1 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3 Sergey Lega <sergey.s.lega@intel.com>
4 Anna Tikhonova <anna.tikhonova@intel.com>
5 Ilya Tocar <ilya.tocar@intel.com>
6 Andrey Turetskiy <andrey.turetskiy@intel.com>
7 Ilya Verbin <ilya.verbin@intel.com>
8 Kirill Yukhin <kirill.yukhin@intel.com>
9 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
14 (avx512fmaskmode): Ditto.
15 (bcstscalarsuff): Ditto.
16 (avx512f_blendm<mode>): Ditto.
17 (cmp_imm_predicate): Ditto.
18 (avx512f_cmp<mode>3): Ditto.
19 (avx512f_vec_dup<mode>): Ditto.
20 (avx512f_vec_dup_mem<mode>): Ditto.
21 (avx512f_vpermi2var<mode>3): Ditto.
22 (avx512f_vpermt2var<mode>3): Ditto.
23 (vec_init<mode>): Ditto.
24 (avx512f_gathersi<mode>): Ditto.
25 (*avx512f_gathersi<mode>): Ditto.
26 (*avx512f_gathersi<mode>_2): Ditto.
27 (avx512f_gatherdi<mode>): Ditto.
28 (*avx512f_gatherdi<mode>): Ditto.
29 (*avx512f_gatherdi<mode>_2): Ditto.
30 (avx512f_scattersi<mode>): Ditto.
31 (*avx512f_scattersi<mode>): Ditto.
32 (avx512f_scatterdi<mode>): Ditto.
33 (*avx512f_scatterdi<mode>): Ditto.
34 (sseintprefix): Extened with wider modes.
35 (VEC_GATHER_IDXSI): Ditto.
36 (VEC_GATHER_IDXDI): Ditto.
37 (VEC_GATHER_SRCDI): Ditto.
39 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
40 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
42 * config/arm/t-aprofile: New file.
43 * config.gcc: Handle --with-multilib-list option.
45 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
47 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
50 2013-10-15 Richard Biener <rguenther@suse.de>
52 * tree-loop-distribution.c (build_empty_rdg): Inline into
54 (rdg_flag_vertex): Inline into single user.
55 (rdg_flag_vertex_and_dependent): Likewise.
56 (build_rdg_partition_for_vertex): Remove processed bitmap.
57 (rdg_build_partitions): Simplify.
59 2013-10-15 Richard Biener <rguenther@suse.de>
61 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
62 Restructure forwarding through conversions and copies to
63 avoid performing copy-propagation the wrong way. Adjust
64 recursion invocations.
65 (forward_propagate_addr_expr): Add argument stating if we
66 are recursing from a single-use.
67 (ssa_forward_propagate_and_combine): Adjust.
69 2013-10-14 David Malcolm <dmalcolm@redhat.com>
71 * dumpfile.h (gcc::dump_manager): New class, to hold state
72 relating to dumpfile management.
73 (get_dump_file_name): Remove in favor of method of dump_manager.
74 (dump_initialized_p): Likewise.
75 (dump_start): Likewise.
76 (dump_finish): Likewise.
77 (dump_switch_p): Likewise.
78 (dump_register): Likewise.
79 (get_dump_file_info): Likewise.
80 * context.c (gcc::context::context): Construct the dump_manager
82 * context.h (gcc::context::get_dumps): New.
83 (gcc::context::m_dumps): New.
84 * coverage.c (coverage_init): Port to dump_manager API.
85 * dumpfile.c (extra_dump_files): Convert to field of
87 (extra_dump_files_in_use): Likewise.
88 (extra_dump_files_alloced): Likewise.
89 (gcc::dump_manager::dump_manager): New.
90 (dump_register): Convert to...
91 (gcc::dump_manager::dump_register): ...method, replacing
92 function-static next_dump with m_next_dump field.
93 (get_dump_file_info): Convert to...
94 (gcc::dump_manager::get_dump_file_info): ...method.
95 (get_dump_file_name): Convert to...
96 (gcc::dump_manager::get_dump_file_name): ...method.
97 (dump_start): Convert to...
98 (gcc::dump_manager::dump_start): ...method.
99 (dump_finish): Convert to...
100 (gcc::dump_manager::dump_finish): ...method.
101 (dump_begin): Replace body with...
102 (gcc::dump_manager::dump_begin): ...new method.
103 (dump_phase_enabled_p): Convert to...
104 (gcc::dump_manager::dump_phase_enabled_p): ...method.
105 (dump_phase_enabled_p): Convert to...
106 (gcc::dump_manager::dump_phase_enabled_p): ...method.
107 (dump_initialized_p): Convert to...
108 (gcc::dump_manager::dump_initialized_p): ...method.
109 (dump_flag_name): Replace body with...
110 (gcc::dump_manager::dump_flag_name): ...new method.
111 (dump_enable_all): Convert to...
112 (gcc::dump_manager::dump_enable_all): ...new method.
113 (opt_info_enable_passes): Convert to...
114 (gcc::dump_manager::opt_info_enable_passes): ...new method.
115 (dump_switch_p_1): Convert to...
116 (gcc::dump_manager::dump_switch_p_1): ...new method.
117 (dump_switch_p): Convert to...
118 (gcc::dump_manager::dump_switch_p): ...new method.
119 (opt_info_switch_p): Port to dump_manager API.
120 (enable_rtl_dump_file): Likewise.
121 * opts-global.c (handle_common_deferred_options): Port to new
123 * passes.c (pass_manager::finish_optimization_passes): Likewise.
124 (pass_manager::register_one_dump_file): Likewise.
125 (pass_manager::register_pass): Likewise.
126 (pass_init_dump_file): Likewise.
127 (pass_fini_dump_file): Likewise.
128 * statistics.c (statistics_early_init): Likewise.
130 2013-10-14 Richard Biener <rguenther@suse.de>
132 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
133 iterative_hash_canonical_type, gimple_canonical_type_hash,
134 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
135 gimple_register_canonical_type, print_gimple_types_stats,
136 free_gimple_type_tables): Move to lto/lto.c
137 (gt-gimple.h): Do not include.
138 * gimple.h (gimple_register_canonical_type,
139 print_gimple_types_stats, free_gimple_type_tables): Remove.
140 * Makefile.in (GTFILES): Remove gimple.c.
142 2013-10-14 Travis Snoozy <quandary@remstate.com>
145 * config/msp430/msp430.c (msp430_option_override): Correct thinko
146 scanning for msp430x targets.
148 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
151 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
152 (registers_ok_for_ldd_peep): Move around.
153 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
154 * config/sparc/sparc.md (widening peepholes): Use it.
156 2013-10-14 Richard Biener <rguenther@suse.de>
160 * gimple.c (iterative_hash_canonical_type): Make sure to
161 record the hash into the correct hashtable slot.
163 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
165 PR rtl-optimization/58662
166 * combine.c (try_combine): Take into account death nodes on I2 when
167 splitting a PARALLEL of two independent SETs. Fix dump message.
169 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
172 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
173 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
174 * config/sh/t-sh (sh_treg_combine.o): New entry.
175 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
176 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
177 (register_sh_passes): New function. Register sh_treg_combine pass.
178 (sh_option_override): Invoke it.
179 (sh_canonicalize_comparison): Handle op0_preserve_value.
180 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
181 opportunities. Canonicalize branch condition.
182 (nott): Allow only if pseudos can be created for non-SH2A.
184 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
187 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
188 (ix86_expand_movmem): Replace copy_addr_to_reg with
189 ix86_copy_addr_to_reg.
190 (ix86_expand_setmem): Likewise.
192 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
194 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
195 provided by ix86_fp_compare_mode instead of CCFPUmode.
197 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
199 * config/aarch64/arm_neon.h
200 (vtbx<1,3>_<psu>8): Fix register constriants.
202 2013-10-11 Jeff Law <law@redhat.com>
204 PR tree-optimization/58640
205 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
206 threading paths that cross over two loop entry points.
208 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
210 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
211 handle vector float as well.
212 (*vsx_le_perm_load_v4si): Likewise.
213 (*vsx_le_perm_store_v2di): Likewise.
214 (*vsx_le_perm_store_v4si): Likewise.
216 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
218 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
219 directly to circumvent subtract from splat{31} workaround.
220 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
222 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
223 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
224 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
225 (altivec_vperm_<mode>): Convert to define_insn_and_split to
226 separate big and little endian logic.
227 (*altivec_vperm_<mode>_internal): New define_insn.
228 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
229 separate big and little endian logic.
230 (*altivec_vperm_<mode>_uns_internal): New define_insn.
231 (vec_permv16qi): Add little endian logic.
233 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
235 * doc/extend.texi (returns_nonnull): Remove arguments.
237 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
238 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
239 Sergey Lega <sergey.s.lega@intel.com>
240 Anna Tikhonova <anna.tikhonova@intel.com>
241 Ilya Tocar <ilya.tocar@intel.com>
242 Andrey Turetskiy <andrey.turetskiy@intel.com>
243 Ilya Verbin <ilya.verbin@intel.com>
244 Kirill Yukhin <kirill.yukhin@intel.com>
245 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
247 * config/i386/sse.md (VI48F_256_512): New.
248 (avx2_permvar<mode>): Change to ...
249 (<avx2_avx512f>_permvar<mode>): This.
251 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
252 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
253 Sergey Lega <sergey.s.lega@intel.com>
254 Anna Tikhonova <anna.tikhonova@intel.com>
255 Ilya Tocar <ilya.tocar@intel.com>
256 Andrey Turetskiy <andrey.turetskiy@intel.com>
257 Ilya Verbin <ilya.verbin@intel.com>
258 Kirill Yukhin <kirill.yukhin@intel.com>
259 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
261 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
262 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
263 * config/i386/sse.md (VI4_AVX): New.
264 (sf2simodelower): Ditto.
265 (sse2_cvtps2dq): Change to ...
266 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
268 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
269 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
270 Sergey Lega <sergey.s.lega@intel.com>
271 Anna Tikhonova <anna.tikhonova@intel.com>
272 Ilya Tocar <ilya.tocar@intel.com>
273 Andrey Turetskiy <andrey.turetskiy@intel.com>
274 Ilya Verbin <ilya.verbin@intel.com>
275 Kirill Yukhin <kirill.yukhin@intel.com>
276 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
278 * config/i386/sse.md (V_512): New.
280 (vcond<V_512:mode><VF_512:mode>): Ditto.
281 (vcond<V_512:mode><VI_512:mode>): Ditto.
282 (vcondu<V_512:mode><VI_512:mode>): Ditto.
284 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
285 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
286 Sergey Lega <sergey.s.lega@intel.com>
287 Anna Tikhonova <anna.tikhonova@intel.com>
288 Ilya Tocar <ilya.tocar@intel.com>
289 Andrey Turetskiy <andrey.turetskiy@intel.com>
290 Ilya Verbin <ilya.verbin@intel.com>
291 Kirill Yukhin <kirill.yukhin@intel.com>
292 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
294 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
295 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
297 (fma<mode>4): Removed condition.
299 (fnma<mode>4): Ditto.
300 (fnms<mode>4): Ditto.
301 (fma4i_fmadd_<mode>): Ditto.
302 (*fma_fmadd_<mode>): Ditto.
303 (*fma_fmsub_<mode>): Ditto.
304 (*fma_fnmadd_<mode>): Ditto.
305 (*fma_fnmsub_<mode>): Ditto.
306 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
307 (*fma_fmaddsub_<mode>): Ditto.
308 (*fma_fmsubadd_<mode>): Ditto.
309 (*fmai_fmadd_<mode>): Ditto.
310 (*fmai_fmsub_<mode>): Ditto.
311 (*fmai_fnmadd_<mode>): Ditto.
312 (*fmai_fnmsub_<mode>): Ditto.
314 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
315 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
316 Sergey Lega <sergey.s.lega@intel.com>
317 Anna Tikhonova <anna.tikhonova@intel.com>
318 Ilya Tocar <ilya.tocar@intel.com>
319 Andrey Turetskiy <andrey.turetskiy@intel.com>
320 Ilya Verbin <ilya.verbin@intel.com>
321 Kirill Yukhin <kirill.yukhin@intel.com>
322 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
324 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
325 (VI124_256): Changed to ...
326 (VI124_256_48_512): This.
327 (ssepackmode): Extended with wider modes.
328 (<code><mode>3): Changed iterator.
329 (*avx2_<code><mode>3): Ditto.
330 (vec_pack_trunc_<mode>): Ditto.
332 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
333 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
334 Sergey Lega <sergey.s.lega@intel.com>
335 Anna Tikhonova <anna.tikhonova@intel.com>
336 Ilya Tocar <ilya.tocar@intel.com>
337 Andrey Turetskiy <andrey.turetskiy@intel.com>
338 Ilya Verbin <ilya.verbin@intel.com>
339 Kirill Yukhin <kirill.yukhin@intel.com>
340 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
342 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
343 (VI8F_256_512): Ditto.
344 (abs<mode>2): Changed iterator.
345 (avx2_perm<mode>): Changed to ...
346 (<avx2_avx512f>_perm<mode>): This.
347 (avx2_perm<mode>_1): Changed to ...
348 (<avx2_avx512f>_perm<mode>_1): This.
350 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
351 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
352 Sergey Lega <sergey.s.lega@intel.com>
353 Anna Tikhonova <anna.tikhonova@intel.com>
354 Ilya Tocar <ilya.tocar@intel.com>
355 Andrey Turetskiy <andrey.turetskiy@intel.com>
356 Ilya Verbin <ilya.verbin@intel.com>
357 Kirill Yukhin <kirill.yukhin@intel.com>
358 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
360 * config/i386/sse.md (VI48_AVX512F): New.
361 (VI48_AVX2): Changed to ...
362 (VI48_AVX2_48_AVX512F): This.
363 (avx2_ashrv<mode>): Changed to ...
364 (<avx2_avx512f>_ashrv<mode>): This.
365 (avx2_<shift_insn>v<mode>): Changed to ...
366 (<avx2_avx512f>_<shift_insn>v<mode>): This.
368 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
369 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
370 Sergey Lega <sergey.s.lega@intel.com>
371 Anna Tikhonova <anna.tikhonova@intel.com>
372 Ilya Tocar <ilya.tocar@intel.com>
373 Andrey Turetskiy <andrey.turetskiy@intel.com>
374 Ilya Verbin <ilya.verbin@intel.com>
375 Kirill Yukhin <kirill.yukhin@intel.com>
376 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
378 * config/i386/sse.md (VI4_AVX512F): New.
379 (VI8_AVX2_AVX512F): Ditto.
380 (mul<mode>3): Extended with wider modes.
381 (*<sse4_1_avx2>_mul<mode>3): Ditto.
383 (vec_widen_<s>mult_odd_<mode>): Ditto.
385 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
386 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
387 Sergey Lega <sergey.s.lega@intel.com>
388 Anna Tikhonova <anna.tikhonova@intel.com>
389 Ilya Tocar <ilya.tocar@intel.com>
390 Andrey Turetskiy <andrey.turetskiy@intel.com>
391 Ilya Verbin <ilya.verbin@intel.com>
392 Kirill Yukhin <kirill.yukhin@intel.com>
393 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
395 * config/i386/sse.md (VI2_AVX512F): New.
396 (VI124_AVX512F): Ditto.
397 (sseunpackmode): Extended with wider modes.
398 (sseunpackfltmode): Ditto.
399 (vec_unpacks_float_hi_<mode>): Ditto.
400 (vec_unpacks_float_lo_<mode>): Ditto.
401 (vec_unpacku_float_hi_<mode>): Ditto.
402 (vec_unpacku_float_lo_<mode>): Ditto.
403 (vec_unpacks_lo_<mode>): Ditto.
404 (vec_unpacks_hi_<mode>): Ditto.
405 (vec_unpacku_lo_<mode>): Ditto.
406 (vec_unpacku_hi_<mode>): Ditto.
408 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
409 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
410 Sergey Lega <sergey.s.lega@intel.com>
411 Anna Tikhonova <anna.tikhonova@intel.com>
412 Ilya Tocar <ilya.tocar@intel.com>
413 Andrey Turetskiy <andrey.turetskiy@intel.com>
414 Ilya Verbin <ilya.verbin@intel.com>
415 Kirill Yukhin <kirill.yukhin@intel.com>
416 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
418 * config/i386/i386.md (multdiv): New.
419 (multdiv_mnemonic): Ditto.
420 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
421 (<sse>_vm<multdiv_mnemonic><mode>3): This.
422 (<sse>_vmdiv<mode>3): Removed.
424 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
425 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
426 Sergey Lega <sergey.s.lega@intel.com>
427 Anna Tikhonova <anna.tikhonova@intel.com>
428 Ilya Tocar <ilya.tocar@intel.com>
429 Andrey Turetskiy <andrey.turetskiy@intel.com>
430 Ilya Verbin <ilya.verbin@intel.com>
431 Kirill Yukhin <kirill.yukhin@intel.com>
432 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
434 * config/i386/sse.md (V): Extended with wider modes.
436 (ssehalfvecmode): Ditto.
438 (ssepackfltmode): Ditto.
439 (avx_vec_concat<mode>): Ditto.
440 (V_256_512): New iterator.
441 (VF2_512_256): Ditto.
442 (si2dfmode): New attribute.
443 (si2dfmodelower): Ditto.
445 (concat_tg_mode): Ditto.
446 (floatv4siv4df2): Changed to ...
447 (float<si2dfmodelower><mode>2): This.
448 (avx_cvtps2pd256): Changed to ...
449 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
450 (vec_pack_trunc_v4df): Changed to ...
451 (vec_pack_trunc_<mode>): This.
452 (avx_vpermil<mode>): Changed to ...
453 (<sse2_avx_avx512f>_vpermil<mode>): This.
454 (<fixsuffix>fix_truncv8dfv8si2): New.
455 (vec_pack_sfix_trunc_v8df): Ditto.
456 (avx512f_rndscale<mode>): Ditto.
457 (avx512f_roundpd512): Ditto.
458 (vec_pack_ufix_trunc_<mode>): Updated iterator.
460 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
461 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
462 Sergey Lega <sergey.s.lega@intel.com>
463 Anna Tikhonova <anna.tikhonova@intel.com>
464 Ilya Tocar <ilya.tocar@intel.com>
465 Andrey Turetskiy <andrey.turetskiy@intel.com>
466 Ilya Verbin <ilya.verbin@intel.com>
467 Kirill Yukhin <kirill.yukhin@intel.com>
468 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
470 * config/i386/i386.md (any_fix): New iterator.
471 (fixsuffix): New attribute.
472 * config/i386/sse.md (VF1): Extened with wider modes.
476 (sseintvecmodelower): Ditto.
477 (ssescalarmode): Ditto.
478 (ssescalarnum): Ditto.
481 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
482 (<sse>_rcp<mode>2): Change iterator.
483 (rsqrt<mode>2): Ditto.
484 (<sse>_rsqrt<mode>2): Ditto.
485 (avx2_vec_dup<mode>): Ditto.
486 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
487 (round<mode>2_sfix): Ditto.
488 (avx2_pbroadcast<mode>): Ditto.
489 (*andnot<mode>3): Handle XI mode.
490 (*<code><mode>3): Ditto.
491 (AVXTOSSEMODE): Removed.
492 (avx_vpermil<mode>): Changed to ...
493 (<sse2_avx_avx512f>_vpermil<mode>): This.
495 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
496 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
497 Sergey Lega <sergey.s.lega@intel.com>
498 Anna Tikhonova <anna.tikhonova@intel.com>
499 Ilya Tocar <ilya.tocar@intel.com>
500 Andrey Turetskiy <andrey.turetskiy@intel.com>
501 Ilya Verbin <ilya.verbin@intel.com>
502 Kirill Yukhin <kirill.yukhin@intel.com>
503 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
505 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
507 (<sse>_ucomi): Ditto.
508 (sse_cvtss2siq_2): Ditto.
509 (sse2_cvtsd2si): Ditto.
510 (sse2_cvtsd2siq): Ditto.
511 (sse2_cvttsd2si): Ditto.
512 (sse2_cvttsd2siq): Ditto.
513 (<shift_insn><mode>3): Ditto.
514 (sse2_cvtsi2sdq): Update constraint and prefix.
515 (sse_cvtsi2ss): Update prefix.
516 (sse_cvtsi2ssq): Ditto.
518 2013-10-11 Jakub Jelinek <jakub@redhat.com>
520 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
521 ignore internal calls.
523 2013-10-11 Richard Biener <rguenther@suse.de>
525 * tree-pretty-print.c (dump_generic_node): Allow to dump
526 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
527 (OVF) if TREE_OVERFLOW is set.
529 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
531 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
533 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
534 * gimple.h: Likewise.
536 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
538 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
541 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
543 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
544 * configure: Regenerate.
546 2013-10-11 Jakub Jelinek <jakub@redhat.com>
548 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
549 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
550 formatting fixes, use pp_colon instead of pp_character (..., ':'),
551 similarly pp_right_paren.
552 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
553 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
554 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
555 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
557 (omp_declare_simd_clauses_equal,
558 omp_remove_redundant_declare_simd_attrs): New functions.
559 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
560 (walk_tree_1): Handle new OpenMP 4.0 clauses.
561 * tree.h (OMP_LOOP_CHECK): Define.
562 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
563 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
564 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
565 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
566 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
567 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
568 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
569 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
570 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
571 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
572 OMP_CLAUSE_SIMDLEN_EXPR): Define.
573 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
574 (omp_remove_redundant_declare_simd_attrs): New prototype.
575 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
576 GIMPLE_OMP_TEAMS): New codes.
577 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
578 * omp-low.c (struct omp_context): Add cancel_label and cancellable
580 (target_nesting_level): New variable.
581 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
582 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
583 for collapse > 1 static schedule unless ordered.
584 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
585 (determine_parallel_type): Adjust get_ws_args_for caller.
586 (install_var_field): Handle mask & 4 for double indirection.
587 (scan_sharing_clauses): Ignore shared clause on teams construct.
588 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
589 (create_omp_child_function): If inside target or declare target
590 constructs, set "omp declare target" attribute on the child function.
591 (find_combined_for): New function.
592 (scan_omp_parallel): Handle combined loops.
593 (scan_omp_target, scan_omp_teams): New functions.
594 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
595 restrictions and set ctx->cancellable for cancellable constructs.
596 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
597 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
598 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
599 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
600 (omp_clause_aligned_alignment): New function.
601 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
603 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
604 on teams constructs. Handle user defined reductions and new
606 (lower_reduction_clauses): Don't set placeholder to address of ref
607 if it has already the right type.
608 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
609 (expand_parallel_call): Use the new non-_start suffixed builtins,
610 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
611 and GOMP_parallel_end after the call.
612 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
613 (expand_omp_for_init_counts): Handle combined loops.
614 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
616 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
617 end of cancellable loops.
618 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
619 Likewise. Handle collapse > 1 loops.
620 (expand_omp_simd): Handle combined loops.
621 (expand_omp_for): Add inner_stmt argument, adjust callers of
622 expand_omp_for* functions, use expand_omp_for_static*chunk even
623 for collapse > 1 unless ordered.
624 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
625 of cancellable sections.
626 (expand_omp_single): Remove need_barrier variable, just rely on
627 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
628 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
629 (expand_omp_atomic_load, expand_omp_atomic_store,
630 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
631 (expand_omp_target): New function.
632 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
633 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
634 (build_omp_regions_1): Immediately close region for
635 GF_OMP_TARGET_KIND_UPDATE.
636 (maybe_add_implicit_barrier_cancel): New function.
637 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
639 (lower_omp_single): Likewise. Add clobber after the barrier.
640 (lower_omp_taskgroup): New function.
641 (lower_omp_for): Handle combined loops. Adjust
642 lower_rec_input_clauses caller. Handle cancellation.
643 (lower_depend_clauses): New function.
644 (lower_omp_taskreg): Lower depend clauses. Adjust
645 lower_rec_input_clauses caller. Add clobber after the call. Handle
647 (lower_omp_target, lower_omp_teams): New functions.
648 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
649 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
650 and GOMP_cancellation_point calls.
651 (lower_omp): Fold stmts inside of target region.
652 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
653 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
654 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
655 (BT_FN_VOID_OMPFN_PTR_UINT,
656 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
657 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
658 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
659 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
660 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
661 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
662 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
663 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
664 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
665 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
666 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
667 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
668 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
669 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
670 BUILT_IN_GOMP_PARALLEL_END.
671 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
672 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
673 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
674 GF_OMP_FOR_KIND_DISTRIBUTE.
675 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
676 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
677 (dump_gimple_omp_return): Print lhs if it has any.
678 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
679 gimple_omp_atomic_seq_cst_p.
680 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
681 and GIMPLE_OMP_TEAMS.
682 * langhooks.c (lhd_omp_mappable_type): New function.
683 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
684 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
686 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
687 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
688 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
690 (struct gimplify_omp_ctx): Add combined_loop field.
691 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
692 on stmts inside of target region.
693 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
694 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
695 ORT_TARGET and ORT_TARGET_DATA.
696 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
698 (omp_notice_threadprivate_variable): Complain about threadprivate
699 variables in target region.
700 (omp_notice_variable): Complain about vars with non-mappable type
701 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
702 (omp_check_private): Ignore ORT_TARGET* regions.
703 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
704 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
705 (find_combined_omp_for): New function.
706 (gimplify_omp_for): Handle gimplification of combined loops.
707 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
709 (gimplify_omp_target_update): New function.
710 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
711 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
712 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
713 (gimplify_body): If fndecl has "omp declare target" attribute, add
714 implicit ORT_TARGET context around it.
715 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
716 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
717 * tree-nested.c (convert_nonlocal_reference_stmt,
718 convert_local_reference_stmt, convert_gimple_call): Handle
719 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
720 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
721 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
722 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
723 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
724 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
725 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
726 BUILT_IN_GOMP_LOOP_END_CANCEL,
727 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
728 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
729 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
730 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
731 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
732 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
733 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
734 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
735 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
736 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
737 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
738 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
739 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
740 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
741 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
742 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
743 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
744 gimple_build_omp_teams): New functions.
745 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
746 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
747 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
748 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
749 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
750 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
751 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
752 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
753 GF_OMP_ATOMIC_SEQ_CST): New.
754 (gimple_build_omp_taskgroup, gimple_build_omp_target,
755 gimple_build_omp_teams): New prototypes.
756 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
757 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
758 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
759 GIMPLE_OMP_SINGLE as end of range.
760 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
761 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
762 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
763 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
764 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
765 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
766 gimple_omp_target_kind, gimple_omp_target_set_kind,
767 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
768 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
769 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
770 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
771 gimple_omp_teams_set_clauses): New inlines.
772 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
773 and GIMPLE_OMP_TASKGROUP.
774 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
775 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
776 enum omp_clause_proc_bind_kind): New.
777 (union omp_clause_subcode): Add depend_kind, map_kind and
778 proc_bind_kind fields.
779 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
780 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
781 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
782 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
783 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
785 2013-10-10 Teresa Johnson <tejohnson@google.com>
787 * predict.c (tree_estimate_probability): Add new parameter
788 for estimate_bb_frequencies.
789 (estimate_bb_frequencies): Add new parameter to force estimation.
790 (rebuild_frequencies): When max frequency in function is small,
791 recompute counts from frequencies.
792 * predict.h (estimate_bb_frequencies): New parameter.
794 2013-10-10 David Malcolm <dmalcolm@redhat.com>
796 * ipa-inline.c (ipa_inline): Fix leak of "order" when
797 optimizations are disabled.
799 2013-10-10 David Malcolm <dmalcolm@redhat.com>
801 * coverage.c (coverage_finish): Fix leak of da_file_name.
803 2013-10-10 Jan Hubicka <jh@suse.cz>
805 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
806 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
809 2013-10-10 Jakub Jelinek <jakub@redhat.com>
812 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
813 if any labels are in FALLTHRU_BB, use a special label emitted
814 immediately after the asm goto insn rather than label_rtx
816 (expand_asm_stmt): Adjust caller.
817 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
818 edge if the last insn in predecessor is a jump with single successor,
819 but it isn't simplejump_p.
821 2013-10-10 Richard Biener <rguenther@suse.de>
823 PR tree-optimization/58656
824 * tree-ssa-pre.c (phi_translate): Do not cache failed
827 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
829 * gimplify.c: Include expr.h and tm_p.h for targets with special
830 va-arg padding requirements.
832 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
834 * tree-flow.h: Move some prototypes to gimple.h.
835 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
836 * gimple.h: Relocate some prototypes from tree-flow.h
837 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
839 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
840 (build_va_arg_indirect_ref): Relocate and make static.
841 (std_gimplify_va_arg_expr): Relocate here.
842 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
843 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
845 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
847 * doc/md.texi: Document the mnemonic attribute.
849 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
852 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
853 (set_attr_alternative x ...) when searching for user defined
856 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
858 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
860 2013-10-09 Easwaran Raman <eraman@google.com>
862 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
863 * cfgexpand.c (defer_stack_allocation): ...use here
864 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
866 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
868 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
869 (operand_equal_for_value_replacement): New function, extracted from
870 value_replacement and enhanced to catch more cases.
871 (value_replacement): Use operand_equal_for_value_replacement.
873 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
875 * loop-doloop.c (doloop_modify, doloop_optimize): Use
876 get_max_loop_iterations.
878 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
880 * config/arm/aarch-common.c (arm_early_load_addr_dep):
881 Place comment above function.
883 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
885 * tree-flow.h: Remove all remaining prototypes, enums and structs that
886 are not related to tree-cfg.c.
887 * tree-ssa-address.h: New file. Relocate prototypes.
888 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
889 (addr_for_mem_ref): New. Combine call to get_address_description and
890 return addr_for_mem_ref.
891 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
892 * tree-ssa-live.h: Adjust prototypes.
893 * passes.c: Include tree-ssa-live.h.
894 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
895 * graphite.c (graphite_transform_loops): Make static.
896 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
897 make_pass_graphite, pass_data_graphite_transforms,
898 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
899 * ipa-pure-const.c (warn_function_noreturn): Make static.
900 (execute_warn_function_noreturn, gate_warn_function_noreturn,
901 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
902 Relocate from tree-cfg.c
903 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
905 (execute_warn_function_noreturn, gate_warn_function_noreturn,
906 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
907 Move to ipa-pure-const.c.
908 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
909 Relocate from tree-optimize.c.
910 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
911 make_pass_fixup_cfg): Move to tree-cfg.c.
912 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
913 Relocate some prototypes.
914 * tree-data-ref.h (tree_check_data_deps) Add prototype.
915 * tree-dump.c (dump_function_to_file): Remove prototype.
916 Add tree-flow.h to the include file.
917 * tree-dump.h: Remove prototype.
918 * tree-parloops.h: New File. Add prototypes.
919 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
920 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
921 from tree-ssa-loop.c.
922 * tree-predcom.c (run_tree_predictive_commoning,
923 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
924 Relocate here from tree-ssa-loop.c.
925 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
926 ssa_name_values.release ().
927 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
928 (ssa_name_values): Relocate from tree-flow.h.
929 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
930 * tree-ssa-loop.c (run_tree_predictive_commoning,
931 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
932 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
933 make_pass_graphite, pass_data_graphite_transforms,
934 make_pass_graphite_transforms, gate_tree_parallelize_loops,
935 tree_parallelize_loops, pass_data_parallelize_loops,
936 make_pass_parallelize_loops): Move to other files.
937 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
939 * tree.h: Remove prototypes from tree-address.c.
941 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
943 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
944 (struct int_tree_map): Move to tree-hasher.h
945 (SCALE, LABEL, PERCENT): Move to gimple.h
946 * tree-flow-inline.h: Delete. Move functions to other files.
947 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
948 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
949 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
950 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
951 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
952 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
953 * tree-hasher.h: Don't include tree-flow.h.
954 (struct int_tree_map): Relocate from tree-flow.h.
955 * tree-sra.c (contains_view_convert_expr_p): Relocate from
956 tree-flow-inline.h and make static.
957 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
959 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
960 tree-flow-inline.h and make static.
961 * tree.h (is_global_var, may_be_aliased): Relocate from
963 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
964 * value-prof.c: No longer include tree-flow-inline.h.
965 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
967 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
969 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
970 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
971 (enum move_pos): Move to tree-ssa-loop-im.h
972 * cfgloop.h: Move some prototypes.
973 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
974 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
975 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
976 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
977 (loop_containing_stmt): Relocate from tree-flow-inline.h.
978 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
979 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
980 (enum move_pos): Relocate here.
981 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
983 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
984 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
985 make_pass_lim): Relocate here from tree-ssa-loop.c.
986 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
988 (loop_edge_to_cancel, unloop_loops): Make static.
989 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
990 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
991 (tree_complete_unroll, gate_tree_complete_unroll,
992 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
993 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
994 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
995 * tree-ssa-loop-ivopts.c: Remove local prototypes.
996 (stmt_invariant_in_loop_p): Remove unused function.
997 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
998 * tree-ssa-loop-manip.h: New file. Add prototypes.
999 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
1000 (gcov_type_to_double_int): Move to cfgloop.h.
1001 (double_int_cmp, bound_index,
1002 estimate_numbers_of_iterations_loop): Make static.
1003 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
1004 (max_loop_iterations): Factor out get_max_loop_iterations.
1005 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
1007 * tree-ssa-loop-niter.h: New file. Add prototypes.
1008 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
1009 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
1010 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
1011 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
1012 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
1013 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
1014 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
1015 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
1016 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
1017 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
1018 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
1019 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
1020 pass_data_complete_unroll, make_pass_complete_unroll,
1021 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
1022 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
1023 tree-ssa-loop-ivcanon.c.
1024 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
1025 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
1026 tree-ssa-loop-prefetch.c.
1027 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
1029 (get_lsm_tmp_name): Relocate and add suffix parameter.
1030 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
1031 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
1032 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
1033 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
1034 (get_estimated_loop_iterations): Factor out accessor from
1035 estimated_loop_iterations in tree-ssa-loop-niter.c.
1036 (get_max_loop_iterations): Factor out accessor from
1037 _max_loop_iterations in tree-ssa-niter.c.
1038 * loop-unroll.c (decide_unroll_constant_iterations,
1039 decide_unroll_runtime_iterations, decide_peel_simple,
1040 decide_unroll_stupid): Use new get_* accessors.
1042 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
1044 PR tree-optimization/20318
1045 * doc/extend.texi (returns_nonnull): New function attribute.
1046 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
1048 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
1049 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
1051 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
1054 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
1055 false if both components are bitfields.
1057 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1059 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
1060 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
1061 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
1062 * config/aarch64/aarch64.h
1063 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
1064 * config/aarch64/aarch64-simd-builtins.def
1065 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
1067 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1069 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
1071 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1073 * config/aarch64/arm_neon.h (vdiv_f64): Added.
1075 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1077 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
1078 (vneg_f64): New intrinsic.
1079 (vneg_s8): Asm replaced with C.
1080 (vneg_s16): Likewise.
1081 (vneg_s32): Likewise.
1082 (vneg_s64): New intrinsic.
1083 (vnegq_f32): Asm replaced with C.
1084 (vnegq_f64): Likewise.
1085 (vnegq_s8): Likewise.
1086 (vnegq_s16): Likewise.
1087 (vnegq_s32): Likewise.
1088 (vnegq_s64): Likewise.
1090 2013-10-09 Renlin Li <Renlin.Li@arm.com>
1092 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
1094 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1096 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
1097 packed stack special handling.
1098 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
1099 back to fixed stack slots for FPRs saved due to stdarg.
1101 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1103 * config/s390/s390.c (s390_frame_info): Restructure function.
1105 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1107 * config/s390/s390.c (struct s390_frame_layout): New field
1109 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
1110 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
1111 regs_ever_clobbered to char*.
1112 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
1113 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
1114 (s390_register_info): Enable FPR save slots. Move/Copy some
1115 functionality into ...
1116 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
1117 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
1119 (s390_frame_info): Do gpr slot allocation here now. stdarg does
1120 not imply a stack frame.
1121 (s390_init_frame_layout): Remove variable clobbered_regs.
1122 (s390_update_register_info): Remove function.
1123 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
1124 cannot be used for register renaming.
1125 (s390_hard_regno_scratch_ok): New function.
1126 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
1127 (s390_initial_elimination_offset): Change offset calculation of
1128 the return address pointer.
1129 (save_gprs): Deal with only r6 being saved from the call-saved regs.
1130 (restore_gprs): Set frame related flag.
1131 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
1132 (s390_emit_prologue): Call s390_register_info instead of
1133 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
1134 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
1135 (s390_optimize_prologue): Call s390_optimize_register_info.
1136 Try to remove also FPR slot save/restore INSNs. Remove frame
1137 related flags from restore INSNs.
1139 2013-10-08 DJ Delorie <dj@redhat.com>
1141 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
1144 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
1145 avoid conflict with the MI use of %c.
1146 * config/rl78/rl78-real.md: change %c to %C throughout.
1147 * config/rl78/rl78-virt.md: Likewise.
1149 2013-10-08 Jan Hubicka <jh@suse.cz>
1151 * config/i386/i386.c (ix86_option_override_internal): Switch
1152 to SSE math for -ffast-math when target ISA supports SSE2.
1154 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
1156 * tree-flow.h: Remove some prototypes.
1157 * tree.h: Remove some protypes, add a couple.
1158 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
1159 using_eh_for_cleanups_p): Add interface routines for front ends.
1160 * tree-eh.h: New file. Add protoptyes.
1161 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
1162 (add_stmt_to_eh_lp_fn): Make static.
1163 (lower_try_finally): Use new using_eh_for_cleanups_p.
1164 * emit-rtl.c: Include tree-eh.h.
1165 * gimple.h: Include tree-eh.h.
1167 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
1169 PR tree-optimization/58480
1170 * tree-vrp.c (infer_nonnull_range): New function.
1171 (infer_value_range): Call infer_nonnull_range.
1173 2013-10-08 Dehao Chen <dehao@google.com>
1175 PR tree-optimization/58619
1176 * tree-inline.c (copy_phis_for_bb): Combine location data
1179 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1182 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
1183 RTX_FRAME_RELATED_P on INSN.
1185 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1187 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
1188 (altivec_expand_vec_perm_const): Call it.
1190 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1192 * config/rs6000/vector.md (mov<mode>): Emit permuted move
1193 sequences for LE VSX loads and stores at expand time.
1194 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
1196 * config/rs6000/rs6000.c (rs6000_const_vec): New.
1197 (rs6000_gen_le_vsx_permute): New.
1198 (rs6000_gen_le_vsx_load): New.
1199 (rs6000_gen_le_vsx_store): New.
1200 (rs6000_gen_le_vsx_move): New.
1201 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
1202 (*vsx_le_perm_load_v4si): New.
1203 (*vsx_le_perm_load_v8hi): New.
1204 (*vsx_le_perm_load_v16qi): New.
1205 (*vsx_le_perm_store_v2di): New.
1206 (*vsx_le_perm_store_v4si): New.
1207 (*vsx_le_perm_store_v8hi): New.
1208 (*vsx_le_perm_store_v16qi): New.
1209 (*vsx_xxpermdi2_le_<mode>): New.
1210 (*vsx_xxpermdi4_le_<mode>): New.
1211 (*vsx_xxpermdi8_le_V8HI): New.
1212 (*vsx_xxpermdi16_le_V16QI): New.
1213 (*vsx_lxvd2x2_le_<mode>): New.
1214 (*vsx_lxvd2x4_le_<mode>): New.
1215 (*vsx_lxvd2x8_le_V8HI): New.
1216 (*vsx_lxvd2x16_le_V16QI): New.
1217 (*vsx_stxvd2x2_le_<mode>): New.
1218 (*vsx_stxvd2x4_le_<mode>): New.
1219 (*vsx_stxvd2x8_le_V8HI): New.
1220 (*vsx_stxvd2x16_le_V16QI): New.
1222 2013-10-07 Renlin Li <Renlin.Li@arm.com>
1224 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
1226 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1228 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
1229 loop to work also for 31bit ABI.
1230 Save the stack pointer for frame_size > 0.
1232 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1234 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
1235 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
1236 constraint letters from expanders.
1237 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
1238 retry count to general_operand.
1239 ("tabort"): Give operand 0 a mode.
1240 ("tabort_1"): Add mode and constraint letter for operand 0.
1241 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
1243 2013-10-04 Jeff Law <law@redhat.com>
1245 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
1247 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
1249 (thread_across_edge): Here. Call it.
1251 2013-10-04 Cary Coutant <ccoutant@google.com>
1253 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
1254 discarding a location list expression (or a piece of one).
1256 2013-10-03 Jan Hubicka <jh@suse.cz>
1258 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
1259 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
1260 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
1262 2013-10-03 Jan Hubicka <jh@suse.cz>
1264 * config/i386/i386.c (ix86_option_override_internal): Do not enable
1265 accumulate-outgoing-args when producing unwind info.
1267 2013-10-03 Wei Mi <wmi@google.com>
1269 * lra-constraints.c (insert_move_for_subreg): New function
1270 extracted from simplify_operand_subreg.
1271 (simplify_operand_subreg): Add reload for paradoxical subreg.
1273 2013-10-03 Rong Xu <xur@google.com>
1275 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
1276 the candidate of builtin_expect such that we should fix the
1277 size/time estimation.
1278 (estimate_function_body_sizes): Do the acutally size/time fix-up
1281 2013-10-03 Rong Xu <xur@google.com>
1283 * predict.c (tree_predict_by_opcode): Get the probability
1284 for builtin_expect from param builtin_expect_probability.
1285 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
1286 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
1287 * doc/invoke.texi: Add documentation for builtin-expect-probability.
1289 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
1292 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
1293 * calls.c (alloca_call_p): Use get_callee_fndecl.
1294 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
1295 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
1297 (vrp_visit_stmt): Remove duplicated code.
1299 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
1301 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
1302 ceildf2, btruncdf2, instead of vsx_* name.
1304 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
1305 iterators to only do V2DF and V4SF here. Move the DF code to
1306 rs6000.md where it is combined with SF mode. Replace <VSv> with
1307 just 'v' since only vector operations are handled with these insns
1308 after moving the DF support to rs6000.md.
1309 (vsx_sub<mode>3): Likewise.
1310 (vsx_mul<mode>3): Likewise.
1311 (vsx_div<mode>3): Likewise.
1312 (vsx_fre<mode>2): Likewise.
1313 (vsx_neg<mode>2): Likewise.
1314 (vsx_abs<mode>2): Likewise.
1315 (vsx_nabs<mode>2): Likewise.
1316 (vsx_smax<mode>3): Likewise.
1317 (vsx_smin<mode>3): Likewise.
1318 (vsx_sqrt<mode>2): Likewise.
1319 (vsx_rsqrte<mode>2): Likewise.
1320 (vsx_fms<mode>4): Likewise.
1321 (vsx_nfma<mode>4): Likewise.
1322 (vsx_copysign<mode>3): Likewise.
1323 (vsx_btrunc<mode>2): Likewise.
1324 (vsx_floor<mode>2): Likewise.
1325 (vsx_ceil<mode>2): Likewise.
1326 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
1327 (vsx_sminsf3): Likewise.
1328 (vsx_fmadf4): Likewise.
1329 (vsx_fmsdf4): Likewise.
1330 (vsx_nfmadf4): Likewise.
1331 (vsx_nfmsdf4): Likewise.
1332 (vsx_cmpdf_internal1): Likewise.
1334 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
1335 simpler to select whether a target has SPE or traditional floating
1336 point support in iterators.
1337 (TARGET_DF_SPE): Likewise.
1338 (TARGET_SF_FPR): Likewise.
1339 (TARGET_DF_FPR): Likewise.
1340 (TARGET_SF_INSN): Macros to say whether floating point support
1341 exists for a given operation for expanders.
1342 (TARGET_DF_INSN): Likewise.
1344 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
1345 combining of SF/DF mode operations, using both traditional and VSX
1353 (abs<mode>2): Combine SF/DF modes using traditional floating point
1354 instructions. Add support for using the upper DF registers with
1355 VSX support, and SF registers with power8-vector support. Update
1356 expanders for operations supported by both the SPE and traditional
1357 floating point units.
1358 (abs<mode>2_fpr): Likewise.
1359 (nabs<mode>2): Likewise.
1360 (nabs<mode>2_fpr): Likewise.
1361 (neg<mode>2): Likewise.
1362 (neg<mode>2_fpr): Likewise.
1363 (add<mode>3): Likewise.
1364 (add<mode>3_fpr): Likewise.
1365 (sub<mode>3): Likewise.
1366 (sub<mode>3_fpr): Likewise.
1367 (mul<mode>3): Likewise.
1368 (mul<mode>3_fpr): Likewise.
1369 (div<mode>3): Likewise.
1370 (div<mode>3_fpr): Likewise.
1371 (sqrt<mode>3): Likewise.
1372 (sqrt<mode>3_fpr): Likewise.
1373 (fre<Fs>): Likewise.
1374 (rsqrt<mode>2): Likewise.
1375 (cmp<mode>_fpr): Likewise.
1376 (smax<mode>3): Likewise.
1377 (smin<mode>3): Likewise.
1378 (smax<mode>3_vsx): Likewise.
1379 (smin<mode>3_vsx): Likewise.
1380 (negsf2): Delete SF operations that are merged with DF.
1387 (fmasf4_fpr): Likewise.
1388 (fmssf4_fpr): Likewise.
1389 (nfmasf4_fpr): Likewise.
1390 (nfmssf4_fpr): Likewise.
1391 (sqrtsf2): Likewise.
1392 (rsqrtsf_internal1): Likewise.
1393 (smaxsf3): Likewise.
1394 (sminsf3): Likewise.
1395 (cmpsf_internal1): Likewise.
1396 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
1397 (negdf2): Delete DF operations that are merged with SF.
1399 (nabsdf2): Likewise.
1405 (rsqrtdf_internal1): Likewise.
1406 (fmadf4_fpr): Likewise.
1407 (fmsdf4_fpr): Likewise.
1408 (nfmadf4_fpr): Likewise.
1409 (nfmsdf4_fpr): Likewise.
1410 (sqrtdf2): Likewise.
1411 (smaxdf3): Likewise.
1412 (smindf3): Likewise.
1413 (cmpdf_internal1): Likewise.
1414 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
1415 (btrunc<mode>2): Delete separate expander, and combine with the
1416 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
1417 (btrunc<mode>2_fpr): Likewise.
1418 (ceil<mode>2): Likewise.
1419 (ceil<mode>2_fpr): Likewise.
1420 (floor<mode>2): Likewise.
1421 (floor<mode>2_fpr): Likewise.
1422 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
1423 Add support for using the upper registers with VSX and
1424 power8-vector. Move insns to be closer to the define_expands. On
1425 VSX systems, prefer the traditional form of FMA over the VSX
1426 version, since the traditional form allows the target not to
1427 overlap with the inputs.
1428 (fms<mode>4_fpr): Likewise.
1429 (nfma<mode>4_fpr): Likewise.
1430 (nfms<mode>4_fpr): Likewise.
1432 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1433 Richard Earnshaw <richard.earnshaw@arm.com>
1435 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
1436 (struct mult_cost_table): Likewise.
1437 (struct mem_cost_table): Likewise.
1438 (struct fp_cost_table): Likewise.
1439 (struct vector_cost_table): Likewise.
1440 (cpu_cost_table): Likewise.
1441 * config/arm/arm.opt (mold-rts-costs): New option.
1442 (mnew-generic-costs): Likewise.
1443 * config/arm/arm.c (generic_extra_costs): New table.
1444 (cortexa15_extra_costs): Likewise.
1445 (arm_slowmul_tune): Use NULL as new costs.
1446 (arm_fastmul_tune): Likewise.
1447 (arm_strongarm_tune): Likewise.
1448 (arm_xscale_tune): Likewise.
1449 (arm_9e_tune): Likewise.
1450 (arm_v6t2_tune): Likewise.
1451 (arm_cortex_a5_tune): Likewise.
1452 (arm_cortex_a9_tune): Likewise.
1453 (arm_v6m_tune): Likewise.
1454 (arm_fa726te_tune): Likewise.
1455 (arm_cortex_a15_tune): Use cortex15_extra_costs.
1456 (arm_cortex_tune): Use generict_extra_costs.
1457 (shifter_op_p): New function.
1458 (arm_unspec_cost): Likewise.
1459 (LIBCALL_COST): Define.
1460 (arm_new_rtx_costs): New function.
1461 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
1462 table is available. Use old costs otherwise unless mnew-generic-costs
1464 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
1465 (cpu_cost_table): Declare.
1467 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
1470 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
1471 (*subs_mul_imm_<mode>)
1472 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
1473 (*sub_<shift>_<mode>)
1474 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
1475 Remove k constraint.
1477 2013-10-03 Ian Bolton <ian.bolton@arm.com>
1479 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
1481 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
1483 2013-10-02 Teresa Johnson <tejohnson@google.com>
1485 * predict.c (probably_never_executed): New function.
1486 (probably_never_executed_bb_p): Invoke probably_never_executed.
1487 (probably_never_executed_edge_p): Ditto.
1488 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
1489 Treat profile insanities conservatively.
1491 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
1493 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
1495 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
1497 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
1498 use it. Use smaller increase for scratch. Don't increase reject
1499 for early clobber scratch.
1500 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
1501 setting eliminated regs except setting fp from hfp.
1502 (lra_eliminate): Check lra_insn_recog_data on NULL.
1504 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1507 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
1508 setting -mvsx-timode by default until the underlying problem is fixed.
1509 (RS6000_CPU, power7 defaults): Likewise.
1511 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
1513 * config/x-linux (host-linux.o): Remove header dependencies.
1514 Use $(COMPILE) and $(POSTCOMPILE).
1515 * config/t-linux-android (linux-android.o): Ditto.
1517 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
1519 * Makefile.in (expmed.o-warn): Remove.
1521 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1523 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
1524 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
1526 2013-10-02 Teresa Johnson <tejohnson@google.com>
1528 * dojump.c (do_jump_1): Divide probability between
1529 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
1531 2013-10-02 Tom Tromey <tromey@redhat.com>
1533 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
1535 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1537 * tree-flow.h: Remove some prototypes.
1538 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
1539 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
1540 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
1541 mark_virtual_phi_result_for_renaming): Relocate here.
1542 * tree-into-ssa.h: Add prototypes.
1543 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
1544 single_pred_before_succ_order.
1545 (blocks_in_phiopt_order): Rename and move to cfganal.c.
1546 (nonfreeing_call_p) Move to gimple.c.
1547 * cfganal.c (single_pred_before_succ_order): Move and renamed from
1549 * basic-block.h (single_pred_before_succ_order): Add prototype.
1550 * gimple.c (nonfreeing_call_p): Relocate here.
1551 * gimple.h: Add prototype.
1552 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
1553 * tree-ssa-dom.h: New file. Relocate prototypes here.
1554 * tree-ssa.h: Include tree-ssa-dom.h.
1556 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
1558 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
1559 Use $(COMPILE) and $(POSTCOMPILE).
1561 * config/alpha/x-alpha (driver-alpha.o): Ditto.
1563 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1565 * tree-flow.h: Remove some prototypes.
1566 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
1567 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
1568 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
1569 propagate_tree_value*): Move from here to...
1570 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
1571 propagate_tree_value*): Relocate here.
1572 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
1573 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
1574 * gimple-fold.c: Remove gimple-fold.h from include list.
1575 * tree-vrp.c: Remove gimple-fold.h from include list.
1576 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
1577 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
1578 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
1579 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
1580 * sese.c: Add tree-ssa-propagate.h to include list.
1582 2013-10-02 Richard Biener <rguenther@suse.de>
1584 * tree-loop-distribution.c: Include tree-vectorizer.h for
1586 (enum partition_kind): Remove PKIND_REDUCTION.
1587 (struct partition_s): Remove has_writes member, add reduction_p member.
1588 (partition_alloc): Adjust.
1589 (partition_builtin_p): Likewise.
1590 (partition_has_writes): Remove.
1591 (partition_reduction_p): New function.
1592 (partition_merge_into): Likewise.
1593 (generate_code_for_partition): Commonize builtin partition
1595 (rdg_cannot_recompute_vertex_p): Remove.
1596 (already_processed_vertex_p): Likewise.
1597 (rdg_flag_vertex): Do not set has_writes.
1598 (classify_partition): Adjust.
1599 (rdg_build_partitions): Do not set has_writes, treat all
1600 partitions as useful.
1601 (distribute_loop): Record number of library calls generated. Adjust.
1602 (tree_loop_distribution): Report number of loops and library
1603 calls generated as opt-info.
1605 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
1607 * tree-flow.h: Include new .h files. Move prototypes.
1608 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
1609 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
1610 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
1611 * tree-pretty-print.h: Add prototypes from tree-flow.h.
1612 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
1613 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
1614 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
1615 * tree.h (get_ref_base_and_extent): Move prototype out.
1616 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
1618 * gimple-low.h: New File. Add prototypes from tree-flow.h.
1619 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
1620 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
1621 * tree-scalar-evolution.c: Include tree.h.
1622 * sese.c: Include tree.h.
1623 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
1624 * dwarf2out.c: Include tree-dfa.h.
1625 * tree-chrec.c: Include tree.h.
1626 * tree-data-ref.c: Include tree.h.
1628 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
1630 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
1633 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1635 * config/t-sol2 (sol2-c.o): Remove header dependencies.
1636 Use $(COMPILE) and $(POSTCOMPILE).
1637 (sol2-cxx.o): Likewise.
1638 (sol2-stubs.o): Likewise.
1640 * config/x-solaris (host-solaris.o): Likewise.
1642 * config/sparc/t-sparc (sparc.o): Remove.
1643 (sparc-c.o): Remove header dependencies.
1644 Use $(COMPILE) and $(POSTCOMPILE).
1645 * config/sparc/x-sparc: Likewise.
1647 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
1649 * config/arc/arc-opts.h: Add 2013 to Copyright years.
1650 * config/arc/arc700.md: Likewise.
1651 * config/arc/arc-modes.def: Likewise.
1652 * config/arc/arc-simd.h: Likewise.
1653 * config/arc/t-arc-uClibc: Likewise.
1654 * config/arc/t-arc-newlib: Likewise.
1656 2013-10-02 Renlin Li <renlin.li@arm.com>
1658 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
1660 (aarch64_expand_epilogue): Likewise.
1662 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1663 Yufeng Zhang <yufeng.zhang@arm.com>
1665 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
1667 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
1668 'base_in' represent a conversion and legal_cast_p_1 holds; set
1669 'base_in' with the returned value from get_unwidened.
1671 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1673 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
1674 plus_constant is not used.
1676 2013-10-01 Wei Mi <wmi@google.com>
1678 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
1679 * config/i386/i386.md: Add define_peephole2 to
1680 break partial reg stall for cvtss2sd/cvtsd2ss.
1682 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1684 * config/arc/arc.c (pass_arc_ifcvt::clone):
1685 Update for ctxt_ -> m_ctxt change.
1687 2013-10-01 Jeff Law <law@redhat.com>
1689 * tree-ssa-threadupdate.c (struct redirection_data): Delete
1690 outgoing_edge and intermediate_edge fields. Instead store the path.
1691 (redirection_data::hash): Hash on the last edge's destination index.
1692 (redirection_data::equal): Check the entire thread path.
1693 (lookup_redirectio_data): Corresponding changes.
1694 (create_edge_and_update_destination_phis): Likewise.
1695 (thread_single_edge): Likewise.
1697 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1698 Diego Novillo <dnovillo@google.com>
1700 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
1701 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
1702 (vld32wh_insn, vld32wl_insn): Delete commented-out old
1703 versions of these patterns.
1705 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
1706 (__builtin_arc_aligned): Likewise.
1708 * config/arc/arc.md: Expand adc_0 comment stating the intended
1709 purpose and why it isn't ready.
1710 Replace commented out call_value_via_label_mixed with a
1711 plain comment about bl_s.
1713 * config/arc/arc.c (stdio.h): Don't include directly.
1714 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
1715 Remove if (1) condition.
1716 (arc_encode_section_info): Fix comment.
1718 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1720 * config/arc/arc.c (arc_conditional_register_usage):
1721 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
1722 Also set reg_alloc_order for DMA config regs.
1724 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1725 Jeremy Bennett <jeremy.bennett@embecosm.com>
1727 * doc/install.texi (--with-cpu): Mention ARC.
1728 (arc-*-elf32): New paragraph.
1729 (arc-linux-uclibc): Likewise.
1730 * doc/md.texi (Machine Constraints): Add ARC part.
1731 * doc/invoke.texi: (menu): Add ARC Options.
1732 (Machine Dependent Options) <ARC Options>: Add synopsis.
1733 (node ARC Options): Add.
1734 * doc/extend.texi (long_call / short_call attribute): Add ARC.
1735 (ARC Built-in Functions): New section defining
1736 generic ARC built-in functions.
1737 (ARC SIMD Built-in Functions): New section defining SIMD specific
1739 (Declaring Attributes of Functions): Extended
1740 description of short_call and long_call attributes for ARC and
1741 added index entries.
1743 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
1744 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
1745 Joern Rennecke <joern.rennecke@embecosm.com>
1746 Muhammad Khurram Riaz <khurram.riaz@arc.com>
1747 Brendan Kehoe <brendan@zen.org>
1748 Michael Eager <eager@eagercon.com>
1749 Simon Cook <simon.cook@embecosm.com>
1750 Jeremy Bennett <jeremy.bennett@embecosm.com>
1752 * config/arc, common/config/arc: New directories.
1754 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
1755 Brendan Kehoe <brendan@zen.org>
1756 Simon Cook <simon.cook@embecosm.com>
1758 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
1760 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
1762 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
1763 * tree-ssa-coalesce.h: New. Move prototype to here.
1764 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
1765 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
1766 (gimple_can_coalesce_p): Move to...
1767 * gimple.c (gimple_can_coalesce_p): Here.
1769 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
1771 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
1772 (dump_decl_set): Move to gimple.c.
1773 * gimple.h: Don't include tree-ssa-operands.h.
1774 (dump_decl_set): Add prototype.
1775 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
1776 Move to gimple-ssa.h.
1777 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
1778 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
1779 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
1780 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
1781 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
1783 (dump_decl_set): Relocate here.
1784 * gimple-ssa.h: New file.
1785 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
1786 Relocate from gimple.h.
1787 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
1788 * tree-ssa-operands.c (swap_ssa_operands): Rename from
1789 swap_tree_operands and remove non-ssa path.
1790 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
1791 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
1793 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
1794 vect_is_simple_reduction_1): Use swap_ssa_operands.
1795 * tree-flow.h: Move various prototypes to tree-phinodes.h.
1796 (enum need_phi_state): Move to tree-into-ssa.c.
1797 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
1798 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
1799 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
1800 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
1801 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
1802 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
1803 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
1804 num_imm_uses): Move to ssa-iterators.h.
1805 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
1806 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
1808 (op_iter_done, op_iter_next_def, op_iter_next_tree,
1809 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
1810 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
1811 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
1812 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
1813 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
1814 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
1815 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
1816 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
1817 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
1818 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
1819 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
1820 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
1821 (set_phi_nodes): Move to tree-phinodes.h.
1822 * tree-ssa-operands.h (enum ssa_op_iter_type,
1823 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
1824 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
1825 (dump_decl_set): Remove prototype.
1826 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
1827 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
1828 (set_phi_nodes): Relocate from tree-flow-inline.h.
1829 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
1831 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
1832 include list. Temporarily add gimple.h to include list.
1833 * ssa-iterators.h: New file.
1834 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
1835 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
1836 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
1837 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
1838 Relocate from tree-ssa-operands.h.
1839 (delink_imm_use, link_imm_use_to_list, link_imm_use,
1840 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
1841 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
1842 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
1843 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
1844 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
1845 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
1846 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
1847 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
1848 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
1849 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
1850 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
1851 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
1852 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
1853 Relocate from tree-flow-inline.h.
1854 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
1856 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
1859 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
1860 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
1861 Insert '\t' to output template.
1862 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
1863 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
1864 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
1865 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
1867 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
1869 * doc/install.texi (Host/target specific installation notes for GCC):
1870 Put @anchor before @heading.
1871 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
1872 Use @email for email addresses.
1874 2013-10-01 Jeff Law <law@redhat.com>
1876 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
1877 a vec. Only delete the path if we create one without successfully
1878 registering a jump thread.
1879 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
1881 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
1882 (paths): New vector of jump threading paths.
1883 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
1884 (THREAD_PATH): New accessor macro for the entire thread path.
1885 (lookup_redirection_data): Get intermediate and final outgoing edge
1886 from the thread path.
1887 (create_edge_and_update_destination_phis): Copy the threading path.
1888 (ssa_fix_duplicate_block_edges): Get edges and block types from the
1889 jump threading path.
1890 (ssa_redirect_edges): Get edges and block types from the jump threading
1891 path. Free the path vector.
1892 (thread_block): Get edges from the jump threading path. Look at the
1893 entire path to see if we thread to a loop exit. If we cancel a jump
1894 thread request, then free the path vector.
1895 (thread_single_edge): Get edges and block types from the jump threading
1896 path. Free the path vector.
1897 (thread_through_loop_header): Get edges and block types from the jump
1898 threading path. Free the path vector.
1899 (mark_threaded_blocks): Iterate over the vector of paths and store
1900 the path on the appropriate edge. Get edges and block types from the
1901 jump threading path.
1902 (mark_threaded_blocks): Get edges and block types from the jump
1903 threading path. Free the path vector.
1904 (thread_through_all_blocks): Use the vector of paths rather than
1905 a vector of 3-edge sets.
1906 (register_jump_thread): Accept pointer to a path vector rather
1907 than the path vector itself. Store the path vector for later use.
1910 2013-10-01 Jakub Jelinek <jakub@redhat.com>
1911 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1914 * config/s390/s390.c (s390_split_branches): Modify check for table
1916 (s390_chunkify_start): Rearrange table jump insn check in order to
1917 deal with compare and branch insns correctly.
1919 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
1923 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
1924 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
1925 define_insn_and_split.
1926 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
1927 (shiftsi3_compare): New pattern.
1929 * config/arm/unspecs.md (UNSPEC_RRX): New.
1931 2013-10-01 Alan Modra <amodra@gmail.com>
1933 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
1934 casing inout operands.
1936 2013-10-01 Richard Biener <rguenther@suse.de>
1938 PR tree-optimization/58553
1939 * tree-loop-distribution.c (struct partition_s): Add niter member.
1940 (classify_partition): Populate niter member for the partition
1941 and properly identify whether the relevant store happens before
1942 or after the loop exit.
1943 (generate_memset_builtin): Use niter member from the partition.
1944 (generate_memcpy_builtin): Likewise.
1946 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1948 * vec.h (vec_prefix, vec): Prefix member names with "m_".
1949 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
1951 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1953 * basic-block.h (edge_list): Prefix member names with "m_".
1954 * context.h (context): Likewise.
1955 * domwalk.h (dom_walker): Likewise.
1956 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
1957 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1958 * hash-table.h (hash_table): Likewise.
1959 * machmode.h (bit_field_mode_iterator): Likewise.
1960 * pass_manager.h (pass_list): Likewise.
1961 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
1962 * tree-pass.h (pass_data): Likewise.
1963 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
1964 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
1965 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1966 * asan.c (pass_data_asan): Update accordingly.
1967 * cfganal.c (control_dependences::find_control_dependence): Likewise.
1968 (control_dependences::control_dependences): Likewise.
1969 (control_dependences::~control_dependences): Likewise.
1970 (control_dependences::~control_dependences): Likewise.
1971 (control_dependences::get_edges_dependent_on): Likewise.
1972 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
1973 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
1974 * context.c (gcc::context::context): Likewise.
1975 * cprop.c (pass_rtl_cprop::clone): Likewise.
1976 * domwalk.c (dom_walker::walk): Likewise.
1977 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
1978 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
1979 * mode-switching.c (pass_mode_switching::clone): Likewise.
1980 * passes.c (opt_pass::opt_pass): Likewise.
1981 (pass_manager::pass_manager): Likewise.
1982 * predict.c (pass_strip_predict_hints::clone): Likewise.
1983 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
1984 (pass_split_all_insns::clone): Likewise.
1985 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
1987 (bit_field_mode_iterator::next_mode): Likewise.
1988 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
1989 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
1990 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
1991 * tree-complex.c (pass_lower_complex::clone): Likewise.
1992 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
1993 * tree-object-size.c (pass_object_sizes::clone): Likewise.
1994 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
1995 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
1996 (pass_fold_builtins::clone): Likewise.
1997 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
1998 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
1999 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
2000 (pass_cd_dce::clone): Likewise.
2001 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
2002 (pass_phi_only_cprop::clone): Likewise.
2003 * tree-ssa-dse.c (pass_dse::clone): Likewise.
