fnmssf4): Use fnmsf.s on E3V5 architectures.
19365 2013-04-03 Jeff Law <law@redhat.com>
19367 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
19368 (lra-eliminations.o): Likewise.
19370 2013-04-03 Teresa Johnson <tejohnson@google.com>
19372 * gcov-io.c (compute_working_sets): Moved most of body of old
19373 compute_working_sets here from profile.c.
19374 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
19375 (gcov_working_set_t): Moved typedef here from basic-block.h
19376 (compute_working_set): Declare.
19377 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
19378 (get_working_sets): Renamed from compute_working_set,
19379 replace most of body with call to new compute_working_sets.
19380 (get_exec_counts): Replace call to compute_working_sets
19381 to get_working_sets.
19382 * profile.h (get_working_sets): Renamed from compute_working_set.
19383 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
19384 to get_working_sets.
19385 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
19386 * gcov-dump.c (dump_working_sets): New function.
19388 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
19390 * hwint.c (sext_hwi, zext_hwi): New functions.
19391 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
19392 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
19393 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
19394 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
19395 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
19396 (sext_hwi, zext_hwi): New functions.
19398 2013-04-03 Jeff Law <law@redhat.com>
19400 PR tree-optimization/56799
19401 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
19402 back test for widening conversion erroneously dropped in prior change.
19404 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19407 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
19408 instead of next_real_insn.
19410 2013-04-03 Marek Polacek <polacek@redhat.com>
19413 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
19415 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19418 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
19420 (thumb1_output_casesi): Likewise.
19421 (thumb2_output_casesi): Likewise.
19423 2013-04-03 Richard Biener <rguenther@suse.de>
19425 PR tree-optimization/56817
19426 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
19428 (tree_unroll_loops_completely_1): ... new function to manually
19429 walk the loop tree, properly defering outer loops of unrolled
19430 loops to later iterations.
19432 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19434 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
19435 (vectorizable_load): Likewise.
19436 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
19437 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
19439 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19441 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
19444 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
19446 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
19448 2013-04-03 Bin Cheng <bin.cheng@arm.com>
19450 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
19452 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19454 PR tree-optimization/56790
19455 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
19458 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19460 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
19462 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
19463 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
19466 2013-04-03 Jakub Jelinek <jakub@redhat.com>
19469 * tree.h (force_folding_builtin_constant_p): New decl.
19470 * builtins.c (force_folding_builtin_constant_p): New variable.
19471 (fold_builtin_constant_p): Fold immediately also if
19472 force_folding_builtin_constant_p.
19474 2013-04-03 Richard Biener <rguenther@suse.de>
19476 PR tree-optimization/56812
19477 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
19478 DRs of the same interleaving chain are independent.
19480 2013-04-02 Jason Merrill <jason@redhat.com>
19482 * gdbinit.in (pbb): Use debug fn.
19484 2013-04-02 Lawrence Crowl <crowl@google.com>
19486 * sese.h (struct ivtype_map_elt_s): Remove unused.
19487 (extern debug_ivtype_map): Remove unused.
19488 (extern eq_ivtype_map_elts): Remove unused.
19489 * sese.c (debug_ivtype_map): Removed unused.
19490 (debug_ivtype_map_1): Removed unused.
19491 (debug_ivtype_elt): Remove unused.
19492 (eq_ivtype_map_elts): Remove unused.
19495 2013-04-02 Kai Tietz <ktietz@redhat.com>
19498 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
19499 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
19500 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
19502 (legitimize_pe_coff_symbol): Likewise.
19503 (is_imported_p): New helper-function.
19504 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
19505 for Windows x64 targets.
19506 (ix86_expand_prologue): Optimize for pe-coff targets.
19507 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
19508 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
19509 medium/large code-model.
19510 (legitimize_pic_address): Likewise.
19511 (legitimize_tls_address): Likewise.
19512 (ix86_expand_call): Likewise.
19513 (x86_output_mi_thunk): Likewise.
19514 (get_dllimport_decl): Add new beimport argument.
19515 (construct_plt_address): Don't assert for x64 pe-coff targets.
19516 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
19518 (SYMBOL_FLAG_STUBVAR): New macro.
19519 (SYMBOL_REF_STUBVAR_P): Likewise.
19520 * config/i386/winnt.c (stub_list): New structure.
19521 (stub_head): New local variable.
19522 (i386_pe_record_stub): New function.
19523 (i386_pe_file_end): Emit refptr-stubs.
19525 2013-04-02 Jakub Jelinek <jakub@redhat.com>
19527 PR rtl-optimization/56745
19528 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
19529 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
19532 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
19533 and both of them are MEM_REFs, just compare first argument for
19534 equality and attempt to deal even with differing offsets.
19537 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
19538 of gimple_clobber_p to be MEM_REF.
19539 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
19540 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
19541 after gimplification.
19542 * asan.c (get_mem_ref_of_assignment): Don't instrument
19543 gimple_clobber_p stmts.
19544 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
19545 gimple_clobber_p stmt if they have MEM_REF lhs and
19546 are dead because of another gimple_clobber_p stmt.
19547 * tree-ssa-live.c (clear_unused_block_pointer): Treat
19548 gimple_clobber_p stmts like debug stmts.
19549 (remove_unused_locals): Remove clobbers with MEM_REF lhs
19550 that refer to unused VAR_DECLs or uninitialized values.
19551 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
19552 gimple_clobber_p stmts if they refer to removed parameters.
19553 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
19556 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
19558 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
19559 using SWI48 mode attribute.
19561 2013-04-02 Wei Mi <wmi@google.com>
19563 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
19564 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
19565 *<rotate_insn><mode>3_mask in i386.md.
19567 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
19569 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
19571 2013-04-02 Richard Biener <rguenther@suse.de>
19573 PR tree-optimization/56778
19574 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
19575 Runtime alias tests are not supported for gather loads.
19576 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
19577 stmts referenced from SSA operands before updating SSA form.
19579 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
19580 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19582 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
19583 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
19584 * config/arm/cortex-a53.md: New file.
19585 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
19586 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
19587 * config/arm/arm.c (arm_issue_rate): Likewise.
19588 * config/arm/arm-tune.md: Regenerate
19589 * config/arm/arm-tables.opt: Regenerate.
19590 * config/arm/arm-cores.def: Add cortex-a53.
19592 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
19594 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
19597 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
19599 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
19600 scalar load/store operations using B/H registers.
19601 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
19603 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
19605 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
19607 * config/aarch64/aarch64.c
19608 (aarch64_simd_scalar_immediate_valid_for_move): New.
19609 * config/aarch64/aarch64-protos.h
19610 (aarch64_simd_scalar_immediate_valid_for_move): New.
19611 * config/aarch64/constraints.md (Dh, Dq): New.
19612 * config/aarch64/iterators.md (hq): New.
19614 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
19616 * reorg.c (get_branch_condition): Deal with conditional returns.
19617 (fill_simple_delay_slots): Remove dead code dealing with jumps.
19619 2013-04-01 Wei Mi <wmi@google.com>
19621 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
19622 Truncate operand 2 using %b asm operand modifier.
19623 (*<shift_insn><mode>3_mask): Ditto.
19624 (*<rotate_insn><mode>3_mask): Ditto.
19626 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
19628 PR middle-end/56798
19629 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
19631 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
19633 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
19635 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
19637 2013-03-30 Lawrence Crowl <crowl@google.com>
19639 * dse.c (clear_alias_sets): Remove never set.
19640 (disqualified_clear_alias_sets): Remove never set.
19641 (clear_alias_mode_pool): Remove never set.
19642 (dse_step0): Remove condition that is never true.
19643 (canon_address): Remove condition that is never true.
19644 (dse_step7): Remove condition that is never true.
19645 (rest_of_handle_dse): Remove condition that is never true.
19646 (rest_of_handle_dse::did_global): Remove never read from above.
19647 (dse_step2_spill): Remove never called from above.
19648 (dse_step5_spill): Remove never called from above.
19650 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
19652 * doc/md.texi (Standard Names) <casesi>: Update documentation for
19653 JUMP_TABLE_DATA changes.
19654 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
19655 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
19656 (Insns) <jump_table_data>: New entry.
19657 * doc/tm.texi: Regenerate.
19659 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
19661 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
19662 for table jump at the end of a basic block using tablejump_p.
19663 * targhooks.c (default_invalid_within_doloop): Likewise.
19664 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
19665 target hook implementation that is identical to the default hook.
19666 (rs6000_invalid_within_doloop): Remove.
19668 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
19669 unused variable from tablejump_p call.
19671 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
19672 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
19673 (INSN_DELETED_P): Likewise.
19674 (emit_jump_table_data): New prototype.
19675 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
19676 after 4th as unused.
19677 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
19678 * sched-vis.c (print_insn): Likewise.
19679 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
19680 insn for compatibility with back ends that use next_active_insn to
19681 identify jump table data.
19682 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
19683 (remove_insn): Likewise.
19684 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
19686 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
19687 (emit_jump_table_data): New function.
19689 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
19690 basic block, a JUMP_TABLE_DATA never is.
19691 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
19692 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
19693 off from code handling real insns.
19694 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
19695 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
19696 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
19697 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
19698 is not a NONDEBUG_INSN_P.
19699 * ira-costs.c (scan_one_insn): Likewise.
19700 * jump.c (mark_all_labels): Likewise.
19701 (mark_jump_label_1): Likewise.
19702 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
19703 * lra.c (get_insn_freq): Expect all insns reaching here to be in
19705 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
19706 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
19707 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
19708 JUMP_TABLE_DATA_P insns.
19709 (calculate_elim_costs_all_insns): Likewise.
19710 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
19711 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
19712 (delete_output_reload): Code style fixups.
19713 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
19714 insn flags on this non-insn.
19715 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
19716 as scheduling barriers, for pre-change compatibility.
19717 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
19718 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
19720 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
19721 redundant JUMP_TABLE_DATA_P test.
19722 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
19723 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
19724 (frv_for_each_packet): Likewise.
19725 * config/i386/i386.c (min_insn_size): Likewise.
19726 (ix86_avoid_jump_mispredicts): Likewise.
19727 * config/m32r/m32r.c (m32r_is_insn): Likewise.
19728 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
19729 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
19730 (mips16_insn_length): Robustify.
19731 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
19732 (mips16_split_long_branches): Likewise.
19733 * config/pa/pa.c (pa_combine_instructions): Likewise.
19734 * config/rs6000/rs6000.c (get_next_active_insn): Treat
19735 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
19736 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
19737 as contributing to pool range lengths.
19738 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
19739 Remove redundant JUMP_TABLE_DATA_P test.
19740 (sh_loop_align): Likewise.
19741 (split_branches): Likewise.
19742 (sh_insn_length_adjustment): Likewise.
19743 * config/spu/spu.c (get_branch_target): Likewise.
19745 2013-03-29 Jan Hubicka <jh@suse.cz>
19747 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
19748 gcov streaming; stream hot bb threshold to ltrans.
19749 * predict.c (get_hot_bb_threshold): Break out from ....
19750 (maybe_hot_count_p): ... here.
19751 (set_hot_bb_threshold): New function.
19752 * lto-section-in.c (lto_section_name): Add profile.
19753 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
19754 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
19755 and data-streamer.h
19756 (histogram_entry): New structure.
19757 (histogram, histogram_pool): New global vars.
19758 (histogram_hash): New structure.
19759 (histogram_hash::hash): New method.
19760 (histogram_hash::equal): Likewise.
19761 (account_time_size): New function.
19762 (cmp_counts): New function.
19763 (dump_histogram): New function.
19764 (ipa_profile_generate_summary): New function.
19765 (ipa_profile_write_summary): New function.
19766 (ipa_profile_read_summary): New function.
19767 (ipa_profile): Decide on threshold.
19768 (pass_ipa_profile): Add ipa_profile_write_summary and
19769 ipa_profile_read_summary.
19770 * Makefile.in (ipa.o): Update dependencies.
19771 * lto-streamer.h (LTO_section_ipa_profile): New section.
19773 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
19775 * tree.h (VAR_P): New.
19777 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
19780 * doc/invoke.texi ([-fwhole-program]): Fix typo.
19782 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
19784 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
19785 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19786 (control_flow_insn_p): Likewise.
19787 * cfgrtl.c (duplicate_insn_chain): Likewise.
19788 * final.c (get_attr_length_1): Likewise.
19789 (shorten_branches): Likewise.
19790 (final_scan_insn): Likewise.
19791 * function.c (instantiate_virtual_regs): Likewise.
19792 * gcse.c (insert_insn_end_basic_block): Likewise.
19793 * ira-costs.c (scan_one_insn): Likewise.
19794 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
19795 * lra.c (check_rtl): Likewise.
19796 * reload1.c (elimination_costs_in_insn): Likewise.
19797 * reorg.c (follow_jumps): Likewise.
19799 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
19800 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19801 (thumb_far_jump_used_p): Likewise.
19802 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
19803 (workaround_speculation): Likewise.
19804 (add_sched_insns_for_speculation): Likewise.
19805 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
19806 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
19807 (frv_for_each_packet): Likewise.
19808 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
19809 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
19810 (final_emit_insn_group_barriers): Likewise.
19811 * config/m32r/m32r.c (m32r_is_insn): Likewise.
19812 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
19813 (mips16_insn_length): Likewise.
19814 * config/pa/pa.c (pa_reorg): Likewise.
19815 (pa_combine_instructions): Likewise.
19816 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
19817 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
19818 (sh_reorg): Likewise.
19819 (split_branches): Likewise.
19820 * config/spu/spu.c (get_branch_target): Likewise.
19822 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
19825 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
19827 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
19828 Fix declaration name.
19830 2013-03-28 Lawrence Crowl <crowl@google.com>
19832 * graphds.h (struct graph.indicies): Remove unused.
19833 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
19834 (SCOP_ORIGINAL_PDDRS): Remove unused.
19835 * sese.h (extern insert_loop_close_phis): Removed unused.
19836 (extern insert_guard_phis): Removed unused.
19837 (extern ivtype_map_elt_info): Removed unused.
19838 (new_ivtype_map_elt): Removed unused.
19839 * sese.c (ivtype_map_elt_info): Removed unused.
19841 2013-03-28 Lawrence Crowl <crowl@google.com>
19843 * Makefile.in: Add several missing include dependences.
19845 (test-dump.o): New. This object is not added to any executable,
19846 but is present for ad-hoc testing.
19848 (debug (const bitmap_head_def &)): New.
19849 (debug (const bitmap_head_def *)): New.
19851 (extern debug (const bitmap_head_def &)): New.
19852 (extern debug (const bitmap_head_def *)): New.
19854 (debug (edge_def &)): New.
19855 (debug (edge_def *)): New.
19857 (debug (basic_block_def &)): New.
19858 (debug (basic_block_def *)): New.
19860 (dump_node (const_tree, int, FILE *)): Correct source file.
19862 (debug (die_struct &)): New.
19863 (debug (die_struct *)): New.
19865 (extern debug (die_struct &)): New.
19866 (extern debug (die_struct *)): New.
19867 * gimple-pretty-print.c
19868 (debug (gimple_statement_d &)): New.
19869 (debug (gimple_statement_d *)): New.
19870 * gimple-pretty-print.h
19871 (extern debug (gimple_statement_d &)): New.
19872 (extern debug (gimple_statement_d *)): New.
19874 (debug (ira_allocno_copy &)): New.
19875 (debug (ira_allocno_copy *)): New.
19876 (debug (ira_allocno &)): New.
19877 (debug (ira_allocno *)): New.
19879 (extern debug (ira_allocno_copy &)): New.
19880 (extern debug (ira_allocno_copy *)): New.
19881 (extern debug (ira_allocno &)): New.
19882 (extern debug (ira_allocno *)): New.
19884 (debug (live_range &)): New.
19885 (debug (live_range *)): New.
19887 (debug (lra_live_range &)): New.
19888 (debug (lra_live_range *)): New.
19890 (debug (lra_live_range &)): New.
19891 (debug (lra_live_range *)): New.
19893 (debug (omega_pb_d &)): New.
19894 (debug (omega_pb_d *)): New.
19896 (extern debug (omega_pb_d &)): New.
19897 (extern debug (omega_pb_d *)): New.
19899 (debug (const rtx_def &)): New.
19900 (debug (const rtx_def *)): New.
19902 (debug_tree (tree): Move within file.
19903 (debug_raw (const tree_node &)): New.
19904 (debug_raw (const tree_node *)): New.
19905 (dump_tree_via_hooks (const tree_node *, int)): New.
19906 (debug (const tree_node &)): New.
19907 (debug (const tree_node *)): New.
19908 (debug_verbose (const tree_node &)): New.
19909 (debug_verbose (const tree_node *)): New.
19910 (debug_head (const tree_node &)): New.
19911 (debug_head (const tree_node *)): New.
19912 (debug_body (const tree_node &)): New.
19913 (debug_body (const tree_node *)): New.
19914 (debug_vec_tree (tree): Move and reimplement in terms of dump.
19915 (debug (vec<tree, va_gc> &)): New.
19916 (debug (vec<tree, va_gc> *)): New.
19918 (extern debug (const rtx_def &)): New.
19919 (extern debug (const rtx_def *)): New.
19921 (debug_raw (simple_bitmap_def &)): New.
19922 (debug_raw (simple_bitmap_def *)): New.
19923 (debug (simple_bitmap_def &)): New.
19924 (debug (simple_bitmap_def *)): New.
19926 (extern debug (simple_bitmap_def &)): New.
19927 (extern debug (simple_bitmap_def *)): New.
19928 (extern debug_raw (simple_bitmap_def &)): New.
19929 (extern debug_raw (simple_bitmap_def *)): New.
19931 (debug (vinsn_def &)): New.
19932 (debug (vinsn_def *)): New.
19933 (debug_verbose (vinsn_def &)): New.
19934 (debug_verbose (vinsn_def *)): New.
19935 (debug (expr_def &)): New.
19936 (debug (expr_def *)): New.
19937 (debug_verbose (expr_def &)): New.
19938 (debug_verbose (expr_def *)): New.
19939 (debug (vec<rtx> &)): New.
19940 (debug (vec<rtx> *)): New.
19942 (extern debug (vinsn_def &)): New.
19943 (extern debug (vinsn_def *)): New.
19944 (extern debug_verbose (vinsn_def &)): New.
19945 (extern debug_verbose (vinsn_def *)): New.
19946 (extern debug (expr_def &)): New.
19947 (extern debug (expr_def *)): New.
19948 (extern debug_verbose (expr_def &)): New.
19949 (extern debug_verbose (expr_def *)): New.
19950 (extern debug (vec<rtx> &)): New.
19951 (extern debug (vec<rtx> *)): New.
19953 (_list_iter_cond_expr): Make inline instead of static.
19955 (debug (sreal &)): New.
19956 (debug (sreal *)): New.
19958 (extern debug (sreal &)): New.
19959 (extern debug (sreal *)): New.
19961 (extern debug_raw (const tree_node &)): New.
19962 (extern debug_raw (const tree_node *)): New.
19963 (extern debug (const tree_node &)): New.
19964 (extern debug (const tree_node *)): New.
19965 (extern debug_verbose (const tree_node &)): New.
19966 (extern debug_verbose (const tree_node *)): New.
19967 (extern debug_head (const tree_node &)): New.
19968 (extern debug_head (const tree_node *)): New.
19969 (extern debug_body (const tree_node &)): New.
19970 (extern debug_body (const tree_node *)): New.
19971 (extern debug (vec<tree, va_gc> &)): New.
19972 (extern debug (vec<tree, va_gc> *)): New.
19974 (debug (struct loop &)): New.
19975 (debug (struct loop *)): New.
19976 (debug_verbose (struct loop &)): New.
19977 (debug_verbose (struct loop *)): New.
19978 * tree-dump.c: Add header dependence.
19980 (extern debug (struct loop &)): New.
19981 (extern debug (struct loop *)): New.
19982 (extern debug_verbose (struct loop &)): New.
19983 (extern debug_verbose (struct loop *)): New.
19985 (debug (data_reference &)): New.
19986 (debug (data_reference *)): New.
19987 (debug (vec<data_reference_p> &)): New.
19988 (debug (vec<data_reference_p> *)): New.
19989 (debug (vec<ddr_p> &)): New.
19990 (debug (vec<ddr_p> *)): New.
19992 (extern debug (data_reference &)): New.
19993 (extern debug (data_reference *)): New.
19994 (extern debug (vec<data_reference_p> &)): New.
19995 (extern debug (vec<data_reference_p> *)): New.
19996 (extern debug (vec<ddr_p> &)): New.
19997 (extern debug (vec<ddr_p> *)): New.
19999 (debug (pt_solution &)): New.
20000 (debug (pt_solution *)): New.
20002 (extern debug (pt_solution &)): New.
20003 (extern debug (pt_solution *)): New.
20005 (debug (_var_map &)): New.
20006 (debug (_var_map *)): New.
20007 (debug (tree_live_info_d &)): New.
20008 (debug (tree_live_info_d *)): New.
20010 (extern debug (_var_map &)): New.
20011 (extern debug (_var_map *)): New.
20012 (extern debug (tree_live_info_d &)): New.
20013 (extern debug (tree_live_info_d *)): New.
20015 2013-03-28 Jan Hubicka <jh@suse.cz>
20017 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
20019 2013-03-28 Ian Bolton <ian.bolton@arm.com>
20021 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
20022 record only when desired or required.
20024 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
20026 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
20027 *vec_extractv2di_1_rex64. Use x64 isa attribute.
20029 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20031 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
20032 (*andsi3_compare0_uxtw): New pattern.
20033 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
20034 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
20036 2013-03-28 Jan Hubicka <jh@suse.cz>
20038 * data-streamer-in.c (streamer_read_gcov_count): New function.
20039 * gimple-streamer-out.c: Include value-prof.h.
20040 (output_gimple_stmt): Output histogram.
20041 (output_bb): Use streamer_write_gcov_count.
20042 * value-prof.c: Include data-streamer.h
20043 (dump_histogram_value): Add HIST_TYPE_MAX.
20044 (stream_out_histogram_value): New function.
20045 (stream_in_histogram_value): New function.
20046 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
20047 (stream_out_histogram_value, stream_in_histogram_value): Declare.
20048 * data-streamer-out.c (streamer_write_gcov_count): New function.
20049 (streamer_write_gcov_count_stream): New function.
20050 * lto-cgraph.c (lto_output_edge): Update counter streaming.
20051 (lto_output_node): Likewise.
20052 (input_node, input_edge): Likewise.
20053 * lto-streamer-out.c (output_cfg): Update streaming.
20054 * lto-streamer-in.c (input_cfg): Likewise.
20055 * data-streamer.h (streamer_write_gcov_count,
20056 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
20057 * gimple-streamer-in.c: Include value-prof.h
20058 (input_gimple_stmt): Input histograms.
20059 (input_bb): Update profile streaming.
20061 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
20063 * genmodes.c (emit_max_int): New function.
20064 (emit_insn_modes_h): Added call to emit_max_function.
20065 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
20067 * machmode.def: Fixed comment.
20069 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
20071 * combine.c (try_combine): Removed useless assert.
20072 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
20074 2013-03-28 Marek Polacek <polacek@redhat.com>
20075 Richard Biener <rguenther@suse.de>
20077 PR tree-optimization/56695
20078 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
20079 build signed result of a vector comparison.
20080 * tree-cfg.c (verify_gimple_comparison): Check that a result
20081 of a vector comparison has signed type.
20083 2013-03-28 Richard Biener <rguenther@suse.de>
20085 PR tree-optimization/37021
20086 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
20087 do not restrict gaps between groups.
20088 * tree-vect-stmts.c (vectorizable_load): Properly account for
20089 a gap between groups.
20091 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
20093 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
20094 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
20097 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
20099 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
20100 * doc/extend.texi (Named Address Spaces): Ditto.
20101 (Variable Attributes): Ditto.
20103 2013-03-27 Kai Tietz <ktietz@redhat.com>
20105 * config.build: Add support for cygwin x64 target.
20106 * config.gcc: Likewise.
20107 * config.host: Likewise.
20108 * configure.ac: Likewise
20109 * configure: Regenerated.
20111 2013-03-27 Kai Tietz <ktietz@redhat.com>
20113 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
20114 * config/i386/t-cygwin-w64: New file.