2004 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
2005 * tree-ssa-loop.c (pass_lim::clone): Likewise.
2006 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
2007 * tree-ssa-pre.c (pass_fre::clone): Likewise.
2008 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
2009 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
2010 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
2011 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
2012 * tree-vrp.c (pass_vrp::clone): Likewise.
2013 * tsan.c (pass_tsan::clone): Likewise.
2015 2013-09-30 Jakub Jelinek <jakub@redhat.com>
2018 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
2019 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
2022 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
2023 optimization, punt if sign_bit_p looked through any zero extension.
2025 2013-09-30 Teresa Johnson <tejohnson@google.com>
2027 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
2028 Update redirected out edge count in joiner case.
2029 (ssa_redirect_edges): Common the joiner and non-joiner cases
2030 so that joiner case gets profile updates.
2032 2013-09-30 Richard Biener <rguenther@suse.de>
2034 PR tree-optimization/58554
2035 * tree-loop-distribution.c (classify_partition): Require
2036 unconditionally executed stores for memcpy and memset recognition.
2037 (tree_loop_distribution): Calculate dominance info.
2039 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2041 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
2042 (NO_PROFILE_COUNTERS): Likewise.
2043 (PROFILE_HOOK): Likewise.
2044 (FUNCTION_PROFILER): Likewise.
2045 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
2047 2013-09-30 Iain Sandoe <iain@codesourcery.com>
2049 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
2050 calls and defines in TARGET_MACHO conditional.
2051 (load_macho_picbase_di): Likewise.
2052 (reload_macho_picbase): Likewise.
2053 (reload_macho_picbase_si): Likewise.
2054 (reload_macho_picbase_di): Likewise.
2055 (nonlocal_goto_receiver): Likewise.
2057 2013-09-30 Nick Clifton <nickc@redhat.com>
2059 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
2060 that use the MSP430X ISA.
2061 (msp430_option_override): Scan -mmcu command line option for any
2062 MCU name that supports the MSP430X ISA.
2063 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
2064 -mmcu options which enable the MSP430X ISA.
2066 2013-09-30 Richard Biener <rguenther@suse.de>
2069 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
2070 before looking for setjmp-like calls.
2072 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2075 * config/darwin-protos.h (machopic_get_function_picbase): New.
2076 * config/darwin.c (machopic_get_function_picbase): New.
2077 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
2078 label for a new func. (load_macho_picbase_di): Likewise.
2079 (reload_macho_picbase): New expand.
2080 (reload_macho_picbase_si): New insn.
2081 (reload_macho_picbase_di): New insn.
2082 (nonlocal_goto_receiver): New define and split.
2083 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
2084 (unspecv enum): Add UNSPECV_NLGR.
2086 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2088 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
2089 that altivec registers are correctly sized on Darwin.
2091 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2093 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
2094 darwin-driver.o): Use COMPILE and POSTCOMPILE.
2095 * config/x-darwin (host-darwin.o): Likewise.
2096 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
2097 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
2098 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
2100 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
2102 * doc/invoke.texi: Fix usage of @tie{} command.
2104 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
2106 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
2108 2013-09-28 Jan Hubicka <jh@suse.cz>
2110 * config/i386/x86-tune.def: Add documentation for each of the options;
2113 2013-09-28 Jan Hubicka <jh@suse.cz>
2115 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
2117 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
2118 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
2119 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
2120 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
2121 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
2123 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
2125 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
2126 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
2127 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
2128 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
2129 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
2130 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
2131 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
2132 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
2133 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
2134 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
2135 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
2136 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
2137 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
2138 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
2139 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
2140 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
2141 gimple.h, godump.c, graphite-clast-to-gimple.c,
2142 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
2143 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
2144 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
2145 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2146 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
2147 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
2148 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
2149 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
2150 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
2151 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
2152 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
2153 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
2154 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
2155 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
2156 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
2157 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
2158 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
2159 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
2160 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
2161 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
2162 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
2163 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
2164 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
2165 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
2166 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2167 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2168 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
2169 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
2170 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
2171 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2172 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2173 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
2174 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
2175 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2176 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
2177 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
2178 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
2179 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
2180 whitespace before "(".
2182 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
2184 * expr.h (extract_bit_field): Remove packedp parameter.
2185 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
2186 from forward declaration.
2187 (store_split_bit_field): Remove packedp arg from calls to
2188 extract_fixed_bit_field.
2189 (extract_bit_field_1): Remove packedp parameter and packedp
2190 argument from recursive calls and calls to extract_fixed_bit_field.
2191 (extract_bit_field): Remove packedp parameter and corresponding
2192 arg to extract_bit_field_1.
2193 (extract_fixed_bit_field): Remove packedp parameter. Remove code
2195 (extract_split_bit_field): Remove packedp arg from call to
2196 extract_fixed_bit_field.
2197 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
2198 (copy_blkmode_from_reg): Likewise.
2199 (copy_blkmode_to_reg): Likewise.
2200 (read_complex_part): Likewise.
2201 (store_field): Likewise.
2202 (expand_expr_real_1): Likewise.
2203 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
2204 to extract_bit_field.
2205 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
2206 call to extract_bit_field.
2207 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
2208 call to extract_bit_field.
2209 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
2210 and special packedp behavior from -fstrict-volatile-bitfields
2213 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2215 * lra-eliminations.c (init_elim_table): Guard value_p.
2217 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2219 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
2220 DFmode, DImode, and SFmode in the upper VSX registers based on the
2221 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
2222 if -mpower8-vector. Combine -mvsx-timode handling with the rest
2223 of the VSX register handling.
2225 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
2227 (zero_extendsidi2_lfiwzx): Add support for loading into the
2228 Altivec registers with -mpower8-vector. Use wu/wv constraints to
2229 only do VSX memory options on Altivec registers.
2230 (extendsidi2_lfiwax): Likewise.
2231 (extendsfdf2_fpr): Likewise.
2232 (mov<mode>_hardfloat, SF/SD modes): Likewise.
2233 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
2234 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
2235 (movdi_internal64): Likewise.
2237 2013-09-27 Xinliang David Li <davidxl@google.com>
2239 * opts.c (finish_options): Adjust parameters
2240 according to vect cost model.
2241 (common_handle_option): Set dynamic vect cost
2243 targhooks.c (default_add_stmt_cost): Compute stmt cost
2245 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
2246 Use helper function.
2247 * tree-vectorizer.h (unlimited_cost_model): New function.
2248 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
2249 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
2251 (vect_enhance_data_refs_alignment): Ditto.
2252 * flag-types.h: New enum.
2253 * common/config/i386/i386-common.c (ix86_option_init_struct):
2254 No need to initialize vect_cost_model flag.
2255 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
2258 2013-09-27 Diego Novillo <dnovillo@google.com>
2260 * gimple.h (enum ssa_mode): Remove.
2262 2013-09-27 Paulo Matos <pmatos@broadcom.com>
2265 * gcc.dg/pr58463.c: New test.
2267 2013-09-27 Paulo Matos <pmatos@broadcom.com>
2269 * cfgloop.h (number_of_loops): Fix typo in check for null.
2271 2013-09-27 Jakub Jelinek <jakub@redhat.com>
2274 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
2275 are children of outermost saved_cfun's loop, and set it up to
2276 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
2277 if loop != loop0 and SESE region contains bbs that belong to loop0.
2279 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
2281 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
2282 (binary_scale_code_p, get_base_term, get_index_term): New functions.
2283 (set_address_segment, set_address_base, set_address_index)
2284 (set_address_disp): Accept the argument unconditionally.
2285 (baseness): Remove must_be_base_p and must_be_index_p checks.
2286 (decompose_normal_address): Classify as much as possible in the
2289 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
2291 * cse.c (count_reg_usage): Handle INT_LIST.
2292 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
2293 * reginfo.c (reg_scan_mark_refs): Likewise.
2294 * reload1.c (eliminate_regs_1): Likewise.
2296 2013-09-27 Iain Sandoe <iain@codesourcery.com>
2299 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
2302 2013-09-27 Richard Biener <rguenther@suse.de>
2304 PR tree-optimization/58459
2305 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
2306 restriction not propagating into loops.
2308 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
2310 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
2311 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
2312 * doc/tree-ssa.texi (Walking use-def chains): Delete.
2314 2013-09-26 Richard Biener <rguenther@suse.de>
2316 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
2318 2013-09-26 Richard Biener <rguenther@suse.de>
2320 * alias.h (component_uses_parent_alias_set): Rename to ...
2321 (component_uses_parent_alias_set_from): ... this.
2322 * alias.c (component_uses_parent_alias_set): Rename to ...
2323 (component_uses_parent_alias_set_from): ... this and return
2325 (reference_alias_ptr_type_1): Use the result from
2326 component_uses_parent_alias_set_from instead of stripping
2327 components one at a time.
2328 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
2330 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
2332 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
2333 Move prototypes to...
2334 * tree-ssa-ter.h: New File. Move prototypes here.
2335 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
2336 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
2337 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
2338 from is_replaceable_p.
2339 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
2340 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
2341 (process_replaceable): Use ter_is_replaceable_p.
2342 (find_replaceable_in_bb): Use ter_is_replaceable_p.
2343 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
2344 newly refactored ssa_is_replaceable_p.
2345 * cfgexpand.c: Include tree-outof-ssa.h.
2346 * ssaexpand.h: Delete.
2348 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
2350 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
2351 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
2353 (create_gimple_tmp): Delete.
2354 (get_expr_type, build_assign, build_type_cast): Move to...
2355 * gimple-builder.c: New File.
2356 (get_expr_type): Relocate from gimple.c.
2357 (build_assign, build_type_cast): Change to only create ssanames.
2358 * gimple.h: Move prototypes to...
2359 * gimple-builder.h: New File. Here.
2360 * tree-ssa.h: And here.
2361 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
2362 count_uses_and_derefs): Relocate from gimple.c.
2363 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
2364 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
2365 * tree-ssa-math-opts (execute_cse_reciprocals): Use
2366 gimple_replace_ssa_lhs.
2367 * asan.c: Include gimple-builder.h.
2368 * Makefile.in: Add gimple-builder.o.
2370 2013-09-26 Richard Biener <rguenther@suse.de>
2372 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
2373 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
2374 (loe_visit_block): Use gcc_checking_assert.
2375 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
2376 gimple_assign_ssa_name_copy_p.
2377 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
2380 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
2382 * config/rs6000/t-rs6000 (rs6000.o): Remove.
2383 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
2385 2013-09-26 Richard Biener <rguenther@suse.de>
2387 PR tree-optimization/58539
2388 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
2389 the fact that debug statements are not taking part in loop-closed
2392 2013-09-26 Nick Clifton <nickc@redhat.com>
2394 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
2395 time warning message.
2396 (msp430_print_operand_raw): Delete unused letter parameter.
2397 (TARGET_PRINT_OPERAND_ADDRESS): Define.
2398 (msp430_print_operand_address): New function.
2399 (msp430_print_operand): Move address printing code from here to
2401 * config/msp430/msp430.md (movsipsi2): Add comment in generated
2403 (zero_extendpsisi2): Likewise.
2404 (extendpsisi2): New pattern.
2405 (andneghi3): New pattern.
2407 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
2409 * config/aarch64/aarch64.opt (mlra): New option.
2410 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
2411 (TARGET_LRA_P): Define.
2413 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
2415 * expr.c (expand_assignment): Remove obsolete comment.
2417 2013-09-25 Jeff Law <law@redhat.com>
2419 * tree-flow.h (thread_through_all_blocks): Prototype moved into
2420 tree-ssa-threadupdate.h.
2421 (register_jump_thread): Similarly.
2422 * tree-ssa-threadupdate.h: New header file.
2423 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
2424 * tree-vrp.c: Likewise.
2425 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
2426 (thread_around_empty_blocks): Change type of path vector argument to
2427 an edge,type pair from just an edge. Initialize both elements when
2428 appending to a jump threading path. Tweak references to elements
2430 (thread_across_edge): Similarly. Release memory for the elements
2432 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
2433 (dump_jump_thread_path): New function broken out from
2434 register_jump_thread.
2435 (register_jump_thread): Use dump_jump_thread_path. Change type of
2436 path vector entries. Search the path for NULL edges and dump
2437 the path if one is found. Tweak the conversion of path to 3-edge
2438 form to use the block copy type information embedded in the path.
2440 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
2442 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
2444 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
2445 Vladimir Makarov <vmakarov@redhat.com>
2447 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
2448 from the least significant bit.
2449 (strip_address_mutations): Add bitfield operations handling.
2450 (must_be_index_p): Add shifting and rotate operations handling.
2451 (set_address_base): Use must_be_base_p predicate.
2452 (set_address_index): Use must_be_index_p predicate.
2454 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
2455 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2456 Sergey Lega <sergey.s.lega@intel.com>
2457 Anna Tikhonova <anna.tikhonova@intel.com>
2458 Ilya Tocar <ilya.tocar@intel.com>
2459 Andrey Turetskiy <andrey.turetskiy@intel.com>
2460 Ilya Verbin <ilya.verbin@intel.com>
2461 Kirill Yukhin <kirill.yukhin@intel.com>
2462 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2464 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
2466 (ix86_expand_vector_move_misalign): Support new unaligned load and
2467 stores and use new names.
2468 (CODE_FOR_sse2_storedqu): Rename to ...
2469 (CODE_FOR_sse2_storedquv16qi): ... this.
2470 (CODE_FOR_sse2_loaddqu): Rename to ...
2471 (CODE_FOR_sse2_loaddquv16qi): ... this.
2472 (CODE_FOR_avx_loaddqu256): Rename to ...
2473 (CODE_FOR_avx_loaddquv32qi): ... this.
2474 (CODE_FOR_avx_storedqu256): Rename to ...
2475 (CODE_FOR_avx_storedquv32qi): ... this.
2476 * config/i386/i386.md (fpint_logic): New.
2477 * config/i386/sse.md (VMOVE): Extend for AVX512.
2481 (VI_UNALIGNED_LOADSTORE): Ditto.
2482 (sse2_avx_avx512f): Ditto.
2483 (sse2_avx2): Extend for AVX512.
2484 (sse4_1_avx2): Ditto.
2485 (avx2_avx512f): New.
2486 (sse): Extend for AVX512.
2489 (avxsizesuffix): Ditto.
2490 (sseintvecmode): Ditto.
2492 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
2493 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
2494 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
2495 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
2496 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
2497 (<sse2_avx_avx512f>_storedqu<mode): ... this.
2498 (<sse>_movnt<mode>): Replace constraint "x" with "v".
2499 (STORENT_MODE): Extend for AVX512.
2500 (*absneg<mode>2): Replace constraint "x" with "v".
2501 (*mul<mode>3): Ditto.
2502 (*ieee_smin<mode>3): Ditto.
2503 (*ieee_smax<mode>3): Ditto.
2504 (avx_cmp<mode>3): Replace VF with VF_128_256.
2505 (*<sse>_maskcmp<mode>3_comm): Ditto.
2506 (<sse>_maskcmp<mode>3): Ditto.
2507 (<sse>_andnot<mode>3): Extend for AVX512.
2508 (<code><mode>3, anylogic): Replace VF with VF_128_256.
2509 (<code><mode>3, fpint_logic): New.
2510 (*<code><mode>3): Extend for AVX512.
2511 (avx512flogicsuff): New.
2512 (avx512f_<logic><mode>): Ditto.
2513 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
2515 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
2516 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
2517 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
2518 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
2519 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
2520 (xop_vpermil2<mode>3): Ditto.
2521 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
2522 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
2523 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
2524 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
2526 2013-09-25 Tom Tromey <tromey@redhat.com>
2528 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
2529 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
2530 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
2531 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
2532 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
2533 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
2534 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
2535 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
2536 (GRAPHITE_HTAB_H): Remove.
2538 2013-09-25 Tom Tromey <tromey@redhat.com>
2540 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
2542 2013-09-25 Tom Tromey <tromey@redhat.com>
2544 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
2546 2013-09-25 Tom Tromey <tromey@redhat.com>
2548 * config/i386/t-i386 (i386.o): Remove.
2549 (i386-c.o): Use COMPILE and POSTCOMPILE.
2551 2013-09-25 Tom Tromey <tromey@redhat.com>
2553 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
2555 2013-09-25 Tom Tromey <tromey@redhat.com>
2557 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
2558 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
2559 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
2560 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
2561 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
2562 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
2563 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
2564 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
2565 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
2566 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
2567 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
2568 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
2569 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
2570 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
2571 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
2572 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
2573 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
2574 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
2575 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
2576 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
2577 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
2578 (tree-ssa-pre.o, tree-ssa-sccvn.o)
2579 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
2580 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
2581 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
2582 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
2583 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
2584 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
2585 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
2586 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
2587 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
2588 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
2589 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
2590 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
2591 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
2592 (graphite-clast-to-gimple.o, graphite-dependences.o)
2593 (graphite-interchange.o, graphite-poly.o)
2594 (graphite-scop-detection.o, graphite-sese-to-poly.o)
2595 (graphite-optimize-isl.o, tree-vect-loop.o)
2596 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
2597 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
2598 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
2599 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
2600 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
2601 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
2602 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
2603 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
2604 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
2605 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
2606 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
2607 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
2608 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
2609 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
2610 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
2611 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
2612 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
2613 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
2614 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
2615 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
2616 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
2617 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
2618 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
2619 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
2620 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
2621 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
2622 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
2623 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
2624 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
2625 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
2626 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
2627 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
2628 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
2629 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
2630 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
2631 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
2632 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
2633 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
2634 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
2635 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
2636 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
2637 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
2638 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
2639 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
2640 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
2641 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
2642 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
2643 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
2644 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
2645 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
2646 (gcov-dump.o): Remove.
2647 (default-c.o): Use COMPILE and POSTCOMPILE.
2648 (CFLAGS-gcc.o): New variable.
2649 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
2651 2013-09-25 Tom Tromey <tromey@redhat.com>
2653 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
2654 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
2655 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
2656 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
2657 (c-family/c-pragma.o, c-family/c-pretty-print.o)
2658 (c-family/c-semantics.o, c-family/c-ada-spec.o)
2659 (c-family/array-notation-common.o, c-family/stub-objc.o)
2660 (c-family/c-ubsan.o): Remove.
2662 2013-09-25 Tom Tromey <tromey@redhat.com>
2664 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
2666 2013-09-25 Tom Tromey <tromey@redhat.com>
2668 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
2669 to add -DENABLE_SHARED_LIBGCC.
2670 (gcc.o): Don't use subshell.
2672 2013-09-25 Tom Tromey <tromey@redhat.com>
2674 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
2675 * configure.ac: Don't invoke AM_PROG_CC_C_O.
2676 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
2677 * configure, config.in: Rebuild.
2679 2013-09-25 Tom Tromey <tromey@redhat.com>
2681 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
2682 (COMPILE, POSTCOMPILE): New variables.
2683 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
2684 (DEPFILES): New variable.
2685 Include ".Po" files.
2686 * configure.ac: Add checks for dependency checking.
2687 * configure, aclocal.m4: Regenerate.
2689 2013-09-25 Tom Tromey <tromey@redhat.com>
2691 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
2692 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
2694 2013-09-25 Tom Tromey <tromey@redhat.com>
2696 * Makefile.in (generated_files): Add options.h,
2697 target-hooks-def.h, insn-opinit.h,
2698 common/common-target-hooks-def.h, pass-instances.def,
2699 c-family/c-target-hooks-def.h.
2701 2013-09-25 Jeff Law <law@redhat.com>
2703 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
2704 than foo[foo.length () - 1] to access last member in a vec.
2705 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
2707 2013-09-25 Richard Biener <rguenther@suse.de>
2710 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
2712 2013-09-25 Jan Hubicka <jh@suse.cz>
2714 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
2717 2013-09-25 Marek Polacek <polacek@redhat.com>
2720 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
2721 when determining the type name.
2723 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
2725 * config/sh/sh.md: Fix formatting.
2727 2013-09-24 Xinliang David Li <davidxl@google.com>
2729 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
2730 max peel iterations parameter.
2731 * param.def: New parameter.
2732 * doc/invoke.texi: Document New parameter.
2734 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
2736 * gimple-pretty-print.c: Various whitespace tweaks.
2737 * tree-core.h: Likewise.
2738 * tree-pretty-print.c: Likewise.
2739 * tree-ssa-alias.c: Likewise.
2740 * tree-ssa-copy.c: Likewise.
2741 * tree-ssanames.c: Likewise.
2742 * tree-ssanames.h: Likewise.
2743 * tree-vrp.c: Likewise.
2745 2013-09-24 Alan Modra <amodra@gmail.com>
2749 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
2750 for output operands that disallow regs. Don't use EXPAND_WRITE on
2753 2013-09-24 Richard Biener <rguenther@suse.de>
2756 * tree.c (reference_alias_ptr_type): Move ...
2757 * alias.c (reference_alias_ptr_type): ... here and implement
2758 in terms of the new reference_alias_ptr_type_1.
2759 (ref_all_alias_ptr_type_p): New helper.
2760 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
2761 use ref_all_alias_ptr_type_p.
2762 (get_deref_alias_set): Add flag_strict_aliasing check here.
2763 (reference_alias_ptr_type_1): New function, split out from ...
2764 (get_alias_set): ... here.
2765 (alias_ptr_types_compatible_p): New function.
2766 * alias.h (reference_alias_ptr_type): Declare.
2767 (alias_ptr_types_compatible_p): Likewise.
2768 * tree.h (reference_alias_ptr_type): Remove.
2769 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
2770 to compare MEM_REF alias types.
2772 2013-09-24 Richard Biener <rguenther@suse.de>
2774 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
2776 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
2778 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
2779 reload helper function arrays into a single array reg_addr.
2780 (reload_fpr_gpr): Likewise.
2781 (reload_gpr_vsx): Likewise.
2782 (reload_vsx_gpr): Likewise.
2783 (struct rs6000_reg_addr): Likewise.
2784 (reg_addr): Likewise.
2785 (rs6000_debug_reg_global): Change rs6000_vector_reload,
2786 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
2787 (rs6000_init_hard_regno_mode_ok): Likewise.
2788 (rs6000_secondary_reload_direct_move): Likewise.
2789 (rs6000_secondary_reload): Likewise.
2791 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
2792 constraints: wu, ww, and wy. Repurpose wv constraint added during
2793 power8 changes. Put wg constraint in alphabetical order.
2795 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
2796 for future work to add ISA 2.07 VSX single precision support.
2797 (-mvsx-scalar-double): Change default from -1 to 1, update
2798 documentation comment.
2799 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
2800 (-mupper-regs-df): New debug switch to control whether DF values
2801 can go in the traditional Altivec registers.
2802 (-mupper-regs-sf): New debug switch to control whether SF values
2803 can go in the traditional Altivec registers.
2805 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
2807 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
2808 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
2809 Add new constraints, wu/ww/wy. Repurpose wv constraint.
2810 (rs6000_debug_legitimate_address_p): Print if we are running
2811 before, during, or after reload.
2812 (rs6000_secondary_reload): Add a comment.
2813 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
2815 * config/rs6000/constraints.md (wa constraint): Sort w<x>
2816 constraints. Update documentation string.
2817 (wd constraint): Likewise.
2818 (wf constraint): Likewise.
2819 (wg constraint): Likewise.
2820 (wn constraint): Likewise.
2821 (ws constraint): Likewise.
2822 (wt constraint): Likewise.
2823 (wx constraint): Likewise.
2824 (wz constraint): Likewise.
2825 (wu constraint): New constraint for ISA 2.07 SFmode scalar
2827 (ww constraint): Likewise.
2828 (wy constraint): Likewise.
2829 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
2830 the previous submissions.
2831 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
2833 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
2835 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
2836 (REG_BR_PROB): Say that the probability is stored in an int_list.
2837 * reg-notes.def: Update commentary to mention INT_LIST.
2838 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
2839 (INT_LIST): New rtx.
2840 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
2841 * rtlanal.c (int_reg_note_p): New function.
2842 (alloc_reg_note): Assert that the note does not have an int argument.
2843 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
2844 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
2845 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
2846 rather than an INSN_LIST. Handle INT_LIST.
2847 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
2848 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
2849 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
2850 Manipulate them as ints rather than rtxes.
2851 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
2852 * regmove.c (copy_src_to_dest): Likewise.
2853 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
2855 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
2856 into the cases that need it.
2857 * config/arm/arm.c (arm_unwind_emit): Likewise.
2859 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
2860 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
2861 * loop-doloop.c (add_test, doloop_modify): Likewise.
2862 * loop-unswitch.c (compare_and_jump_seq): Likewise.
2863 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
2864 * predict.c (combine_predictions_for_insn): Likewise.
2865 * print-rtl.c (print_rtx): Handle INT_LIST.
2866 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
2867 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
2868 * config/arm/arm.c (emit_unlikely_jump): Likewise.
2869 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
2870 (ix86_split_fp_branch, predict_jump): Likewise.
2871 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
2872 * config/sh/sh.c (expand_cbranchsi4): Likewise.
2873 * config/spu/spu.c (ea_load_store_inline): Likewise.
2875 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
2876 value of a REG_BR_PROB note.
2877 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
2878 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
2879 * emit-rtl.c (try_split): Likewise.
2880 * predict.c (br_prob_note_reliable_p): Likewise.
2881 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
2882 * reorg.c (mostly_true_jump): Likewise.
2883 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
2884 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
2885 * config/i386/i386.c (ix86_print_operand): Likewise.
2886 * config/ia64/ia64.c (ia64_print_operand): Likewise.
2887 * config/mmix/mmix.c (mmix_print_operand): Likewise.
2888 * config/rs6000/rs6000.c (output_cbranch): Likewise.
2889 * config/s390/s390.c (s390_expand_tbegin): Likewise.
2890 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
2891 * config/sparc/sparc.c (output_cbranch): Likewise.
2892 * config/spu/spu.c (get_branch_target): Likewise.
2893 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
2894 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
2896 2013-09-23 Jan Hubicka <jh@suse.cz>
2898 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
2900 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
2901 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
2903 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
2904 (gimple_fold_call): Dump inconsistent devirtualizations; add
2905 sanity check for type based devirtualizations.
2906 * ipa-prop.c: Include ipa-utils.h
2907 (ipa_intraprocedural_devirtualization): Add sanity check.
2908 (try_make_edge_direct_virtual_call): Likewise.
2910 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
2912 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
2913 assignment statements.
2915 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
2917 * gimple-pretty-print.c (dump_ssaname_info): New function.
2918 (dump_gimple_phi): Call it.
2919 (pp_gimple_stmt_1): Likewise.
2920 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
2921 (range_info_def): Declare.
2922 * tree-pretty-print.c (pp_double_int): New function.
2923 (dump_generic_node): Call it.
2924 * tree-pretty-print.h (pp_double_int): Declare.
2925 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
2926 * tree-ssanames.h (range_info_def): New structure.
2927 (value_range_type): Move definition here.
2928 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
2930 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
2932 (set_range_info): New function.
2933 (get_range_info): Likewise.
2934 (duplicate_ssa_name_range_info): Likewise.
2935 (duplicate_ssa_name_fn): Check pointer type and call
2936 duplicate_ssa_name_range_info.
2937 * tree-ssa-copy.c (fini_copy_prop): Likewise.
2938 * tree-vrp.c (value_range_type): Remove definition, now in
2940 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
2941 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
2942 (SSA_NAME_RANGE_INFO): New macro.
2944 2013-09-23 Richard Biener <rguenther@suse.de>
2946 PR tree-optimization/58464
2947 * tree-ssa-pre.c (phi_trans_lookup): Remove.
2948 (phi_trans_add): Change to add conditionally on being not present.
2949 (phi_translate_1): Remove recursion detection here.
2950 (phi_translate): Pre-seed the cache with NULL to catch
2951 recursion here in a more generic way.
2952 (bitmap_find_leader): Adjust comment.
2953 (get_representative_for): Dump value-numbers.
2954 (create_expression_by_pieces): Likewise.
2955 (insert_into_preds_of_block): Likewise.
2957 2013-09-23 Christian Bruel <christian.bruel@st.com>
2960 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
2961 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
2963 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
2966 2013-09-20 Renlin Li <renlin.li@arm.com>
2968 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2970 (aarch64_expand_epilogue): Likewise.
2971 (aarch64_legitimize_reload_address): Likewise.
2973 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
2975 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
2976 NULL_TREE before pushing them onto the vector. Likewise for labels.
2978 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
2980 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
2981 comparison operators when -fno-trapping-math is in effect.
2982 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
2983 comparison operators in TFmode and assert that unsupported operators
2985 (ia64_print_operand): Likewise.
2987 2013-09-21 Jan Hubicka <jh@suse.cz>
2989 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
2990 (sse_typeless_stores): Enable for core
2991 (sse_load0_by_pxor): Likewise.
2992 (four_jump_limit): Disable for core.
2993 (pad_returns): Likewise.
2994 (avoid_vector_decode): Likewise.
2995 (fuse_cmp_and_branch): Enable for cores.
2996 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
2998 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
3001 * config/pa/pa.c (pa_option_override): Disable auto increment and
3002 decrement instructions until reload is completed.
3004 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
3005 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
3006 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
3008 2013-09-20 DJ Delorie <dj@redhat.com>
3009 Nick Clifton <nickc@redhat.com>
3011 * config/rl78/rl78.c: Various whitespace and comment tweaks.
3012 (need_to_save): Save bank 0 on interrupts.
3013 (characterize_address): Strip far address wrappers.
3014 (rl78_as_legitimate_address): Likewise.
3015 (transcode_memory_rtx): Likewise.
3016 (rl78_peep_movhi_p): Disable this peephole after devirt.
3017 (rl78_propogate_register_origins): Forget all origins when a
3019 * config/rl78/rl78-virt.md: Various whitespace tweaks.
3020 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
3022 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
3023 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
3024 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
3026 * config/rl78/constraints.md: Various whitespace and paren tweaks.
3028 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
3030 * config/pa/pa.md: In "scc" insn patterns, change output template to
3031 handle const0_rtx in reg_or_0_operand operands.
3033 2013-09-20 Martin Husemann <martin@NetBSD.org>
3036 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
3037 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
3039 2013-09-20 Richard Biener <rguenther@suse.de>
3042 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
3043 remembering SSA name version and block index.
3044 (new_scev_info_str): Adjust.
3045 (hash_scev_info): Likewise. Also hash the block index.
3046 (eq_scev_info): Adjust.
3047 (find_var_scev_info): Likewise.
3048 (struct instantiate_cache_entry): Remove.
3049 (struct instantiate_cache_type): Use a htab to map name, block
3051 (instantiate_cache_type::~instantiate_cache_type): Adjust.
3052 (get_instantiated_value_entry): Likewise.
3053 (hash_idx_scev_info, eq_idx_scev_info): New functions.
3054 (instantiate_scev_name): Adjust.
3056 2013-09-20 Jeff Law <law@redhat.com>
3058 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
3060 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
3062 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
3063 Call aarch64_simd_expand_args to update op[argc].
3065 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3067 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
3070 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3072 * gengtype.c (file_rules): Added rule for *.cc files.
3073 (get_output_file_with_visibility): Give fatal message when no
3076 2013-09-20 Renlin Li <renlin.li@arm.com>
3078 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
3079 (aarch64_expand_epilogue): Likewise.
3080 (aarch64_legitimize_reload_address): Likewise.
3082 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
3085 * expr.c (expand_assignment): Remove misalignp code path.
3087 2013-09-20 Marek Polacek <polacek@redhat.com>
3090 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
3091 TYPE_PRECISION. Add asserts.
3093 2013-09-20 Richard Biener <rguenther@suse.de>
3095 PR tree-optimization/58453
3096 * tree-loop-distribution.c (distribute_loop): Apply the cost
3097 model for -ftree-loop-distribute-patterns, too.
3099 2013-09-20 Richard Biener <rguenther@suse.de>
3102 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
3103 make type comparison less strict.
3105 2013-09-20 Alan Modra <amodra@gmail.com>
3107 * configure: Regenerate.
3108 * aclocal.m4: Regenerate.
3110 2013-09-20 Marek Polacek <polacek@redhat.com>
3113 * doc/extend.texi: Document that attribute used is meant to be used
3114 on variables with static storage duration.
3116 2013-09-19 Jakub Jelinek <jakub@redhat.com>
3118 PR tree-optimization/58472
3119 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
3120 simd_lane_access set inv_p = false.
3121 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
3122 the simduid magic VAR_DECL.
3124 2013-09-19 Jan Hubicka <jh@suse.cz>
3126 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
3128 2013-09-17 Jeff Law <law@redhat.com>
3130 * tree-ssa-dom.c (record_temporary_equivalences): New function
3131 split out of dom_opt_dom_walker::after_dom_children.
3132 (dom_opt_dom_walker::thread_across_edge): Move common code
3133 in here from dom_opt_dom_walker::after_dom_children.
3134 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
3136 2013-09-19 Jan Hubicka <jh@suse.cz>
3138 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
3139 (TARGET_GENERIC): Use PROCESOR_GENERIC
3140 (enum processor_type): Unify generic32 and 64.
3141 * i386.md (cpu): Likewise.
3142 * x86-tune.def (use_leave): Enable for generic32.
3143 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
3144 * athlon.md: Change generic64 to generic in all occurences.
3145 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
3146 (ix86_target_macros_internal): Likewise.
3147 * driver-i386.c (host_detect_local_cpu): Likewise.
3148 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
3150 (generic_memcpy, generic_memset, generic_cost): This one.
3151 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
3152 (m_GENERIC32, m_GENERIC64): Remove.
3153 (m_GENERIC): Turn into one flag.
3154 (processor_target): Unify generic tunnings.
3155 (ix86_option_override_internal): Replace generic32/64 by generic.
3156 (ix86_issue_rate): Likewise.
3157 (ix86_adjust_cost): Likewise.
3159 2013-09-19 Jan Hubicka <jh@suse.cz>
3161 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
3162 of speculative flag.
3164 2013-09-19 Jakub Jelinek <jakub@redhat.com>
3166 * omp-low.c (expand_omp_sections): Always pass len - 1 to
3167 GOMP_sections_start, even if !exit_reachable.
3169 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
3171 * lra-constraints.c (need_for_all_save_p): Use macro
3172 HARD_REGNO_CALL_PART_CLOBBERED.
3173 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
3174 set up pseudo conflict hard regs.
3176 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
3179 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
3182 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
3184 PR rtl-optimization/58438
3185 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
3186 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
3187 from previous subpasses.
3189 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
3191 * arm.c (arm_get_frame_offsets): Validate architecture supports
3192 LDRD/STRD before accepting the tuning preference.
3193 (arm_expand_prologue): Likewise.
3194 (arm_expand_epilogue): Likewise.
3196 2013-09-18 Richard Biener <rguenther@suse.de>
3198 PR tree-optimization/58417
3199 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
3200 have chrecs with symbols defined in the loop as operands.
3201 (chrec_fold_multiply): Likewise.
3202 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
3203 parameters before folding binary operations.
3204 (struct instantiate_cache_entry_hasher): Remove.
3205 (struct instantiate_cache_type): Use a pointer-map.
3206 (instantiate_cache_type::instantiate_cache_type): New function.
3207 (instantiate_cache_type::get): Likewise.
3208 (instantiate_cache_type::set): Likewise.
3209 (instantiate_cache_type::~instantiate_cache_type): Adjust.
3210 (get_instantiated_value_entry): Likewise.
3211 (global_cache): New global.
3212 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
3213 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
3214 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
3215 (instantiate_scev_name): Adjust.
3216 (instantiate_scev): Construct global instead of local cache.
3217 (resolve_mixers): Likewise.
3219 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
3220 Paolo Carlini <paolo.carlini@oracle.com>
3223 * doc/implement-cxx.texi: Fix references to the C++ standards.
3225 2013-09-18 Jakub Jelinek <jakub@redhat.com>
3227 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
3228 * tree-vect-data-refs.c (vect_analyze_data_refs): For
3229 simd_lane_access drs, update also DR_ALIGNED_TO.
3231 2013-09-18 Marek Polacek <polacek@redhat.com>
3234 * doc/extend.texi: Document no_sanitize_undefined attribute.
3235 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
3236 no_sanitize_undefined attribute.
3238 2013-09-18 Nick Clifton <nickc@redhat.com>
3240 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
3241 (ASM_DECLARE_FUNCTION_NAME): Define.
3243 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
3245 * compare-elim.c (find_comparison_dom_walker): New class
3246 (find_comparisons_in_bb): Rename to
3247 find_comparison_dom_walker::before_dom_children
3248 (find_comparisons): Adjust
3249 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
3251 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
3252 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
3253 (init_walk_dominator_tree): Remove declaration.
3254 (fini_walk_dominator_tree): Remove declaration.
3255 * fwprop.c (single_def_use_dom_walker): New class
3256 (single_def_use_enter_block): Convert to
3257 single_def_use_dom_walker::before_dom_children.
3258 (single_def_use_leave_block): Convert to
3259 single_def_use_dom_walker::after_dom_children.
3260 (build_single_def_use_links): Adjust.
3261 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
3263 (find_candidates_in_block): Convert to
3264 find_candidates_dom_walker::before_dom_children.
3265 (execute_strength_reduction): Adjust.
3266 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
3267 (sese_dom_walker): New class.
3268 (sese_dom_walker::sese_dom_walker): New constructor.
3269 (sese_dom_walker::~sese_dom_walker): New destructor.
3270 (build_sese_conditions_before): Convert to
3271 sese_dom_walker::before_dom_children.
3272 (build_sese_conditions_after): Convert to
3273 sese_dom_walker::after_dom_children.
3274 (build_poly_scop): Adjust
3275 * tree-into-ssa.c (rewrite_dom_walker): New class
3276 (rewrite_enter_block): Convert to
3277 rewrite_dom_walker::before_dom_children.
3278 (rewrite_leave_block): Convert to
3279 rewrite_dom_walker::after_dom_children.
3280 (rewrite_update_dom_walker): New class.
3281 (rewrite_update_enter_block): Convert to
3282 rewrite_update_dom_walker::before_dom_children.
3283 (rewrite_update_leave_block): Convert to
3284 rewrite_update_dom_walker::after_dom_children.
3285 (rewrite_blocks, rewrite_into_ssa): Adjust.
3286 (mark_def_dom_walker): New class.
3287 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
3288 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
3289 (mark_def_sites_blocks): Convert to
3290 mark_def_dom_walker::before_dom_children.
3291 (mark_def_site_blocks): Remove.
3292 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
3293 (tree_ssa_dominator_optimize): Adjust.
3294 (dom_thread_across_edge): Convert to method
3295 dom_opt_dom_walker::thread_across_edge.
3296 (dom_opt_enter_block): Convert to member function
3297 dom_opt_dom_walker::before_dom_children.
3298 (dom_opt_leave_block): Convert to member function
3299 dom_opt_dom_walker::after_dom_children.
3300 * tree-ssa-dse.c (dse_dom_walker): New class.
3301 (dse_enter_block): Convert to member function
3302 dse_dom_walker::before_dom_children.
3303 (tree_ssa_dse): Adjust.
3304 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
3305 (determine_invariantness_stmt): Convert to method
3306 invariantness_dom_walker::before_dom_children.
3307 (determine_invariantness): Remove
3308 (move_computations_dom_walker): New class.
3309 (move_computations_stmt): Convert to method
3310 move_computations_dom_walker::before_dom_children.
3311 (move_computations, tree_ssa_lim): Adjust.
3312 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
3313 (nt_init_block): Make method
3314 notrappping_dom_walker::before_dom_children.
3315 (nt_fini_block): Make
3316 method nontrapping_dom_walker::after_dom_children.
3317 (get_non_trapping): Adjust.
3318 * tree-ssa-pre.c (eliminate_dom_walker): New class.
3319 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
3320 (eliminate_leave_block): Make method.
3321 eliminate_dom_walker::after_dom_children.
3323 * tree-ssa-strlen.c (strlen_dom_walker): New class.
3324 (strlen_enter_block): Make method
3325 strlen_dom_walker::before_dom_children.
3326 (strlen_leave_block): Make
3327 method strlen_dom_walker::after_dom_children.
3328 (tree_ssa_strlen): Adjust.
3329 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
3330 (tree_ssa_uncprop): Adjust.
3331 (uncprop_leave_block): Make method
3332 uncprop_dom_walker::after_dom_children.
3333 (uncprop_leave_block): Make method
3334 uncprop_dom_walker::before_dom_children.
3336 2013-09-18 Bin Cheng <bin.cheng@arm.com>
3338 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
3339 branch in same basic block. Check both src and dest of the move insn.
3341 2013-09-17 Nick Clifton <nickc@redhat.com>
3343 * config/rl78/rl78-real.md (bf): New pattern.
3345 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
3346 (rl78_print_operand): Do not put a # before a %B.
3347 * config/rl78/rl78.opt: Tweak doc strings.
3349 2013-09-17 DJ Delorie <dj@redhat.com>
3351 * config/rl78/constraints.md (Wcv): Allow up to $r31.
3352 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
3353 (rl78_option_override): Likewise, if -mallregs.
3354 (is_virtual_register): Likewise.
3355 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
3356 (REGNO_OK_FOR_BASE_P): Likewise.
3357 * config/rl78/rl78.opt (-mallregs): New.
3359 2013-09-17 Nick Clifton <nickc@redhat.com>
3361 * config/rl78/rl78.c (need_to_save): Change return type to bool.
3362 For interrupt functions: save all call clobbered registers if the
3363 interrupt handler is not a leaf function.
3364 (rl78_expand_prologue): Always recompute the frame information.
3365 For interrupt functions: only select bank 0 if one of the bank 0
3366 registers is going to be psuhed.
3368 2013-09-17 DJ Delorie <dj@redhat.com>
3370 * config/rl78/constraints.md: For each W* constraint, rename to C*
3371 and create a W* constraint that checks for an optional ES: prefix
3373 * config/rl78/rl78.md (UNS_ES_ADDR): New.
3374 (es_addr): New. Used to wrap far addresses.
3375 * config/rl78/rl78-protos.h (rl78_es_addr): New.
3376 (rl78_es_base): New.
3377 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
3378 wrapped far addresses.
3379 (rl78_print_operand_1): Unwrap far addresses before processing.
3380 (rl78_lo16): Wrap far addresses in unspecs.
3381 (rl78_es_addr): New.
3382 (rl78_es_base): New.
3383 (insn_ok_now): Check for not-yet-wrapped far addresses.
3384 (transcode_memory_rtx): Properly re-wrap far addresses.
3386 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
3388 * config/sparc/t-rtems: Add leon3 multilibs.
3390 2013-09-17 Cong Hou <congh@google.com>
3392 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
3393 when checking the dot production pattern. The type of rhs operand
3394 of multiply is now checked correctly.
3396 2013-09-17 Jeff Law <law@redhat.com>
3398 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3399 edge implied equivalences into successor phis.
3400 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
3401 here from tree-ssa-threadedge.c.
3402 (mark_threaded_blocks): When threading through a joiner, if both
3403 successors of the joiner's clone reach the same block, verify the
3404 PHI arguments are equal. If not, cancel the jump threading request.
3405 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
3406 tree-ssa-threadupdate.c
3407 (thread_across_edge): Don't check PHI argument equality when
3408 threading through joiner block here.
3410 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
3412 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
3413 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
3414 (warn_uninit, warn_uninitialized_vars,
3415 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
3416 Move to tree-ssa-uninit.c.
3417 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
3418 (has_undefined_value_p): New. Pass dependant parts of
3419 ssa_undefined_value_p.
3420 (uninit_undefined_value_p): Use has_undefined_value_p.
3421 (warn_uninit, warn_uninitialized_vars,
3422 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
3423 Move from tree-ssa.c.
3424 * tree-ssa.h: Adjust prototypes.
3426 2013-09-17 Jan Hubicka <jh@suse.cz>
3429 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
3430 * ipa-inline.c (can_inline_edge_p): Do not downgrade
3431 FUNCTION_NOT_OPTIMIZED.
3432 * ipa-inline-analysis.c (compute_inline_parameters): Function
3433 not optimized is not inlinable unless it is alwaysinline.
3434 (inline_analyze_function): Force calls in not optimized
3435 function not inlinable.
3437 2013-09-17 Jan Hubicka <jh@suse.cz>
3440 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
3442 * ipa.c (function_and_variable_visibility): Likewise.
3443 * ipa-profile.c (ipa_profile): Likewise.
3445 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
3448 * cgraph.c (cgraph_function_body_availability): Check for ifunc
3449 attribute, and don't inline the resolver in this case.
3451 2013-09-17 Teresa Johnson <tejohnson@google.com>
3453 * coverage.c (get_coverage_counts): Add missing newline.
3455 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3457 PR tree-optimization/58088
3458 * fold-const.c (mask_with_trailing_zeros): New function.
3459 (fold_binary_loc): Make sure we don't recurse infinitely
3460 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
3461 Use mask_with_trailing_zeros where appropriate.
3463 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
3465 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
3466 of 'prev' var to get better distance estimation.
3468 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
3470 * tree-inline.h (struct copy_body_data): Add transform_parameter.
3471 * tree-inline.c (is_parameter_of): New predicate.
3472 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
3473 a parameter has been remapped.
3474 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
3475 (optimize_inline_calls): Initialize transform_parameter.
3476 (copy_gimple_seq_and_replace_locals): Likewise.
3477 (tree_function_versioning): Likewise.
3478 (maybe_inline_call_in_expr): Likewise.
3480 2013-09-17 Nick Clifton <nickc@redhat.com>
3482 * config/msp430/msp430-protos.h: Add prototypes for new functions.
3483 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
3485 (is_attr_func): New function.
3486 (msp430_is_interrupt_func): New function.
3487 (is_naked_func): New function.
3488 (is_reentrant_func): New function.
3489 (is_critical_func): New function.
3490 (msp430_start_function): Add annotations for function attributes.
3491 (msp430_attr): New function.
3492 (msp430_attribute_table): New.
3493 (msp430_function_section): New function.
3494 (TARGET_ASM_FUNCTION_SECTION): Define.
3495 (msp430_builtin): New enum.
3496 (msp430_init_builtins): New function.
3497 (msp430_builtin_devl): New function.
3498 (msp430_expand_builtin): New function.
3499 (TARGET_INIT_BUILTINS): Define.
3500 (TARGET_EXPAND_BUILTINS): Define.
3501 (TARGET_BUILTIN_DECL): Define.
3502 (msp430_expand_prologue): Add support for naked, interrupt,
3503 critical and reentrant functions.
3504 (msp430_expand_epilogue): Likewise.
3505 (msp430_print_operand): Handle 'O' character.
3506 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
3508 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
3509 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
3510 (pushm): Use a 'n' rather than an 'i' constraint.
3511 (msp_return): Add generation of the interrupt return instruction.
3512 (disable_interrupts): New pattern.
3513 (enable_interrupts): New pattern.
3514 (push_intr_state): New pattern.
3515 (pop_intr_state): New pattern.
3516 (bic_SR): New pattern.
3517 (bis_SR): New pattern.
3518 * doc/extend.texi: Document MSP430 function attributes and builtin
3521 2013-09-17 Richard Biener <rguenther@suse.de>
3523 PR tree-optimization/58432
3524 * tree-loop-distribution.c (tree_loop_distribution): Also
3525 scan PHIs for outside loop uses and seed a partition from them.
3527 2013-09-17 Bin Cheng <bin.cheng@arm.com>
3529 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
3530 (restructure_reference): Call backtrace_base_for_ref.
3532 2013-09-17 Alan Modra <amodra@gmail.com>
3535 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
3538 2013-09-16 DJ Delorie <dj@redhat.com>
3540 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
3541 vregs location for RL78/G10.
3542 (rl78_expand_prologue): Avoid SEL on G10.
3543 (rl78_expand_epilogue): Likewise.
3544 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
3545 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
3546 __RL78_G10__ when appropriate.
3547 (ASM_SPEC): Pass -mg10 along to the assembler.
3548 * config/rl78/rl78.md (sel_rb): Disable for G10.
3549 * config/rl78/rl78.opt: Add -mg10 option.
3550 * config/rl78/t-rl78: Add -mg10 multilib.
3552 2013-09-16 Xinliang David Li <davidxl@google.com>
3554 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
3555 * omp-low.c (omp_max_vf): Ditto.
3556 (expand_omp_simd): Ditto.
3557 * tree-vectorizer.c (vectorize_loops): Ditto.
3558 (gate_vect_slp): Ditto.
3559 (gate_increase_alignment): Ditto.
3560 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
3561 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
3562 (gate_tree_vectorize): Name change.
3563 (tree_vectorize): Ditto.
3564 (pass_vectorize::gate): Call new function.
3565 (pass_vectorize::execute): Ditto.
3566 * opts.c: O3 default setting change.
3567 (finish_options): Check new flag.
3568 * doc/invoke.texi: Document new flags.
3569 * common.opt: New flags.
3571 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
3573 * doc/tm.texi.in (Cond Exec Macros): Remove node.
3574 (Condition Code): Don't reference it.
3575 * doc/tm.texi: Regenerate.
3577 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
3580 * lra-constraints.c (undo_optional_reloads): Consider all optional
3581 reload even if it did not get a hard reg.
3583 2013-09-16 Teresa Johnson <tejohnson@google.com>
3585 * dumpfile.c (dump_loc): Remove newline emission.
3586 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
3587 emission to dump_printf_loc calls where missing.
3588 (vect_mark_for_runtime_alias_test): Ditto.
3589 (vect_analyze_data_ref_dependence): Ditto.
3590 (vect_analyze_data_ref_dependences): Ditto.
3591 (vect_slp_analyze_data_ref_dependence): Ditto.
3592 (vect_slp_analyze_data_ref_dependences): Ditto.
3593 (vect_compute_data_ref_alignment): Ditto.
3594 (vect_update_misalignment_for_peel): Ditto.
3595 (vect_verify_datarefs_alignment): Ditto.
3596 (vector_alignment_reachable_p): Ditto.
3597 (vect_get_data_access_cost): Ditto.
3598 (vect_enhance_data_refs_alignment): Ditto.
3599 (vect_find_same_alignment_drs): Ditto.
3600 (vect_analyze_data_refs_alignment): Ditto.
3601 (vect_analyze_group_access): Ditto.
3602 (vect_analyze_data_ref_access): Ditto.
3603 (vect_analyze_data_ref_accesses): Ditto.
3604 (vect_prune_runtime_alias_test_list): Ditto.
3605 (vect_analyze_data_refs): Ditto.
3606 (vect_create_addr_base_for_vector_ref): Ditto.
3607 (vect_create_data_ref_ptr): Ditto.
3608 (vect_grouped_store_supported): Ditto.
3609 (vect_grouped_load_supported): Ditto.
3610 * value-prof.c (check_counter): Ditto.
3611 (check_ic_target): Ditto.
3612 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
3613 (vect_recog_widen_mult_pattern): Ditto.
3614 (vect_recog_widen_sum_pattern): Ditto.
3615 (vect_recog_over_widening_pattern): Ditto.
3616 (vect_recog_widen_shift_pattern): Ditto.
3617 (vect_recog_rotate_pattern): Ditto.
3618 (vect_recog_vector_vector_shift_pattern): Ditto.
3619 (vect_recog_divmod_pattern): Ditto.
3620 (vect_recog_mixed_size_cond_pattern): Ditto.
3621 (vect_recog_bool_pattern): Ditto.
3622 (vect_pattern_recog_1): Ditto.
3623 (vect_pattern_recog): Ditto.
3624 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
3625 (vect_is_simple_iv_evolution): Ditto.
3626 (vect_analyze_scalar_cycles_1): Ditto.
3627 (vect_get_loop_niters): Ditto.
3628 (vect_analyze_loop_1): Ditto.
3629 (vect_analyze_loop_form): Ditto.
3630 (vect_analyze_loop_operations): Ditto.
3631 (vect_analyze_loop_2): Ditto.
3632 (vect_analyze_loop): Ditto.
3633 (report_vect_op): Ditto.
3634 (vect_is_slp_reduction): Ditto.
3635 (vect_is_simple_reduction_1): Ditto.
3636 (vect_get_known_peeling_cost): Ditto.
3637 (vect_estimate_min_profitable_iters): Ditto.
3638 (vect_model_reduction_cost): Ditto.
3639 (vect_model_induction_cost): Ditto.
3640 (get_initial_def_for_induction): Ditto.
3641 (vect_create_epilog_for_reduction): Ditto.
3642 (vectorizable_reduction): Ditto.
3643 (vectorizable_induction): Ditto.
3644 (vectorizable_live_operation): Ditto.
3645 (vect_loop_kill_debug_uses): Ditto.
3646 (vect_transform_loop): Ditto.
3647 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
3648 (vect_stmt_relevant_p): Ditto.
3649 (process_use): Ditto.
3650 (vect_mark_stmts_to_be_vectorized): Ditto.
3651 (vect_model_simple_cost): Ditto.
3652 (vect_model_promotion_demotion_cost): Ditto.
3653 (vect_model_store_cost): Ditto.
3654 (vect_get_store_cost): Ditto.
3655 (vect_model_load_cost): Ditto.
3656 (vect_get_load_cost): Ditto.
3657 (vect_init_vector_1): Ditto.
3658 (vect_get_vec_def_for_operand): Ditto.
3659 (vect_finish_stmt_generation): Ditto.
3660 (vectorizable_call): Ditto.
3661 (vectorizable_conversion): Ditto.
3662 (vectorizable_assignment): Ditto.
3663 (vectorizable_shift): Ditto.
3664 (vectorizable_operation): Ditto.
3665 (vectorizable_store): Ditto.
3666 (vectorizable_load): Ditto.
3667 (vectorizable_condition): Ditto.
3668 (vect_analyze_stmt): Ditto.
3669 (vect_transform_stmt): Ditto.
3670 (vect_is_simple_use): Ditto.
3671 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
3672 (vect_can_advance_ivs_p): Ditto.
3673 (vect_update_ivs_after_vectorizer): Ditto.
3674 (vect_do_peeling_for_loop_bound): Ditto.
3675 (vect_gen_niters_for_prolog_loop): Ditto.
3676 (vect_update_inits_of_drs): Ditto.
3677 (vect_create_cond_for_alias_checks): Ditto.
3678 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
3679 (vect_build_slp_tree_1): Ditto.
3680 (vect_supported_load_permutation_p): Ditto.
3681 (vect_analyze_slp_instance): Ditto.
3682 (vect_analyze_slp): Ditto.
3683 (vect_make_slp_decision): Ditto.
3684 (vect_detect_hybrid_slp): Ditto.
3685 (vect_bb_vectorization_profitable_p): Ditto.
3686 (vect_slp_analyze_bb_1): Ditto.
3687 (vect_update_slp_costs_according_to_vf): Ditto.
3688 (vect_get_mask_element): Ditto.
3689 (vect_transform_slp_perm_load): Ditto.
3690 (vect_schedule_slp_instance): Ditto.
3691 (vect_schedule_slp): Ditto.
3692 (vect_slp_transform_bb): Ditto.
3693 * profile.c (read_profile_edge_counts): Ditto.
3694 (compute_branch_probabilities): Ditto.
3695 * coverage.c (get_coverage_counts): Ditto.
3697 2013-09-16 Diego Novillo <dnovillo@google.com>
3699 * tree-core.h: Add missing comment lines from refactoring of tree.h.
3701 2013-09-16 Jan Hubicka <jh@suse.cz>
3703 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
3704 abstract functions; for static functions check the presence of body.
3706 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3708 * config/aarch64/aarch64-simd-builtins.def (fma): New.
3709 * config/aarch64/aarch64-simd.md
3710 (aarch64_mla_elt<mode>): New.
3711 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3712 (aarch64_mls_elt<mode>): Likewise.
3713 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3714 (aarch64_fma4_elt<mode>): Likewise.
3715 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3716 (aarch64_fma4_elt_to_128v2df): Likewise.
3717 (aarch64_fma4_elt_to_64df): Likewise.
3718 (fnma<mode>4): Likewise.
3719 (aarch64_fnma4_elt<mode>): Likewise.
3720 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
3721 (aarch64_fnma4_elt_to_128v2df): Likewise.
3722 (aarch64_fnma4_elt_to_64df): Likewise.
3723 * config/aarch64/iterators.md (VDQSF): New.
3724 * config/aarch64/arm_neon.h
3725 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
3726 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
3728 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3730 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
3731 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3732 (aarch64_mul3_elt_to_128df): Likewise.
3733 (aarch64_mul3_elt_to_64v2df): Likewise.
3734 * config/aarch64/iterators.md (VEL): Also handle DFmode.
3736 (VMUL_CHANGE_NLANES) Likewise.
3739 * config/aarch64/arm_neon.h
3740 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
3742 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
3744 * config/aarch64/arm_neon.h
3745 (vcvtx_high_f32_f64): Fix parameters.
3747 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3748 Uros Bizjak <ubizjak@gmail.com>
3750 * config/alpha.c: Include tree-ssanames.h.
3752 2013-09-16 Richard Biener <rguenther@suse.de>
3754 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
3755 (dot_rdg_1): Handle control_dd.
3756 (create_edge_for_control_dependence): New function.
3757 (create_rdg_edges): Add control dependences if asked for.
3758 (build_rdg): Likewise.
3759 (generate_loops_for_partition): If there are not necessary
3760 control stmts remove all their dependencies.
3761 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
3762 (distribute_loop): Pass on control dependences.
3763 (tree_loop_distribution): Compute control dependences and remove
3764 restriction on number of loop nodes.
3766 2013-09-16 Jakub Jelinek <jakub@redhat.com>
3768 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
3771 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
3773 * cse.c (try_const_anchors): Punt on CC modes.
3775 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3777 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
3779 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
3782 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
3785 2013-09-14 Iain Sandoe <iain@codesourcery.com>
3788 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
3790 (darwin_objc1_section): Likewise.
3791 (darwin_file_end): Emit Image Info section when required.
3793 2013-09-14 Jan Hubicka <jh@suse.cz>
3795 * tree-into-ssa.c (gate_into_ssa): New.
3796 (pass_data_build_ssa): Use it.
3797 * cgraph.h (expand_thunk): Update prototype.
3798 * cgraphunit.c (analyze_function): Expand thunks early.
3799 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
3800 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
3801 set lowered flag; do not add new function.
3802 (assemble_thunks_and_aliases): Update.
3803 * tree-ssa.c (gate_init_datastructures): New gate.
3804 (pass_data_init_datastructures): Use it.
3806 2013-09-14 Iain Sandoe <iain@codesourcery.com>
3809 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
3810 xmm register set described in the psABI.
3812 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
3814 * dwarf2out.c (should_emit_struct_debug): Add check
3815 for type_decl variable is not NULL.
3817 2013-09-13 Jacek Caban <jacek@codeweavers.com>
3819 * config.gcc: Use new winnt-c.c target hooks
3820 * config/t-winnt: New file
3821 * config/winnt-c.c: New file
3822 * doc/tm.texi.in: Document new hook
3823 * doc/tm.texi: Regenerated
3825 2013-09-13 Jan Hubicka <jh@suse.cz>
3828 * ipa-inline.c (check_callers): New function.
3829 (check_caller_edge): Remove.
3830 (want_inline_function_to_all_callers_p): Also permit alises that are
3832 (inline_to_all_callers): Terminate the walk when devirtualization
3833 introduce new calls.
3835 2013-09-13 Jan Hubicka <jh@suse.cz>
3837 * ipa-inline-analysis.c (struct growth_data): Add node.
3838 (do_estimate_growth_1): Fix detection of recursion.
3840 2013-09-13 Jakub Jelinek <jakub@redhat.com>
3842 PR tree-optimization/58392
3843 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
3844 to avoid shadowing of outer loop variable. If
3845 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
3846 replace_by_duplicate_decl simduid of loops that have it set and
3847 set dest_cfun->has_simduid_loops and/or
3848 dest_cfun->has_force_vect_loops.
3849 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
3850 instead of maybe_lookup_decl.
3851 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
3852 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
3853 call. Copy over force_vect and copy and remap simduid. Set
3854 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
3855 (copy_cfg_body): Remove blocks_to_copy argument. Use
3856 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
3857 caller. Don't set cfun->has_simduid_loops and/or
3858 cfun->has_force_vect_loops here.
3859 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
3861 (expand_call_inline, tree_function_versioning): Adjust copy_body
3864 2013-09-13 Martin Jambor <mjambor@suse.cz>
3867 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
3868 the assert if the edge was a speculative one.
3870 2013-09-13 Richard Biener <rguenther@suse.de>
3872 * tree-data-ref.h (known_dependences_p): Move here ...
3873 * tree-loop-distribution.c (known_dependences_p): ... from here.
3874 (dump_rdg_component, debug_rdg_component): Remove.
3876 (generate_loops_for_partition): Use gimple_uid instead of
3877 relying on matching stmt visit order.
3878 (rdg_build_partitions): Take starting stmt vector.
3879 (ldist_gen): Merge into ...
3880 (distribute_loop): ... this function. Do not compute starting
3882 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
3884 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3886 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
3887 Set type attribute correctly. Set predicable_short_it attribute.
3888 (cmpsi_shiftsi): Remove %? from output template.
3890 2013-09-13 Richard Biener <rguenther@suse.de>
3892 * tree-loop-distribution.c (struct rdg_component,
3893 rdg_defs_used_in_other_loops_p, free_rdg_components,
3894 rdg_build_components): Remove.
3895 (stmts_from_loop): Do not record virtual PHIs.
3896 (generate_loops_for_partition): Skip virtual PHIs.
3897 (build_rdg_partition_for_component): Rename to ...
3898 (build_rdg_partition_for_vertex): ... this and adjust.
3899 (rdg_build_partitions): Take a vector of starting vertices
3900 instead of components. Remove unnecessary leftover handling.
3901 (ldist_gen): Do not build components or record other stores.
3902 (distribute_loop): Do not distribute loops containing stmts
3905 2013-09-13 Christian Bruel <christian.bruel@st.com>
3908 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
3910 2013-09-13 Kai Tietz <ktietz@redhat.com>
3912 * config.gcc: Separate cases for mingw and cygwin targets,
3913 and add 64-bit cygwin target case.
3915 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
3916 dll-export inline-functions.
3917 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
3920 2013-09-13 Jeff Law <law@redhat.com>
3924 2013-09-06 Jeff Law <law@redhat.com>
3926 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3927 edge implied equivalences into successor phis.
3929 2013-09-12 DJ Delorie <dj@redhat.com>
3931 * config/rl78/rl78-virt.md: Change from | to \; for asm line
3934 2013-09-12 Brooks Moses <bmoses@google.com>
3937 * Makefile.in: Do not install driver binaries in $(target)/bin.
3939 2013-09-12 DJ Delorie <dj@redhat.com>
3941 * config/rl78/rl78.opt (mrelax): New.
3942 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
3943 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
3945 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
3947 (rl78_expand_epilogue): Likewise.
3949 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
3952 * lra-eliminations.c (remove_reg_equal_offset_note): New.
3953 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
3954 pointer elimination with using remove_reg_equal_offset_note.
3956 2013-09-12 DJ Delorie <dj@redhat.com>
3958 * config/msp430/: New port.
3959 * config.gcc (msp430): Added.
3960 * doc/invoke.texi: Document MSP430 options.
3961 * doc/install.texi: Document msp430-elf
3962 * doc/md.texi: Document msp430-elf
3963 * doc/contrib.texi: Document msp430-elf
3965 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
3968 2013-09-12 Martin Jambor <mjambor@suse.cz>
3971 * ipa-prop.c (remove_described_reference): Give up if the edge in the
3972 reference descriptor is NULL.
3973 (ipa_edge_removal_hook): If owning a reference descriptor, set its
3976 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3978 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
3979 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
3980 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
3981 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
3982 make_temp_ssa_name): move to tree-ssanames.h
3983 * tree-ssa-alias.h: Move prototype.
3984 * tree-ssa.h: Include tree-ssanames.h.
3985 * tree-ssanames.c (FREE_SSANAMES): Move to here.
3986 * tree-ssanames.h: New. Move items from tree-flow*.h
3987 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
3989 2013-09-12 Richard Biener <rguenther@suse.de>
3991 PR tree-optimization/58404
3992 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
3993 propagate non-invariant addresses into dereferences wrapped
3994 in component references.
3996 2013-09-12 Richard Biener <rguenther@suse.de>
3998 PR tree-optimization/58402
3999 * passes.def: Move pass_late_warn_uninitialized later.
4001 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
4003 * tree-ssa.h: New. Move content from tree-flow.h and
4005 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
4006 Move prototypes belonging to tree-ssa.c.
4007 * tree-flow-inline.h (redirect_edge_var_map_def,
4008 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
4010 * gimple.h: Adjust prototypes.
4011 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
4013 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
4014 * tree.h: Move prototype to tree-ssa.h.