20115 * config/i386/cygwin-w64.h: New file.
20116 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
20117 and add support for x64-cygwin target.
20118 (CPP_SPEC): Likewise.
20119 (CXX_WRAP_SPEC_LIST): Undefine before define.
20120 (LIBGCJ_SONAME): Use 15 as version.
20122 2013-03-27 Richard Biener <rguenther@suse.de>
20124 PR tree-optimization/56716
20125 * tree-ssa-structalias.c (perform_var_substitution): Adjust
20126 dumping for ref nodes.
20128 2013-03-27 Martin Jambor <mjambor@suse.cz>
20130 PR tree-optimization/55334
20131 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
20132 restricted pointers to arrays.
20134 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
20136 * Makefile.in (.SUFFIXES): Add .cc.
20137 (.c.o): Apply same recipe for implicit rule .cc.o.
20139 2013-03-27 Richard Biener <rguenther@suse.de>
20141 PR tree-optimization/37021
20142 * tree-vect-data-refs.c (vect_check_strided_load): Allow
20143 REALPART/IMAGPART_EXPRs around the supported refs.
20144 * tree-ssa-structalias.c (find_func_aliases): Assume that
20145 floating-point values are not used to transfer pointers.
20147 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
20149 * target.def (TARGET_HAS_IFUNC_P): New target hook.
20150 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
20151 * doc/tm.texi: Regenerate.
20152 * targhooks.h (default_has_ifunc_p): New.
20153 * targhooks.c (default_has_ifunc_p): Ditto.
20154 * config/linux-protos.h: New file.
20155 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
20156 hook for linux which disables support of indirect functions in android.
20157 * config/linux-android.c: New file.
20158 * config/t-linux-android.c: Ditto.
20159 * config.gcc: Added new object file linux-android.o.
20160 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
20161 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
20162 * varasm.c (do_assemble_alias): Likewise.
20163 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
20164 doesn't support indirect functions.
20165 * configure: Regenerate.
20167 2013-03-27 Bin Cheng <bin.cheng@arm.com>
20170 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
20171 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
20174 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20176 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
20178 2013-03-27 Terry Guo <terry.guo@arm.com>
20180 * config/arm/arm-cores.def: Added core cortex-r7.
20181 * config/arm/arm-tune.md: Regenerated.
20182 * config/arm/arm-tables.opt: Regenerated.
20183 * doc/invoke.texi: Added entry for core cortex-r7.
20185 2013-03-27 Walter Lee <walt@tilera.com>
20187 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
20188 double-decrement of next_scratch_regno.
20190 2013-03-27 Walter Lee <walt@tilera.com>
20192 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
20194 (insn_v1mulus): Ditto.
20195 (insn_v2muls): Ditto.
20197 2013-03-27 Walter Lee <walt@tilera.com>
20199 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
20200 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
20202 2013-03-27 Walter Lee <walt@tilera.com>
20204 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
20205 (*sibcall_value): Ditto.
20207 2013-03-27 Walter Lee <walt@tilera.com>
20209 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
20210 (insn_mnz_v8qi): ... this ...
20211 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
20213 (insn_v<n>mnz): Replaced by ...
20214 (insn_v1mnz): ... this ...
20215 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
20217 (insn_mz_<mode>): Replaced by ...
20218 (insn_mz_v8qi): ... this ...
20219 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
20221 (insn_v<n>mz): Replaced by ...
20222 (insn_v1mz): ... this ...
20223 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
20226 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
20228 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
20230 2013-03-26 Roland McGrath <mcgrathr@google.com>
20232 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
20233 than fprintf with a non-constant, non-format string.
20235 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
20237 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
20238 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
20239 operand 0 predicate.
20240 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
20241 attribute. Use general_x64nomem_operand as operand 1 predicate.
20242 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
20243 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
20244 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
20245 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
20246 (mov<mode>_insv_1): Remove expander. Merge insn with
20247 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
20248 Use general_x64nomem_operand as operand 1 predicate.
20249 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
20250 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
20251 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
20252 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
20253 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
20254 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
20255 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
20256 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
20257 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
20258 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
20259 (general_x64nomem_operand): Ditto.
20261 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
20263 * config/rtems.opt: Add -pthread option.
20265 2013-03-26 Richard Biener <rguenther@suse.de>
20267 * alias.c (find_base_term): Avoid redundant and not used recursion.
20268 (base_alias_check): Get the initial base term from the caller.
20269 (true_dependence_1): Compute and pass base terms to base_alias_check.
20270 (write_dependence_p): Likewise.
20271 (may_alias_p): Likewise.
20273 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
20275 * config/aarch64/aarch64.c (aarch64_classify_address): Support
20276 PC-relative load in SI modes and above only.
20278 2013-03-26 Xinyu Qi <xyqi@marvell.com>
20280 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
20281 * config/arm/iwmmxt.md (WCGR0): Update.
20282 (WCGR1, WCGR2, WCGR3): Likewise.
20284 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
20286 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
20287 Use x64 and nox64 isa attributes.
20289 2013-03-26 Richard Biener <rguenther@suse.de>
20291 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
20292 alignment computations and rely on get_object_alignment_1
20293 for the !TYPE_P case.
20294 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
20296 2013-03-26 Walter Lee <walt@tilera.com>
20298 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
20299 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
20301 2013-03-25 Jeff Law <law@redhat.com>
20303 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
20304 check for INTEGRAL_TYPE_P that was missing due to checking in
20305 wrong version of prior patch.
20307 2013-03-25 Walter Lee <walt@tilera.com>
20309 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
20310 TILEGX_INSN_SHUFFLEBYTES1.
20311 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
20313 (tilegx_builtins): Ditto.
20314 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
20316 2013-03-25 Walter Lee <walt@tilera.com>
20318 * config/tilegx/tilegx.md (floatsisf2): New pattern.
20319 (floatunssisf2): New pattern.
20320 (floatsidf2): New pattern.
20321 (floatunssidf2): New pattern.
20323 2013-03-25 Walter Lee <walt@tilera.com>
20325 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
20326 tests for constraint J, K, N, P.
20328 2013-03-25 Walter Lee <walt@tilera.com>
20330 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
20331 Use indirect/pcrel encoding.
20332 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
20335 2013-03-25 Steve Ellcey <sellcey@mips.com>
20337 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
20338 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
20339 * config/mips/mips.c (mips_option_override): Set IMADD default.
20340 * config/mips/mips.h (PTF_AVOID_IMADD): New.
20341 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
20342 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
20343 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
20344 * doc/invoke.texi (-mimadd/-mno-imadd): New.
20346 2013-03-25 Jeff Law <law@redhat.com>
20348 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
20349 slightly to avoid creating and folding useless trees. Simplify
20350 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
20352 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
20354 * config/i386/i386.md (*zero_extendsidi2): Merge with
20355 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
20356 * config/i386/predicates.md (x86_64_zext_operand): Rename from
20357 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
20358 targets. Clarify comment.
20360 2013-03-25 Martin Jambor <mjambor@suse.cz>
20362 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
20363 pass-through jump functions differently.
20364 (ipa_read_jump_function): Likewise. Also use setter functions to set
20367 2013-03-25 Martin Jambor <mjambor@suse.cz>
20369 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
20370 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
20372 (ipa_get_indirect_edge_target): New function.
20373 (devirtualization_time_bonus): New parameter known_aggs, pass it to
20374 ipa_get_indirect_edge_target. Update all callers.
20375 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
20376 ipa_get_indirect_edge_target_1 instead of calling
20377 ipa_get_indirect_edge_target.
20378 (create_specialized_node): Pass aggvlas to
20379 ipcp_discover_new_direct_edges.
20381 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20383 * config/arm/arm.md (f_sels, f_seld): New types.
20384 (*cmov<mode>): New pattern.
20385 * config/arm/predicates.md (arm_vsel_comparison_operator): New
20388 2013-03-25 Kai Tietz <ktietz@redhat.com>
20390 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
20391 POSIX-printf for mingw-hosted builds.
20393 2013-03-25 Richard Biener <rguenther@suse.de>
20395 PR middle-end/56694
20396 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
20397 must-not-throw stmt location.
20399 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20401 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
20402 Emit load-acquire versions when acq is true.
20403 (arm_emit_store_exclusive): Add rel parameter.
20404 Emit store-release versions when rel is true.
20405 (arm_split_compare_and_swap): Use acquire-release instructions
20407 of barriers when appropriate.
20408 (arm_split_atomic_op): Likewise.
20409 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
20410 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
20411 (VUNSPEC_SLX): Likewise.
20412 (VUNSPEC_LDA): Likewise.
20413 (VUNSPEC_STL): Likewise.
20414 * config/arm/sync.md (atomic_load<mode>): New pattern.
20415 (atomic_store<mode>): Likewise.
20416 (arm_load_acquire_exclusive<mode>): Likewise.
20417 (arm_load_acquire_exclusivesi): Likewise.
20418 (arm_load_acquire_exclusivedi): Likewise.
20419 (arm_store_release_exclusive<mode>): Likewise.
20421 2013-03-25 Catherine Moore <clm@codesourcery.com>
20423 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
20424 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
20425 * config/mip/predicates.md (lwsp_swsp_operand,
20426 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
20427 sb16_operand, db4_operand, db7_operand, ib3_operand,
20428 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
20429 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
20430 andi16_operand): New predicates.
20431 * config/mips/mips.md (compression): New attribute.
20432 (enabled): New attribute.
20433 (length): Consider compression in computing length.
20434 (shift_compression): New code attribute.
20435 (*add<mode>3): New operands. Record compression.
20436 (sub<mode>3): Likewise.
20437 (one_cmpl<mode>2): Likewise.
20438 (*and<mode>3): Likewise.
20439 (*ior<mode>3): Likewise.
20440 (unnamed pattern for xor): Likewise.
20441 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
20442 (*<optab><mode>3): Likewise.
20443 (*mov<mode>_internal: Likewise.
20444 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
20445 (mips_unsigned_immediate_p): New.
20446 (umips_lwsp_swsp_address_p): New.
20447 (m16_based_address_p): New.
20448 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
20449 (mips_unsigned_immediate_p): New prototype.
20450 (lwsp_swsp_address_p): New prototype.
20451 (m16_based_address_p): New prototype.
20452 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
20453 (mips_signed_immediate_p): New function.
20454 (m16_based_address_p): New function.
20455 (lwsp_swsp_address_p): New function.
20456 (mips_print_operand_punctuation): Recognize short delay slot insns
20457 for microMIPS.add<mode>3"
20459 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20462 * config/arm/iterators.md (v_cmp_result): New mode attribute.
20463 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
20465 2013-03-25 Richard Biener <rguenther@suse.de>
20467 PR tree-optimization/56689
20468 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
20471 2013-03-25 Richard Biener <rguenther@suse.de>
20473 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
20475 (memory_references): Likewise.
20476 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
20477 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
20478 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
20479 (gather_mem_refs_in_loops): Fold into ...
20480 (analyze_memory_references): ... this. Move initialization
20481 to tree_ssa_lim_initialize.
20482 (fill_always_executed_in): Rename to ...
20483 (fill_always_executed_in_1): ... this.
20484 (fill_always_executed_in): Move contains_call computation to
20485 this new function from ...
20486 (tree_ssa_lim_initialize): ... here.
20487 (tree_ssa_lim): Call fill_always_executed_in.
20489 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
20491 * postreload.c (reload_combine): Fix code detecting returns.
20493 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
20495 * function.c (emit_use_return_register_into_block): On cc0 targets,
20496 do not emit the sequence between cc0 setter and user.
20498 2013-03-25 Kai Tietz <ktietz@redhat.com>
20500 * config/i386/predicates.md (local_symbolic_operand): Interpret
20501 dll-imported symbols as none-local.
20503 2013-03-25 Richard Biener <rguenther@suse.de>
20505 * tree-ssa-loop-im.c (struct depend): Remove.
20506 (struct lim_aux_data): Make depends a vec of gimples.
20507 (free_lim_aux_data): Adjust.
20508 (add_dependency): Likewise.
20509 (set_level): Likewise.
20511 2013-03-25 Richard Biener <rguenther@suse.de>
20513 PR middle-end/56434
20514 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
20515 the pointer returned by calls with ECF_MALLOC set.
20517 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
20519 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
20521 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
20523 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
20524 using MMXMODE mode iterator.
20525 (*move<mode>_internal): Merge with *movv2sf_internal and
20526 *movv2sf_internal_rex64 using MMXMODE mode iterator.
20528 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
20530 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
20531 (record_last_mem_set_info): Likewise.
20533 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
20534 of XNEWVEC followed by memset.
20535 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
20537 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
20539 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
20540 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
20541 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
20542 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
20543 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
20544 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
20545 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
20546 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
20547 BARRIER_P instead of GET_CODE.
20549 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20551 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
20552 inaccuracy in the probing code.
20554 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
20555 (ctrapdi4): Likewise.
20557 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20559 * calls.c (expand_call): Add missing guard to code handling return
20560 of non-BLKmode structures in MSB.
20561 * function.c (expand_function_end): Likewise.
20563 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20565 * combine.c (try_combine): Adjust comment. Do not add the set of
20566 insn #0 if the destination indirectly is set or dies in insn #2.
20567 Tidy up code to distribute a new note.
20569 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
20571 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
20572 also for alternatives 16 and 17.
20574 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
20576 * config/i386/sse.md (*mov<mode>_internal): Merge with
20577 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
20578 Emit insn template depending on type attribute. Use
20579 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
20580 movd instead of movq mnemonic for interunit moves. Rewrite mode
20581 attribute calculation. Remove unit attribute calculation.
20582 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
20583 Set prefix_data16 attribute for DImode ssemov types.
20584 Use Ym instead of y for SSE-MMX conversion alternatives.
20585 Reorder operand constraints.
20587 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
20589 * df.h (df_insn_delete): Adjust prototype.
20590 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
20591 and let it decide whether mark the basic block dirty.
20592 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
20593 * df-scan.c (df_insn_info_delete): New helper function, split
20594 off from df_insn_delete.
20595 (df_scan_free_bb_info): Use it.
20596 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
20598 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
20599 that the insn is actually an insn and it has a non-NULL basic block.
20600 Do not mark basic block dirty if only deleting a DEBUG_INSN.
20602 2013-03-22 Richard Biener <rguenther@suse.de>
20604 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
20606 (mem_ref_alloc): Do not allocate them.
20607 (refs_independent_p): Do not query or maintain a cache.
20609 2013-03-22 Richard Biener <rguenther@suse.de>
20611 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
20612 (gather_mem_refs_in_loops): Do not compute it.
20613 (analyze_memory_references): Do not allocate it.
20614 (tree_ssa_lim_finalize): Do not free it.
20615 (for_all_locs_in_loop): Do not query all_refs_in_loop.
20617 2013-03-22 Richard Biener <rguenther@suse.de>
20619 * is-a.h (as_a): Use gcc_checking_assert.
20621 2013-03-22 Ian Bolton <ian.bolton@arm.com>
20623 * config/aarch64/aarch64.c (aarch64_print_operand): New
20624 format specifier for printing a constant in hex.
20625 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
20626 format specifier for printing second operand.
20628 2013-03-22 Richard Biener <rguenther@suse.de>
20630 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
20632 (gather_mem_refs_in_loops): Perform store accumulation here.
20633 (create_vop_ref_mapping_loop): Remove.
20634 (create_vop_ref_mapping): Likewise.
20635 (analyze_memory_references): Initialize refs_stored_in_loop.
20636 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
20637 (record_indep_loop): Remove.
20638 (record_dep_loop): New function.
20639 (ref_indep_loop_p_1): Adjust to only walk over references
20640 in the loop, not its subloops.
20641 (ref_indep_loop_p): Rename to ...
20642 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
20643 maintaining a more fine-grained cache.
20644 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
20645 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
20647 2013-03-22 Richard Biener <rguenther@suse.de>
20649 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
20650 (struct mem_ref): Make accesses_in_loop a vec of a vec of
20651 aggregate mem_ref_loc.
20652 (free_mem_ref_locs): Inline into ...
20653 (memref_free): ... this and adjust.
20654 (mem_ref_alloc): Adjust.
20655 (mem_ref_locs_alloc): Remove.
20656 (record_mem_ref_loc): Adjust.
20657 (get_all_locs_in_loop): Rewrite into ...
20658 (for_all_locs_in_loop): ... this iterator.
20659 (rewrite_mem_ref_loc): New functor.
20660 (rewrite_mem_refs): Use for_all_locs_in_loop.
20661 (sm_set_flag_if_changed): New functor.
20662 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
20663 (ref_always_accessed): New functor.
20664 (ref_always_accessed_p): Use for_all_locs_in_loop.
20666 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20668 * tree-pass.h (PROP_gimple_lvec): New.
20669 * passes.c (dump_properties): Handle PROP_gimple_lvec.
20670 (init_optimization_passes): Move pass_lower_vector.
20671 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
20673 (pass_lower_vector): Provide PROP_gimple_lvec.
20674 (pass_lower_vector_ssa): Likewise.
20675 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
20677 2013-03-21 Mark Wielaard <mjw@redhat.com>
20679 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
20681 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
20683 * config/i386/i386.md (*movdi_internal): Disparage slightly
20684 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
20685 conversion alternatives.
20687 2013-03-21 Jakub Jelinek <jakub@redhat.com>
20689 PR middle-end/48087
20690 * diagnostic.def (DK_WERROR): New kind.
20691 * diagnostic.h (werrorcount): Define.
20692 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
20693 promoted to DK_ERROR, increment DK_WERROR counter instead of
20695 * toplev.c (toplev_main): Call print_ignored_options even if
20696 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
20697 even if just werrorcount is non-zero.
20700 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
20702 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
20703 (string_cst_pool_decl): New function.
20704 (optimize_one_addr_into_implicit_ptr): New function.
20705 (resolve_addr_in_expr): Optimize DWARF location expression
20706 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
20707 which doesn't live in memory, but has DW_AT_location or
20708 DW_AT_const_value, or refers to a string literal, into
20709 DW_OP_GNU_implicit_pointer.
20710 (optimize_location_into_implicit_ptr): New function.
20711 (resolve_addr): If removing DW_AT_location of a variable because
20712 it was DW_OP_addr of address of the variable, but the variable doesn't
20713 live in memory, try to emit const value attribute for the initializer.
20715 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20717 * tree.h (VECTOR_TYPE_P): New macro.
20718 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
20719 TYPE_MODE): Use it.
20720 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
20721 VEC_COND_EXPR cannot be lvalues.
20722 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
20724 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20726 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
20727 Restrict the transformation to equal modes.
20729 2013-03-21 Richard Biener <rguenther@suse.de>
20731 PR tree-optimization/39326
20732 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
20733 (MEM_ANALYZABLE): Adjust.
20734 (record_mem_ref_loc): Move bitmap ops ...
20735 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
20736 unanalyzable refs, do not record locations for it.
20737 (analyze_memory_references): Allocate ref zero as shared
20739 (refs_independent_p): Do not test for unanalyzed mems here.
20740 (ref_indep_loop_p_1): Special-case disambiguation against
20741 the unanalyzed ref.
20742 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
20744 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
20746 * config/arm/arm-protos.h (tune_params): Add
20747 prefer_neon_for_64bits field.
20748 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
20749 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
20750 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
20751 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
20752 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
20753 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
20754 (arm_option_override): Handle -mneon-for-64bits new option.
20755 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
20756 (prefer_neon_for_64bits): Declare new variable.
20757 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
20758 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
20759 (arch_enabled): Handle new arch types. Remove support for onlya8
20761 (one_cmpldi2): Use new arch names.
20762 (zero_extend<mode>di2, extend<mode>di2): Ditto.
20763 * config/arm/arm.opt (mneon-for-64bits): Add option.
20764 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
20765 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
20766 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
20768 * doc/invoke.texi (-mneon-for-64bits): Document.
20770 2013-03-21 Richard Biener <rguenther@suse.de>
20772 PR tree-optimization/39326
20773 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
20774 (sort_bbs_in_loop_postorder_cmp): New function.
20775 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
20777 2013-03-21 Richard Biener <rguenther@suse.de>
20779 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
20780 (vect_insert_into_interleaving_chain): Likewise.
20781 (vect_drs_dependent_in_basic_block): Inline ...
20782 (vect_slp_analyze_data_ref_dependence): ... here. New function,
20784 (vect_analyze_data_ref_dependence): ... here. Simplify.
20785 (vect_check_interleaving): Simplify.
20786 (vect_analyze_data_ref_dependences): Likewise. Split out ...
20787 (vect_slp_analyze_data_ref_dependences): ... this new function.
20788 (dr_group_sort_cmp): New function.
20789 (vect_analyze_data_ref_accesses): Compute data-reference groups
20790 here instead of in vect_analyze_data_ref_dependence. Use
20791 a more efficient algorithm.
20792 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
20793 vect_slp_analyze_data_ref_dependences. Call
20794 vect_analyze_data_ref_accesses earlier.
20795 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
20796 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
20797 (vect_slp_analyze_data_ref_dependences): New prototype.
20799 2013-03-21 Richard Biener <rguenther@suse.de>
20801 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
20802 ref is stored in the loop.
20803 (find_refs_for_sm): Walk only over all stores.
20804 (store_motion_loop): Allocate from lim_bitmap_obstack.
20805 (store_motion): Likewise.
20807 2013-03-21 Richard Biener <rguenther@suse.de>
20809 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
20810 Update virtual SSA form.
20812 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20814 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
20815 * configure: Regenerate.
20816 * config.in: Regenerate.
20817 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
20818 if !HAVE_LD_EH_FRAME_CIEV3.
20820 2013-03-21 Richard Biener <rguenther@suse.de>
20822 * tree-cfg.c (verify_expr_no_block): New function.
20823 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
20824 nor DECL_VALUE_EXPR have locations with associated blocks.
20825 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
20826 (clear_unused_block_pointer): Remove code dealing with
20827 blocks in DECL_DEBUG_EXPR locations.
20829 2013-03-21 Richard Biener <rguenther@suse.de>
20831 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
20832 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
20833 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
20834 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
20835 instead of DECL_DEBUG_EXPR_IS_FROM.
20836 * gimplify.c (gimplify_modify_expr): Likewise.
20837 * tree-cfg.c (verify_expr_location_1): Likewise.
20838 * tree-complex.c (create_one_component_var): Likewise.
20839 * tree-sra.c (create_access_replacement): Likewise.
20840 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
20841 (clear_unused_block_pointer): Likewise.
20842 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
20843 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
20844 * var-tracking.c (var_debug_decl): Likewise.
20845 (track_expr_p): Likewise.
20846 * tree-inline.c (add_local_variables): Likewise. Set
20847 DECL_HAS_DEBUG_EXPR_P after copying it.
20848 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
20849 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
20851 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
20854 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
20855 * configure: Regenerate.
20856 * config.in: Regenerate.
20857 * config/i386/i386.md (*movdf_internal): Use
20858 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
20859 movd instead of movq mnemonic for interunit moves.
20860 (*movdi_internal): Ditto.
20862 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20864 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
20865 (abd<mode>_3): New pattern.
20866 (aba<mode>_3): New pattern.
20867 (fabd<mode>_3): New pattern.
20869 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20871 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
20872 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
20873 occurrence of REGISTER_PREFIX as its empty string.
20875 2013-03-20 Jeff Law <law@redhat.com>
20877 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
20878 addititional equivalences for equality comparisons between an SSA_NAME
20879 and a constant where the SSA_NAME was set from a widening conversion.
20881 2013-03-20 Walter Lee <walt@tilera.com>
20883 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
20885 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20887 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
20888 depending on type attribute.
20889 (*movti_internal): Ditto.
20890 (*movtf_internal): Ditto.
20891 (*movxf_internal): Ditto.
20892 (*movdf_internal): Ditto.
20893 (*movsf_internal): Ditto.
20895 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20897 * config/i386/i386.md (*movti_internal): Set prefix attribute to
20898 maybe_vex for sselog1 and ssemov types.
20899 (*movdi_internal): Reorder operand constraints.
20900 (*movsi_internal): Ditto. Set prefix attribute to
20901 maybe_vex for sselog1 and ssemov types.
20902 (*movtf_internal): Set prefix attribute to maybe_vex
20903 for sselog1 and ssemov types.