4015 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
4016 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
4017 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
4018 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
4019 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
4020 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
4021 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
4022 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
4023 graphite-blocking.c, graphite-clast-to-gimple.c,
4024 graphite-dependences.c, graphite-interchange.c,
4025 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
4026 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
4027 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
4028 ipa-reference.c, ipa-split.c, ipa-utils.c,
4029 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
4030 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
4031 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
4032 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4033 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
4034 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
4035 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
4036 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
4037 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
4038 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
4039 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
4040 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4041 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
4042 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
4043 value-prof.c, var-tracking.c,
4044 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
4046 2013-09-12 Richard Biener <rguenther@suse.de>
4048 PR tree-optimization/58396
4049 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
4050 (build_rdg): Take a loop-nest parameter, fix memleaks.
4051 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
4053 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4055 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
4058 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
4060 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
4061 The FMA instruction names should have a 'v' prefix.
4063 2013-09-12 Richard Biener <rguenther@suse.de>
4065 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
4066 (dot_rdg): Use popen instead of system in optional code.
4067 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
4068 (already_processed_vertex_p): Adjust.
4069 (has_anti_or_output_dependence, predecessor_has_mem_write,
4070 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
4071 rdg_flag_uses): Remove.
4072 (rdg_flag_vertex): Simplify.
4073 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
4075 (build_rdg_partition_for_component): Process the first vertex
4076 of a component only.
4077 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
4079 2013-09-12 Alan Modra <amodra@gmail.com>
4081 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
4083 2013-09-11 DJ Delorie <dj@redhat.com>
4084 Nick Clifton <nickc@redhat.com>
4086 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
4087 (rl78_stack_based_mem): New.
4088 * config/rl78/constraints.md (Iv08): New.
4097 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
4099 (movsi): Change from expand to insn-and-split.
4100 (ashrsi3): Clobber AX.
4104 * config/rl78/rl78.md (CC_REG): Fix.
4105 (addsi3): Allow memory and immediate operands.
4106 (addsi3_internal): Split into...
4107 (addsi3_internal_virt): ...new, and ...
4108 (addsi3_internal_real): ...new.
4110 (subsi3_internal_virt): New.
4111 (subsi3_internal_real): New.
4112 (mulsi3): Add memory operand.
4113 (mulsi3_rl78): Likewise.
4114 (mulsi3_g13): Likewise.
4115 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
4116 (cbranchqi4_real): Add more constraint options.
4117 (cbranchhi4_real): Expand pattern.
4118 (cbranchhi4_real_signed): New.
4119 (cbranchhi4_real_inverted): New.
4120 (cbranchsi4_real_lt): New.
4121 (cbranchsi4_real_ge): New.
4122 (cbranchsi4_real_signed): New.
4123 (cbranchsi4_real): New.
4125 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
4127 (lshrsi3_virt): Likewise.
4128 (ashlsi3_virt): Likewise.
4129 (cbranchqi4_virt_signed): New.
4130 (cbranchhi4_virt_signed): New.
4131 (cbranchsi4_virt): New.
4132 * config/rl78/rl78.c: Whitespace fixes throughout.
4133 (move_elim_pass): New.
4134 (pass_data_rl78_move_elim): New.
4135 (pass_rl78_move_elim): New.
4136 (make_pass_rl78_move_elim): New.
4137 (rl78_devirt_info): Run devirt earlier.
4138 (rl78_move_elim_info): New.
4139 (rl78_asm_file_start): Register it.
4140 (rl78_split_movsi): New.
4141 (rl78_as_legitimate_address): Allow virtual base registers when
4143 (rl78_addr_space_convert): Remove spurious debug stuff.
4144 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
4145 (rl78_print_operand): More cases for not printing '#'.
4146 (rl78_expand_compare): Remove most of the logic.
4147 (content_memory): New.
4148 (clear_content_memory): New.
4149 (get_content_index): New.
4150 (get_content_name): New.
4151 (display_content_memory): New.
4152 (update_content): New.
4153 (record_content): New.
4154 (already_contains): New.
4155 (insn_ok_now): Re-recog insns with virtual registers.
4156 (add_postponed_content_update): New.
4157 (process_postponed_content_update): New.
4158 (gen_and_emit_move): New.
4159 (transcode_memory_rtx): Record new location content.
4160 Use gen_and_emit_move.
4161 (force_into_acc): New.
4162 (move_to_acc): Use gen_and_emit_move.
4163 (move_from_acc): Likewise.
4164 (move_acc_to_reg): Likewise.
4165 (move_to_x): Likewise.
4166 (move_to_hl): Likewise.
4167 (move_to_de): Likewise.
4168 (rl78_alloc_physical_registers_op1): Record location content.
4169 (has_constraint): New.
4170 (rl78_alloc_physical_registers_op2): Record location content.
4172 (rl78_alloc_physical_registers_ro1): Likewise.
4173 (rl78_alloc_physical_registers_cmp): Likewise.
4174 (rl78_alloc_physical_registers_umul): Likewise.
4175 (rl78_alloc_address_registers_macax): New.
4176 (rl78_alloc_physical_registers): Initialize and set location
4177 content memory as needed.
4178 (rl78_reorg): Make sure split2 is called.
4179 (rl78_rtx_costs): New.
4181 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
4183 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
4184 for (not (neg ...)) and (neg (not ...)) cases.
4186 2013-09-11 Richard Biener <rguenther@suse.de>
4189 * passes.def: Split critical edges before late uninit warning passes.
4190 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
4192 2013-09-11 Jakub Jelinek <jakub@redhat.com>
4194 PR tree-optimization/58385
4195 * fold-const.c (build_range_check): If both low and high are NULL,
4196 use omit_one_operand_loc to preserve exp side-effects.
4198 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4200 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
4202 2013-09-11 Richard Biener <rguenther@suse.de>
4204 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
4205 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
4206 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
4207 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
4208 stmts_from_loop, known_dependences_p, build_empty_rdg,
4209 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
4210 * tree-loop-distribution.c: ... here.
4211 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
4212 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
4213 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
4214 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
4215 * tree-loop-distribution.c: ... here.
4216 * tree-loop-distribution.c: Include gimple-pretty-print.h.
4217 (struct partition_s): Add loops member.
4218 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
4219 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
4220 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
4222 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4223 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4224 Sergey Lega <sergey.s.lega@intel.com>
4225 Anna Tikhonova <anna.tikhonova@intel.com>
4226 Ilya Tocar <ilya.tocar@intel.com>
4227 Andrey Turetskiy <andrey.turetskiy@intel.com>
4228 Ilya Verbin <ilya.verbin@intel.com>
4229 Kirill Yukhin <kirill.yukhin@intel.com>
4230 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4232 * config/i386/constraints.md (k): New.
4234 * config/i386/i386.c (const regclass_map): Add new mask registers.
4235 (dbx_register_map): Ditto.
4236 (dbx64_register_map): Ditto.
4237 (svr4_dbx_register_map): Ditto.
4238 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
4240 (ix86_preferred_reload_class): Disable constants for mask registers.
4241 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
4242 (ix86_hard_regno_mode_ok): Support new mask registers.
4243 (x86_order_regs_for_local_alloc): Ditto.
4244 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
4245 (FIXED_REGISTERS): Add new mask registers.
4246 (CALL_USED_REGISTERS): Ditto.
4247 (REG_ALLOC_ORDER): Ditto.
4248 (VALID_MASK_REG_MODE): New.
4249 (FIRST_MASK_REG): Ditto.
4250 (LAST_MASK_REG): Ditto.
4251 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
4252 (MAYBE_MASK_CLASS_P): New.
4253 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
4254 (REG_CLASS_CONTENTS): Ditto.
4255 (MASK_REGNO_P): New.
4256 (ANY_MASK_REG_P): Ditto.
4257 (HI_REGISTER_NAMES): Add new mask registers.
4258 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
4259 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
4261 (attribute "type"): Add mskmov, msklog.
4262 (attribute "length_immediate"): Support them.
4263 (attribute "memory"): Ditto.
4264 (attribute "prefix_0f"): Ditto.
4265 (*movhi_internal): Support new mask registers.
4266 (*movqi_internal): Ditto.
4267 (define_split): Split out clobber pattern is a logic
4268 insn on mask registers.
4269 (*k<logic><mode>): New.
4270 (*andhi_1): Extend to support mask regs.
4271 (*andqi_1): Extend to support mask regs.
4273 (define_split): Split and-not to and and not if operands
4275 (*<code><mode>_1): Separate HI mode to new pattern...
4276 (*<code>hi_1): This.
4277 (*<code>qi_1): Extend to support mask regs.
4279 (kortestzhi): Ditto.
4280 (kortestchi): Ditto.
4282 (*one_cmpl<mode>2_1): Remove HImode and handle it...
4283 (*one_cmplhi2_1): ...Here, now with mask registers support.
4284 (*one_cmplqi2_1): Support new mask registers.
4285 (HI/QImode arithmetics splitter): Don't split if mask registers
4287 (HI/QImode not splitter): Ditto.
4288 * config/i386/predicated.md (mask_reg_operand): New.
4289 (general_reg_operand): Ditto.
4291 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4293 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
4294 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
4296 2013-09-10 Jeff Law <law@redhat.com>
4298 PR tree-optimization/58380
4299 * tree-ssa-threadupdate.c (thread_block): Recognize another case
4300 of threading through a buried loop header.
4302 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
4303 return value for single successor case.
4305 2013-09-10 Jan Hubicka <jh@suse.cz>
4307 * ipa-devirt.c (ipa_devirt): Enable with LTO.
4309 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
4312 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
4313 support conditional execution.
4314 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
4316 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
4318 * lra.c (lra): Clear lra_optional_reload_pseudos before every
4320 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
4321 Check destination too to check move insn.
4322 (undo_optional_reloads): Add check that the original peudo did not
4323 changed its allocation and the optional reload was inherited on last
4324 inheritance pass. Break loop after deciding to keep optional reload.
4325 (lra_undo_inheritance): Add check that inherited pseudo still in
4328 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
4330 * config/aarch64/aarch64.md (generic_sched): New.
4331 * config/aarch64/aarch64-generic.md (load): Make conditional
4332 on generic_sched attribute.
4333 (nonload): Likewise.
4335 2013-09-10 Jan Hubicka <jh@suse.cz>
4337 * lto-cgraph.c: Include ipa-utils.h.
4338 (compute_ltrans_boundary): Also add possible targets into the boundary.
4340 2013-09-10 Jan Hubicka <jh@suse.cz>
4342 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
4343 VAR_DECL of vtable rather than full expression.
4345 2013-09-10 Jan Hubicka <jh@suse.cz>
4346 Paolo Carlini <paolo.carlini@oracle.com>
4348 * cgraphunit.c (analyze_functions): Save input_location, set it
4349 to UNKNOWN_LOCATION and restore it at the end.
4351 2013-09-10 Martin Jambor <mjambor@suse.cz>
4353 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
4354 represented by a thunk.
4356 2013-09-10 Jeff Law <law@redhat.com>
4358 PR tree-optimization/58343
4359 * tree-ssa-threadupdate.c (thread_block): Identify and disable
4360 jump threading requests through loop headers buried in the middle
4361 of a jump threading path.
4363 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
4364 in return value/type.
4366 2013-09-10 Jakub Jelinek <jakub@redhat.com>
4368 PR rtl-optimization/58365
4369 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
4370 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
4373 2013-09-10 Richard Biener <rguenther@suse.de>
4375 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
4376 * tree-data-ref.c (create_rdg_vertices): Collect all data
4377 references, signal failure to the caller, use data-ref API.
4378 (build_rdg): Compute data references only once. Maintain lifetime
4379 of data references and data dependences from within RDG.
4380 (free_rdg): Free dependence relations.
4381 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
4382 inventing extra dependences.
4383 (distribute_loop): Update for RDG API changes.
4385 2013-09-10 Kai Tietz <ktietz@redhat.com>
4387 * doc/invoke.texi (fms-extensions): Document changed
4388 behavior for ms-abi targets.
4389 * config/i386/i386.c (ix86_option_override_internal):
4390 Set default value of option -fms-extension for ms-abi targets.
4392 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4394 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
4396 2013-09-10 Alan Modra <amodra@gmail.com>
4399 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
4401 2013-09-10 Alan Modra <amodra@gmail.com>
4403 * config/rs6000/predicates.md (add_cint_operand): New.
4404 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
4405 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
4406 using add_cint_operand.
4407 (largetoc_high_plus_aix): Likewise.
4409 2013-09-09 Jakub Jelinek <jakub@redhat.com>
4411 PR tree-optimization/58364
4412 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
4413 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
4414 the current range can't be an unconditional true or false.
4416 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
4418 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
4420 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
4422 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
4424 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
4427 * doc/invoke.texi (-Wdelete-incomplete): Document it.
4429 2013-09-09 Ian Bolton <ian.bolton@arm.com>
4431 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
4432 NO_REGS for immediate that can't be moved directly into FP_REGS.
4434 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4436 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
4437 comparison with negated operand.
4438 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
4441 2013-09-09 Richard Biener <rguenther@suse.de>
4444 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
4445 of a subloop record all its block as affecting loop-closed SSA form.
4447 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4449 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
4450 of an rtx/bitpos pair.
4451 (store_fixed_bit_field): Update accordingly.
4453 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4455 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
4457 * builtins.c (expand_errno_check): Likewise.
4458 * dwarf2cfi.c (init_return_column_size): Likewise.
4459 * except.c (sjlj_mark_call_sites): Likewise.
4460 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
4461 * lra-constraints.c (emit_inc): Likewise.
4462 * ree.c (combine_set_extension): Likewise.
4463 * regmove.c (fixup_match_2): Likewise.
4464 * reload1.c (inc_for_reload): Likewise.
4466 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4468 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
4469 (force_to_mode, simplify_shift_const_1, simplify_comparison):
4470 Use gen_int_mode with the mode of the associated simplify_* call.
4471 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
4472 * expmed.c (expand_shift_1): Likewise.
4473 * function.c (instantiate_virtual_regs_in_insn): Likewise.
4474 * loop-iv.c (iv_number_of_iterations): Likewise.
4475 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
4476 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
4478 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4480 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
4481 of the associated expand_* call.
4482 (asan_emit_stack_protection): Likewise.
4483 * builtins.c (round_trampoline_addr): Likewise.
4484 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
4485 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
4486 (emit_store_flag): Likewise.
4487 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
4488 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
4490 * function.c (instantiate_virtual_regs_in_insn): Likewise.
4491 * ifcvt.c (noce_try_store_flag_constants): Likewise.
4492 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
4493 * modulo-sched.c (generate_prolog_epilog): Likewise.
4494 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
4495 (expand_ctz, expand_ffs, expand_unop): Likewise.
4497 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4499 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
4500 of the associated gen_rtx_* call.
4501 * caller-save.c (init_caller_save): Likewise.
4502 * combine.c (find_split_point, make_extraction): Likewise.
4503 (make_compound_operation): Likewise.
4504 * dwarf2out.c (mem_loc_descriptor): Likewise.
4505 * explow.c (plus_constant, probe_stack_range): Likewise.
4506 * expmed.c (expand_mult_const): Likewise.
4507 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
4508 * reload1.c (init_reload): Likewise.
4509 * valtrack.c (cleanup_auto_inc_dec): Likewise.
4510 * var-tracking.c (adjust_mems): Likewise.
4511 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
4512 rather than gen_rtx_fmt_ee.
4514 2013-09-09 Jan Hubicka <jh@suse.cz>
4517 * value-prof.c (gimple_ic): Copy also abnormal edges.
4519 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
4521 * asan.c (asan_shadow_cst): Use gen_int_mode.
4523 2013-09-08 Jan Hubicka <jh@suse.cz>
4525 * ipa-profile.c: Add toplevel comment.
4526 (ipa_propagate_frequency_1): Be more conservative when profile is read.
4527 (contains_hot_call_p): New function.
4528 (ipa_propagate_frequency): Set frequencies based on counts when
4530 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
4531 profile; do not tamper with profile after inlining if it is read.
4533 2013-09-08 Jan Hubicka <jh@suse.cz>
4535 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
4538 2013-09-08 Jan Hubicka <jh@suse.cz>
4540 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
4543 2013-09-08 Jeff Law <law@redhat.com>
4546 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
4549 2013-09-08 Andi Kleen <ak@linux.intel.com>
4551 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
4553 2013-09-08 Jan Hubicka <jh@suse.cz>
4555 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
4557 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
4559 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
4560 for non-debug insns.
4561 * lra.c (new_insn_reg): Take the containing insn as a parameter.
4562 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
4563 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
4566 2013-09-08 Jan Hubicka <jh@suse.cz>
4568 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
4569 targets and devirtualize to BUILT_IN_UNREACHABLE.
4570 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
4571 * ipa.c (walk_polymorphic_call_targets): New function.
4572 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
4573 functions; use the new timevar.
4574 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
4575 was removed from the program.
4576 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
4577 not consider it in the walk when its vtable is dead.
4578 (possible_polymorphic_call_targets_1): Pass anonymous flag to
4580 (devirt_variable_node_removal_hook): New function.
4581 (possible_polymorphic_call_targets): Also register
4582 devirt_variable_node_removal_hook.
4583 (ipa_devirt): Do not do non-speculative devirtualization.
4584 (gate_ipa_devirt): One execute if devirtualizing speculatively.
4586 2013-09-08 Jan Hubicka <jh@suse.cz>
4588 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
4589 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
4590 varpool_remove_variable_insertion_hook): Declare.
4591 * varpool.c (varpool_node_hook_list): New structure.
4592 (first_varpool_node_removal_hook,
4593 first_varpool_variable_insertion_hook): New variables.
4594 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
4595 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
4596 varpool_remove_variable_insertion_hook,
4597 varpool_call_variable_insertion_hooks): New functions.
4598 (varpool_remove_node): Use it.
4600 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
4603 * diagnostic.c (diagnostic_build_prefix): When s.file is
4604 "<built-in>" don't output line and column numbers.
4606 2013-09-06 Jan Hubicka <jh@suse.cz>
4608 * cgraphunit.c (expand_thunk): Get body before touching arguments.
4609 * lto-streamer-out.c: Stream thunks, too.
4610 * lto-streamer-in.c (input_function): Pop cfun here
4611 (lto_read_body): Instead of here.
4613 2013-09-06 Caroline Tice <cmtice@google.com>
4615 * doc/install.texi: Add documentation for the --enable-vtable-verify
4616 and the --disable-libvtv configure options.
4618 2013-09-06 Jeff Law <law@redhat.com>
4620 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4621 edge implied equivalences into successor phis.
4623 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
4625 * resource.c (mark_referenced_resources): Handle COND_EXEC.
4627 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
4629 * resource.c (mark_target_live_regs): Compute resources taking
4630 into account if a call is predicated or not.
4632 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
4634 * toplev.c (output_stack_usage): Be prepared for suffixes created by
4635 the compiler in the function names.
4637 2013-09-06 Jan Hubicka <jh@suse.cz>
4640 * ipa-inline.c (has_caller_p): New function.
4641 (want_inline_function_to_all_callers_p): Use it.
4642 (sum_callers, inline_to_all_callers): Break out from ...
4643 (ipa_inline): ... here.
4645 2013-09-06 Jan Hubicka <jh@suse.cz>
4647 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
4648 only when AVX is enabled.
4650 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4652 * config/aarch64/aarch64.md
4653 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
4654 is fpsimd_<load/store>2.
4655 (load_pair<mode>): Likewise.
4656 (store_pair<mode>): Likewise.
4658 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4660 * config/arm/types.md (type): Add "mrs" type.
4661 * config/aarch64/aarch64.md
4662 (aarch64_load_tp_hard): Make type "mrs".
4664 (load_tp_hard): Make type "mrs".
4665 * config/arm/cortex-a15.md: Update with new attributes.
4666 * config/arm/cortex-a5.md: Update with new attributes.
4667 * config/arm/cortex-a53.md: Update with new attributes.
4668 * config/arm/cortex-a7.md: Update with new attributes.
4669 * config/arm/cortex-a8.md: Update with new attributes.
4670 * config/arm/cortex-a9.md: Update with new attributes.
4671 * config/arm/cortex-m4.md: Update with new attributes.
4672 * config/arm/cortex-r4.md: Update with new attributes.
4673 * config/arm/fa526.md: Update with new attributes.
4674 * config/arm/fa606te.md: Update with new attributes.
4675 * config/arm/fa626te.md: Update with new attributes.
4676 * config/arm/fa726te.md: Update with new attributes.
4678 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4680 * config/aarch64/aarch64.md
4681 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
4682 (*movtf_aarch64): Likewise.
4684 (thumb1_movdi_insn): Use "multiple" for type where more than one
4685 instruction is used for a move.
4686 (*arm32_movhf): Likewise.
4687 (*thumb_movdf_insn): Likewise.
4689 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4691 * config/arm/types.md (type): Rename fcpys to fmov.
4693 (*arm_movsi_vfp): Rename type fcpys as fmov.
4694 (*thumb2_movsi_vfp): Likewise
4695 (*movhf_vfp_neon): Likewise
4696 (*movhf_vfp): Likewise
4697 (*movsf_vfp): Likewise
4698 (*thumb2_movsf_vfp): Likewise
4699 (*movsfcc_vfp): Likewise
4700 (*thumb2_movsfcc_vfp): Likewise
4701 * config/aarch64/aarch64-simd.md
4702 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
4703 * config/aarch64/aarch64.md
4704 (*movsi_aarch64): Replace type mov_reg with fmovs.
4705 (*movdi_aarch64): Likewise
4706 (*movsf_aarch64): Likewise
4707 (*movdf_aarch64): Likewise
4709 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
4710 * config/arm/iwmmxt.md
4711 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
4712 * config/arm/arm1020e.md: Update with new attributes.
4713 * config/arm/cortex-a15-neon.md: Update with new attributes.
4714 * config/arm/cortex-a5.md: Update with new attributes.
4715 * config/arm/cortex-a53.md: Update with new attributes.
4716 * config/arm/cortex-a7.md: Update with new attributes.
4717 * config/arm/cortex-a8-neon.md: Update with new attributes.
4718 * config/arm/cortex-a9.md: Update with new attributes.
4719 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4720 * config/arm/cortex-r4f.md: Update with new attributes.
4721 * config/arm/marvell-pj4.md: Update with new attributes.
4722 * config/arm/vfp11.md: Update with new attributes.
4724 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4726 * config/aarch64/aarch64.md
4727 (*madd<mode>): Fix type attribute.
4728 (*maddsi_uxtw): Likewise.
4729 (*msub<mode>): Likewise.
4730 (*msubsi_uxtw): Likewise.
4731 (<su_optab>maddsidi4): Likewise.
4732 (<su_optab>msubsidi4): Likewise.
4734 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4736 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
4737 * config/arm/arm.md (core_cycles): Remove fdiv.
4738 * config/arm/vfp.md:
4739 (*sqrtsf2_vfp): Update for attribute changes.
4740 (*sqrtdf2_vfp): Likewise.
4741 * config/aarch64/aarch64.md:
4742 (sqrt<mode>2): Update for attribute changes.
4743 * config/arm/arm1020e.md: Update with new attributes.
4744 * config/arm/cortex-a15-neon.md: Update with new attributes.
4745 * config/arm/cortex-a5.md: Update with new attributes.
4746 * config/arm/cortex-a53.md: Update with new attributes.
4747 * config/arm/cortex-a7.md: Update with new attributes.
4748 * config/arm/cortex-a8-neon.md: Update with new attributes.
4749 * config/arm/cortex-a9.md: Update with new attributes.
4750 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4751 * config/arm/cortex-r4f.md: Update with new attributes.
4752 * config/arm/marvell-pj4.md: Update with new attributes.
4753 * config/arm/vfp11.md: Update with new attributes.
4755 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4757 * config/arm/types.md
4758 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
4759 * config/aarch64/aarch64.md
4760 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
4762 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
4763 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
4764 (float<GPI:mode><GPF:mode>2): Likewise.
4766 (*truncsisf2_vfp): Update with new attributes.
4767 (*truncsidf2_vfp): Likewise.
4768 (fixuns_truncsfsi2): Likewise.
4769 (fixuns_truncdfsi2): Likewise.
4770 (*floatsisf2_vfp): Likewise.
4771 (*floatsidf2_vfp): Likewise.
4772 (floatunssisf2): Likewise.
4773 (floatunssidf2): Likewise.
4774 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
4775 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
4776 * config/arm/arm1020e.md: Update with new attributes.
4777 * config/arm/cortex-a15-neon.md: Update with new attributes.
4778 * config/arm/cortex-a5.md: Update with new attributes.
4779 * config/arm/cortex-a53.md: Update with new attributes.
4780 * config/arm/cortex-a7.md: Update with new attributes.
4781 * config/arm/cortex-a8-neon.md: Update with new attributes.
4782 * config/arm/cortex-a9.md: Update with new attributes.
4783 * config/arm/cortex-m4-fpu.md: Update with new attributes.
4784 * config/arm/cortex-r4f.md: Update with new attributes.
4785 * config/arm/marvell-pj4.md: Update with new attributes.
4786 * config/arm/vfp11.md: Update with new attributes.
4788 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4790 * config/aarch64/arm_neon.h
4791 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
4792 (vqtbx<1,2,3,4><q>_s8): Likewise.
4794 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4796 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
4797 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
4798 (add<mode>3): Add type attribute.
4799 (add<mode>3): Likewise.
4800 (usadd<mode>3): Likewise.
4801 (ssadd<mode>3): Likewise.
4802 (sub<mode>3): Likewise.
4803 (sub<mode>3): Likewise.
4804 (ussub<mode>3): Likewise.
4805 (sssub<mode>3): Likewise.
4806 (ssmulsa3): Likewise.
4807 (usmulusa3): Likewise.
4808 (arm_usatsihi): Likewise.
4810 (*movdi_vfp): Add types for all instructions.
4811 (*movdi_vfp_cortexa8): Likewise.
4812 (*movhf_vfp_neon): Likewise.
4813 (*movhf_vfp): Likewise.
4814 (*movdf_vfp): Likewise.
4815 (*thumb2_movdf_vfp): Likewise.
4816 (*thumb2_movdfcc_vfp): Likewise.
4817 * config/arm/arm.md: Add type attribute to all insn patterns.
4818 (*thumb1_adddi3): Add type attribute.
4819 (*arm_adddi3): Likewise.
4820 (*adddi_sesidi_di): Likewise.
4821 (*adddi_zesidi_di): Likewise.
4822 (*thumb1_addsi3): Likewise.
4823 (addsi3_compare0): Likewise.
4824 (*addsi3_compare0_scratch): Likewise.
4825 (*compare_negsi_si): Likewise.
4826 (cmpsi2_addneg): Likewise.
4827 (*addsi3_carryin_<optab>): Likewise.
4828 (*addsi3_carryin_alt2_<optab>): Likewise.
4829 (*addsi3_carryin_clobercc_<optab>): Likewise.
4830 (*subsi3_carryin): Likewise.
4831 (*subsi3_carryin_const): Likewise.
4832 (*subsi3_carryin_compare): Likewise.
4833 (*subsi3_carryin_compare_const): Likewise.
4834 (*arm_subdi3): Likewise.
4835 (*thumb_subdi3): Likewise.
4836 (*subdi_di_zesidi): Likewise.
4837 (*subdi_di_sesidi): Likewise.
4838 (*subdi_zesidi_di): Likewise.
4839 (*subdi_sesidi_di): Likewise.
4840 (*subdi_zesidi_ze): Likewise.
4841 (thumb1_subsi3_insn): Likewise.
4842 (*arm_subsi3_insn): Likewise.
4843 (*anddi3_insn): Likewise.
4844 (*anddi_zesidi_di): Likewise.
4845 (*anddi_sesdi_di): Likewise.
4846 (*ne_zeroextracts): Likewise.
4847 (*ne_zeroextracts): Likewise.
4848 (*ite_ne_zeroextr): Likewise.
4849 (*ite_ne_zeroextr): Likewise.
4850 (*anddi_notdi_di): Likewise.
4851 (*anddi_notzesidi): Likewise.
4852 (*anddi_notsesidi): Likewise.
4853 (andsi_notsi_si): Likewise.
4854 (thumb1_bicsi3): Likewise.
4855 (*iordi3_insn): Likewise.
4856 (*iordi_zesidi_di): Likewise.
4857 (*iordi_sesidi_di): Likewise.
4858 (*thumb1_iorsi3_insn): Likewise.
4859 (*xordi3_insn): Likewise.
4860 (*xordi_zesidi_di): Likewise.
4861 (*xordi_sesidi_di): Likewise.
4862 (*arm_xorsi3): Likewise.
4863 (*andsi_iorsi3_no): Likewise.
4864 (*smax_0): Likewise.
4865 (*smax_m1): Likewise.
4866 (*arm_smax_insn): Likewise.
4867 (*smin_0): Likewise.
4868 (*arm_smin_insn): Likewise.
4869 (*arm_umaxsi3): Likewise.
4870 (*arm_uminsi3): Likewise.
4871 (*minmax_arithsi): Likewise.
4872 (*minmax_arithsi_): Likewise.
4873 (*satsi_<SAT:code>): Likewise.
4874 (arm_ashldi3_1bit): Likewise.
4875 (arm_ashrdi3_1bit): Likewise.
4876 (arm_lshrdi3_1bit): Likewise.
4877 (*arm_negdi2): Likewise.
4878 (*thumb1_negdi2): Likewise.
4879 (*arm_negsi2): Likewise.
4880 (*thumb1_negsi2): Likewise.
4881 (*negdi_extendsid): Likewise.
4882 (*negdi_zero_extend): Likewise.
4883 (*arm_abssi2): Likewise.
4884 (*thumb1_abssi2): Likewise.
4885 (*arm_neg_abssi2): Likewise.
4886 (*thumb1_neg_abss): Likewise.
4887 (one_cmpldi2): Likewise.
4888 (extend<mode>di2): Likewise.
4889 (*compareqi_eq0): Likewise.
4890 (*arm_extendhisi2addsi): Likewise.
4891 (*arm_movdi): Likewise.
4892 (*thumb1_movdi_insn): Likewise.
4893 (*arm_movt): Likewise.
4894 (*thumb1_movsi_insn): Likewise.
4895 (pic_add_dot_plus_four): Likewise.
4896 (pic_add_dot_plus_eight): Likewise.
4897 (tls_load_dot_plus_eight): Likewise.
4898 (*thumb1_movhi_insn): Likewise.
4899 (*thumb1_movsf_insn): Likewise.
4900 (*movdf_soft_insn): Likewise.
4901 (*thumb_movdf_insn): Likewise.
4902 (cbranchsi4_insn): Likewise.
4903 (cbranchsi4_scratch): Likewise.
4904 (*negated_cbranchsi4): Likewise.
4905 (*tbit_cbranch): Likewise.
4906 (*tlobits_cbranch): Likewise.
4907 (*tstsi3_cbranch): Likewise.
4908 (*cbranchne_decr1): Likewise.
4909 (*addsi3_cbranch): Likewise.
4910 (*addsi3_cbranch_scratch): Likewise.
4911 (*arm_cmpdi_insn): Likewise.
4912 (*arm_cmpdi_unsig): Likewise.
4913 (*arm_cmpdi_zero): Likewise.
4914 (*thumb_cmpdi_zero): Likewise.
4915 (*deleted_compare): Likewise.
4916 (*mov_scc): Likewise.
4917 (*mov_negscc): Likewise.
4918 (*mov_notscc): Likewise.
4919 (*cstoresi_eq0_thumb1_insn): Likewise.
4920 (cstoresi_nltu_thumb1): Likewise.
4921 (cstoresi_ltu_thu): Likewise.
4922 (thumb1_addsi3_addgeu): Likewise.
4923 (*arm_jump): Likewise.
4924 (*thumb_jump): Likewise.
4925 (*check_arch2): Likewise.
4926 (arm_casesi_internal): Likewise.
4927 (thumb1_casesi_dispatch): Likewise.
4928 (*arm_indirect_jump): Likewise.
4929 (*thumb1_indirect_jump): Likewise.
4931 (*and_scc): Likewise.
4932 (*ior_scc): Likewise.
4933 (*compare_scc): Likewise.
4934 (*cond_move): Likewise.
4935 (*cond_arith): Likewise.
4936 (*cond_sub): Likewise.
4937 (*cmp_ite0): Likewise.
4938 (*cmp_ite1): Likewise.
4939 (*cmp_and): Likewise.
4940 (*cmp_ior): Likewise.
4941 (*ior_scc_scc): Likewise.
4942 (*ior_scc_scc_cmp): Likewise.
4943 (*and_scc_scc): Likewise.
4944 (*and_scc_scc_cmp): Likewise.
4945 (*and_scc_scc_nod): Likewise.
4946 (*negscc): Likewise.
4947 (movcond_addsi): Likewise.
4948 (movcond): Likewise.
4949 (*ifcompare_plus_move): Likewise.
4950 (*if_plus_move): Likewise.
4951 (*ifcompare_move_plus): Likewise.
4952 (*if_move_plus): Likewise.
4953 (*ifcompare_arith_arith): Likewise.
4954 (*if_arith_arith): Likewise.
4955 (*ifcompare_arith_move): Likewise.
4956 (*if_arith_move): Likewise.
4957 (*ifcompare_move_arith): Likewise.
4958 (*if_move_arith): Likewise.
4959 (*ifcompare_move_not): Likewise.
4960 (*if_move_not): Likewise.
4961 (*ifcompare_not_move): Likewise.
4962 (*if_not_move): Likewise.
4963 (*ifcompare_shift_move): Likewise.
4964 (*if_shift_move): Likewise.
4965 (*ifcompare_move_shift): Likewise.
4966 (*if_move_shift): Likewise.
4967 (*ifcompare_shift_shift): Likewise.
4968 (*ifcompare_not_arith): Likewise.
4969 (*ifcompare_arith_not): Likewise.
4970 (*if_arith_not): Likewise.
4971 (*ifcompare_neg_move): Likewise.
4972 (*if_neg_move): Likewise.
4973 (*ifcompare_move_neg): Likewise.
4974 (*if_move_neg): Likewise.
4975 (prologue_thumb1_interwork): Likewise.
4976 (*cond_move_not): Likewise.
4977 (*sign_extract_onebit): Likewise.
4978 (*not_signextract_onebit): Likewise.
4979 (stack_tie): Likewise.
4980 (align_4): Likewise.
4981 (align_8): Likewise.
4982 (consttable_end): Likewise.
4983 (consttable_1): Likewise.
4984 (consttable_2): Likewise.
4985 (consttable_4): Likewise.
4986 (consttable_8): Likewise.
4987 (consttable_16): Likewise.
4988 (*thumb1_tablejump): Likewise.
4989 (prefetch): Likewise.
4990 (force_register_use): Likewise.
4991 (thumb_eh_return): Likewise.
4992 (load_tp_hard): Likewise.
4993 (load_tp_soft): Likewise.
4994 (tlscall): Likewise.
4995 (*arm_movtas_ze): Likewise.
4996 (*arm_rev): Likewise.
4997 (*arm_revsh): Likewise.
4998 (*arm_rev16): Likewise.
4999 * config/arm/thumb2.md
5000 (*thumb2_smaxsi3): Likewise.
5001 (*thumb2_sminsi3): Likewise.
5002 (*thumb32_umaxsi3): Likewise.
5003 (*thumb2_uminsi3): Likewise.
5004 (*thumb2_negdi2): Likewise.
5005 (*thumb2_abssi2): Likewise.
5006 (*thumb2_neg_abss): Likewise.
5007 (*thumb2_movsi_insn): Likewise.
5008 (tls_load_dot_plus_four): Likewise.
5009 (*thumb2_movhi_insn): Likewise.
5010 (*thumb2_mov_scc): Likewise.
5011 (*thumb2_mov_negs): Likewise.
5012 (*thumb2_mov_negs): Likewise.
5013 (*thumb2_mov_nots): Likewise.
5014 (*thumb2_mov_nots): Likewise.
5015 (*thumb2_movsicc_): Likewise.
5016 (*thumb2_movsfcc_soft_insn): Likewise.
5017 (*thumb2_indirect_jump): Likewise.
5018 (*thumb2_and_scc): Likewise.
5019 (*thumb2_ior_scc): Likewise.
5020 (*thumb2_ior_scc_strict_it): Likewise.
5021 (*thumb2_cond_move): Likewise.
5022 (*thumb2_cond_arith): Likewise.
5023 (*thumb2_cond_ari): Likewise.
5024 (*thumb2_cond_sub): Likewise.
5025 (*thumb2_negscc): Likewise.
5026 (*thumb2_movcond): Likewise.
5027 (thumb2_casesi_internal): Likewise.
5028 (thumb2_casesi_internal_pic): Likewise.
5029 (*thumb2_alusi3_short): Likewise.
5030 (*thumb2_mov<mode>_shortim): Likewise.
5031 (*thumb2_addsi_short): Likewise.
5032 (*thumb2_subsi_short): Likewise.
5033 (thumb2_addsi3_compare0): Likewise.
5034 (*thumb2_cbz): Likewise.
5035 (*thumb2_cbnz): Likewise.
5036 (*thumb2_one_cmplsi2_short): Likewise.
5037 (*thumb2_negsi2_short): Likewise.
5038 (*orsi_notsi_si): Likewise.
5039 * config/arm/arm1020e.md: Update with new attributes.
5040 * config/arm/arm1026ejs.md: Update with new attributes.
5041 * config/arm/arm1136jfs.md: Update with new attributes.
5042 * config/arm/arm926ejs.md: Update with new attributes.
5043 * config/arm/cortex-a15.md: Update with new attributes.
5044 * config/arm/cortex-a5.md: Update with new attributes.
5045 * config/arm/cortex-a53.md: Update with new attributes.
5046 * config/arm/cortex-a7.md: Update with new attributes.
5047 * config/arm/cortex-a8.md: Update with new attributes.
5048 * config/arm/cortex-a9.md: Update with new attributes.
5049 * config/arm/cortex-m4.md: Update with new attributes.
5050 * config/arm/cortex-r4.md: Update with new attributes.
5051 * config/arm/fa526.md: Update with new attributes.
5052 * config/arm/fa606te.md: Update with new attributes.
5053 * config/arm/fa626te.md: Update with new attributes.
5054 * config/arm/fa726te.md: Update with new attributes.
5056 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5058 * config/aarch64/aarch64-simd.md
5059 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
5060 <vwx> iterator to ensure correct register choice.
5061 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
5062 (aarch64_sqdmull_n<mode>): Likewise.
5063 (aarch64_sqdmull2_n<mode>_internal): Likewise.
5064 * config/aarch64/arm_neon.h
5065 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
5066 (vml<as><q>_n_<su>16): Likewise.
5067 (vml<as>l_high_lane<q>_<su>16): Likewise.
5068 (vml<as>l_high_n_<su>16): Likewise.
5069 (vml<as>l_lane<q>_<su>16): Likewise.
5070 (vml<as>l_n_<su>16): Likewise.
5071 (vmul<q>_lane<q>_<su>16): Likewise.
5072 (vmul<q>_n_<su>16): Likewise.
5073 (vmull_lane<q>_<su>16): Likewise.
5074 (vmull_n_<su>16): Likewise.
5075 (vmull_high_lane<q>_<su>16): Likewise.
5076 (vmull_high_n_<su>16): Likewise.
5077 (vqrdmulh<q>_n_s16): Likewise.
5079 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
5081 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
5082 have the correct lane parameter.
5084 2013-09-06 Richard Biener <rguenther@suse.de>
5086 * cfganal.c (control_dependences::~control_dependences):
5087 Properly free all of the vector.
5089 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
5092 * config/i386/i386.c (ix86_conditional_register_usage):
5093 Proper initialize extended SSE registers.
5095 2013-09-06 Jan Hubicka <jh@suse.cz>
5097 PR tree-optimization/58311
5098 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
5100 2013-09-06 Jan Hubicka <jh@suse.cz>
5102 * Makefile.in (tree-sra.o): Update dependencies.
5103 * tree-sra.c: Include ipa-utils.h
5104 (scan_function): Use recursive_call_p.
5105 (has_caller_p): New function.
5106 (cgraph_for_node_and_aliases): Count also callers of aliases.
5108 2013-09-06 Jan Hubicka <jh@suse.cz>
5111 * cgraph.h (symtab_semantically_equivalent_p): Declare.
5112 * tree-tailcall.c: Include ipa-utils.h.
5113 (find_tail_calls): Use it.
5114 * ipa-pure-const.c (check_call): Likewise.
5115 * ipa-utils.c (recursive_call_p): New function.
5116 * ipa-utils.h (recursive_call_p): Dclare.
5117 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
5118 (symtab_semantically_equivalent_p): New function.
5119 * Makefile.in (tree-tailcall.o): Update dependencies.
5121 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5123 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
5126 2013-09-06 Richard Biener <rguenther@suse.de>
5128 * lto-streamer.h (lto_global_var_decls): Remove.
5129 * Makefile.in (OBJS): Remove lto-symtab.o.
5130 (lto-symtab.o): Remove.
5131 (GTFILES): Remove lto-symtab.c
5132 * lto-symtab.c: Move to lto/
5134 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5136 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
5137 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
5138 (UNSPEC_FPINT_RINT): New constant definitions.
5139 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
5140 definition with 2 attributes.
5141 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
5142 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
5145 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5147 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
5148 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
5149 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
5150 attribute to "z196_alone".
5151 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
5154 2013-09-06 Richard Biener <rguenther@suse.de>
5156 * basic-block.h (class control_dependences): New.
5157 * tree-ssa-dce.c (control_dependence_map): Remove.
5159 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
5160 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
5161 find_pdom, find_control_dependence, find_all_control_dependences):
5163 (mark_control_dependent_edges_necessary,
5164 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
5165 tree_dce_done, perform_tree_ssa_dce): Adjust.
5166 * cfganal.c (set_control_dependence_map_bit,
5167 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
5168 find_all_control_dependences): Move from tree-ssa-dce.c and
5169 implement as methods of control_dependences class.
5170 (control_dependences::control_dependences): New.
5171 (control_dependences::~control_dependences): Likewise.
5172 (control_dependences::get_edges_dependent_on): Likewise.
5173 (control_dependences::get_edge): Likewise.
5175 2013-09-04 Jan Hubicka <jh@suse.cz>
5177 * tree.c (types_same_for_odr): Drop overactive check.
5178 * ipa-devirt.c (hash_type_name): Likewise.
5180 2013-09-04 Jan Hubicka <jh@suse.cz>
5182 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
5183 (analyze_functions): ... here.
5185 2013-09-04 Jan Hubicka <jh@suse.cz>
5188 * cgraphunit.c (analyze_functions): Clear AUX fields
5189 after processing; initialize assembler name has.
5191 2013-09-05 Jeff Law <law@redhat.com>
5193 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
5194 from thread_around_empty_block. Record threading path into PATH.
5195 Recurse if threading through the initial block is successful.
5196 (thread_across_edge): Corresponding changes to slightly simplify.
5198 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
5200 * config/aarch64/aarch64.md
5201 (type): Remove frecpe, frecps, frecpx.
5202 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
5203 fix to be a TARGET_SIMD instruction.
5204 (aarch64_frecps): Remove.
5205 * config/aarch64/aarch64-simd.md
5206 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
5207 (aarch64_frecps<mode>): Handle all float/vector of float modes.
5209 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
5210 Sofiane Naci <sofiane.naci@arm.com>
5212 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
5213 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
5214 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
5215 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
5216 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
5217 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
5218 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
5219 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
5220 into "clz, "rbit". Rename "shift" to "shift_imm".
5221 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
5222 changes. Update for attribute changes all occurrences of arlo_* and
5224 * config/arm/arm-fixed.md: Update for attribute changes
5225 all occurrences of arlo_* types.
5226 * config/arm/thumb2.md: Update for attribute changes all occurrences
5228 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
5229 (cortexa7_older_only): Likewise.
5230 (cortexa7_younger): Likewise.
5231 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5232 (1020alu_shift_op): Likewise.
5233 (1020alu_shift_reg_op): Likewise.
5234 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5235 (alu_shift_op): Likewise.
5236 (alu_shift_reg_op): Likewise.
5237 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5238 (11_alu_shift_op): Likewise.
5239 (11_alu_shift_reg_op): Likewise.
5240 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5241 (9_alu_shift_reg_op): Likewise.
5242 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
5244 (cortex_a15_alu_shift): Likewise.
5245 (cortex_a15_alu_shift_reg): Likewise.
5246 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
5248 (cortex_a5_alu_shift): Likewise.
5249 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
5251 (cortex_a53_alu_shift): Likewise.
5252 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
5254 (cortex_a7_alu_reg): Likewise.
5255 (cortex_a7_alu_shift): Likewise.
5256 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
5258 (cortex_a8_alu_shift): Likewise.
5259 (cortex_a8_alu_shift_reg): Likewise.
5260 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5261 (cortex_a9_dp_shift): Likewise.
5262 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
5264 * config/arm/cortex-r4.md
5265 (cortex_r4_alu): Update for attribute changes.
5266 (cortex_r4_mov): Likewise.
5267 (cortex_r4_alu_shift_reg): Likewise.
5268 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5269 (526_alu_shift_op): Likewise.
5270 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5271 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5272 (626te_alu_shift_op): Likewise.
5273 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
5274 (726te_alu_shift_op): Likewise.
5275 (726te_alu_shift_reg_op): Likewise.
5276 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5277 (mp626_alu_shift_op): Likewise.
5278 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
5279 (pj4_alu_conds): Likewise.
5280 (pj4_shift): Likewise.
5281 (pj4_shift_conds): Likewise.
5282 (pj4_alu_shift): Likewise.
5283 (pj4_alu_shift_conds): Likewise.
5284 * config/aarch64/aarch64.md: Update for attribute change
5285 all occurrences of arlo_* and shift* types.
5287 2013-09-05 Mike Stump <mikestump@comcast.net>
5289 * tree.h: Move documentation for tree_function_decl to tree-core.h
5290 with the declaration.
5292 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
5295 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
5296 looking for widest mode.
5298 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
5300 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
5302 2013-09-05 Richard Biener <rguenther@suse.de>
5304 PR tree-optimization/58137
5305 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
5306 Do not create vectors of pointers.
5307 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
5308 types for the components of the vector initializer.
5309 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
5310 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
5312 2013-09-05 Martin Jambor <mjambor@suse.cz>
5314 * ipa-prop.c (remove_described_reference): Accept missing references,
5315 return false if that hppens, otherwise return true.
5316 (cgraph_node_for_jfunc): New function.
5317 (try_decrement_rdesc_refcount): Likewise.
5318 (try_make_edge_direct_simple_call): Use them.
5319 (ipa_edge_removal_hook): Remove references from rdescs.
5320 (ipa_edge_duplication_hook): Clone rdescs and their references
5321 when the new edge has the same caller as the old one.
5322 * cgraph.c (cgraph_resolve_speculation): Remove speculative
5323 reference before removing any edges.
5325 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
5327 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
5329 * thumb2.md (thumb2_storewb_parisi): New pattern.
5331 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
5333 * config/aarch64/aarch64-option-extensions.def: Add
5334 AARCH64_OPT_EXTENSION of 'crc'.
5335 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
5336 (AARCH64_ISA_CRC): Ditto.
5337 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
5338 description of the CRC extension.
5340 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
5342 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
5343 * config/rs6000/linux.h: Ditto.
5344 * alpha/linux.h: Ditto.
5345 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
5346 no_c99_libc_has_function.
5347 * config/c6x/uclinux-elf.h: Ditto.
5348 * config/lm32/uclinux-elf.h: Ditto.
5349 * config/m68k/uclinux.h: Ditto.
5350 * config/moxie/uclinux.h: Ditto.
5351 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
5352 (crisv32-*-linux*, cris-*-linux*): Ditto.
5353 * config/bfin/bfin.c: Include "tm_p.h".
5355 2013-09-05 Richard Biener <rguenther@suse.de>
5357 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
5358 check for a definition without a basic-block.
5360 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
5361 Sofiane Naci <sofiane.naci@arm.com>
5363 * config/aarch64/aarch64.md
5364 (*movti_aarch64): Rename r_2_f and f_2_r.
5365 (*movsf_aarch64): Likewise.
5366 (*movdf_aarch64): Likewise.
5367 (*movtf_aarch64): Likewise.
5368 (aarch64_movdi_<mode>low): Likewise.
5369 (aarch64_movdi_<mode>high): Likewise.
5370 (aarch64_mov<mode>high_di): Likewise.
5371 (aarch64_mov<mode>low_di): Likewise.
5372 (aarch64_movtilow_tilow): Likewise.
5373 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
5374 values to config/arm/types.md
5375 (attribute "conds"): Update for attribute change.
5376 (anddi3_insn): Likewise.
5377 (iordi3_insn): Likewise.
5378 (xordi3_insn): Likewise.
5379 (one_cmpldi2): Likewise.
5380 * config/arm/types.md (type): Add Neon types.
5381 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
5382 use "type" attribute.
5383 (movmisalign<mode>_neon_store): Likewise.
5384 (movmisalign<mode>_neon_load): Likewise.
5385 (vec_set<mode>_internal): Likewise.
5386 (vec_setv2di_internal): Likewise.
5387 (vec_extract<mode>): Likewise.
5388 (vec_extractv2di): Likewise.
5389 (add<mode>3_neon): Likewise.
5390 (adddi3_neon): Likewise.
5391 (sub<mode>3_neon): Likewise.
5392 (subdi3_neon): Likewise.
5393 (mul<mode>3_neon): Likewise.
5394 (mul<mode>3add<mode>_neon): Likewise.
5395 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
5396 (fma<VCVTF:mode>4)): Likewise.
5397 (fma<VCVTF:mode>4_intrinsic): Likewise.
5398 (fmsub<VCVTF:mode>4)): Likewise.
5399 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
5400 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
5401 (ior<mode>3): Likewise.
5402 (and<mode>3): Likewise.
5403 (anddi3_neon): Likewise.
5404 (orn<mode>3_neon): Likewise.
5405 (orndi3_neon): Likewise.
5406 (bic<mode>3_neon): Likewise.
5407 (bicdi3_neon): Likewise.
5408 (xor<mode>3): Likewise.
5409 (one_cmpl<mode>2): Likewise.
5410 (abs<mode>2): Likewise.
5411 (neg<mode>2): Likewise.
5412 (umin<mode>3_neon): Likewise.
5413 (umax<mode>3_neon): Likewise.
5414 (smin<mode>3_neon): Likewise.
5415 (smax<mode>3_neon): Likewise.
5416 (vashl<mode>3): Likewise.
5417 (vashr<mode>3_imm): Likewise.
5418 (vlshr<mode>3_imm): Likewise.
5419 (ashl<mode>3_signed): Likewise.
5420 (ashl<mode>3_unsigned): Likewise.
5421 (neon_load_count): Likewise.
5422 (ashldi3_neon_noclobber): Likewise.
5423 (signed_shift_di3_neon): Likewise.
5424 (unsigned_shift_di3_neon): Likewise.
5425 (ashrdi3_neon_imm_noclobber): Likewise.
5426 (lshrdi3_neon_imm_noclobber): Likewise.
5427 (widen_ssum<mode>3): Likewise.
5428 (widen_usum<mode>3): Likewise.
5429 (quad_halves_<code>v4si): Likewise.
5430 (quad_halves_<code>v4sf): Likewise.
5431 (quad_halves_<code>v8hi): Likewise.
5432 (quad_halves_<code>v16qi): Likewise.
5433 (reduc_splus_v2di): Likewise.
5434 (neon_vpadd_internal<mode>): Likewise.
5435 (neon_vpsmin<mode>): Likewise.
5436 (neon_vpsmax<mode>): Likewise.
5437 (neon_vpumin<mode>): Likewise.
5438 (neon_vpumax<mode>): Likewise.
5439 (ss_add<mode>_neon): Likewise.
5440 (us_add<mode>_neon): Likewise.
5441 (ss_sub<mode>_neon): Likewise.
5442 (us_sub<mode>_neon): Likewise.
5443 (neon_vadd<mode>_unspec): Likewise.
5444 (neon_vaddl<mode>): Likewise.
5445 (neon_vaddw<mode>): Likewise.
5446 (neon_vhadd<mode>): Likewise.
5447 (neon_vqadd<mode>): Likewise.
5448 (neon_vaddhn<mode>): Likewise.
5449 (neon_vmul<mode>): Likewise.
5450 (neon_vmla<mode>): Likewise.
5451 (neon_vmlal<mode>): Likewise.
5452 (neon_vmls<mode>): Likewise.
5453 (neon_vmlsl<mode>): Likewise.
5454 (neon_vqdmulh<mode>): Likewise.
5455 (neon_vqdmlal<mode>): Likewise.
5456 (neon_vqdmlsl<mode>): Likewise.
5457 (neon_vmull<mode>): Likewise.
5458 (neon_vqdmull<mode>): Likewise.
5459 (neon_vsub<mode>_unspec): Likewise.
5460 (neon_vsubl<mode>): Likewise.
5461 (neon_vsubw<mode>): Likewise.
5462 (neon_vqsub<mode>): Likewise.
5463 (neon_vhsub<mode>): Likewise.
5464 (neon_vsubhn<mode>): Likewise.
5465 (neon_vceq<mode>): Likewise.
5466 (neon_vcge<mode>): Likewise.
5467 (neon_vcgeu<mode>): Likewise.
5468 (neon_vcgt<mode>): Likewise.
5469 (neon_vcgtu<mode>): Likewise.
5470 (neon_vcle<mode>): Likewise.
5471 (neon_vclt<mode>): Likewise.
5472 (neon_vcage<mode>): Likewise.
5473 (neon_vcagt<mode>): Likewise.
5474 (neon_vtst<mode>): Likewise.
5475 (neon_vabd<mode>): Likewise.
5476 (neon_vabdl<mode>): Likewise.
5477 (neon_vaba<mode>): Likewise.
5478 (neon_vabal<mode>): Likewise.
5479 (neon_vmax<mode>): Likewise.
5480 (neon_vmin<mode>): Likewise.
5481 (neon_vpaddl<mode>): Likewise.
5482 (neon_vpadal<mode>): Likewise.
5483 (neon_vpmax<mode>): Likewise.
5484 (neon_vpmin<mode>): Likewise.
5485 (neon_vrecps<mode>): Likewise.
5486 (neon_vrsqrts<mode>): Likewise.
5487 (neon_vqabs<mode>): Likewise.
5488 (neon_vqneg<mode>): Likewise.
5489 (neon_vcls<mode>): Likewise.
5490 (clz<mode>2): Likewise.
5491 (popcount<mode>2): Likewise.
5492 (neon_vrecpe): Likewise.
5493 (neon_vrsqrte): Likewise.
5494 (neon_vget_lane<mode>_sext_internal): Likewise.
5495 (neon_vget_lane<mode>_zext_internal): Likewise.
5496 (neon_vdup_n<mode>): Likewise.
5497 (neon_vdup_nv2di): Likewise.
5498 (neon_vdpu_lane<mode>_internal): Likewise.
5499 (neon_vswp<mode>): Likewise.
5500 (float<mode><V_cvtto>2): Likewise.
5501 (floatuns<mode><V_cvtto>2): Likewise.
5502 (fix_trunc<mode><V_cvtto>)2): Likewise
5503 (fixuns_trunc<mode><V_cvtto)2): Likewise.
5504 (neon_vcvt<mode>): Likewise.
5505 (neon_vcvtv4sfv4hf): Likewise.
5506 (neon_vcvtv4hfv4sf): Likewise.
5507 (neon_vcvt_n<mode>): Likewise.
5508 (neon_vmovn<mode>): Likewise.
5509 (neon_vqmovn<mode>): Likewise.
5510 (neon_vqmovun<mode>): Likewise.
5511 (neon_vmovl<mode>): Likewise.
5512 (neon_vmul_lane<mode>): Likewise.
5513 (neon_vmull_lane<mode>): Likewise.
5514 (neon_vqdmull_lane<mode>): Likewise.
5515 (neon_vqdmulh_lane<mode>): Likewise.
5516 (neon_vmla_lane<mode>): Likewise.
5517 (neon_vmlal_lane<mode>): Likewise.
5518 (neon_vqdmlal_lane<mode>): Likewise.
5519 (neon_vmls_lane<mode>): Likewise.
5520 (neon_vmlsl_lane<mode>): Likewise.
5521 (neon_vqdmlsl_lane<mode>): Likewise.
5522 (neon_vext<mode>): Likewise.
5523 (neon_vrev64<mode>): Likewise.
5524 (neon_vrev32<mode>): Likewise.
5525 (neon_vrev16<mode>): Likewise.
5526 (neon_vbsl<mode>_internal): Likewise.
5527 (neon_vshl<mode>): Likewise.
5528 (neon_vqshl<mode>): Likewise.
5529 (neon_vshr_n<mode>): Likewise.
5530 (neon_vshrn_n<mode>): Likewise.
5531 (neon_vqshrn_n<mode>): Likewise.
5532 (neon_vqshrun_n<mode>): Likewise.
5533 (neon_vshl_n<mode>): Likewise.
5534 (neon_vqshl_n<mode>): Likewise.
5535 (neon_vqshlu_n<mode>): Likewise.
5536 (neon_vshll_n<mode>): Likewise.
5537 (neon_vsra_n<mode>): Likewise.
5538 (neon_vsri_n<mode>): Likewise.
5539 (neon_vsli_n<mode>): Likewise.
5540 (neon_vtbl1v8qi): Likewise.
5541 (neon_vtbl2v8qi): Likewise.
5542 (neon_vtbl3v8qi): Likewise.
5543 (neon_vtbl4v8qi): Likewise.
5544 (neon_vtbx1v8qi): Likewise.
5545 (neon_vtbx2v8qi): Likewise.
5546 (neon_vtbx3v8qi): Likewise.
5547 (neon_vtbx4v8qi): Likewise.
5548 (neon_vtrn<mode>_internal): Likewise.
5549 (neon_vzip<mode>_internal): Likewise.
5550 (neon_vuzp<mode>_internal): Likewise.
5551 (neon_vld1<mode>): Likewise.
5552 (neon_vld1_lane<mode>): Likewise.
5553 (neon_vld1_dup<mode>): Likewise.
5554 (neon_vld1_dupv2di): Likewise.
5555 (neon_vst1<mode>): Likewise.
5556 (neon_vst1_lane<mode>): Likewise.
5557 (neon_vld2<mode>): Likewise.
5558 (neon_vld2_lane<mode>): Likewise.
5559 (neon_vld2_dup<mode>): Likewise.
5560 (neon_vst2<mode>): Likewise.
5561 (neon_vst2_lane<mode>): Likewise.
5562 (neon_vld3<mode>): Likewise.
5563 (neon_vld3qa<mode>): Likewise.
5564 (neon_vld3qb<mode>): Likewise.
5565 (neon_vld3_lane<mode>): Likewise.
5566 (neon_vld3_dup<mode>): Likewise.
5567 (neon_vst3<mode>): Likewise.
5568 (neon_vst3qa<mode>): Likewise.
5569 (neon_vst3qb<mode>): Likewise.
5570 (neon_vst3_lane<mode>): Likewise.
5571 (neon_vld4<mode>): Likewise.
5572 (neon_vld4qa<mode>): Likewise.
5573 (neon_vld4qb<mode>): Likewise.
5574 (neon_vld4_lane<mode>): Likewise.
5575 (neon_vld4_dup<mode>): Likewise.
5576 (neon_vst4<mode>): Likewise.
5577 (neon_vst4qa<mode>): Likewise.
5578 (neon_vst4qb<mode>): Likewise.
5579 (neon_vst4_lane<mode>): Likewise.
5580 (neon_vec_unpack<US>_lo_<mode>): Likewise.
5581 (neon_vec_unpack<US>_hi_<mode>): Likewise.
5582 (neon_vec_<US>mult_lo_<mode>): Likewise.
5583 (neon_vec_<US>mult_hi_<mode>): Likewise.
5584 (neon_vec_<US>shiftl_<mode>): Likewise.
5585 (neon_unpack<US>_<mode>): Likewise.
5586 (neon_vec_<US>mult_<mode>): Likewise.
5587 (vec_pack_trunc_<mode>): Likewise.
5588 (neon_vec_pack_trunk_<mode>): Likewise.
5589 (neon_vabd<mode>_2): Likewise.
5590 (neon_vabd<mode>_3): Likewise.
5591 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5592 (thumb2_movsi_vfp): Likewise.
5593 (movdi_vfp): Likewise.
5594 (movdi_vfp_cortexa8): Likewise.
5595 (movhf_vfp_neon): Likewise.
5596 (movhf_vfp): Likewiwse.
5597 (movsf_vfp): Likewiwse.
5598 (thumb2_movsf_vfp): Likewiwse.
5599 (movdf_vfp): Likewise.
5600 (thumb2_movdf_vfp): Likewise.
5601 (movsfcc_vfp): Likewise.
5602 (thumb2_movsfcc_vfp): Likewise.
5603 (movdfcc_vfp): Likewise.
5604 (thumb2_movdfcc_vfp): Likewise.
5605 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
5606 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
5607 (v10_v2c): Likewise.
5608 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
5610 (cortex_a15_neon_int_2): Likewise.
5611 (cortex_a15_neon_int_3): Likewise.
5612 (cortex_a15_neon_int_4): Likewise.
5613 (cortex_a15_neon_int_5): Likewise.
5614 (cortex_a15_neon_vqneg_vqabs): Likewise.
5615 (cortex_a15_neon_vmov): Likewise.
5616 (cortex_a15_neon_vaba): Likewise.
5617 (cortex_a15_neon_vaba_qqq): Likewise.
5618 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5619 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5620 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
5621 scalar_64_32_long_scalar): Likewise.
5622 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5623 (cortex_a15_neon_mla_qqq_8_16): Likewise.
5624 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5625 lotype_qdd_64_32_long): Likewise.
5626 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5627 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5628 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
5629 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5630 (cortex_a15_neon_shift_1): Likewise.
5631 (cortex_a15_neon_shift_2): Likewise.
5632 (cortex_a15_neon_shift_3): Likewise.
5633 (cortex_a15_neon_vshl_ddd): Likewise.
5634 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5635 (cortex_a15_neon_vsra_vrsra): Likewise.
5636 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
5637 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
5638 (cortex_a15_neon_fp_vmul_ddd): Likewise.
5639 (cortex_a15_neon_fp_vmul_qqd): Likewise.
5640 (cortex_a15_neon_fp_vmla_ddd): Likewise.
5641 (cortex_a15_neon_fp_vmla_qqq): Likewise.
5642 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
5643 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
5644 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5645 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5646 (cortex_a15_neon_bp_simple): Likewise.
5647 (cortex_a15_neon_bp_2cycle): Likewise.
5648 (cortex_a15_neon_bp_3cycle): Likewise.
5649 (cortex_a15_neon_vld1_1_2_regs): Likewise.
5650 (cortex_a15_neon_vld1_3_4_regs): Likewise.
5651 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5652 (cortex_a15_neon_vld2_4_regs): Likewise.
5653 (cortex_a15_neon_vld3_vld4): Likewise.
5654 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5655 (cortex_a15_neon_vst1_3_4_regs): Likewise.
5656 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
5657 (cortex_a15_neon_vst3_vst4): Likewise.
5658 (cortex_a15_neon_vld1_vld2_lane): Likewise.
5659 (cortex_a15_neon_vld3_vld4_lane" 10
5660 (cortex_a15_neon_vst1_vst2_lane): Likewise.
5661 (cortex_a15_neon_vst3_vst4_lane): Likewise.
5662 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
5663 (cortex_a15_neon_ldm_2): Likewise.0
5664 (cortex_a15_neon_stm_2): Likewise.
5665 (cortex_a15_neon_mcr): Likewise.
5666 (cortex_a15_neon_mcr_2_mcrr): Likewise.
5667 (cortex_a15_neon_mrc): Likewise.
5668 (cortex_a15_neon_mrrc): Likewise.
5669 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5671 (cortex_a15_alu_shift): Likewise.
5672 (cortex_a15_alu_shift_reg): Likewise.
5673 (cortex_a15_mult32): Likewise.
5674 (cortex_a15_mult64): Likewise.
5675 (cortex_a15_block): Likewise.
5676 (cortex_a15_branch): Likewise.
5677 (cortex_a15_load1): Likewise.
5678 (cortex_a15_load3): Likewise.
5679 (cortex_a15_store1): Likewise.
5680 (cortex_a15_store3): Likewise.
5681 (cortex_a15_call): Likewise.
5682 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
5683 (cortex_a5_f2r): Likewise.
5684 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
5686 (cortex_a53_f2r): Likewise.
5687 * config/arm/cortex-a7.md
5688 (cortex_a7_branch): Update for attribute change.