20904 (*movdf_internal): Ditto. Set prefix_data16 attribute for
20905 DImode ssemov types. Reorder operand constraints.
20906 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
20907 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
20908 attribute for SImode ssemov types. Reorder operand constraints.
20910 2013-03-20 Martin Jambor <mjambor@suse.cz>
20912 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
20913 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
20915 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
20917 * config/rs6000/predicates.md (indexed_address, update_address_mem
20918 update_indexed_address_mem): New predicates.
20919 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
20920 attribute for load/store instructions.
20921 * config/rs6000/dfp.md (movsd_store): Likewise.
20922 (movsd_load): Likewise.
20923 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
20924 (unnamed HI->DI extend define_insn): Likewise.
20925 (unnamed SI->DI extend define_insn): Likewise.
20926 (unnamed QI->SI extend define_insn): Likewise.
20927 (unnamed QI->HI extend define_insn): Likewise.
20928 (unnamed HI->SI extend define_insn): Likewise.
20929 (unnamed HI->SI extend define_insn): Likewise.
20930 (extendsfdf2_fpr): Likewise.
20931 (movsi_internal1): Likewise.
20932 (movsi_internal1_single): Likewise.
20933 (movhi_internal): Likewise.
20934 (movqi_internal): Likewise.
20935 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
20936 attribute for load/store instructions.
20937 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
20939 (mov<mode>_softfloat): Likewise.
20940 (mov<mode>_hardfloat32): Likewise.
20941 (mov<mode>_hardfloat64): Likewise.
20942 (mov<mode>_softfloat64): Likewise.
20943 (movdi_internal32): Likewise.
20944 (movdi_internal64): Likewise.
20945 (probe_stack_<mode>): Likewise.
20947 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
20949 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
20950 floating point, and decimal floating point to reload iterator.
20952 * config/rs6000/constraints.md (wl constraint): New constraints to
20953 return FLOAT_REGS if certain options are used to reduce the number
20954 of separate patterns that exist in the file.
20955 (wx constraint): Likewise.
20956 (wz constraint): Likewise.
20958 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20959 -mdebug=reg, print wg, wl, wx, and wz constraints.
20960 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20961 Initialize the reload functions for 64-bit binary/decimal floating
20963 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
20964 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
20965 create the buffer on the stack to overcome not having a 32-bit
20967 (rs6000_emit_move): Likewise.
20968 (rs6000_secondary_memory_needed_rtx): Likewise.
20969 (rs6000_alloc_sdmode_stack_slot): Likewise.
20970 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
20971 via xxlxor, just like DFmode 0.0.
20973 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
20974 define as 1 if we are running on a power7 or newer.
20975 (enum r6000_reg_class_enum): Add new constraints.
20977 * config/rs6000/dfp.md (movsd): Delete, combine with binary
20978 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
20979 with other moves by using conditional constraits (wg). Use LFIWZX
20980 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
20981 (movsd splitter): Likewise.
20982 (movsd_hardfloat): Likewise.
20983 (movsd_softfloat): Likewise.
20985 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
20986 binary and decimal floating point moves.
20987 (fmove_ok): New attributes to combine binary and decimal floating
20988 point moves, and to combine power6x (mfpgpr) moves along normal
20990 (real_value_to_target): Likewise.
20991 (f32_lr): Likewise.
20992 (f32_lm): Likewise.
20993 (f32_li): Likewise.
20994 (f32_sr): Likewise.
20995 (f32_sm): Likewise.
20996 (f32_si): Likewise.
20997 (movsf): Combine binary and decimal floating point moves. Combine
20998 power6x (mfpgpr) moves with other moves by using conditional
20999 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
21000 (mov<mode> for SFmode/SDmode); Likewise.
21001 (SFmode/SDmode splitters): Likewise.
21002 (movsf_hardfloat): Likewise.
21003 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
21004 (movsf_softfloat): Likewise.
21005 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
21007 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
21008 wx and wz constraints.
21010 * config/rs6000/constraints.md (wg constraint): New constraint to
21011 return FLOAT_REGS if -mmfpgpr (power6x) was used.
21013 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
21016 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
21017 -mdebug=reg, print wg, wl, wx, and wz constraints.
21018 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
21019 Initialize the reload functions for 64-bit binary/decimal floating
21021 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
21022 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
21023 create the buffer on the stack to overcome not having a 32-bit
21025 (rs6000_emit_move): Likewise.
21026 (rs6000_secondary_memory_needed_rtx): Likewise.
21027 (rs6000_alloc_sdmode_stack_slot): Likewise.
21028 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
21029 via xxlxor, just like DFmode 0.0.
21031 * config/rs6000/dfp.md (movdd): Delete, combine with binary
21032 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
21033 with other moves by using conditional constraits (wg). Use LFIWZX
21034 and STFIWX for loading SDmode on power7.
21035 (movdd splitters): Likewise.
21036 (movdd_hardfloat32): Likewise.
21037 (movdd_softfloat32): Likewise.
21038 (movdd_hardfloat64_mfpgpr): Likewise.
21039 (movdd_hardfloat64): Likewise.
21040 (movdd_softfloat64): Likewise.
21042 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
21043 64-bit binary and decimal floating point moves.
21044 (FMOVE64X): Likewise.
21045 (movdf): Combine 64-bit binary and decimal floating point moves.
21046 Combine power6x (mfpgpr) moves with other moves by using
21047 conditional constraits (wg).
21048 (mov<mode> for DFmode/DDmode): Likewise.
21049 (DFmode/DDmode splitters): Likewise.
21050 (movdf_hardfloat32): Likewise.
21051 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
21052 (movdf_softfloat32): Likewise.
21053 (movdf_hardfloat64_mfpgpr): Likewise.
21054 (movdf_hardfloat64): Likewise.
21055 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
21056 (movdf_softfloat64): Likewise.
21057 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
21058 (reload_<mode>_load): Move to later in the file so they aren't in
21059 the middle of the floating point move insns.
21060 (reload_<mode>_store): Likewise.
21062 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
21065 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
21066 constraint if -mdebug=reg.
21067 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
21068 Enable using dd reload support if needed.
21070 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
21071 binary and decimal floating point moves in rs6000.md.
21072 (movtd_internal): Likewise.
21074 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
21075 decimal floating point moves.
21077 (movtf_internal): Likewise.
21078 (mov<mode>_internal, TDmode/TFmode): Likewise.
21079 (movtf_softfloat): Likewise.
21080 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
21082 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
21083 movdi_internal64, using wg constraint for move direct operations.
21084 (movdi_internal64): Likewise.
21086 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
21087 MODES_TIEABLE_P for selected modes. Print the numerical value of
21088 the various virtual registers. Use GPR/FPR first/last values,
21089 instead of hard coding the register numbers. Print which modes
21090 have reload functions registered.
21091 (rs6000_option_override_internal): If -mdebug=reg, trace the options
21092 settings before/after setting cpu, target and subtarget settings.
21093 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
21094 and for secondary reload failures in rs6000_secondary_reload_inner.
21095 (rs6000_secondary_reload_fail): Likewise.
21096 (rs6000_secondary_reload_inner): Likewise.
21098 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
21099 macros for first/last GPR and FPR registers.
21100 (LAST_GPR_REGNO): Likewise.
21101 (FIRST_FPR_REGNO): Likewise.
21102 (LAST_FPR_REGNO): Likewise.
21104 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
21105 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
21106 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
21107 (vcond<mode><mode>): Likewise.
21108 (vcondu<mode><mode>): Likewise.
21109 (vector_gtu<mode>): Likewise.
21110 (vector_gte<mode>): Likewise.
21111 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
21112 to prevent the compiler from converting DImode operations to TImode.
21113 (ior<mode>3): Likewise.
21114 (and<mode>3): Likewise.
21115 (one_cmpl<mode>2): Likewise.
21116 (nor<mode>3): Likewise.
21117 (andc<mode>3): Likewise.
21119 * config/rs6000/constraints.md (wt constraint): New constraint
21120 that returns VSX_REGS if TImode is allowed in VSX registers.
21122 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
21123 constant under VSX.
21125 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
21126 similar to TImode, but it is restricted to being in the GPRs.
21128 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
21129 TImode to occupy a single VSX register.
21131 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
21132 -mvsx-timode for power7/power8.
21133 (power7 cpu): Likewise.
21134 (power8 cpu): Likewise.
21136 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
21137 sure that TFmode/TDmode take up two registers if they are ever
21138 allowed in the upper VSX registers.
21139 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
21141 (rs6000_init_hard_regno_mode_ok): Likewise.
21142 (rs6000_debug_reg_global): Add debugging for PTImode and wt
21143 constraint. Print if LRA is turned on.
21144 (rs6000_option_override_internal): Give an error if -mvsx-timode
21145 and VSX is not enabled.
21146 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
21147 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
21148 to reg+offset addressing. Use PTImode when checking offset
21149 addresses for validity.
21150 (reg_offset_addressing_ok_p): Likewise.
21151 (rs6000_legitimate_offset_address_p): Likewise.
21152 (rs6000_legitimize_address): Likewise.
21153 (rs6000_legitimize_reload_address): Likewise.
21154 (rs6000_legitimate_address_p): Likewise.
21155 (rs6000_eliminate_indexed_memrefs): Likewise.
21156 (rs6000_emit_move): Likewise.
21157 (rs6000_secondary_reload): Likewise.
21158 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
21159 reloads to fpr registers to continue to use reg+offset addressing,
21160 but 64-bit reloads to altivec registers need reg+reg addressing.
21161 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
21162 it. Treat LO_SUM like a PLUS operation.
21163 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
21164 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
21165 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
21166 registers to share a register with a smaller sized type, since VSX
21167 puts scalars in the upper 64-bits.
21168 (print_operand): Add support for PTImode.
21169 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
21170 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
21171 registers, but don't have arithmetic support.
21172 (rs6000_memory_move_cost): Add test for VSX.
21173 (rs6000_opt_masks): Add -mvsx-timode.
21175 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
21178 (VSr): Use wt constraint for TImode.
21179 (VSv): Drop TImode support.
21180 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
21181 (vsx_movti_64bit): Likewise.
21182 (vsx_movti_32bit): Likewise.
21183 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
21184 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
21185 one '?' on the appropriate output constraint. Do not allow TImode
21186 logical operations on 32-bit systems.
21187 (vsx_ior<mode>3): Likewise.
21188 (vsx_xor<mode>3): Likewise.
21189 (vsx_one_cmpl<mode>2): Likewise.
21190 (vsx_nor<mode>3): Likewise.
21191 (vsx_andc<mode>3): Likewise.
21192 (vsx_concat_<mode>): Likewise.
21193 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
21195 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
21196 OPTION_MASK_VSX_TIMODE.
21197 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
21198 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
21200 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
21201 (TI2 iterator): New iterator for TImode, PTImode.
21202 (wd mode attribute): Add values for vector types.
21203 (movti_string): Replace TI move operations with operations for TImode
21204 and PTImode. Add support for TImode being allowed in VSX registers.
21205 (mov<mode>_string, TImode/PTImode): Likewise.
21206 (movti_ppc64): Likewise.
21207 (mov<mode>_ppc64, TImode/PTImode): Likewise.
21208 (TI mode splitters): Likewise.
21210 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
21213 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
21215 PR tree-optimization/56355
21216 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
21217 Also handle integers with undefined overflow.
21219 2013-03-20 Catherine Moore <clm@codesourcery.com>
21220 Maciej W. Rozycki <macro@codesourcery.com>
21221 Tom de Vries <tom@codesourcery.com>
21222 Nathan Sidwell <nathan@codesourcery.com>
21223 Iain Sandoe <iain@codesourcery.com>
21224 Nathan Froyd <froydnj@codesourcery.com>
21225 Chao-ying Fu <fu@mips.com>
21227 * doc/extend.texi: (micromips, nomicromips, nocompression):
21228 Document new function attributes.
21229 * doc/invoke.texi (minterlink-compressed, mmicromips,
21230 m14k, m14ke, m14kec): Document new options.
21231 (minterlink-mips16): Update documentation.
21232 * doc/md.texi (ZC, ZD): Document new constraints.
21233 * configure.ac (gcc_cv_as_micromips): Check if linker
21234 supports the .set micromips directive.
21235 * configure: Regenerate.
21236 * config.in: Regenerate.
21237 * config/mips/mips-tables.opt: Regenerate.
21238 * config/mips/micromips.md: New file.
21239 * constraints.md (ZC, ZD): New constraints.
21240 * config/mips/predicates.md (movep_src_register): New predicate.
21241 (movep_src_operand): New predicate.
21242 (non_volatile_mem_operand): New predicate.
21243 * config/mips/mips.md (multimem): New type.
21244 (length): Differentiate between 17-bit and 18-bit branch offsets.
21245 (MOVEP1, MOVEP2): New mode iterator.
21246 (mov_<load>l): Use ZC constraint.
21247 (mov_<load>r): Likewise.
21248 (mov_<store>l): Likewise.
21249 (mov_<store>r): Likewise.
21250 (*branch_equality<mode>_inverted): Add microMIPS support.
21251 (*branch_equality<mode>): Likewise.
21252 (*jump_absolute): Likewise.
21253 (indirect_jump_<mode>): Likewise.
21254 (tablejump_<mode>): Likewise.
21255 (<optab>_internal): Likewise.
21256 (sibcall_internal): Likewise.
21257 (sibcall_value_internal): Likewise.
21258 (prefetch): Use constraint ZD.
21259 * config/mips/mips.opt (minterlink-compressed): New option.
21260 (minterlink-mips16): Now an alias for minterlink-compressed.
21261 (mmicromips): New option.
21262 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21263 (compare_and_swap_12): Likewise.
21264 (sync_add<mode>): Likewise.
21265 (sync_<optab>_12): Likewise.
21266 (sync_old_<optab>_12): Likewise.
21267 (sync_new_<optab>_12): Likewise.
21268 (sync_nand_12): Likewise.
21269 (sync_old_nand_12): Likewise.
21270 (sync_new_nand_12): Likewise.
21271 (sync_sub<mode>): Likewise.
21272 (sync_old_add<mode>): Likewise.
21273 (sync_old_sub<mode>): Likewise.
21274 (sync_new_add<mode>): Likewise.
21275 (sync_new_sub<mode>): Likewise.
21276 (sync_<optab><mode>): Likewise.
21277 (sync_old_<optab><mode>): Likewise.
21278 (sync_new_<optab><mode>): Likewise.
21279 (sync_nand<mode>): Likewise.
21280 (sync_old_nand<mode>): Likewise.
21281 (sync_new_nand<mode>): Likewise.
21282 (sync_lock_test_and_set<mode>): Likewise.
21283 (test_and_set_12): Likewise.
21284 (atomic_compare_and_swap<mode>): Likewise.
21285 (atomic_exchange<mode>_llsc): Likewise.
21286 (atomic_fetch_add<mode>_llsc): Likewise.
21287 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21288 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21289 (umips_save_restore_pattern_p): Likewise.
21290 (umips_load_store_pair_p): Likewise.
21291 (umips_output_load_store_pair): Likewise.
21292 (umips_movep_target_p): Likewise.
21293 (umips_12bit_offset_address_p): Likewise.
21294 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21295 (mips_base_mips16): Rename this...
21296 (mips_base_compression_flags): ...to this. Update all uses.
21297 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21298 (mips_mips16_decl_p): Delete.
21299 (mips_nomips16_decl_p): Delete.
21300 (mips_get_compress_on_flags): New function.
21301 (mips_get_compress_off_flags): New function.
21302 (mips_get_compress_mode): New function.
21303 (mips_get_compress_on_name): New function.
21304 (mips_get_compress_off_name): New function.
21305 (mips_insert_attributes): Support multiple compression types.
21306 (mips_merge_decl_attributes): Likewise.
21307 (umips_12bit_offset_address_p): New function.
21308 (mips_start_function_definition): Emit .set micromips directive.
21309 (mips_call_may_need_jalx_p): New function.
21310 (mips_function_ok_for_sibcall): Add microMIPS support.
21311 (mips_print_operand_punctuation): Support short delay slots and
21313 (umips_swm_mask, umips_swm_encoding): New.
21314 (umips_build_save_restore): New function.
21315 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21316 (was_mips16_p): Remove.
21317 (old_compression_mode): New.
21318 (mips_set_compression_mode): New function.
21319 (mips_set_current_function): Add microMIPS support.
21320 (mips_option_override): Likewise.
21321 (umips_save_restore_pattern_p): New function.
21322 (umips_output_save_restore): New function.
21323 (umips_load_store_pair_p_1): New function.
21324 (umips_load_store_pair_p): New function.
21325 (umips_output_load_store_pair_1): New function.
21326 (umips_output_load_store_pair): New function.
21327 (umips_movep_target_p) New function.
21328 (mips_prepare_pch_save): Add microMIPS support.
21329 * config/mips/mips.h (TARGET_COMPRESSION): New.
21330 (TARGET_CPU_CPP_BUILTINS): Update macro
21331 to use new compression flags and to support microMIPS.
21332 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21333 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21334 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21335 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21336 (ASM_SPEC): Support mmicromips and mno-micromips.
21337 (M16STORE_REG_P): New macro.
21338 (MIPS_CALL): Support TARGET_MICROMIPS.
21339 (MICROMIPS_J): New macro.
21340 (mips_base_mips16): Rename this...
21341 (mips_base_compression_flags): ...to this.
21342 (UMIPS_12BIT_OFFSET_P): New macro.
21343 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21344 (MULTILIB_DIRNAMES): Likewise.
21345 2013-03-20 Richard Biener <rguenther@suse.de>
21347 PR tree-optimization/56661
21348 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
21349 the result does not have to be distinct.
21351 2013-03-20 Richard Biener <rguenther@suse.de>
21353 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
21356 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21357 Steven Bosscher <steven@gcc.gnu.org>
21359 PR rtl-optimization/56605
21360 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
21362 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
21365 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
21366 that require movd instead of movq.
21368 2013-03-20 Richard Biener <rguenther@suse.de>
21370 * tree-ssa-structalias.c (struct variable_info): Add pointer
21371 to the first field of an aggregate with sub-vars. Make
21372 this and the pointer to the next subfield its ID.
21373 (vi_next): New function.
21374 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
21375 storedanything_id, integer_id): Increment by one.
21376 (new_var_info, get_call_vi, lookup_call_clobber_vi,
21377 get_call_clobber_vi): Adjust.
21378 (solution_set_expand): Simplify and speedup.
21379 (solution_set_add): Inline into ...
21380 (set_union_with_increment): ... this. Adjust accordingly.
21381 (do_sd_constraint): Likewise.
21382 (do_ds_constraint): Likewise.
21383 (do_complex_constraint): Simplify.
21384 (build_pred_graph): Adjust.
21385 (solve_graph): Likewise. Simplify and speedup.
21386 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
21387 get_constraint_for_component_ref, get_constraint_for_1,
21388 first_vi_for_offset, first_or_preceding_vi_for_offset,
21389 create_function_info_for, create_variable_info_for_1,
21390 create_variable_info_for, intra_create_variable_infos): Adjust.
21391 (init_base_vars): Push NULL for ID zero.
21392 (compute_points_to_sets): Adjust.
21394 2013-03-20 Richard Biener <rguenther@suse.de>
21396 * cfgloop.c (verify_loop_structure): Streamline and avoid
21397 ICEing on corrupt loop tree.
21398 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
21401 2013-03-20 Richard Biener <rguenther@suse.de>
21403 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
21404 check whether an SSA update is needed.
21406 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
21408 * config/mips/constraints.md (T): Rename to...
21412 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
21413 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
21415 2013-03-19 Ian Bolton <ian.bolton@arm.com>
21417 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
21418 (*subsi3_carryin_uxtw): Likewise.
21420 2013-03-19 Ian Bolton <ian.bolton@arm.com>
21422 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
21423 (*rorsi3_insn_uxtw): Likewise.
21425 2013-03-19 Ian Bolton <ian.bolton@arm.com>
21427 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
21428 (*extrsi5_insn_uxtw): Likewise.
21430 2013-03-19 Richard Biener <rguenther@suse.de>
21432 PR tree-optimization/56273
21433 * passes.c (init_optimization_passes): Move second VRP after DOM.
21435 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
21437 * config/i386/i386.md (*movti_internal): Merge from
21438 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
21439 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
21440 nox64 isa attributes.
21442 2013-03-18 Richard Biener <rguenther@suse.de>
21444 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
21446 (merge_node_constraints): Likewise.
21447 (build_succ_graph): Likewise.
21448 (valid_graph_edge): Inline into single caller.
21449 (unify_nodes): Likewise. Use bitmap_set_bit return value
21451 (scc_visit): Fix formatting and variable use.
21452 (do_sd_constraint): Use gcc_checking_assert.
21453 (do_ds_constraint): Likewise.
21454 (do_complex_constraint): Likewise.
21455 (condense_visit): Likewise. Cleanup.
21456 (dump_pred_graph): New function.
21457 (perform_var_substitution): Dump the pred-graph before
21458 variable substitution.
21459 (find_equivalent_node): Use gcc_checking_assert.
21460 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
21462 2013-03-18 Richard Biener <rguenther@suse.de>
21464 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
21465 Remove cond_expr_stmt_list argument and do not gimplify the
21467 (vect_loop_versioning): Adjust.
21468 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
21469 Cleanup to use less temporaries.
21470 (vect_create_data_ref_ptr): Cleanup.
21472 2013-03-18 Jakub Jelinek <jakub@redhat.com>
21474 PR tree-optimization/56635
21475 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
21476 require types_compatible_p types.
21478 2013-03-18 Nick Clifton <nickc@redhat.com>
21480 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
21481 spurious backslash.
21483 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
21484 Add missing line to comment describing function.
21486 2013-03-18 Richard Biener <rguenther@suse.de>
21488 PR tree-optimization/56210
21489 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21490 Handle string / character search functions.
21491 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
21493 2013-03-18 Richard Biener <rguenther@suse.de>
21495 PR middle-end/56483
21496 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
21497 and implement properly.
21498 * gimple.h (gimple_cond_single_var_p): Remove.
21500 2013-03-18 Richard Biener <rguenther@suse.de>
21502 * tree-data-ref.h (find_data_references_in_loop): Declare.
21503 * tree-data-ref.c (get_references_in_stmt): Use a stack
21504 vector pre-allocated in the callers.
21505 (find_data_references_in_stmt): Adjust.
21506 (graphite_find_data_references_in_stmt): Likewise.
21507 (create_rdg_vertices): Likewise.
21508 (find_data_references_in_loop): Export.
21509 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
21510 Compute dependences here...
21511 (vect_analyze_data_refs): ...not here. When we encounter
21512 a non-vectorizable data reference in basic-block vectorization
21513 truncate the data reference vector. Do not bother to
21514 fixup data-dependence information for gather loads.
21515 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
21516 of data references, as reported.
21518 2013-03-18 Richard Biener <rguenther@suse.de>
21520 PR tree-optimization/3713
21521 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
21522 has_constants and expr.
21523 (stmt_has_constants): Properly valueize SSA names when deciding
21524 whether the stmt has constants.
21526 2013-03-18 Richard Biener <rguenther@suse.de>
21528 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
21529 whole function when there is nothing to do.
21530 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
21531 * tree-vectorizer.c (vectorize_loops): Update virtual and
21532 loop-closed SSA once.
21533 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
21535 2013-03-18 Richard Biener <rguenther@suse.de>
21537 PR middle-end/56113
21538 * domwalk.c (bb_postorder): New global static.
21539 (cmp_bb_postorder): New function.
21540 (walk_dominator_tree): Replace scheme imposing an order for
21541 visiting dominator sons by one sorting them at the time they
21542 are pushed on the stack.
21544 2013-03-18 Richard Biener <rguenther@suse.de>
21546 PR tree-optimization/39326
21547 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
21548 (struct mem_ref): Replace mem member with ao_ref typed member.
21549 (MEM_ANALYZABLE): Adjust.
21550 (memref_eq): Likewise.
21551 (mem_ref_alloc): Likewise.
21552 (gather_mem_refs_stmt): Likewise.
21553 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
21554 (execute_sm_if_changed_flag_set): Adjust.
21555 (execute_sm): Likewise.
21556 (ref_always_accessed_p): Likewise.
21557 (refs_independent_p): Likewise.
21558 (can_sm_ref_p): Likewise.
21560 2013-03-18 Jakub Jelinek <jakub@redhat.com>
21563 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
21564 return 1 even for !unsignedp.
21566 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21568 * config/i386/i386.md (isa): Add x64 and nox64.
21569 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
21570 (*pushtf): Enable *roF alternative for x64 isa only.
21571 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
21572 mode attribute of integer alternatives to DImode for TARGET_64BIT.
21573 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
21574 (*movtf_internal): Merge from *movtf_internal_rex64 and
21575 *movtf_internal_sse. Use x64 and nox64 isa attributes.
21576 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
21577 nox64 isa attributes.
21578 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
21579 nox64 isa attributes.
21580 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
21582 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21584 * config/alpha/alpha.c (TARGET_LRA_P): New define.
21586 2013-03-17 Jakub Jelinek <jakub@redhat.com>
21589 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
21590 class names. Remove trailing comma after "ALL_REGS".
21592 2013-03-16 Jan Hubicka <jh@suse.cz>
21594 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
21595 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
21596 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
21597 of cgraph_get_create_node.
21598 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
21600 2013-03-16 Jason Merrill <jason@redhat.com>
21603 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
21604 with DW_AT_default_value.
21606 2013-03-16 Jakub Jelinek <jakub@redhat.com>
21608 * BASE-VER: Set to 4.9.0.
21610 2013-03-14 Andi Kleen <ak@linux.intel.com>
21613 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
21614 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
21615 Document _x* TSX intrinsics.
21617 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
21618 David Holsgrove <david.holsgrove@xilinx.com>
21620 * configure.ac: Add MicroBlaze TLS support detection.
21621 * configure: Regenerate.
21622 * config/microblaze/microblaze-protos.h
21623 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
21624 symbol_mentioned_p, label_mentioned_p): Add prototypes.
21625 * config/microblaze/microblaze.c (microblaze_address_type): Add
21626 ADDRESS_TLS and tls_reloc address types.
21627 (microblaze_address_info): Add tls_reloc.
21628 (TARGET_HAVE_TLS): Define.
21629 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
21630 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
21631 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
21632 load_tls_operand, microblaze_call_tls_get_addr,
21633 microblaze_legitimize_tls_address): New functions.
21634 (microblaze_classify_unspec): Handle UNSPEC_TLS.
21635 (get_base_reg): Use microblaze_tls_symbol_p.
21636 (microblaze_classify_address): Handle TLS.
21637 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
21638 label_mentioned_p and microblaze_tls_referenced_p.
21639 (microblaze_legitimize_address): Handle TLS.
21640 (microblaze_address_insns): Handle ADDRESS_TLS.
21641 (pic_address_needs_scratch): Handle TLS.
21642 (print_operand_address): Handle TLS.
21643 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
21644 (microblaze_expand_move): Handle TLS.
21645 (microblaze_legitimate_constant_p): Check
21646 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
21647 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
21648 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
21649 (PIC_OFFSET_TABLE_REGNUM): Set.
21650 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
21651 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
21652 (addsi3, movsi_internal2, movdf_internal): Update constraints
21653 * config/microblaze/predicates.md (arith_plus_operand): Define
21654 (move_operand): Redefine as move_src_operand,
21655 check microblaze_tls_referenced_p.
21657 2013-03-14 Ian Bolton <ian.bolton@arm.com>
21659 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
21660 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
21662 2013-03-14 Ian Bolton <ian.bolton@arm.com>
21664 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
21667 2013-03-14 Jakub Jelinek <jakub@redhat.com>
21669 PR tree-optimization/53265
21670 * common.opt (Waggressive-loop-optimizations): New option.
21671 * tree-ssa-loop-niter.c: Include tree-pass.h.
21672 (do_warn_aggressive_loop_optimizations): New function.
21673 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
21674 if number_of_latch_executions returned constant.
21675 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
21676 early. If number_of_latch_executions returned constant, set
21677 nb_iterations_upper_bound back to it.
21678 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
21680 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
21681 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
21683 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
21684 (MULTILIB_OSDIRNAMES): Set.
21685 * genmultilib: If defaultosdirname doesn't start with :: , set
21686 defaultosdirname2 instead, clear it and emit two . multilib_raw
21687 entries instead of just one.
21689 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
21691 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
21692 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
21693 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
21694 (SUBTARGET_OVERRIDE_OPTIONS): New.
21696 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
21699 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
21700 (musermode): Convert to Var(TARGET_USERMODE).
21701 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
21702 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
21703 * config/sh/sh.c (sh_option_override): Use
21704 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
21705 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
21707 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
21709 (udivsi3_i4_single, divsi3_i4_single): Use
21710 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
21712 2013-03-13 Dave Korn <dave.korn.cygwin@....>
21714 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
21717 2013-03-13 Richard Biener <rguenther@suse.de>
21719 PR tree-optimization/56608
21720 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
21721 calls when vectorizing basic-blocks.
21723 2013-03-13 Jakub Jelinek <jakub@redhat.com>
21726 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
21729 2013-03-12 Jakub Jelinek <jakub@redhat.com>
21731 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
21733 2013-03-11 Jan Hubicka <jh@suse.cz>
21736 * lto-streamer-out.c (output_symbol_p): Skip references from
21737 constructors of external variables.
21739 2013-03-11 Jan Hubicka <jh@suse.cz>
21741 PR middle-end/56571
21742 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
21744 * emit-rtl.c (verify_rtx_sharing): Likewise.
21745 (copy_insn_1): Likewise.
21746 * rtl.c (copy_rtx): Likewise.
21748 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
21751 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
21752 output_operand_lossage message.
21754 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
21757 * arm.c (shift_op): Validate RTL pattern on the fly.
21758 (arm_print_operand, case 'S'): Don't use shift_operator to validate
21761 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21764 * config/pa/pa.md (call_value): Check for calls to powf and direct to
21765 new call patterns that clobber %fr12.
21766 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
21767 split and postreload patterns.
21768 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
21769 registers %fr12 and %fr12R as call used.
21771 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
21773 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
21774 (canon_address, record_store, replace_read, check_mem_read_rtx,
21775 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
21776 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
21777 rest_of_handle_dse): Likewise.
21779 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
21781 PR middle-end/56524
21782 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
21784 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
21785 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
21786 (save_optabs_if_changed): Replace with...
21787 (init_tree_optimization_optabs): ...this.
21788 * optabs.c (save_optabs_if_changed): Rename to...
21789 (init_tree_optimization_optabs): ...this. Take the optimization node
21790 as argument. Do nothing if the base optabs are already correct.
21791 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
21792 to recompute optabs.
21793 * function.h (function): Remove optabs field.
21794 * function.c (invoke_set_current_function_hook): Call
21795 init_tree_optimization_optabs. Use the result to initialize
21798 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
21800 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
21801 if GTMA_HAS_NO_INSTRUMENTATION.
21802 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
21803 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
21804 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
21805 * gimple-pretty-print.c (dump_gimple_transaction): Handle
21806 GTMA_HAS_NO_INSTRUMENTATION.
21808 2013-03-08 Jakub Jelinek <jakub@redhat.com>
21810 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
21813 2013-03-08 Marek Polacek <polacek@redhat.com>
21814 Jakub Jelinek <jakub@redhat.com>
21816 PR tree-optimization/56478
21817 * predict.c (is_comparison_with_loop_invariant_p): Change the
21818 type of loop_step to tree.
21819 (predict_loops): Adjust.
21820 (predict_iv_comparison): Perform the computations on double_ints.
21822 2013-03-08 Richard Biener <rguenther@suse.de>
21824 PR tree-optimization/56570
21825 * tree-cfg.c (verify_expr_location_1): Verify locations for
21827 * tree-sra.c (create_access_replacement): Strip locations
21828 from DECL_DEBUG_EXPRs.
21830 2013-03-08 Richard Biener <rguenther@suse.de>
21832 * tree-inline.c (expand_call_inline): Do not associate
21833 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
21834 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
21836 2013-03-08 Richard Biener <rguenther@suse.de>
21838 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
21839 or block changes with -Og. Fix for location / block encoding
21840 changes and PHI arguments with locations.
21842 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
21844 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
21846 (struct output_info): Likewise.
21847 (register_overhead): Remove bad gcc_assert.
21848 (bitmap_find_bit): If there is only a single bitmap element, do not
21849 count a miss as a search.
21850 (print_statistics): Update for counter type changes.
21851 (dump_bitmap_statistics): Likewise. Print headers such that they
21852 are properly lined up with the printed counters.
21854 2013-03-07 Jakub Jelinek <jakub@redhat.com>
21856 PR tree-optimization/56559
21857 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
21858 check that it has only a single use.
21860 2013-03-07 Richard Biener <rguenther@suse.de>
21862 * doc/invoke.texi (fwhole-program): Discourage use in combination
21865 2013-03-06 Jakub Jelinek <jakub@redhat.com>
21867 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
21869 PR tree-optimization/56539
21870 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
21871 instead of GSI_CONTINUE_LINKING as last argument to
21872 force_gimple_operand_gsi. Adjust function comment.
21874 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
21877 PR middle-end/56548
21878 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
21879 promoted mode, convert the result back to the original mode.
21881 2013-03-06 Richard Biener <rguenther@suse.de>
21883 PR middle-end/56294
21884 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
21885 (insert_updated_phi_nodes_compare_uids): New function.
21886 (update_ssa): Sort symbols_to_rename after UID before
21887 traversing it to insert PHI nodes.
21889 2013-03-06 Richard Biener <rguenther@suse.de>
21891 PR middle-end/50494
21892 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
21893 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
21896 2013-02-13 Richard Biener <rguenther@suse.de>
21899 * varasm.c (output_constant_def_1): Get the decl representing
21900 the constant as argument.
21901 (output_constant_def): Wrap output_constant_def_1.
21902 (make_decl_rtl): Use output_constant_def_1 with the decl
21903 representing the constant.
21904 (build_constant_desc): Optionally re-use a decl already
21905 representing the constant.
21906 (tree_output_constant_def): Adjust.
21908 2013-03-06 Joey Ye <joey.ye@arm.com>
21911 * gcc.c (convert_white_space): New function.
21912 (main): Handles white space in function name.
21914 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
21917 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
21918 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
21919 to SH_DIV_CALL_TABLE for TARGET_SH2.
21920 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
21922 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
21923 call-table options.
21925 2013-03-05 Sterling Augustine <saugustine@google.com>
21926 Cary Coutant <ccoutant@google.com>
21929 * dwarf2out.c (resolve_addr): Don't call
21930 remove_loc_list_addr_table_entries a second time for the same
21933 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21936 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
21937 (avoid_complex_debug_insns): New function.
21938 (expand_debug_locations): Call it.
21940 PR rtl-optimization/56484
21941 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
21942 lifetimes of hard registers on small register class machines.
21944 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21946 * config/microblaze/microblaze-protos.h: Rename
21947 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
21948 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
21950 (microblaze_fast_interrupt_function_p): New function.
21951 (microblaze_is_interrupt_handler): Rename to
21952 microblaze_is_interrupt_variant and add fast_interrupt check.
21953 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
21954 (save_restore_insns): Likewise.
21955 (compute_frame_size): Likewise.
21956 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
21957 (microblaze_globalize_label): Likewise.
21958 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
21959 * config/microblaze/microblaze.md: Use wrapper
21960 microblaze_is_interrupt_variant.
21962 2013-03-05 Kai Tietz <ktietz@redhat.com>
21964 * sdbout.c (sdbout_one_type): Switch to current function's section
21965 supporting cold/hot.
21967 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21969 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
21972 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21974 PR middle-end/56461
21975 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
21976 if VALGRIND_GET_VBITS is defined, temporarily make object
21977 memory all defined, and restore previous valgrind addressability
21978 and definability afterwards. Free this_object at the end.
21980 PR middle-end/56461
21981 * lra.c (lra): Call lra_clear_live_ranges if live_p,
21982 right before calling lra_create_live_ranges, also call it
21983 when clearing live_p. Only call lra_clear_live_ranges
21984 at the end if live_p.
21986 PR middle-end/56461
21987 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
21989 2013-03-05 Richard Biener <rguenther@suse.de>
21991 PR tree-optimization/56521
21992 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
21995 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
21998 * except.h (remove_unreachable_eh_regions): New prototype.
21999 * except.c (remove_eh_handler_splicer): New function, split out
22000 of remove_eh_handler.
22001 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
22002 warning about running it on many EH regions one at a time.
22003 (remove_unreachable_eh_regions_worker): New function, walk the
22004 EH tree in depth-first order and remove non-marked regions.
22005 (remove_unreachable_eh_regions): New function.
22006 * tree-eh.c (mark_reachable_handlers): New function, split out
22007 from remove_unreachable_handlers.
22008 (remove_unreachable_handlers): Use mark_reachable_handlers and
22009 remove_unreachable_eh_regions.
22010 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
22011 and remove_unreachable_eh_regions.
22013 2013-03-05 Richard Biener <rguenther@suse.de>
22015 PR middle-end/56525
22016 * loop-init.c (fix_loop_structure): Remove loops in two stages,
22017 not freeing them until the end.
22019 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22021 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
22023 2013-03-05 Richard Biener <rguenther@suse.de>
22025 PR tree-optimization/56270
22026 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
22027 of loads after scheduling an SLP instance.
22029 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22031 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
22033 (check_gcc_parallelize): Run guality.exp as a separate job from
22034 vect.exp with unsorted.exp and $(dg_target_exps) separately from
22035 struct-layout-1.exp with stackalign.exp.
22037 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
22039 PR middle-end/56461
22040 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
22041 load_index sbitmap even if some bit in it isn't set.
22043 PR middle-end/56461
22044 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
22045 (discover_iteration_bound_by_body_walk): Change queues to
22046 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
22047 spelling in comment. Call safe_push on queues[bound_index] directly.
22048 Release queues[queue_index] in every iteration unconditionally.
22049 Release bounds vector.
22051 PR middle-end/56461
22052 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
22053 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
22054 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
22057 2013-03-05 Richard Biener <rguenther@suse.de>
22060 * tree-inline.c (remap_blocks_to_null): New function.
22061 (expand_call_inline): When expanding a call stmt without
22062 an associated block inline remap all callee blocks to NULL.
22064 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22066 PR rtl-optimization/56494
22067 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
22068 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
22069 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
22071 PR middle-end/56461
22072 * sel-sched-ir.c (free_sched_pools): Release
22073 succs_info_pool.stack[succs_info_pool.max_top] vectors too
22074 if succs_info_pool.max_top isn't -1.
22077 * opts.c (opts_obstack, opts_concat): Moved to...
22078 * opts-common.c (opts_obstack, opts_concat): ... here.
22080 2013-03-04 Jakub Jelinek <jakub@redhat.com>
22082 PR middle-end/56461
22083 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
22085 2013-03-04 Martin Jambor <mjambor@suse.cz>
22087 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
22088 all appropriate places.
22090 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
22092 PR tree-optimization/56424
22093 * ipa-split.c (split_function): Do not set the RSO flag if result is
22094 not by reference and its type is a register type.
22096 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
22098 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
22099 (microblaze_legitimate_pic_operand): Likewise
22100 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
22101 new function microblaze_legitimate_pic_operand
22102 * config/microblaze/microblaze-protos.h
22103 (microblaze_legitimate_pic_operand): Declare.
22105 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22107 * config/microblaze/predicates.md (call_insn_simple_operand):
22108 New predicate for supported rtx code types.
22109 * config/microblaze/microblaze.md (call_internal1): Use
22110 call_insn_simple_operand predicate.
22112 2013-03-04 Jakub Jelinek <jakub@redhat.com>
22114 PR middle-end/56461
22115 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
22116 partitions.ordered_remove.
22118 PR middle-end/56461
22119 * tree-vect-stmts.c (vectorizable_conversion): Don't call
22120 vec_oprnds0.create (1) for modifier == NONE.
22122 PR middle-end/56461
22123 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
22124 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
22125 vec_oprnds1 right before pushing anything to it for
22128 PR middle-end/56461
22129 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
22130 set nbbs to 0 instead of having separate code path.
22131 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
22132 instead of false as last argument if returning NULL.
22134 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
22136 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
22137 the attribute is now called "target" instead of "option".
22138 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
22139 * doc/tm.texi.in (Target Attributes): Likewise document the correct
22140 attribute/pragma name for TARGET_OPTION_VALID_P and
22141 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
22142 * doc/tm.texi: Regenerated.
22144 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
22146 * config/microblaze/microblaze.c:
22147 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
22148 * config/microblaze/microblaze.h: Add -mxl-reorder to
22150 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
22151 instructions emitted if TARGET_REORDER.
22152 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
22153 or 0 for -m/-mno case, but initialises as 2 to detect default use case
22156 2013-03-01 Xinliang David Li <davidxl@google.com>
22158 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
22161 2013-03-01 Jakub Jelinek <jakub@redhat.com>
22163 PR middle-end/56461
22164 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
22165 vector even when returning true. Fix up function comment formatting.
22167 PR middle-end/56461
22168 * ira-build.c (ira_loop_nodes_count): New variable.
22169 (create_loop_tree_nodes): Initialize it.
22170 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
22172 PR middle-end/56461
22173 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
22174 method on dr_chain and result_chain.
22175 * tree-vect-stmts.c (vectorizable_store): Only call
22176 result_chain.create if j == 0.
22178 PR middle-end/56461
22179 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
22180 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
22181 before overwriting it.
22183 2013-03-01 Tobias Burnus <burnus@net-b.de>
22185 * doc/extended.texi (C Extensions): Change order in @menu
22187 (Other MIPS Built-in Functions): Move last MIPS entry before
22188 "picoChip Built-in Functions".
22189 (SH Built-in Functions): Move after RX Built-in Functions.
22190 * doc/gcc.texi (Introduction): Change order in @menu
22192 * doc/md.texi (Constraints): Ditto.
22193 * gty.texi (Type Information): Ditto.
22194 (User-provided marking routines for template types): Make
22196 * doc/invoke.texi (AArch64 Options): Move before
22197 "Adapteva Epiphany Options".
22199 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
22200 Jakub Jelinek <jakub@redhat.com>
22203 * asan.c (gate_asan): Lookup no_sanitize_address instead of
22204 no_address_safety_analysis attribute.
22205 * doc/extend.texi (no_address_safety_attribute): Rename to
22206 no_sanitize_address attribute, mention no_address_safety_analysis
22207 attribute as deprecated alias.
22209 2013-02-28 Jakub Jelinek <jakub@redhat.com>
22211 PR middle-end/56461
22212 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
22213 type to vec<vec<tree> > *.
22214 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
22215 to be vec<tree> instead of vec<tree> *, set vec_defs
22216 to vNULL and call vec_defs.create (number_of_vects), adjust other
22218 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
22219 vectorizable_condition): Adjust vect_get_slp_defs callers.
22221 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22223 * config/aarch64/aarch64.c
22224 (aarch64_float_const_representable): Remove unused variable.
22226 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22228 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
22230 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22232 * config/aarch64/aarch64-builtins.c
22233 (aarch64_init_simd_builtins): Make static.
22235 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22237 * config/aarch64/aarch64.c
22238 (aarch64_simd_make_constant): Make static.
22240 2013-02-28 Martin Jambor <mjambor@suse.cz>
22242 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
22243 with no initialization to the RHS of debug statements.
22245 2013-02-28 Martin Jambor <mjambor@suse.cz>
22247 PR tree-optimization/56294
22248 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
22250 (get_access_replacement): Do not call create_access_replacement.
22251 Assert a replacement exists.
22252 (get_repl_default_def_ssa_name): Create the replacement declaration
22255 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22257 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
22258 final_end_function.
22260 2013-02-28 Marek Polacek <polacek@redhat.com>
22262 PR rtl-optimization/56466
22263 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
22264 if we're changing a loop.
22265 (peel_loops_completely): Likewise.
22267 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
22270 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
22272 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
22275 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
22276 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
22277 INTX_FTYPE_FX, FX_FTYPE_INTX.
22278 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
22280 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
22282 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
22283 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
22284 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
22285 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
22286 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
22287 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
22288 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
22289 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
22290 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
22291 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
22292 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
22293 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
22294 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
22295 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
22296 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
22297 (avrxmega6): Increase max flash segments from 5 to 6.
22298 * config/avr/t-multilib: Regenerate.
22299 * config/avr/avr-tables.opt: Regenerate.
22300 * doc/avr-mmcu.texi: Regenerate.
22302 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
22304 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
22305 (avr_device_to_arch): Rename to avr_device_to_ld.
22306 (avr_device_to_as): New prototype.
22307 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
22308 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
22309 * config/avr/driver-avr.c (avr_device_to_as): New.
22310 (avr_device_to_arch): Rename to avr_device_to_ld.
22312 2013-02-27 Jakub Jelinek <jakub@redhat.com>
22314 PR middle-end/56461
22315 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
22316 method on dr_chain and result_chain.
22318 PR middle-end/56461
22319 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
22320 pointer_set_destroy on not_executed_last_iteration.
22322 PR middle-end/56461
22323 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
22325 PR middle-end/56461
22326 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
22327 FOR_EACH_DEFINED_FUNCTION when freeing state.
22329 PR middle-end/56461
22330 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
22332 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
22335 PR middle-end/56461
22336 * ipa-cp.c (decide_whether_version_node): Call vec_free on
22337 known_aggs[i].items and release known_aggs vector.
22339 PR middle-end/56461
22340 * ipa-reference.c (propagate): Free node_info even for alias nodes.
22342 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22344 * config/microblaze/microblaze.c (microblaze_emit_compare):
22345 Use xor for EQ/NE comparisions.
22346 * config/microblaze/microblaze.md (cstoresf4): Add constraints
22347 (cbranchsf4): Adjust operator to comparison_operator.
22349 2013-02-27 Jakub Jelinek <jakub@redhat.com>
22351 PR middle-end/56461
22352 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
22354 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
22355 vec_safe_push, always update *slot.
22356 (redirect_edge_var_map_clear): Use vec_free.
22357 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
22358 (free_var_map_entry): Use vec_free.
22359 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
22360 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
22362 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
22364 PR middle-end/45472
22365 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
22366 when the may_trap_p bit of the exprs being merged differs.
22367 Reorder tests for speculativeness in the logical and operator.
22369 2013-02-27 Jakub Jelinek <jakub@redhat.com>
22371 * incpath.c (add_standard_paths): Use reconcat instead of concat
22372 where appropriate and avoid leaking memory.
22374 * opts.h: Include obstack.h.
22375 (opts_concat): New prototype.
22376 (opts_obstack): New declaration.
22377 * opts.c (opts_concat): New function.
22378 (opts_obstack): New variable.
22379 (init_options_struct): Call gcc_init_obstack on opts_obstack.
22380 (finish_options): Use opts_concat instead of concat
22381 and XOBNEWVEC instead of XNEWVEC.
22382 * opts-common.c (generate_canonical_option, decode_cmdline_option,
22383 generate_option): Likewise.
22384 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
22385 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
22388 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
22389 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
22391 2013-02-26 Jakub Jelinek <jakub@redhat.com>
22393 PR middle-end/56461
22394 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
22396 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
22398 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
22399 (arm_block_move_unaligned_straight): Likewise.
22400 (arm_adjust_block_mem): Likewise.
22402 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
22405 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
22406 temp, cond and label.
22407 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
22410 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
22411 * config/c6x/c6x.h (dbx_register_map): Update declaration.
22414 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
22415 of prologue/epilogue functions.
22418 * config/tilegx/tilegx.c (tilegx_expand_prologue):
22419 Remove unused variable cfa_offset.
22420 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
22423 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
22424 type promotion to unsigned.
22427 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
22428 for HOST_WIDE_INT of 32 bit / same size as int.
22429 (arm_block_move_unaligned_straight): Likewise.
22430 (arm_adjust_block_mem): Likewise.
22433 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
22436 2013-02-26 Marek Polacek <polacek@redhat.com>
22438 PR tree-optimization/56426
22439 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
22441 2013-02-26 Richard Biener <rguenther@suse.de>
22444 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
22445 unused variable loops.