5689 (cortex_a7_call): Likewise.
5690 (cortex_a7_alu_imm): Likewise.
5691 (cortex_a7_alu_reg): Likewise.
5692 (cortex_a7_alu_shift): Likewise.
5693 (cortex_a7_mul): Likewise.
5694 (cortex_a7_load1): Likewise.
5695 (cortex_a7_store1): Likewise.
5696 (cortex_a7_load2): Likewise.
5697 (cortex_a7_store2): Likewise.
5698 (cortex_a7_load3): Likewise.
5699 (cortex_a7_store3): Likewise.
5700 (cortex_a7_load4): Likewise.
5701 (cortex_a7_store4): Likewise.
5702 (cortex_a7_fpalu): Likewise.
5703 (cortex_a7_fconst): Likewise.
5704 (cortex_a7_fpmuls): Likewise.
5705 (cortex_a7_neon_mul): Likewise.
5706 (cortex_a7_fpmacs): Likewise.
5707 (cortex_a7_neon_mla: Likewise.
5708 (cortex_a7_fpmuld: Likewise.
5709 (cortex_a7_fpmacd: Likewise.
5710 (cortex_a7_fpfmad: Likewise.
5711 (cortex_a7_fdivs: Likewise.
5712 (cortex_a7_fdivd: Likewise.
5713 (cortex_a7_r2f: Likewise.
5714 (cortex_a7_f2r: Likewise.
5715 (cortex_a7_f_flags: Likewise.
5716 (cortex_a7_f_loads: Likewise.
5717 (cortex_a7_f_loadd: Likewise.
5718 (cortex_a7_f_stores: Likewise.
5719 (cortex_a7_f_stored: Likewise.
5720 (cortex_a7_neon): Likewise.
5721 * config/arm/cortex-a8-neon.md
5722 (cortex_a8_neon_mrc): Update for attribute change.
5723 (cortex_a8_neon_mrrc): Likewise.
5724 (cortex_a8_neon_int_1): Likewise.
5725 (cortex_a8_neon_int_2): Likewise.
5726 (cortex_a8_neon_int_3): Likewise.
5727 (cortex_a8_neon_int_4): Likewise.
5728 (cortex_a8_neon_int_5): Likewise.
5729 (cortex_a8_neon_vqneg_vqabs): Likewise.
5730 (cortex_a8_neon_vmov): Likewise.
5731 (cortex_a8_neon_vaba): Likewise.
5732 (cortex_a8_neon_vaba_qqq): Likewise.
5733 (cortex_a8_neon_vsma): Likewise.
5734 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5735 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5736 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
5737 long_scalar): Likewise.
5738 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5739 (cortex_a8_neon_mla_qqq_8_16): Likewise.
5740 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5741 long_scalar_qdd_64_32_long): Likewise.
5742 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5743 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5744 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
5745 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5746 (cortex_a8_neon_shift_1): Likewise.
5747 (cortex_a8_neon_shift_2): Likewise.
5748 (cortex_a8_neon_shift_3): Likewise.
5749 (cortex_a8_neon_vshl_ddd): Likewise.
5750 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5751 (cortex_a8_neon_vsra_vrsra): Likewise.
5752 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
5753 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
5754 (cortex_a8_neon_fp_vsum): Likewise.
5755 (cortex_a8_neon_fp_vmul_ddd): Likewise.
5756 (cortex_a8_neon_fp_vmul_qqd): Likewise.
5757 (cortex_a8_neon_fp_vmla_ddd): Likewise.
5758 (cortex_a8_neon_fp_vmla_qqq): Likewise.
5759 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
5760 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
5761 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5762 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5763 (cortex_a8_neon_bp_simple): Likewise.
5764 (cortex_a8_neon_bp_2cycle): Likewise.
5765 (cortex_a8_neon_bp_3cycle): Likewise.
5766 (cortex_a8_neon_ldr): Likewise.
5767 (cortex_a8_neon_str): Likewise.
5768 (cortex_a8_neon_vld1_1_2_regs): Likewise.
5769 (cortex_a8_neon_vld1_3_4_regs): Likewise.
5770 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5771 (cortex_a8_neon_vld2_4_regs): Likewise.
5772 (cortex_a8_neon_vld3_vld4): Likewise.
5773 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5774 (cortex_a8_neon_vst1_3_4_regs): Likewise.
5775 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
5776 (cortex_a8_neon_vst3_vst4): Likewise.
5777 (cortex_a8_neon_vld1_vld2_lane): Likewise.
5778 (cortex_a8_neon_vld3_vld4_lane): Likewise.
5779 (cortex_a8_neon_vst1_vst2_lane): Likewise.
5780 (cortex_a8_neon_vst3_vst4_lane): Likewise.
5781 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
5782 (cortex_a8_neon_mcr): Likewise.
5783 (cortex_a8_neon_mcr_2_mcrr): Likewise.
5784 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5785 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
5787 (ca9_neon_mrrc): Likewise.
5788 (cortex_a9_neon_int_1): Likewise.
5789 (cortex_a9_neon_int_2): Likewise.
5790 (cortex_a9_neon_int_3): Likewise.
5791 (cortex_a9_neon_int_4): Likewise.
5792 (cortex_a9_neon_int_5): Likewise.
5793 (cortex_a9_neon_vqneg_vqabs): Likewise.
5794 (cortex_a9_neon_vmov): Likewise.
5795 (cortex_a9_neon_vaba): Likewise.
5796 (cortex_a9_neon_vaba_qqq): Likewise.
5797 (cortex_a9_neon_vsma): Likewise.
5798 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5799 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
5800 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
5801 long_scalar): Likewise.
5802 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
5803 (cortex_a9_neon_mla_qqq_8_16): Likewise.
5804 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
5805 long_scalar_qdd_64_32_long): Likewise.
5806 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
5807 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
5808 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
5809 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
5810 (cortex_a9_neon_shift_1): Likewise.
5811 (cortex_a9_neon_shift_2): Likewise.
5812 (cortex_a9_neon_shift_3): Likewise.
5813 (cortex_a9_neon_vshl_ddd): Likewise.
5814 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
5815 (cortex_a9_neon_vsra_vrsra): Likewise.
5816 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
5817 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
5818 (cortex_a9_neon_fp_vsum): Likewise.
5819 (cortex_a9_neon_fp_vmul_ddd): Likewise.
5820 (cortex_a9_neon_fp_vmul_qqd): Likewise.
5821 (cortex_a9_neon_fp_vmla_ddd): Likewise.
5822 (cortex_a9_neon_fp_vmla_qqq): Likewise.
5823 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
5824 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
5825 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5826 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5827 (cortex_a9_neon_bp_simple): Likewise.
5828 (cortex_a9_neon_bp_2cycle): Likewise.
5829 (cortex_a9_neon_bp_3cycle): Likewise.
5830 (cortex_a9_neon_ldr): Likewise.
5831 (cortex_a9_neon_str): Likewise.
5832 (cortex_a9_neon_vld1_1_2_regs): Likewise.
5833 (cortex_a9_neon_vld1_3_4_regs): Likewise.
5834 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5835 (cortex_a9_neon_vld2_4_regs): Likewise.
5836 (cortex_a9_neon_vld3_vld4): Likewise.
5837 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5838 (cortex_a9_neon_vst1_3_4_regs): Likewise.
5839 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
5840 (cortex_a9_neon_vst3_vst4): Likewise.
5841 (cortex_a9_neon_vld1_vld2_lane): Likewise.
5842 (cortex_a9_neon_vld3_vld4_lane): Likewise.
5843 (cortex_a9_neon_vst1_vst2_lane): Likewise.
5844 (cortex_a9_neon_vst3_vst4_lane): Likewise.
5845 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5846 (cortex_a9_neon_mcr): Likewise.
5847 (cortex_a9_neon_mcr_2_mcrr): Likewise.
5848 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
5849 (cortex_a9_fps): Likewise.
5850 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
5852 (cortex_m4_fmuls): Likewise.
5853 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
5855 (cortex_r4_mrc): Likewise.
5856 * config/arm/iterators.md: Update comment referring to neon_type.
5857 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
5858 (iwmmxt_movsi_insn): Likewise.
5859 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
5861 (pj4_core_to_vfp): Likewise.
5862 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
5864 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
5865 (vfp_fstore): Likewise.
5866 * doc/md.texi: Change references to neon_type to refer to type.
5868 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5870 * tree.h (DECL_BUILT_IN): Fix typo in comment.
5872 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
5874 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
5877 2013-09-04 Easwaran Raman <eraman@google.com>
5880 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
5881 (build_and_add_sum): Use it.
5882 (appears_later_in_bb): Simplify code.
5884 2013-09-04 Teresa Johnson <tejohnson@google.com>
5886 * dumpfile.c (dump_finish): Don't close stderr/stdout.
5888 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
5890 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
5892 2013-09-04 Jan Hubicka <jh@suse.cz>
5894 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
5895 * ipa-devirt.c: Include diganostic.h
5896 (odr_type_d): Add types and types_set.
5897 (hash_type_name): Work for types with vtables during LTO.
5898 (odr_hasher::remove): Fix comment; destroy types_set.
5899 (add_type_duplicate): New function,
5900 (get_odr_type): Use it.
5901 (dump_type_inheritance_graph): Dump type duplicates.
5902 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
5904 * tree.c (types_same_for_odr): Give exact answers on types with
5907 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5909 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
5910 explaining their differences.
5912 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
5914 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
5916 2013-09-03 Jeff Law <law@redhat.com>
5918 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
5919 when not threading through a joiner block. Pass joiner/no joiner
5920 state to register_jump_thread.
5921 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
5922 state from argument rather than implying on path length.
5923 Dump the entire jump thread path into debugging dump.
5924 * tree-flow.h (register_jump_thread): Update prototype.
5926 2013-08-29 Xinliang David Li <davidxl@google.com>
5928 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5929 Remove a trivial gcc_assert.
5931 2013-08-29 Xinliang David Li <davidxl@google.com>
5933 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
5934 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
5935 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5936 Delay base decl alignment adjustment.
5937 * tree-vectorizer.c (vect_destroy_datarefs): New function.
5938 * tree-vectorizer.h: New data structure.
5939 (set_dr_misalignment): New function.
5940 (dr_misalignment): Ditto.
5941 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
5942 (vectorizable_load): Ditto.
5943 (ensure_base_align): New function.
5944 (vectorize_loops): Add dbg_cnt support.
5945 (execute_vect_slp): Ditto.
5946 * dbgcnt.def: New debug counter.
5947 * Makefile: New dependency.
5949 2013-09-03 Meador Inge <meadori@codesourcery.com>
5953 2013-08-30 Meador Inge <meadori@codesourcery.com>
5955 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
5957 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
5959 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
5960 function descriptor.
5962 2013-09-03 Richard Biener <rguenther@suse.de>
5964 * tree-affine.c (add_elt_to_tree): Fix association issue,
5965 avoid useless converts and make sure to always return a
5966 properly typed result.
5968 2013-09-03 Richard Biener <rguenther@suse.de>
5971 * fold-const.c (negate_expr_p): Fix division case.
5972 (negate_expr): Likewise.
5974 2013-09-03 Richard Biener <rguenther@suse.de>
5977 * tree-streamer-out.c: Include tm.h.
5978 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
5980 2013-09-03 Jan Hubicka <jh@suse.cz>
5982 * tree-profile.c (tree_profiling): Cleanup CFG when done.
5984 2013-09-03 Alan Modra <amodra@gmail.com>
5986 * config.gcc (powerpc*-*-linux*): Add support for little-endian
5987 multilibs to big-endian target and vice versa.
5988 * config/rs6000/t-linux64: Use := assignment on all vars.
5989 (MULTILIB_EXTRA_OPTS): Remove fPIC.
5990 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
5991 * config/rs6000/t-linux64le: New file.
5992 * config/rs6000/t-linux64bele: New file.
5993 * config/rs6000/t-linux64lebe: New file.
5995 2013-09-02 Jan Hubicka <jh@suse.cz>
5997 * ipa-inline-transform.c (inline_transform): Do not
5998 optimize_inline_calls when not optimizing.
6000 2013-09-02 Jan Hubicka <jh@suse.cz>
6002 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
6003 duplicated nodes for assembler names.
6004 * symtab.c (symtab_unregister_node): Do not attempt to unlink
6005 hard registers from assembler name hash.
6007 2013-09-02 Jan Hubicka <jh@suse.cz>
6009 * ipa-split.c (execute_split_functions): Split externally visible
6010 functions called once.
6012 2013-09-02 Martin Jambor <mjambor@suse.cz>
6015 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
6016 linked list. When finding the correct duplicate, also consider also
6017 the caller in additon to its inlined_to node.
6019 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
6021 * config/aarch64/aarch64-simd-builtins.def
6022 (dup_lane_scalar): Remove.
6023 * config/aarch64/aarch64-simd.md
6024 (aarch64_simd_dup): Add 'w->w' alternative.
6025 (aarch64_dup_lane<mode>): Allow for VALL.
6026 (aarch64_dup_lane_scalar<mode>): Remove.
6027 (aarch64_dup_lane_<vswap_width_name><mode>): New.
6028 (aarch64_get_lane_signed<mode>): Add w->w altenative.
6029 (aarch64_get_lane_unsigned<mode>): Likewise.
6030 (aarch64_get_lane<mode>): Likewise.
6031 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
6032 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
6033 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
6034 (VCON): Change container of V2SF.
6035 (vswap_width_name): Likewise.
6036 * config/aarch64/arm_neon.h
6037 (__aarch64_vdup_lane_any): New.
6038 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6039 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
6040 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6042 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
6045 * expr.c (emit_move_complex): Do not move complex FP values as parts if
6046 the source or the destination is a single hard register.
6048 2013-09-02 Richard Biener <rguenther@suse.de>
6051 * tree-scalar-evolution.c (instantiate_scev_name): Allow
6054 2013-09-02 Richard Biener <rguenther@suse.de>
6056 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
6057 arithmetic to sizetype.
6059 2013-09-02 Bin Cheng <bin.cheng@arm.com>
6061 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
6062 Find auto-increment use both before and after candidate.
6064 2013-09-02 Marek Polacek <polacek@redhat.com>
6066 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
6068 2013-09-01 Jan Hubicka <jh@suse.cz>
6070 * Makefile.in: Add ipa-profile.o
6071 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
6072 * cgraph.c (struct cgraph_propagate_frequency_data,
6073 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
6074 ipa-profile.c; replace cgraph_ by ipa_ prefix.
6075 * cgraph.h (cgraph_propagate_frequency): Remove.
6076 * ipa-inline-analysis.c: Include ipa-utils.h;
6077 drop duplicated cfgloop.h.
6078 (inline_update_callee_summaries): Update.
6079 * ipa-profile.c: New file.
6080 * ipa-utils.h (ipa_propagate_frequency): Declare.
6081 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
6082 data-streamer.h, value-prof.h.
6083 (symtab_remove_unreachable_nodes): Update profile.
6084 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
6085 account_time_size, cmp_counts, dump_histogram,
6086 ipa_profile_generate_summary, ipa_profile_write_summary,
6087 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
6088 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
6089 Move to ipa-profile.c.
6091 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
6093 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
6095 2013-09-01 Jan Hubicka <jh@suse.cz>
6097 * common.opt (fdevirtualize-speculatively): New function.
6098 * invoke.texi (fdevirtualize-speculatively): Document.
6099 * ipa-devirt.c: Include ipa-inline.h
6100 (likely_target_p): New function.
6101 (ipa_devirt): New function.
6102 (gate_ipa_devirt): New function.
6103 (pass_data_ipa_devirt): New static var.
6104 (pass_ipa_devirt): Likewise.
6105 (make_pass_ipa_devirt): New function.
6106 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
6107 (common_handle_option): Disable devirtualization when
6108 value range profiling is available.
6109 * passes.def (pass_ipa_devirt): Add.
6110 * timever.def (TV_IPA_DEVIRT): New timevar.
6111 * tree-pass.h (make_pass_ipa_devirt):
6113 2013-09-01 Iain Sandoe <iain@codesourcery.com>
6115 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
6116 include sanitize(undefined).
6118 2013-08-31 Diego Novillo <dnovillo@google.com>
6120 * Makefile.in (TREE_CORE_H): Define.
6122 (GTFILES): Add tree-core.h.
6123 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
6125 * tree-core.h: New file.
6126 Move all data structures, enum, typedefs, global
6127 declarations and constants from ...
6130 2013-08-31 Jan Hubicka <jh@suse.cz>
6132 * bulitins.c (expand_builtin): Do not early exit for gcov
6133 instrumented functions.
6135 2013-08-31 Marek Polacek <polacek@redhat.com>
6137 * ubsan.c: Include tm_p.h.
6139 2013-08-31 Jan Hubicka <jh@suse.cz>
6141 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
6144 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
6145 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
6146 * tree-cfg.c (verify_gimple_label): ... here.
6147 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
6148 (ipa_merge_profiles): New function.
6149 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
6150 (lto_input_function_body): Likewise.
6151 * ipa-utils.h (ipa_merge_profiles): Declare.
6152 * lto-streamer.h (lto_input_function_body): Update prototype.
6153 (emit_label_in_global_context_p): Remove.
6154 * lto-symtab.c: Include ipa-utils.h
6155 (lto_cgraph_replace_node): Use ipa_merge_profiles.
6157 2013-08-31 Jan Hubicka <jh@suse.cz>
6159 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
6161 2013-08-31 Jan Hubicka <jh@suse.cz>
6163 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
6165 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
6167 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
6168 "cmp" RTX before signed_comparison_operator check to account
6171 2013-08-30 Jan Hubicka <jh@suse.cz>
6173 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
6174 (detect_type_change_1): Rename to ...
6175 (detect_type_change): ... this one; early return on non-polymorphic
6177 (detect_type_change_ssa): Add comp_type parameter; update
6178 use of detect_type_change.
6179 (compute_complex_assign_jump_func): Add param_type parameter;
6180 update use of detect_type_change_ssa.
6181 (compute_complex_ancestor_jump_func): Likewise.
6182 (ipa_get_callee_param_type): New function.
6183 (ipa_compute_jump_functions_for_edge): Compute parameter type;
6184 update calls to the jump function computation functions.
6186 2013-08-30 Teresa Johnson <tejohnson@google.com>
6187 Steven Bosscher <steven@gcc.gnu.org>
6189 * cfgrtl.c (fixup_new_cold_bb): New routine.
6190 (commit_edge_insertions): Invoke fixup_partitions.
6191 (find_partition_fixes): New routine.
6192 (fixup_partitions): Ditto.
6193 (verify_hot_cold_block_grouping): Update comments.
6194 (rtl_verify_edges): Invoke find_partition_fixes.
6195 (rtl_verify_bb_pointers): Update comments.
6196 (rtl_verify_bb_layout): Ditto.
6197 * basic-block.h (probably_never_executed_edge_p): Declare.
6198 (fixup_partitions): Ditto.
6199 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
6200 * bb-reorder.c (sanitize_hot_paths): New function.
6201 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
6203 * predict.c (probably_never_executed_edge_p): New routine.
6204 * cfg.c (check_bb_profile): Add partition insanity warnings.
6206 2013-08-30 Meador Inge <meadori@codesourcery.com>
6208 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
6210 2013-08-30 Marek Polacek <polacek@redhat.com>
6212 * Makefile.in (ubsan.o): Add.
6213 (c-family/c-ubsan.o): Add.
6214 (builtins.o): Add ubsan.h dependency.
6215 * ubsan.h: New file.
6216 * ubsan.c: New file.
6217 * common.opt: Add -fsanitize=undefined option.
6218 (flag_sanitize): Add variable.
6219 (fsanitize=): Add option. Add Driver.
6220 (fsanitize=thread): Remove option.
6221 (fsanitize=address): Likewise.
6222 (static-libubsan): New option.
6223 * doc/invoke.texi: Document the new flag and -static-libubsan.
6224 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
6225 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
6226 * builtin-attrs.def (ATTR_COLD): Define.
6227 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
6228 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
6229 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
6230 * flag-types.h (sanitize_code): New enum.
6231 * opts.c (common_handle_option): Parse command line arguments
6232 of -fsanitize=. Add -fsanitize=unreachable option.
6233 * varasm.c (get_variable_section): Adjust.
6234 (assemble_noswitch_variable): Likewise.
6235 (assemble_variable): Likewise.
6236 (output_constant_def_contents): Likewise.
6237 (categorize_decl_for_section): Likewise.
6238 (place_block_symbol): Likewise.
6239 (output_object_block): Likewise.
6240 * builtins.def: Likewise.
6241 * toplev.c (compile_file): Likewise.
6242 (process_options): Likewise.
6243 * cppbuiltin.c: Likewise.
6244 * tsan.c (tsan_pass): Likewise.
6245 (tsan_gate): Likewise.
6246 (tsan_gate_O0): Likewise.
6247 * cfgexpand.c (partition_stack_vars): Likewise.
6248 (expand_stack_vars): Likewise.
6249 (defer_stack_allocation): Likewise.
6250 (expand_used_vars): Likewise.
6251 * cfgcleanup.c (old_insns_match_p): Likewise.
6252 * asan.c (asan_finish_file): Likewise.
6253 (asan_instrument): Likewise.
6254 (gate_asan): Likewise.
6255 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
6256 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
6257 (asan_global_struct): Use pointer_sized_int_node instead
6258 calling build_nonstandard_integer_type.
6259 (initialize_sanitizer_builtins): Likewise.
6260 (asan_finish_file): Likewise.
6261 * gcc.c: Document %{%:function(args):X}.
6262 (static_spec_functions): Add sanitize.
6263 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
6264 store funcval != NULL there.
6265 (do_spec_1): Adjust handle_spec_function caller.
6266 (handle_braces): Allow %:function(args) as condition.
6267 (sanitize_spec_function): New function.
6268 (ADD_STATIC_LIBUBSAN_LIBS): Define.
6269 (LIBUBSAN_SPEC): Likewise.
6270 (LIBUBSAN_EARLY_SPEC): Likewise.
6271 (SANITIZER_SPEC): Handle libubsan.
6272 (SANITIZER_EARLY_SPEC): Likewise.
6273 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
6274 instead of fsanitize=address.
6275 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
6276 instead of fsanitize=address*.
6277 * builtins.c: Include ubsan.h.
6278 (fold_builtin_0): Instrument __builtin_unreachable.
6279 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
6280 instead of flag_asan.
6281 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
6282 (pointer_sized_int_node): Define.
6283 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
6285 2013-08-30 Mike Stump <mikestump@comcast.net>
6287 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
6290 2013-08-29 Jan Hubicka <jh@suse.cz>
6292 * cgraph.c (cgraph_function_body_availability): Handle weakref
6294 * passes.def: Remove pass_fixup_cfg.
6295 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
6296 track when we need to remove functions.
6297 (gate_ipa_inline): Execute inlining always; add comment why.
6298 (pass_data_ipa_inline): Remove TODO_remove_functions.
6299 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
6300 do not produce summaries.
6301 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
6302 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
6303 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
6304 constant pool and vtable.
6306 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
6308 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
6309 New arm_neon.h's internal macros to specify 64-bit constants.
6310 Avoid using stdint.h's macros.
6312 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
6314 * recog.c (verify_changes): Verify that changes[i].old is non-zero
6315 before applying REG_P.
6317 2013-08-30 Jakub Jelinek <jakub@redhat.com>
6319 PR tree-optimization/58277
6320 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
6321 after seeing too many stmts with vdef in between dombb and current
6322 bb, invalidate everything.
6324 2013-08-30 Richard Biener <rguenther@suse.de>
6326 * fold-const.c (fold_single_bit_test): Fix overflow test.
6328 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
6330 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
6331 and which can live in a register, always retrieve the value on entry.
6332 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
6333 passed by invisible reference specially.
6334 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
6335 (vt_add_function_parameter): Correctly deal with a parameter passed by
6336 invisible reference.
6338 2013-08-30 Jan Hubicka <jh@suse.cz>
6340 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
6342 2013-08-30 Richard Biener <rguenther@suse.de>
6344 PR tree-optimization/58228
6345 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
6346 allow invariant loads in nested loop vectorization.
6348 2013-08-30 Richard Biener <rguenther@suse.de>
6350 PR tree-optimization/58223
6351 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
6352 (has_anti_or_output_dependence): ... this and adjust to also
6353 look for output dependences.
6354 (mark_nodes_having_upstream_mem_writes): Adjust.
6355 (rdg_flag_uses): Likewise.
6357 2013-08-30 Richard Biener <rguenther@suse.de>
6359 PR tree-optimization/58010
6360 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
6361 assert that we have a loop-closed PHI.
6363 2013-08-29 Jan Hubicka <jh@suse.cz>
6365 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
6366 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
6367 * lto-section-in.c (lto_free_function_in_decl_state): New function.
6368 (lto_free_function_in_decl_state_for_node): New function.
6370 2013-08-29 Xinliang David Li <davidxl@google.com>
6372 * loop-unroll.c (report_unroll_peel): Minor message change.
6373 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
6374 Emit alignment peeling message with default -fopt-info.
6375 (vect_loop_versioning): Emit loop version info message.
6376 * tree-vectorizer.c (vectorize_loops): Minor message change.
6377 (execute_vect_slp): Ditto.
6379 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
6381 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
6382 of the clone from the DECL_NAME of the original function.
6384 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
6386 * passes.c (register_pass): Add overload.
6387 * tree-pass.h (register_pass): Forward declare it. Add comment.
6389 2013-08-29 Jan Hubicka <jh@suse.cz>
6391 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
6392 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
6393 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
6394 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
6395 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
6396 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
6397 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
6398 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
6400 2013-08-29 Teresa Johnson <tejohnson@google.com>
6402 * dumpfile.c (dump_loc): Output column number.
6403 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
6404 * doc/invoke.texi: Document optall -fopt-info flag.
6405 * profile.c (read_profile_edge_counts): Use new dump framework.
6406 (compute_branch_probabilities): Ditto.
6407 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
6408 when pass not in any opt group.
6409 * pass_manager.h (pass_manager::get_pass_profile): New method.
6410 * value-prof.c (check_counter): Use new dump framework.
6411 (check_ic_target): Ditto.
6412 * coverage.c (get_coverage_counts): Ditto.
6413 (coverage_init): Setup new dump framework.
6415 2013-08-29 Richard Biener <rguenther@suse.de>
6417 PR tree-optimization/58246
6418 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
6419 handle the dominance check inside a basic-block.
6421 2013-08-29 Richard Biener <rguenther@suse.de>
6424 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
6425 of default defs that appear in abnormal PHI nodes.
6427 2013-08-29 Richard Biener <rguenther@suse.de>
6429 PR tree-optimization/57685
6430 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
6431 single-use operands to avoid exponential complexity.
6433 2013-08-28 Dehao Chen <dehao@google.com>
6435 * ipa-inline.c (edge_badness): Fix integer underflow.
6437 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
6439 * gtm-builtins.def (_ITM_free): Declare leaf.
6441 2013-08-28 Jakub Jelinek <jakub@redhat.com>
6444 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
6445 (*tls_local_dynamic_base_64_largepic): Likewise.
6446 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
6447 Remove predicate from call operand.
6448 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
6449 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
6451 2013-08-28 Jeff Law <law@redhat.com>
6453 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
6454 checks for the number of predecessors and successors allowed.
6455 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
6456 which require copying a joiner block if there is a request which
6457 is a subpath that requires no joiner block copying.
6459 2013-08-28 Jan Hubicka <jh@suse.cz>
6461 * lto-streamer-out.c (DFS_write_tree_body): Drop
6462 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
6463 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
6464 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
6465 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
6466 Do not read DECL_ERROR_ISSUED.
6467 (unpack_ts_decl_with_vis_value_fields): Do not read
6469 (lto_input_ts_binfo_tree_pointers): Do not read
6470 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
6471 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
6472 write DECL_ERROR_ISSUED..
6473 (pack_ts_decl_with_vis_value_fields): Do not write
6475 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
6476 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
6477 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
6478 * tree.h (tree_decl_common): Update comment.
6479 (DECL_ERROR_ISSUED): Remove.
6481 2013-08-28 Jakub Jelinek <jakub@redhat.com>
6484 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
6486 2013-08-28 Jan Hubicka <jh@suse.cz>
6488 * builtins.def (free): Declare leaf.
6490 2013-08-27 David Malcolm <dmalcolm@redhat.com>
6493 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
6494 * configure: Regenerate.
6496 2013-08-27 Martin Jambor <mjambor@suse.cz>
6498 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
6499 (ipa_ancestor_jf_data): Likewise.
6500 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
6501 (ipa_get_jf_pass_through_type_preserved): New function.
6502 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
6503 (ipa_get_jf_ancestor_type_preserved): New function.
6504 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
6505 (ipa_get_jf_ancestor_result): Likewise.
6506 (propagate_vals_accross_pass_through): Use
6507 ipa_get_jf_pass_through_result to do all the value mappings.
6508 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
6509 type_preserved flag.
6510 (ipa_set_jf_cst_copy): New function.
6511 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
6512 (ipa_set_jf_arith_pass_through): Likewise.
6513 (ipa_set_ancestor_jf): Likewise.
6514 (compute_complex_assign_jump_func): Set type_preserved instead of
6516 (ipa_compute_jump_functions_for_edge): Likewise.
6517 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
6518 (update_jump_functions_after_inlining): Update type_preserved.
6519 Explicitely create jump functions when combining one with pass_through.
6520 (ipa_write_jump_function): Stream the type_preserved flags.
6521 (ipa_read_jump_function): Likewise.
6523 2013-08-27 Jakub Jelinek <jakub@redhat.com>
6524 Aldy Hernandez <aldyh@redhat.com>
6526 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
6527 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
6528 * function.h (struct function): Add has_force_vect_loops and
6530 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
6531 * gimple.c (gimple_build_omp_critical): Add KIND argument and
6533 * gimple.def: Update CLAUSES comments.
6534 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
6535 (gimple_build_omp_for): Add argument to prototype.
6536 (gimple_omp_for_kind): New.
6537 (gimple_omp_for_set_kind): New.
6538 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
6539 GOVD_DATA_SHARE_CLASS.
6540 (enum omp_region_type): Add ORT_SIMD.
6541 (gimple_add_tmp_var): Handle ORT_SIMD.
6542 (gimplify_var_or_parm_decl): Same.
6543 (is_gimple_stmt): Same.
6544 (omp_firstprivatize_variable): Same.
6545 (omp_add_variable): Only use splay_tree_insert if lookup failed.
6546 (omp_notice_variable): Handle ORT_SIMD.
6547 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
6548 (omp_check_private): Handle ORT_SIMD.
6549 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
6551 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
6552 Handle OMP_CLAUSE_LASTPRIVATE.
6553 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
6555 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
6556 (gimplify_expr): Handle OMP_SIMD.
6557 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
6558 (expand_GOMP_SIMD_VF): New.
6559 (expand_GOMP_SIMD_LAST_LANE): New.
6560 * internal-fn.def (GOMP_SIMD_LANE): New.
6561 (GOMP_SIMD_VF): New.
6562 (GOMP_SIMD_LAST_LANE): New.
6563 * omp-low.c: Include target.h.
6564 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
6566 (check_omp_nesting_restrictions): Same.
6568 (lower_rec_simd_input_clauses): New.
6569 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
6571 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
6572 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
6573 (expand_omp_build_assign): New.
6574 (expand_omp_for_init_counts): New.
6575 (expand_omp_for_init_vars): New.
6576 (extract_omp_for_update_vars): New.
6577 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
6578 and rewrite accordingly.
6579 (expand_omp_simd): New.
6580 (expand_omp_for): Use expand_omp_simd.
6581 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
6582 (lower_omp_for): Do not lower the body.
6583 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
6585 * tree-flow.h (find_omp_clause): Remove prototype.
6586 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
6587 forcing vectorization of the loop, or if flag_tree_vectorize.
6588 (gate_tree_if_conversion): Similarly.
6589 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
6590 gimple_build_omp_for.
6591 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
6592 * tree-parloops (create_parallel_loop): Pass kind argument to
6593 gimple_build_omp_for.
6594 * tree-pretty-print.c (dump_omp_clause): Add cases for
6595 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
6596 OMP_CLAUSE__SIMDUID_.
6597 (dump_generic_node): Handle OMP_SIMD.
6598 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
6599 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
6600 unroll OMP_SIMD loops here.
6601 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
6602 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
6604 (vect_analyze_data_refs): Handle simd loops.
6605 * tree-vect-loop.c (vectorizable_live_operation): Handle
6607 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
6608 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
6609 (vectorizable_load): Same.
6610 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
6611 (struct simduid_to_vf): New.
6612 (simduid_to_vf::hash): New.
6613 (simduid_to-vf::equal): New.
6614 (struct simd_array_to_simduid): New.
6615 (simd_array_to_simduid::hash): New.
6616 (simd_array_to_simduid::equal): New.
6617 (adjust_simduid_builtins): New.
6618 (struct note_simd_array_uses_struct): New.
6619 (note_simd_array_uses_cb): New.
6620 (note_simd_array_uses): New.
6621 (vectorize_loops): Handle simd hints and adjust simd builtins
6623 * tree-vectorizer.h (struct _stmt_vec_info): Add
6624 simd_lane_access_p field.
6625 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
6626 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
6627 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
6628 (omp_clause_code_name): Same.
6629 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
6630 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
6631 * tree.def (OMP_SIMD): New entry.
6632 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
6633 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
6634 (OMP_CLAUSE_DECL): Adjust range for new clauses.
6635 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
6636 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
6637 (OMP_CLAUSE_LINEAR_STEP): New.
6638 (OMP_CLAUSE_SAFELEN_EXPR): New.
6639 (OMP_CLAUSE__SIMDUID__DECL): New.
6640 (find_omp_clause): New prototype.
6642 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
6644 * config/i386/driver-i386.c (host_detect_local_cpu): Update
6645 Haswell processor detection.
6647 2013-08-27 Christian Widmer <shadow@umbrox.de>
6650 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
6651 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
6652 AVX2 capable processors.
6654 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
6656 * config/aarch64/arm_neon.h: Replace all inline asm implementations
6657 of vget_low_* with implementations in terms of other intrinsics.
6659 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
6662 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
6663 values to -1, 0 and 1.
6665 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
6667 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
6668 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
6669 (<optab><mode>3_insn): Remove.
6670 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
6671 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
6672 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
6673 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
6674 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
6675 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
6676 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
6677 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
6678 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
6679 (ror<mode>3_insn): Likewise.
6680 * config/aarch64/predicates.md (aarch64_simd_register): New.
6682 2013-08-27 Richard Biener <rguenther@suse.de>
6684 PR tree-optimization/57521
6685 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
6686 one edge is non-critical.
6687 (find_phi_replacement_condition): Make sure to use a non-critical
6688 edge. Cleanup and remove old bug workarounds.
6689 (bb_postdominates_preds): Remove.
6690 (if_convertible_loop_p_1): Do not compute post-dominators.
6691 (combine_blocks): Do not free post-dominators.
6692 (main_tree_if_conversion): Likewise.
6693 (pass_data_if_conversion): Add TODO_verify_ssa.
6695 2013-08-27 DJ Delorie <dj@redhat.com>
6697 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
6699 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
6701 * function.c (assign_parm_find_data_types): Set passed_mode and
6702 nominal_mode to the TYPE_MODE of nominal_type for the built
6703 pointer type in case of the struct-pass-by-reference.
6705 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6707 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
6708 (UINT16_TYPE): Change default to "unsigned int".
6710 * config/avr/avr.opt (mfract-convert-truncate): New option.
6711 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
6712 is set, round negative fractional integers according to n1169
6713 when converting to integer types.
6715 2013-08-26 Jan Hubicka <jh@suse.cz>
6717 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
6718 methods can not be called indirectly when their address is not taken.
6720 2013-08-26 Jan Hubicka <jh@suse.cz>
6722 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
6725 2013-08-26 Jan Hubicka <jh@suse.cz>
6727 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
6730 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6732 * reload.c (find_valid_class): Allow classes that do not include
6733 FIRST_PSEUDO_REGISTER - 1.
6735 2013-08-26 Jan Hubicka <jh@suse.cz>
6737 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
6738 fix edge count/frequency when speculation failed; fix type check
6739 for the direct call.
6741 2013-08-26 Jan Hubicka <jh@suse.cz>
6743 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
6745 2013-08-26 Jan Hubicka <jh@suse.cz>
6747 * ipa-inline-transform.c (inline_transform): Be ready for basic block
6748 to be changed by edge redirection.
6750 2013-08-26 Jan Hubicka <jh@suse.cz>
6752 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
6753 formating; add sanity check.
6754 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
6755 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
6756 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
6757 (resolve_noninline_speculation): Update callee keys, too.
6759 2013-08-26 Jan Hubicka <jh@suse.cz>
6761 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
6762 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
6764 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
6766 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
6769 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
6771 * config/i386/i386.c (ix86_debug_options): Remove prototype.
6772 (x86_64_elf_select_section): Ditto.
6773 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
6775 (ix86_pass_by_reference): Ditto.
6776 (output_set_got): Ditto.
6777 (ix86_unary_operator_ok): Ditto.
6778 (ix86_expand_builtin): Ditto.
6780 2013-08-23 Jan Hubicka <jh@suse.cz>
6782 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
6784 2013-08-23 Jan Hubicka <jh@suse.cz>
6786 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
6787 (tree_decl_with_vis): Add FINAL field.
6789 2013-08-23 Jeff Law <law@redhat.com>
6791 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
6792 the debugging dump when the expression is fully redundant.
6794 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
6796 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
6797 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
6798 * pretty-print.c (pp_formatted_text_data): Likewise.
6799 (pp_write_text_to_stream): Likewise.
6800 (pp_write_text_as_dot_label_to_stream): Likewise.
6801 (pp_append_r): Likewise.
6802 (pp_format): Likewise.
6803 (pp_flush): Likewise.
6804 (pp_clear_output_area): Likewise.
6805 (pp_append_text): Likewise.
6806 (pp_formatted_text): Likewise.
6807 (pp_remaining_character_count_for_line): Likewise.
6808 (pp_newline): Likewise.
6809 (pp_character): Likewise.
6810 (output_buffer::~output_buffer): Define.
6811 (pretty_printer::~pretty_printer): Destruct output buffer.
6812 * pretty-print.h (output_buffer::~output_buffer): Declare.
6813 (pretty_printer::~pretty_printer): Declare virtual.
6815 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
6818 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
6819 HOST_WIDE_INT_M1U): New macros.
6820 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
6821 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
6822 unsigned -1 for lshift.
6823 * cse.c (cse_insn): Likewise.
6824 * double-int.c (rshift_double, lshift_double): Likewise.
6825 * builtins.c (fold_builtin_bitop): Likewise.
6826 * combine.c (force_to_mode): Likewise.
6827 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
6828 * simplify-rtx.c (simplify_const_unary_operation,
6829 simplify_const_binary_operation): Likewise.
6830 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
6831 check_va_list_escapes): Likewise.
6832 * rtlanal.c (nonzero_bits1): Likewise.
6833 * expmed.c (expand_smod_pow2): Likewise.
6834 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
6836 2013-08-23 Jan Hubicka <jh@suse.cz>
6838 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
6839 as having address taken.
6841 2013-08-23 Jan Hubicka <jh@suse.cz>
6843 * ipa-utils.h (method_class_type): Declare.
6844 * ipa-devirt.c (method_class_type): Export.
6846 * cgraphunit.c (analyze_functions): Do basic devirtualization;
6847 do not walk base classes of anonymous types.
6849 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
6851 PR rtl-optimization/58220
6853 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
6854 handle SEQUENCE insns properly.
6856 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
6858 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
6860 (pp_newline_and_indent): Likewise.
6861 (pp_separate_with): Likewise.
6862 * pretty-print.c (pp_newline_and_flush): Define.
6863 (pp_newline_and_indent): Likewise.
6864 (pp_separate_with): Likewise.
6866 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6869 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
6870 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
6872 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
6874 * config/i386/predicates.md (ext_sse_reg_operand): New.
6875 * config/i386/i386.md (*movti_internal): Use
6876 predicate to determine if EVEX is needed.
6877 (*movsi_internal): Ditto.
6878 (*movdf_internal): Ditto.
6879 (*movsf_internal): Ditto.
6880 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
6882 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6884 PR tree-optimization/58209
6885 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
6886 (find_tail_calls): Give up for pointer result types if m is non-NULL.
6887 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
6888 emit POINTER_PLUS_EXPR.
6889 (create_tailcall_accumulator): For pointer result type accumulate in
6892 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
6894 * configure.ac: Add backslashes missing from the last change.
6895 * configure: Regenerate.
6897 2013-08-22 Jan Hubicka <jh@susue.cz>
6899 * ipa.c (function_and_variable_visibility): First remember function
6900 was global and then make it local.
6902 2013-08-22 Julian Brown <julian@codesourcery.com>
6904 * configure.ac: Add aarch64 to list of arches which use "nop" in
6906 * configure: Regenerate.
6908 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6910 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
6911 gnu_libc_has_function.
6912 * config/s390/tpf.h: Likewise.
6914 2013-08-22 Jan Hubicka <jh@susue.cz>
6916 * timevar.c (validate_phases): Add cast.
6918 2013-08-22 Jan Hubicka <jh@susue.cz>
6920 * timevar.c (validate_phases): Use size_t for memory.
6921 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
6923 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
6925 * pretty-print.h (output_buffer::output_buffer): Declare.
6926 (pretty_printer::pretty_printer): Likewise.
6927 (pp_construct): Remove.
6928 * pretty-print.c (output_buffer::output_buffer): Define.
6929 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
6930 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
6931 (print_gimple_expr): Likewise.
6932 (print_gimple_seq): Likewise.
6933 (gimple_dump_bb): Likewise.
6934 * sched-vis.c (dump_value_slim): Likewise.
6935 (dump_insn_slim): Likewise.
6936 (dump_rtl_slim): Likewise.
6937 (str_pattern_slim): Likewise.
6938 * tree-mudflap.c (mf_varname_tree): Likewise.
6939 * graph.c (print_graph_cfg): Likewise.
6940 (start_graph_dump): Likewise.
6941 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
6943 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
6944 pretty printer initialization.
6945 * coretypes.h (diagnostic_context): Remove superflous type alias
6947 (pretty_printer): Likewise. Declare directly as a class.
6948 (pretty_print_info): Remove declaration as class.
6949 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
6950 and pp_clear_output_area.
6951 (asan_add_global): Likewise.
6953 2013-08-22 Jan Hubicka <jh@suse.cz>
6955 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
6956 * ipa-utils.h (update_type_inheritance_graph): Declare.
6957 (possible_polymorphic_call_target_p): Declare.
6958 (possible_polymorphic_call_target_p): New.
6959 * ipa-devirt.c: Update toplevel comments.
6960 (cached_polymorphic_call_targets): Move up.
6961 (odr_type_d): Move ID down.
6962 (polymorphic_type_binfo_p): Update comment.
6963 (odr_hasher::remove): Likewise;
6964 (get_odr_type): Set anonymous_namespace.
6965 (dump_odr_type): Dump it.
6966 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
6967 (maybe_record_node): Record node in cached_polymorphic_call_targets.
6968 (record_binfo): Add comment.
6969 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
6971 (devirt_node_removal_hook): Do not iCE when cache is freed.
6972 (possible_polymorphic_call_target_p): New predicate.
6973 (update_type_inheritance_graph): New function.
6975 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
6976 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6977 Sergey Lega <sergey.s.lega@intel.com>
6978 Anna Tikhonova <anna.tikhonova@intel.com>
6979 Ilya Tocar <ilya.tocar@intel.com>
6980 Andrey Turetskiy <andrey.turetskiy@intel.com>
6981 Ilya Verbin <ilya.verbin@intel.com>
6982 Kirill Yukhin <kirill.yukhin@intel.com>
6983 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6985 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
6986 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
6987 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
6988 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
6989 (OPTION_MASK_ISA_AVX2_UNSET): Update.
6990 (OPTION_MASK_ISA_AVX512F_UNSET): New.
6991 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
6992 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
6993 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
6994 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
6995 OPT_mavx512pf, OPT_mavx512er cases.
6996 * config/i386/constraints.md (v): New constraint.
6997 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
6998 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
6999 (bit_AVX512CD): New.
7000 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7001 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
7002 * config/i386/i386-c.c (ix86_target_macros_internal):
7003 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
7005 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
7006 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
7007 * config/i386/i386.c (regclass_map, dbx_register_map)
7008 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
7009 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
7010 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
7012 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
7013 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
7014 -mavx512pf options. Fix formatting.
7015 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
7016 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
7017 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
7018 -mavx512cd, -mavx512pf options.
7019 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
7020 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
7021 (ix86_preferred_output_reload_class): Replace SSE_REGS with
7023 (ix86_hard_regno_mode_ok): Support 512-bit registers.
7024 (ix86_set_reg_reg_cost): Ditto.
7025 (x86_order_regs_for_local_alloc): Ditto.
7026 (MAX_VECT_LEN): Extend to 64-byte.
7027 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
7028 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
7029 (TARGET_AVX512ER, TARGET_AVX512CD): New.
7030 (BIGGEST_ALIGNMENT): Extend to 512-bits.
7031 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
7032 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
7033 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
7034 (SSE_REG_MODE_P): Support new modes.
7035 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
7036 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
7037 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
7038 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
7039 (REG_CLASS_CONTENTS): Add new registers.
7040 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
7041 (EXT_REX_SSE_REGNO_P): New.
7042 (HI_REGISTER_NAMES): Add new registers.
7043 * config/i386/i386.md: Define constants for new registers.
7044 (mode): Add new 512-bit modes.
7045 (prefix): Support evex prefix.
7046 (isa): Support avx512f, noavx512f, fma_avx512f.
7047 (ssemodesuffix): Add new 512-bit modes.
7049 (*movxi_internal_avx512f): Ditto.
7050 (*movdi_internal): Replace constraint "x" with the new constraint "v".
7052 (*movsi_internal): Likewise.
7053 (*movdf_internal): Likewise.
7054 (*movsf_internal): Likewise.
7055 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
7056 (<code><mode>3): Likewise.
7057 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
7059 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
7060 with the new constraint "v".
7061 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
7063 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
7064 with the new constraint "v".
7065 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
7066 (<sse2>_storedqu<avxsizesuffix>): Likewise.
7067 (*<plusminus_insn><mode>3): Likewise.
7068 (<sse>_vm<plusminus_insn><mode>3): Likewise.
7069 (*mul<mode>3): Likewise.
7070 (<sse>_vmmul<mode>3): Likewise.
7071 (<sse>_div<mode>3): Likewise.
7072 (<sse>_vmdiv<mode>3): Likewise.
7073 (<sse>_sqrt<mode>2): Likewise.
7074 (<sse>_vmsqrt<mode>2): Likewise.
7075 (*<code><mode>3_finite): Likewise.
7076 (*<code><mode>3) <smaxmin>: Likewise.
7077 (<sse>_vm<code><mode>3): Likewise.
7078 (*<code><mode>3) <any_logic>: Likewise.
7079 (*fma_fmadd_<mode>): Likewise.
7080 (*fma_fmsub_<mode>): Likewise.
7081 (*fma_fnmadd_<mode>): Likewise.
7082 (*fma_fnmsub_<mode>): Likewise.
7083 (*fma_fmaddsub_<mode>): Likewise.
7084 (*fma_fmsubadd_<mode>): Likewise.
7085 (*fmai_fmadd_<mode>): Likewise.
7086 (*fmai_fmsub_<mode>): Likewise.
7087 (*fmai_fnmadd_<mode>): Likewise.
7088 (*fmai_fnmsub_<mode>): Likewise.
7089 (sse_cvtsi2ss): Likewise.
7090 (sse_cvtsi2ssq): Likewise.
7091 (sse_cvtss2si): Likewise.
7092 (sse_cvtss2si_2): Likewise.
7093 (sse_cvtss2siq): Likewise.
7094 (sse_cvtss2siq_2): Likewise.
7095 (sse_cvttss2si): Likewise.
7096 (sse_cvtss2siq_2): Likewise.
7097 (float<sseintvecmodelower><mode>2): Likewise.
7098 (sse2_cvtsd2si_2): Likewise.
7099 (sse2_cvtsd2siq_2): Likewise.
7100 (*<plusminus_insn><mode>3): Likewise.
7101 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
7102 (*<sse4_1_avx2>_mul<mode>3): Likewise.
7103 (ashr<mode>3): Likewise.
7104 (<shift_insn><mode>3): Likewise.
7105 (avx2_<code><mode>3): Likewise.
7106 (*avx2_<code><mode>3): Likewise.
7107 (*andnot<mode>3): Likewise.
7108 (*<code><mode>3) <any_logic>: Likewise.
7109 (abs<mode>2): Likewise.
7110 (avx2_permvar<mode>): Likewise.
7111 (avx2_perm<mode>_1): Likewise.
7112 (*avx_vpermilp<mode>): Likewise.
7113 (avx_vpermilvar<mode>3): Likewise.
7114 (avx2_ashrv<mode>): Likewise.
7115 (avx2_<shift_insn>v<mode>): Likewise.
7116 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
7118 * doc/rtl.texi: Document XImode.
7120 2013-08-21 Jeff Law <law@redhat.com>
7122 * tree-flow.h (register_jump_thread): Pass vector of edges
7123 instead of each important edge.
7124 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
7125 thread path into a vector and pass that to register_jump_thread.
7126 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
7127 passed in edge vector to the current 3-edge form.
7130 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
7132 * dce.c (fini_dce): Call df_analyze again just in case
7133 delete_unmarked_insns removed anything.
7135 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
7137 * reload.h (struct reg_equivs): Rename to ..
7138 (struct reg_equivs_s): .. this.
7140 2013-08-20 Martin Liska <marxin.liska@gmail.com>
7142 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
7144 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7146 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
7148 2013-08-21 Jeff Law <law@redhat.com>
7150 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
7151 simplify assignments too. If the RHS collapses to a singleton
7152 range, then return the value for the range.
7154 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
7156 * config/i386/sse.md (V16): Rename to...
7158 (mov<mode>): Update iterator name.
7159 (*mov<mode>_internal): Ditto.
7160 (push<mode>1): Ditto.
7161 (movmisalign<mode>): Ditto.
7163 2013-08-20 Jan Hubicka <jh@suse.cz>
7166 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
7167 entry for direct edges.
7168 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
7170 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7172 Revert my last two changes, r201865 and r201864:
7175 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7177 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
7178 instances can own GC refs.
7180 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
7181 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
7182 (gcc::context::gt_pch_nx): Likewise.
7183 (gcc::context::gt_pch_nx): Likewise.
7184 * ggc.h (gt_ggc_mx <T>): New.
7185 (gt_pch_nx_with_op <T>): New.
7186 (gt_pch_nx <T>): New.
7187 * passes.c (opt_pass::gt_ggc_mx): New.
7188 (opt_pass::gt_pch_nx): New.
7189 (opt_pass::gt_pch_nx_with_op): New.
7190 (pass_manager::gt_ggc_mx): New.
7191 (pass_manager::gt_pch_nx): New.
7192 (pass_manager::gt_pch_nx_with_op): New.
7193 (pass_manager::operator new): Use
7194 ggc_internal_cleared_alloc_stat rather than xcalloc.
7195 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
7196 (pass_manager::gt_ggc_mx): New.
7197 (pass_manager::gt_pch_nx): New.
7198 (pass_manager::gt_pch_nx_with_op): New.
7199 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
7200 (opt_pass::operator new): New.
7201 (opt_pass::gt_ggc_mx): New.
7202 (opt_pass::gt_pch_nx): New.
7203 (opt_pass::gt_pch_nx_with_op): New.
7206 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7208 * Makefile.in (GTFILES): Add context.h.
7209 * context.c (gcc::context::operator new): New.
7210 (gcc::context::gt_ggc_mx): New.
7211 (gcc::context::gt_pch_nx): New.
7212 (gcc::context::gt_pch_nx): New.
7213 * context.h (gcc::context): Add GTY((user)) marking.
7214 (gcc::context::operator new): New.
7215 (gcc::context::gt_ggc_mx): New.
7216 (gcc::context::gt_pch_nx): New.
7217 (gcc::context::gt_pch_nx): New.
7218 (g): Add GTY marking.
7219 (gt_ggc_mx (gcc::context *)): New.
7220 (gt_pch_nx (gcc::context *)): New.
7221 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
7222 void *cookie)): New.
7223 * gengtype.c (open_base_files) <ifiles>: Add context.h.
7225 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
7227 * dce.c (fini_dce): Call df_analyze again just in case
7228 delete_unmarked_insns removed anything.
7230 2013-08-20 Teresa Johnson <tejohnson@google.com>
7232 PR rtl-optimizations/57451
7233 * final.c (reemit_insn_block_notes): Prevent lexical blocks
7234 from crossing split section boundaries.
7236 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7238 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
7239 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
7240 with MULTLIB_DEFAULTS.
7242 2013-08-20 Nick Clifton <nickc@redhat.com>
7244 * target.def (narrow_volatile_bitfield): Note that the default
7245 value is false, not !TARGET_STRICT_ALIGN.
7246 * doc/tm.texi: Regenerate.
7248 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
7250 Fix LIB_SPEC for systems without libpthread.
7252 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
7253 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
7255 * config/i386/linux-common.h: Likewise.
7256 * config/mips/linux-common.h: Likewise.
7258 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
7260 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
7263 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7265 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
7266 instances can own GC refs.
7268 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
7269 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
7270 (gcc::context::gt_pch_nx): Likewise.
7271 (gcc::context::gt_pch_nx): Likewise.
7272 * ggc.h (gt_ggc_mx <T>): New.
7273 (gt_pch_nx_with_op <T>): New.
7274 (gt_pch_nx <T>): New.
7275 * passes.c (opt_pass::gt_ggc_mx): New.
7276 (opt_pass::gt_pch_nx): New.
7277 (opt_pass::gt_pch_nx_with_op): New.
7278 (pass_manager::gt_ggc_mx): New.
7279 (pass_manager::gt_pch_nx): New.
7280 (pass_manager::gt_pch_nx_with_op): New.
7281 (pass_manager::operator new): Use
7282 ggc_internal_cleared_alloc_stat rather than xcalloc.
7283 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
7284 (pass_manager::gt_ggc_mx): New.
7285 (pass_manager::gt_pch_nx): New.
7286 (pass_manager::gt_pch_nx_with_op): New.
7287 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
7288 (opt_pass::operator new): New.
7289 (opt_pass::gt_ggc_mx): New.
7290 (opt_pass::gt_pch_nx): New.
7291 (opt_pass::gt_pch_nx_with_op): New.
7293 2013-08-20 David Malcolm <dmalcolm@redhat.com>
7295 * Makefile.in (GTFILES): Add context.h.
7296 * context.c (gcc::context::operator new): New.
7297 (gcc::context::gt_ggc_mx): New.
7298 (gcc::context::gt_pch_nx): New.
7299 (gcc::context::gt_pch_nx): New.
7300 * context.h (gcc::context): Add GTY((user)) marking.
7301 (gcc::context::operator new): New.
7302 (gcc::context::gt_ggc_mx): New.
7303 (gcc::context::gt_pch_nx): New.
7304 (gcc::context::gt_pch_nx): New.
7305 (g): Add GTY marking.
7306 (gt_ggc_mx (gcc::context *)): New.
7307 (gt_pch_nx (gcc::context *)): New.
7308 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
7309 void *cookie)): New.
7310 * gengtype.c (open_base_files) <ifiles>: Add context.h.
7312 2013-08-20 Alan Modra <amodra@gmail.com>
7315 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
7316 (rs6000_emit_epilogue): Likewise.
7318 2013-08-19 Dehao Chen <dehao@google.com>
7320 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
7322 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
7323 Jakub Jelinek <jakub@redhat.com>
7325 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
7326 (BUILT_IN_FABSD64): Likewise.
7327 (BUILT_IN_FABSD128): Likewise.
7328 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
7329 (fold_builtin_1): Likewise.
7330 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
7331 destination and source operands.
7332 (*abstd2_fpr): Likewise.
7333 (*nabstd2_fpr): Likewise.
7335 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
7337 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
7340 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
7342 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
7345 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
7347 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
7348 * builtins.c (default_libc_has_function): New.
7349 (gnu_libc_has_function): Ditto.
7350 (no_c99_libc_has_function): Ditto.
7351 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
7352 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
7353 (fold_builtin_sincos): Likewise.
7354 (fold_builtin_cexp): Likewise.
7355 * builtins.def (DEF_C94_BUILTIN): Likewise.
7356 (DEF_C99_BUILTIN): Likewise.
7357 (DEF_C99_C90RES_BUILTIN): Likewise.
7358 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
7359 definitions to using this define.
7360 * config/darwin-protos.h (darwin_libc_has_function): New.
7361 * config/darwin.c (darwin_libc_has_function): Ditto.
7362 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
7363 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
7364 * config/darwin.h: Ditto.
7365 * config/elfos.h: Ditto.
7366 * config/freebsd.h: Ditto.
7367 * config/i386/cygming.h: Ditto.
7368 * config/i386/djgpp.h: Ditto.
7369 * config/i386/i386-interix.h: Ditto.
7370 * config/microblaze/microblaze.h: Ditto.
7371 * config/mmix/mmix.h: Ditto.
7372 * config/gnu-user.h: Ditto.
7373 * config/ia64/hpux.h: Ditto.
7374 * config/pa/pa-hpux.h: Ditto.
7375 * config/pdp11/pdp11.h: Ditto.
7376 * config/picochip/picochip.h: Ditto.
7377 * config/linux.h: Ditto.
7378 * config/netbsd.h: Ditto.
7379 * config/openbsd.h: Ditto.
7380 * config/rs6000/aix43.h: Ditto.
7381 * config/rs6000/aix51.h: Ditto.
7382 * config/rs6000/aix52.h: Ditto.
7383 * config/rs6000/aix53.h: Ditto.
7384 * config/rs6000/aix61.h: Ditto.
7385 * config/rs6000/darwin.h: Ditto.
7386 * config/rs6000/linux.h: Ditto.
7387 * config/rs6000/linux64.h: Ditto.
7388 * config/s390/tpf.h: Ditto.
7389 * config/sol2-10.h: Ditto.
7390 * config/sol2.h: Ditto.
7391 * config/vms/vms.h: Ditto.
7392 * config/vxworks.h: Ditto.
7393 * config/linux-android.c (linux_android_libc_has_function):
7394 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
7395 * config/linux-protos.h (linux_android_libc_has_function):
7397 * config/i386/i386.c (ix86_libc_has_function): New.
7398 * config/i386/i386-protos.h
7399 (ix86_libc_has_function): New declaration.
7400 * config/i386/i386.md
7401 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
7402 ("isinf<mode>2): Likewise.
7403 * convert.c (convert_to_integer): Using new target hook
7404 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
7405 TARGET_C99_FUNCTIONS.
7406 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
7407 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
7408 * coretypes.h (function_class): New enum for different
7409 classes of functions.
7410 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
7411 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
7412 (TARGET_HAS_SINCOS): Likewise.
7413 (TARGET_LIBC_HAS_FUNCTION): New.
7414 * doc/tm.texi: Regenerated.
7415 * targhooks.h (default_libc_has_function): New declaration.
7416 (no_c99_libc_has_function): Ditto.
7417 (gnu_libc_has_function): Ditto.
7418 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
7419 and TARGET_HAS_SINCOS.
7421 2013-08-18 Jan Hubicka <jh@suse.cz>
7423 * Makeifle-in (ipa-devirt.o): New.
7424 (GTFILES): Add ipa-utils.h and ipa-devirt.c
7425 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
7426 (analyze_functions): Look into possible targets of polymorphic call.
7427 * dumpfile.c (dump_files): Add type-inheritance dump.
7428 * dumpfile.h (TDI_inheritance): New.
7429 * ipa-devirt.c: New file.
7430 * ipa-utils.h (odr_type_d): Forward declare.
7431 (odr_type): New type.
7432 (build_type_inheritance_graph): Declare.
7433 (possible_polymorphic_call_targets): Declare and introduce inline
7434 variant when only edge is pased.
7435 (dump_possible_polymorphic_call_targets): Likewise.
7436 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
7437 * tree.c (type_in_anonymous_namespace_p): Break out from ...
7438 (types_same_for_odr): ... here.
7439 * tree.h (type_in_anonymous_namespace_p): Declare.
7441 2013-08-18 Jakub Jelinek <jakub@redhat.com>
7443 PR tree-optimization/58006
7444 * tree-parloops.c (take_address_of): Don't ICE if get_name
7446 (eliminate_local_variables_stmt): Remove clobber stmts.
7448 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
7450 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
7451 error message is issued for an alias to undefined symbol.
7453 2013-08-18 Jan Hubicka <jh@suse.cz>
7455 * cgraph.c (cgraph_create_indirect_edge): Discover
7456 polymorphic calls and record basic info into indirect_info.
7457 * gimple-fold.c (gimple_fold_call): When doing BINFO based
7458 devirtualization, ignore objc function calls.
7459 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
7460 call with no parm index info.
7461 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
7462 * tree.c (virtual_method_call_p): New function.
7463 * tree.h (virtual_method_call_p): Declare.
7465 2013-08-16 Jan Hubicka <jh@suse.cz>
7468 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
7470 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
7472 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
7475 2013-08-16 David Malcolm <dmalcolm@redhat.com>
7477 * gengtype.c (type_for_name): Add special-case support for
7478 locating types within the "gcc::" namespace.
7479 (open_base_files): Emit a "using namespace gcc" directive.
7481 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7484 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
7485 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
7487 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
7488 array instead of each individual operand as a separate argument.
7489 (emit_fusion_gpr_load): Likewise.
7490 (expand_fusion_gpr_load): Add new function declaration.
7492 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
7493 signature to have the operands passed as an array, instead of as
7494 separate arguments. Allow ZERO_EXTEND to be in the memory
7495 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
7496 depend on the register live/dead flags when peepholes are run.
7497 (expand_fusion_gpr_load): New function to be called from the
7498 peephole2 pass, to change the register that addis sets to be the
7500 (emit_fusion_gpr_load): Change the calling signature to have the
7501 operands passed as an array, instead of as separate arguments.
7502 Allow ZERO_EXTEND to be in the memory address, and also
7503 SIGN_EXTEND if -mpower8-fusion-sign.
7505 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
7507 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
7508 adjust the register addis loads up in the peephole2 pass. Do not
7509 depend on the register live/dead state when the peephole pass is done.
7511 2013-08-16 David Malcolm <dmalcolm@redhat.com>
7513 * gengtype.c (create_user_defined_type): Ensure that the kind
7514 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
7515 declaration is seen before the GTY((user)) marking.
7517 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
7520 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
7522 2013-08-16 Jan Hubicka <jh@suse.cz>
7524 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
7525 arugment expected_type.
7526 (gimple_fold_call): Use it.
7527 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
7528 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
7529 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
7530 (try_make_edge_direct_virtual_call): Likewise.
7531 * tree.c (obj_type_ref_class): New.
7532 * tree.h (obj_type_ref_class): Use it.
7534 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
7536 * sched-vis.c (rtl_slim_pp_initialized): Remove.
7537 (rtl_slim_pp): Likewise.
7538 (init_rtl_slim_pretty_print): Likewise.
7539 (dump_value_slim): Don't call it. Use local pretty printer.
7540 (dump_insn_slim): Likewise.
7541 (dump_rtl_slim): Likewise.
7542 (str_pattern_slim): Likewise.
7543 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
7546 2013-08-16 Jakub Jelinek <jakub@redhat.com>
7548 PR tree-optimization/58164
7549 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
7550 walk gimple_goto_dest of GIMPLE_GOTO.
7552 PR tree-optimization/58165
7553 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
7554 bi_call must be the last stmt in a bb, don't split_block, instead
7555 use fallthru edge from it and give up if there is none.
7556 Release conds vector when returning early.
7558 2013-08-14 Xinliang David Li <davidxl@google.com>
7560 * config/i386/i386.c (ix86_option_override_internal):
7561 Remove unused variable and field.
7563 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7566 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
7567 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
7568 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
7569 and Linux, correct BLKmode alignment when 128-bit alignment is
7570 required and compatibility flag is not set.
7571 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
7572 for zero-size arguments when compatibility flag is not set.
7574 2013-08-14 Jakub Jelinek <jakub@redhat.com>
7576 PR tree-optimization/58145
7577 * tree-sra.c (build_ref_for_offset): If prev_base has
7578 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
7580 2013-08-14 Xinliang David Li <davidxl@google.com>
7582 * config/i386/i386.c (ix86_option_override_internal):
7583 Fix uninitialized variable error.
7585 2013-08-14 Xinliang David Li <davidxl@google.com>
7587 * config/i386/i386.opt: Define two new options.