22447 2013-02-26 Jakub Jelinek <jakub@redhat.com>
22449 PR tree-optimization/56448
22450 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
22451 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
22452 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
22453 later operands of the references, or even first operand for
22454 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
22456 PR tree-optimization/56443
22457 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
22458 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
22459 to type_for_mode langhook.
22461 2013-02-25 Matt Turner <mattst88@gmail.com>
22463 * doc/invoke.texi: Document r4700.
22465 2013-02-25 Richard Biener <rguenther@suse.de>
22467 PR tree-optimization/56175
22468 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
22470 (simplify_bitwise_binary): ... here. Also guard the conversion
22471 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
22473 2013-02-25 Catherine Moore <clm@codesourcery.com>
22476 2013-02-24 Catherine Moore <clm@codesourcery.com>
22477 Maciej W. Rozycki <macro@codesourcery.com>
22478 Tom de Vries <tom@codesourcery.com>
22479 Nathan Sidwell <nathan@codesourcery.com>
22480 Iain Sandoe <iain@codesourcery.com>
22481 Nathan Froyd <froydnj@codesourcery.com>
22482 Chao-ying Fu <fu@mips.com>
22484 * doc/extend.texi: (micromips, nomicromips, nocompression):
22485 Document new function attributes.
22486 * doc/invoke.texi (minterlink-compressed, mmicromips,
22487 m14k, m14ke, m14kec): Document new options.
22488 (minterlink-mips16): Update documentation.
22489 * doc/md.texi (ZC, ZD): Document new constraints.
22490 * configure.ac (gcc_cv_as_micromips): Check if linker
22491 supports the .set micromips directive.
22492 * configure: Regenerate.
22493 * config.in: Regenerate.
22494 * config/mips/mips-tables.opt: Regenerate.
22495 * config/mips/micromips.md: New file.
22496 * constraints.md (ZC, AD): New constraints.
22497 * config/mips/predicates.md (movep_src_register): New predicate.
22498 (movep_src_operand): New predicate.
22499 (non_volatile_mem_operand): New predicate.
22500 * config/mips/mips.md (multimem): New type.
22501 (length): Differentiate between 17-bit and 18-bit branch offsets.
22502 (MOVEP1, MOVEP2): New mode iterator.
22503 (mov_<load>l): Use ZC constraint.
22504 (mov_<load>r): Likewise.
22505 (mov_<store>l): Likewise.
22506 (mov_<store>r): Likewise.
22507 (*branch_equality<mode>_inverted): Add microMIPS support.
22508 (*branch_equality<mode>): Likewise.
22509 (*jump_absolute): Likewise.
22510 (indirect_jump_<mode>): Likewise.
22511 (tablejump_<mode>): Likewise.
22512 (<optab>_internal): Likewise.
22513 (sibcall_internal): Likewise.
22514 (sibcall_value_internal): Likewise.
22515 (prefetch): Use constraint ZD.
22516 * config/mips/mips.opt (minterlink-compressed): New option.
22517 (minterlink-mips16): Now an alias for minterlink-compressed.
22518 (mmicromips): New option.
22519 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22520 (compare_and_swap_12): Likewise.
22521 (sync_add<mode>): Likewise.
22522 (sync_<optab>_12): Likewise.
22523 (sync_old_<optab>_12): Likewise.
22524 (sync_new_<optab>_12): Likewise.
22525 (sync_nand_12): Likewise.
22526 (sync_old_nand_12): Likewise.
22527 (sync_new_nand_12): Likewise.
22528 (sync_sub<mode>): Likewise.
22529 (sync_old_add<mode>): Likewise.
22530 (sync_old_sub<mode>): Likewise.
22531 (sync_new_add<mode>): Likewise.
22532 (sync_new_sub<mode>): Likewise.
22533 (sync_<optab><mode>): Likewise.
22534 (sync_old_<optab><mode>): Likewise.
22535 (sync_new_<optab><mode>): Likewise.
22536 (sync_nand<mode>): Likewise.
22537 (sync_old_nand<mode>): Likewise.
22538 (sync_new_nand<mode>): Likewise.
22539 (sync_lock_test_and_set<mode>): Likewise.
22540 (test_and_set_12): Likewise.
22541 (atomic_compare_and_swap<mode>): Likewise.
22542 (atomic_exchange<mode>_llsc): Likewise.
22543 (atomic_fetch_add<mode>_llsc): Likewise.
22544 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22545 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22546 (umips_save_restore_pattern_p): Likewise.
22547 (umips_load_store_pair_p): Likewise.
22548 (umips_output_load_store_pair): Likewise.
22549 (umips_movep_target_p): Likewise.
22550 (umips_12bit_offset_address_p): Likewise.
22551 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22552 (mips_base_mips16): Rename this...
22553 (mips_base_compression_flags): ...to this. Update all uses.
22554 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22555 (mips_mips16_decl_p): Delete.
22556 (mips_nomips16_decl_p): Delete.
22557 (mips_get_compress_on_flags): New function.
22558 (mips_get_compress_off_flags): New function.
22559 (mips_get_compress_mode): New function.
22560 (mips_get_compress_on_name): New function.
22561 (mips_get_compress_off_name): New function.
22562 (mips_insert_attributes): Support multiple compression types.
22563 (mips_merge_decl_attributes): Likewise.
22564 (umips_12bit_offset_address_p): New function.
22565 (mips_start_function_definition): Emit .set micromips directive.
22566 (mips_call_may_need_jalx_p): New function.
22567 (mips_function_ok_for_sibcall): Add microMIPS support.
22568 (mips_print_operand_punctuation): Support short delay slots and
22570 (umips_swm_mask, umips_swm_encoding): New.
22571 (umips_build_save_restore): New function.
22572 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22573 (was_mips16_p): Remove.
22574 (old_compression_mode): New.
22575 (mips_set_compression_mode): New function.
22576 (mips_set_current_function): Add microMIPS support.
22577 (mips_option_override): Likewise.
22578 (umips_save_restore_pattern_p): New function.
22579 (umips_output_save_restore): New function.
22580 (umips_load_store_pair_p_1): New function.
22581 (umips_load_store_pair_p): New function.
22582 (umips_output_load_store_pair_1): New function.
22583 (umips_output_load_store_pair): New function.
22584 (umips_movep_target_p) New function.
22585 (mips_prepare_pch_save): Add microMIPS support.
22586 * config/mips/mips.h (TARGET_COMPRESSION): New.
22587 (TARGET_CPU_CPP_BUILTINS): Update macro
22588 to use new compression flags and to support microMIPS.
22589 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22590 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22591 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22592 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22593 (ASM_SPEC): Support mmicromips and mno-micromips.
22594 (M16STORE_REG_P): New macro.
22595 (MIPS_CALL): Support TARGET_MICROMIPS.
22596 (MICROMIPS_J): New macro.
22597 (mips_base_mips16): Rename this...
22598 (mips_base_compression_flags): ...to this.
22599 (UMIPS_12BIT_OFFSET_P): New macro.
22600 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22601 (MULTILIB_DIRNAMES): Likewise.
22603 2013-02-25 Tom de Vries <tom@codesourcery.com>
22605 PR rtl-optimization/56131
22606 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
22607 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
22608 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
22610 2013-02-25 Tobias Burnus <burnus@net-b.de>
22612 * doc/invoke.texi (-fsanitize=): Move from optimization
22613 to debugging options.
22615 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
22617 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
22619 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
22620 Alexander Monakov <amonakov@ispras.ru>
22622 PR middle-end/56077
22623 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
22624 flush pending lists also on non-jumps. Adjust comment.
22626 2013-02-24 Catherine Moore <clm@codesourcery.com>
22627 Maciej W. Rozycki <macro@codesourcery.com>
22628 Tom de Vries <tom@codesourcery.com>
22629 Nathan Sidwell <nathan@codesourcery.com>
22630 Iain Sandoe <iain@codesourcery.com>
22631 Nathan Froyd <froydnj@codesourcery.com>
22632 Chao-ying Fu <fu@mips.com>
22634 * doc/extend.texi: (micromips, nomicromips, nocompression):
22635 Document new function attributes.
22636 * doc/invoke.texi (minterlink-compressed, mmicromips,
22637 m14k, m14ke, m14kec): Document new options.
22638 (minterlink-mips16): Update documentation.
22639 * doc/md.texi (ZC, ZD): Document new constraints.
22640 * configure.ac (gcc_cv_as_micromips): Check if linker
22641 supports the .set micromips directive.
22642 * configure: Regenerate.
22643 * config.in: Regenerate.
22644 * config/mips/mips-tables.opt: Regenerate.
22645 * config/mips/micromips.md: New file.
22646 * constraints.md (ZC, AD): New constraints.
22647 * config/mips/predicates.md (movep_src_register): New predicate.
22648 (movep_src_operand): New predicate.
22649 (non_volatile_mem_operand): New predicate.
22650 * config/mips/mips.md (multimem): New type.
22651 (length): Differentiate between 17-bit and 18-bit branch offsets.
22652 (MOVEP1, MOVEP2): New mode iterator.
22653 (mov_<load>l): Use ZC constraint.
22654 (mov_<load>r): Likewise.
22655 (mov_<store>l): Likewise.
22656 (mov_<store>r): Likewise.
22657 (*branch_equality<mode>_inverted): Add microMIPS support.
22658 (*branch_equality<mode>): Likewise.
22659 (*jump_absolute): Likewise.
22660 (indirect_jump_<mode>): Likewise.
22661 (tablejump_<mode>): Likewise.
22662 (<optab>_internal): Likewise.
22663 (sibcall_internal): Likewise.
22664 (sibcall_value_internal): Likewise.
22665 (prefetch): Use constraint ZD.
22666 * config/mips/mips.opt (minterlink-compressed): New option.
22667 (minterlink-mips16): Now an alias for minterlink-compressed.
22668 (mmicromips): New option.
22669 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22670 (compare_and_swap_12): Likewise.
22671 (sync_add<mode>): Likewise.
22672 (sync_<optab>_12): Likewise.
22673 (sync_old_<optab>_12): Likewise.
22674 (sync_new_<optab>_12): Likewise.
22675 (sync_nand_12): Likewise.
22676 (sync_old_nand_12): Likewise.
22677 (sync_new_nand_12): Likewise.
22678 (sync_sub<mode>): Likewise.
22679 (sync_old_add<mode>): Likewise.
22680 (sync_old_sub<mode>): Likewise.
22681 (sync_new_add<mode>): Likewise.
22682 (sync_new_sub<mode>): Likewise.
22683 (sync_<optab><mode>): Likewise.
22684 (sync_old_<optab><mode>): Likewise.
22685 (sync_new_<optab><mode>): Likewise.
22686 (sync_nand<mode>): Likewise.
22687 (sync_old_nand<mode>): Likewise.
22688 (sync_new_nand<mode>): Likewise.
22689 (sync_lock_test_and_set<mode>): Likewise.
22690 (test_and_set_12): Likewise.
22691 (atomic_compare_and_swap<mode>): Likewise.
22692 (atomic_exchange<mode>_llsc): Likewise.
22693 (atomic_fetch_add<mode>_llsc): Likewise.
22694 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22695 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22696 (umips_save_restore_pattern_p): Likewise.
22697 (umips_load_store_pair_p): Likewise.
22698 (umips_output_load_store_pair): Likewise.
22699 (umips_movep_target_p): Likewise.
22700 (umips_12bit_offset_address_p): Likewise.
22701 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22702 (mips_base_mips16): Rename this...
22703 (mips_base_compression_flags): ...to this. Update all uses.
22704 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22705 (mips_mips16_decl_p): Delete.
22706 (mips_nomips16_decl_p): Delete.
22707 (mips_get_compress_on_flags): New function.
22708 (mips_get_compress_off_flags): New function.
22709 (mips_get_compress_mode): New function.
22710 (mips_get_compress_on_name): New function.
22711 (mips_get_compress_off_name): New function.
22712 (mips_insert_attributes): Support multiple compression types.
22713 (mips_merge_decl_attributes): Likewise.
22714 (umips_12bit_offset_address_p): New function.
22715 (mips_start_function_definition): Emit .set micromips directive.
22716 (mips_call_may_need_jalx_p): New function.
22717 (mips_function_ok_for_sibcall): Add microMIPS support.
22718 (mips_print_operand_punctuation): Support short delay slots and
22720 (umips_swm_mask, umips_swm_encoding): New.
22721 (umips_build_save_restore): New function.
22722 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22723 (was_mips16_p): Remove.
22724 (old_compression_mode): New.
22725 (mips_set_compression_mode): New function.
22726 (mips_set_current_function): Add microMIPS support.
22727 (mips_option_override): Likewise.
22728 (umips_save_restore_pattern_p): New function.
22729 (umips_output_save_restore): New function.
22730 (umips_load_store_pair_p_1): New function.
22731 (umips_load_store_pair_p): New function.
22732 (umips_output_load_store_pair_1): New function.
22733 (umips_output_load_store_pair): New function.
22734 (umips_movep_target_p) New function.
22735 (mips_prepare_pch_save): Add microMIPS support.
22736 * config/mips/mips.h (TARGET_COMPRESSION): New.
22737 (TARGET_CPU_CPP_BUILTINS): Update macro
22738 to use new compression flags and to support microMIPS.
22739 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22740 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22741 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22742 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22743 (ASM_SPEC): Support mmicromips and mno-micromips.
22744 (M16STORE_REG_P): New macro.
22745 (MIPS_CALL): Support TARGET_MICROMIPS.
22746 (MICROMIPS_J): New macro.
22747 (mips_base_mips16): Rename this...
22748 (mips_base_compression_flags): ...to this.
22749 (UMIPS_12BIT_OFFSET_P): New macro.
22750 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22751 (MULTILIB_DIRNAMES): Likewise.
22753 2013-02-24 Jakub Jelinek <jakub@redhat.com>
22756 * target-globals.c (save_target_globals): For init_reg_sets and
22757 target_reinit remporarily set this_fn_optabs to this_target_optabs.
22759 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
22761 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
22762 * config/aarch64/t-aarch64
22763 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
22765 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
22767 PR inline-asm/56148
22768 * lra-constraints.c (process_alt_operands): Reload operand
22769 conflicting with earlier clobber only if no more other conflicting
22772 2013-02-22 Jakub Jelinek <jakub@redhat.com>
22775 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
22776 if not linking a shared library.
22778 2013-02-22 Seth LaForge <sethml@google.com>
22780 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
22782 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
22784 * config/arm/arm.md (split for extendsidi): Update condition.
22785 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
22786 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
22787 (qhs_zextenddi_cstr): Likewise.
22789 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22791 PR middle-end/56420
22792 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
22793 avoid signed wrapping.
22794 (expand_mult): Handle properly multiplication by
22795 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
22796 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
22797 in the compiler if coeff is HOST_WIDE_INT_MIN.
22798 (expand_divmod): Don't make ext_op1 static, change it's type to
22799 uhwi. Avoid undefined behavior in -INTVAL (op1).
22801 PR rtl-optimization/50339
22802 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
22804 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
22805 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
22806 into splitting_ashiftrt field.
22807 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
22809 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
22812 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
22814 PR middle-end/56108
22815 * trans-mem.c (execute_tm_mark): Do not expand transactions that
22816 are sure to go irrevocable.
22818 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
22820 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
22821 scalars are valid operands.
22823 2013-02-21 Martin Jambor <mjambor@suse.cz>
22825 PR tree-optimization/56310
22826 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
22827 only matching indices and non-negative final offsets.
22828 (intersect_aggregates_with_edge): Pass src_idx to
22829 agg_replacements_to_vector. Pass src_idx insstead of index to
22830 intersect_with_agg_replacements.
22832 2013-02-21 Martin Jambor <mjambor@suse.cz>
22834 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
22835 instead of hard-wired defaults.
22837 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
22839 * doc/invoke.texi (MIPS Options): Update documentation of the
22840 floating-point multiply-accumulate instruction restrictions.
22842 2013-02-21 Kostya Serebryany <kcc@google.com>
22844 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
22845 asan_shadow_offset on x86_64 linux.
22847 2013-02-21 Richard Biener <rguenther@suse.de>
22849 PR tree-optimization/56415
22851 2013-02-11 Richard Biener <rguenther@suse.de>
22853 PR tree-optimization/56273
22854 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22857 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22860 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
22863 PR inline-asm/56405
22864 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
22865 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
22867 2013-02-20 Jan Hubicka <jh@suse.cz>
22869 PR tree-optimization/56265
22870 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
22871 when target is referenced for first time.
22873 2013-02-20 Richard Biener <rguenther@suse.de>
22875 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
22876 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
22877 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
22878 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
22879 not return anything.
22880 (rename_ssa_copies): Do not remove unused locals.
22881 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
22882 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
22883 * passes.c (execute_function_todo): Do not schedule unused locals
22884 removal if cleanup_tree_cfg did something.
22885 * tree-ssa-live.c (remove_unused_locals): Dump statistics
22886 about the number of removed locals.
22888 2013-02-20 Richard Biener <rguenther@suse.de>
22890 PR tree-optimization/56398
22891 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
22893 2013-02-20 Martin Jambor <mjambor@suse.cz>
22895 PR tree-optimization/55334
22896 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
22897 restricted pointers to arrays.
22899 2013-02-20 Richard Biener <rguenther@suse.de>
22900 Jakub Jelinek <jakub@redhat.com>
22902 PR tree-optimization/56396
22903 * tree-ssa-ccp.c (n_const_val): New static variable.
22904 (get_value): Return NULL for SSA names we don't have a lattice
22906 (ccp_initialize): Initialize n_const_val.
22907 * tree-ssa-copy.c (n_copy_of): New static variable.
22908 (init_copy_prop): Initialize n_copy_of.
22909 (get_value): Return NULL_TREE for SSA names we don't have a
22912 2013-02-20 Martin Jambor <mjambor@suse.cz>
22914 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
22916 2013-02-20 Richard Biener <rguenther@suse.de>
22918 * genpreds.c (write_lookup_constraint): Do not compare first
22919 letter of the constraint again.
22921 2013-02-20 Richard Biener <rguenther@suse.de>
22923 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
22925 (get_use_iv_cost): Terminate hashtable walk when coming across
22928 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
22930 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
22931 reassociation for avx2 targets.
22933 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22935 * config/microblaze/microblaze.c: microblaze_has_clz = 0
22936 Add version check for v8.10.a to enable microblaze_has_clz
22937 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
22938 version and TARGET_PATTERN_COMPARE check
22939 * config/microblaze/microblaze.md: New clzsi2 instruction
22941 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22943 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
22944 function before branching.
22946 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22948 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
22950 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
22952 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22954 PR middle-end/55889
22955 * sel-sched.c: Include ira.h.
22956 (implicit_clobber_conflict_p): New function.
22957 (moveup_expr): Use it.
22958 * Makefile.in (sel-sched.o): Depend on ira.h.
22960 2013-02-19 Richard Biener <rguenther@suse.de>
22962 PR tree-optimization/56384
22963 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
22964 (vn_hash_type): Split out from ...
22965 (vn_hash_constant_with_type): ... here.
22966 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
22967 (vn_phi_eq): Compare types from vn_phi_s structure.
22968 (vn_phi_lookup): Populate vn_phi_s type.
22969 (vn_phi_insert): Likewise.
22971 2013-02-19 Jakub Jelinek <jakub@redhat.com>
22973 PR tree-optimization/56350
22974 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
22975 if haven't found reduction or nested cycle operand, rather than
22976 asserting we must find it.
22978 PR tree-optimization/56381
22979 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
22982 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
22983 Jakub Jelinek <jakub@redhat.com>
22986 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
22987 (swap_optab_enable): Same.
22988 (init_all_optabs): Use argument instead of global.
22989 * tree.h (struct tree_optimization_option): New field target_optabs.
22990 * expr.h (init_all_optabs): Add argument to prototype.
22991 (TREE_OPTIMIZATION_OPTABS): New.
22992 (save_optabs_if_changed): Protoize.
22993 * optabs.h: Declare this_fn_optabs.
22994 * optabs.c (save_optabs_if_changed): New.
22995 Declare this_fn_optabs.
22996 (init_optabs): Add argument to init_all_optabs() call.
22997 * function.c (invoke_set_current_function_hook): Handle per
22999 * function.h (struct function): New field optabs.
23000 * config/mips/mips.c (mips_set_mips16_mode): Handle when
23001 optimization_current_node has changed.
23002 * target-globals.h (save_target_globals_default_opts): Protoize.
23003 * target-globals.c (save_target_globals_default_opts): New.
23005 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23008 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
23009 registers %fr12 and %fr12R as call used.
23012 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
23013 and HImode, require all displacements to be an integer multiple of
23015 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
23016 only allow QImode and HImode when reload is in progress and strict is
23017 true. Likewise for symbolic addresses. Use base14_operand to check
23018 displacements in REG+BASE addresses.
23020 2013-02-18 Richard Biener <rguenther@suse.de>
23022 PR tree-optimization/56366
23023 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23024 handle sign-conversion of outer-loop initial induction value.
23026 2013-02-18 Richard Biener <rguenther@suse.de>
23028 PR middle-end/56349
23029 * cfghooks.c (merge_blocks): If we merge a latch into another
23030 block adjust references to it.
23031 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
23032 (verify_loop_structure): Verify that a recorded latch is in fact
23035 2013-02-18 Richard Biener <rguenther@suse.de>
23037 PR tree-optimization/56321
23038 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
23039 order SSA name release and virtual operand unlinking.
23041 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23043 * config/microblaze/microblaze.md (save_stack_block): Define.
23044 (restore_stack_block): Likewise.
23046 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23048 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
23049 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
23050 * config/microblaze/microblaze.c (microblaze_option_override):
23051 Bail out early for PIC modes when target does not support PIC.
23053 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23055 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
23056 Replace with a microblaze version.
23057 (microblaze_trampoline_init): Adapt for microblaze.
23058 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
23061 2013-02-16 Jakub Jelinek <jakub@redhat.com>
23062 Dodji Seketeli <dodji@redhat.com>
23065 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
23066 (instrument_mem_region_access): Do not forget to always put
23067 instrumentation of the of 'base' and 'base + len' in a "if (len !=
23068 0) statement, even for cases where either 'base' or 'base + len'
23069 are not instrumented -- because they have been previously
23070 instrumented. Simplify the logic by putting all the statements
23071 instrument 'base + len' inside a sequence, and then insert that
23072 sequence right before the current insertion point. Then, to
23073 instrument 'base + len', just get an iterator on that statement.
23074 And do not forget to update the pointer to iterator the function
23075 received as argument.
23077 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
23079 PR rtl-optimization/56348
23080 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
23082 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
23084 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
23085 (clean_graph_dump_file): Pass base to start_graph_dump.
23087 2013-02-14 Richard Henderson <rth@redhat.com>
23090 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
23092 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
23094 * collect2-aix.h: Define F_LOADONLY.
23096 2013-02-14 Richard Biener <rguenther@suse.de>
23099 * varasm.c (output_constant_def_1): Get the decl representing
23100 the constant as argument.
23101 (output_constant_def): Wrap output_constant_def_1.
23102 (make_decl_rtl): Use output_constant_def_1 with the decl
23103 representing the constant.
23104 (build_constant_desc): Optionally re-use a decl already
23105 representing the constant.
23106 (tree_output_constant_def): Adjust.
23108 2013-02-14 Dodji Seketeli <dodji@redhat.com>
23111 * asan.c (instrument_builtin_call): Really put the length of the
23112 second source argument into src1_len.
23114 2013-02-13 Jakub Jelinek <jakub@redhat.com>
23116 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
23117 argument. If it is false, don't create edge from then_bb to
23119 (insert_if_then_before_iter): Pass true to it.
23120 (build_check_stmt): Pass false to it.
23121 (transform_statements): Flush hash table only on extended basic
23122 block boundaries, rather than at the beginning of every bb.
23123 Don't flush hash table on nonfreeing_call_p calls.
23124 * tree-flow.h (nonfreeing_call_p): New prototype.
23125 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
23127 2013-02-13 David S. Miller <davem@davemloft.net>
23129 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
23131 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
23134 * ira.c (max_regno_before_ira): Move from ...
23136 (fix_reg_equiv_init): Use max_regno_before_ira instead of
23139 2013-02-13 Jakub Jelinek <jakub@redhat.com>
23141 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
23143 2013-02-13 Richard Biener <rguenther@suse.de>
23146 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
23147 globals in MEM_REFs.