7588 * config/i386/x86-tune.def: Add arch selector field in macros.
7589 * config/i386/i386.h: Adjust macro definition.
7590 * config/i386/i386.c (ix86_option_override_internal):
7592 (parse_mtune_ctrl_str): New function.
7593 (set_ix86_tune_features): New function.
7594 (ix86_function_specific_restore): Call the new helper function.
7596 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
7598 PR rtl-optimization/57662
7599 * sel-sched.c (code_motion_process_successors): When the current insn
7600 is removed after the recursive traversal, break from the loop.
7601 Add comments and debug printouts.
7603 2013-08-14 Jakub Jelinek <jakub@redhat.com>
7604 Alexandre Oliva <aoliva@redhat.com>
7607 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
7608 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
7609 there also UNSPEC_PLTOFF.
7611 2013-08-14 Marek Polacek <polacek@redhat.com>
7613 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
7614 NUM_CONDITIONS bit positions.
7616 2013-08-13 Cary Coutant <ccoutant@google.com>
7618 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
7619 (attr_checksum): Hash vector contents instead of pointer.
7620 (attr_checksum_ordered): Likewise.
7622 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
7624 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
7625 when Pmode != word_mode. Add length_address attribute.
7626 (sse3_monitor_<mode>): Merge from sse3_monitor and
7627 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
7628 Pmode != word_mode. Update insn length attribute.
7629 * config/i386/i386.c (ix86_option_override_internal): Update
7630 ix86_gen_monitor selection for merged sse3_monitor insn.
7632 2013-08-13 Julian Brown <julian@codesourcery.com>
7634 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
7635 perform invalid legitimization on greater-than-word-size modes for
7638 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
7640 * ira.c (setup_class_translate_array): Use aclass instead of cl
7641 for classes not fully covered by allocno classes.
7643 2013-08-13 Jakub Jelinek <jakub@redhat.com>
7645 PR tree-optimization/57661
7646 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
7647 * tree-inline.c (tree_function_versioning): Initialize it.
7648 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
7649 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
7650 that is not being copied.
7653 * asan.c (instrument_strlen_call): Fix typo in comment.
7654 Use char * type even for the lhs of POINTER_PLUS_EXPR.
7656 2013-08-13 Steve Ellcey <sellcey@mips.com>
7658 * config/mips/mips.md (prefetch): Use lw instead of ld on
7659 loongson in 32bit mode.
7661 2013-08-13 Nick Clifton <nickc@redhat.com>
7663 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
7665 2013-08-13 Jan Hubicka <jh@suse.cz>
7667 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
7668 introduced edge; fix typo in sanity check.
7669 (cgraph_resolve_speculation): Export; improve diagnostic.
7670 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
7671 speculation at type mismatch.
7672 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
7673 (cgraph_resolve_speculation): Declare.
7674 (symtab_can_be_discarded): New function.
7675 * value-prof.c (gimple_ic_transform): Remove actual transform code.
7676 * ipa-inline-transform.c (speculation_removed): New global var.
7677 (clone_inlined_nodes): See if speculation can be removed.
7678 (inline_call): If speculations was removed, we growths may not match.
7679 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
7680 (speculation_useful_p): New function.
7681 (resolve_noninline_speculation): New function.
7682 (inline_small_functions): Resolve useless speculations.
7683 * ipa-inline.h (speculation_useful_p): Declare
7684 * ipa.c (can_replace_by_local_alias): Simplify.
7685 (ipa_profile): Produce speculative calls in non-lto, too;
7686 add simple cost model; produce local aliases.
7688 2013-08-13 David Malcolm <dmalcolm@redhat.com>
7690 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
7693 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
7695 * config/i386/i386.c (ix86_function_versions): Use error + inform.
7697 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
7699 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
7700 iterator instead of X87MODEF.
7702 2013-08-12 Perez Read <netfirewall@gmail.com>
7705 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
7706 operand 0 for intel asm alternative.
7707 (*movabs<mode>_2): Ditto for operand 1.
7709 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
7711 * config/aarch64/arm_none.h
7712 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
7714 2013-08-12 Nick Clifton <nickc@redhat.com>
7716 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
7718 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
7720 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
7721 expand for QI/HImode operand to produce more effictive code for
7722 unsigned char(short) --> float(double) conversion.
7724 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
7726 * doc/invoke.texi: Mention that -ftls-model does not force the final
7729 2013-08-12 Marek Polacek <polacek@redhat.com>
7730 Marc Glisse <marc.glisse@inria.fr>
7732 PR tree-optimization/57980
7733 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
7734 when creating -1 constant.
7736 2013-08-10 Jan Hubicka <jh@suse.cz>
7738 Workaround binutils PR14342.
7739 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
7740 (gimple_init_edge_profiler): Likewise.
7741 (gimple_gen_ic_func_profiler): Likewise.
7743 2013-08-09 Jan Hubicka <jh@suse.cz>
7745 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
7747 2013-08-09 Xinliang David Li <davidxl@google.com>
7749 * config/i386/stringop.def: New file.
7750 * config/i386/stringop.opt: New file.
7751 * config/i386/i386-opts.h: Include stringopt.def.
7752 * config/i386/i386.opt: Include stringopt.opt.
7753 * config/i386/i386.c (ix86_option_override_internal):
7754 Override default size based stringop inline strategies with options.
7755 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
7758 2013-08-09 Jan Hubicka <jh@suse.cz>
7760 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
7762 2013-08-09 Jan Hubicka <jh@suse.cz>
7764 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
7766 (dump_cgraph_node): Dump profile-id.
7767 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
7768 and common_target_probability.
7769 * lto-cgraph.c (lto_output_edge): Stream common targets.
7770 (lto_output_node): Stream profile ids.
7771 (input_node): Stream profile ids.
7772 (input_edge): Stream common targets.
7773 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
7774 * ipa.c: Include value-prof.h
7775 (ipa_profile_generate_summary): Turn indirect call statement histograms
7776 into common targets.
7777 (ipa_profile): Turn common targets into speculative edges.
7779 2013-08-09 Jan Hubicka <jh@suse.cz>
7781 * cgraph.h (cgraph_node): Add profile_id.
7782 * value-prof.c (cgraph_node_map): Turn into pointer_map.
7783 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
7784 (del_node_map): Update.
7785 (find_func_by_funcdef_no): Replace by ...
7786 (find_func_by_profile_id): ... this one.
7787 (gimple_ic_transform): Do not remove useful histograms when
7788 speculation is not done; dump info when indirect call removal
7790 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
7791 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
7792 (__gcov_indirect_call_profiler_v2): .. this one.
7793 * profile.h (init_node_map): Update.
7794 * coverage.c (coverage_compute_profile_id): New function.
7795 * coverage.h (coverage_compute_profile_id): Declare.
7796 * tree-profile.c (init_ic_make_global_vars): Make
7797 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
7798 (gimple_init_edge_profiler): Update prototype of
7799 __gcov_indirect_call_profiler.
7800 (gimple_gen_ic_func_profiler): Simplify.
7801 (tree_profiling): Use init_node_map
7803 2013-08-09 Jan Hubicka <jh@suse.cz>
7805 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
7806 non-speculative refs.
7807 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
7808 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
7809 (cgraph_set_call_stmt): Likewise.
7810 (cgraph_create_edge_1): Fix release checking compilatoin;
7812 (cgraph_free_edge): Free indirect info.
7813 (cgraph_turn_edge_to_speculative): New function.
7814 (cgraph_speculative_call_info): New function.
7815 (cgraph_make_edge_direct): Return direct edge; handle speculation.
7816 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
7817 (dump_cgraph_node): Dump speculation.
7818 (verify_edge_count_and_frequency): Accept speculative edges.
7819 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
7820 (verify_cgraph_node): Handle speculation.
7821 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
7822 (cgraph_set_call_stmt): Update prototype.
7823 (cgraph_make_edge_direct): Update prototype.
7824 (cgraph_speculative_call_info): Declare.
7825 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
7826 to change; update call of ipa_find_references.
7827 * ipa-ref.c (ipa_record_reference): Fix return value; clear
7828 lto_stmt_uid and speculative flags.
7829 (ipa_dump_references): Dump speculation.
7830 (ipa_clone_references): Clone speculative flag.
7831 (ipa_clone_referring): Likewise.
7832 (ipa_clone_ref): New function.
7833 (ipa_find_reference): Look into lto_stmt_uids
7834 (ipa_clear_stmts_in_references): Do not clear speculative calls.
7835 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
7836 (ipa_find_reference): Update declaration.
7837 (ipa_clone_ref): Declare.
7838 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
7839 stream speculative flag.
7840 (lto_output_ref): Stream statements uids and speculation.
7841 (input_ref): Likewise.
7842 (input_edge): Stream speuclation.
7843 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
7844 (cgraph_set_call_stmt_including_clones): Handle speculation.
7845 * ipa-inline.c (heap_edge_removal_hook): New function.
7846 (inline_small_functions): Register it.
7847 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
7848 also initialize refs.
7849 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
7851 (try_make_edge_direct_simple_call): Likewise.
7852 (try_make_edge_direct_simple_call): Likewise.
7853 (update_indirect_edges_after_inlining): Likewise.
7854 (remove_described_reference): Look proper lto_stmt_uid.
7855 (propagate_controlled_uses): Likewise.
7856 (propagate_controlled_uses): Liekwise.
7857 * tree-inline.c (copy_bb): Copy speculative edges.
7858 (redirect_all_calls): New function.
7859 (copy_cfg_body): Do redirection after loop info is updated.
7860 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
7862 2013-08-09 Jan Hubicka <jh@suse.cz>
7864 * lto-streamer-out.c (output_function): Renumber PHIs.
7865 * lto-streamer-in.c (input_function): Likewise.
7867 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
7869 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
7870 (get_lane_unsigned): Likewise.
7871 (dup_lane_scalar): Likewise.
7872 (get_lane): enable for VALL.
7873 * config/aarch64/aarch64-simd.md
7874 (aarch64_dup_lane_scalar<mode>): Remove.
7875 (aarch64_get_lane_signed<mode>): Likewise.
7876 (aarch64_get_lane_unsigned<mode>): Likewise.
7877 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
7878 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
7879 (aarch64_get_lane<mode>): Enable for all vector modes.
7880 (aarch64_get_lanedi): Remove misleading constraints.
7881 * config/aarch64/arm_neon.h
7882 (__aarch64_vget_lane_any): Define.
7883 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
7884 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
7885 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
7886 * config/aarch64/iterators.md (VDQQH): New.
7890 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
7892 * configure.ac: Add GAS check for LEON instructions on SPARC.
7893 * configure: Regenerate.
7894 * config.in: Likewise.
7895 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
7897 * config/sparc/sparc.opt (LEON, LEON3): New masks.
7898 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
7900 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
7901 (AS_LEON_FLAG): New macro.
7902 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
7903 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
7904 Deal with LEON and LEON3 for the memory model.
7905 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
7906 (atomic_compare_and_swap<mode>_1): Likewise.
7907 (*atomic_compare_and_swap<mode>_1): Likewise.
7909 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7911 * config/arm/neon.md (vcond): Fix floating-point vector
7912 comparisons against 0.
7914 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
7916 * lra-constraints.c (emit_spill_move): Remove assert.
7917 (process_alt_operands): Add more debugging
7918 output. Increase reject for spilling into memory. Decrease
7919 reject for reloading scratch.
7920 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
7922 2013-08-08 Steve Ellcey <sellcey@mips.com>
7924 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
7925 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
7926 micromips incompatible. Add nan2008.
7927 (MULTILIB_DIRNAMES): Add nan2008.
7928 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7929 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
7930 and micromips incompatible. Add nan2008.
7931 (MULTILIB_DIRNAMES): Add nan2008.
7932 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7934 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
7936 PR rtl-optimization/58079
7937 * combine.c (combine_simplify_rtx): Avoid using SUBST if
7938 simplify_comparison has widened a comparison with an integer.
7940 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7942 * config/arm/neon.md (movmisalign<mode>): Disable when we
7943 don't allow unaligned accesses.
7944 (*movmisalign<mode>_neon_store): Likewise.
7945 (*movmisalign<mode>_neon_load): Likewise.
7946 (*movmisalign<mode>_neon_store): Likewise.
7947 (*movmisalign<mode>_neon_load): Likewise.
7949 2013-08-08 Jan Hubicka <jh@suse.cz>
7951 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
7952 (make_pass_rebuild_cgraph_edges): Also clear references.
7953 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
7954 * ipa-inline-transform.c (inline_transform): Remove all references
7956 * cgraphunit.c (expand_function): Remove all references after
7958 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
7959 (ipa_find_reference): Rewrite to iterator.
7960 (remove_stmt_references): Likewise.
7961 (ipa_clear_stmts_in_references): New function.
7962 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
7963 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
7965 * ipa-split.c (split_function): Remove references in split function.
7967 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7970 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
7971 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
7973 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7976 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
7977 suggested mode for the assignment isn't compatible with the
7980 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7983 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
7985 2013-08-07 Xinliang David Li <davidxl@google.com>
7987 * config/i386/i386.opt: New option -mtune-ctrl=.
7988 * config/i386/x86-tune.def: New file.
7989 * config/i386/i386.h: include x86-tune.def.
7990 * config/i386/i386.c (ix86_option_override_internal):
7991 Parsing -mtune-ctrl= option and set tune features.
7993 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
7996 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
7997 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
8000 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
8002 * rtl.h (update_alignments): Declare.
8003 * final.c (grow_label_align): New function extracted from...
8004 (shorten_branches): ...here. Call it.
8005 (update_alignments): New function.
8006 * reorg.c (sibling_labels): New variable.
8007 (get_label_before): Add SIBLING parameter. If it is non-zero, push
8008 the new label along with it onto the sibling_labels vector.
8009 (fill_simple_delay_slots): Adjust call to get_label_before.
8010 (fill_slots_from_thread): Likewise.
8011 (relax_delay_slots): Likewise.
8012 (make_return_insns): Likewise.
8013 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
8015 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
8017 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
8018 document its semantics.
8019 (diagnostic_report_diagnostic): Adjust accordingly.
8021 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8023 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
8024 (sparc_option_override): ...and port to new C++ pass API.
8025 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
8027 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
8029 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
8031 2013-08-06 Caroline Tice <cmtice@google.com>
8033 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
8034 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
8035 * tree-pass.h: Add pass_vtable_verify.
8036 * varasm.c (assemble_variable): Add code to properly set the comdat
8037 section and name for the .vtable_map_vars section.
8038 (assemble_vtyv_preinit_initializer): New function.
8039 (default_sectin_type_flags): Make sure .vtable_map_vars section has
8041 * output.h: Add function decl for assemble_vtv_preinit_initializer.
8042 * vtable-verify.c: New file.
8043 * vtable-verify.h: New file.
8044 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
8045 initialiation levels.
8046 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
8047 * passes.def: Insert pass_vtable_verify.
8048 * aclocal.m4: Reorder includes.
8049 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
8050 -fvtv-counts options.
8051 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
8052 as appropriate, if -fvtable-verify=... is used.
8053 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
8054 -fvtable-verify=... is used.
8055 * Makefile.in (OBJS): Add vtable-verify.o to list.
8056 (vtable-verify.o): Add new build rule.
8057 (GTFILES): Add vtable-verify.c to list.
8058 * common.opt (fvtable-verify=): New flag.
8059 (vtv_priority): Values for fvtable-verify= flag.
8060 (fvtv-counts): New flag.
8061 (fvtv-debug): New flag.
8062 * tree.h (save_vtable_map_decl): New extern function decl.
8064 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8066 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
8067 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
8068 (pass_data_rl78_devirt): ...new pass_data instance and...
8069 (make_pass_rl78_devirt): ...new function.
8070 (rl78_asm_file_start): Port pass registration to new C++ API.
8072 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8074 * coretypes.h (rtl_opt_pass): Add.
8075 (gcc::context): Add.
8076 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
8077 (epiphany_init): Port to new C++ pass API.
8078 (epiphany_optimize_mode_switching): Likewise.
8079 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
8080 (pass_manager::get_pass_mode_switching): New.
8081 (pass_manager::get_pass_peephole2): New.
8082 * mode-switching.c (pass_mode_switching): Add clone method.
8083 * recog.c (pass_peephole2): Add clone method.
8084 (pass_split_all_insns): Add clone method.
8086 2013-08-06 David Malcolm <dmalcolm@redhat.com>
8088 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
8089 (mips_option_override): ...here, porting to new C++ API for passes.
8091 2013-08-06 Jan Hubicka <jh@suse.cz>
8093 * cgraph.c (cgraph_get_body): New function based on lto.c
8095 * cgraph.h (cgraph_get_body): Declare.
8096 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
8098 * cgraphunit.c (expand_function): Get body prior expanding.
8099 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
8100 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
8102 * passes.c (do_per_function_toporder): Get body.
8103 * tree-inline.c (expand_call_inline): Get body prior inlining it.
8104 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
8106 2013-08-06 Martin Jambor <mjambor@suse.cz>
8109 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
8110 is not re-finalized. Rename second parameter to no_collect.
8112 2013-08-06 Martin Jambor <mjambor@suse.cz>
8115 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
8116 MEM_REF has proper alignment information.
8118 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
8121 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
8123 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
8124 rtx (*) (rtx, ...) with insn_gen_fn.
8125 * genoutput.c (output_insn_data): Cast gen_? function pointers to
8126 insn_gen_fn::stored_funcptr. Add initializer braces.
8128 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8130 Rewrite how instances of passes are cloned to remove assumptions
8131 about their sizes (thus allowing pass subclasses to have
8132 additional data fields, albeit non-GC-managed ones at this point).
8134 * passes.c (make_pass_instance): Now that passes have clone
8135 methods, rewrite this function to eliminate XNEW and memcpy
8136 calls that used hardcoded sizes. Since this function no longer
8137 creates pass instances, rename it to...
8138 (add_pass_instance): ...this. Document the old way that passes were
8139 numbered and flagged, and rework this function to continue using it.
8140 (next_pass_1): Add an initial_pass argument for use by
8142 (position_pass): When adding multiple instances of a pass, use
8143 the pass's clone method, rather than relying on the XNEW/memcpy
8144 within the former make_pass_instance (now add_pass_instance).
8145 (pass_manager::pass_manager): When invoking next_pass_1, also supply
8146 the initial instance of the current pass within the pass manager.
8148 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8150 This is the automated part of the conversion of passes from C
8151 structs to C++ classes.
8153 Patch autogenerated by refactor_passes.py from
8154 https://github.com/davidmalcolm/gcc-refactoring-scripts
8155 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
8157 * asan.c (pass_asan): Convert from a global struct to a subclass of
8158 gimple_opt_pass along with...
8159 (pass_data_asan): ...new pass_data instance and...
8160 (make_pass_asan): ...new function.
8161 (pass_asan_O0): Convert from a global struct to a subclass of
8162 gimple_opt_pass along with...
8163 (pass_data_asan_O0): ...new pass_data instance and...
8164 (make_pass_asan_O0): ...new function.
8165 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
8166 subclass of rtl_opt_pass along with...
8167 (pass_data_inc_dec): ...new pass_data instance and...
8168 (make_pass_inc_dec): ...new function.
8169 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
8170 a subclass of rtl_opt_pass along with...
8171 (pass_data_reorder_blocks): ...new pass_data instance and...
8172 (make_pass_reorder_blocks): ...new function.
8173 (pass_duplicate_computed_gotos): Convert from a global struct to a
8174 subclass of rtl_opt_pass along with...
8175 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
8176 (make_pass_duplicate_computed_gotos): ...new function.
8177 (pass_partition_blocks): Convert from a global struct to a subclass of
8178 rtl_opt_pass along with...
8179 (pass_data_partition_blocks): ...new pass_data instance and...
8180 (make_pass_partition_blocks): ...new function.
8181 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
8182 struct to a subclass of rtl_opt_pass along with...
8183 (pass_data_branch_target_load_optimize1): ...new pass_data instance
8185 (make_pass_branch_target_load_optimize1): ...new function.
8186 (pass_branch_target_load_optimize2): Convert from a global struct to a
8187 subclass of rtl_opt_pass along with...
8188 (pass_data_branch_target_load_optimize2): ...new pass_data instance
8190 (make_pass_branch_target_load_optimize2): ...new function.
8191 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
8192 of rtl_opt_pass along with...
8193 (pass_data_jump): ...new pass_data instance and...
8194 (make_pass_jump): ...new function.
8195 (pass_jump2): Convert from a global struct to a subclass of
8196 rtl_opt_pass along with...
8197 (pass_data_jump2): ...new pass_data instance and...
8198 (make_pass_jump2): ...new function.
8199 * cfgexpand.c (pass_expand): Convert from a global struct to a
8200 subclass of rtl_opt_pass along with...
8201 (pass_data_expand): ...new pass_data instance and...
8202 (make_pass_expand): ...new function.
8203 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
8204 of rtl_opt_pass along with...
8205 (pass_data_free_cfg): ...new pass_data instance and...
8206 (make_pass_free_cfg): ...new function.
8207 (pass_into_cfg_layout_mode): Convert from a global struct to a
8208 subclass of rtl_opt_pass along with...
8209 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
8210 (make_pass_into_cfg_layout_mode): ...new function.
8211 (pass_outof_cfg_layout_mode): Convert from a global struct to a
8212 subclass of rtl_opt_pass along with...
8213 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
8214 (make_pass_outof_cfg_layout_mode): ...new function.
8215 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
8216 struct to a subclass of gimple_opt_pass along with...
8217 (pass_data_build_cgraph_edges): ...new pass_data instance and...
8218 (make_pass_build_cgraph_edges): ...new function.
8219 (pass_rebuild_cgraph_edges): Convert from a global struct to a
8220 subclass of gimple_opt_pass along with...
8221 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
8222 (make_pass_rebuild_cgraph_edges): ...new function.
8223 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
8224 subclass of gimple_opt_pass along with...
8225 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
8227 (make_pass_remove_cgraph_callee_edges): ...new function.
8228 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
8229 struct to a subclass of rtl_opt_pass along with...
8230 (pass_data_stack_adjustments): ...new pass_data instance and...
8231 (make_pass_stack_adjustments): ...new function.
8232 * combine.c (pass_combine): Convert from a global struct to a subclass
8233 of rtl_opt_pass along with...
8234 (pass_data_combine): ...new pass_data instance and...
8235 (make_pass_combine): ...new function.
8236 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
8237 global struct to a subclass of rtl_opt_pass along with...
8238 (pass_data_compare_elim_after_reload): ...new pass_data instance
8240 (make_pass_compare_elim_after_reload): ...new function.
8241 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
8242 of rtl_opt_pass along with...
8243 (pass_data_rtl_cprop): ...new pass_data instance and...
8244 (make_pass_rtl_cprop): ...new function.
8245 * cse.c (pass_cse): Convert from a global struct to a subclass of
8246 rtl_opt_pass along with...
8247 (pass_data_cse): ...new pass_data instance and...
8248 (make_pass_cse): ...new function.
8249 (pass_cse2): Convert from a global struct to a subclass of
8250 rtl_opt_pass along with...
8251 (pass_data_cse2): ...new pass_data instance and...
8252 (make_pass_cse2): ...new function.
8253 (pass_cse_after_global_opts): Convert from a global struct to a
8254 subclass of rtl_opt_pass along with...
8255 (pass_data_cse_after_global_opts): ...new pass_data instance and...
8256 (make_pass_cse_after_global_opts): ...new function.
8257 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
8258 of rtl_opt_pass along with...
8259 (pass_data_ud_rtl_dce): ...new pass_data instance and...
8260 (make_pass_ud_rtl_dce): ...new function.
8261 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
8262 rtl_opt_pass along with...
8263 (pass_data_fast_rtl_dce): ...new pass_data instance and...
8264 (make_pass_fast_rtl_dce): ...new function.
8265 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
8266 a subclass of rtl_opt_pass along with...
8267 (pass_data_df_initialize_opt): ...new pass_data instance and...
8268 (make_pass_df_initialize_opt): ...new function.
8269 (pass_df_initialize_no_opt): Convert from a global struct to a
8270 subclass of rtl_opt_pass along with...
8271 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
8272 (make_pass_df_initialize_no_opt): ...new function.
8273 (pass_df_finish): Convert from a global struct to a subclass of
8274 rtl_opt_pass along with...
8275 (pass_data_df_finish): ...new pass_data instance and...
8276 (make_pass_df_finish): ...new function.
8277 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
8278 rtl_opt_pass along with...
8279 (pass_data_rtl_dse1): ...new pass_data instance and...
8280 (make_pass_rtl_dse1): ...new function.
8281 (pass_rtl_dse2): Convert from a global struct to a subclass of
8282 rtl_opt_pass along with...
8283 (pass_data_rtl_dse2): ...new pass_data instance and...
8284 (make_pass_rtl_dse2): ...new function.
8285 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
8286 subclass of rtl_opt_pass along with...
8287 (pass_data_dwarf2_frame): ...new pass_data instance and...
8288 (make_pass_dwarf2_frame): ...new function.
8289 * except.c (pass_set_nothrow_function_flags): Convert from a global
8290 struct to a subclass of rtl_opt_pass along with...
8291 (pass_data_set_nothrow_function_flags): ...new pass_data instance
8293 (make_pass_set_nothrow_function_flags): ...new function.
8294 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
8295 subclass of rtl_opt_pass along with...
8296 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
8298 (make_pass_convert_to_eh_region_ranges): ...new function.
8299 * final.c (pass_compute_alignments): Convert from a global struct to a
8300 subclass of rtl_opt_pass along with...
8301 (pass_data_compute_alignments): ...new pass_data instance and...
8302 (make_pass_compute_alignments): ...new function.
8303 (pass_final): Convert from a global struct to a subclass of
8304 rtl_opt_pass along with...
8305 (pass_data_final): ...new pass_data instance and...
8306 (make_pass_final): ...new function.
8307 (pass_shorten_branches): Convert from a global struct to a subclass of
8308 rtl_opt_pass along with...
8309 (pass_data_shorten_branches): ...new pass_data instance and...
8310 (make_pass_shorten_branches): ...new function.
8311 (pass_clean_state): Convert from a global struct to a subclass of
8312 rtl_opt_pass along with...
8313 (pass_data_clean_state): ...new pass_data instance and...
8314 (make_pass_clean_state): ...new function.
8315 * function.c (pass_instantiate_virtual_regs): Convert from a global
8316 struct to a subclass of rtl_opt_pass along with...
8317 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
8318 (make_pass_instantiate_virtual_regs): ...new function.
8319 (pass_leaf_regs): Convert from a global struct to a subclass of
8320 rtl_opt_pass along with...
8321 (pass_data_leaf_regs): ...new pass_data instance and...
8322 (make_pass_leaf_regs): ...new function.
8323 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
8324 subclass of rtl_opt_pass along with...
8325 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
8327 (make_pass_thread_prologue_and_epilogue): ...new function.
8328 (pass_match_asm_constraints): Convert from a global struct to a
8329 subclass of rtl_opt_pass along with...
8330 (pass_data_match_asm_constraints): ...new pass_data instance and...
8331 (make_pass_match_asm_constraints): ...new function.
8332 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
8333 subclass of rtl_opt_pass along with...
8334 (pass_data_rtl_fwprop): ...new pass_data instance and...
8335 (make_pass_rtl_fwprop): ...new function.
8336 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
8337 rtl_opt_pass along with...
8338 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
8339 (make_pass_rtl_fwprop_addr): ...new function.
8340 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
8341 rtl_opt_pass along with...
8342 (pass_data_rtl_pre): ...new pass_data instance and...
8343 (make_pass_rtl_pre): ...new function.
8344 (pass_rtl_hoist): Convert from a global struct to a subclass of
8345 rtl_opt_pass along with...
8346 (pass_data_rtl_hoist): ...new pass_data instance and...
8347 (make_pass_rtl_hoist): ...new function.
8348 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
8349 subclass of gimple_opt_pass along with...
8350 (pass_data_lower_cf): ...new pass_data instance and...
8351 (make_pass_lower_cf): ...new function.
8352 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
8353 from a global struct to a subclass of gimple_opt_pass along with...
8354 (pass_data_strength_reduction): ...new pass_data instance and...
8355 (make_pass_strength_reduction): ...new function.
8356 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
8357 of rtl_opt_pass along with...
8358 (pass_data_rtl_ifcvt): ...new pass_data instance and...
8359 (make_pass_rtl_ifcvt): ...new function.
8360 (pass_if_after_combine): Convert from a global struct to a subclass of
8361 rtl_opt_pass along with...
8362 (pass_data_if_after_combine): ...new pass_data instance and...
8363 (make_pass_if_after_combine): ...new function.
8364 (pass_if_after_reload): Convert from a global struct to a subclass of
8365 rtl_opt_pass along with...
8366 (pass_data_if_after_reload): ...new pass_data instance and...
8367 (make_pass_if_after_reload): ...new function.
8368 * init-regs.c (pass_initialize_regs): Convert from a global struct to
8369 a subclass of rtl_opt_pass along with...
8370 (pass_data_initialize_regs): ...new pass_data instance and...
8371 (make_pass_initialize_regs): ...new function.
8372 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
8373 of ipa_opt_pass_d along with...
8374 (pass_data_ipa_cp): ...new pass_data instance and...
8375 (make_pass_ipa_cp): ...new function.
8376 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
8377 global struct to a subclass of gimple_opt_pass along with...
8378 (pass_data_inline_parameters): ...new pass_data instance and...
8379 (make_pass_inline_parameters): ...new function.
8380 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
8381 subclass of gimple_opt_pass along with...
8382 (pass_data_early_inline): ...new pass_data instance and...
8383 (make_pass_early_inline): ...new function.
8384 (pass_ipa_inline): Convert from a global struct to a subclass of
8385 ipa_opt_pass_d along with...
8386 (pass_data_ipa_inline): ...new pass_data instance and...
8387 (make_pass_ipa_inline): ...new function.
8388 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
8389 struct to a subclass of gimple_opt_pass along with...
8390 (pass_data_local_pure_const): ...new pass_data instance and...
8391 (make_pass_local_pure_const): ...new function.
8392 (pass_ipa_pure_const): Convert from a global struct to a subclass of
8393 ipa_opt_pass_d along with...
8394 (pass_data_ipa_pure_const): ...new pass_data instance and...
8395 (make_pass_ipa_pure_const): ...new function.
8396 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
8397 to a subclass of ipa_opt_pass_d along with...
8398 (pass_data_ipa_reference): ...new pass_data instance and...
8399 (make_pass_ipa_reference): ...new function.
8400 * ipa-split.c (pass_split_functions): Convert from a global struct to
8401 a subclass of gimple_opt_pass along with...
8402 (pass_data_split_functions): ...new pass_data instance and...
8403 (make_pass_split_functions): ...new function.
8404 (pass_feedback_split_functions): Convert from a global struct to a
8405 subclass of gimple_opt_pass along with...
8406 (pass_data_feedback_split_functions): ...new pass_data instance and...
8407 (make_pass_feedback_split_functions): ...new function.
8408 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
8409 global struct to a subclass of simple_ipa_opt_pass along with...
8410 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
8412 (make_pass_ipa_function_and_variable_visibility): ...new function.
8413 (pass_ipa_free_inline_summary): Convert from a global struct to a
8414 subclass of simple_ipa_opt_pass along with...
8415 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
8416 (make_pass_ipa_free_inline_summary): ...new function.
8417 (pass_ipa_whole_program_visibility): Convert from a global struct to a
8418 subclass of ipa_opt_pass_d along with...
8419 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
8421 (make_pass_ipa_whole_program_visibility): ...new function.
8422 (pass_ipa_profile): Convert from a global struct to a subclass of
8423 ipa_opt_pass_d along with...
8424 (pass_data_ipa_profile): ...new pass_data instance and...
8425 (make_pass_ipa_profile): ...new function.
8426 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
8427 ipa_opt_pass_d along with...
8428 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
8429 (make_pass_ipa_cdtor_merge): ...new function.
8430 * ira.c (pass_ira): Convert from a global struct to a subclass of
8431 rtl_opt_pass along with...
8432 (pass_data_ira): ...new pass_data instance and...
8433 (make_pass_ira): ...new function.
8434 (pass_reload): Convert from a global struct to a subclass of
8435 rtl_opt_pass along with...
8436 (pass_data_reload): ...new pass_data instance and...
8437 (make_pass_reload): ...new function.
8438 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
8439 subclass of rtl_opt_pass along with...
8440 (pass_data_cleanup_barriers): ...new pass_data instance and...
8441 (make_pass_cleanup_barriers): ...new function.
8442 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
8443 of rtl_opt_pass along with...
8444 (pass_data_loop2): ...new pass_data instance and...
8445 (make_pass_loop2): ...new function.
8446 (pass_rtl_loop_init): Convert from a global struct to a subclass of
8447 rtl_opt_pass along with...
8448 (pass_data_rtl_loop_init): ...new pass_data instance and...
8449 (make_pass_rtl_loop_init): ...new function.
8450 (pass_rtl_loop_done): Convert from a global struct to a subclass of
8451 rtl_opt_pass along with...
8452 (pass_data_rtl_loop_done): ...new pass_data instance and...
8453 (make_pass_rtl_loop_done): ...new function.
8454 (pass_rtl_move_loop_invariants): Convert from a global struct to a
8455 subclass of rtl_opt_pass along with...
8456 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
8457 (make_pass_rtl_move_loop_invariants): ...new function.
8458 (pass_rtl_unswitch): Convert from a global struct to a subclass of
8459 rtl_opt_pass along with...
8460 (pass_data_rtl_unswitch): ...new pass_data instance and...
8461 (make_pass_rtl_unswitch): ...new function.
8462 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
8463 subclass of rtl_opt_pass along with...
8464 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
8466 (make_pass_rtl_unroll_and_peel_loops): ...new function.
8467 (pass_rtl_doloop): Convert from a global struct to a subclass of
8468 rtl_opt_pass along with...
8469 (pass_data_rtl_doloop): ...new pass_data instance and...
8470 (make_pass_rtl_doloop): ...new function.
8471 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
8472 a subclass of rtl_opt_pass along with...
8473 (pass_data_lower_subreg): ...new pass_data instance and...
8474 (make_pass_lower_subreg): ...new function.
8475 (pass_lower_subreg2): Convert from a global struct to a subclass of
8476 rtl_opt_pass along with...
8477 (pass_data_lower_subreg2): ...new pass_data instance and...
8478 (make_pass_lower_subreg2): ...new function.
8479 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
8480 struct to a subclass of ipa_opt_pass_d along with...
8481 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
8482 (make_pass_ipa_lto_gimple_out): ...new function.
8483 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
8484 of ipa_opt_pass_d along with...
8485 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
8486 (make_pass_ipa_lto_finish_out): ...new function.
8487 * mode-switching.c (pass_mode_switching): Convert from a global struct
8488 to a subclass of rtl_opt_pass along with...
8489 (pass_data_mode_switching): ...new pass_data instance and...
8490 (make_pass_mode_switching): ...new function.
8491 * modulo-sched.c (pass_sms): Convert from a global struct to a
8492 subclass of rtl_opt_pass along with...
8493 (pass_data_sms): ...new pass_data instance and...
8494 (make_pass_sms): ...new function.
8495 * omp-low.c (pass_expand_omp): Convert from a global struct to a
8496 subclass of gimple_opt_pass along with...
8497 (pass_data_expand_omp): ...new pass_data instance and...
8498 (make_pass_expand_omp): ...new function.
8499 (pass_lower_omp): Convert from a global struct to a subclass of
8500 gimple_opt_pass along with...
8501 (pass_data_lower_omp): ...new pass_data instance and...
8502 (make_pass_lower_omp): ...new function.
8503 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
8504 of gimple_opt_pass along with...
8505 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
8506 (make_pass_diagnose_omp_blocks): ...new function.
8507 * passes.c (pass_early_local_passes): Convert from a global struct to
8508 a subclass of simple_ipa_opt_pass along with...
8509 (pass_data_early_local_passes): ...new pass_data instance and...
8510 (make_pass_early_local_passes): ...new function.
8511 (pass_all_early_optimizations): Convert from a global struct to a
8512 subclass of gimple_opt_pass along with...
8513 (pass_data_all_early_optimizations): ...new pass_data instance and...
8514 (make_pass_all_early_optimizations): ...new function.
8515 (pass_all_optimizations): Convert from a global struct to a subclass
8516 of gimple_opt_pass along with...
8517 (pass_data_all_optimizations): ...new pass_data instance and...
8518 (make_pass_all_optimizations): ...new function.
8519 (pass_all_optimizations_g): Convert from a global struct to a subclass
8520 of gimple_opt_pass along with...
8521 (pass_data_all_optimizations_g): ...new pass_data instance and...
8522 (make_pass_all_optimizations_g): ...new function.
8523 (pass_rest_of_compilation): Convert from a global struct to a subclass
8524 of rtl_opt_pass along with...
8525 (pass_data_rest_of_compilation): ...new pass_data instance and...
8526 (make_pass_rest_of_compilation): ...new function.
8527 (pass_postreload): Convert from a global struct to a subclass of
8528 rtl_opt_pass along with...
8529 (pass_data_postreload): ...new pass_data instance and...
8530 (make_pass_postreload): ...new function.
8531 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
8532 subclass of rtl_opt_pass along with...
8533 (pass_data_gcse2): ...new pass_data instance and...
8534 (make_pass_gcse2): ...new function.
8535 * postreload.c (pass_postreload_cse): Convert from a global struct to
8536 a subclass of rtl_opt_pass along with...
8537 (pass_data_postreload_cse): ...new pass_data instance and...
8538 (make_pass_postreload_cse): ...new function.
8539 * predict.c (pass_profile): Convert from a global struct to a subclass
8540 of gimple_opt_pass along with...
8541 (pass_data_profile): ...new pass_data instance and...
8542 (make_pass_profile): ...new function.
8543 (pass_strip_predict_hints): Convert from a global struct to a subclass
8544 of gimple_opt_pass along with...
8545 (pass_data_strip_predict_hints): ...new pass_data instance and...
8546 (make_pass_strip_predict_hints): ...new function.
8547 * recog.c (pass_peephole2): Convert from a global struct to a subclass
8548 of rtl_opt_pass along with...
8549 (pass_data_peephole2): ...new pass_data instance and...
8550 (make_pass_peephole2): ...new function.
8551 (pass_split_all_insns): Convert from a global struct to a subclass of
8552 rtl_opt_pass along with...
8553 (pass_data_split_all_insns): ...new pass_data instance and...
8554 (make_pass_split_all_insns): ...new function.
8555 (pass_split_after_reload): Convert from a global struct to a subclass
8556 of rtl_opt_pass along with...
8557 (pass_data_split_after_reload): ...new pass_data instance and...
8558 (make_pass_split_after_reload): ...new function.
8559 (pass_split_before_regstack): Convert from a global struct to a
8560 subclass of rtl_opt_pass along with...
8561 (pass_data_split_before_regstack): ...new pass_data instance and...
8562 (make_pass_split_before_regstack): ...new function.
8563 (pass_split_before_sched2): Convert from a global struct to a subclass
8564 of rtl_opt_pass along with...
8565 (pass_data_split_before_sched2): ...new pass_data instance and...
8566 (make_pass_split_before_sched2): ...new function.
8567 (pass_split_for_shorten_branches): Convert from a global struct to a
8568 subclass of rtl_opt_pass along with...
8569 (pass_data_split_for_shorten_branches): ...new pass_data instance
8571 (make_pass_split_for_shorten_branches): ...new function.
8572 * ree.c (pass_ree): Convert from a global struct to a subclass of
8573 rtl_opt_pass along with...
8574 (pass_data_ree): ...new pass_data instance and...
8575 (make_pass_ree): ...new function.
8576 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
8577 subclass of rtl_opt_pass along with...
8578 (pass_data_stack_regs): ...new pass_data instance and...
8579 (make_pass_stack_regs): ...new function.
8580 (pass_stack_regs_run): Convert from a global struct to a subclass of
8581 rtl_opt_pass along with...
8582 (pass_data_stack_regs_run): ...new pass_data instance and...
8583 (make_pass_stack_regs_run): ...new function.
8584 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
8585 subclass of rtl_opt_pass along with...
8586 (pass_data_cprop_hardreg): ...new pass_data instance and...
8587 (make_pass_cprop_hardreg): ...new function.
8588 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
8589 subclass of rtl_opt_pass along with...
8590 (pass_data_reginfo_init): ...new pass_data instance and...
8591 (make_pass_reginfo_init): ...new function.
8592 * regmove.c (pass_regmove): Convert from a global struct to a subclass
8593 of rtl_opt_pass along with...
8594 (pass_data_regmove): ...new pass_data instance and...
8595 (make_pass_regmove): ...new function.
8596 * regrename.c (pass_regrename): Convert from a global struct to a
8597 subclass of rtl_opt_pass along with...
8598 (pass_data_regrename): ...new pass_data instance and...
8599 (make_pass_regrename): ...new function.
8600 * reorg.c (pass_delay_slots): Convert from a global struct to a
8601 subclass of rtl_opt_pass along with...
8602 (pass_data_delay_slots): ...new pass_data instance and...
8603 (make_pass_delay_slots): ...new function.
8604 (pass_machine_reorg): Convert from a global struct to a subclass of
8605 rtl_opt_pass along with...
8606 (pass_data_machine_reorg): ...new pass_data instance and...
8607 (make_pass_machine_reorg): ...new function.
8608 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
8609 of rtl_opt_pass along with...
8610 (pass_data_sched): ...new pass_data instance and...
8611 (make_pass_sched): ...new function.
8612 (pass_sched2): Convert from a global struct to a subclass of
8613 rtl_opt_pass along with...
8614 (pass_data_sched2): ...new pass_data instance and...
8615 (make_pass_sched2): ...new function.
8616 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
8617 to a subclass of rtl_opt_pass along with...
8618 (pass_data_stack_ptr_mod): ...new pass_data instance and...
8619 (make_pass_stack_ptr_mod): ...new function.
8620 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
8621 to a subclass of rtl_opt_pass along with...
8622 (pass_data_rtl_store_motion): ...new pass_data instance and...
8623 (make_pass_rtl_store_motion): ...new function.
8624 * tracer.c (pass_tracer): Convert from a global struct to a subclass
8625 of gimple_opt_pass along with...
8626 (pass_data_tracer): ...new pass_data instance and...
8627 (make_pass_tracer): ...new function.
8628 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
8629 to a subclass of gimple_opt_pass along with...
8630 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
8631 (make_pass_diagnose_tm_blocks): ...new function.
8632 (pass_lower_tm): Convert from a global struct to a subclass of
8633 gimple_opt_pass along with...
8634 (pass_data_lower_tm): ...new pass_data instance and...
8635 (make_pass_lower_tm): ...new function.
8636 (pass_tm_init): Convert from a global struct to a subclass of
8637 gimple_opt_pass along with...
8638 (pass_data_tm_init): ...new pass_data instance and...
8639 (make_pass_tm_init): ...new function.
8640 (pass_tm_mark): Convert from a global struct to a subclass of
8641 gimple_opt_pass along with...
8642 (pass_data_tm_mark): ...new pass_data instance and...
8643 (make_pass_tm_mark): ...new function.
8644 (pass_tm_edges): Convert from a global struct to a subclass of
8645 gimple_opt_pass along with...
8646 (pass_data_tm_edges): ...new pass_data instance and...
8647 (make_pass_tm_edges): ...new function.
8648 (pass_tm_memopt): Convert from a global struct to a subclass of
8649 gimple_opt_pass along with...
8650 (pass_data_tm_memopt): ...new pass_data instance and...
8651 (make_pass_tm_memopt): ...new function.
8652 (pass_ipa_tm): Convert from a global struct to a subclass of
8653 simple_ipa_opt_pass along with...
8654 (pass_data_ipa_tm): ...new pass_data instance and...
8655 (make_pass_ipa_tm): ...new function.
8656 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
8657 subclass of gimple_opt_pass along with...
8658 (pass_data_call_cdce): ...new pass_data instance and...
8659 (make_pass_call_cdce): ...new function.
8660 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
8661 subclass of gimple_opt_pass along with...
8662 (pass_data_build_cfg): ...new pass_data instance and...
8663 (make_pass_build_cfg): ...new function.
8664 (pass_split_crit_edges): Convert from a global struct to a subclass of
8665 gimple_opt_pass along with...
8666 (pass_data_split_crit_edges): ...new pass_data instance and...
8667 (make_pass_split_crit_edges): ...new function.
8668 (pass_warn_function_return): Convert from a global struct to a
8669 subclass of gimple_opt_pass along with...
8670 (pass_data_warn_function_return): ...new pass_data instance and...
8671 (make_pass_warn_function_return): ...new function.
8672 (pass_warn_function_noreturn): Convert from a global struct to a
8673 subclass of gimple_opt_pass along with...
8674 (pass_data_warn_function_noreturn): ...new pass_data instance and...
8675 (make_pass_warn_function_noreturn): ...new function.
8676 (pass_warn_unused_result): Convert from a global struct to a subclass
8677 of gimple_opt_pass along with...
8678 (pass_data_warn_unused_result): ...new pass_data instance and...
8679 (make_pass_warn_unused_result): ...new function.
8680 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
8681 a subclass of gimple_opt_pass along with...
8682 (pass_data_merge_phi): ...new pass_data instance and...
8683 (make_pass_merge_phi): ...new function.
8684 * tree-complex.c (pass_lower_complex): Convert from a global struct to
8685 a subclass of gimple_opt_pass along with...
8686 (pass_data_lower_complex): ...new pass_data instance and...
8687 (make_pass_lower_complex): ...new function.
8688 (pass_lower_complex_O0): Convert from a global struct to a subclass of
8689 gimple_opt_pass along with...
8690 (pass_data_lower_complex_O0): ...new pass_data instance and...
8691 (make_pass_lower_complex_O0): ...new function.
8692 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
8693 subclass of gimple_opt_pass along with...
8694 (pass_data_lower_eh): ...new pass_data instance and...
8695 (make_pass_lower_eh): ...new function.
8696 (pass_refactor_eh): Convert from a global struct to a subclass of
8697 gimple_opt_pass along with...
8698 (pass_data_refactor_eh): ...new pass_data instance and...
8699 (make_pass_refactor_eh): ...new function.
8700 (pass_lower_resx): Convert from a global struct to a subclass of
8701 gimple_opt_pass along with...
8702 (pass_data_lower_resx): ...new pass_data instance and...
8703 (make_pass_lower_resx): ...new function.
8704 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
8705 of gimple_opt_pass along with...
8706 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
8707 (make_pass_lower_eh_dispatch): ...new function.
8708 (pass_cleanup_eh): Convert from a global struct to a subclass of
8709 gimple_opt_pass along with...
8710 (pass_data_cleanup_eh): ...new pass_data instance and...
8711 (make_pass_cleanup_eh): ...new function.
8712 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
8713 to a subclass of simple_ipa_opt_pass along with...
8714 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
8715 (make_pass_ipa_lower_emutls): ...new function.
8716 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
8717 a subclass of gimple_opt_pass along with...
8718 (pass_data_if_conversion): ...new pass_data instance and...
8719 (make_pass_if_conversion): ...new function.
8720 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
8721 subclass of gimple_opt_pass along with...
8722 (pass_data_build_ssa): ...new pass_data instance and...
8723 (make_pass_build_ssa): ...new function.
8724 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
8725 global struct to a subclass of gimple_opt_pass along with...
8726 (pass_data_loop_distribution): ...new pass_data instance and...
8727 (make_pass_loop_distribution): ...new function.
8728 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
8729 subclass of gimple_opt_pass along with...
8730 (pass_data_mudflap_1): ...new pass_data instance and...
8731 (make_pass_mudflap_1): ...new function.
8732 (pass_mudflap_2): Convert from a global struct to a subclass of
8733 gimple_opt_pass along with...
8734 (pass_data_mudflap_2): ...new pass_data instance and...
8735 (make_pass_mudflap_2): ...new function.
8736 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
8737 subclass of gimple_opt_pass along with...
8738 (pass_data_mudflap_1): ...new pass_data instance and...
8739 (make_pass_mudflap_1): ...new function.
8740 (pass_mudflap_2): Convert from a global struct to a subclass of
8741 gimple_opt_pass along with...
8742 (pass_data_mudflap_2): ...new pass_data instance and...
8743 (make_pass_mudflap_2): ...new function.
8744 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
8745 gimple_opt_pass along with...
8746 (pass_data_nrv): ...new pass_data instance and...
8747 (make_pass_nrv): ...new function.
8748 (pass_return_slot): Convert from a global struct to a subclass of
8749 gimple_opt_pass along with...
8750 (pass_data_return_slot): ...new pass_data instance and...
8751 (make_pass_return_slot): ...new function.
8752 * tree-object-size.c (pass_object_sizes): Convert from a global struct
8753 to a subclass of gimple_opt_pass along with...
8754 (pass_data_object_sizes): ...new pass_data instance and...
8755 (make_pass_object_sizes): ...new function.
8756 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
8757 global struct to a subclass of gimple_opt_pass along with...
8758 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
8760 (make_pass_cleanup_cfg_post_optimizing): ...new function.
8761 (pass_fixup_cfg): Convert from a global struct to a subclass of
8762 gimple_opt_pass along with...
8763 (pass_data_fixup_cfg): ...new pass_data instance and...
8764 (make_pass_fixup_cfg): ...new function.
8765 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
8766 (make_pass_mudflap_1): ...new function.
8767 (pass_mudflap_2): Replace declaration with that of...
8768 (make_pass_mudflap_2): ...new function.
8769 (pass_asan): Replace declaration with that of...
8770 (make_pass_asan): ...new function.
8771 (pass_asan_O0): Replace declaration with that of...
8772 (make_pass_asan_O0): ...new function.
8773 (pass_tsan): Replace declaration with that of...
8774 (make_pass_tsan): ...new function.
8775 (pass_tsan_O0): Replace declaration with that of...
8776 (make_pass_tsan_O0): ...new function.
8777 (pass_lower_cf): Replace declaration with that of...
8778 (make_pass_lower_cf): ...new function.
8779 (pass_refactor_eh): Replace declaration with that of...
8780 (make_pass_refactor_eh): ...new function.
8781 (pass_lower_eh): Replace declaration with that of...
8782 (make_pass_lower_eh): ...new function.
8783 (pass_lower_eh_dispatch): Replace declaration with that of...
8784 (make_pass_lower_eh_dispatch): ...new function.
8785 (pass_lower_resx): Replace declaration with that of...
8786 (make_pass_lower_resx): ...new function.
8787 (pass_build_cfg): Replace declaration with that of...
8788 (make_pass_build_cfg): ...new function.
8789 (pass_early_tree_profile): Replace declaration with that of...
8790 (make_pass_early_tree_profile): ...new function.
8791 (pass_cleanup_eh): Replace declaration with that of...
8792 (make_pass_cleanup_eh): ...new function.
8793 (pass_sra): Replace declaration with that of...
8794 (make_pass_sra): ...new function.
8795 (pass_sra_early): Replace declaration with that of...
8796 (make_pass_sra_early): ...new function.
8797 (pass_early_ipa_sra): Replace declaration with that of...
8798 (make_pass_early_ipa_sra): ...new function.
8799 (pass_tail_recursion): Replace declaration with that of...
8800 (make_pass_tail_recursion): ...new function.
8801 (pass_tail_calls): Replace declaration with that of...
8802 (make_pass_tail_calls): ...new function.
8803 (pass_tree_loop): Replace declaration with that of...
8804 (make_pass_tree_loop): ...new function.
8805 (pass_tree_loop_init): Replace declaration with that of...
8806 (make_pass_tree_loop_init): ...new function.
8807 (pass_lim): Replace declaration with that of...
8808 (make_pass_lim): ...new function.
8809 (pass_tree_unswitch): Replace declaration with that of...
8810 (make_pass_tree_unswitch): ...new function.
8811 (pass_predcom): Replace declaration with that of...
8812 (make_pass_predcom): ...new function.
8813 (pass_iv_canon): Replace declaration with that of...
8814 (make_pass_iv_canon): ...new function.
8815 (pass_scev_cprop): Replace declaration with that of...
8816 (make_pass_scev_cprop): ...new function.
8817 (pass_empty_loop): Replace declaration with that of...
8818 (make_pass_empty_loop): ...new function.
8819 (pass_record_bounds): Replace declaration with that of...
8820 (make_pass_record_bounds): ...new function.
8821 (pass_graphite): Replace declaration with that of...
8822 (make_pass_graphite): ...new function.
8823 (pass_graphite_transforms): Replace declaration with that of...
8824 (make_pass_graphite_transforms): ...new function.
8825 (pass_if_conversion): Replace declaration with that of...
8826 (make_pass_if_conversion): ...new function.
8827 (pass_loop_distribution): Replace declaration with that of...
8828 (make_pass_loop_distribution): ...new function.
8829 (pass_vectorize): Replace declaration with that of...
8830 (make_pass_vectorize): ...new function.
8831 (pass_slp_vectorize): Replace declaration with that of...
8832 (make_pass_slp_vectorize): ...new function.
8833 (pass_complete_unroll): Replace declaration with that of...
8834 (make_pass_complete_unroll): ...new function.
8835 (pass_complete_unrolli): Replace declaration with that of...
8836 (make_pass_complete_unrolli): ...new function.
8837 (pass_parallelize_loops): Replace declaration with that of...
8838 (make_pass_parallelize_loops): ...new function.
8839 (pass_loop_prefetch): Replace declaration with that of...
8840 (make_pass_loop_prefetch): ...new function.
8841 (pass_iv_optimize): Replace declaration with that of...
8842 (make_pass_iv_optimize): ...new function.
8843 (pass_tree_loop_done): Replace declaration with that of...
8844 (make_pass_tree_loop_done): ...new function.
8845 (pass_ch): Replace declaration with that of...
8846 (make_pass_ch): ...new function.
8847 (pass_ccp): Replace declaration with that of...
8848 (make_pass_ccp): ...new function.
8849 (pass_phi_only_cprop): Replace declaration with that of...
8850 (make_pass_phi_only_cprop): ...new function.
8851 (pass_build_ssa): Replace declaration with that of...
8852 (make_pass_build_ssa): ...new function.
8853 (pass_build_alias): Replace declaration with that of...
8854 (make_pass_build_alias): ...new function.
8855 (pass_build_ealias): Replace declaration with that of...
8856 (make_pass_build_ealias): ...new function.
8857 (pass_dominator): Replace declaration with that of...
8858 (make_pass_dominator): ...new function.
8859 (pass_dce): Replace declaration with that of...
8860 (make_pass_dce): ...new function.
8861 (pass_dce_loop): Replace declaration with that of...
8862 (make_pass_dce_loop): ...new function.
8863 (pass_cd_dce): Replace declaration with that of...
8864 (make_pass_cd_dce): ...new function.
8865 (pass_call_cdce): Replace declaration with that of...
8866 (make_pass_call_cdce): ...new function.
8867 (pass_merge_phi): Replace declaration with that of...
8868 (make_pass_merge_phi): ...new function.
8869 (pass_split_crit_edges): Replace declaration with that of...
8870 (make_pass_split_crit_edges): ...new function.
8871 (pass_pre): Replace declaration with that of...
8872 (make_pass_pre): ...new function.
8873 (pass_profile): Replace declaration with that of...
8874 (make_pass_profile): ...new function.
8875 (pass_strip_predict_hints): Replace declaration with that of...
8876 (make_pass_strip_predict_hints): ...new function.
8877 (pass_lower_complex_O0): Replace declaration with that of...
8878 (make_pass_lower_complex_O0): ...new function.
8879 (pass_lower_complex): Replace declaration with that of...
8880 (make_pass_lower_complex): ...new function.
8881 (pass_lower_vector): Replace declaration with that of...
8882 (make_pass_lower_vector): ...new function.
8883 (pass_lower_vector_ssa): Replace declaration with that of...
8884 (make_pass_lower_vector_ssa): ...new function.
8885 (pass_lower_omp): Replace declaration with that of...
8886 (make_pass_lower_omp): ...new function.
8887 (pass_diagnose_omp_blocks): Replace declaration with that of...
8888 (make_pass_diagnose_omp_blocks): ...new function.
8889 (pass_expand_omp): Replace declaration with that of...
8890 (make_pass_expand_omp): ...new function.
8891 (pass_expand_omp_ssa): Replace declaration with that of...
8892 (make_pass_expand_omp_ssa): ...new function.
8893 (pass_object_sizes): Replace declaration with that of...
8894 (make_pass_object_sizes): ...new function.
8895 (pass_strlen): Replace declaration with that of...
8896 (make_pass_strlen): ...new function.
8897 (pass_fold_builtins): Replace declaration with that of...
8898 (make_pass_fold_builtins): ...new function.
8899 (pass_stdarg): Replace declaration with that of...
8900 (make_pass_stdarg): ...new function.
8901 (pass_early_warn_uninitialized): Replace declaration with that of...
8902 (make_pass_early_warn_uninitialized): ...new function.
8903 (pass_late_warn_uninitialized): Replace declaration with that of...
8904 (make_pass_late_warn_uninitialized): ...new function.
8905 (pass_cse_reciprocals): Replace declaration with that of...
8906 (make_pass_cse_reciprocals): ...new function.
8907 (pass_cse_sincos): Replace declaration with that of...
8908 (make_pass_cse_sincos): ...new function.
8909 (pass_optimize_bswap): Replace declaration with that of...
8910 (make_pass_optimize_bswap): ...new function.
8911 (pass_optimize_widening_mul): Replace declaration with that of...
8912 (make_pass_optimize_widening_mul): ...new function.
8913 (pass_warn_function_return): Replace declaration with that of...
8914 (make_pass_warn_function_return): ...new function.
8915 (pass_warn_function_noreturn): Replace declaration with that of...
8916 (make_pass_warn_function_noreturn): ...new function.
8917 (pass_cselim): Replace declaration with that of...
8918 (make_pass_cselim): ...new function.
8919 (pass_phiopt): Replace declaration with that of...
8920 (make_pass_phiopt): ...new function.
8921 (pass_forwprop): Replace declaration with that of...
8922 (make_pass_forwprop): ...new function.
8923 (pass_phiprop): Replace declaration with that of...
8924 (make_pass_phiprop): ...new function.
8925 (pass_tree_ifcombine): Replace declaration with that of...
8926 (make_pass_tree_ifcombine): ...new function.
8927 (pass_dse): Replace declaration with that of...
8928 (make_pass_dse): ...new function.
8929 (pass_nrv): Replace declaration with that of...
8930 (make_pass_nrv): ...new function.
8931 (pass_rename_ssa_copies): Replace declaration with that of...
8932 (make_pass_rename_ssa_copies): ...new function.
8933 (pass_sink_code): Replace declaration with that of...
8934 (make_pass_sink_code): ...new function.
8935 (pass_fre): Replace declaration with that of...
8936 (make_pass_fre): ...new function.
8937 (pass_check_data_deps): Replace declaration with that of...
8938 (make_pass_check_data_deps): ...new function.
8939 (pass_copy_prop): Replace declaration with that of...
8940 (make_pass_copy_prop): ...new function.
8941 (pass_vrp): Replace declaration with that of...
8942 (make_pass_vrp): ...new function.
8943 (pass_uncprop): Replace declaration with that of...
8944 (make_pass_uncprop): ...new function.
8945 (pass_return_slot): Replace declaration with that of...
8946 (make_pass_return_slot): ...new function.
8947 (pass_reassoc): Replace declaration with that of...
8948 (make_pass_reassoc): ...new function.
8949 (pass_rebuild_cgraph_edges): Replace declaration with that of...
8950 (make_pass_rebuild_cgraph_edges): ...new function.
8951 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
8952 (make_pass_remove_cgraph_callee_edges): ...new function.
8953 (pass_build_cgraph_edges): Replace declaration with that of...
8954 (make_pass_build_cgraph_edges): ...new function.
8955 (pass_local_pure_const): Replace declaration with that of...
8956 (make_pass_local_pure_const): ...new function.
8957 (pass_tracer): Replace declaration with that of...
8958 (make_pass_tracer): ...new function.
8959 (pass_warn_unused_result): Replace declaration with that of...
8960 (make_pass_warn_unused_result): ...new function.
8961 (pass_diagnose_tm_blocks): Replace declaration with that of...
8962 (make_pass_diagnose_tm_blocks): ...new function.
8963 (pass_lower_tm): Replace declaration with that of...
8964 (make_pass_lower_tm): ...new function.
8965 (pass_tm_init): Replace declaration with that of...
8966 (make_pass_tm_init): ...new function.
8967 (pass_tm_mark): Replace declaration with that of...
8968 (make_pass_tm_mark): ...new function.
8969 (pass_tm_memopt): Replace declaration with that of...
8970 (make_pass_tm_memopt): ...new function.
8971 (pass_tm_edges): Replace declaration with that of...
8972 (make_pass_tm_edges): ...new function.
8973 (pass_split_functions): Replace declaration with that of...
8974 (make_pass_split_functions): ...new function.
8975 (pass_feedback_split_functions): Replace declaration with that of...
8976 (make_pass_feedback_split_functions): ...new function.
8977 (pass_strength_reduction): Replace declaration with that of...
8978 (make_pass_strength_reduction): ...new function.
8979 (pass_ipa_lower_emutls): Replace declaration with that of...
8980 (make_pass_ipa_lower_emutls): ...new function.
8981 (pass_ipa_function_and_variable_visibility): Replace declaration with
8983 (make_pass_ipa_function_and_variable_visibility): ...new function.
8984 (pass_ipa_tree_profile): Replace declaration with that of...
8985 (make_pass_ipa_tree_profile): ...new function.
8986 (pass_early_local_passes): Replace declaration with that of...
8987 (make_pass_early_local_passes): ...new function.
8988 (pass_ipa_whole_program_visibility): Replace declaration with that
8990 (make_pass_ipa_whole_program_visibility): ...new function.
8991 (pass_ipa_lto_gimple_out): Replace declaration with that of...
8992 (make_pass_ipa_lto_gimple_out): ...new function.
8993 (pass_ipa_increase_alignment): Replace declaration with that of...
8994 (make_pass_ipa_increase_alignment): ...new function.
8995 (pass_ipa_inline): Replace declaration with that of...
8996 (make_pass_ipa_inline): ...new function.
8997 (pass_ipa_free_lang_data): Replace declaration with that of...
8998 (make_pass_ipa_free_lang_data): ...new function.
8999 (pass_ipa_free_inline_summary): Replace declaration with that of...
9000 (make_pass_ipa_free_inline_summary): ...new function.
9001 (pass_ipa_cp): Replace declaration with that of...
9002 (make_pass_ipa_cp): ...new function.
9003 (pass_ipa_reference): Replace declaration with that of...
9004 (make_pass_ipa_reference): ...new function.
9005 (pass_ipa_pure_const): Replace declaration with that of...
9006 (make_pass_ipa_pure_const): ...new function.
9007 (pass_ipa_pta): Replace declaration with that of...
9008 (make_pass_ipa_pta): ...new function.
9009 (pass_ipa_lto_finish_out): Replace declaration with that of...
9010 (make_pass_ipa_lto_finish_out): ...new function.
9011 (pass_ipa_tm): Replace declaration with that of...
9012 (make_pass_ipa_tm): ...new function.
9013 (pass_ipa_profile): Replace declaration with that of...
9014 (make_pass_ipa_profile): ...new function.
9015 (pass_ipa_cdtor_merge): Replace declaration with that of...
9016 (make_pass_ipa_cdtor_merge): ...new function.
9017 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
9019 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9020 (pass_init_datastructures): Replace declaration with that of...
9021 (make_pass_init_datastructures): ...new function.
9022 (pass_fixup_cfg): Replace declaration with that of...
9023 (make_pass_fixup_cfg): ...new function.
9024 (pass_expand): Replace declaration with that of...
9025 (make_pass_expand): ...new function.
9026 (pass_instantiate_virtual_regs): Replace declaration with that of...
9027 (make_pass_instantiate_virtual_regs): ...new function.
9028 (pass_rtl_fwprop): Replace declaration with that of...
9029 (make_pass_rtl_fwprop): ...new function.
9030 (pass_rtl_fwprop_addr): Replace declaration with that of...
9031 (make_pass_rtl_fwprop_addr): ...new function.
9032 (pass_jump): Replace declaration with that of...
9033 (make_pass_jump): ...new function.
9034 (pass_jump2): Replace declaration with that of...
9035 (make_pass_jump2): ...new function.
9036 (pass_lower_subreg): Replace declaration with that of...
9037 (make_pass_lower_subreg): ...new function.
9038 (pass_cse): Replace declaration with that of...
9039 (make_pass_cse): ...new function.
9040 (pass_fast_rtl_dce): Replace declaration with that of...
9041 (make_pass_fast_rtl_dce): ...new function.
9042 (pass_ud_rtl_dce): Replace declaration with that of...
9043 (make_pass_ud_rtl_dce): ...new function.
9044 (pass_rtl_dce): Replace declaration with that of...
9045 (make_pass_rtl_dce): ...new function.
9046 (pass_rtl_dse1): Replace declaration with that of...
9047 (make_pass_rtl_dse1): ...new function.
9048 (pass_rtl_dse2): Replace declaration with that of...
9049 (make_pass_rtl_dse2): ...new function.
9050 (pass_rtl_dse3): Replace declaration with that of...
9051 (make_pass_rtl_dse3): ...new function.
9052 (pass_rtl_cprop): Replace declaration with that of...
9053 (make_pass_rtl_cprop): ...new function.
9054 (pass_rtl_pre): Replace declaration with that of...
9055 (make_pass_rtl_pre): ...new function.
9056 (pass_rtl_hoist): Replace declaration with that of...
9057 (make_pass_rtl_hoist): ...new function.
9058 (pass_rtl_store_motion): Replace declaration with that of...
9059 (make_pass_rtl_store_motion): ...new function.
9060 (pass_cse_after_global_opts): Replace declaration with that of...
9061 (make_pass_cse_after_global_opts): ...new function.
9062 (pass_rtl_ifcvt): Replace declaration with that of...
9063 (make_pass_rtl_ifcvt): ...new function.
9064 (pass_into_cfg_layout_mode): Replace declaration with that of...
9065 (make_pass_into_cfg_layout_mode): ...new function.
9066 (pass_outof_cfg_layout_mode): Replace declaration with that of...
9067 (make_pass_outof_cfg_layout_mode): ...new function.
9068 (pass_loop2): Replace declaration with that of...
9069 (make_pass_loop2): ...new function.
9070 (pass_rtl_loop_init): Replace declaration with that of...
9071 (make_pass_rtl_loop_init): ...new function.
9072 (pass_rtl_move_loop_invariants): Replace declaration with that of...
9073 (make_pass_rtl_move_loop_invariants): ...new function.
9074 (pass_rtl_unswitch): Replace declaration with that of...
9075 (make_pass_rtl_unswitch): ...new function.
9076 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
9077 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9078 (pass_rtl_doloop): Replace declaration with that of...
9079 (make_pass_rtl_doloop): ...new function.
9080 (pass_rtl_loop_done): Replace declaration with that of...
9081 (make_pass_rtl_loop_done): ...new function.
9082 (pass_web): Replace declaration with that of...
9083 (make_pass_web): ...new function.
9084 (pass_cse2): Replace declaration with that of...
9085 (make_pass_cse2): ...new function.
9086 (pass_df_initialize_opt): Replace declaration with that of...
9087 (make_pass_df_initialize_opt): ...new function.
9088 (pass_df_initialize_no_opt): Replace declaration with that of...
9089 (make_pass_df_initialize_no_opt): ...new function.
9090 (pass_reginfo_init): Replace declaration with that of...
9091 (make_pass_reginfo_init): ...new function.
9092 (pass_inc_dec): Replace declaration with that of...
9093 (make_pass_inc_dec): ...new function.
9094 (pass_stack_ptr_mod): Replace declaration with that of...
9095 (make_pass_stack_ptr_mod): ...new function.
9096 (pass_initialize_regs): Replace declaration with that of...
9097 (make_pass_initialize_regs): ...new function.
9098 (pass_combine): Replace declaration with that of...
9099 (make_pass_combine): ...new function.
9100 (pass_if_after_combine): Replace declaration with that of...
9101 (make_pass_if_after_combine): ...new function.
9102 (pass_ree): Replace declaration with that of...
9103 (make_pass_ree): ...new function.
9104 (pass_partition_blocks): Replace declaration with that of...
9105 (make_pass_partition_blocks): ...new function.
9106 (pass_match_asm_constraints): Replace declaration with that of...
9107 (make_pass_match_asm_constraints): ...new function.
9108 (pass_regmove): Replace declaration with that of...
9109 (make_pass_regmove): ...new function.
9110 (pass_split_all_insns): Replace declaration with that of...
9111 (make_pass_split_all_insns): ...new function.
9112 (pass_fast_rtl_byte_dce): Replace declaration with that of...
9113 (make_pass_fast_rtl_byte_dce): ...new function.
9114 (pass_lower_subreg2): Replace declaration with that of...
9115 (make_pass_lower_subreg2): ...new function.
9116 (pass_mode_switching): Replace declaration with that of...
9117 (make_pass_mode_switching): ...new function.
9118 (pass_sms): Replace declaration with that of...
9119 (make_pass_sms): ...new function.
9120 (pass_sched): Replace declaration with that of...
9121 (make_pass_sched): ...new function.
9122 (pass_ira): Replace declaration with that of...
9123 (make_pass_ira): ...new function.
9124 (pass_reload): Replace declaration with that of...
9125 (make_pass_reload): ...new function.
9126 (pass_clean_state): Replace declaration with that of...
9127 (make_pass_clean_state): ...new function.
9128 (pass_branch_prob): Replace declaration with that of...
9129 (make_pass_branch_prob): ...new function.
9130 (pass_value_profile_transformations): Replace declaration with that
9132 (make_pass_value_profile_transformations): ...new function.
9133 (pass_postreload_cse): Replace declaration with that of...
9134 (make_pass_postreload_cse): ...new function.
9135 (pass_gcse2): Replace declaration with that of...
9136 (make_pass_gcse2): ...new function.
9137 (pass_split_after_reload): Replace declaration with that of...
9138 (make_pass_split_after_reload): ...new function.
9139 (pass_branch_target_load_optimize1): Replace declaration with that
9141 (make_pass_branch_target_load_optimize1): ...new function.
9142 (pass_thread_prologue_and_epilogue): Replace declaration with that
9144 (make_pass_thread_prologue_and_epilogue): ...new function.
9145 (pass_stack_adjustments): Replace declaration with that of...
9146 (make_pass_stack_adjustments): ...new function.
9147 (pass_peephole2): Replace declaration with that of...
9148 (make_pass_peephole2): ...new function.
9149 (pass_if_after_reload): Replace declaration with that of...
9150 (make_pass_if_after_reload): ...new function.
9151 (pass_regrename): Replace declaration with that of...
9152 (make_pass_regrename): ...new function.
9153 (pass_cprop_hardreg): Replace declaration with that of...
9154 (make_pass_cprop_hardreg): ...new function.
9155 (pass_reorder_blocks): Replace declaration with that of...
9156 (make_pass_reorder_blocks): ...new function.
9157 (pass_branch_target_load_optimize2): Replace declaration with that
9159 (make_pass_branch_target_load_optimize2): ...new function.
9160 (pass_leaf_regs): Replace declaration with that of...
9161 (make_pass_leaf_regs): ...new function.
9162 (pass_split_before_sched2): Replace declaration with that of...
9163 (make_pass_split_before_sched2): ...new function.
9164 (pass_compare_elim_after_reload): Replace declaration with that of...
9165 (make_pass_compare_elim_after_reload): ...new function.
9166 (pass_sched2): Replace declaration with that of...
9167 (make_pass_sched2): ...new function.
9168 (pass_stack_regs): Replace declaration with that of...
9169 (make_pass_stack_regs): ...new function.
9170 (pass_stack_regs_run): Replace declaration with that of...
9171 (make_pass_stack_regs_run): ...new function.
9172 (pass_df_finish): Replace declaration with that of...
9173 (make_pass_df_finish): ...new function.
9174 (pass_compute_alignments): Replace declaration with that of...
9175 (make_pass_compute_alignments): ...new function.
9176 (pass_duplicate_computed_gotos): Replace declaration with that of...
9177 (make_pass_duplicate_computed_gotos): ...new function.
9178 (pass_variable_tracking): Replace declaration with that of...
9179 (make_pass_variable_tracking): ...new function.
9180 (pass_free_cfg): Replace declaration with that of...
9181 (make_pass_free_cfg): ...new function.
9182 (pass_machine_reorg): Replace declaration with that of...
9183 (make_pass_machine_reorg): ...new function.
9184 (pass_cleanup_barriers): Replace declaration with that of...
9185 (make_pass_cleanup_barriers): ...new function.
9186 (pass_delay_slots): Replace declaration with that of...
9187 (make_pass_delay_slots): ...new function.
9188 (pass_split_for_shorten_branches): Replace declaration with that of...
9189 (make_pass_split_for_shorten_branches): ...new function.
9190 (pass_split_before_regstack): Replace declaration with that of...
9191 (make_pass_split_before_regstack): ...new function.
9192 (pass_convert_to_eh_region_ranges): Replace declaration with that
9194 (make_pass_convert_to_eh_region_ranges): ...new function.
9195 (pass_shorten_branches): Replace declaration with that of...
9196 (make_pass_shorten_branches): ...new function.
9197 (pass_set_nothrow_function_flags): Replace declaration with that of...
9198 (make_pass_set_nothrow_function_flags): ...new function.
9199 (pass_dwarf2_frame): Replace declaration with that of...
9200 (make_pass_dwarf2_frame): ...new function.
9201 (pass_final): Replace declaration with that of...
9202 (make_pass_final): ...new function.
9203 (pass_rtl_seqabstr): Replace declaration with that of...
9204 (make_pass_rtl_seqabstr): ...new function.
9205 (pass_release_ssa_names): Replace declaration with that of...
9206 (make_pass_release_ssa_names): ...new function.
9207 (pass_early_inline): Replace declaration with that of...
9208 (make_pass_early_inline): ...new function.
9209 (pass_inline_parameters): Replace declaration with that of...
9210 (make_pass_inline_parameters): ...new function.
9211 (pass_update_address_taken): Replace declaration with that of...
9212 (make_pass_update_address_taken): ...new function.
9213 (pass_convert_switch): Replace declaration with that of...
9214 (make_pass_convert_switch): ...new function.
9215 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
9216 to a subclass of simple_ipa_opt_pass along with...
9217 (pass_data_ipa_tree_profile): ...new pass_data instance and...
9218 (make_pass_ipa_tree_profile): ...new function.
9219 * tree-sra.c (pass_sra_early): Convert from a global struct to a
9220 subclass of gimple_opt_pass along with...
9221 (pass_data_sra_early): ...new pass_data instance and...
9222 (make_pass_sra_early): ...new function.
9223 (pass_sra): Convert from a global struct to a subclass of
9224 gimple_opt_pass along with...
9225 (pass_data_sra): ...new pass_data instance and...
9226 (make_pass_sra): ...new function.
9227 (pass_early_ipa_sra): Convert from a global struct to a subclass of
9228 gimple_opt_pass along with...
9229 (pass_data_early_ipa_sra): ...new pass_data instance and...
9230 (make_pass_early_ipa_sra): ...new function.
9231 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
9232 subclass of gimple_opt_pass along with...
9233 (pass_data_ccp): ...new pass_data instance and...
9234 (make_pass_ccp): ...new function.
9235 (pass_fold_builtins): Convert from a global struct to a subclass of
9236 gimple_opt_pass along with...
9237 (pass_data_fold_builtins): ...new pass_data instance and...
9238 (make_pass_fold_builtins): ...new function.
9239 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
9240 subclass of gimple_opt_pass along with...
9241 (pass_data_copy_prop): ...new pass_data instance and...
9242 (make_pass_copy_prop): ...new function.
9243 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
9244 global struct to a subclass of gimple_opt_pass along with...
9245 (pass_data_rename_ssa_copies): ...new pass_data instance and...
9246 (make_pass_rename_ssa_copies): ...new function.
9247 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
9248 subclass of gimple_opt_pass along with...
9249 (pass_data_dce): ...new pass_data instance and...
9250 (make_pass_dce): ...new function.
9251 (pass_dce_loop): Convert from a global struct to a subclass of
9252 gimple_opt_pass along with...
9253 (pass_data_dce_loop): ...new pass_data instance and...
9254 (make_pass_dce_loop): ...new function.
9255 (pass_cd_dce): Convert from a global struct to a subclass of
9256 gimple_opt_pass along with...
9257 (pass_data_cd_dce): ...new pass_data instance and...
9258 (make_pass_cd_dce): ...new function.
9259 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
9260 subclass of gimple_opt_pass along with...
9261 (pass_data_dominator): ...new pass_data instance and...
9262 (make_pass_dominator): ...new function.
9263 (pass_phi_only_cprop): Convert from a global struct to a subclass of
9264 gimple_opt_pass along with...
9265 (pass_data_phi_only_cprop): ...new pass_data instance and...
9266 (make_pass_phi_only_cprop): ...new function.
9267 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
9268 subclass of gimple_opt_pass along with...
9269 (pass_data_dse): ...new pass_data instance and...
9270 (make_pass_dse): ...new function.
9271 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
9272 a subclass of gimple_opt_pass along with...
9273 (pass_data_forwprop): ...new pass_data instance and...
9274 (make_pass_forwprop): ...new function.
9275 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
9276 struct to a subclass of gimple_opt_pass along with...
9277 (pass_data_tree_ifcombine): ...new pass_data instance and...
9278 (make_pass_tree_ifcombine): ...new function.
9279 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
9280 subclass of gimple_opt_pass along with...
9281 (pass_data_ch): ...new pass_data instance and...
9282 (make_pass_ch): ...new function.
9283 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
9284 subclass of gimple_opt_pass along with...
9285 (pass_data_tree_loop): ...new pass_data instance and...
9286 (make_pass_tree_loop): ...new function.
9287 (pass_tree_loop_init): Convert from a global struct to a subclass of
9288 gimple_opt_pass along with...
9289 (pass_data_tree_loop_init): ...new pass_data instance and...
9290 (make_pass_tree_loop_init): ...new function.
9291 (pass_lim): Convert from a global struct to a subclass of
9292 gimple_opt_pass along with...
9293 (pass_data_lim): ...new pass_data instance and...
9294 (make_pass_lim): ...new function.
9295 (pass_tree_unswitch): Convert from a global struct to a subclass of
9296 gimple_opt_pass along with...
9297 (pass_data_tree_unswitch): ...new pass_data instance and...
9298 (make_pass_tree_unswitch): ...new function.
9299 (pass_predcom): Convert from a global struct to a subclass of
9300 gimple_opt_pass along with...
9301 (pass_data_predcom): ...new pass_data instance and...
9302 (make_pass_predcom): ...new function.
9303 (pass_vectorize): Convert from a global struct to a subclass of
9304 gimple_opt_pass along with...
9305 (pass_data_vectorize): ...new pass_data instance and...
9306 (make_pass_vectorize): ...new function.
9307 (pass_graphite): Convert from a global struct to a subclass of
9308 gimple_opt_pass along with...
9309 (pass_data_graphite): ...new pass_data instance and...
9310 (make_pass_graphite): ...new function.
9311 (pass_graphite_transforms): Convert from a global struct to a subclass
9312 of gimple_opt_pass along with...
9313 (pass_data_graphite_transforms): ...new pass_data instance and...
9314 (make_pass_graphite_transforms): ...new function.
9315 (pass_check_data_deps): Convert from a global struct to a subclass of
9316 gimple_opt_pass along with...
9317 (pass_data_check_data_deps): ...new pass_data instance and...
9318 (make_pass_check_data_deps): ...new function.
9319 (pass_iv_canon): Convert from a global struct to a subclass of
9320 gimple_opt_pass along with...
9321 (pass_data_iv_canon): ...new pass_data instance and...
9322 (make_pass_iv_canon): ...new function.
9323 (pass_scev_cprop): Convert from a global struct to a subclass of
9324 gimple_opt_pass along with...
9325 (pass_data_scev_cprop): ...new pass_data instance and...
9326 (make_pass_scev_cprop): ...new function.
9327 (pass_record_bounds): Convert from a global struct to a subclass of
9328 gimple_opt_pass along with...
9329 (pass_data_record_bounds): ...new pass_data instance and...
9330 (make_pass_record_bounds): ...new function.
9331 (pass_complete_unroll): Convert from a global struct to a subclass of
9332 gimple_opt_pass along with...
9333 (pass_data_complete_unroll): ...new pass_data instance and...
9334 (make_pass_complete_unroll): ...new function.
9335 (pass_complete_unrolli): Convert from a global struct to a subclass of
9336 gimple_opt_pass along with...
9337 (pass_data_complete_unrolli): ...new pass_data instance and...
9338 (make_pass_complete_unrolli): ...new function.
9339 (pass_parallelize_loops): Convert from a global struct to a subclass
9340 of gimple_opt_pass along with...
9341 (pass_data_parallelize_loops): ...new pass_data instance and...
9342 (make_pass_parallelize_loops): ...new function.
9343 (pass_loop_prefetch): Convert from a global struct to a subclass of
9344 gimple_opt_pass along with...
9345 (pass_data_loop_prefetch): ...new pass_data instance and...
9346 (make_pass_loop_prefetch): ...new function.
9347 (pass_iv_optimize): Convert from a global struct to a subclass of
9348 gimple_opt_pass along with...
9349 (pass_data_iv_optimize): ...new pass_data instance and...
9350 (make_pass_iv_optimize): ...new function.
9351 (pass_tree_loop_done): Convert from a global struct to a subclass of
9352 gimple_opt_pass along with...
9353 (pass_data_tree_loop_done): ...new pass_data instance and...
9354 (make_pass_tree_loop_done): ...new function.
9355 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
9356 struct to a subclass of gimple_opt_pass along with...
9357 (pass_data_cse_reciprocals): ...new pass_data instance and...
9358 (make_pass_cse_reciprocals): ...new function.
9359 (pass_cse_sincos): Convert from a global struct to a subclass of
9360 gimple_opt_pass along with...
9361 (pass_data_cse_sincos): ...new pass_data instance and...
9362 (make_pass_cse_sincos): ...new function.
9363 (pass_optimize_bswap): Convert from a global struct to a subclass of
9364 gimple_opt_pass along with...
9365 (pass_data_optimize_bswap): ...new pass_data instance and...
9366 (make_pass_optimize_bswap): ...new function.
9367 (pass_optimize_widening_mul): Convert from a global struct to a
9368 subclass of gimple_opt_pass along with...
9369 (pass_data_optimize_widening_mul): ...new pass_data instance and...
9370 (make_pass_optimize_widening_mul): ...new function.
9371 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
9372 subclass of gimple_opt_pass along with...
9373 (pass_data_phiopt): ...new pass_data instance and...
9374 (make_pass_phiopt): ...new function.
9375 (pass_cselim): Convert from a global struct to a subclass of
9376 gimple_opt_pass along with...
9377 (pass_data_cselim): ...new pass_data instance and...
9378 (make_pass_cselim): ...new function.
9379 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
9380 subclass of gimple_opt_pass along with...
9381 (pass_data_phiprop): ...new pass_data instance and...
9382 (make_pass_phiprop): ...new function.
9383 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
9384 subclass of gimple_opt_pass along with...
9385 (pass_data_pre): ...new pass_data instance and...
9386 (make_pass_pre): ...new function.
9387 (pass_fre): Convert from a global struct to a subclass of
9388 gimple_opt_pass along with...
9389 (pass_data_fre): ...new pass_data instance and...
9390 (make_pass_fre): ...new function.
9391 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
9392 subclass of gimple_opt_pass along with...
9393 (pass_data_reassoc): ...new pass_data instance and...
9394 (make_pass_reassoc): ...new function.
9395 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
9396 subclass of gimple_opt_pass along with...
9397 (pass_data_sink_code): ...new pass_data instance and...
9398 (make_pass_sink_code): ...new function.
9399 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
9400 subclass of gimple_opt_pass along with...
9401 (pass_data_strlen): ...new pass_data instance and...
9402 (make_pass_strlen): ...new function.
9403 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
9404 struct to a subclass of gimple_opt_pass along with...
9405 (pass_data_build_alias): ...new pass_data instance and...
9406 (make_pass_build_alias): ...new function.
9407 (pass_build_ealias): Convert from a global struct to a subclass of
9408 gimple_opt_pass along with...
9409 (pass_data_build_ealias): ...new pass_data instance and...
9410 (make_pass_build_ealias): ...new function.
9411 (pass_ipa_pta): Convert from a global struct to a subclass of
9412 simple_ipa_opt_pass along with...
9413 (pass_data_ipa_pta): ...new pass_data instance and...
9414 (make_pass_ipa_pta): ...new function.
9415 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
9416 subclass of gimple_opt_pass along with...
9417 (pass_data_uncprop): ...new pass_data instance and...
9418 (make_pass_uncprop): ...new function.
9419 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
9420 global struct to a subclass of gimple_opt_pass along with...
9421 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
9422 (make_pass_late_warn_uninitialized): ...new function.
9423 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
9424 to a subclass of gimple_opt_pass along with...
9425 (pass_data_init_datastructures): ...new pass_data instance and...
9426 (make_pass_init_datastructures): ...new function.
9427 (pass_early_warn_uninitialized): Convert from a global struct to a
9428 subclass of gimple_opt_pass along with...
9429 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
9430 (make_pass_early_warn_uninitialized): ...new function.
9431 (pass_update_address_taken): Convert from a global struct to a
9432 subclass of gimple_opt_pass along with...
9433 (pass_data_update_address_taken): ...new pass_data instance and...
9434 (make_pass_update_address_taken): ...new function.
9435 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
9436 struct to a subclass of gimple_opt_pass along with...
9437 (pass_data_release_ssa_names): ...new pass_data instance and...
9438 (make_pass_release_ssa_names): ...new function.
9439 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
9440 subclass of gimple_opt_pass along with...
9441 (pass_data_stdarg): ...new pass_data instance and...
9442 (make_pass_stdarg): ...new function.
9443 * tree-switch-conversion.c (pass_convert_switch): Convert from a
9444 global struct to a subclass of gimple_opt_pass along with...
9445 (pass_data_convert_switch): ...new pass_data instance and...
9446 (make_pass_convert_switch): ...new function.
9447 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
9448 to a subclass of gimple_opt_pass along with...
9449 (pass_data_tail_recursion): ...new pass_data instance and...
9450 (make_pass_tail_recursion): ...new function.
9451 (pass_tail_calls): Convert from a global struct to a subclass of
9452 gimple_opt_pass along with...
9453 (pass_data_tail_calls): ...new pass_data instance and...
9454 (make_pass_tail_calls): ...new function.
9455 * tree-vect-generic.c (pass_lower_vector): Convert from a global
9456 struct to a subclass of gimple_opt_pass along with...
9457 (pass_data_lower_vector): ...new pass_data instance and...
9458 (make_pass_lower_vector): ...new function.
9459 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
9460 gimple_opt_pass along with...
9461 (pass_data_lower_vector_ssa): ...new pass_data instance and...
9462 (make_pass_lower_vector_ssa): ...new function.
9463 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
9464 to a subclass of gimple_opt_pass along with...
9465 (pass_data_slp_vectorize): ...new pass_data instance and...
9466 (make_pass_slp_vectorize): ...new function.
9467 (pass_ipa_increase_alignment): Convert from a global struct to a
9468 subclass of simple_ipa_opt_pass along with...
9469 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
9470 (make_pass_ipa_increase_alignment): ...new function.
9471 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
9472 gimple_opt_pass along with...
9473 (pass_data_vrp): ...new pass_data instance and...
9474 (make_pass_vrp): ...new function.
9475 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
9476 subclass of simple_ipa_opt_pass along with...
9477 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
9478 (make_pass_ipa_free_lang_data): ...new function.
9479 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
9480 gimple_opt_pass along with...
9481 (pass_data_tsan): ...new pass_data instance and...
9482 (make_pass_tsan): ...new function.
9483 (pass_tsan_O0): Convert from a global struct to a subclass of
9484 gimple_opt_pass along with...
9485 (pass_data_tsan_O0): ...new pass_data instance and...
9486 (make_pass_tsan_O0): ...new function.
9487 * var-tracking.c (pass_variable_tracking): Convert from a global
9488 struct to a subclass of rtl_opt_pass along with...
9489 (pass_data_variable_tracking): ...new pass_data instance and...
9490 (make_pass_variable_tracking): ...new function.
9491 * web.c (pass_web): Convert from a global struct to a subclass of
9492 rtl_opt_pass along with...
9493 (pass_data_web): ...new pass_data instance and...
9494 (make_pass_web): ...new function.
9495 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
9496 declaration with that of...
9497 (make_pass_mode_switch_use): ...new function.
9498 (pass_resolve_sw_modes): Replace declaration with that of...
9499 (make_pass_resolve_sw_modes): ...new function.
9500 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
9501 from a global struct to a subclass of rtl_opt_pass along with...
9502 (pass_data_mode_switch_use): ...new pass_data instance and...
9503 (make_pass_mode_switch_use): ...new function.
9504 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
9505 from a global struct to a subclass of rtl_opt_pass along with...
9506 (pass_data_resolve_sw_modes): ...new pass_data instance and...
9507 (make_pass_resolve_sw_modes): ...new function.
9508 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
9509 struct to a subclass of rtl_opt_pass along with...
9510 (pass_data_insert_vzeroupper): ...new pass_data instance and...
9511 (make_pass_insert_vzeroupper): ...new function.
9512 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
9513 global struct to a subclass of rtl_opt_pass along with...
9514 (pass_data_work_around_errata): ...new pass_data instance and...
9515 (make_pass_work_around_errata): ...new function.
9516 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
9517 struct to a subclass of rtl_opt_pass along with...
9518 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
9519 (make_pass_mips_machine_reorg2): ...new function.
9521 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9523 * passes.c (pass_manager::operator new): New.
9525 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9527 Handwritten part of conversion of passes to C++ classes.
9529 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
9530 (toplev.o): Add dep on PASS_MANAGER_H.
9531 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
9532 of early local pases to reflect this moving from a global to a
9533 member of gcc::pass_manager.
9534 (cgraph_add_new_function): Likewise.
9535 * lto-cgraph.c (lto_output_node): Update for conversion of
9536 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
9537 * passes.c (opt_pass::clone): New.
9538 (opt_pass::gate): New.
9539 (opt_pass::execute): New.
9540 (opt_pass::opt_pass): New.
9541 (pass_manager::execute_early_local_passes): New.
9542 (pass_manager::execute_pass_mode_switching): new.
9543 (finish_optimization_passes): Convert to...
9544 (pass_manager::finish_optimization_passes): ...this.
9545 (finish_optimization_passes): Update for conversion of passes to
9547 (register_dump_files_1): Use has_gate since we cannot portably
9548 check a vtable entry against NULL.
9549 (dump_one_pass): Likewise.
9550 (ipa_write_summaries_2): Likewise.
9551 (ipa_write_optimization_summaries_1): Likewise.
9552 (ipa_read_summaries_1): Likewise.
9553 (ipa_read_optimization_summaries_1): Likewise.
9554 (execute_ipa_stmt_fixups): Likewise.
9555 (pass_manager::pass_manager): Rewrite pass-creation, invoking
9556 pass-creation functions rather than wiring up globals, and
9557 storing the results in fields of pass_manager generated using
9559 (pass_manager::dump_profile_report): Update for conversion of
9560 passes to C++ classes.
9561 (pass_manager::execute_ipa_summary_passes): Likewise.
9562 (execute_one_ipa_transform_pass): Likewise.
9563 (execute_one_pass): Use has_gate and has_execute since we cannot
9564 portably check a vtable entry against NULL.
9565 * pass_manager.h (pass_manager::finish_optimization_passes): New.
9566 (pass_manager): Use pass-instances.def to add fields for the
9567 various pass instances.
9568 * toplev.c (finalize): Update for move of
9569 finish_optimization_passes to a method of gcc::pass_manager.
9570 * toplev.h (finish_optimization_passes): Move to method of class
9572 * tree-pass.h (struct pass_data): New.
9573 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
9574 (opt_pass::gate): Convert to virtual function.
9575 (opt_pass::~opt_pass): New.
9576 (opt_pass::clone): New.
9577 (opt_pass::execute): Convert to virtual function.
9578 (opt_pass::opt_pass): New.
9579 (opt_pass::ctxt_): new.
9580 (gimple_opt_pass): Convert to subclass of opt_pass.
9581 (gimple_opt_pass::gimple_opt_pass): New.
9582 (rtl_opt_pass): Convert to subclass of opt_pass.
9583 (rtl_opt_pass::rtl_opt_pass): New.
9584 (ipa_opt_pass_d): Convert to subclass of opt_pass.
9585 (ipa_opt_pass_d::ipa_opt_pass_d): New.
9586 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
9587 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
9588 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
9589 invocation of pass_mode_switching to reflect this moving from a
9590 global to a member of gcc::pass_manager.
9591 (ix86_option_override): Rework how pass_insert_vzeroupper is
9592 added to the pass_manager to reflect autogenerated changes.
9593 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
9595 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
9597 PR rtl-optimization/57708
9598 * recog.c (peep2_find_free_register): Validate all regs in a
9601 2013-08-05 Jan Hubicka <jh@suse.cz>
9604 * cgraph.c (verify_cgraph_node): Accept local flags from other
9606 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
9607 (function_and_variable_visibility): Likewise.
9608 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
9610 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9612 * graph.c (init_graph_slim_pretty_print): Remove.
9613 (print_graph_cfg): Do not call it. Use local pretty printer.
9614 (start_graph_dump): Likewise.
9616 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9618 * gimple-pretty-print.c (buffer): Remove.
9619 (initialized): Likewise.
9620 (maybe_init_pretty_print): Likewise.
9621 (print_gimple_stmt): Do not call it. Use non-static local
9622 pretty_printer variable.
9623 (print_gimple_expr): Likewise.
9624 (print_gimple_seq): Likewise.
9625 (gimple_dump_bb): Likewise.
9627 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
9629 * asan.c (asan_pp): Remove.
9630 (asan_pp_initialized): Likewise.
9631 (asan_pp_initialize): Likewise.
9632 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
9633 (asan_emit_stack_protection): Tidy. Use local pretty printer.
9634 (asan_add_global): Likewise.
9636 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
9638 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
9639 * diagnostic.h (diagnostic_flush_buffer): Adjust.
9640 * pretty-print.c (pp_formatted_text_data): Likewise.
9641 (pp_indent): Rename from pp_base_indent.
9642 (pp_format): Rename from pp_base_format.
9643 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
9644 (pp_format_verbatim): Rename from pp_base_format_verbatim.
9645 (pp_flush): Rename from pp_base_flush.
9646 (pp_set_line_maximum_length): Rename from
9647 pp_base_set_line_maximum_length.
9648 (pp_clear_output_area): Rename from pp_base_clear_output_area.
9649 (pp_set_prefix): Rename from pp_base_set_prefix.
9650 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
9651 (pp_emit_prefix): Rename from pp_base_emit_prefix.
9652 (pp_append_text): Rename from pp_base_append_text.
9653 (pp_formatted_text): Rename from pp_base_formatted_text.
9654 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
9655 (pp_remaining_character_count_for_line): Rename from
9656 pp_base_remaining_character_count_for_line.
9657 (pp_newline): Rename from pp_base_newline.
9658 (pp_character): Rename from pp_base_character.
9659 (pp_string): Rename from pp_base_string.
9660 (pp_maybe_space): Rename from pp_base_maybe_space.
9661 * asan.c (asan_pp_string): Adjust.
9662 (asan_emit_stack_protection): Likewise.
9663 (asan_add_global): Likewise.
9664 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
9665 * tree-mudflap.c (mf_varname_tree): Likewise.
9666 * tree-pretty-print.c (pp_tree_identifier): Rename from
9667 pp_base_tree_identifier.
9668 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
9669 Declare as function.
9671 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
9673 * pretty-print.h (pp_bar_bar): New.
9674 (pp_ampersand_ampersand): Likewise.
9675 (pp_less_equal): Likewise.
9676 (pp_greater_equal): Likewise.
9677 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
9678 printer functions instead of pp_string or operators and punctuators.
9679 (dump_gimple_call): Likewise.
9680 (dump_gimple_omp_for): Likewise.
9681 (dump_gimple_transaction): Likewise.
9682 (dump_gimple_phi): Likewise.
9683 (pp_gimple_stmt_1): Likewise.
9684 * sched-vis.c (print_insn): Likewise.
9685 * tree-mudflap.c (mf_varname_tree): Likewise.
9686 * tree-pretty-print.c (dump_block_node): Likewise.
9687 (dump_generic_node): Likewise.
9689 2013-08-02 Jan Hubicka <jh@suse.cz>
9691 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
9693 * lto-streamer-out.c (tree_is_indexable): Results decls and
9694 parm decls are not indexable.
9695 (DFS_write_tree_body): Do not follow args and results.
9696 (hash_tree): Likewise.
9697 (output_functions): Rearrange so struct function is needed
9698 only when real body is output; be able to also ouptut abstract
9699 functions; output DECL_ARGUMENTS and DECL_RESULT.
9700 (lto_output): When not in WPA, ale store abstract functions.
9701 (write_symbol): Do not care about RESULT_DECL.
9702 (output_symbol_p): Handle correctly sbtract decls.
9703 * lto-streamer-in.c (input_function): Rearrange so struct
9704 function can be NULL at entry; allow streaming of
9705 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
9706 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
9707 sanity check during LTO.
9708 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
9709 RESULT_DECl and DECL_ARGUMENTS.
9710 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
9713 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
9715 * pretty-print.h (pp_underscore): New.
9717 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
9718 printer functions instead of pp_character.
9719 (dump_binary_rhs): Likewise.
9720 (dump_ternary_rhs): Likewise.
9721 (dump_gimple_call_args): Likewise.
9722 (pp_points_to_solution): Likewise.
9723 (dump_gimple_call): Likewise.
9724 (dump_gimple_switch): Likewise.
9725 (dump_gimple_cond): Likewise.
9726 (dump_gimple_bind): Likewise.
9727 (dump_gimple_try): Likewise.
9728 (dump_gimple_omp_for): Likewise.
9729 (dump_gimple_omp_continue): Likewise.
9730 (dump_gimple_omp_single): Likewise.
9731 (dump_gimple_omp_sections): Likewise.
9732 (dump_gimple_omp_block): Likewise.
9733 (dump_gimple_omp_critical): Likewise.
9734 (dump_gimple_transaction): Likewise.
9735 (dump_gimple_asm): Likewise.
9736 (dump_gimple_phi): Likewise.
9737 (dump_gimple_omp_parallel): Likewise.
9738 (dump_gimple_omp_task): Likewise.
9739 (dump_gimple_omp_atomic_load): Likewise.
9740 (dump_gimple_omp_atomic_store): Likewise.
9741 (dump_gimple_mem_ops): Likewise.
9742 (pp_gimple_stmt_1): Likewise.
9743 (pp_cfg_jump): Likewise.
9744 (dump_implicit_edges): Likewise.
9745 (gimple_dump_bb_for_graph): Likewise.
9746 * graph.c (draw_cfg_node): Likewise.
9747 * langhooks.c (lhd_print_error_function): Likewise.
9748 * sched-vis.c (print_exp): Likewise.
9749 (print_value): Likewise.
9750 (print_pattern): Likewise.
9751 (print_insn): Likewise.
9752 (rtl_dump_bb_for_graph): Likewise.
9753 * tree-pretty-print.c (dump_function_declaration): Likewise.
9754 (dump_array_domain): Likewise.
9755 (dump_omp_clause): Likewise.
9756 (dump_location): Likewise.
9757 (dump_generic_node): Likewise.
9758 (print_struct_decl): Likewise.
9759 * diagnostic.c (diagnostic_show_locus): Use pp_space.
9761 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9763 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
9764 candidate table when replacing a candidate statement.
9765 (replace_rhs_if_not_dup): Likewise.
9766 (replace_one_candidate): Likewise.
9768 2013-08-02 Jan Hubicka <jh@suse.cz>
9769 Martin Liska <marxin.liska@gmail.com>
9771 * cgraphunit.c (add_new_function): Fix logic when adding from
9773 (assemble_thunk): Rename to ...
9774 (expand_thunk); .. this one; export; get it working with
9775 general functions; make produced gimple valid.
9776 * cgraph.h (expand_thunk): Declare.
9778 2013-08-02 Jan Hubicka <jh@suse.cz>
9780 * ipa-cp.c (gather_context_independent_values): Use
9781 ipa_get_param_move_cost.
9782 (get_replacement_map): Remove PARAM; move parameter folding
9784 (create_specialized_node): Update.
9785 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
9786 assert that we have gimple body; update move_cost.
9787 (count_formal_params): Assert that we have gimple body.
9788 (ipa_dump_param): New function.
9789 (ipa_alloc_node_params): Break out from ...
9790 (ipa_initialize_node_params): ... here.
9791 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
9792 (ipa_write_node_info): Stream move costs.
9793 (ipa_read_node_info): Read move costs.
9794 (ipa_update_after_lto_read): Do not recompute node params.
9795 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
9796 (ipa_get_param): Check we are not in WPA.
9797 (ipa_get_param_move_cost): New.
9798 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
9799 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
9800 parm numbers to be present.
9802 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
9804 PR rtl-optimization/58048
9805 * lra-constraints.c (process_alt_operands): Don't check asm
9806 operand on register.
9808 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
9810 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
9811 the implied StoreLoad barrier for atomic operations if before.
9813 2013-08-02 Jan Hubicka <jh@suse.cz>
9814 Martin Liska <marxin.liska@gmail.com>
9816 * cgraph.c (cgraph_function_body_availability): Do not check
9818 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
9819 symtab_node_availability): Declare.
9820 * ipa.c (can_replace_by_local_alias): New.
9821 (function_and_variable_visibility): Use it.
9822 * symtab.c (symtab_for_node_and_aliases,
9823 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
9825 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
9827 PR rtl-optimization/57963
9828 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
9829 (lra_constraints): Use them.
9831 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
9833 * config/arm/types.md (define_attr "type"): Add "load_acq"
9835 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
9837 (cortex_a53_store1): Likewise.
9839 2013-08-01 Jan Hubicka <jh@suse.cz>
9841 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
9842 partitions are not needed.
9844 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
9846 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
9847 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
9848 MAYBE_NON_Q_CLASS_P where appropriate.
9850 2013-08-01 Jan Hubicka <jh@suse.cz>
9852 * cgraph.h (release_function_body): Declare.
9853 * tree.c (free_lang_data_in_decl): Free, parameters and return values
9854 of unused delcarations.
9856 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9858 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
9859 RTL form when subtracting a constant.
9861 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9863 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
9864 Generate canonical plus rtx with negated immediate instead of minus
9866 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
9868 2013-08-01 Jan Hubicka <jh@suse.cz>
9870 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
9871 (cgraph_release_function_body): Likewise.
9872 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
9873 * cgraph.h (cgrpah_node): Rename abstract_and_needed
9874 to used_as_abstract_origin.
9875 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
9876 symbols used as abstract origins.
9877 * cgraphunit.c (analyze_functions): Update.
9878 * ipa.c (symtab_remove_unreachable_nodes): Recompute
9879 used_as_abstract_origin.
9880 * tree-inline.c (tree_function_versioning): Update
9881 used_as_abstract_origin; be ready for DECL_RESULT and
9882 DECL_ARGUMENTS to be NULL.
9884 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
9885 for abstract functions.
9886 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
9889 2013-08-01 Jan Hubicka <jh@suse.cz>
9891 * profile.c (compute_value_histograms): Fix thinko.
9893 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9895 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
9896 aarch-common-protos.h to extra_headers.
9897 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9898 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
9899 * config/aarch64/t-aarch64 (aarch-common.o): Define.
9901 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9903 * config/aarch64/aarch64.md (define_attr "type"): Delete.
9904 Include "../arm/types.md". Define "type" attribute for all patterns.
9905 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
9908 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
9910 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
9911 to support power8 load fusion.
9912 (fusion_gpr_mem_load): Likewise.
9914 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
9916 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
9917 declarations for power8 load fusion.
9918 (emit_fusion_gpr_load): Likewise.
9920 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
9921 tuning for power8, turn on fusion mode by default. Turn on sign
9922 extending fusion mode if normal fusion mode is on, and we are at
9924 (fusion_gpr_load_p): New function, return true if we can fuse an
9925 addis instruction with a dependent load to a GPR.
9926 (emit_fusion_gpr_load): Emit the instructions for power8 load
9929 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
9930 (VSX load fusion peepholes): New peepholes to fuse together an
9931 addi instruction with a VSX load instruction.
9933 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
9934 peepholes to fuse an addis instruction with a load to a GPR base
9935 register. If we are supporting sign extending fusions, convert
9936 sign extending loads to zero extending loads and add an explicit
9939 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9941 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
9942 aarch-common-protos.h to extra_headers.
9943 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9944 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
9945 (arm_early_store_addr_dep): Likewise.
9946 (arm_no_early_alu_shift_dep): Likewise.
9947 (arm_no_early_alu_shift_value_dep): Likewise.
9948 (arm_no_early_mul_dep): Likewise.
9949 (arm_no_early_store_addr_dep): Likewise.
9950 (arm_mac_accumulator_is_mul_result): Likewise.
9951 (arm_mac_accumulator_is_result): Likewise.
9952 * config/arm/aarch-common.c: ... here. New file.
9953 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
9955 (arm_early_store_addr_dep): Likewise.
9956 (arm_no_early_alu_shift_dep): Likewise.
9957 (arm_no_early_alu_shift_value_dep): Likewise.
9958 (arm_no_early_mul_dep): Likewise.
9959 (arm_no_early_store_addr_dep): Likewise.
9960 (arm_mac_accumulator_is_mul_result): Likewise.
9961 (arm_mac_accumulator_is_result): Likewise.
9962 * config/arm/aarch-common-protos.h: ... here. New file.
9963 * config/arm/t-arm (aarch-common.o): Define.
9965 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9967 * config/arm/arm.md: Include new file "types.md".
9968 (define_attr "type"): Move from here to ...
9969 (define_attr "mul32"): Likewise.
9970 (define_attr "mul64"): Likewise.
9971 * config/arm/types.md: ... here. New file.
9973 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
9975 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
9976 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
9978 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9980 * gen-pass-instances.awk: Fix offset of substr().
9982 2013-07-31 David Malcolm <dmalcolm@redhat.com>
9984 * Makefile.in (pass-instances.def): New.
9985 (passes.o): Replace dependency on passes.def with one on
9988 * gen-pass-instances.awk: New.
9990 * passes.c (pass_manager::pass_manager): Use pass-instances.def
9991 rather than passes.def, updating local definition of NEXT_PASS
9992 macro to add an extra NUM parameter (currently unused).
9994 2013-07-30 David Malcolm <dmalcolm@redhat.com>
9996 * Makefile.in (PASS_MANAGER_H): New.
9997 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
9998 (passes.o): Likewise.
9999 (statistics.o): Likewise.
10000 (cgraphunit.o): Likewise.
10001 (context.o): Depend on PASS_MANAGER_H.
10003 * pass_manager.h: New.
10005 * cgraphunit.c (cgraph_add_new_function): Update for moves
10006 of globals to fields of pass_manager.
10007 (analyze_function): Likewise.
10008 (expand_function): Likewise.
10009 (ipa_passes): Likewise.
10010 (compile): Likewise.
10012 * context.c (context::context): New.
10013 * context.h (context::context): New.
10014 (context::get_passes): New.
10015 (context::passes_): New.
10017 * lto-cgraph.c (input_node): Update for moves of globals to
10018 fields of pass_manager.
10020 * passes.c (all_passes): Remove, in favor of a field of the
10021 same name within the new class pass_manager.
10022 (all_small_ipa_passes): Likewise.
10023 (all_lowering_passes): Likewise.
10024 (all_regular_ipa_passes): Likewise.
10025 (all_late_ipa_passes): Likewise.
10026 (all_lto_gen_passes): Likewise.
10027 (passes_by_id): Likewise.
10028 (passes_by_id_size): Likewise.
10029 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
10030 the new class pass_manager.
10031 (set_pass_for_id): Convert to...
10032 (pass_manager::set_pass_for_id): ...method.
10033 (get_pass_for_id): Convert to...
10034 (pass_manager::get_pass_for_id): ...method.
10035 (register_one_dump_file): Move body of implementation into...
10036 (pass_manager::register_one_dump_file): ...here.
10037 (register_dump_files_1): Convert to...
10038 (pass_manager::register_dump_files_1): ...method.
10039 (register_dump_files): Convert to...
10040 (pass_manager::register_dump_files): ...method.
10041 (create_pass_tab): Update for moves of globals to fields of
10043 (dump_passes): Move body of implementation into...
10044 (pass_manager::dump_passes): ...here.
10045 (register_pass): Move body of implementation into...
10046 (pass_manager::register_pass): ...here.
10047 (init_optimization_passes): Convert into...
10048 (pass_manager::pass_manager): ...constructor for new
10049 pass_manager class, and initialize the pass_lists array.
10050 (check_profile_consistency): Update for moves of globals to
10051 fields of pass_manager.
10052 (dump_profile_report): Move body of implementation into...
10053 (pass_manager::dump_profile_report): ...here.
10054 (ipa_write_summaries_1): Update for moves of pass lists from
10055 being globals to fields of pass_manager.
10056 (ipa_write_optimization_summaries): Likewise.
10057 (ipa_read_summaries): Likewise.
10058 (ipa_read_optimization_summaries): Likewise.
10059 (execute_all_ipa_stmt_fixups): Likewise.
10061 * statistics.c (statistics_fini): Update for moves of globals to
10062 fields of pass_manager.
10064 * toplev.c (general_init): Replace call to
10065 init_optimization_passes with construction of the pass_manager
10068 * tree-pass.h (all_passes): Remove, in favor of a field of the
10069 same name within the new class pass_manager.
10070 (all_small_ipa_passes): Likewise.
10071 (all_lowering_passes): Likewise.
10072 (all_regular_ipa_passes): Likewise.
10073 (all_lto_gen_passes): Likewise.
10074 (all_late_ipa_passes): Likewise.
10075 (passes_by_id): Likewise.
10076 (passes_by_id_size): Likewise.
10077 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
10078 the new class pass_manager.
10079 (get_pass_for_id): Remove.
10081 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
10083 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
10086 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
10088 * arm.md (mulhi3): New expand pattern.
10090 2013-07-30 Jan Hubicka <jh@suse.cz>
10091 Martin Liska <marxin.liska@gmail.com>
10093 * profile.c (compute_value_histograms): Do not ICE when
10094 there is mismatch only on some counters.
10096 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10098 PR rtl-optimization/57637
10099 * function.c (move_insn_for_shrink_wrap): Also check the
10100 GEN set of the LIVE problem for the liveness analysis
10101 if it exists, otherwise give up.
10103 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10105 PR tree-optimization/57993
10106 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
10107 replaced statement in the candidate table.
10108 (phi_add_costs): Return infinite cost when the hidden basis does
10109 not dominate all phis on which the candidate is dependent.
10110 (replace_one_candidate): Record replaced statement in the
10113 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
10115 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
10116 (ashlv2si3): New expander.
10117 (*ashlv2si3_i): New define_insn_and_split.
10118 * predicates.md (float_operation): Allow patterns with three
10119 basic sub-patterns.
10121 PR rtl-optimization/58021
10122 * mode-switching.c (create_pre_exit): Always split off preceding
10123 insns if we are not at the basic block head.
10125 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
10127 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
10128 (UCLIBC_DYNAMIC_LINKER): New macro.
10129 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
10131 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
10132 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
10134 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
10135 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
10136 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
10137 for SF and DF modes. Use ieee_quad_format for TF mode.
10138 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
10139 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
10140 (mips_option_override): Handle `-mnan=legacy'.
10141 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
10142 `-mabs=2008' and `-mnan=2008'.
10143 (OPTION_DEFAULT_SPECS): Add "nan" default.
10144 (ASM_SPEC): Handle `-mnan='.
10145 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
10146 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
10147 comment accordingly.
10148 (neg<mode>2): Likewise.
10149 * config/mips/mips.opt (mabs, mnan): New options.
10150 * doc/install.texi (Configuration): Document `--with-nan=' option.
10151 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
10153 (MIPS Options): Document them.
10154 * config.gcc <mips*-*-*>: Handle `--with-nan='.
10155 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
10156 * configure: Regenerate.
10157 * config.in: Regenerate.
10159 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
10161 * config/i386/i386.md (float post-reload splitters): Do not check
10162 for subregs of SSE registers.
10164 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
10165 H.J. Lu <hongjiu.lu@intel.com>
10169 * config/i386/i386.md (post-reload splitter
10170 to avoid partial SSE reg dependency stalls): New pattern.
10172 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
10174 * config/s390/s390.md ("movcc"): Swap load and store instructions.
10176 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
10178 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
10179 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
10181 2013-07-26 Cary Coutant <ccoutant@google.com>
10183 * dwarf2out.c (die_checksum_ordered): Don't include template
10184 instantiations in signature.
10185 (is_template_parameter): New function.
10186 (is_template_instantiation): New function.
10187 (generate_skeleton_bottom_up): Don't include template instantiations
10189 (generate_skeleton): Likewise.
10190 (break_out_comdat_types): Move recursive call to break out nested
10192 (prune_unused_types_mark_generic_parms_dies): Call
10193 is_template_parameter.
10195 2013-07-26 Ian Bolton <ian.bolton@arm.com>
10197 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
10198 uses vector registers.
10199 * config/aarch64/iterators.md: Add attributes rtn and vas.
10201 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10202 Richard Earnshaw <richard.earnshaw@arm.com>
10204 * combine.c (simplify_comparison): Re-canonicalize operands
10206 * config/arm/arm.md (movcond_addsi): New splitter.
10208 2013-07-25 Sterling Augustine <saugustine@google.com>
10210 * dwarf2out.c (size_of_pubnames): Move code to...
10211 (include_pubname_in_output): ...here. New.
10212 (want_pubnames): Rearrange.
10213 (output_pubnames): Call include_pubname_in_output. Move assertion.
10215 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
10217 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
10219 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
10222 * doc/extend.texi: Remove obsolete builtins. Fix
10223 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
10225 2013-07-25 Jan Hubicka <jh@suse.cz>
10227 * cgraph.c (release_function_body): Break out from ...
10228 (cgraph_release_function_body): ... this one; also release DECL_RESULT
10229 and DECL_ARGUMENTS.
10230 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
10231 old_tree in the map.
10232 (create_specialized_node): Update.
10233 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
10235 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
10236 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
10237 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
10238 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
10241 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10243 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
10244 addsi3_carryin_alt2_<optab>): Correct output template.
10246 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10248 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
10249 Adjust for arm_restrict_it.
10250 Remove trailing whitespace.
10252 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
10254 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
10255 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
10257 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
10259 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
10261 PR rtl-optimization/57960
10262 * lra-constraints.c (process_alt_operands): Use the right mode
10263 when checking strict_low.
10265 2013-07-25 Jan Hubicka <jh@suse.cz>
10267 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
10268 * cgraph.c (cgraph_remove_node): Do not release function body
10269 when in cgraph streaming.
10270 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
10271 in other partitions are not considered reachable; fix handling of
10274 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10276 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
10278 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10283 * config/arm/arm.md ("*sibcall_insn): Replace use of
10284 Ss with US. Adjust output for v5 and v4t.
10285 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
10287 * config/arm/constraints.md ("Ss"): Rename to US.
10289 2013-07-25 Terry Guo <terry.guo@arm.com>
10291 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
10292 shift_add/shift_sub0/shift_sub1 RTXs.
10294 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
10295 Anton Blanchard <anton@au1.ibm.com>
10297 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
10298 (altivec_vpks<VI_char>ss): Likewise.
10299 (altivec_vpks<VI_char>us): Likewise.
10300 (altivec_vpku<VI_char>us): Likewise.
10301 (altivec_vpku<VI_char>um): Likewise.
10303 2013-07-24 David Malcolm <dmalcolm@redhat.com>
10305 Introduce context class.
10307 * Makefile.in (CONTEXT_H): New.
10308 (OBJS): Add context.o.
10309 (toplev.o): Add CONTEXT_H to dependencies.
10312 * toplev.c (general_init): Create the singleton gcc::context instance.
10318 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
10320 PR rtl-optimization/57968
10321 * mode-switching.c (create_pre_exit): Allow instructions that
10322 don't set a return register to need a non-exit mode.
10324 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10325 Anton Blanchard <anton@au1.ibm.com>
10327 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
10328 operands to vperm for little endian.
10329 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
10330 of lvsl to create the control mask for a vperm for little endian.
10332 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10333 Anton Blanchard <anton@au1.ibm.com>
10335 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
10336 two operands for little-endian.
10338 2013-07-23 Steve Ellcey <sellcey@mips.com>
10340 * config/mips/mips.c (mips_case_values_threshold): New.
10341 (TARGET_CASE_VALUES_THRESHOLD): Define.
10343 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10344 Anton Blanchard <anton@au1.ibm.com>
10346 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
10347 selection of field for vector splat in little endian mode.
10349 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10351 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
10352 expanders to rs6000.md.
10353 (ior<mode>3): Likewise.
10354 (and<mode>3): Likewise.
10355 (one_cmpl<mode>2): Likewise.
10356 (nor<mode>3): Likewise.
10357 (andc<mode>3): Likewise.
10358 (eqv<mode>3): Likewise.
10359 (nand<mode>3): Likewise.
10360 (orc<mode>3): Likewise.
10362 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
10365 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
10366 to split multi-word logical operations.
10367 (rs6000_split_logical_di): Likewise.
10368 (rs6000_split_logical): Likewise.
10370 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
10371 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
10372 and allow TImode operations in 32-bit.
10373 (vsx_and<mode>3_64bit): Likewise.
10374 (vsx_ior<mode>3_32bit): Likewise.
10375 (vsx_ior<mode>3_64bit): Likewise.
10376 (vsx_xor<mode>3_32bit): Likewise.
10377 (vsx_xor<mode>3_64bit): Likewise.
10378 (vsx_one_cmpl<mode>2_32bit): Likewise.
10379 (vsx_one_cmpl<mode>2_64bit): Likewise.
10380 (vsx_nor<mode>3_32bit): Likewise.
10381 (vsx_nor<mode>3_64bit): Likewise.
10382 (vsx_andc<mode>3_32bit): Likewise.
10383 (vsx_andc<mode>3_64bit): Likewise.
10384 (vsx_eqv<mode>3_32bit): Likewise.
10385 (vsx_eqv<mode>3_64bit): Likewise.
10386 (vsx_nand<mode>3_32bit): Likewise.
10387 (vsx_nand<mode>3_64bit): Likewise.
10388 (vsx_orc<mode>3_32bit): Likewise.
10389 (vsx_orc<mode>3_64bit): Likewise.
10391 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
10392 logical types in GPRs.
10394 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
10395 logical insns to rs6000.md, and allow TImode operations in
10397 (altivec_ior<mode>3): Likewise.
10398 (altivec_xor<mode>3): Likewise.
10399 (altivec_one_cmpl<mode>2): Likewise.
10400 (altivec_nor<mode>3): Likewise.
10401 (altivec_andc<mode>3): Likewise.
10403 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
10404 attributes for moving the 128-bit logical operations into
10406 (BOOL_REGS_OUTPUT): Likewise.
10407 (BOOL_REGS_OP1): Likewise.
10408 (BOOL_REGS_OP2): Likewise.
10409 (BOOL_REGS_UNARY): Likewise.
10410 (BOOL_REGS_AND_CR0): Likewise.
10411 (one_cmpl<mode>2): Add support for DI logical operations on
10412 32-bit, splitting the operations to 32-bit.
10413 (anddi3): Likewise.
10414 (iordi3): Likewise.
10415 (xordi3): Likewise.
10416 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
10417 changes to combine the 32/64-bit code, allow logical operations on
10418 TI mode in 32-bit, and to use similar match_operator patterns like
10419 scalar mode uses. Combine the Altivec and VSX code for logical
10420 operations, and move it here.
10421 (ior<mode>3, 128-bit types): Likewise.
10422 (xor<mode>3, 128-bit types): Likewise.
10423 (one_cmpl<mode>3, 128-bit types): Likewise.
10424 (nor<mode>3, 128-bit types): Likewise.
10425 (andc<mode>3, 128-bit types): Likewise.
10426 (eqv<mode>3, 128-bit types): Likewise.
10427 (nand<mode>3, 128-bit types): Likewise.
10428 (orc<mode>3, 128-bit types): Likewise.
10429 (and<mode>3_internal): Likewise.
10430 (bool<mode>3_internal): Likewise.
10431 (boolc<mode>3_internal1): Likewise.
10432 (boolc<mode>3_internal2): Likewise.
10433 (boolcc<mode>3_internal1): Likewise.
10434 (boolcc<mode>3_internal2): Likewise.
10435 (eqv<mode>3_internal1): Likewise.
10436 (eqv<mode>3_internal2): Likewise.
10437 (one_cmpl1<mode>3_internal): Likewise.
10439 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
10441 * config/microblaze/microblaze.c (microblaze_expand_prologue):
10442 Rename flag_stack_usage to flag_stack_usage_info.
10444 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
10446 * config/microblaze/sync.md: New file.
10447 * config/microblaze/microblaze.md: Include sync.md
10448 * config/microblaze/microblaze.c: Add print_operand 'y'.
10449 * config/microblaze/constraints.md: Add memory_contraint
10450 'Q' which is a single register.
10452 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
10454 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
10456 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
10458 * reload.c (find_reloads): Exit loop once we find this operand
10459 cannot be reloaded somehow for this alternative.
10461 * reload.c (find_reloads): Exit loop once we find a hard register.
10463 * rtlanal.c (computed_jump_p): Exit loop once we find label
10466 * i386.c (ix86_pad_returns): Exit loop after setting replace.
10468 * cfgloopmanip.c (remove_path): Exit loop after setting
10471 * gensupport.c (subst_dup): Avoid loop if code is not
10472 MATCH_DUP nor MATCH_OP_DUP.
10474 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
10476 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
10478 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10480 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
10481 true for SP_REGNUM if mode == ptr_mode.
10482 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
10483 with value R0_REGNUM + 31.
10485 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10487 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
10488 pad pointer-typed argument downward.
10490 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10492 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
10493 and __ILP32__ when the ILP32 model is in use.
10495 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10497 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
10498 (aarch64_load_symref_appropriately): In the case of
10499 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
10500 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
10501 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
10502 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
10503 if 'mode' doesn't equal to 'ptr_mode'.
10504 (aarch64_output_mi_thunk): Add an assertion on the alignment of
10505 'vcall_offset'; change to call aarch64_emit_move differently depending
10506 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
10507 to calculate the upper bound of 'vcall_offset'.
10508 (aarch64_cannot_force_const_mem): Change to also return true if
10510 (aarch64_legitimize_reload_address): In the case of large
10511 displacements, add new local variable 'xmode' and an assertion
10512 based on it; change to use 'xmode' to generate the new rtx and
10514 (aarch64_asm_trampoline_template): Change to generate the template
10515 differently depending on TARGET_ILP32 or not; change to use
10516 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
10517 (aarch64_trampoline_size): Removed.
10518 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
10519 and replace immediate literals with it. Change to use 'ptr_mode'
10520 instead of 'DImode' and call convert_memory_address if the mode
10521 of 'fnaddr' doesn't equal to 'ptr_mode'.
10522 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
10524 (aarch64_elf_asm_destructor): Likewise.
10525 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
10526 on TARGET_ILP32 instead of aarch64_trampoline_size.
10527 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
10528 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
10529 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
10530 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
10531 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
10532 (storewb_pair<GPI:mode>_<P:mode>): ... this.
10533 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
10534 depending on the value of 'mode'.
10535 (add_losym_<mode>): New.
10536 (ldr_got_small_<mode>): New, based on ldr_got_small.
10537 (ldr_got_small): Remove.
10538 (ldr_got_small_sidi): New.
10539 * config/aarch64/iterators.md (P): New.
10540 (PTR): Change to 'ptr_mode' in the condition.
10542 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
10544 * config.gcc (aarch64*-*-*): Support --with-abi.
10545 (aarch64*-*-elf): Support --with-multilib-list.
10546 (aarch64*-*-linux*): Likewise.
10547 (supported_defaults): Add abi to aarch64*-*-*.
10548 * configure.ac: Mention AArch64 for --with-multilib-list.
10549 * configure: Re-generated.
10550 * config/aarch64/biarchilp32.h: New file.
10551 * config/aarch64/biarchlp64.h: New file.
10552 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
10554 (MULTILIB_DEFAULTS): Ditto.
10555 (DRIVER_SELF_SPECS): Ditto.
10556 (ASM_SPEC): Update to also substitute -mabi.
10557 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
10558 file whose name depends on -mabi= and -mbig-endian.
10559 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
10561 (POINTER_SIZE): New define.
10562 (POINTERS_EXTEND_UNSIGNED): Ditto.
10563 (enum aarch64_abi_type): New enumeration tag.
10564 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
10565 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
10566 (TARGET_ILP32): New define.
10567 * config/aarch64/aarch64.opt (mabi): New.
10568 (aarch64_abi): New.
10569 (ilp32, lp64): New values for -mabi.
10570 * config/aarch64/t-aarch64 (comma): New define.
10571 (MULTILIB_OPTIONS): Ditto.
10572 (MULTILIB_DIRNAMES): Ditto.
10573 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
10574 * doc/invoke.texi: Document -mabi for AArch64.
10576 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
10578 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
10580 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10581 Anton Blanchard <anton@au1.ibm.com>
10583 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
10584 endianness when selecting field to splat.
10586 2013-07-22 Eric Christopher <echristo@gmail.com>
10588 * dwarf2out.c (die_odr_checksum): New function to use
10589 CHECKSUM_ macros and ULEB128 for DIE tag.
10590 (generate_type_signature): Use.
10592 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
10594 * config.gcc (sparc*-*-*): Accept leon3 processor.
10595 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
10596 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
10597 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
10598 * config/sparc/sparc.opt (enum processor_type): Add leon3.
10599 (mfix-ut699): Adjust comment.
10600 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
10601 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
10602 (CPP_CPU_SPEC): Likewise.
10603 (ASM_CPU_SPEC): Likewise.
10604 * config/sparc/sparc.c (leon3_cost): New constant.
10605 (sparc_option_override): Add leon3 support.
10606 (mem_ref): New function.
10607 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
10608 (sparc_do_work_around_errata): Look into the instruction in the delay
10609 slot and adjust accordingly. Add fix for the data cache nullify issues
10610 of the UT699. Change insertion position for the NOP.
10611 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
10612 (leon3_load): New reservation.
10613 (leon_store): Bump latency to 2.
10614 (grfpu): New automaton.
10615 (grfpu_alu): New unit.
10616 (grfpu_ds): Likewise.
10617 (leon_fp_alu): Adjust.
10618 (leon_fp_mult): Delete.
10619 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
10620 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
10621 * config/sparc/sparc.md (cpu): Add leon3.
10622 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
10623 (swapsi): Likewise.
10624 (atomic_test_and_set): Likewise.
10625 (ldstub): Likewise.
10627 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
10629 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
10630 default for R5900 targets.
10631 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
10632 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
10633 * config/mips/mips.c (mips_option_override): Report an error for
10634 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
10635 for -march=r5900 -mhard-float.
10637 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
10639 * df-problems.c (can_move_insns_across): Exit loop once we
10640 find a non-fixed, non-global register.
10642 * ipa-pure-const.c (propagate_nothrow): Exit loop after
10645 * omega.c (omega_eliminate_red): Break after setting red_found.
10646 (omega_problem_has_red_equations): Similarly after setting found.
10647 (omega_query_variable): Similarly after setting coupled.
10649 2013-07-22 Marek Polacek <polacek@redhat.com>
10651 * gimplify.c: Don't include gimple.h twice.
10653 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10655 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
10656 instead of TARGET_THUMB1.
10657 (Pz): New constraint.
10658 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
10660 (compare_negsi_si): Likewise.
10661 (compare_addsi2_op0): Likewise.
10662 (compare_addsi2_op1): Likewise.
10663 (addsi3_carryin_<optab>): Likewise.
10664 (addsi3_carryin_alt2_<optab>): Likewise.
10665 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
10666 for arm_restrict_it.
10667 (subsi3_carryin): Likewise.
10668 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
10669 (minmax_arithsi): Disable for arm_restrict_it.
10670 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
10671 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
10672 (satsi_<SAT:code>_shift): Likewise.
10673 (arm_shiftsi3): Add alternative for 16-bit encoding.
10674 (arm32_movhf): Disable for arm_restrict_it.
10675 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
10676 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
10678 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
10680 * config/arm/arm.md (attribute "insn"): Delete.
10681 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
10682 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
10683 (not_shiftsi): Update for attribute change.
10684 (not_shiftsi_compare0): Likewise.
10685 (not_shiftsi_compare0_scratch): Likewise.
10686 (arm_one_cmplsi2): Likewise.
10687 (thumb1_one_cmplsi2): Likewise.
10688 (notsi_compare0): Likewise.
10689 (notsi_compare0_scratch): Likewise.