23149 2013-02-13 Richard Biener <rguenther@suse.de>
23151 * loop-init.c (loop_optimizer_init): Clear loop state when
23152 re-initializing preserved loops.
23153 * loop-unswitch.c (unswitch_single_loop): Return whether
23154 we unswitched the loop. Do not verify loop state here.
23155 (unswitch_loops): When we unswitched a loop discover new loops.
23157 2013-02-13 Kostya Serebryany <kcc@google.com>
23159 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
23161 * sanitizer.def: Rename __asan_init to __asan_init_v1.
23163 2013-02-12 Dodji Seketeli <dodji@redhat.com>
23165 Avoid instrumenting duplicated memory access in the same basic block
23166 * Makefile.in (asan.o): Add new dependency on hash-table.h
23167 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
23168 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
23169 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
23170 (free_mem_ref_resources, has_mem_ref_been_instrumented)
23171 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
23172 (get_mem_ref_of_assignment): New functions.
23173 (get_mem_refs_of_builtin_call): Extract from
23174 instrument_builtin_call and tweak a little bit to make it fit with
23176 (instrument_builtin_call): Use the new
23177 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
23178 of is_gimple_builtin_call.
23179 (instrument_derefs, instrument_mem_region_access): Insert the
23180 instrumented memory reference into the hash table.
23181 (maybe_instrument_assignment): Renamed instrument_assignment into
23182 this, and change it to advance the iterator when instrumentation
23183 actually happened and return true in that case. This makes it
23184 homogeneous with maybe_instrument_assignment, and thus give a
23185 chance to callers to be more 'regular'.
23186 (transform_statements): Clear the memory reference hash table
23187 whenever we enter a new BB, when we cross a function call, or when
23188 we are done transforming statements. Use
23189 maybe_instrument_assignment instead of instrumentation. No more
23190 need to special case maybe_instrument_assignment and advance the
23191 iterator after calling it; it's now handled just like
23192 maybe_instrument_call. Update comment.
23194 2013-02-13 Richard Biener <rguenther@suse.de>
23196 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
23197 Fix loop discovery code.
23199 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
23201 PR inline-asm/56148
23202 * lra-constraints.c (process_alt_operands): Match early clobber
23203 operand with itself. Check conflicts with earlyclobber only if
23204 the operand is not reloaded. Prefer to reload conflicting operand
23205 if earlyclobber and matching operands are the same.
23207 2013-02-12 Richard Biener <rguenther@suse.de>
23210 * lto-streamer-out.c (write_symbol): Do not output symbols
23211 for hard register variables.
23213 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
23216 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
23217 (umulsidi3_insn, mulsidi3_insn): New insns.
23219 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
23221 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
23222 (struct tune_params): Add vec_costs field.
23223 * config/arm/arm.c (arm_builtin_vectorization_cost)
23224 (arm_add_stmt_cost): New functions.
23225 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
23226 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
23227 (arm_default_vec_cost): New struct of type cpu_vec_costs.
23228 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
23229 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
23230 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
23231 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
23233 2013-02-12 Richard Biener <rguenther@suse.de>
23236 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
23237 decls again if possible.
23239 2013-02-12 Richard Biener <rguenther@suse.de>
23241 PR middle-end/56288
23242 * tree-ssa.c (verify_ssa_name): Fix check, move
23243 SSA_NAME_IN_FREE_LIST check up.
23245 2013-02-12 Jakub Jelinek <jakub@redhat.com>
23246 Steven Bosscher <steven@gcc.gnu.org>
23248 PR rtl-optimization/56151
23249 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
23250 equal to op0 or op1, and last_insn pattern is CODE operation
23251 with MEM dest and one of the operands matches that MEM.
23253 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
23255 * doc/extend.texi: Document Function Multiversioning and "default"
23256 parameter string to target attribute.
23257 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
23258 target attribute parameter is "default".
23259 (ix86_compare_version_priority): Remove checks for target attribute.
23260 (ix86_mangle_function_version_assembler_name): Change error to sorry.
23261 Remove check for target attribute equal to NULL. Add assert.
23262 (ix86_generate_version_dispatcher_body): Change error to sorry.
23264 2013-02-11 Iain Sandoe <iain@codesourcery.com>
23265 Jack Howarth <howarth@bromo.med.uc.edu>
23266 Patrick Marlier <patrick.marlier@gmail.com>
23269 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
23270 define ENDFILE_SPEC as TM_DESTRUCTOR.
23271 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
23273 2013-02-11 Alexander Potapenko <glider@google.com>
23274 Jack Howarth <howarth@bromo.med.uc.edu>
23275 Jakub Jelinek <jakub@redhat.com>
23278 * config/darwin.c (cdtor_record): Rename ctor_record.
23279 (sort_cdtor_records): Rename sort_ctor_records.
23280 (finalize_dtors): New routine to sort destructors by
23281 priority before use in assemble_integer.
23282 (machopic_asm_out_destructor): Use finalize_dtors if needed.
23284 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
23286 PR rtl-optimization/56275
23287 * simplify-rtx.c (avoid_constant_pool_reference): Check that
23288 offset is non-negative and less than cmode size before
23289 calling simplify_subreg.
23291 2013-02-11 Richard Biener <rguenther@suse.de>
23293 PR tree-optimization/56264
23294 * cfgloop.h (fix_loop_structure): Adjust prototype.
23295 * loop-init.c (fix_loop_structure): Return the number of
23296 newly discovered loops.
23297 * tree-cfgcleanup.c (repair_loop_structures): When new loops
23298 are discovered, do a full loop-closed SSA rewrite.
23300 2013-02-11 Richard Biener <rguenther@suse.de>
23302 PR tree-optimization/56273
23303 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
23305 (check_array_ref): Fix missing newline in dumps.
23306 (search_for_addr_array): Likewise.
23308 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
23310 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
23312 2013-02-09 Jakub Jelinek <jakub@redhat.com>
23315 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
23317 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
23319 PR rtl-optimization/56246
23320 * lra-constraints.c (simplify_operand_subreg): Try to reuse
23322 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
23323 constraints are satisfied.
23325 2013-02-08 Jeff Law <law@redhat.com>
23328 * emit-rtl.c (reg_is_parm_p): New function.
23329 * regs.h (reg_is_parm_p): New prototype.
23330 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
23331 callee-clobbered registers.
23333 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
23336 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
23337 If there is no implicit builtin declaration, just return NULL.
23339 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
23341 * config/i386/sse.md (FMAMODEM): New mode iterator.
23342 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
23343 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
23345 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
23347 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
23348 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
23349 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
23351 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23353 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
23354 (microblaze*-*-elf): Likewise.
23355 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
23357 * config/microblaze/microblaze-c.c: Add builtin defines for
23358 _LITTLE_ENDIAN and _BIG_ENDIAN.
23359 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
23360 add to TARGET_DEFAULT flags.
23361 Expand ASM_SPEC and LINK_SPEC.
23362 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
23363 * config/microblaze/microblaze.md: Update extendsidi2 and
23364 movdi_internal instructions to use low-order / high-order reg
23366 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
23367 options and inversemask / mask of LITTLE_ENDIAN.
23368 * config/microblaze/t-microblaze: Expand multilib options to
23369 include mlittle-endian (le) and update exceptions patterns.
23371 2013-02-08 Jakub Jelinek <jakub@redhat.com>
23373 PR rtl-optimization/56195
23374 * lra-constraints.c (get_reload_reg): Don't reuse regs
23375 if they have smaller mode than requested, if they have
23376 wider mode than requested, try to return a SUBREG.
23378 PR tree-optimization/56250
23379 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
23380 if type is unsigned and code isn't MULT_EXPR.
23382 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
23384 PR tree-optimization/56064
23385 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
23386 bits according to mode.
23387 * fixed-value.h (fixed_from_double_int)
23388 (const_fixed_from_double_int): Adjust comments.
23390 2013-02-08 Richard Biener <rguenther@suse.de>
23393 * lto-streamer.h (struct data_in): Remove current_file, current_line
23394 and current_col members.
23395 * lto-streamer-out.c (lto_output_location): Stream changed bits
23396 en-block for efficiency.
23397 * lto-streamer-in.c (clear_line_info): Remove.
23398 (lto_input_location): Cache current file, line and column
23399 globally via local statics. Read changed bits en-block.
23400 (input_function): Do not call clear_line_info.
23401 (lto_read_body): Likewise.
23402 (lto_input_toplevel_asms): Likewise.
23404 2013-02-08 Michael Matz <matz@suse.de>
23406 PR tree-optimization/52448
23407 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
23408 (nt_call_phase): New static.
23409 (add_or_mark_expr): Only mark accesses with newer phase than any
23411 (nonfreeing_call_p): New.
23412 (nt_init_block): Update nt_call_phase, mark blocks as visited.
23413 (nt_fini_block): Keep blocks marked as visited.
23414 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
23416 2013-02-08 Richard Biener <rguenther@suse.de>
23418 * ira.c (ira): Free broken dominator information.
23420 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
23422 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
23424 2013-02-08 Marek Polacek <polacek@redhat.com>
23426 * cfgloop.c (verify_loop_structure): Add more checking of headers.
23428 2013-02-08 Richard Biener <rguenther@suse.de>
23430 PR middle-end/56181
23431 * cfgloop.h (flow_loops_find): Adjust.
23432 (bb_loop_header_p): Declare.
23433 * cfgloop.c (bb_loop_header_p): New function split out from ...
23434 (flow_loops_find): ... here. Adjust function signature,
23435 support incremental loop structure update.
23436 (verify_loop_structure): Cleanup. Verify a loop is a loop.
23437 * cfgloopmanip.c (fix_loop_structure): Move ...
23438 * loop-init.c (fix_loop_structure): ... here.
23439 (apply_loop_flags): Split out from ...
23440 (loop_optimizer_init): ... here.
23441 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
23442 in incremental mode, only remove dead loops here.
23444 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
23447 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
23448 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
23449 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
23450 (*round<mode>3.libgcc): New insns for fixed-modes.
23451 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
23452 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
23453 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
23454 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
23455 implementations. Define to __builtin_avr_absFX,
23456 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
23457 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
23458 __builtin_avr_countlsFX, respectively.
23459 * config/avr/avr-c.c (target.h): Include it.
23460 (enum avr_builtin_id): New enum.
23461 (avr_resolve_overloaded_builtin): New static function.
23462 (avr_register_target_pragmas): Use it to set
23463 targetm.resolve_overloaded_builtin.
23464 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
23465 tree nodes used by DEF_BUILTIN.
23466 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
23467 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
23468 <AVR_BUILTIN_xxBITS>: Same.
23470 2013-02-08 Richard Biener <rguenther@suse.de>
23472 * cfgloop.c (verify_loop_structure): Properly handle
23473 a loop exiting to another loop header.
23474 * ira-int.h (ira_loops): Remove.
23475 * ira.c (ira_loops): Remove.
23476 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
23477 (do_reload): Use loop_optimizer_finalize.
23478 * ira-build.c (create_loop_tree_nodes): Use get_loops and
23479 number_of_loops to access the loop tree.
23480 (more_one_region_p): Likewise.
23481 (finish_loop_tree_nodes): Likewise.
23482 (rebuild_regno_allocno_maps): Likewise.
23483 (mark_loops_for_removal): Likewise.
23484 (mark_all_loops_for_removal): Likewise.
23485 (remove_unnecessary_regions): Likewise.
23486 (ira_build): Likewise.
23487 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
23489 2013-02-08 Richard Biener <rguenther@suse.de>
23491 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
23492 * ipa-pure-const.c (analyze_function): Avoid calling
23493 mark_irreducible_loops twice.
23494 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
23496 2013-02-07 David S. Miller <davem@davemloft.net>
23498 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
23500 * var-tracking.c (vt_add_function_parameter): Test the presence of
23501 HAVE_window_save properly and do not remap argument registers when
23502 we have a leaf function.
23504 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
23507 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
23509 * config/i386/i386.c (ix86_print_operand): Ditto.
23511 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
23513 * lra-constraints.c (process_alt_operands): Fix recently added comment.
23515 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
23517 PR rtl-optimization/56225
23518 * lra-constraints.c (process_alt_operands): Check that reload hard
23519 reg can hold value for strict_low_part.
23521 2013-02-07 Jakub Jelinek <jakub@redhat.com>
23524 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
23525 dwarf2out_end_function.
23526 (in_first_function_p, maybe_at_text_label_p,
23527 first_loclabel_num_not_at_text_label): New variables.
23528 (dwarf2out_var_location): In the first function find out
23529 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
23530 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
23533 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
23535 PR rtl-optimization/56178
23536 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
23537 SUBREG of a register. Tidy up related block of code.
23538 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
23539 note if the source is a register or a SUBREG of a register.
23541 2013-02-07 Jakub Jelinek <jakub@redhat.com>
23544 * config/rs6000/rs6000.md (ptrm): New mode attr.
23545 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
23546 call_value_indirect_aix<pttrsize>,
23547 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
23550 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
23552 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
23553 if -bnortl. Convert to strcmp and strncmp.
23555 2013-02-07 Alan Modra <amodra@gmail.com>
23558 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
23559 addresses won't wrap when offsetting.
23560 (rs6000_secondary_reload): Provide secondary reloads needed for
23561 wrapping LO_SUM addresses.
23563 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
23565 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
23566 MACH, just __MACH__.
23568 2013-02-06 Richard Biener <rguenther@suse.de>
23570 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
23571 instead of calling fix_loop_structure.
23573 2013-02-06 Jakub Jelinek <jakub@redhat.com>
23575 PR middle-end/56217
23576 * omp-low.c (use_pointer_for_field): Return false if
23577 lower_send_shared_vars doesn't generate any copy-out code.
23579 2013-02-06 Tom de Vries <tom@codesourcery.com>
23581 PR rtl-optimization/56131
23582 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
23583 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
23584 of the label is NULL. Add comment.
23586 2013-02-05 Jakub Jelinek <jakub@redhat.com>
23588 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
23591 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
23592 (STATIC_LIBTSAN_LIBS): Likewise.
23593 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
23594 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
23595 is defined, don't add anything else beyond that.
23596 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
23597 (LINK_COMMAND_SPEC): Use them.
23599 PR tree-optimization/56205
23600 * tree-stdarg.c (check_all_va_list_escapes): Return true if
23601 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
23602 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
23604 2013-02-05 Richard Biener <rguenther@suse.de>
23606 PR tree-optimization/53342
23607 PR tree-optimization/53185
23608 * tree-vectorizer.h (vect_check_strided_load): Remove.
23609 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
23610 not disallow peeling for vectorized strided loads.
23611 (vect_check_strided_load): Make static and simplify.
23612 (vect_analyze_data_refs): Adjust.
23613 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
23614 correctly when vectorizing strided loads.
23616 2013-02-05 Richard Biener <rguenther@suse.de>
23618 * doc/install.texi: Refer to ISL, not PPL.
23620 2013-02-05 Jan Hubicka <jh@suse.cz>
23622 PR tree-optimization/55789
23623 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
23625 2013-02-05 Jan Hubicka <jh@suse.cz>
23627 PR tree-optimization/55789
23628 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
23629 the dead call anyway.
23631 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
23634 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
23636 2013-02-04 Alexander Potapenko <glider@google.com>
23637 Jack Howarth <howarth@bromo.med.uc.edu>
23638 Jakub Jelinek <jakub@redhat.com>
23641 * config/darwin.c (sort_ctor_records): Stabilized qsort
23642 on constructor priority by using original position.
23643 (finalize_ctors): New routine to sort constructors by
23644 priority before use in assemble_integer.
23645 (machopic_asm_out_constructor): Use finalize_ctors if needed.
23647 2013-02-04 Jakub Jelinek <jakub@redhat.com>
23650 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
23651 about visibility on artificial decls.
23652 * config/sol2.c (solaris_assemble_visibility): Likewise.
23654 2013-02-04 Kai Tietz <ktietz@redhat.com>
23657 * config/i386/i386.c (function_value_ms_64): Add additional valtype
23658 argument and improve checking of return-argument types for 16-byte
23660 (ix86_function_value_1): Add additional valtype argument on call
23661 of function_value_64.
23662 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
23663 handling infunction_value_64 function.
23665 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
23667 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
23669 2013-02-04 Richard Biener <rguenther@suse.de>
23671 PR tree-optimization/56188
23672 * tree-ssa-structalias.c (label_visit): Consider case with
23673 initially non-empty points-to set.
23674 (perform_var_substitution): Dump node mapping and clean up.
23676 2013-02-04 Richard Guenther <rguenther@suse.de>
23679 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
23680 node prevail as last resort.
23681 (lto_symtab_merge_decls): Remove guard on LTRANS here.
23682 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
23684 2013-02-04 Richard Biener <rguenther@suse.de>
23686 PR tree-optimization/56113
23687 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
23689 (equiv_class_lookup_or_add): ... this.
23690 (label_visit): Adjust and fix error in previous patch.
23691 (perform_var_substitution): Adjust.
23693 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
23695 * config/sh/divtab.c: Fix formatting and comments throughout the file.
23696 * config/sh/sh4-300.md: Likewise.
23697 * config/sh/sh4a.md: Likewise.
23698 * config/sh/constraints.md: Likewise.
23699 * config/sh/sh.md: Likewise.
23700 * config/sh/netbsd-elf.h: Likewise.
23701 * config/sh/predicates.md: Likewise.
23702 * config/sh/sh-protos.h: Likewise.
23703 * config/sh/ushmedia.h: Likewise.
23704 * config/sh/linux.h: Likewise.
23705 * config/sh/sh.c: Likewise.
23706 * config/sh/superh.h: Likewise.
23707 * config/sh/elf.h: Likewise.
23708 * config/sh/sh4.md: Likewise.
23709 * config/sh/sh.h: Likewise.
23711 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23713 * config/pa/constraints.md: Adjust unused letters. Change "T"
23714 constraint to match_test floating_point_store_memory_operand().
23715 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
23716 (base14_operand): New.
23717 (floating_point_store_memory_operand): New.
23718 (integer_store_memory_operand): Revise to use base14_operand and
23719 reg_plus_base_memory_operand.
23720 (move_dest_operand): Allow symbolic_memory_operands.
23721 (symbolic_memory_operand): Check for LO_SOM.
23722 (symbolic_operand): Change default case to break.
23723 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
23724 CONST_DOUBLE values to be reloaded by putting them into memory when
23725 the destination is a floating point register.
23726 (movdf): Remove code to handle CONST_DOUBLE.
23728 (reload_indf_r1): New.
23729 (reload_insf_r1): New.
23730 Consistently use "Q" and "T" constraints with integer and floating
23731 point move instructions, respectively.
23732 (movdi): Remove FAIL.
23733 Change predicate for source operand unamed DImode move from
23734 general_operand to move_src_operand.
23735 (umulsidi3): Change predicate for destination operand to
23737 Likewise for similar unamed patterns.
23738 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
23739 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
23740 (hppa_legitimize_address): Simplify mask calculation.
23741 (pa_emit_move_sequence): Revised handling of secondary reloads from
23742 REG+D addresses for floating point loads and stores. Directly handle
23743 loading CONST0_RTX (mode) to a floating point register.
23744 (pa_secondary_reload): Handle reloading DF and SFmode constant values
23745 to floating point registers. Don't restrict secondary reloads to
23746 floating point registers to integer modes. Revise some comments and
23748 (TARGET_LEGITIMATE_ADDRESS_P): Define.
23749 (pa_legitimate_address_p): New.
23750 (pa_legitimize_reload_address): New.
23751 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
23752 (STRICT_REG_OK_FOR_BASE_P): New.
23753 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
23754 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
23756 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
23757 Andrew Dixie <andrewd@gentrack.com>
23759 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
23762 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
23764 * expmed.c (extract_bit_field_1): Pass the full width of the
23765 structure to get_best_reg_extraction_insn.
23767 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
23770 * configure.ac (use_cxa_atexit): Add AIX.
23771 * configure: Regenerate.
23773 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
23775 2013-02-01 Jakub Jelinek <jakub@redhat.com>
23778 * final.c (need_profile_function): New variable.
23779 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
23780 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
23781 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
23782 notes, targetm.asm_out.function_prologue doesn't emit anything,
23783 HAVE_prologue and profiler should be emitted before prologue,
23784 set need_profile_function instead of emitting it.
23785 (final_scan_insn): If need_profile_function, emit
23786 profile_function on the first NOTE_INSN_BASIC_BLOCK or
23787 NOTE_INSN_FUNCTION_BEG note.
23789 2013-02-01 Richard Henderson <rth@redhat.com>
23791 * config/rs6000/rs6000.md (smulditi3): New.
23794 * config/alpha/alpha.md (umulditi3): New.
23796 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
23798 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
23799 (ASM_OUTPUT_ALIGNED_LOCAL): New.
23801 2013-02-01 Richard Biener <rguenther@suse.de>
23803 PR tree-optimization/56113
23804 * tree-ssa-structalias.c (label_visit): Reduce work for
23805 single-predecessor nodes.
23807 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
23809 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
23810 range isn't testing for zero.
23812 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
23814 PR middle-end/56113
23815 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
23817 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
23818 Nick Clifton <nickc@redhat.com>
23820 * config/v850/constraints.md (Q): Define as a memory constraint.
23821 * config/v850/predicates.md (label_ref_operand): New predicate.
23822 (e3v5_shift_operand): New predicate.
23823 (ior_operator): New predicate.
23824 * config/v850/t-v850: Add e3v5 multilib.
23825 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
23826 (v850_gen_movdi): Prototype.
23827 * config/v850/v850.c: Add support for e3v5 architecture.
23828 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
23830 (construct_save_jarl): Add e3v5 long JARL support.
23831 (v850_adjust_insn_length): New function. Adjust length of call
23832 insns when using e3v5 instructions.
23833 (v850_gen_movdi): New function: Generate instructions to move a
23835 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
23836 (CPP_SPEC): Define __v850e3v5__ as appropriate.
23837 (TARGET_USE_FPU): Enable for e3v5.
23838 (CONST_OK_FOR_W): New macro.
23839 (ADJUST_INSN_LENGTH): Define.
23840 * config/v850/v850.md (UNSPEC_LOOP): Define.
23841 (attr cpu): Add v850e3v5.
23842 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
23843 (movdi): New pattern.
23844 (movdi_internal): New pattern.
23845 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
23846 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
23847 (cstoresf4): Likewise.
23848 (cstoredf4): Likewise.
23849 (insv): New pattern.
23850 (rotlso3_a): New pattern.
23851 (rotlsi3_b): New pattern
23852 (rotlsi3_v850e3v5): New pattern.
23853 (doloop_begin): New pattern.
23854 (fix_loop_counter): New pattern.
23855 (doloop_end): New pattern.
23856 (branch_normal): Add e3v5 long branch support.
23857 (branch_invert): Likewise.
23858 (branch_z_normal): Likewise.
23859 (branch_z_invert): Likewise.
23860 (branch_nz_normal): Likewise.
23861 (branch_nz_invert): Likewise.
23862 (call_internal_short): Add e3v5 register-indirect JARL support.
23863 (call_internal_long): Likewise.
23864 (call_value_internal_short): Likewise.
23865 (call_value_internal_long): Likewise.
23866 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
23867 (mloop): New option.
23868 * config.gcc: Add support for configuring v840e3v5 target.
23869 * doc/invoke.texi: Document new v850 specific command line options.
23871 2013-01-31 Paul Koning <ni1d@arrl.net>
23875 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
23876 children if parent is a class.
23877 (prune_unused_types_prune): Don't add DW_AT_declaration.
23879 2013-01-31 Richard Biener <rguenther@suse.de>
23881 PR tree-optimization/56157
23882 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
23883 match up operand with SLP child.
23885 2013-01-31 Jason Merrill <jason@redhat.com>
23888 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
23889 parameters the first time.
23890 (gen_scheduled_generic_parms_dies): Check completeness here.
23892 2013-01-31 Richard Biener <rguenther@suse.de>
23894 PR middle-end/53073
23895 * common.opt (faggressive-loop-optimizations): New flag,
23896 enabled by default.
23897 * doc/invoke.texi (faggressive-loop-optimizations): Document.
23898 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
23899 infer_loop_bounds_from_undefined by it.
23901 2013-01-31 Richard Biener <rguenther@suse.de>
23903 PR tree-optimization/56150
23904 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
23905 visit virtual operands.
23906 (find_uses_to_rename_bb): Likewise.
23908 2013-01-31 Richard Biener <rguenther@suse.de>
23910 PR tree-optimization/56150
23911 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
23912 mixed store non-store stmts.
23914 2013-01-30 Jakub Jelinek <jakub@redhat.com>
23917 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
23918 LIBASAN_EARLY_SPEC is defined.
23919 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
23920 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
23922 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
23925 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
23926 invalid args instead of ICEing on it.
23927 (ix86_valid_target_attribute_tree): Return error_mark_node if
23928 ix86_valid_target_attribute_inner_p failed.
23929 (ix86_valid_target_attribute_p): Return false only if
23930 ix86_valid_target_attribute_tree returned error_mark_node. Allow
23931 target("default") attribute.
23932 (sorted_attr_string): Change argument from const char * to tree,
23933 merge in all target attribute arguments rather than just one.
23934 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
23935 instead of free. Avoid using strcat.
23936 (ix86_mangle_function_version_assembler_name): Mangle
23937 target("default") as if no target attribute is present. Adjust
23938 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
23939 instead of xmalloc and XDELETEVEC instead of free.
23940 (ix86_function_versions): Don't return true if one of the decls
23941 doesn't have target attribute. If they don't and one of the decls
23942 is DECL_FUNCTION_VERSIONED, report an error. Adjust
23943 sorted_attr_string caller. Use XDELETEVEC instead of free.
23944 (ix86_supports_function_versions): Remove.
23945 (make_name): Fix up formatting.
23946 (make_dispatcher_decl): Remove resolver_name and its initialization.
23947 Avoid leaking memory.
23948 (is_function_default_version): Return true if there is
23949 target("default") attribute rather than no target attribute at all.
23950 (make_resolver_func): Avoid leaking memory.
23951 (ix86_generate_version_dispatcher_body): Likewise.
23952 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
23953 * target.def (supports_function_versions): Remove.
23954 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
23955 * doc/tm.texi: Regenerated.
23957 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
23959 PR rtl-optimization/56144
23960 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
23961 for values with side effects.
23963 2013-01-30 Richard Biener <rguenther@suse.de>
23965 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
23966 (sparseset_pop): Likewise.
23967 * cfganal.c (compute_idf): Likewise. Increase work-stack size
23968 to be able to use quick_push in the worker loop.
23970 2013-01-30 Marek Polacek <polacek@redhat.com>
23972 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
23974 2013-01-30 Richard Biener <rguenther@suse.de>
23977 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
23979 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
23981 PR tree-optimization/56064
23982 * fixed-value.c (fixed_from_double_int): New function.
23983 * fixed-value.h (fixed_from_double_int): New prototype.
23984 (const_fixed_from_double_int): New static inline function.
23985 * fold-const.c (native_interpret_fixed): New static function.
23986 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
23987 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
23988 (native_encode_fixed): New static function.
23989 (native_encode_expr) <FIXED_CST>: Use it.
23990 (native_interpret_int): Move double_int worker code to...
23991 * double-int.c (double_int::from_buffer): ...this new static method.
23992 * double-int.h (double_int::from_buffer): Prototype it.
23994 2013-01-30 Richard Biener <rguenther@suse.de>
23996 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
23997 New pointer-map and obstack.
23998 (init_alias_vars): Allocate pointer-map and obstack.
23999 (delete_points_to_sets): Free them.
24000 (find_what_var_points_to): Cache result.
24001 (find_what_p_points_to): Adjust for changed interface of
24002 find_what_var_points_to.
24003 (compute_points_to_sets): Likewise.
24004 (ipa_pta_execute): Likewise.
24006 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24008 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
24009 * configure: Regenerate.
24010 * config.in: Regenerate.
24011 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
24012 #nobits/#progbits if supported.
24014 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
24017 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
24018 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
24019 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
24021 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24023 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
24024 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
24026 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24028 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
24029 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
24031 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24033 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
24035 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
24036 * config/arm/cortex-a7.md: New bypasses using
24037 arm_mac_accumulator_is_result.
24039 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24041 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
24042 (cortex_a7_neon_mla): Likewise.
24043 (cortex_a7_fpfmad): New reservation.
24044 (cortex_a7_fpmacs): Use ffmas and update required units.
24045 (cortex_a7_fpmuld): Update required units and latency.
24046 (cortex_a7_fpmacd): Likewise.
24047 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
24048 (cortex_a7_neon). Likewise.
24049 (bypass) Update participating units.
24051 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24053 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
24054 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
24056 * config/arm/vfp11.md (vfp_farith): Use ffmas.
24057 (vfp_fmul): Use ffmad.
24058 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
24059 (cortex_r4_fmacd): Use ffmad.
24060 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
24061 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
24062 (cortex_a9_fmacd): Use ffmad.
24063 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
24064 (cortex_a8_vfp_macd): Use ffmad.
24065 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
24066 (cortex_a5_fpmacd): Use ffmad.
24067 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
24068 (cortex_a15_vfp_macd): Use ffmad.
24069 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
24071 2013-01-29 Jason Merrill <jason@redhat.com>
24074 * varasm.c (default_assemble_visibility): Don't warn about
24075 visibility on artificial decls.
24077 2013-01-29 Richard Biener <rguenther@suse.de>
24079 PR tree-optimization/56113
24080 * tree-ssa-structalias.c (equiv_class_lookup): Also return
24082 (label_visit): Free duplicate bitmaps and record the leader instead.
24083 (perform_var_substitution): Adjust.
24085 2013-01-29 Richard Biener <rguenther@suse.de>
24087 PR tree-optimization/55270
24088 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
24089 the CFG, schedule loops for fixup.
24091 2013-01-29 Nick Clifton <nickc@redhat.com>
24093 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
24096 2013-01-28 Leif Ekblad <leif@rdos.net>
24098 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
24099 * config/i386/i386.h (TARGET_RDOS): New macro.
24100 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
24101 * config/i386/i386.c (ix86_option_override_internal): For 64bit
24102 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
24103 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
24104 DEFAULT_LARGE_SECTION_THRESHOLD.
24105 * config/i386/i386.md (R14_REG, R15_REG): New constants.
24106 * config/i386/rdos.h: New file.
24107 * config/i386/rdos64.h: New file.
24109 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
24112 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
24115 2013-01-28 Jakub Jelinek <jakub@redhat.com>
24117 PR rtl-optimization/56117
24118 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
24119 call cselib_lookup_from_insn on the MEM before calling
24120 add_insn_mem_dependence.
24122 2013-01-28 Richard Biener <rguenther@suse.de>
24124 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
24125 to a stmt that didn't have one.
24126 (copy_phis_for_bb): Likewise for PHI arguments.
24127 (copy_debug_stmt): Likewise for debug stmts.
24129 2013-01-28 Richard Biener <rguenther@suse.de>
24131 PR tree-optimization/56034
24132 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
24133 (partition_builtin_p): Adjust.
24134 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
24135 it is the last partition.
24136 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
24137 up the vertex for the definition.
24138 (classify_partition): Classify whether a partition is a
24139 PKIND_REDUCTION, thus has uses outside of the loop.
24140 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
24141 Merge all PKIND_REDUCTION partitions into the last partition.
24142 (tree_loop_distribution): Seed partitions from reductions as well.
24144 2013-01-28 Jakub Jelinek <jakub@redhat.com>
24146 PR tree-optimization/56125
24147 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
24148 pow(x,c) into sqrt(x) * powi(x, n/2) or
24149 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
24150 optimizing for size.
24151 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
24152 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
24155 PR tree-optimization/56094
24156 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
24157 to UNKNOWN_LOCATION while gimplifying expr.
24159 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
24162 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
24163 operand 0 in movabs insn template for -masm=intel asm alternative.
24164 (*movabs<mode>_2): Ditto for operand 1.
24166 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
24169 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
24172 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24174 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
24177 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24179 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
24180 Undef to avoid warning.
24182 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24184 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
24185 * configure: Regenerate.
24187 2013-01-25 Jakub Jelinek <jakub@redhat.com>
24189 PR tree-optimization/56098
24190 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
24191 for stmts with volatile ops.
24192 (cond_store_replacement): Don't optimize if assign has volatile ops.
24193 (cond_if_else_store_replacement_1): Don't optimize if either
24194 then_assign or else_assign have volatile ops.
24195 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
24198 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
24200 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
24202 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
24204 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
24205 missing ':' in asm example.
24207 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
24209 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
24210 entries into lane and laneq entries.
24211 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
24212 Remove AdvSIMD scalar modes.
24213 (aarch64_sq<r>dmulh_laneq<mode>): New.
24214 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
24216 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
24217 builtin implementations to relfect changes in RTL in aarch64-simd.md.
24218 * config/aarch64/iterators.md (VCOND): New.
24221 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
24224 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
24225 Add NULL LIBNAME argument to existing definitions.
24226 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
24227 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
24228 * config/avr/avr.c (DEF_BUILTIN): Same.
24229 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
24230 (avr_expand_builtin): Expand to a vanilla call if a libgcc
24231 implementation is available (DECL_ASSEMBLER_NAME is set).
24232 (avr_fold_absfx): New static function.
24233 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
24234 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
24235 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
24236 AVR_BUILTIN_ABSLLK.
24237 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
24238 (abshk, absk, abslk, absllk): Provide as static inline functions.
24240 2013-01-25 Marek Polacek <polacek@redhat.com>
24242 PR tree-optimization/56035
24243 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
24245 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
24247 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
24248 (*movtf_internal_rex64): Add (!o,C) alternative
24249 (*movxf_internal_rex64): Ditto.
24250 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
24252 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
24254 * doc/invoke.texi: fix typo.
24255 * doc/objc.texi: fix typo.
24257 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
24259 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
24260 for the first two alternatives.
24262 2013-01-24 Diego Novillo <dnovillo@google.com>
24264 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
24265 (ggc-zone.o): Remove.
24266 * configure.ac: Remove option --with-gc.
24267 * configure: Re-generate.
24268 * doc/install.texi: Remove documentation for --with-gc.
24269 * gengtype.c (write_enum_defn): Remove. Update all users.
24270 (write_Types_process_field): Remove generation of gt_e_* argument.
24271 (output_type_enum): Remove. Update all users.
24272 (write_enum_defn): Remove. Update all users.
24273 (enum alloc_zone): Remove. Update all users.
24274 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
24275 * ggc-common.c (ggc_splay_alloc): Remove first argument.
24276 Update all callers.
24277 (struct ptr_data): Remove field TYPE. Update all users.
24278 (gt_pch_note_object): Remove argument TYPE. Update all users.
24279 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
24281 * ggc-none.c (ggc_alloc_typed_stat): Remove.
24282 (struct alloc_zone): Remove.
24283 (ggc_internal_alloc_zone_stat): Remove.
24284 (ggc_internal_cleared_alloc_zone_stat): Remove.
24285 * ggc-page.c (ggc_alloc_typed_stat): Remove.
24286 (ggc_pch_count_object): Remove last argument. Update all users.
24287 (ggc_pch_alloc_object): Remove last argument. Update all users.
24288 (struct alloc_zone): Remove.
24289 * ggc-zone.c: Remove.
24290 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
24291 (struct alloc_zone): Remove.
24292 (ggc_alloc_typed_stat): Remove.
24293 (ggc_alloc_typed): Remove.
24294 (ggc_splay_alloc): Remove first argument.
24295 (rtl_zone): Remove. Update all users.
24296 (tree_zone): Remove. Update all users.
24297 (tree_id_zone): Remove. Update all users.
24298 (ggc_internal_zone_alloc_stat): Remove. Update all users.
24299 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
24300 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
24301 * tree-ssanames.c: Remove references to zone allocator in comments.
24303 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
24305 * config/avr/avr.c (avr_out_fract): Make register numbers that
24306 might be outside of source operand signed.
24308 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
24310 * config/i386/constraints.md (Yf): New constraint.
24311 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
24312 of f constraint to conditionaly disable x87 register preferences.
24313 (*movdf_internal): Ditto.
24314 (*movsf_internal): Ditto.
24316 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
24318 PR inline-asm/55934
24319 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
24320 that have operands with impossible constraints.
24321 Add a FIXME for a speed-up opportunity.
24322 * lra-constraints.c (process_alt_operands): Verify that a class
24323 selected from constraints on asms is valid for the operand mode.
24324 (curr_insn_transform): Remove incorrect comment.
24326 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
24328 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
24329 TOC operand is a valid symbol ref in the constant pool.
24331 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24333 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
24335 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
24338 * config/avr/stdfix.h: New file.
24339 * t-avr (stdfix-gcc.h): New rule to build it.
24340 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
24342 2013-01-23 Kostya Serebryany <kcc@google.com>
24344 * config/darwin.h: remove dependency on
24345 CoreFoundation (asan on Mac OS).
24347 2013-01-23 Jakub Jelinek <jakub@redhat.com>
24350 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
24351 instead of cmpdi_operand for first comparison operand.
24352 Don't assert that comparison operands aren't both constants.
24354 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
24356 * doc/install.texi (Downloading the Source): Update references to
24357 downloading separate components.
24359 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
24361 * doc/extend.texi (__int128): Improve grammar.
24363 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
24366 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
24367 alternative to (o,r).
24368 (*movdi_internal_rex64): Remove (!o,n) alternative.
24369 (DImode immediate->memory splitter): Remove.
24370 (DImode immediate->memory peephole2): Remove.
24371 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
24372 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
24373 alternative to (!o,*r).
24374 (*movtf_internal_sse): New pattern.
24375 (*movxf_internal_rex64): New pattern.
24376 (*movxf_internal): Disable for TARGET_64BIT.
24377 (*movdf_internal_rex64): Remove (!o,F) alternative.
24379 2013-01-22 Jakub Jelinek <jakub@redhat.com>
24381 PR middle-end/56074
24382 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
24383 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
24384 * tree-vect-loop-manip.c (find_loop_location): Also ignore
24385 stmt locations where LOCATION_LOCUS of the stmt location is
24386 UNKNOWN_LOCATION or BUILTINS_LOCATION.
24389 * config/i386/i386.md (UNSPEC_STOS): New.
24390 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
24391 *strsetqi_1): Add UNSPEC_STOS.
24393 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
24396 * doc/invoke.texi: Remove left over -Wsynth example.
24398 2013-01-21 Jakub Jelinek <jakub@redhat.com>
24400 PR tree-optimization/56051
24401 * fold-const.c (fold_binary_loc): Don't fold
24402 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
24403 a narrowing conversion, or widening conversion from signed
24406 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
24408 PR rtl-optimization/56023
24409 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
24410 dependent on debug instruction.
24412 2013-01-21 Martin Jambor <mjambor@suse.cz>
24414 PR middle-end/56022
24415 * function.c (allocate_struct_function): Call
24416 invoke_set_current_function_hook earlier.
24418 2013-01-21 Jakub Jelinek <jakub@redhat.com>
24420 * reload1.c (init_reload): Only initialize reload_obstack
24421 during the first call.
24423 2013-01-21 Marek Polacek <polacek@redhat.com>
24425 * cfgloop.c (verify_loop_structure): Fix up grammar.
24427 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
24429 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
24430 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
24432 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24435 * config/arm/marvell-pj4.md: Update copyright year.
24436 Fix up use of alu to alu_reg and simple_alu_imm.
24438 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
24440 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
24442 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
24445 * lra-constraints.c (curr_insn_transform): Don't reuse original
24446 insn for secondary memory move when memory mode should be different.
24448 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24450 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
24451 atomic_storedi_1): New patterns.
24453 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
24455 btver2 pipeline descriptions.
24456 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
24458 * config/i386/i386.md (btver2_decode): New type attributes.
24459 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
24461 * config/i386/btver2.md: New file describing btver2 pipelines.
24463 2013-01-19 Andrew Pinski <apinski@cavium.com>
24465 PR tree-optimization/52631
24466 * tree-ssa-sccvn (visit_use): Before looking up the original
24467 statement, try looking up the simplified expression.
24469 2013-01-19 Anthony Green <green@moxielogic.com>
24471 * config/moxie/moxie.c (moxie_expand_prologue): Set
24472 current_function_static_stack_size.
24474 2013-01-18 Jakub Jelinek <jakub@redhat.com>
24476 PR tree-optimization/56029
24477 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
24478 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
24480 2013-01-18 Sharad Singhai <singhai@google.com>
24482 PR tree-optimization/55995
24483 * dumpfile.c (dump_loc): Print location only if available.
24484 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
24486 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
24489 * lra-constraints.c (curr_insn_transform): Reuse original insn for
24490 secondary memory move.
24491 (inherit_reload_reg): Use rclass instead of cl for
24492 check_secondary_memory_needed_p.
24494 2013-01-18 Jakub Jelinek <jakub@redhat.com>
24496 PR middle-end/56015
24497 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
24498 the case where writing real complex part of target modifies op1.
24500 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
24502 * config/aarch64/aarch64-simd.md
24503 (aarch64_vcond_internal<mode>): Handle unordered cases.
24504 * config/aarch64/iterators.md (v_cmp_result): New.
24506 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
24507 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24509 * config/arm/marvell-pj4.md: New file.
24510 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
24511 * config/arm/arm.md (generic_sched): Add marvell_pj4.
24512 (generic_vfp): Likewise.
24513 * config/arm/arm-cores.def: Add marvell-pj4.
24514 * config/arm/arm-tune.md: Regenerate.
24515 * config/arm/arm-tables.opt: Regenerate.
24516 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
24517 * doc/invoke.texi: Document marvell-pj4.
24519 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
24521 * config/aarch64/arm_neon.h: Map scalar types to standard types.
24523 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24528 * var-tracking.c (negative_power_of_two_p): New.
24529 (global_get_addr_cache, local_get_addr_cache): New.
24530 (get_addr_from_global_cache, get_addr_from_local_cache): New.
24531 (vt_canonicalize_addr): Rewrite using the above. Adjust the
24533 (vt_stack_offset_p): Remove.
24534 (vt_canon_true_dep): Always canonicalize loc's address.
24535 (clobber_overlapping_mems): Make sure we have a MEM.
24536 (local_get_addr_clear_given_value): New.
24537 (val_reset): Clear local cached entries.
24538 (compute_bb_dataflow): Create and release the local cache.
24539 Disable duplicate MEMs clobbering.
24540 (emit_notes_in_bb): Clobber MEMs likewise.
24541 (vt_emit_notes): Create and release the local cache.
24542 (vt_initialize, vt_finalize): Create and release the global
24543 cache, respectively.
24544 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
24546 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24548 PR libmudflap/53359
24549 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
24550 not found in the symtab.
24552 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24555 PR rtl-optimization/55547
24556 PR rtl-optimization/53827
24559 * alias.c (offset_overlap_p): New, factored out of...
24560 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
24561 the conservative special case for symbolic constants. Don't
24562 adjust zero sizes on alignment.
24564 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
24566 PR rtl-optimization/52573
24567 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
24568 REG_UNUSED for the same register.
24570 2013-01-17 Richard Biener <rguenther@suse.de>
24571 Marek Polacek <polacek@redhat.com>
24573 PR rtl-optimization/55833
24574 * loop-unswitch.c (unswitch_loops): Move loop verification...
24575 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
24576 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
24577 Set it to true when we're removing a loop from hierarchy tree in
24578 an irreducible region.
24579 (fix_bb_placements): Adjust caller.
24580 (fix_loop_placements): Likewise.
24582 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
24584 * config/avr/builtins.def (DEF_BUILTIN): Factor out
24585 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
24586 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
24587 Remove ID. Adjust comments.
24588 * config/avr/avr-c.c (avr_builtin_name): Remove.
24589 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
24590 * config/avr/avr.c (avr_tolower): New static function.
24591 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
24592 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
24593 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
24596 2013-01-17 Jan Hubicka <jh@suse.cz>
24598 PR tree-optimization/55273
24599 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
24601 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
24604 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
24605 store through atomic_store<mode>_1.
24606 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
24608 2013-01-17 Martin Jambor <mjambor@suse.cz>
24610 PR tree-optimizations/55264
24611 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
24612 for virtual methods.
24613 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
24614 virtual methods before inlining is over.
24615 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
24617 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
24620 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
24622 PR rtl-optimization/56005
24623 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
24624 pending reads for prefetch.
24626 2013-01-16 Ian Bolton <ian.bolton@arm.com>
24628 * config/aarch64/aarch64.md
24629 (*cstoresi_neg_uxtw): New pattern.
24630 (*cmovsi_insn_uxtw): New pattern.
24631 (*<optab>si3_uxtw): New pattern.
24632 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
24633 (*<optab>si3_insn_uxtw): New pattern.
24634 (*bswapsi2_uxtw): New pattern.
24636 2013-01-16 Richard Biener <rguenther@suse.de>
24638 * tree-inline.c (tree_function_versioning): Remove set but
24639 never used variable.
24641 2013-01-16 Richard Biener <rguenther@suse.de>
24643 PR tree-optimization/55964
24644 * tree-flow.h (rename_variables_in_loop): Remove.
24645 (rename_variables_in_bb): Likewise.
24646 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
24647 (copy_loop_before): Adjust and delete update-ssa status.
24648 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
24649 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
24650 (rename_variables_in_loop): Remove.
24651 (slpeel_update_phis_for_duplicate_loop): Likewise.
24652 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
24653 use available cfg machinery instead of duplicating it.
24654 Update PHI nodes and perform poor-mans SSA update here.
24655 (slpeel_tree_peel_loop_to_edge): Adjust.
24657 2013-01-16 Richard Biener <rguenther@suse.de>
24659 PR tree-optimization/54767
24660 PR tree-optimization/53465
24661 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
24662 (vrp_visit_phi_node): For PHI arguments coming via backedges
24663 drop all symbolical range information.
24664 (execute_vrp): Compute backedges.
24666 2013-01-16 Richard Biener <rguenther@suse.de>
24668 * doc/install.texi: Update CLooG and ISL requirements to
24671 2013-01-16 Christian Bruel <christian.bruel@st.com>
24674 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
24675 (broken_move): Handle UNSPECV_SP_SWITCH_B.
24676 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
24678 2013-01-16 DJ Delorie <dj@redhat.com>
24680 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
24681 (UNSPECV_SP_SWITCH_E): New.
24682 (sp_switch_1): Change to an unspec.
24683 (sp_switch_2): Change to an unspec. Don't use post-inc when we
24686 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
24688 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
24689 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
24690 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
24691 (expand_mem_thread_fence): Ditto.
24692 (expand_mem_signal_fence): Ditto.
24693 (expand_atomic_load): Ditto.
24694 (expand_atomic_store): Ditto.
24696 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
24698 PR rtl-optimization/55547
24699 PR rtl-optimization/53827
24702 * alias.c (memrefs_conflict_p): Set sizes to negative after
24705 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24708 * function.c (thread_prologue_and_epilogue_insns): Always
24709 add crtl->drap_reg to set_up_by_prologue.set, even if
24710 stack_realign_drap is false.
24712 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
24714 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
24715 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
24716 *call): Fix indention.
24718 2013-01-15 Tom de Vries <tom@codesourcery.com>
24721 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
24724 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
24726 PR rtl-optimization/55153
24727 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
24729 2013-01-15 Martin Jambor <mjambor@suse.cz>
24731 PR tree-optimization/55920
24732 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
24733 accesses as grp_to_be_debug_replaced.
24735 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24737 PR tree-optimization/55920
24738 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
24739 there is non-useless type conversion needed from debug rhs to lhs,
24740 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
24742 2013-01-15 Joseph Myers <joseph@codesourcery.com>
24743 Mikael Pettersson <mikpe@it.uu.se>
24746 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
24748 (ASM_OUTPUT_CASE_LABEL): Remove.
24749 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
24750 * final.c (shorten_branches): Update alignment of labels before
24751 jump tables if CASE_VECTOR_SHORTEN_MODE.
24753 2013-01-15 Richard Biener <rguenther@suse.de>
24756 * system.h: Do not include gmp.h for building host tools.
24758 2013-01-15 Richard Biener <rguenther@suse.de>
24760 PR middle-end/55882
24761 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
24762 account for bitpos when computing alignment.