10690 (thumb1_movdi_insn): Likewise.
10691 (arm_movsi_insn): Likewise.
10692 (movhi_insn_arch4): Likewise.
10693 (movhi_bytes): Likewise.
10694 (arm_movqi_insn): Likewise.
10695 (thumb1_movqi_insn): Likewise.
10696 (arm32_movhf): Likewise.
10697 (thumb1_movhf): Likewise.
10698 (arm_movsf_soft_insn): Likewise.
10699 (thumb1_movsf_insn): Likewise.
10700 (thumb_movdf_insn): Likewise.
10701 (movsicc_insn): Likewise.
10702 (movsfcc_soft_insn): Likewise.
10703 (and_scc): Likewise.
10704 (cond_move): Likewise.
10705 (if_move_not): Likewise.
10706 (if_not_move): Likewise.
10707 (if_shift_move): Likewise.
10708 (if_move_shift): Likewise.
10709 (if_shift_shift): Likewise.
10710 (if_not_arith): Likewise.
10711 (if_arith_not): Likewise.
10712 (cond_move_not): Likewise.
10713 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
10714 (neon_mov<mode>): Likewise.
10715 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
10716 (thumb2_movsi_vfp): Likewise.
10717 (movsf_vfp): Likewise.
10718 (thumb2_movsf_vfp): Likewise.
10719 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
10721 (cortexa7_older_only): Likewise.
10722 (cortexa7_younger): Likewise.
10723 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
10724 (1020alu_shift_op): Likewise.
10725 (1020alu_shift_reg_op): Likewise.
10726 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
10727 (alu_shift_op): Likewise.
10728 (alu_shift_reg_op): Likewise.
10729 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
10730 (11_alu_shift_op): Likewise.
10731 (11_alu_shift_reg_op): Likewise.
10732 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
10733 (9_alu_shift_reg_op): Likewise.
10734 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10736 (cortex_a15_alu_shift): Likewise.
10737 (cortex_a15_alu_shift_reg): Likewise.
10738 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
10739 (cortex_a5_alu_shift): Likewise.
10740 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
10742 (cortex_a53_alu_shift): Likewise.
10743 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
10745 (cortex_a7_alu_reg): Likewise.
10746 (cortex_a7_alu_shift): Likewise.
10747 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10748 (cortex_a8_alu_shift): Likewise.
10749 (cortex_a8_alu_shift_reg): Likewise.
10750 (cortex_a8_mov): Likewise.
10751 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10752 (cortex_a9_dp_shift): Likewise.
10753 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
10754 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
10755 (cortex_r4_mov): Likewise.
10756 (cortex_r4_alu_shift): Likewise.
10757 (cortex_r4_alu_shift_reg): Likewise.
10758 * config/arm/fa526.md (526_alu_op): Update for attribute change.
10759 (526_alu_shift_op): Likewise.
10760 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
10761 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
10762 (626te_alu_shift_op): Likewise.
10763 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
10764 (726te_alu_op): Likewise.
10765 (726te_alu_shift_op): Likewise.
10766 (726te_alu_shift_reg_op): Likewise.
10767 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
10768 (mp626_alu_shift_op): Likewise.
10769 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
10770 (pj4_alu_e1_conds): Likewise.
10771 (pj4_alu): Likewise.
10772 (pj4_alu_conds): Likewise.
10773 (pj4_shift): Likewise.
10774 (pj4_shift_conds): Likewise.
10775 (pj4_alu_shift): Likewise.
10776 (pj4_alu_shift_conds): Likewise.
10778 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10780 * config/arm/predicates.md (shiftable_operator_strict_it):
10782 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
10783 Disable cond_exec version for arm_restrict_it.
10784 (thumb2_smaxsi3): Convert to generate cond_exec.
10785 (thumb2_sminsi3): Likewise.
10786 (thumb32_umaxsi3): Likewise.
10787 (thumb2_uminsi3): Likewise.
10788 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
10789 (thumb2_neg_abssi2): Likewise.
10790 (thumb2_mov_scc): Add alternative for 16-bit encoding.
10791 (thumb2_movsicc_insn): Adjust alternatives.
10792 (thumb2_mov_negscc): Disable for arm_restrict_it.
10793 (thumb2_mov_negscc_strict_it): New pattern.
10794 (thumb2_mov_notscc_strict_it): New pattern.
10795 (thumb2_mov_notscc): Disable for arm_restrict_it.
10796 (thumb2_ior_scc): Likewise.
10797 (thumb2_ior_scc_strict_it): New pattern.
10798 (thumb2_cond_move): Adjust for arm_restrict_it.
10799 (thumb2_cond_arith): Disable for arm_restrict_it.
10800 (thumb2_cond_arith_strict_it): New pattern.
10801 (thumb2_cond_sub): Adjust for arm_restrict_it.
10802 (thumb2_movcond): Likewise.
10803 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
10804 (thumb2_zero_extendhisi2_v6): Likewise.
10805 (thumb2_zero_extendqisi2_v6): Likewise.
10806 (orsi_notsi_si): Likewise.
10807 (orsi_not_shiftsi_si): Likewise.
10809 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
10811 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
10812 instruction sequence is 1 byte shorter.
10814 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
10816 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
10817 it is not needed after split.
10819 2013-07-20 Iain Sandoe <iain@codesourcery.com>
10822 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
10823 second label for nonlocal goto receivers. Don't output pic base labels
10824 unless we're producing PIC; mark that action unreachable().
10825 (ix86_save_reg): If the function contains a nonlocal label, save the
10827 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
10828 * config/darwin.c (emitted_pic_label_num): New GTY.
10829 (update_pic_label_number_if_needed): New.
10830 (machopic_output_function_base_name): Adjust for nonlocal receiver
10832 (machopic_should_output_picbase_label): New.
10833 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
10834 (nonlocal_goto_receiver): New insn and split.
10836 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
10838 * config/aarch64/aarch64-builtins.c
10839 (aarch64_fold_builtin): Fold abs in all modes.
10840 * config/aarch64/aarch64-simd-builtins.def
10841 (abs): Enable for all modes.
10842 * config/aarch64/arm_neon.h
10843 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
10844 (vabs_f64): Add missing intrinsic.
10846 2013-07-19 Ian Bolton <ian.bolton@arm.com>
10848 * config/aarch64/arm_neon.h (vabs_s64): New function
10850 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
10853 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
10854 * config/avr/avr.md (adjust_len): Add `round'.
10855 * config/avr/avr-protos.h (avr_out_round): New prototype.
10856 (avr_out_plus): Add `out_label' argument.
10857 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
10858 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
10859 Handle the case where `insn' is just a pattern.
10860 (avr_out_bitop): Handle the case where `insn' is just a pattern.
10861 (avr_out_round): New function.
10862 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
10864 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
10866 * config/microblaze/microblaze.c (microblaze_expand_prologue):
10867 Add check for flag_stack_usage to handle -fstack-usage support
10869 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
10871 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
10872 interaction for new Power8 flags and VSX.
10874 2013-07-18 Sriraman Tallam <tmsriram@google.com>
10876 PR middle-end/57698
10877 * tree-inline.c (expand_call_inline): Emit errors during
10878 early_inlining only if optimization is not turned on.
10880 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10884 * passes.c (init_optimization_passes): Move the construction of
10885 the pass hierarchy into a new passes.def file.
10887 * Makefile.in (passes.o): Add dependency on passes.def.
10889 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10891 * passes.c (init_optimization_passes): Introduce macros for
10892 constructing the tree of passes (INSERT_PASSES_AFTER,
10893 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
10894 TERMINATE_PASS_LIST).
10896 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
10897 Wei Mi <wmi@google.com>
10899 PR rtl-optimization/57878
10900 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
10902 (reload_pseudo_compare_func): Check nregs first for reload
10905 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10907 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
10909 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
10911 * read-rtl.c (validate_const_int): Once an invalid character is
10912 seen, quit the loop.
10914 * gengtype.c (write_roots): Similarly once we find the "deletable"
10915 or "if_marked" option.
10917 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10919 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
10920 "xtab" and "sat". Move value "clz" from here to ...
10921 (attriubte "type"): ... here.
10922 (satsi_<SAT:code>): Delete "insn" attribute.
10923 (satsi_<SAT:code>_shift): Likewise.
10924 (arm_zero_extendqisi2addsi): Likewise.
10925 (arm_extendqisi2addsi): Likewise.
10926 (clzsi2): Update for attribute changes.
10927 (rbitsi2): Likewise.
10928 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
10930 (arm_usatsihi): Likewise.
10931 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10933 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10935 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
10936 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
10937 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
10938 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
10939 in alphabetical order.
10940 (attribute "core_cycles"): Update for attribute changes.
10941 (arm_addsi3): Likewise.
10942 (addsi3_compare0): Likewise.
10943 (addsi3_compare0_scratch): Likewise.
10944 (addsi3_compare_op1): Likewise.
10945 (addsi3_compare_op2): Likewise.
10946 (compare_addsi2_op0): Likewise.
10947 (compare_addsi2_op1): Likewise.
10948 (addsi3_carryin_shift_<optab>): Likewise.
10949 (subsi3_carryin_shift): Likewise.
10950 (rsbsi3_carryin_shift): Likewise.
10951 (arm_subsi3_insn): Likewise.
10952 (subsi3_compare0): Likewise.
10953 (subsi3_compare): Likewise.
10954 (arm_andsi3_insn): Likewise.
10955 (thumb1_andsi3_insn): Likewise.
10956 (andsi3_compare0): Likewise.
10957 (andsi3_compare0_scratch): Likewise.
10958 (zeroextractsi_compare0_scratch
10959 (andsi_not_shiftsi_si): Likewise.
10960 (iorsi3_insn): Likewise.
10961 (iorsi3_compare0): Likewise.
10962 (iorsi3_compare0_scratch): Likewise.
10963 (arm_xorsi3): Likewise.
10964 (thumb1_xorsi3_insn): Likewise.
10965 (xorsi3_compare0): Likewise.
10966 (xorsi3_compare0_scratch): Likewise.
10967 (satsi_<SAT:code>_shift): Likewise.
10969 (arm_shiftsi3): Likewise.
10970 (shiftsi3_compare0): Likewise.
10971 (not_shiftsi): Likewise.
10972 (not_shiftsi_compare0): Likewise.
10973 (not_shiftsi_compare0_scratch): Likewise.
10974 (arm_one_cmplsi2): Likewise.
10975 (thumb_one_complsi2): Likewise.
10976 (notsi_compare0): Likewise.
10977 (notsi_compare0_scratch): Likewise.
10978 (thumb1_zero_extendhisi2): Likewise.
10979 (arm_zero_extendhisi2): Likewise.
10980 (arm_zero_extendhisi2_v6): Likewise.
10981 (arm_zero_extendhisi2addsi): Likewise.
10982 (thumb1_zero_extendqisi2): Likewise.
10983 (thumb1_zero_extendqisi2_v6): Likewise.
10984 (arm_zero_extendqisi2): Likewise.
10985 (arm_zero_extendqisi2_v6): Likewise.
10986 (arm_zero_extendqisi2addsi): Likewise.
10987 (thumb1_extendhisi2): Likewise.
10988 (arm_extendhisi2): Likewise.
10989 (arm_extendhisi2_v6): Likewise.
10990 (arm_extendqisi): Likewise.
10991 (arm_extendqisi_v6): Likewise.
10992 (arm_extendqisi2addsi): Likewise.
10993 (thumb1_extendqisi2): Likewise.
10994 (thumb1_movdi_insn): Likewise.
10995 (arm_movsi_insn): Likewise.
10996 (movsi_compare0): Likewise.
10997 (movhi_insn_arch4): Likewise.
10998 (movhi_bytes): Likewise.
10999 (arm_movqi_insn): Likewise.
11000 (thumb1_movqi_insn): Likewise.
11001 (arm32_movhf): Likewise.
11002 (thumb1_movhf): Likewise.
11003 (arm_movsf_soft_insn): Likewise.
11004 (thumb1_movsf_insn): Likewise.
11005 (movdf_soft_insn): Likewise.
11006 (thumb_movdf_insn): Likewise.
11007 (arm_cmpsi_insn): Likewise.
11008 (cmpsi_shiftsi): Likewise.
11009 (cmpsi_shiftsi_swp): Likewise.
11010 (arm_cmpsi_negshiftsi_si): Likewise.
11011 (movsicc_insn): Likewise.
11012 (movsfcc_soft_insn): Likewise.
11013 (arith_shiftsi): Likewise.
11014 (arith_shiftsi_compare0
11015 (arith_shiftsi_compare0_scratch
11016 (sub_shiftsi): Likewise.
11017 (sub_shiftsi_compare0
11018 (sub_shiftsi_compare0_scratch
11019 (and_scc): Likewise.
11020 (cond_move): Likewise.
11021 (if_plus_move): Likewise.
11022 (if_move_plus): Likewise.
11023 (if_move_not): Likewise.
11024 (if_not_move): Likewise.
11025 (if_shift_move): Likewise.
11026 (if_move_shift): Likewise.
11027 (if_shift_shift): Likewise.
11028 (if_not_arith): Likewise.
11029 (if_arith_not): Likewise.
11030 (cond_move_not): Likewise.
11031 (thumb1_ashlsi3): Set type attribute.
11032 (thumb1_ashrsi3): Likewise.
11033 (thumb1_lshrsi3): Likewise.
11034 (thumb1_rotrsi3): Likewise.
11035 (shiftsi3_compare0_scratch): Likewise.
11036 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
11037 (neon_mov<mode>): Likewise.
11038 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
11040 (thumb2_movsi_insn): Likewise.
11041 (thumb2_cmpsi_neg_shiftsi): Likewise.
11042 (thumb2_extendqisi_v6): Likewise.
11043 (thumb2_zero_extendhisi2_v6): Likewise.
11044 (thumb2_zero_extendqisi2_v6): Likewise.
11045 (thumb2_shiftsi3_short): Likewise.
11046 (thumb2_addsi3_compare0_scratch): Likewise.
11047 (orsi_not_shiftsi_si): Likewise.
11048 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
11049 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
11051 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11052 (1020alu_shift_op): Likewise.
11053 (1020alu_shift_reg_op): Likewise.
11054 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11055 (alu_shift_op): Likewise.
11056 (alu_shift_reg_op): Likewise.
11057 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11058 (11_alu_shift_op): Likewise.
11059 (11_alu_shift_reg_op): Likewise.
11060 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11061 (9_alu_shift_reg_op): Likewise.
11062 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11064 (cortex_a15_alu_shift): Likewise.
11065 (cortex_a15_alu_shift_reg): Likewise.
11066 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
11068 (cortex_a5_alu_shift): Likewise.
11069 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
11071 (cortex_a53_alu_shift): Likewise.
11072 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11074 (cortex_a7_alu_reg): Likewise.
11075 (cortex_a7_alu_shift): Likewise.
11076 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
11078 (cortex_a8_alu_shift): Likewise.
11079 (cortex_a8_alu_shift_reg): Likewise.
11080 (cortex_a8_mov): Likewise.
11081 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11082 (cortex_a9_dp_shift): Likewise.
11083 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
11085 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
11087 (cortex_r4_mov): Likewise.
11088 (cortex_r4_alu_shift): Likewise.
11089 (cortex_r4_alu_shift_reg): Likewise.
11090 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11091 (526_alu_shift_op): Likewise.
11092 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11093 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11094 (626te_alu_shift_op): Likewise.
11095 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
11096 (726te_alu_op): Likewise.
11097 (726te_alu_shift_op): Likewise.
11098 (726te_alu_shift_reg_op): Likewise.
11099 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11100 (mp626_alu_shift_op): Likewise.
11101 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
11102 (pj4_alu_e1_conds): Likewise.
11103 (pj4_alu): Likewise.
11104 (pj4_alu_conds): Likewise.
11105 (pj4_shift): Likewise.
11106 (pj4_shift_conds): Likewise.
11107 (pj4_alu_shift): Likewise.
11108 (pj4_alu_shift_conds): Likewise.
11109 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11111 (cortexa7_older_only): Likewise.
11112 (cortexa7_younger): Likewise.
11114 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11116 * ipa-pure-const.c (generate_summary): Rename to...
11117 (pure_const_generate_summary): ... this.
11119 2013-07-17 Iain Sandoe <iain@codesourcery.com>
11121 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
11123 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
11126 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
11129 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11131 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
11132 enabled without -march=zEC12.
11133 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
11136 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11138 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
11139 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
11140 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
11141 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
11142 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
11143 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
11144 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
11146 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
11147 and ISA_HAS_NMADD3_NMSUB3.
11148 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
11149 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
11150 (nmsub4<mode>, nmsub3<mode>): Likewise.
11151 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
11153 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11155 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
11156 TARGET_MIPS5400 checking.
11158 2013-07-16 Jakub Jelinek <jakub@redhat.com>
11159 Peter Bergner <bergner@vnet.ibm.com>
11161 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
11162 registers in the comment.
11163 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
11164 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
11165 rather than FIRST_PSEUDO_REGISTERS.
11167 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
11169 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
11170 enable extra ISA flags with TARGET_HTM.
11172 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11174 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
11177 2013-07-15 Cong Hou <congh@google.com>
11179 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
11180 in compare function for sorting.
11182 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
11184 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
11185 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
11186 * config/rs6000/rs6000.opt: Add -mhtm option.
11187 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
11188 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
11189 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11190 __HTM__ if the HTM instructions are available.
11191 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
11192 htm_spr_reg_operand): New define_predicates.
11193 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
11194 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
11196 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
11197 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
11198 HTM builtin functions.
11199 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
11200 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
11201 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
11202 (rs6000_builtin_mask_calculate): Likewise.
11203 (rs6000_option_override_internal): Likewise.
11204 (bdesc_htm): Add new HTM builtin support.
11205 (htm_spr_num): New function.
11206 (htm_spr_regno): Likewise.
11207 (rs6000_htm_spr_icode): Likewise.
11208 (htm_expand_builtin): Likewise.
11209 (htm_init_builtins): Likewise.
11210 (rs6000_expand_builtin): Add support for HTM builtin functions.
11211 (rs6000_init_builtins): Likewise.
11212 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
11214 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
11215 (TARGET_HTM, MASK_HTM): Define macros.
11216 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
11217 (FIXED_REGISTERS): Likewise.
11218 (CALL_USED_REGISTERS): Likewise.
11219 (CALL_REALLY_USED_REGISTERS): Likewise.
11220 (REG_ALLOC_ORDER): Likewise.
11221 (enum reg_class): Likewise.
11222 (REG_CLASS_NAMES): Likewise.
11223 (REG_CLASS_CONTENTS): Likewise.
11224 (REGISTER_NAMES): Likewise.
11225 (ADDITIONAL_REGISTER_NAMES): Likewise.
11226 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
11227 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
11228 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
11229 * config/rs6000/htm.md: New file.
11230 * config/rs6000/htmintrin.h: New file.
11231 * config/rs6000/htmxlintrin.h: New file.
11233 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
11235 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
11236 Define SYMBOL_TINY_GOT, update comment.
11237 * config/aarch64/aarch64.c
11238 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
11239 (aarch64_expand_mov_immediate): Likewise.
11240 (aarch64_print_operand): Likewise.
11241 (aarch64_classify_symbol): Likewise.
11242 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
11243 (ldr_got_tiny): Define.
11245 2013-07-13 Tobias Grosser <tobias@grosser.es>
11247 PR tree-optimization/54094
11248 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
11249 scheduling dimension for the parallelism check from the polyhedral
11250 information in the AST.
11251 * graphite-dependences.c (carries_deps): Do not assume the schedule is
11254 2013-07-13 Jason Merrill <jason@redhat.com>
11256 * print-tree.c (debug_vec_tree): Use debug_raw.
11257 (debug_raw (vec<tree, va_gc> &)): New.
11258 (debug_raw (vec<tree, va_gc> *)): New.
11259 * tree.h: Declare them.
11261 2013-07-13 Bin Cheng <bin.cheng@arm.com>
11263 * ifcvt.c (ifcvt_after_combine): New static variable.
11264 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
11266 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
11267 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
11268 rest_of_handle_if_after_reload): Pass new argument for if_convert.
11270 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
11272 * config/mips/mips.c (mips_expand_call): Remove empty statement.
11274 2013-07-12 Michael Matz <matz@suse.de>
11276 PR middle-end/55771
11277 * convert.c (convert_to_real): Reject non-float inner types.
11279 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
11281 * config/aarch64/aarch64-protos.h
11282 (aarch64_simd_immediate_valid_for_move): Remove.
11283 * config/aarch64/aarch64.c (simd_immediate_info): New member.
11284 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
11286 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
11288 2013-07-11 Steve Ellcey <sellcey@mips.com>
11290 * config/mips/mips.c (mips_conditional_register_usage): Do not
11291 use t[0-7] registers in MIPS16 mode when optimizing for size.
11293 2013-07-11 Sriraman Tallam <tmsriram@google.com>
11295 * config/i386/i386.c (dispatch_function_versions): Fix array
11296 indexing of function_version_info to match actual_versions.
11298 2013-07-11 Teresa Johnson <tejohnson@google.com>
11300 * vec.h (struct va_gc): Move release out-of-line.
11301 (va_gc::release): Call ggc_free on released vec.
11303 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
11305 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
11306 Require GOT register as additional operand in UNSPEC.
11307 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
11308 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
11309 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
11310 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
11311 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
11312 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
11313 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
11315 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
11318 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
11319 name seen by assembler/linker rather if available.
11321 2013-07-11 Andreas Schwab <schwab@suse.de>
11323 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
11325 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
11327 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
11329 2013-07-10 Joseph Myers <joseph@codesourcery.com>
11331 * doc/tm.texi.in: Move hook documentation to ....
11332 * target.def: ... here.
11334 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
11335 text on @hook line.
11336 * doc/tm.texi: Regenerate.
11338 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
11341 * doc/invoke.texi: Document -Wconditionally-supported.
11343 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
11346 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
11349 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
11352 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
11353 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
11354 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
11355 Remove duplicate devices.
11356 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
11357 * config/avr/t-multilib: Regenerate.
11358 * config/avr/avr-tables.opt: Regenerate.
11359 * doc/avr-mmcu.texi: Regenerate.
11361 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
11364 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
11366 2013-07-10 Graham Stott <graham.stott@btinternet.com>
11368 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
11369 the cost of MULT when optimizing for size.
11371 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11373 * config/cr16/cr16-protos.h: Don't include target.h.
11375 2013-07-09 Joseph Myers <joseph@codesourcery.com>
11377 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
11378 adjust register size for TDmode and TFmode for VSX registers.
11380 2013-07-08 Kai Tietz <ktietz@redhat.com>
11383 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
11384 hook_bool_const_tree_true.
11386 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11388 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
11389 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
11390 * config/s390/s390.md: Define FPR*_REGNUM constants.
11391 Fix FPR2_REGNUM constant (18 -> 17).
11392 ("*trunc<BFP:mode><DFP_ALL:mode>2")
11393 ("*trunc<DFP_ALL:mode><BFP:mode>2")
11394 ("trunc<BFP:mode><DFP_ALL:mode>2")
11395 ("trunc<DFP_ALL:mode><BFP:mode>2")
11396 ("*extend<BFP:mode><DFP_ALL:mode>2")
11397 ("*extend<DFP_ALL:mode><BFP:mode>2")
11398 ("extend<BFP:mode><DFP_ALL:mode>2")
11399 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
11402 2013-07-08 Graham Stott <graham.stott@btinternet.com>
11404 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
11406 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11408 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
11409 and cfun_fpr_bit_p to cfun_fpr_save_p.
11410 (s390_frame_area, s390_register_info, s390_frame_info)
11411 (s390_emit_prologue, s390_emit_epilogue)
11412 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
11414 * config/s390/s390.h: Define *_REGNUM macros for floating point
11417 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
11419 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
11421 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
11423 PR rtl-optimization/57786
11424 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
11425 and break out of the loop when it is set to false.
11427 2013-07-08 Jakub Jelinek <jakub@redhat.com>
11430 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
11431 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
11432 (const_int 63)) 0)).
11433 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
11434 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
11435 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
11437 PR rtl-optimization/57829
11438 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
11439 mask bits outside of mode are just sign-extension from mode to HWI.
11441 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
11443 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
11444 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
11445 adjust_address instead of change_address to keep info about alignment.
11446 (emit_strmov): Remove.
11447 (emit_memmov): New function.
11448 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
11449 (expand_movmem_epilogue): Likewise and return updated rtx for
11451 (expand_constant_movmem_prologue): Likewise and return updated rtx for
11452 destination and source.
11453 (decide_alignment): Refactor, handle vector_loop.
11454 (ix86_expand_movmem): Likewise.
11455 (ix86_expand_setmem): Likewise.
11456 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
11458 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
11460 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
11461 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
11463 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
11465 * config/i386/sse.md (sse_movlhps): Change alternative 3
11466 of operand 2 to "m".
11468 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
11471 * config/i386/sse.md (iptr): New mode attribute.
11472 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
11473 (<sse>_vm<plusminus_insn><mode>3): Ditto.
11474 (<sse>_vmmul<mode>3): Ditto.
11475 (<sse>_vmdiv<mode>3): Ditto.
11476 (sse_vmrcpv4sf2): Ditto.
11477 (<sse>_vmsqrt<mode>2): Ditto.
11478 (sse_vmrsqrtv4sf2): Ditto.
11479 (<sse>_vm<code><mode>3): Ditto.
11480 (avx_vmcmp<mode>3): Ditto.
11481 (<sse>_vmmaskcmp<mode>3): Ditto.
11482 (<sse>_comi): Ditto.
11483 (<sse>_ucomi): Ditto.
11484 (*xop_vmfrcz_<mode>): Ditto.
11485 (*fmai_fmadd_<mode>): Ditto.
11486 (*fmai_fmsub_<mode>): Ditto.
11487 (*fmai_fnmadd_<mode>): Ditto.
11488 (*fmai_fnmsub_<mode>): Ditto.
11489 (*fma4i_vmfmadd_<mode>): Ditto.
11490 (*fma4i_vmfmsub_<mode>): Ditto.
11491 (*fma4i_vmfnmadd_<mode>): Ditto.
11492 (*fma4i_vmfnmsub_<mode>): Ditto.
11493 (*xop_vmfrcz_<mode>): Ditto.
11494 (sse_cvtps2pi): Ditto.
11495 (sse_cvttps2pi): Ditto.
11496 (sse_cvtss2si): Ditto.
11497 (sse_cvtss2si_2): Ditto.
11498 (sse_cvtss2siq_2): Ditto.
11499 (sse_cvttss2si): Ditto.
11500 (sse_cvttss2siq): Ditto.
11501 (sse_cvtsd2si): Ditto.
11502 (sse_cvtsd2si_2): Ditto.
11503 (sse_cvtsd2siq_2): Ditto.
11504 (sse_cvttsd2si): Ditto.
11505 (sse_cvttsd2siq): Ditto.
11506 (sse_cvtsd2ss): Ditto.
11507 (sse_cvtss2sd): Ditto.
11508 (avx2_pbroadcast<mode>): Ditto.
11509 (avx2_pbroadcast<mode>_1): Ditto.
11510 (*avx_vperm_broadcast_v4sf): Ditto.
11512 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
11513 (sse_movlhps): Ditto.
11514 (sse_storehps): Ditto.
11515 (sse_loadhps): Ditto.
11516 (sse_storelps): Ditto.
11517 (sse_loadlps): Ditto.
11518 (*vec_concatv4sf): Ditto.
11519 (*vec_interleave_highv2df): Ditto.
11520 (*vec_interleave_lowv2df): Ditto.
11521 (*vec_extractv2df_1_sse): Ditto.
11522 (*vec_extractv2df_0_sse): Ditto.
11523 (sse2_storelpd): Ditto.
11524 (sse2_loadlpd): Ditto.
11525 (sse2_movsd): Ditto.
11526 (*vec_concatv4si): Ditto.
11527 (vec_concatv2di): Ditto.
11529 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
11530 for Intel asm dialect.
11531 (mmx_punpcklwd): Ditto.
11532 (mmx_punpckldq): Ditto.
11534 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
11535 for intel assembler dialect.
11537 2013-07-06 Jakub Jelinek <jakub@redhat.com>
11540 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
11541 for BUILT_IN_C{LZ,LRSB}*.
11542 * tree.h (CASE_INT_FN): Add FN##IMAX case.
11543 * tree-vrp.c (extract_range_basic): Handle
11544 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
11545 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
11546 fall thru to code calling set_value*.
11547 * builtins.c (expand_builtin): Remove *IMAX cases.
11548 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
11549 if width is bigger than 2*HWI.
11551 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
11553 PR rtl-optimization/55342
11554 * lra-int.h (lra_subreg_reload_pseudos): New.
11555 * lra.c: Add undoing optional reloads to the block diagram.
11556 (lra_subreg_reload_pseudos): New.
11557 (lra_optional_reload_pseudos): Change comments.
11558 (lra): Init and clear lra_subreg_reload_pseudos. Clear
11559 lra_optional_reload_pseudos after undo transformations.
11560 * lra-assigns.c (pseudo_prefix_title): New.
11561 (lra_setup_reg_renumber): Use it.
11562 (spill_for): Ditto. Check subreg reload pseudos too.
11563 (assign_by_spills): Consider subreg reload pseudos too.
11564 * lra-constraints.c (simplify_operand_subreg): Use
11565 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
11566 (curr_insn_transform): Recognize and do optional reloads.
11567 (undo_optional_reloads): New.
11568 (lra_undo_inheritance): Call undo_optional_reloads.
11570 2013-07-05 Thomas Quinot <quinot@adacore.com>
11572 * tree-complex.c (expand_complex_operations_1): Fix typo.
11574 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
11576 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
11577 (tune_params): New member 'const vec_costs'.
11578 * config/aarch64/aarch64.c (generic_vector_cost): New.
11579 (generic_tunings): New member 'generic_vector_cost'.
11580 (aarch64_builtin_vectorization_cost): New.
11581 (aarch64_add_stmt_cost): New.
11582 (TARGET_VECTORIZE_ADD_STMT_COST): New.
11583 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
11585 2013-07-03 Jakub Jelinek <jakub@redhat.com>
11588 * config/i386/predicates.md (vsib_address_operand): Disallow
11589 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
11591 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
11593 PR middle-end/55030
11594 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
11595 expand_builtin_setjmp_receiver.
11596 (expand_label): Adjust, call expand_builtin_setjmp_receiver
11597 with NULL for the label parameter.
11598 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
11599 the frame-pointer. Adjust comments.
11600 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
11601 only if LABEL is non-NULL.
11603 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
11605 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
11606 (ARM_ABI_AAPCS64): Ditto.
11608 (ARM_DEFAULT_ABI): Ditto.
11610 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
11612 * config/aarch64/aarch64-builtins.c
11613 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
11614 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
11616 * config/aarch64/aarch64-simd.md
11617 (aarch64_ld1<VALL:mode>): New.
11618 (aarch64_st1<VALL:mode>): Likewise.
11619 * config/aarch64/arm_neon.h
11620 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
11622 2013-07-02 Sriraman Tallam <tmsriram@google.com>
11624 * config/i386/i386.c (gate_insert_vzeroupper): Check if
11626 (ix86_option_override_internal):Turn on all -mavx target flags by
11627 default as they are dependent on AVX anyway.
11629 2013-07-02 Cary Coutant <ccoutant@google.com>
11631 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
11632 deterministic hash.
11633 (loc_checksum_ordered): Likewise.
11634 (hash_loc_operands): Remove inline keyword.
11636 2013-07-02 Jakub Jelinek <jakub@redhat.com>
11638 PR tree-optimization/57741
11639 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
11640 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
11641 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
11642 Allow REAL_CST step_exprs if flag_associative_math.
11643 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
11645 2013-07-02 Ian Bolton <ian.bolton@arm.com>
11647 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
11649 2013-07-02 Ian Bolton <ian.bolton@arm.com>
11651 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
11653 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11655 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
11657 (iorsi3_insn): Likewise.
11658 (arm_xorsi3): Likewise.
11660 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
11662 * arm.md (attribute "wtype"): Delete. Move attribute values from here
11664 (attribute "type"): ... here, and prefix with "wmmx_".
11665 (attribute "core_cycles"): Update for attribute changes.
11666 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
11667 (tbcstv4hi): Likewise.
11668 (tbcstv2si): Likewise.
11669 (iwmmxt_iordi3): Likewise.
11670 (iwmmxt_xordi3): Likewise.
11671 (iwmmxt_anddi3): Likewise.
11672 (iwmmxt_nanddi3): Likewise.
11673 (iwmmxt_arm_movdi): Likewise.
11674 (iwmmxt_movsi_insn): Likewise.
11675 (mov<mode>_internal): Likewise.
11676 (and<mode>3_iwmmxt): Likewise.
11677 (ior<mode>3_iwmmxt): Likewise.
11678 (xor<mode>3_iwmmxt): Likewise.
11679 (add<mode>3_iwmmxt): Likewise.
11680 (ssaddv8qi3): Likewise.
11681 (ssaddv4hi3): Likewise.
11682 (ssaddv2si3): Likewise.
11683 (usaddv8qi3): Likewise.
11684 (usaddv4hi3): Likewise.
11685 (usaddv2si3): Likewise.
11686 (sub<mode>3_iwmmxt): Likewise.
11687 (sssubv8qi3): Likewise.
11688 (sssubv4hi3): Likewise.
11689 (sssubv2si3): Likewise.
11690 (ussubv8qi3): Likewise.
11691 (ussubv4hi3): Likewise.
11692 (ussubv2si3): Likewise.
11693 (mulv4hi3_iwmmxt): Likewise.
11694 (smulv4hi3_highpart): Likewise.
11695 (umulv4hi3_highpart): Likewise.
11696 (iwmmxt_wmacs): Likewise.
11697 (iwmmxt_wmacsz): Likewise.
11698 (iwmmxt_wmacu): Likewise.
11699 (iwmmxt_wmacuz): Likewise.
11700 (iwmmxt_clrdi): Likewise.
11701 (iwmmxt_clrv8qi): Likewise.
11702 (iwmmxt_clr4hi): Likewise.
11703 (iwmmxt_clr2si): Likewise.
11704 (iwmmxt_uavgrndv8qi3): Likewise.
11705 (iwmmxt_uavgrndv4hi3): Likewise.
11706 (iwmmxt_uavgv8qi3): Likewise.
11707 (iwmmxt_uavgv4hi3): Likewise.
11708 (iwmmxt_tinsrb): Likewise.
11709 (iwmmxt_tinsrh): Likewise.
11710 (iwmmxt_tinsrw): Likewise.
11711 (iwmmxt_textrmub): Likewise.
11712 (iwmmxt_textrmsb): Likewise.
11713 (iwmmxt_textrmuh): Likewise.
11714 (iwmmxt_textrmsh): Likewise.
11715 (iwmmxt_textrmw): Likewise.
11716 (iwmxxt_wshufh): Likewise.
11717 (eqv8qi3): Likewise.
11718 (eqv4hi3): Likewise.
11719 (eqv2si3): Likewise.
11720 (gtuv8qi3): Likewise.
11721 (gtuv4hi3): Likewise.
11722 (gtuv2si3): Likewise.
11723 (gtv8qi3): Likewise.
11724 (gtv4hi3): Likewise.
11725 (gtv2si3): Likewise.
11726 (smax<mode>3_iwmmxt): Likewise.
11727 (umax<mode>3_iwmmxt): Likewise.
11728 (smin<mode>3_iwmmxt): Likewise.
11729 (umin<mode>3_iwmmxt): Likewise.
11730 (iwmmxt_wpackhss): Likewise.
11731 (iwmmxt_wpackwss): Likewise.
11732 (iwmmxt_wpackdss): Likewise.
11733 (iwmmxt_wpackhus): Likewise.
11734 (iwmmxt_wpackwus): Likewise.
11735 (iwmmxt_wpackdus): Likewise.
11736 (iwmmxt_wunpckihb): Likewise.
11737 (iwmmxt_wunpckihh): Likewise.
11738 (iwmmxt_wunpckihw): Likewise.
11739 (iwmmxt_wunpckilb): Likewise.
11740 (iwmmxt_wunpckilh): Likewise.
11741 (iwmmxt_wunpckilw): Likewise.
11742 (iwmmxt_wunpckehub): Likewise.
11743 (iwmmxt_wunpckehuh): Likewise.
11744 (iwmmxt_wunpckehuw): Likewise.
11745 (iwmmxt_wunpckehsb): Likewise.
11746 (iwmmxt_wunpckehsh): Likewise.
11747 (iwmmxt_wunpckehsw): Likewise.
11748 (iwmmxt_wunpckelub): Likewise.
11749 (iwmmxt_wunpckeluh): Likewise.
11750 (iwmmxt_wunpckeluw): Likewise.
11751 (iwmmxt_wunpckelsb): Likewise.
11752 (iwmmxt_wunpckelsh): Likewise.
11753 (iwmmxt_wunpckelsw): Likewise.
11754 (ror<mode>3): Likewise.
11755 (ashr<mode>3_iwmmxt): Likewise.
11756 (lshr<mode>3_iwmmxt): Likewise.
11757 (ashl<mode>3_iwmmxt): Likewise.
11758 (ror<mode>3_di): Likewise.
11759 (ashr<mode>3_di): Likewise.
11760 (lshr<mode>3_di): Likewise.
11761 (ashl<mode>3_di): Likewise.
11762 (iwmmxt_wmadds): Likewise.
11763 (iwmmxt_wmaddu): Likewise.
11764 (iwmmxt_tmia): Likewise.
11765 (iwmmxt_tmiaph): Likewise.
11766 (iwmmxt_tmiabb): Likewise.
11767 (iwmmxt_tmiatb): Likewise.
11768 (iwmmxt_tmiabt): Likewise.
11769 (iwmmxt_tmiatt): Likewise.
11770 (iwmmxt_tmovmskb): Likewise.
11771 (iwmmxt_tmovmskh): Likewise.
11772 (iwmmxt_tmovmskw): Likewise.
11773 (iwmmxt_waccb): Likewise.
11774 (iwmmxt_wacch): Likewise.
11775 (iwmmxt_waccw): Likewise.
11776 (iwmmxt_waligni): Likewise.
11777 (iwmmxt_walignr): Likewise.
11778 (iwmmxt_walignr0): Likewise.
11779 (iwmmxt_walignr1): Likewise.
11780 (iwmmxt_walignr2): Likewise.
11781 (iwmmxt_walignr3): Likewise.
11782 (iwmmxt_wsadb): Likewise.
11783 (iwmmxt_wsadh): Likewise.
11784 (iwmmxt_wsadbz): Likewise.
11785 (iwmmxt_wsadhz): Likewise.
11786 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
11787 (iwmmxt_wabsdiffb): Likewise.
11788 (iwmmxt_wabsdiffh): Likewise.
11789 (iwmmxt_wabsdiffw): Likewise.
11790 (iwmmxt_waddsubhx): Likewise
11791 (iwmmxt_wsubaddhx): Likewise.
11792 (addc<mode>3): Likewise.
11793 (iwmmxt_avg4): Likewise.
11794 (iwmmxt_avg4r): Likewise.
11795 (iwmmxt_wmaddsx): Likewise.
11796 (iwmmxt_wmaddux): Likewise.
11797 (iwmmxt_wmaddsn): Likewise.
11798 (iwmmxt_wmaddun): Likewise.
11799 (iwmmxt_wmulwsm): Likewise.
11800 (iwmmxt_wmulwum): Likewise.
11801 (iwmmxt_wmulsmr): Likewise.
11802 (iwmmxt_wmulumr): Likewise.
11803 (iwmmxt_wmulwsmr): Likewise.
11804 (iwmmxt_wmulwumr): Likewise.
11805 (iwmmxt_wmulwl): Likewise.
11806 (iwmmxt_wqmulm): Likewise.
11807 (iwmmxt_wqmulwm): Likewise.
11808 (iwmmxt_wqmulmr): Likewise.
11809 (iwmmxt_wqmulwmr): Likewise.
11810 (iwmmxt_waddbhusm): Likewise.
11811 (iwmmxt_waddbhusl): Likewise.
11812 (iwmmxt_wqmiabb): Likewise.
11813 (iwmmxt_wqmiabt): Likewise.
11814 (iwmmxt_wqmiatb): Likewise.
11815 (iwmmxt_wqmiatt): Likewise.
11816 (iwmmxt_wqmiabbn): Likewise.
11817 (iwmmxt_wqmiabtn): Likewise.
11818 (iwmmxt_wqmiatbn): Likewise.
11819 (iwmmxt_wqmiattn): Likewise.
11820 (iwmmxt_wmiabb): Likewise.
11821 (iwmmxt_wmiabt): Likewise.
11822 (iwmmxt_wmiatb): Likewise.
11823 (iwmmxt_wmiatt): Likewise.
11824 (iwmmxt_wmiabbn): Likewise.
11825 (iwmmxt_wmiabtn): Likewise.
11826 (iwmmxt_wmiatbn): Likewise.
11827 (iwmmxt_wmiattn): Likewise.
11828 (iwmmxt_wmiawbb): Likewise.
11829 (iwmmxt_wmiawbt): Likewise.
11830 (iwmmxt_wmiawtb): Likewise.
11831 (iwmmxt_wmiawtt): Likewise.
11832 (iwmmxt_wmiawbbn): Likewise.
11833 (iwmmxt_wmiawbtn): Likewise.
11834 (iwmmxt_wmiawtbn): Likewise.
11835 (iwmmxt_wmiawttn): Likewise.
11836 (iwmmxt_wmerge): Likewise.
11837 (iwmmxt_tandc<mode>3): Likewise.
11838 (iwmmxt_torc<mode>3): Likewise.
11839 (iwmmxt_torvsc<mode>3): Likewise.
11840 (iwmmxt_textrc<mode>3): Likewise.
11841 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
11842 (wmmxt_pack): Likewise.
11843 (wmmxt_mult_c1): Likewise.
11844 (wmmxt_mult_c2): Likewise.
11845 (wmmxt_alu_c1): Likewise.
11846 (wmmxt_alu_c2): Likewise.
11847 (wmmxt_alu_c3): Likewise.
11848 (wmmxt_transfer_c1): Likewise.
11849 (wmmxt_transfer_c2): Likewise.
11850 (wmmxt_transfer_c3): Likewise.
11851 (marvell_f_iwmmxt_wstr): Likewise.
11852 (marvell_f_iwmmxt_wldr): Likewise.
11854 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
11856 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
11858 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11861 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11862 * lra-constraints.c (need_for_split_p): Check call used hard regs
11863 living through calls.
11865 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
11866 call used regs for call insn.
11868 2013-06-28 Jakub Jelinek <jakub@redhat.com>
11871 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
11872 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
11874 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11876 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
11878 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11880 * lra-constraints.c (need_for_split_p): Check call used hard regs
11881 living through calls.
11883 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
11886 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
11887 to tie with any other modes. Eliminate Altivec vector mode tests,
11888 since these are a subset of ALTIVEC or VSX vector modes. Simplify
11889 code, to return 0 if testing MODE2 for a condition, if we've
11890 already tested MODE1 for the same condition.
11892 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11894 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
11897 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11899 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
11900 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
11902 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11904 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
11906 (aarch64_symbolic_constant_p): Remove.
11907 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
11908 static. Fix line length and white space.
11909 (aarch64_symbolic_constant_p): Remove.
11910 * config/aarch64/predicates.md (aarch64_valid_symref):
11911 Use aarch64_classify_symbol_expression.
11913 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11915 * config/arm/constraints.md (Ts): New constraint.
11916 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
11918 (compare_scc): Use "Ts" constraint for operand 0.
11919 (ior_scc_scc): Likewise.
11920 (and_scc_scc): Likewise.
11921 (and_scc_scc_nodom): Likewise.
11922 (ior_scc_scc_cmp): Likewise for operand 7.
11923 (and_scc_scc_cmp): Likewise.
11924 * config/arm/thumb2.md (thumb2_movsi_insn):
11925 Add alternatives for 16-bit encodings.
11926 (thumb2_movhi_insn): Likewise.
11927 (thumb2_movsicc_insn): Likewise.
11928 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
11929 (thumb2_negscc): Use "Ts" constraint.
11930 Move mvn instruction outside cond_exec block.
11931 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
11932 for 16-bit encodings.
11934 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11936 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
11938 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
11939 (mulsi3subsi): Likewise.
11940 (mulsidi3adddi): Likewise.
11941 (mulsidi3_v6): Likewise.
11942 (umulsidi3_v6): Likewise.
11943 (umulsidi3adddi_v6): Likewise.
11944 (smulsi3_highpart_v6): Likewise.
11945 (umulsi3_highpart_v6): Likewise.
11946 (mulhisi3tb): Likewise.
11947 (mulhisi3bt): Likewise.
11948 (mulhisi3tt): Likewise.
11949 (maddhisi4): Likewise.
11950 (maddhisi4tb): Likewise.
11951 (maddhisi4tt): Likewise.
11952 (maddhidi4): Likewise.
11953 (maddhidi4tb): Likewise.
11954 (maddhidi4tt): Likewise.
11955 (zeroextractsi_compare0_scratch): Likewise.
11956 (insv_zero): Likewise.
11957 (insv_t2): Likewise.
11958 (anddi_notzesidi_di): Likewise.
11959 (anddi_notsesidi_di): Likewise.
11960 (andsi_notsi_si): Likewise.
11961 (iordi_zesidi_di): Likewise.
11962 (xordi_zesidi_di): Likewise.
11963 (andsi_iorsi3_notsi): Likewise.
11964 (smax_0): Likewise.
11965 (smax_m1): Likewise.
11966 (smin_0): Likewise.
11967 (not_shiftsi): Likewise.
11968 (unaligned_loadsi): Likewise.
11969 (unaligned_loadhis): Likewise.
11970 (unaligned_loadhiu): Likewise.
11971 (unaligned_storesi): Likewise.
11972 (unaligned_storehi): Likewise.
11973 (extv_reg): Likewise.
11974 (extzv_t2): Likewise.
11975 (divsi3): Likewise.
11976 (udivsi3): Likewise.
11977 (arm_zero_extendhisi2addsi): Likewise.
11978 (arm_zero_extendqisi2addsi): Likewise.
11979 (compareqi_eq0): Likewise.
11980 (arm_extendhisi2_v6): Likewise.
11981 (arm_extendqisi2addsi): Likewise.
11982 (arm_movt): Likewise.
11983 (thumb2_ldrd): Likewise.
11984 (thumb2_ldrd_base): Likewise.
11985 (thumb2_ldrd_base_neg): Likewise.
11986 (thumb2_strd): Likewise.
11987 (thumb2_strd_base): Likewise.
11988 (thumb2_strd_base_neg): Likewise.
11989 (arm_negsi2): Add alternative for 16-bit encoding.
11990 (arm_one_cmplsi2): Likewise.
11992 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11994 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
11995 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
11996 (movdfcc): Likewise.
11997 * config/arm/vfp.md (*thumb2_movsf_vfp):
11998 Disable predication for arm_restrict_it.
11999 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
12000 (*thumb2_movdfcc_vfp): Likewise.
12001 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
12002 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
12003 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
12004 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
12005 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
12006 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
12007 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
12008 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
12009 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
12010 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
12011 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
12012 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
12013 Disable predication for arm_restrict_it.
12015 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
12017 * config/i386/bmiintrin.h (_bextr_u32): New.
12018 (_bextr_u64): Ditto.
12020 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
12022 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
12023 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
12024 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
12025 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
12026 * config/mips/n32-elf.h: ...this new file.
12028 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
12031 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
12032 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
12034 2013-06-27 Catherine Moore <clm@codesourcery.com>
12036 * config/mips/mips-tables.opt: Regenerate.
12037 * config/mips/mips-cpus.def: Add m14ke and m14kec.
12038 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
12039 * doc/invoke.texi: Add -m14kc.
12041 2013-06-27 Jakub Jelinek <jakub@redhat.com>
12044 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
12045 constraints of operand 1 and 2.
12048 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
12049 to match RTL canonicalization. Swap predicates and
12050 constraints of operand 1 and 2.
12052 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
12054 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
12055 Process OP_INOUT regs for splitting too.
12057 2013-06-27 Jakub Jelinek <jakub@redhat.com>
12059 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
12060 decl before the loop, initialize to NULL.
12061 (vectorizable_load): Initialize ptr_incr to NULL.
12063 2013-06-27 Martin Jambor <mjambor@suse.cz>
12066 * ipa-ref.h (ipa_maybe_record_reference): Declare.
12067 * ipa-ref.c (ipa_maybe_record_reference): New function.
12068 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
12069 * ipa-cp.c (create_specialized_node): Record potential references from
12071 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
12073 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12075 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
12076 parameter 'mode' of type 'enum machine_mode mode'; change to pass
12077 'mode' to force_reg.
12078 (aarch64_add_offset): Update calls to aarch64_force_temporary.
12079 (aarch64_expand_mov_immediate): Likewise.
12081 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12083 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
12084 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
12086 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12088 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
12089 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12090 (struct machine_function): Add tbegin_p.
12091 (s390_canonicalize_comparison): Fold CC mode compares to
12092 conditional jump if possible.
12093 (s390_emit_jump): Return the emitted jump.
12094 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
12095 Handle CCRAWmode compares.
12096 (s390_option_override): Default to -mhtm if available.
12097 (s390_reg_clobbered_rtx): Handle floating point regs as well.
12098 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
12099 FPRs instead of df_regs_ever_live_p.
12100 (s390_optimize_nonescaping_tx): New function.
12101 (s390_init_frame_layout): Extend clobbered_regs array to cover
12103 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
12104 (s390_expand_tbegin): New function.
12105 (enum s390_builtin): New enum definition.
12106 (code_for_builtin): New array definition.
12107 (s390_init_builtins): New function.
12108 (s390_expand_builtin): New function.
12109 (TARGET_INIT_BUILTINS): Define.
12110 (TARGET_EXPAND_BUILTIN): Define.
12111 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
12112 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
12113 (s390_alc_comparison): Likewise.
12114 * config/s390/s390-modes.def: Add CCRAWmode.
12115 * config/s390/s390.h (processor_flags): Add PF_TX.
12116 (TARGET_CPU_HTM): Define macro.
12117 (TARGET_HTM): Define macro.
12118 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
12119 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
12120 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12121 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
12122 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
12123 (TBEGIN_MASK, TBEGINC_MASK): New constants.
12124 ("*cc_to_int"): Move up.
12125 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
12126 constants other than 0.
12127 ("*ccraw_to_int"): New insn and splitter definition.
12128 ("tbegin", "tbegin_nofloat", "tbegin_retry")
12129 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
12130 ("tx_assist"): New expander.
12131 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
12132 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
12133 * config/s390/s390.opt: Add -mhtm option.
12134 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
12135 * config/s390/htmxlintrin.h: New file.
12136 * config/s390/htmintrin.h: New file.
12137 * config/s390/s390intrin.h: New file.
12138 * doc/extend.texi: Document htm builtins.
12139 * config.gcc: Add the new header files to extra_headers.
12141 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
12143 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
12144 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
12146 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12147 Pat Haugen <pthaugen@us.ibm.com>
12148 Peter Bergner <bergner@vnet.ibm.com>
12150 * config/rs6000/power8.md: New.
12151 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
12152 setting for power8 entry.
12153 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
12154 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
12155 test for Power4/Power5 only.
12156 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
12158 (force_new_group): Adjust comment.
12159 * config/rs6000/rs6000.md: Include power8.md.
12161 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
12163 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
12164 * config/arm/arm-protos.h (arm_max_conditional_execute): New
12166 (tune_params): Update comment.
12167 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
12168 (arm_max_conditional_execute): New function.
12169 (thumb2_final_prescan_insn): Use max_insn_skipped and
12170 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
12172 2013-06-25 Jakub Jelinek <jakub@redhat.com>
12174 PR tree-optimization/57705
12175 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
12176 SSA_NAME step, provided that it is not defined inside the loop.
12177 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
12178 (get_initial_def_for_induction): Handle SSA_NAME IV step.
12180 2013-06-25 Martin Jambor <mjambor@suse.cz>
12182 PR middle-end/57670
12183 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
12184 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
12186 (ipa_note_param_call): Initialize member_ptr flag.
12187 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
12188 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
12189 (ipa_write_indirect_edge_info): Stream member_ptr flag.
12190 (ipa_read_indirect_edge_info): Likewise.
12192 2013-06-25 Richard Biener <rguenther@suse.de>
12194 PR middle-end/56977
12195 * passes.c (init_optimization_passes): Move pass_fold_builtins
12196 and pass_dce earlier with -Og.
12198 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
12200 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
12201 <BIT_FIELD_REF>: Remove trailing TAB.
12202 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
12205 2013-06-24 Martin Jambor <mjambor@suse.cz>
12207 PR tree-optimization/57358
12208 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
12209 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
12210 (ipa_analyze_params_uses): Generate pessimistic info when true.
12212 2013-06-24 Martin Jambor <mjambor@suse.cz>
12214 PR tree-optimization/57539
12215 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
12216 global.inlined_to of the new node to it. All callers changed.
12217 * ipa-inline-transform.c (clone_inlined_nodes): New variable
12218 inlining_into, pass it to cgraph_clone_node.
12219 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
12220 ipa_free_edge_args_substructures.
12221 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
12222 rdesc linked list. Do not assert rdesc edges have inlined caller.
12223 Assert we have found an rdesc in the rdesc list.
12225 2013-06-24 Richard Biener <rguenther@suse.de>
12227 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
12228 (pointer_set_lookup): Declare.
12229 (class pointer_map): New template class implementing a
12230 generic pointer to T map.
12231 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
12232 pointer_map<T>::contains, pointer_map<T>::insert,
12233 pointer_map<T>::traverse): New functions.
12234 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
12235 (pointer_set_lookup): New function.
12236 (pointer_set_contains): Use pointer_set_lookup.
12237 (pointer_set_insert): Likewise.
12238 (insert_aux): Remove.
12239 (struct pointer_map_t): Embed a pointer_set_t.
12240 (pointer_map_create): Adjust.
12241 (pointer_map_destroy): Likewise.
12242 (pointer_map_contains): Likewise.
12243 (pointer_map_insert): Likewise.
12244 (pointer_map_traverse): Likewise.
12245 * tree-streamer.h (struct streamer_tree_cache_d): Use a
12246 pointer_map<unsigned> instead of a pointer_map_t.
12247 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
12248 (streamer_tree_cache_lookup): Likewise.
12249 (streamer_tree_cache_create): Likewise.
12250 (streamer_tree_cache_delete): Likewise.
12251 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
12252 pointer_map<unsigned> instead of a pointer_map_t.
12253 (lto_init_tree_ref_encoder): Adjust.
12254 (lto_destroy_tree_ref_encoder): Likewise.
12255 * lto-section-out.c (lto_output_decl_index): Likewise.
12256 (lto_record_function_out_decl_state): Likewise.
12257 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
12259 2013-06-24 Richard Biener <rguenther@suse.de>
12261 PR tree-optimization/57488
12262 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
12264 2013-06-24 Alan Modra <amodra@gmail.com>
12266 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
12267 (gen_easy_altivec_constant): Likewise.
12268 * config/rs6000/predicates.md (easy_vector_constant_add_self,
12269 easy_vector_constant_msb): Likewise.
12271 2013-06-23 Jakub Jelinek <jakub@redhat.com>
12274 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
12275 add missing return true.
12277 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
12280 * config/sh/predicates.md (general_extend_operand): Invoke
12281 general_movsrc_operand for memory operands.
12282 (general_movsrc_operand): Allow reg+reg addressing, do not use
12283 general_operand for memory operands.
12285 2013-06-23 Sriraman Tallam <tmsriram@google.com>
12287 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
12288 when current target options does not apply.
12289 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
12290 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
12291 * config/i386/bmiintrin.h: Pass appropriate target
12292 attributes to header.
12293 * config/i386/mmintrin.h: Ditto.
12294 * config/i386/nmmintrin.h: Ditto.
12295 * config/i386/avx2intrin.h: Ditto.
12296 * config/i386/fxsrintrin.h: Ditto.
12297 * config/i386/tbmintrin.h: Ditto.
12298 * config/i386/xsaveintrin.h: Ditto.
12299 * config/i386/f16cintrin.h: Ditto.
12300 * config/i386/xtestintrin.h: Ditto.
12301 * config/i386/xsaveoptintrin.h: Ditto.
12302 * config/i386/bmi2intrin.h: Ditto.
12303 * config/i386/lzcntintrin.h: Ditto.
12304 * config/i386/smmintrin.h: Ditto.
12305 * config/i386/wmmintrin.h: Ditto.
12306 * config/i386/x86intrin.h: Remove all header include guards.
12307 * config/i386/prfchwintrin.h: Ditto.
12308 * config/i386/pmmintrin.h: Ditto.
12309 * config/i386/tmmintrin.h: Ditto.
12310 * config/i386/xmmintrin.h: Ditto.
12311 * config/i386/popcntintrin.h: Ditto.
12312 * config/i386/rdseedintrin.h: Ditto.
12313 * config/i386/ammintrin.h: Ditto.
12314 * config/i386/emmintrin.h: Ditto.
12315 * config/i386/immintrin.h: Remove all header include guards.
12316 * config/i386/fma4intrin.h: Ditto.
12317 * config/i386/lwpintrin.h: Ditto.
12318 * config/i386/xopintrin.h: Ditto.
12319 * config/i386/ia32intrin.h: Ditto.
12320 * config/i386/avxintrin.h: Ditto.
12321 * config/i386/rtmintrin.h: Ditto.
12322 * config/i386/fmaintrin.h: Ditto.
12323 * config/i386/mm3dnow.h: Ditto.
12325 2013-06-22 Sriraman Tallam <tmsriram@google.com>
12327 * common/config/i386/i386-common.c: Handle LZCNT.
12329 2013-06-22 Andi Kleen <ak@linux.intel.com>
12331 * doc/extend.texi: Use __atomic_store_n instead of
12332 __atomic_store in HLE example.
12334 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
12336 * config/sh/sh.c: Remove <cstdlib> workaround.
12338 2013-06-21 Andi Kleen <ak@linux.intel.com>
12340 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
12342 2013-06-21 Andi Kleen <ak@linux.intel.com>
12344 * doc/extend.texi: Document that __atomic_clear and
12345 __atomic_test_and_set should only be used with bool.
12347 2013-06-20 Jan Hubicka <jh@suse.cz>
12349 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
12350 types_same_for_odr.
12351 * tree.c (decls_same_for_odr): New function.
12352 (same_for_edr): New function.
12353 (types_same_for_odr): New function.
12354 (get_binfo_at_offset): Use it.
12355 * tree.h (types_same_for_odr): Declare.
12357 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
12358 Jason Merrill <jason@redhat.com>
12360 * system.h: Include <cstdlib> as well as <stdlib.h>.
12362 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
12365 * config/i386/i386.c (construct_container): Report error if
12366 long double is used with disabled x87 float returns.
12368 2013-06-20 Jan Hubicka <jh@suse.cz>
12370 * lto-cgraph.c (input_symtab): Do not set cgraph state.
12372 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
12374 PR rtl-optimization/57425
12375 PR rtl-optimization/57569
12376 * alias.c (write_dependence_p): Remove parameters mem_mode and
12377 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
12378 Changed all callers.
12379 (canon_anti_dependence): Get comments and semantics in sync.
12380 Add parameter mem_canonicalized. Changed all callers.
12381 * rtl.h (canon_anti_dependence): Update prototype.
12383 2013-06-20 Richard Biener <rguenther@suse.de>
12385 * data-streamer-in.c (streamer_read_uhwi): Optimize single
12386 byte case, inline streamer_read_uchar and defer section
12389 2013-06-20 Richard Biener <rguenther@suse.de>
12391 PR tree-optimization/57584
12392 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
12393 SSA names into the expanded expression that take part in
12394 abnormal coalescing.
12396 2013-06-19 Sharad Singhai <singhai@google.com>
12398 * gcov.c (print_usage): Handle new option.
12399 (process_args): Ditto.
12400 (get_gcov_intermediate_filename): New function.
12401 (output_intermediate_file): New function.
12402 (output_gcov_file): New function
12403 (generate_results): Handle new option.
12404 (release_function): Relase demangled name.
12405 (read_graph_file): Handle demangled name.
12406 (output_lines): Ditto.
12407 * doc/gcov.texi: Document gcov intermediate format.
12409 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
12412 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
12413 (lra_emit_add): Use the functions. Add comment about Y as an
12416 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
12419 * collect2.c (collect_atexit): New.
12420 (collect_exit): Delete.
12421 (main): Register collect_atexit with atexit.
12422 (collect_wait): Change collect_exit to exit.
12424 * collect2.h (collect_exit): Delete.
12425 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
12427 2013-06-19 Wei Mi <wmi@google.com>
12429 PR rtl-optimization/57518
12430 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
12431 if regno is used in paradoxical subreg.
12432 (update_equiv_regs): Check pdx_subregs[regno] before
12433 set a reg to be equivalent with a mem.
12435 2013-06-19 Matthias Klose <doko@ubuntu.com>
12438 * file-find.h (find_a_file): Add a mode parameter.
12439 * file-find.c (find_a_file): Likewise.
12440 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
12441 with X_OK for the executables.
12442 * collect2.c (main): Call find_a_file with X_OK.
12444 2013-06-19 Steve Ellcey <sellcey@mips.com>
12447 * config/mips/mips.md (casesi_internal_mips16_<mode>):
12448 Use NEXT_INSN instead of next_real_insn.
12450 2013-06-19 Jan Hubicka <jh@suse.cz>
12452 * cgraph.h (const_value_known_p): Replace by ...
12453 (ctor_for_folding): .. this one.
12454 * cgraphunit.c (process_function_and_variable_attributes): Use it.
12455 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
12456 * expr.c (expand_expr_real_1): Likewise.
12457 (string_constant): Likewise.
12458 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
12459 * ipa.c (process_references): Likewise.
12460 (symtab_remove_unreachable_nodes): Likewise.
12461 * ipa-inline-analysis.c (param_change_prob): Likewise.
12462 * gimple-fold.c (canonicalize_constructor_val): Likewise.
12463 (get_base_constructor): Likwise.
12464 * varpool.c (varpool_remove_node): Likewise.
12465 (varpool_remove_initializer): LIkewise.
12466 (dump_varpool_node): LIkwise.
12467 (const_value_known_p): Rewrite to ...
12468 (ctor_for_folding): ... this one.
12470 2013-06-19 Jakub Jelinek <jakub@redhat.com>
12473 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
12474 PERSONALITY in $PATH derived prefixes.
12476 2013-06-19 Jeff Law <law@redhat.com>
12478 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
12481 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
12482 (simplify_bitwise_binary): Use it to simpify certain binary ops on
12485 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
12487 * config/arm/vfp.md: Move VFP instruction classification documentation
12489 * config/arm/arm.md: ... here. Update instruction classification
12492 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
12494 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
12495 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
12496 pattern. Use more efficient sequences on ARMv5 and Thumb2.
12498 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
12501 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
12502 with NEXT_INSN. Use tablejump_p to check for jump table data
12505 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
12508 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
12509 that now in C++ the value is correct per the C++ standards.
12511 2013-06-19 Richard Biener <rguenther@suse.de>
12513 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
12514 for global context.
12516 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12519 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12522 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
12523 with next_active_insn.
12525 2013-06-18 Sriraman Tallam <tmsriram@google.com>
12527 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
12528 functions are inlined during failures to flag an error.
12529 * tree-inline.c (expand_call_inline): Allow the error to be flagged
12530 in early inline pass.
12532 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
12534 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
12537 2013-06-18 Julian Brown <julian@codesourcery.com>
12539 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
12540 Permit virtual register pre-reload if !strict.
12541 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
12543 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
12545 * config/arm/neon.md (movmisalign<mode>): Use
12546 neon_perm_struct_or_reg_operand instead of
12547 neon_struct_or_register_operand.
12548 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
12549 neon_permissive_struct_operand instead of neon_struct_operand.
12550 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
12551 neon_vector_mem_operand.
12552 * config/arm/predicates.md (neon_struct_operand): Adjust call to
12553 neon_vector_mem_operand.
12554 (neon_permissive_struct_operand): New.
12555 (neon_struct_or_register_operand): Rename to...
12556 (neon_perm_struct_or_reg_operand): This. Adjust call to
12557 neon_vector_mem_operand.
12559 2013-06-18 Richard Biener <rguenther@suse.de>
12561 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
12562 * lto-streamer.h: Include pointer-set.h.
12563 (struct lto_decl_slot): Remove.
12564 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
12565 Remove next_index entry.
12566 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
12567 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
12568 (lto_init_tree_ref_encoder): Adjust.