24764 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
24766 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
24767 (ix86_target_macros_internal): Likewise.
24769 * config/i386/i386.c (m_CORE2I7): Removed.
24770 (m_CORE_HASWELL): New macro.
24771 (m_CORE_ALL): Likewise.
24772 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
24773 (initial_ix86_arch_features): Likewise.
24774 (processor_target_table): Initializations for Core avx2.
24775 (cpu_names): New names "core-avx2".
24776 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
24777 PROCESSOR_CORE_HASWELL.
24778 (ix86_issue_rate): New case.
24779 (ia32_multipass_dfa_lookahead): Likewise.
24780 (ix86_sched_init_global): Likewise.
24782 * config/i386/i386.h (TARGET_HASWELL): New macro.
24783 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
24784 (processor_type): New PROCESSOR_HASWELL.
24786 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24788 PR tree-optimization/55955
24789 * tree-vect-loop.c (vectorizable_reduction): Give up early on
24790 *SHIFT_EXPR and *ROTATE_EXPR codes.
24792 PR tree-optimization/48766
24793 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
24794 -ftrapv disable -fwrapv.
24796 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24799 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
24800 etc. to 1 and not to __flash.
24801 Use LL suffix for __INT24_MAX__ with -mint8.
24802 Use ULL suffix for __UINT24_MAX__ with -mint8.
24804 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24806 * config/avr/avr-arch.h
24807 (struct base_arch_s): Use typedef avr_arch_t instead.
24808 (struct arch_info_s): Use typedef avr_arch_info_t instead.
24809 (struct mcu_type_s): Use typedef avr_mcu_t instead.
24810 * config/avr/avr.c: Same.
24811 * config/avr/avr-devices.c: Same.
24812 * config/avr/driver-avr.c: Same.
24813 * config/avr/gen-avr-mmcu-texi.c: Same.
24814 * config/avr/avr-mcus.def: Adjust comment.
24816 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
24818 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
24819 * config/aarch64/iterators.md (VALLDI): New.
24821 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
24822 Andi Kleen <ak@linux.intel.com>
24825 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
24826 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
24829 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24831 * config/avr/avr-stdint.h: Remove trailing blanks.
24832 * config/avr/avr-log.h: Same.
24833 * config/avr/avr-arch.h: Same.
24834 * config/avr/avr-devices.c: Same.
24835 * config/avr/avr-dimode.md: Same.
24836 * config/avr/predicates.md: Same.
24837 * config/avr/avr-c.c: Same. And fix typo.
24839 * config/avr/avr-protos.h: Same. And:
24840 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
24841 (init_cumulative_args): Rename to avr_init_cumulative_args.
24842 (expand_prologue): Rename to avr_expand_prologue.
24843 (expand_epilogue): Rename to avr_expand_epilogue.
24844 (adjust_insn_length): Rename to avr_adjust_insn_length.
24845 (notice_update_cc): Rename to avr_notice_update_cc.
24846 (final_prescan_insn): Rename to avr_final_prescan_insn.
24847 * config/avr/avr.c: Same.
24848 * config/avr/avr.h: Same.
24849 * config/avr/avr.md: Remove trailing blanks.
24850 (prologue): Use avr_expand_prologue.
24851 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
24853 2013-01-14 Richard Biener <rguenther@suse.de>
24855 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
24856 verify_location, collect_subblocks): New functions.
24857 (verify_gimple_in_cfg): Verify that locations only reference
24858 BLOCKs in the functions BLOCK tree.
24860 2013-01-14 Richard Biener <rguenther@suse.de>
24862 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
24864 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
24866 (insert_out_of_ssa_copy_on_edge): Likewise.
24867 (rewrite_close_phi_out_of_ssa): Likewise.
24868 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
24870 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
24871 propagated constants.
24872 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
24875 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24877 * config/avr/avr-modes.def: Add GPL copyright notice.
24879 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
24881 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
24882 MEMMODEL_MASK to determine memory model.
24883 (atomic_store<mode>): Ditto from operands[2].
24884 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
24886 2013-01-13 Jakub Jelinek <jakub@redhat.com>
24889 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
24890 (fold_gimple_assign): Don't call unshare_expr here.
24891 (fold_ctor_reference): Call unshare_expr.
24893 2013-01-13 Terry Guo <terry.guo@arm.com>
24895 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
24896 * doc/fragments.texi: Document MULTILIB_REUSE.
24897 * gcc.c (multilib_reuse): New internal spec.
24898 (set_multilib_dir): Also search multilib from multilib_reuse.
24899 * genmultilib (tmpmultilib3): Refactor code.
24900 (tmpmultilib4): Ditto.
24901 (multilib_reuse): New multilib argument.
24903 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
24905 * Makefile.in: Update copyright.
24907 2013-01-12 Tom de Vries <tom@codesourcery.com>
24909 PR middle-end/55890
24910 * calls.c (expand_call): Check if arg_nr is valid.
24912 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
24914 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
24915 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
24916 documentation. Add missing '__' in front of
24917 __builtin_ia32_packssdw256.
24919 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24922 * config/s390/s390.c (s390_preferred_reload_class): Do not return
24923 NO_REGS for larl operands.
24924 (s390_reload_larl_operand): Use s390_load_address instead of
24927 2013-01-11 Richard Biener <rguenther@suse.de>
24929 * tree-cfg.c (verify_node_sharing_1): Split out from ...
24930 (verify_node_sharing): ... here.
24931 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
24933 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
24935 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
24936 Substitute TREECHECKING.
24937 * configure: Regenerate.
24938 * Makefile.in (TREECHECKING): New.
24940 2013-01-11 Richard Guenther <rguenther@suse.de>
24942 PR tree-optimization/44061
24943 * tree-vrp.c (extract_range_basic): Compute zero as
24944 value-range for __builtin_constant_p of function parameters.
24946 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
24948 Update copyright years.
24950 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
24952 PR rtl-optimization/55672
24953 * lra-eliminations.c (mark_not_eliminable): Permit addition with
24954 const to be eliminable.
24956 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
24958 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
24959 * configure: Regenerate.
24961 2013-01-10 Richard Biener <rguenther@suse.de>
24963 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
24965 2013-01-10 Richard Biener <rguenther@suse.de>
24968 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
24969 locations for virtual PHI arguments.
24970 (rewrite_update_phi_arguments): Likewise.
24972 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
24974 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
24977 2013-01-10 Jakub Jelinek <jakub@redhat.com>
24979 PR tree-optimization/55921
24980 * tree-complex.c (expand_complex_asm): New function.
24981 (expand_complex_operations_1): Call it for GIMPLE_ASM.
24983 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24986 * config/s390/s390.c (s390_symref_operand_p)
24987 (s390_loadrelative_operand_p): Merge the two functions.
24988 (s390_check_qrst_address, print_operand_address): Add parameters
24989 to s390_loadrelative_operand_p invokation.
24990 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
24991 (s390_reload_larl_operand, s390_secondary_reload): Use
24992 s390_loadrelative_operand_p instead of s390_symref_operand_p.
24993 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
24995 2013-01-09 Mike Stump <mikestump@comcast.net>
24997 * dse.c (record_store): Remove unnecessary assert.
24999 2013-01-09 Jan Hubicka <jh@suse.cz>
25001 PR tree-optimization/55569
25002 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
25003 * cfgloop.h (scale_loop_profile): Likewise.
25005 2013-01-09 Jan Hubicka <jh@suse.cz>
25008 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
25010 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
25012 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
25014 PR middle-end/55114
25015 * expr.h (maybe_emit_group_store): Declare.
25016 * expr.c (maybe_emit_group_store): New function.
25017 * builtins.c (expand_builtin_int_roundingfn): Call it.
25018 (expand_builtin_int_roundingfn_2): Likewise.
25020 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
25022 PR rtl-optimization/55829
25023 * lra-constraints.c (match_reload): Add code for absent output.
25024 (curr_insn_transform): Add code for reloads of matched inputs
25027 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
25029 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
25030 attribute of movddup insn to DF.
25031 (*vec_interleave_lowv2df): Ditto.
25032 (vec_dupv2df): Ditto.
25034 2013-01-09 Jan Hubicka <jh@suse.cz>
25036 PR tree-optimiation/55875
25037 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
25038 EVERY_ITERATION parameter.
25039 (number_of_iterations_exit): Check if exit is executed every iteration.
25040 (idx_infer_loop_bounds): Similarly here.
25041 (n_of_executions_at_most): Simplify
25042 to only test for cases where statement is dominated by the
25043 particular bound; handle correctly the "postdominance" test.
25044 (scev_probably_wraps_p): Use max loop iterations info
25045 as a global bound first.
25047 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
25048 Nick Clifton <nickc@redhat.com>
25050 * config/v850/v850.md (cbranchsf4): New pattern.
25051 (cstoresf4): New pattern.
25052 (cbranchdf4): New pattern.
25053 (cstoredf4): New pattern.
25054 (movsicc): Disallow floating point comparisons.
25055 (cmpsf_le_insn): Fix order of operators.
25056 (cmpsf_lt_insn): Likewise.
25057 (cmpsf_eq_insn): Likewise.
25058 (cmpdf_le_insn): Likewise.
25059 (cmpdf_lt_insn): Likewise.
25060 (cmpdf_eq_insn): Likewise.
25061 (cmpsf_ge_insn): Use LE comparison.
25062 (cmpdf_ge_insn): Likewise.
25063 (cmpsf_gt_insn): Use LT comparison.
25064 (cmpdf_gt_insn): Likewise.
25065 (cmpsf_ne_insn): Delete pattern.
25066 (cmpdf_ne_insn): Delete pattern.
25067 * config/v850/v850.c (v850_gen_float_compare): Use
25068 gen_cmpdf_eq_insn for NE comparison.
25069 (v850_float_z_comparison_operator)
25070 (v850_float_nz_comparison_operator): Move from here ...
25071 * config/v850/predicates.md: ... to here. Move GT and GE
25072 comparisons into v850_float_z_comparison_operator.
25073 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
25075 (v850_float_nz_comparison_operator): Likewise.
25077 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25079 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
25080 with calls to gen_insvsi/gen_insvdi.
25082 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25084 * config/i386/i386.c (initial_ix86_tune_features): Set up
25085 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
25087 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
25088 Jakub Jelinek <jakub@redhat.com>
25090 PR tree-optimization/48189
25091 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
25092 If nitercst is 0, don't predict the exit edge.
25094 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25096 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
25097 in asm_fprintf with reg_names.
25098 (aarch64_print_operand_address): Likewise.
25099 (aarch64_return_addr): Likewise.
25100 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
25102 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25104 * config/pa/pa.h (VAL_U6_BITS_P): Define.
25105 (INT_U6_BITS): Likewise.
25106 * config/pa/predicates.md (uint6_operand): New predicate.
25107 (shift5_operand, shift6_operand): Likewise.
25108 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
25110 (lshrdi3): Use shift6_operand.
25111 (shrpsi4, shrpdi4): New insn patterns.
25112 (extzv): Delete expander.
25113 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
25114 predicates in unamed zero extract patterns. Tighten common constraint.
25115 (extv): Delete expander.
25116 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
25117 predicates in unamed sign extract patterns. Tighten common constraint.
25118 (insv): Delete expander.
25119 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
25120 predicates in unamed insert patterns. Tighten common constraint.
25121 Change uint32_operand predicate to uint6_operand predicate in unamed
25122 DImode pattern to insert constant values of type 1...1xxxx.
25124 2013-01-04 Jan Hubicka <jh@suse.cz>
25126 PR tree-optimization/55823
25127 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
25130 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25131 Uros Bizjak <ubizjak@gmail.com>
25133 PR rtl-optimization/55845
25134 * df-problems.c (can_move_insns_across): Stop scanning at
25135 volatile_insn_p source instruction or give up if
25136 across_from .. across_to range contains any volatile_insn_p
25139 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
25141 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
25142 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
25144 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
25145 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
25147 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25150 * asan.c (asan_clear_shadow): New function.
25151 (asan_emit_stack_protection): Use it.
25153 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
25155 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
25156 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
25157 with tab instead of space.
25159 2013-01-08 Nick Clifton <nickc@redhat.com>
25161 * config/rl78/rl78.c (rl78_expand_prologue): Always select
25162 register bank 0 at the start of an interrupt handler.
25163 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
25166 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
25168 * config/aarch64/aarch64-simd.md
25169 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
25170 (aarch64_simd_bsl): Likewise.
25171 (aarch64_vcond_internal<mode>): Likewise.
25172 (vcond<mode><mode>): Likewise.
25173 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
25174 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
25176 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
25178 * config/aarch64/aarch64-builtins.c
25179 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
25181 2013-01-08 Martin Jambor <mjambor@suse.cz>
25184 * tree-sra.c (analyze_access_subtree): Return true also after
25185 potentially creating a debug-only replacement.
25187 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25189 PR middle-end/55890
25190 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
25192 PR tree-optimization/54120
25193 * tree-vrp.c (range_fits_type_p): Don't allow
25194 src_precision < precision from signed vr to unsigned_p
25195 if vr->min or vr->max is negative.
25196 (simplify_float_conversion_using_ranges): Test can_float_p
25197 against CODE_FOR_nothing.
25199 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25200 Richard Biener <rguenther@suse.de>
25202 PR middle-end/55851
25203 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
25204 types instead of just INTEGER_TYPE types.
25206 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
25208 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
25211 2013-01-07 Steve Ellcey <sellcey@mips.com>
25214 * config/mips/mips.opt: Change mad to mmad to match documentation.
25216 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25219 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
25220 .progmemx.data now.
25222 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25225 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
25226 (avr_addrspace_t): Add .section_name field.
25227 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
25229 (avr_addrspace): Same. Initialize .section_name. Remove last
25230 NULL entry. Put __memx into .progmemx.data.
25231 (progmem_section_prefix): Remove.
25232 (avr_asm_init_sections): No need to initialize progmem_section.
25233 (avr_asm_named_section): Use avr_addrspace[].section_name to get
25234 section name prefix.
25235 (avr_asm_select_section): Ditto. And use get_unnamed_section to
25236 retrieve the progmem section.
25237 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
25238 boundary to run over avr_addrspace[].
25239 (avr_register_target_pragmas): Ditto.
25241 2013-01-06 Jakub Jelinek <jakub@redhat.com>
25243 * varasm.c (output_constant_def_contents): For asan_protect_global
25244 protected strings, adjust DECL_ALIGN if needed, before testing for
25246 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
25247 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
25249 (output_object_block): For asan protected decls, emit asan padding
25250 after their contents.
25251 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
25252 (asan_finish_file): Test it here instead.
25254 2013-01-07 Nick Clifton <nickc@redhat.com>
25255 Matthias Klose <doko@debian.org>
25256 Doug Kwan <dougkwan@google.com>
25257 H.J. Lu <hongjiu.lu@intel.com>
25260 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
25262 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
25264 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
25266 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
25268 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
25270 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25273 * doc/install.texi (Cross-Compiler-Specific Options): Document
25276 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
25278 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
25279 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
25280 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
25281 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
25282 vqmovun_high_s64): Fix source operand number and update copyright.
25284 2013-01-07 Richard Biener <rguenther@suse.de>
25286 PR middle-end/55890
25287 * gimple.h (gimple_call_builtin_p): New overload.
25288 * gimple.c (validate_call): New function.
25289 (gimple_call_builtin_p): Likewise.
25290 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25291 Use gimple_call_builtin_p.
25292 (find_func_clobbers): Likewise.
25293 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
25294 (strlen_optimize_stmt): Likewise.
25296 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
25298 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
25299 (vld1q_dup_*): Likewise.
25300 (vld1_*): Likewise.
25301 (vld1q_*): Likewise.
25302 (vld1_lane_*): Likewise.
25303 (vld1q_lane_*): Likewise.
25305 2013-01-07 Richard Biener <rguenther@suse.de>
25307 * lto-streamer.h (LTO_minor_version): Bump to 2.
25309 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
25311 * config/aarch64/aarch64-protos.h
25312 (aarch64_const_double_zero_rtx_p): Rename to...
25313 (aarch64_float_const_zero_rtx_p): ...this.
25314 (aarch64_float_const_representable_p): New.
25315 (aarch64_output_simd_mov_immediate): Likewise.
25316 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
25317 move immediate case.
25318 * config/aarch64/aarch64.c
25319 (aarch64_const_double_zero_rtx_p): Rename to...
25320 (aarch64_float_const_zero_rtx_p): ...this.
25321 (aarch64_print_operand): Allow printing of new constants.
25322 (aarch64_valid_floating_const): New.
25323 (aarch64_legitimate_constant_p): Check for valid floating-point
25325 (aarch64_simd_valid_immediate): Likewise.
25326 (aarch64_vect_float_const_representable_p): New.
25327 (aarch64_float_const_representable_p): Likewise.
25328 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
25329 (aarch64_output_simd_mov_immediate): New.
25330 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
25331 (*movdf_aarch64): Likewise.
25332 * config/aarch64/constraints.md (Ufc): New.
25333 (Y): call aarch64_float_const_zero_rtx.
25334 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
25336 2013-01-07 Richard Biener <rguenther@suse.de>
25338 PR tree-optimization/55888
25339 PR tree-optimization/55862
25340 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
25341 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
25342 not if it is contained therein.
25344 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25346 * config/avr/t-avr: Typo.
25348 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25351 * config/avr/t-avr: Don't automatically rebuild
25352 $(srcdir)/config/avr/t-multilib
25353 $(srcdir)/config/avr/avr-tables.opt
25354 $(srcdir)/doc/avr-mmcu.texi
25355 (avr-mcus): New phony target to build them on request.
25356 (s-avr-mlib, s-avr-mmcu-texi): Remove.
25357 * avr/avr-mcus.def: Adjust comments.
25359 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
25361 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
25363 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
25365 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
25367 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
25369 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
25371 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
25373 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
25374 to generate profiling.
25375 * config/rs6000/aix64.h (LIB_SPEC): Same.
25377 2013-01-04 Andrew Pinski <apinski@cavium.com>
25379 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
25381 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
25383 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
25385 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
25387 (ix86_expand_move): Ditto.
25388 (ix86_zero_extend_to_Pmode): Ditto.
25389 (ix86_expand_call): Ditto.
25390 (ix86_expand_special_args_builtin): Ditto.
25391 (ix86_expand_builtin): Ditto.
25393 2013-01-04 Richard Biener <rguenther@suse.de>
25395 PR tree-optimization/55862
25396 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
25397 translating them through PHI nodes.
25399 2013-01-04 Martin Jambor <mjambor@suse.cz>
25401 PR tree-optimization/55755
25402 * tree-sra.c (sra_modify_assign): Do not check that an access has no
25403 children when trying to avoid producing a VIEW_CONVERT_EXPR.
25405 2013-01-04 Marek Polacek <polacek@redhat.com>
25407 PR middle-end/55859
25408 * opts.c (default_options_optimization): Clarify error message.
25410 2013-01-04 Richard Biener <rguenther@suse.de>
25412 PR middle-end/55863
25413 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
25416 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25419 * config/pa/pa.md (movsi): Revert previous change.
25420 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
25423 2013-01-03 Richard Henderson <rth@redhat.com>
25425 * config/i386/i386.c (ix86_expand_move): Always assign to op1
25426 after eliminating TLS symbols.
25428 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
25431 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
25432 * graphite-poly.c (debug_gmp_value): Likewise.
25434 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
25437 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
25438 selected code model, define __code_mode_small__, __code_model_medium__,
25439 __code_model_large__, __code_model_32__ or __code_model_kernel__.
25440 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
25441 xchg temporary register with %k. Declare temporary register as
25443 [__x86_64__]: For medium and large code models, preserve %rbx register.
25445 2013-01-03 Richard Biener <rguenther@suse.de>
25447 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
25448 (dump_subscript): Adjust.
25449 (finalize_ddr_dependent): Do not dump redundant info.
25450 (analyze_siv_subscript): Adjust.
25451 (subscript_dependence_tester): Likewise.
25452 (compute_affine_dependence): Likewise.
25454 2013-01-03 Richard Biener <rguenther@suse.de>
25457 2013-01-03 Richard Biener <rguenther@suse.de>
25459 PR tree-optimization/55857
25460 * tree-vect-stmts.c (vectorizable_load): Do not setup
25461 re-alignment for invariant loads.
25463 2013-01-02 Richard Biener <rguenther@suse.de>
25465 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
25466 invariant load do not generate a vector load from the scalar location.
25468 2013-01-03 Richard Biener <rguenther@suse.de>
25470 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
25471 for not vectorizing.
25472 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
25473 not build INDIRECT_REFs, call get_name once only.
25474 (vect_create_data_ref_ptr): Likewise. Dump base object kind
25475 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
25477 2013-01-03 Richard Biener <rguenther@suse.de>
25479 PR tree-optimization/55857
25480 * tree-vect-stmts.c (vectorizable_load): Do not setup
25481 re-alignment for invariant loads.
25483 2013-01-03 Richard Biener <rguenther@suse.de>
25486 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
25487 prefer a built-in decl.
25489 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25491 * gcc.c (process_command): Update copyright notice dates.
25492 * gcov.c (print_version): Likewise.
25493 * gcov-dump.c (print_version): Likewise.
25495 PR rtl-optimization/55838
25496 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
25497 iv0.step, iv1.step and step.
25499 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25500 Marc Glisse <marc.glisse@inria.fr>
25502 PR tree-optimization/55832
25503 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
25504 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
25505 integer_{one,zero}_node.
25507 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25510 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
25511 * var-tracking.c (reverse_op): Don't add reverse ops to
25512 VALUEs that have already
25513 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
25515 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
25517 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
25519 2013-01-02 Teresa Johnson <tejohnson@google.com>
25521 * dumpfile.c (dump_loc): Print filename with location.
25522 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
25523 new location_t parameter to emit complete unroll message with
25524 new dump framework.
25525 (canonicalize_loop_induction_variables): Compute loops location
25526 and pass to try_unroll_loop_completely.
25527 * loop-unroll.c (report_unroll_peel): New function.
25528 (peel_loops_completely): Use new dump format with location
25529 for main dumpfile message, and invoke report_unroll_peel on success.
25530 (decide_unrolling_and_peeling): Ditto.
25531 (decide_peel_once_rolling): Remove old dumpfile message subsumed
25532 by report_unroll_peel.
25533 (decide_peel_completely): Ditto.
25534 (decide_unroll_constant_iterations): Ditto.
25535 (decide_unroll_runtime_iterations): Ditto.
25536 (decide_peel_simple): Ditto.
25537 (decide_unroll_stupid): Ditto.
25538 * cfgloop.c (get_loop_location): New function.
25539 * cfgloop.h (get_loop_location): Declare.
25541 2013-01-02 Sriraman Tallam <tmsriram@google.com>
25543 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
25546 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25548 PR middle-end/55198
25549 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
25550 BLKmode objects when EXPAND_MEMORY is specified.
25552 2013-01-02 Sriraman Tallam <tmsriram@google.com>
25554 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
25556 (fold_builtin_cpu): Do not share cpu model decls across statements.
25558 2013-01-02 Jason Merrill <jason@redhat.com>
25561 * tree.c (build_array_type_1): Revert earlier change.
25563 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
25565 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
25567 * config/aarch64/aarch64-tune.md: Re-generate.
25569 2013-01-02 Richard Biener <rguenther@suse.de>
25571 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
25572 invariant load do not generate a vector load from the scalar location.
25574 2013-01-02 Richard Biener <rguenther@suse.de>
25577 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
25578 * configure: Regenerate.
25580 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
25582 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
25583 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
25584 (expand_builtin_int_roundingfn_2): Keep the original target around
25585 for the fallback case.
25587 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
25589 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
25590 to be clear for sign changes.
25592 2013-01-01 Jan Hubicka <jh@suse.cz>
25594 * ipa-inline-analysis.c: Fix formatting.
25596 2013-01-01 Jakub Jelinek <jakub@redhat.com>
25598 PR tree-optimization/55831
25599 * tree-vect-loop.c (get_initial_def_for_induction): Use
25600 gsi_after_labels instead of gsi_start_bb.
25602 Copyright (C) 2013 Free Software Foundation, Inc.
25604 Copying and distribution of this file, with or without modification,
25605 are permitted in any medium without royalty provided the copyright
25606 notice and this notice are preserved.