12569 (lto_destroy_tree_ref_encoder): Likewise.
12570 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
12571 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
12572 (lto_output_decl_index): Adjust.
12573 (lto_new_out_decl_state): Likewise.
12574 (lto_record_function_out_decl_state): Likewise.
12575 * lto-streamer-out.c (copy_function): Likewise.
12577 2013-06-18 Richard Biener <rguenther@suse.de>
12579 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
12580 * cgraphunit.c: Include cfgloop.h.
12581 (init_lowered_empty_function): Initialize the loop tree.
12582 (assemble_thunk): Insert new BBs into loops.
12584 2013-06-18 Richard Biener <rguenther@suse.de>
12586 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
12587 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
12588 the map from cache entry to cache index optional.
12589 (streamer_tree_cache_replace_tree): Adjust accordingly.
12590 (streamer_tree_cache_append): Likewise.
12591 (streamer_tree_cache_delete): Likewise.
12592 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
12593 streamer cache map from cache entry to cache index.
12594 * lto-streamer-out.c (create_output_block): Adjust.
12596 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
12598 * config/arm/arm.md (attribute "insn"): Move multiplication and
12599 division attributes to...
12600 (attribute "type"): ... here. Remove mult.
12601 (attribute "mul32"): New attribute.
12602 (attribute "mul64"): Add umaal.
12603 (*arm_mulsi3): Update attributes.
12604 (*arm_mulsi3_v6): Likewise.
12605 (*thumb_mulsi3): Likewise.
12606 (*thumb_mulsi3_v6): Likewise.
12607 (*mulsi3_compare0): Likewise.
12608 (*mulsi3_compare0_v6): Likewise.
12609 (*mulsi_compare0_scratch): Likewise.
12610 (*mulsi_compare0_scratch_v6): Likewise.
12611 (*mulsi3addsi): Likewise.
12612 (*mulsi3addsi_v6): Likewise.
12613 (*mulsi3addsi_compare0): Likewise.
12614 (*mulsi3addsi_compare0_v6): Likewise.
12615 (*mulsi3addsi_compare0_scratch): Likewise.
12616 (*mulsi3addsi_compare0_scratch_v6): Likewise.
12617 (*mulsi3subsi): Likewise.
12618 (*mulsidi3adddi): Likewise.
12619 (*mulsi3addsi_v6): Likewise.
12620 (*mulsidi3adddi_v6): Likewise.
12621 (*mulsidi3_nov6): Likewise.
12622 (*mulsidi3_v6): Likewise.
12623 (*umulsidi3_nov6): Likewise.
12624 (*umulsidi3_v6): Likewise.
12625 (*umulsidi3adddi): Likewise.
12626 (*umulsidi3adddi_v6): Likewise.
12627 (*smulsi3_highpart_nov6): Likewise.
12628 (*smulsi3_highpart_v6): Likewise.
12629 (*umulsi3_highpart_nov6): Likewise.
12630 (*umulsi3_highpart_v6): Likewise.
12631 (mulhisi3): Likewise.
12632 (*mulhisi3tb): Likewise.
12633 (*mulhisi3bt): Likewise.
12634 (*mulhisi3tt): Likewise.
12635 (maddhisi4): Likewise.
12636 (*maddhisi4tb): Likewise.
12637 (*maddhisi4tt): Likewise.
12638 (maddhidi4): Likewise.
12639 (*maddhidi4tb): Likewise.
12640 (*maddhidi4tt): Likewise.
12641 (divsi3): Likewise.
12642 (udivsi3): Likewise.
12643 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
12644 (thumb2_mulsi_short_compare0): Likewise.
12645 (thumb2_mulsi_short_compare0_scratch): Likewise.
12646 * config/arm/arm1020e.md (1020mult1): Update attribute change.
12647 (1020mult2): Likewise.
12648 (1020mult3): Likewise.
12649 (1020mult4): Likewise.
12650 (1020mult5): Likewise.
12651 (1020mult6): Likewise.
12652 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
12654 (cortex_a15_mult64): Likewise.
12655 (cortex_a15_sdiv): Likewise.
12656 (cortex_a15_udiv): Likewise.
12657 * config/arm/arm1026ejs.md (mult1): Update attribute change.
12663 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
12664 (pj4_ir_div): Likewise.
12665 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
12666 (11_mult2): Likewise.
12667 (11_mult3): Likewise.
12668 (11_mult4): Likewise.
12669 (11_mult5): Likewise.
12670 (11_mult6): Likewise.
12671 (11_mult7): Likewise.
12672 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
12673 (cortex_a8_mla): Likewise.
12674 (cortex_a8_mull): Likewise.
12675 (cortex_a8_smulwy): Likewise.
12676 (cortex_a8_smlald): Likewise.
12677 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
12678 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
12679 (cortex_r4_mul_3): Likewise.
12680 (cortex_r4_mla_4): Likewise.
12681 (cortex_r4_mla_3): Likewise.
12682 (cortex_r4_smlald): Likewise.
12683 (cortex_r4_mull): Likewise.
12684 (cortex_r4_sdiv): Likewise.
12685 (cortex_r4_udiv): Likewise.
12686 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
12687 (cortex_a7_idiv): Likewise.
12688 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
12689 (9_mult2): Likewise.
12690 (9_mult3): Likewise.
12691 (9_mult4): Likewise.
12692 (9_mult5): Likewise.
12693 (9_mult6): Likewise.
12694 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
12695 (cortex_a53_sdiv): Likewise.
12696 (cortex_a53_udiv): Likewise.
12697 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
12698 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
12699 (mp626_mult2): Likewise.
12700 (mp626_mult3): Likewise.
12701 (mp626_mult4): Likewise.
12702 * config/arm/fa526.md (526_mult1): Update attribute change.
12703 (526_mult2): Likewise.
12704 * config/arm/arm-generic.md (mult): Update attribute change.
12705 (mult_ldsched_strongarm): Likewise.
12706 (mult_ldsched): Likewise.
12707 (multi_cycle): Likewise.
12708 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
12709 * config/arm/fa606te.md (606te_mult1): Update attribute change.
12710 (606te_mult2): Likewise.
12711 (606te_mult3): Likewise.
12712 (606te_mult4): Likewise.
12713 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
12714 (cortex_a9_mac16): Likewise.
12715 (cortex_a9_multiply): Likewise.
12716 (cortex_a9_mac): Likewise.
12717 (cortex_a9_multiply_long): Likewise.
12718 * config/arm/fa626te.md (626te_mult1): Update attribute change.
12719 (626te_mult2): Likewise.
12720 (626te_mult3): Likewise.
12721 (626te_mult4): Likewise.
12723 2013-06-18 Richard Biener <rguenther@suse.de>
12726 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
12728 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12731 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
12732 with next_active_insn.
12734 2013-06-18 Alan Modra <amodra@gmail.com>
12736 * config/rs6000/rs6000.h (enum data_align): New.
12737 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
12738 (DATA_ABI_ALIGNMENT): Define.
12739 (CONSTANT_ALIGNMENT): Correct comment.
12740 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
12741 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
12743 2013-06-17 David Malcolm <dmalcolm@redhat.com>
12745 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
12746 ATTRIBUTE_UNUSED marking.
12748 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
12750 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
12751 alternative and update.
12752 (aarch64_dup_lanedi): Delete.
12753 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
12754 * config/aarch64/aarch64-simd-builtins.def: Update.
12756 2013-06-17 Richard Biener <rguenther@suse.de>
12758 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
12759 (lto_input_scc): Declare.
12760 (lto_input_tree_1): Likewise.
12761 (struct lto_stats_d): Add num_tree_bodies_output and
12762 num_pickle_refs_output.
12763 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
12764 (lto_read_tree_1): Split out from ...
12765 (lto_read_tree): ... this.
12766 (lto_input_scc): New function.
12767 (lto_input_tree_1): Split out from ...
12768 (lto_input_tree): ... this. Handle LTO_tree_scc.
12769 (lto_data_in_create): Create the streamer cache without hashes.
12770 * lto-streamer-out.c (create_output_block): Create the streamer
12771 cache with hashes when not doing WPA.
12772 (lto_write_tree_1): Split out from ...
12773 (lto_write_tree): ... this.
12774 (get_symbol_initial_value): New function.
12775 (lto_output_tree_1): Split out from ...
12776 (lto_output_tree): ... this. Write trees as series of SCCs
12777 using a DFS walk via DFS_write_tree.
12778 (struct sccs, struct scc_entry): New types.
12779 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
12780 (DFS_write_tree_body): New function.
12781 (DFS_write_tree): Likewise.
12782 (hash_tree): Likewise.
12783 (scc_entry_compare): Likewise.
12784 (hash_scc): Likewise.
12785 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
12786 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
12787 TREE_CHAIN as regular reference.
12788 (streamer_read_integer_cst): Remove.
12789 (streamer_get_pickled_tree): Adjust.
12790 * tree-streamer-out.c (streamer_write_chain): Disable streaming
12791 of DECL_EXTERNALs in BLOCK_VARS for now.
12792 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
12794 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
12795 Add hash value argument and record that if hashes are recorded
12797 (streamer_tree_cache_insert_1): Adjust.
12798 (streamer_tree_cache_insert): Likewise.
12799 (streamer_tree_cache_insert_at): Rename to ...
12800 (streamer_tree_cache_replace_tree): ... this and adjust.
12801 (streamer_tree_cache_append): Adjust.
12802 (record_common_node): Likewise.
12803 (streamer_tree_cache_create): Add argument whether to
12804 record hash values together with trees.
12805 (streamer_tree_cache_delete): Adjust.
12806 * tree-streamer.h (struct streamer_tree_cache_d): Add
12808 (streamer_read_integer_cst): Remove.
12809 (streamer_tree_cache_insert): Adjust.
12810 (streamer_tree_cache_append): Likewise.
12811 (streamer_tree_cache_insert_at): Rename to ...
12812 (streamer_tree_cache_replace_tree): ... this and adjust.
12813 (streamer_tree_cache_create): Add argument whether to record hashes.
12814 (streamer_tree_cache_get): Rename to ...
12815 (streamer_tree_cache_get_tree): ... this.
12816 (streamer_tree_cache_get_hash): New function.
12817 * tree.c (cache_integer_cst): New function.
12818 * tree.h (cache_integer_cst): Declare.
12819 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
12820 * lto-symtab.c (lto_varpool_replace_node): Only release
12821 DECL_INITIAL of non-prevailing decls.
12822 * varpool.c (varpool_remove_initializer): Do not release
12823 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
12825 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12827 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
12828 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
12829 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
12830 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
12831 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
12832 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
12833 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
12834 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
12835 instead of TARGET_64BIT.
12836 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
12837 Require ISA_HAS_<D>DIV.
12839 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
12841 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
12842 (mips*-*-linux*): Move default with_llsc setting to where other
12844 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
12846 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
12847 Likewise. Remove default with_tune setting. Move default float
12848 setting to its own block. Handle with_llsc in the same block as above.
12850 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
12852 PR rtl-optimization/57425
12853 PR rtl-optimization/57569
12854 * alias.c (write_dependence_p): Add new parameters mem_mode,
12855 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
12856 Changed all callers.
12857 (canon_anti_dependence): New function.
12858 * cse.c (check_dependence): Use canon_anti_dependence.
12859 * cselib.c (cselib_invalidate_mem): Likewise.
12860 * rtl.h (canon_anti_dependence): Declare.
12862 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12864 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
12865 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
12866 ".set mips3" for 64-bit targets.
12868 2013-06-15 Dehao Chen <dehao@google.com>
12870 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
12871 * gimple-low.c (gimple_check_call_matching_types): Likewise.
12872 (gimple_check_call_args): Likewise.
12873 * value-prof.c (check_ic_target): Likewise.
12874 * ipa-inline.c (early_inliner): Likewise.
12875 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
12876 * cgraph.c (cgraph_create_edge_1): Likewise.
12877 (cgraph_make_edge_direct): Likewise.
12879 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
12882 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
12883 rs6000_output_move_128bit to handle emitting quad memory
12884 operations. Set attribute length to 8 bytes.
12886 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
12888 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
12890 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
12891 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
12892 (aarch64_<su>mlsl<mode>): Likewise.
12894 2013-06-14 Mike Stump <mikestump@comcast.net>
12896 * Makefile.in (TARGET_H): Add insn-codes.h.
12898 2013-06-14 Alan Modra <amodra@gmail.com>
12900 PR middle-end/57134
12901 PR middle-end/57586
12902 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
12903 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
12904 bitfield expansion when EXPAND_MEMORY.
12905 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
12907 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
12909 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
12910 test for clearing quad memory on 32-bit later.
12912 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12914 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
12915 (fold_negate_expr): Likewise.
12916 (fold_real_zero_addition_p): Handle vectors.
12917 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
12919 2013-06-14 Alan Modra <amodra@gmail.com>
12921 * varasm.c (force_const_mem): Revert 2013-06-07 change.
12923 2013-06-13 Jan Hubicka <jh@suse.cz>
12925 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12926 Local comdats are not externally visible.
12927 * symtab.c (dump_symtab_base): Dump externally visible.
12928 (verify_symtab_base): Verify back links in the symtab hash.
12930 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12932 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
12933 CONVERT_EXPR as equal nodes.
12935 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12937 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
12939 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12941 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
12942 Generalize to complex and vector.
12943 * tree.c (build_all_ones_cst): New function.
12944 * tree.h (build_all_ones_cst): Declare it.
12946 2013-06-13 Alan Modra <amodra@gmail.com>
12948 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
12949 * config/rs6000/rs6000.md (signbittf2): New insn.
12950 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
12951 (abstf2_internal, cmptf_internal2): Likewise.
12952 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
12954 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12955 Pat Haugen <pthaugen@us.ibm.com>
12956 Peter Bergner <bergner@vnet.ibm.com>
12958 * config/rs6000/rs6000.c (emit_load_locked): Add support for
12959 power8 byte, half-word, and quad-word atomic instructions.
12960 (emit_store_conditional): Likewise.
12961 (rs6000_expand_atomic_compare_and_swap): Likewise.
12962 (rs6000_expand_atomic_op): Likewise.
12964 * config/rs6000/sync.md (larx): Add new modes for power8.
12966 (AINT): New mode iterator to include TImode as well as normal
12967 integer modes on power8.
12968 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
12969 that VSX registers are not considered. Use AINT mode iterator
12970 instead of INT1 to allow inclusion of quad word atomic operations
12972 (load_locked<mode>): Likewise.
12973 (store_conditional<mode>): Likewise.
12974 (atomic_compare_and_swap<mode>): Likewise.
12975 (atomic_exchange<mode>): Likewise.
12976 (atomic_nand<mode>): Likewise.
12977 (atomic_fetch_<fetchop_name><mode>): Likewise.
12978 (atomic_nand_fetch<mode>): Likewise.
12979 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
12981 (ATOMIC): On power8, add QImode, HImode modes.
12982 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
12983 modes that promote to SImode.
12984 (load_lockedti): Convert TImode arguments to PTImode, so that we
12985 get a guaranteed even/odd register pair.
12986 (load_lockedpti): Likewise.
12987 (store_conditionalti): Likewise.
12988 (store_conditionalpti): Likewise.
12990 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
12991 atomic load/store instructions.
12994 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
12996 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
12998 (insn_count): New attribute, with most cases extracted from...
12999 (length): ...here. Redefine most cases in terms of insn_count.
13000 (single_insn): Delete.
13001 (can_delay): Use insn_count to check for single instructions.
13002 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
13003 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
13004 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
13005 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
13006 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
13007 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
13008 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
13009 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
13010 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
13011 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
13012 rather than "length".
13013 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
13014 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
13015 Use "insn_count" rather than "length".
13016 * config/mips/mips-dsp.md
13017 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
13018 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
13021 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
13023 PR tree-optimization/57361
13024 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
13026 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
13028 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
13030 (aarch64_simd_combine<mode>): New instruction expansion.
13031 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
13032 function prototype.
13033 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
13034 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
13036 2013-06-12 Jan Hubicka <jh@suse.cz>
13038 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
13039 decl has when in streaming stage.
13040 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
13041 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
13043 2013-06-12 Roland Stigge <stigge@antcom.de>
13046 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
13048 2013-06-12 Jakub Jelinek <jakub@redhat.com>
13050 PR tree-optimization/57537
13051 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
13052 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
13054 2013-06-12 Richard Biener <rguenther@suse.de>
13056 * data-streamer.h (streamer_write_char_stream): CSE
13057 obs->current_pointer.
13058 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
13059 streamer_write_char_stream manually and optimize the resulting loop.
13060 (streamer_write_hwi_stream): Likewise.
13062 2013-06-12 Jan Hubicka <jh@suse.cz>
13064 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
13065 * cgraph.h (varpool_create_empty_node): Declare.
13066 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
13068 * symtab.c (symtab_unregister_node): Be lax about missin entries
13070 (symtab_get_node): Update comment.
13071 * varpool.c (varpool_create_empty_node): Break out from ...
13072 (varpool_node_for_decl): ... here.
13073 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
13075 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
13077 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
13078 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
13079 part. Use straight-line flow at the end.
13080 <COMPONENT_REF>: Remove superfluous else.
13081 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
13083 2013-06-12 Jakub Jelinek <jakub@redhat.com>
13086 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
13087 target hook even for !TREE_PUBLIC decls. If no resolution info
13088 is available, return false for common and external decls.
13090 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13092 * config/rl78/constraints.md (U): New constraint.
13093 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
13096 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
13099 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
13100 to allow returning address to AT_PLATFORM name.
13102 2013-06-11 Jan Hubicka <jh@suse.cz>
13104 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
13105 * cgraph.h (symtab_node_base): Add weakref flag.
13106 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
13107 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
13108 (output_weakrefs): Use weakref flag.
13109 * fold-const.c (simple_operand_p): Handle WEAK.
13110 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
13111 * ipa.c (varpool_externally_visible_p): Drop weakref.
13112 (function_and_variable_visibility): Update comment; fix weakref
13113 sanity checks; do not clear DECL_WEAK on them.
13114 * lto-cgraph.c (lto_output_node): update.
13115 (lto_output_varpool_node): Update.
13116 (input_overwrite_node): Update.
13117 (input_node): Update.
13118 (input_varpool_node): Update.
13119 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
13120 (lto_symtab_merge_symbols): Add sanity check.
13121 (lto_symtab_prevailing_decl): Do not special case weakrefs.
13122 * passes.c (rest_of_decl_compilation): Set static flag, too.
13123 * symtab.c (dump_symtab_base): Dump weakref.
13124 (verify_symtab_base): Sanity check weakrefs.
13125 (symtab_make_decl_local): Remove duplicated code.
13126 (symtab_alias_ultimate_target): Simplify.
13127 * varpool.c (varpool_create_variable_alias): Set weakref flag.
13129 2013-06-11 Tom de Vries <tom@codesourcery.com>
13131 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
13132 sequence_vect == NULL.
13134 2013-06-11 DJ Delorie <dj@redhat.com>
13136 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
13137 (rl78_unwind_word_mode): New.
13139 2013-06-11 David Malcolm <dmalcolm@redhat.com>
13141 * final.c (debug_prefix_maps): Make static.
13143 2013-06-11 David Malcolm <dmalcolm@redhat.com>
13145 * function.c (initial_trampoline): Remove stray copy.
13147 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
13149 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
13151 2013-06-11 Martin Jambor <mjambor@suse.cz>
13153 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
13154 within bounds at the beginning of the function.
13156 2013-06-11 Alan Modra <amodra@gmail.com>
13158 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
13160 (get_named_section): Don't NULL !DECL_P decl.
13162 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
13164 * doc/invoke.texi (core-avx2): Document.
13166 (atom): Updated with MOVBE.
13168 2013-06-11 Richard Biener <rguenther@suse.de>
13170 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
13172 2013-06-11 Anton Blanchard <anton@samba.org>
13174 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
13175 correct shift value in little-endian mode.
13177 2013-06-11 Jakub Jelinek <jakub@redhat.com>
13180 * varasm.c (get_variable_align): Move #endif to the right place.
13182 2013-06-10 Cary Coutant <ccoutant@google.com>
13184 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
13185 for hash so that hash table traversal order is deterministic.
13187 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
13188 Pat Haugen <pthaugen@us.ibm.com>
13189 Peter Bergner <bergner@vnet.ibm.com>
13191 * config/rs6000/vector.md (GPR move splitter): Do not split moves
13192 of vectors in GPRS if they are direct moves or quad word load or
13195 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
13197 (direct_move_p): Likewise.
13198 (quad_load_store_p): Likewise.
13200 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
13201 classes into bins based on the physical register type.
13202 (reg_class_to_reg_type): Likewise.
13203 (IS_STD_REG_TYPE): Likewise.
13204 (IS_FP_VECT_REG_TYPE): Likewise.
13205 (reload_fpr_gpr): Arrays to determine what insn to use if we can
13206 use direct move instructions.
13207 (reload_gpr_vsx): Likewise.
13208 (reload_vsx_gpr): Likewise.
13209 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
13210 information that is a simplification of register classes. Also
13211 precalculate direct move reload helpers.
13212 (direct_move_p): New function to return true if the operation can
13213 be done as a direct move instruciton.
13214 (quad_load_store_p): New function to return true if the operation
13215 is a quad memory operation.
13216 (rs6000_legitimize_address): If quad memory, only allow register
13217 indirect for TImode addresses.
13218 (rs6000_legitimate_address_p): Likewise.
13219 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
13220 (rs6000_reload_register_type): Likewise.
13221 (register_to_reg_type): Return register type.
13222 (rs6000_secondary_reload_simple_move): New helper function for
13223 secondary reload and secondary memory needed to identify anything
13224 that is a simple move, and does not need reloading.
13225 (rs6000_secondary_reload_direct_move): New helper function for
13226 secondary reload to identify cases that can be done with several
13227 instructions via the direct move instructions.
13228 (rs6000_secondary_reload_move): New helper function for secondary
13229 reload to identify moves between register types that can be done.
13230 (rs6000_secondary_reload): Add support for quad memory operations
13231 and for direct move.
13232 (rs6000_secondary_memory_needed): Likewise.
13233 (rs6000_debug_secondary_memory_needed): Change argument names.
13234 (rs6000_output_move_128bit): New function to return the move to
13235 use for 128-bit moves, including knowing about the various
13236 limitations of quad memory operations.
13238 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
13239 memory operations. call rs6000_output_move_128bit for the actual
13240 instruciton(s) to generate.
13241 (vsx_movti_64bit): Likewise.
13243 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
13244 (UNSPEC_P8V_MTVSRWZ): Likewise.
13245 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
13246 (UNSPEC_P8V_MTVSRD): Likewise.
13247 (UNSPEC_P8V_XXPERMDI): Likewise.
13248 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
13249 (UNSPEC_FUSION_GPR): Likewise.
13250 (FMOVE128_GPR): New iterator for direct move.
13251 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
13252 (f32_sv): Likewise.
13253 (f32_dm): Likewise.
13254 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
13255 loads and direct move instructions.
13256 (zero_extendsidi2_lfiwzx): Likewise.
13257 (extendsidi2_lfiwax): Likewise.
13258 (extendsidi2_nocell): Likewise.
13259 (floatsi<mode>2_lfiwax): Likewise.
13260 (lfiwax): Likewise.
13261 (floatunssi<mode>2_lfiwzx): Likewise.
13262 (lfiwzx): Likewise.
13263 (fix_trunc<mode>_stfiwx): Likewise.
13264 (fixuns_trunc<mode>_stfiwx): Likewise.
13265 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
13266 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
13267 (parity<mode>2_cmpb): Set length/type attr.
13268 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
13269 for 'mr.' to fast_compare.
13270 (bpermd_<mode>): Change type attr to popcnt.
13271 (p8_fmrgow_<mode>): New insns for power8 direct move support.
13272 (p8_mtvsrwz_1): Likewise.
13273 (p8_mtvsrwz_2): Likewise.
13274 (reload_fpr_from_gpr<mode>): Likewise.
13275 (p8_mtvsrd_1): Likewise.
13276 (p8_mtvsrd_2): Likewise.
13277 (p8_xxpermdi_<mode>): Likewise.
13278 (reload_vsx_from_gpr<mode>): Likewise.
13279 (reload_vsx_from_gprsf): Likewise.
13280 (p8_mfvsrd_3_<mode>): LIkewise.
13281 (reload_gpr_from_vsx<mode>): Likewise.
13282 (reload_gpr_from_vsxsf): Likewise.
13283 (p8_mfvsrd_4_disf): Likewise.
13284 (multi-word GPR splits): Do not split direct moves or quad memory
13287 2013-06-10 David Malcolm <dmalcolm@redhat.com>
13289 * tree-into-ssa.c (interesting_blocks): Make static.
13291 2013-06-10 Jakub Jelinek <jakub@redhat.com>
13294 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
13295 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
13296 Use DATA_ABI_ALIGNMENT for that case instead if defined.
13297 (get_variable_align): New function.
13298 (get_variable_section, emit_bss, emit_common,
13299 assemble_variable_contents, place_block_symbol): Use
13300 get_variable_align instead of DECL_ALIGN.
13301 (assemble_noswitch_variable): Add align argument, use it
13302 instead of DECL_ALIGN.
13303 (assemble_variable): Adjust caller. Use get_variable_align
13304 instead of DECL_ALIGN.
13305 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
13307 (DATA_ABI_ALIGNMENT): Define.
13308 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
13309 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
13310 opt is false, only return the psABI mandated alignment increase.
13311 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
13312 (DATA_ABI_ALIGNMENT): ... this.
13313 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
13314 (DATA_ABI_ALIGNMENT): ... this.
13315 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
13316 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
13317 (DATA_ABI_ALIGNMENT): ... this.
13318 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
13319 * doc/tm.texi: Regenerated.
13321 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
13323 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
13324 cmp_code to construct REG_EQUAL note.
13326 2013-06-09 Jakub Jelinek <jakub@redhat.com>
13329 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
13330 that operands[2] doesn't overlap with operands[0].
13332 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
13333 Jan Hubicka <jh@suse.cz>
13335 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
13336 hack to mark symbols as used.
13338 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
13340 PR rtl-optimization/57559
13341 * lra-constraints.c (process_alt_operands): Don't discourage
13342 memory with known offset for offsetable memory constraint.
13343 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
13345 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
13347 * varasm.c (struct oc_local_state): Reorder fields.
13348 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
13349 and adjust accordingly.
13350 (output_constructor): Reorder initialization code and adjust call to
13351 output_constructor_bitfield.
13353 2013-06-07 Jan Hubicka <jh@suse.cz>
13355 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
13357 2013-06-07 David Malcolm <dmalcolm@redhat.com>
13359 * tree-object-size.c (unknown): Make const.
13361 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13363 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
13364 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
13365 for last alternative in the cpu_facility attribute.
13367 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13370 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
13371 (xordi3): Change operand 2 constraint to arm_xordi_operand.
13372 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
13373 * config/arm/constraints.md (Dg): New constraint.
13374 * config/arm/neon.md (xordi3_neon): Remove.
13375 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
13376 * config/arm/predicates.md (arm_xordi_operand): New predicate.
13378 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13380 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
13381 Clean up alternatives.
13383 2013-06-07 Alan Modra <amodra@gmail.com>
13385 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
13388 2013-06-07 Alan Modra <amodra@gmail.com>
13390 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
13392 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13394 * config/arm/constraints.md (Df): New constraint.
13395 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
13396 Correct length attribute for last two alternatives.
13398 2013-06-07 Alan Modra <amodra@gmail.com>
13400 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
13401 override user -mfp-in-toc.
13402 (offsettable_ok_by_alignment): Consider just the current access
13403 rather than the whole object, unless BLKmode. Handle
13404 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
13405 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
13406 for -mcmodel=medium.
13407 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
13408 override user -mfp-in-toc or -msum-in-toc. Default to
13409 -mno-fp-in-toc for -mcmodel=medium.
13411 2013-06-06 DJ Delorie <dj@redhat.com>
13413 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
13414 TARGET_VALID_POINTER_MODE.
13416 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
13417 Pat Haugen <pthaugen@us.ibm.com>
13418 Peter Bergner <bergner@vnet.ibm.com>
13420 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
13421 Document new power8 builtins.
13423 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
13424 condition code register, to allow 128-bit logical operations to be
13425 done in the VSX or GPR registers.
13426 (nor<mode>3): Use the canonical form for nor.
13427 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
13428 vclz*, and vpopcnt* vector instructions.
13429 (nand<mode>3): Likewise.
13430 (orc<mode>3): Likewise.
13431 (clz<mode>2): LIkewise.
13432 (popcount<mode>2): Likewise.
13434 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
13435 that only the GPRs are recognized.
13437 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
13438 support for new power8 builtins.
13440 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
13442 (xscvdpspn): Likewise.
13448 (vpopcnt): Likewise.
13449 (vpopcntb): Likewise.
13450 (vpopcnth): Likewise.
13451 (vpopcntw): Likewise.
13452 (vpopcntd): Likewise.
13454 (vmrgew): Likewise.
13455 (vmrgow): Likewise.
13457 (eqv_v16qi3): Likewise.
13458 (eqv_v8hi3): Likewise.
13459 (eqv_v4si3): Likewise.
13460 (eqv_v2di3): Likewise.
13461 (eqv_v4sf3): Likewise.
13462 (eqv_v2df3): Likewise.
13464 (nand_v16qi3): Likewise.
13465 (nand_v8hi3): Likewise.
13466 (nand_v4si3): Likewise.
13467 (nand_v2di3): Likewise.
13468 (nand_v4sf3): Likewise.
13469 (nand_v2df3): Likewise.
13471 (orc_v16qi3): Likewise.
13472 (orc_v8hi3): Likewise.
13473 (orc_v4si3): Likewise.
13474 (orc_v2di3): Likewise.
13475 (orc_v4sf3): Likewise.
13476 (orc_v2df3): Likewise.
13478 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
13479 allow power8 quad mode in 64-bit.
13480 (rs6000_builtin_vectorized_function): Add support to vectorize
13481 ISA 2.07 count leading zeros, population count builtins.
13482 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
13483 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
13484 (builtin_function_type): Add vgbbd builtin function which takes an
13486 (altivec_expand_vec_perm_const): Add support for new power8 merge
13489 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
13490 that does not include TImdoe for use with 32-bit.
13491 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
13493 (UNSPEC_VSX_CVDPSPN): Likewise.
13494 (vsx_xscvdpspn): Likewise.
13495 (vsx_xscvspdpn): Likewise.
13496 (vsx_xscvdpspn_scalar): Likewise.
13497 (vsx_xscvspdpn_directmove): Likewise.
13498 (vsx_and<mode>3): Split logical operations into 32-bit and
13499 64-bit. Add support to do logical operations on TImode as well as
13500 VSX vector types. Allow logical operations to be done in either
13501 VSX registers or in general purpose registers in 64-bit mode. Add
13502 splitters if GPRs were used. For AND, add clobber of CCmode to
13503 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
13505 (vsx_and<mode>3_32bit): Likewise.
13506 (vsx_and<mode>3_64bit): Likewise.
13507 (vsx_ior<mode>3): Likewise.
13508 (vsx_ior<mode>3_32bit): Likewise.
13509 (vsx_ior<mode>3_64bit): Likewise.
13510 (vsx_xor<mode>3): Likewise.
13511 (vsx_xor<mode>3_32bit): Likewise.
13512 (vsx_xor<mode>3_64bit): Likewise.
13513 (vsx_one_cmpl<mode>2): Likewise.
13514 (vsx_one_cmpl<mode>2_32bit): Likewise.
13515 (vsx_one_cmpl<mode>2_64bit): Likewise.
13516 (vsx_nor<mode>3): Likewise.
13517 (vsx_nor<mode>3_32bit): Likewise.
13518 (vsx_nor<mode>3_64bit): Likewise.
13519 (vsx_andc<mode>3): Likewise.
13520 (vsx_andc<mode>3_32bit): Likewise.
13521 (vsx_andc<mode>3_64bit): Likewise.
13522 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
13523 and xxlorc instructions.
13524 (vsx_eqv<mode>3_64bit): Likewise.
13525 (vsx_nand<mode>3_32bit): Likewise.
13526 (vsx_nand<mode>3_64bit): Likewise.
13527 (vsx_orc<mode>3_32bit): Likewise.
13528 (vsx_orc<mode>3_64bit): Likewise.
13530 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
13532 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
13534 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
13535 (p8_vmrgow): Likewise.
13536 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
13537 GPRs to be split under VSX.
13538 (p8v_clz<mode>2): Add power8 count leading zero support.
13539 (p8v_popcount<mode>2): Add power8 population count support.
13540 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
13543 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
13546 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
13548 (vec_nand): Likewise.
13549 (vec_vclz): Likewise.
13550 (vec_vclzb): Likewise.
13551 (vec_vclzd): Likewise.
13552 (vec_vclzh): Likewise.
13553 (vec_vclzw): Likewise.
13554 (vec_vgbbd): Likewise.
13555 (vec_vmrgew): Likewise.
13556 (vec_vmrgow): Likewise.
13557 (vec_vpopcnt): Likewise.
13558 (vec_vpopcntb): Likewise.
13559 (vec_vpopcntd): Likewise.
13560 (vec_vpopcnth): Likewise.
13561 (vec_vpopcntw): Likewise.
13563 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13565 PR rtl-optimization/57468
13566 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
13569 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13571 PR rtl-optimization/57459
13572 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
13573 type when setting live regs.
13575 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
13577 * config/s390/s390.opt (mlra): New option.
13578 * config/s390/s390.c (s390_decompose_address): Check displacement
13579 for all registers for LRA.
13580 (s390_secondary_reload): Don't used secondary reloads for LRA.
13581 (s390_lra_p): New function.
13582 (TARGET_LRA_P): Define.
13583 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
13584 of attribute cpu_facility to zarch for the last alternative.
13585 (*cmpmem_short): Ditto.
13587 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
13589 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
13590 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
13591 (arm_expand_prologue): Likewise.
13593 2013-06-06 Teresa Johnson <tejohnson@google.com>
13596 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
13597 as this is now done by redirect_edge_and_branch_force.
13598 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
13599 barriers, and fix interaction with splitting.
13600 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
13601 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
13602 reflect changes made in the routine.
13603 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
13604 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
13605 since this is called in cfglayout mode, and replace partition fixup
13606 with assert as that is now done by force_nonfallthru_and_redirect.
13607 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
13608 already be marked with region crossing note.
13609 (insert_section_boundary_note): Make non-static, gate on flag
13610 has_bb_partition, rewrite to also check for multiple partitions.
13611 (rest_of_handle_reorder_blocks): Remove call to
13612 insert_section_boundary_note, now done later during free_cfg.
13613 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
13614 * bb-reorder.h (insert_section_boundary_note): Declare.
13615 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
13616 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
13617 invoke insert_section_boundary_note.
13618 (try_redirect_by_replacing_jump): Remove unnecessary
13619 check for region crossing note.
13620 (fixup_partition_crossing): New function.
13621 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
13622 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
13623 in non-cfglayout mode.
13624 (force_nonfallthru_and_redirect): Fixup partition boundaries,
13625 remove old code that tried to do this. Emit barrier correctly
13626 when we are in cfglayout mode.
13627 (last_bb_in_partition): New function.
13628 (rtl_split_edge): Correctly fixup partition boundaries.
13629 (commit_one_edge_insertion): Remove old code that tried to
13630 fixup region crossing edge since this is now handled in
13631 split_block, and set up insertion point correctly since
13632 block may now end in a jump.
13633 (verify_hot_cold_block_grouping): Guard against checking when not in
13634 linearized RTL mode.
13635 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
13637 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
13638 rtl_verify_flow_info, so not called in cfglayout mode.
13639 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
13640 (fixup_reorder_chain): Remove old code that attempted to fixup region
13641 crossing note as this is now handled in force_nonfallthru_and_redirect.
13642 (duplicate_insn_chain): Don't duplicate switch section notes.
13643 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
13645 * basic-block.h (emit_barrier_after_bb): Declare.
13647 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13649 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
13650 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
13651 arm_usatsihi): Adjust alternatives for arm_restrict_it.
13653 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13655 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
13657 * config/arm/ldmstm.md: Regenerate.
13659 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13661 * config/arm/sync.md (atomic_loaddi_1):
13662 Disable predication for arm_restrict_it.
13663 (arm_load_exclusive<mode>): Likewise.
13664 (arm_load_exclusivesi): Likewise.
13665 (arm_load_exclusivedi): Likewise.
13666 (arm_load_acquire_exclusive<mode>): Likewise.
13667 (arm_load_acquire_exclusivesi): Likewise.
13668 (arm_load_acquire_exclusivedi): Likewise.
13669 (arm_store_exclusive<mode>): Likewise.
13670 (arm_store_exclusive<mode>): Likewise.
13671 (arm_store_release_exclusivedi): Likewise.
13672 (arm_store_release_exclusive<mode>): Likewise.
13674 2013-06-06 Richard Biener <rguenther@suse.de>
13676 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
13678 (lto_tag_is_tree_code_p): Adjust.
13679 (lto_tag_is_gimple_code_p): Likewise.
13680 (lto_gimple_code_to_tag): Likewise.
13681 (lto_tag_to_gimple_code): Likewise.
13682 (lto_tree_code_to_tag): Likewise.
13683 (lto_tag_to_tree_code): Likewise.
13684 * data-streamer.h (streamer_write_hwi_in_range): Use
13685 uhwi streaming to stream the normalized range.
13686 (streamer_read_hwi_in_range): Likewise.
13688 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13690 * config/arm/arm.md (enabled_for_depr_it): New attribute.
13691 (predicable_short_it): Likewise.
13692 (predicated): Likewise.
13693 (enabled): Handle above.
13694 (define_cond_exec): Set predicated attribute to yes.
13696 2013-06-05 Mike Stump <mikestump@comcast.net>
13698 * gdbinit.in (__FUNCTION__): Add.
13700 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
13702 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
13703 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
13705 2013-06-05 Jan Hubicka <jh@suse.cz>
13707 * varasm.c (mark_decl_referenced): Revert the removal until targets
13710 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
13712 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
13713 instead of mark_decl_referenced.
13715 2013-06-05 Jan Hubicka <jh@suse.cz>
13717 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
13718 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
13719 and symtab_used_from_object_file_p.
13720 (cgraph_make_node_local_1): Clear forced_by_abi.
13721 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
13722 * cgraph.h (symtab_node_base): Add forced_by_abi.
13723 (decide_is_variable_needed): Remove.
13724 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
13725 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
13726 (decide_is_symbol_needed): ... this one; handle symbols in general;
13727 always analyze virtuals; honnor forced_by_abi.
13728 (cgraph_finalize_function): Update.
13729 (varpool_finalize_decl): Update.
13730 (symbol_defined_and_needed): Remove.
13731 (analyze_functions): Update.
13732 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13733 output_refs, input_overwrite_node): Handle forced_by_abi.
13734 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
13735 (address_taken_from_non_vtable_p): ... this one.
13736 (comdat_can_be_unshared_p_1): New function.
13737 (cgraph_comdat_can_be_unshared_p): Rename to ...
13738 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
13739 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
13740 (function_and_variable_visibility): Clear forced_by_abi as needed.
13741 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
13742 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
13743 * symtab.c (dump_symtab_base): Dump forced_by_abi.
13744 * varpool.c (decide_is_variable_needed): Remove.
13746 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13748 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
13749 (arm_option_override): Override arm_restrict_it where appropriate.
13750 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
13751 * config/arm/arm.opt (mrestrict-it): New command-line option.
13752 * doc/invoke.texi: Document -mrestrict-it.
13754 2013-06-05 David Malcolm <dmalcolm@redhat.com>
13756 * tsan.c (tsan_atomic_table): Make const.
13758 2013-06-05 Richard Biener <rguenther@suse.de>
13760 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
13761 index associated with the tree we are supposed to replace.
13762 * tree-streamer-out.c (pack_ts_base_value_fields): Output
13763 TREE_ASM_WRITTEN as zero for everything but SSA names.
13765 2013-06-05 David Malcolm <dmalcolm@redhat.com>
13767 * tree-ssa-structalias.c (call_stmt_vars): Make static.
13769 2013-06-04 Jan Hubicka <jh@suse.cz>
13771 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
13772 (input_node, input_varpool_node): Handle correctly external same
13774 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
13775 nodes at ltrans stage.
13777 2013-06-04 Jan Hubicka <jh@suse.cz>
13779 * ipa-inline.c (update_caller_keys): Fix availability test.
13780 (update_callee_keys): Likewise.
13781 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
13782 to follow ELF standard.
13784 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
13786 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
13787 (mips64r5900el-*-elf*): New configurations.
13788 * config/mips/mips-cpus.def (r5900): New processor.
13789 * config/mips/mips-tables.opt: Regenerate.
13790 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
13791 (mips_issue_rate): Handle PROCESSOR_R5900.
13792 (mips_reorg_process_insns): Force reorder mode for the R5900.
13793 * config/mips/mips.h (TARGET_MIPS5900): Define.
13794 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
13796 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
13798 * config/mips/mips.md (processor): Add r5900.
13799 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
13801 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13803 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
13804 into function to generate MOVI instruction.
13805 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
13806 (aarch64_preferred_simd_mode): Turn into wrapper.
13807 (aarch64_output_scalar_simd_mov_immediate): New function.
13808 * config/aarch64/aarch64-protos.h: Add prototype for above.
13810 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13812 * config/aarch64/aarch64.c (simd_immediate_info): Remove
13813 element_char member.
13814 (sizetochar): Return signed char.
13815 (aarch64_simd_valid_immediate): Remove elchar and other
13816 unnecessary variables.
13817 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
13818 Calculate element_char as required.
13819 * config/aarch64/aarch64-protos.h: Update and move prototype
13820 for aarch64_output_simd_mov_immediate.
13821 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
13824 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13826 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
13827 information completed by aarch64_simd_valid_immediate.
13828 (aarch64_legitimate_constant_p): Update arguments.
13829 (aarch64_simd_valid_immediate): Work with struct rather than many
13831 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
13832 (aarch64_simd_make_constant): Update arguments.
13833 (aarch64_output_simd_mov_immediate): Work with struct rather than
13834 many pointers. Output immediate directly rather than as operand.
13835 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13837 * config/aarch64/constraints.md (Dn): Update arguments.
13839 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13841 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
13843 (aarch64_simd_immediate_valid_for_move): Remove.
13844 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
13845 (aarch64_simd_make_constant): Update call.
13846 (aarch64_output_simd_mov_immediate): Update call.
13847 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13849 * config/aarch64/constraints.md (Dn): Update call.
13851 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13853 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
13854 return type to bool for prototype.
13855 (aarch64_legitimate_constant_p): Check for true instead of not -1.
13856 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
13857 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
13859 2013-06-04 Catherine Moore <clm@codesourcery.com>
13861 * config/mips/mips.opt (meva): New.
13862 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
13863 (ASM_SPEC): Handle -meva.
13864 * doc/invoke.texi (meva): Document.
13866 2013-06-04 Alan Modra <amodra@gmail.com>
13868 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
13871 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13873 * rtl.def: Add extra fourth optional field to define_cond_exec.
13874 * gensupport.c (process_one_cond_exec): Process attributes from
13876 * doc/md.texi: Document fourth field in define_cond_exec.
13878 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
13880 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
13881 out the processing order as in store_bit_field_1.
13883 2013-06-04 Jan Hubicka <jh@suse.cz>
13885 PR middle-end/57500
13886 * cgraphunit.c (cgraph_process_same_body_aliases): Create
13887 non-VAR_DECL node if it does not exist yet.
13889 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
13891 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
13892 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
13893 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
13894 target_cpu_default setting.
13896 2013-06-03 Teresa Johnson <tejohnson@google.com>
13898 * dumpfile.c (opt_info_switch_p): Change -fopt-info
13899 default to -fopt-info=optimized instead of all.
13900 * doc/invoke.texi: Ditto.
13901 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
13902 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
13903 (execute_vect_slp): Emit BB vectorization success under
13904 MSG_OPTIMIZED_LOCATIONS.
13905 * tree-vect-slp.c (vect_slp_transform_bb): Change
13906 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
13907 * tree-vect-loop.c (vect_transform_loop): Ditto.
13909 2013-06-03 Jason Merrill <jason@redhat.com>
13912 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
13913 Use TARGET_EXPR for C++.
13915 2013-06-03 Jakub Jelinek <jakub@redhat.com>
13917 PR rtl-optimization/57268
13918 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
13919 if DEBUG_INSN_P (insn).
13922 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
13924 PR rtl-optimization/57268
13925 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13926 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13928 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13930 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
13931 (ix86_avoid_lea_for_addr): Likewise.
13932 (exact_dependency_1): Likewise.
13933 (ix86_adjust_cost): Likewise.
13934 (swap_top_of_ready_list): Fix formatting and !reload_completed check
13936 (do_reorder_for_imul): Fix typo, formatting and
13937 !reload_completed check removed.
13938 (ix86_sched_reorder): Fix typo and formatting.
13939 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
13942 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
13944 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
13946 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
13948 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
13949 <STRING_CST>: Likewise.
13950 <VECTOR_CST>: Likewise.
13952 2013-06-01 Janus Weil <janus@gcc.gnu.org>
13953 Mikael Morin <mikael@gcc.gnu.org>
13955 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
13956 * config.in: Regenerated.
13957 * configure: Regenerated.
13959 2013-06-01 Jan Hubicka <jh@suse.cz>
13961 PR middle-end/57366
13962 * cgraphunit.c (compile): When weakref is not supported,
13963 set up transparent aliases before final output pass.
13964 * varasm.c (assemble_alias): Do not try to do it here.
13966 2013-06-01 Jan Hubicka <jh@suse.cz>
13968 PR middle-end/57467
13969 * passes.c (for_per_function): Skip unanalyzed functions.
13971 2013-06-01 Jan Hubicka <jh@suse.cz>
13973 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
13974 (lto_symtab_merge_symbols_1): ... this one.
13975 (lto_symtab_merge_cgraph_nodes): Rename to ...
13976 (lto_symtab_merge_symbols): ... this one; simplify.
13977 * cgraph.c (same_body_aliases_done): Rename to ...
13978 (cpp_implicit_aliases_done): ... this one.
13979 (cgraph_create_function_alias): Update.
13980 (cgraph_same_body_alias): Update.
13981 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
13982 (verify_edge_corresponds_to_fndecl): Simplify.
13983 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
13984 (cgraph_node): Remove same_body_alias.
13985 (varpool_node): Remove alias_of and extra_name_alias.
13986 (same_body_aliases_done): Rename to ..
13987 (cpp_implicit_aliases_done): ... this one.
13988 (symtab_alias_ultimate_target): Add default parameter.
13989 (symtab_resolve_alias): New function.
13990 (fixup_same_cpp_alias_visibility): Declare.
13991 (cgraph_function_node): Add default parameter.
13992 (cgraph_node_asm_name): Likewise.
13993 (cgraph_function_or_thunk_node): Add default parameter; do
13994 not ICE when it is NULL.
13995 (varpool_variable_node): Likewise.
13996 * tree-emutls.c (create_emultls_var): Update.
13997 (ipa_lower_emutls): Update.
13998 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
13999 (cgraph_reset_node): Reset alias info.
14000 (cgraph_finalize_function): Update.
14001 (fixup_same_cpp_alias_visibility): Move to symtab.c.
14002 (analyze_function): Simplify.
14003 (cgraph_process_same_body_aliases): Simplify.
14004 (analyze_functions): Fixup same body aliases.
14005 (handle_alias_pairs): Simplify.
14006 (assemble_thunk): Update.
14007 (assemble_thunks_and_aliases): Update.
14008 (output_weakrefs): Rewrite.
14009 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
14010 (lto_output_varpool_node): Likewise.
14011 (compute_ltrans_boundary): Remve assert.
14012 (get_alias_symbol): New functoin.
14013 (input_node): Rewrite alias handling.
14014 (input_varpool_node): Likewise.
14015 * ipa-pure-const.c (propagate_pure_const): Fix formating.
14016 * ipa.c (process_references): Handle weakrefs correctly.
14017 (symtab_remove_unreachable_nodes): Likewise.
14018 * trans-mem.c (get_cg_data): Update.
14019 (ipa_tm_create_version_alias): Update.
14020 (ipa_tm_execute): Update.
14021 * symtab.c (dump_symtab_base): Dump aliases.
14022 (verify_symtab_base): Verify aliases.
14023 (symtab_node_availability): New function.
14024 (symtab_alias_ultimate_target): Simplify.
14025 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
14026 handle all the fixup cases.
14027 (symtab_resolve_alias): New function.
14028 * passes.c (ipa_write_summaries): Handle weakrefs.
14029 * varpool.c (varpool_analyze_node): Simplify.
14030 (assemble_aliases): Update.
14031 (varpool_create_variable_alias): Simplify.
14032 (varpool_extra_name_alias): Simplify.
14033 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
14034 (lto_symtab_merge_symbols): ... this one.
14036 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
14039 PR rtl-optimization/57268
14040 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14041 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14043 2013-06-01 Tobias Burnus <burnus@net-b.de>
14045 Partially reverted:
14046 2013-05-31 Tobias Burnus <burnus@net-b.de>
14048 PR middle-end/57073
14049 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
14052 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14054 PR rtl-optimization/57268
14055 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14056 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14058 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14060 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
14061 unordered comparison operators when -fno-trapping-math is in effect
14063 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
14064 and implement unordered comparison operators properly on the e500.
14066 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14068 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
14069 for constant scalar integers.
14070 (simplify_relational_operation_1): Likewise.
14072 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
14074 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
14075 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
14078 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14079 Igor Zamyatin <igor.zamyatin@intel.com>
14081 Silvermont (SLM) architecture performance tuning.
14082 * config/i386/i386.h (enum ix86_tune_indices): Add
14083 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
14084 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
14086 * config/i386/i386.c (initial_ix86_tune_features)
14087 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
14088 (ix86_lea_outperforms): Handle Silvermont tuning.
14089 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
14091 (ix86_use_lea_for_mov): Likewise.
14092 (ix86_avoid_lea_for_addr): Likewise.
14093 (ix86_lea_for_add_ok): Likewise.
14094 (exact_dependency_1): New function.
14095 (exact_store_load_dependency): Likewise.
14096 (ix86_adjust_cost): Handle Silvermont tuning.
14097 (do_reoder_for_imul): Likewise.
14098 (swap_top_of_ready_list): New function.
14099 (ix86_sched_reorder): Changed to handle Silvermont tuning.
14101 * config/i386/i386.md (peepholes that split memory operand in fp
14104 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14106 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
14107 Remove un-necessary braces.
14109 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14111 * config/aarch64/aarch64.c (aarch64_classify_symbol):
14112 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
14114 2013-05-31 Tobias Burnus <burnus@net-b.de>
14116 PR middle-end/57073
14117 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
14119 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14122 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
14123 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
14124 * config/arm/neon.md (iordi3_neon): Remove.
14125 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
14126 * config/arm/predicates.md (imm_for_neon_logic_operand):
14127 Move to earlier in the file.
14128 (neon_logic_op2): Likewise.
14129 (arm_iordi_operand_neon): New predicate.
14131 2013-05-31 Richard Biener <rguenther@suse.de>
14133 PR tree-optimization/57478
14134 PR tree-optimization/57453
14135 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
14138 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14140 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
14141 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
14143 2013-05-30 Tobias Burnus <burnus@net-b.de>
14144 Thomas Koenig <tkoenig@gcc.gnu.org>
14146 PR middle-end/57073
14147 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
14148 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
14150 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
14152 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
14154 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
14156 * target.def (register_usage_leveling_p): New hook.
14157 * targhooks.c (default_register_usage_leveling_p): New.
14158 * targhooks.h (default_register_usage_leveling_p): New prototype.
14159 * lra-assigns.c (register_usage_leveling_p): Use the hook.
14160 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
14161 * doc/tm.texi: Update.
14162 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
14164 2013-05-30 Ian Bolton <ian.bolton@arm.com>
14166 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
14167 (*insv_reg<mode>): New define_insn.
14169 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
14171 PR rtl-optimization/57439
14172 * postreload.c (move2add_valid_value_p): Check that we have
14173 a zero subreg_regno_offset when accessing the register in
14174 the requested mode.
14176 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14177 Igor Zamyatin <igor.zamyatin@intel.com>
14179 Silvermont (SLM) architecture pipeline model, tuning and
14181 * config.gcc: Add slm config options and target.
14183 * config/i386/slm.md: New.
14185 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
14187 * config/i386/i386-c.c (ix86_target_macros_internal): New case
14189 (ix86_target_macros_internal): Likewise.
14191 * config/i386/i386.c (slm_cost): New cost.
14192 (m_SLM): New macro flag.
14193 (initial_ix86_tune_features): Set m_SLM.
14194 (x86_accumulate_outgoing_args): Likewise.
14195 (x86_arch_always_fancy_math_387): Likewise.
14196 (processor_target_table): Add slm cost.
14197 (cpu_names): Add slm cpu name.
14198 (x86_option_override_internal): Set SLM ISA.
14199 (ix86_issue_rate): New case PROCESSOR_SLM.
14200 (ia32_multipass_dfa_lookahead): Likewise.
14201 (fold_builtin_cpu): Add slm.
14203 * config/i386/i386.h (TARGET_SLM): New target macro.
14204 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
14205 (processor_type): Add PROCESSOR_SLM.
14207 * config/i386/i386.md (cpu): Add new value "slm".
14208 (slm.md): Include slm.md.
14210 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
14211 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14213 * config/arm/arm-protos.h: Add and update function protos.
14214 * config/arm/arm.c (use_simple_return_p): New added.
14215 (thumb2_expand_return): Check simple_return flag.
14216 * config/arm/arm.md: Add simple_return and conditional simple_return.
14217 * config/arm/iterators.md: Add iterator for return and simple_return.
14219 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14221 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
14222 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
14223 (arm_emit_vfp_multi_reg_pop): Likewise.
14224 (thumb2_emit_ldrd_pop): Likewise.
14225 (arm_expand_epilogue): Add misc REG_CFA notes.
14226 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
14228 2013-05-29 Lawrence Crowl <crowl@google.com>
14230 * config/arm/t-arm: Update for below.
14232 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
14233 Change type to hash_table. Update dependent calls and types.
14235 * config/i386/t-cygming: Update for below.
14237 * config/i386/t-interix: Update for below.
14239 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
14240 Change type to hash_table. Update dependent calls and types.
14241 (i386_find_on_wrapper_list::wrappers): Likewise.
14243 * config/ia64/t-ia64: Update for below.
14245 * config/ia64/ia64.c (bundle_state_table):
14246 Change type to hash_table. Update dependent calls and types.
14248 * config/mips/mips.c (mips_reorg_process_insns::htab):
14249 Change type to hash_table. Update dependent calls and types.
14251 * config/sol2.c (solaris_comdat_htab):
14252 Change type to hash_table. Update dependent calls and types.
14254 * config/t-sol2: Update for above.
14256 2013-05-29 Teresa Johnson <tejohnson@google.com>
14258 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
14259 functions are not yet marked as defined.
14261 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
14262 Pat Haugen <pthaugen@us.ibm.com>
14263 Peter Bergner <bergner@vnet.ibm.com>
14265 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
14269 (vrotl<mode>3): Likewise.
14270 (vashl<mode>3): Likewise.
14271 (vlshr<mode>3): Likewise.
14272 (vashr<mode>3): Likewise.
14274 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14275 support for power8 V2DI builtins.
14277 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
14278 power8 V2DI builtins.
14279 (vupkhsw): Likewise.
14280 (vupklsw): Likewise.
14281 (vaddudm): Likewise.
14282 (vminsd): Likewise.
14283 (vmaxsd): Likewise.
14284 (vminud): Likewise.
14285 (vmaxud): Likewise.
14286 (vpkudum): Likewise.
14287 (vpksdss): Likewise.
14288 (vpkudus): Likewise.
14289 (vpksdus): Likewise.
14294 (vsubudm): Likewise.
14295 (vcmpequd): Likewise.
14296 (vcmpgtsd): Likewise.
14297 (vcmpgtud): Likewise.
14298 (vcmpequd_p): Likewise.
14299 (vcmpgtsd_p): Likewise.
14300 (vcmpgtud_p): Likewise.
14301 (vupkhsw): Likewise.
14302 (vupklsw): Likewise.
14303 (vaddudm): Likewise.
14304 (vmaxsd): Likewise.
14305 (vmaxud): Likewise.
14306 (vminsd): Likewise.
14307 (vminud): Likewise.
14308 (vpksdss): Likewise.
14309 (vpksdus): Likewise.
14310 (vpkudum): Likewise.
14311 (vpkudus): Likewise.
14316 (vsubudm): Likewise.
14318 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
14319 support for power8 V2DI instructions.
14321 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
14322 power8 V2DI instructions. Combine pack and unpack insns to use an
14323 iterator for each mode. Check whether a particular mode supports
14324 Altivec instructions instead of just checking TARGET_ALTIVEC.
14325 (UNSPEC_VPKUWUM): Likewise.
14326 (UNSPEC_VPKSHSS): Likewise.
14327 (UNSPEC_VPKSWSS): Likewise.
14328 (UNSPEC_VPKUHUS): Likewise.
14329 (UNSPEC_VPKSHUS): Likewise.
14330 (UNSPEC_VPKUWUS): Likewise.
14331 (UNSPEC_VPKSWUS): Likewise.
14332 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
14333 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
14334 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
14335 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
14336 (UNSPEC_VUPKHSB): Likewise.
14337 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
14338 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
14339 (UNSPEC_VUPKHSH): Likewise.
14340 (UNSPEC_VUPKLSB): Likewise.
14341 (UNSPEC_VUPKLSH): Likewise.
14343 (VI_char): Likewise.
14344 (VI_scalar): Likewise.
14345 (VI_unit): Likewise.
14347 (VP_small): Likewise.
14348 (VP_small_lc): Likewise.
14349 (VU_char): Likewise.
14350 (add<mode>3): Likewise.
14351 (altivec_vaddcuw): Likewise.
14352 (altivec_vaddu<VI_char>s): Likewise.
14353 (altivec_vadds<VI_char>s): Likewise.
14354 (sub<mode>3): Likewise.
14355 (altivec_vsubcuw): Likewise.
14356 (altivec_vsubu<VI_char>s): Likewise.
14357 (altivec_vsubs<VI_char>s): Likewise.
14358 (altivec_vavgs<VI_char>): Likewise.
14359 (altivec_vcmpbfp): Likewise.
14360 (altivec_eq<mode>): Likewise.
14361 (altivec_gt<mode>): Likewise.
14362 (altivec_gtu<mode>): Likewise.
14363 (umax<mode>3): Likewise.
14364 (smax<mode>3): Likewise.
14365 (umin<mode>3): Likewise.
14366 (smin<mode>3): Likewise.
14367 (altivec_vpkuhum): Likewise.
14368 (altivec_vpkuwum): Likewise.
14369 (altivec_vpkshss): Likewise.
14370 (altivec_vpkswss): Likewise.
14371 (altivec_vpkuhus): Likewise.
14372 (altivec_vpkshus): Likewise.
14373 (altivec_vpkuwus): Likewise.
14374 (altivec_vpkswus): Likewise.
14375 (altivec_vpks<VI_char>ss): Likewise.
14376 (altivec_vpks<VI_char>us): Likewise.
14377 (altivec_vpku<VI_char>us): Likewise.
14378 (altivec_vpku<VI_char>um): Likewise.
14379 (altivec_vrl<VI_char>): Likewise.
14380 (altivec_vsl<VI_char>): Likewise.
14381 (altivec_vsr<VI_char>): Likewise.
14382 (altivec_vsra<VI_char>): Likewise.
14383 (altivec_vsldoi_<mode>): Likewise.
14384 (altivec_vupkhsb): Likewise.
14385 (altivec_vupkhs<VU_char>): Likewise.
14386 (altivec_vupkls<VU_char>): Likewise.
14387 (altivec_vupkhsh): Likewise.
14388 (altivec_vupklsb): Likewise.
14389 (altivec_vupklsh): Likewise.
14390 (altivec_vcmpequ<VI_char>_p): Likewise.
14391 (altivec_vcmpgts<VI_char>_p): Likewise.
14392 (altivec_vcmpgtu<VI_char>_p): Likewise.
14393 (abs<mode>2): Likewise.
14394 (vec_unpacks_hi_v16qi): Likewise.
14395 (vec_unpacks_hi_v8hi): Likewise.
14396 (vec_unpacks_lo_v16qi): Likewise.
14397 (vec_unpacks_hi_<VP_small_lc>): Likewise.
14398 (vec_unpacks_lo_v8hi): Likewise.
14399 (vec_unpacks_lo_<VP_small_lc>): Likewise.
14400 (vec_pack_trunc_v8h): Likewise.
14401 (vec_pack_trunc_v4si): Likewise.
14402 (vec_pack_trunc_<mode>): Likewise.
14404 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
14406 (vec_vmaxsd): Likewise.
14407 (vec_vmaxud): Likewise.
14408 (vec_vminsd): Likewise.
14409 (vec_vminud): Likewise.
14410 (vec_vpksdss): Likewise.
14411 (vec_vpksdus): Likewise.
14412 (vec_vpkudum): Likewise.
14413 (vec_vpkudus): Likewise.
14414 (vec_vrld): Likewise.
14415 (vec_vsld): Likewise.
14416 (vec_vsrad): Likewise.
14417 (vec_vsrd): Likewise.
14418 (vec_vsubudm): Likewise.
14419 (vec_vupkhsw): Likewise.
14420 (vec_vupklsw): Likewise.
14422 2013-05-29 Jan Hubicka <jh@suse.cz>
14424 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
14425 flags; reorder rest of fields in more consistent way.
14426 (varpool_node): Remove analyzed, finalized and alias.
14427 (cgraph_ndoe): Likewise.
14428 (symtab_alias_ultimate_target): New function.
14429 (cgraph_function_node): Move offline.
14430 (cgraph_reset_node): Declare.
14431 (cgraph_comdat_can_be_unshared_p): Remove.
14432 (varpool_remove_initializer): Declare.
14433 (varpool_first_defined_variable, varpool_next_defined_variable
14434 cgraph_first_defined_function, cgraph_next_defined_function): Update.
14435 (cgraph_function_with_gimple_body_p): Update.
14436 (varpool_all_refs_explicit_p): Update.
14437 (symtab_alias_target): New function.
14438 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
14439 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
14440 (cgraph_function_or_thunk_node): Simplify using
14441 symtab_alias_ultimate_target.
14442 (varpool_variable_node): Likewise.
14443 * cgraph.c (cgraph_create_function_alias): Update.
14444 (cgraph_add_thunk): Update.
14445 (cgraph_remove_node): Update.
14446 (dump_cgraph_node): Do not dump removed flags.
14447 (cgraph_function_body_availability): Update.
14448 (cgraph_propagate_frequency): Update.
14449 (verify_cgraph_node): Check sanity of local flag.
14450 (cgraph_function_node): Move here from cgraph.h; revamp for
14451 cgraph_function_or_thunk_node.
14452 * lto-symtab.c (lto_varpool_replace_node): Update.
14453 (lto_symtab_resolve_can_prevail_p): Update.
14454 (lto_symtab_merge_cgraph_nodes): Update.
14455 * ipa-cp.c (determine_versionability, initialize_node_lattices,
14456 propagate_constants_accross_call, devirtualization_time_bonus,
14457 ipcp_propagate_stage): Update.
14458 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
14459 * ipa-inline-transform.c (clone_inlined_nodes,
14460 preserve_function_body_p): Update.
14461 * ipa-reference.c (propagate): Update.
14462 (write_node_summary_p): Update.
14463 * toplev.c (wrapup_global_declaration_2): Update.
14464 * cgraphunit.c (cgraph_analyze_function): Rename to ...
14465 (analyze_function) ... this one.
14466 (cgraph_process_new_functions): Update.
14467 (cgraph_reset_node): Export.
14468 (cgraph_finalize_function): Update.
14469 (cgraph_add_new_function): Update.
14470 (process_function_and_variable_attributes): Update.
14471 (varpool_finalize_decl): Update.
14472 (symbol_finalized): Remove.
14473 (symbol_finalized_and_needed): Rename to ...
14474 (symbol_defined_and_needed): ... update.
14475 (cgraph_analyze_functions): Update.
14476 (handle_alias_pairs): Update.
14477 (mark_functions_to_output): Update.
14478 (assemble_thunk): Update.
14479 (output_in_order): Update.
14480 (output_weakrefs): Update.
14481 (finalize_compilation_unit): Update.
14482 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
14483 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
14484 input_node, input_varpool_node): Update.
14485 * dbxout.c (dbxout_expand_expr): Update.
14486 * cgraphclones.c (cgraph_clone_node): Update.
14487 (cgraph_copy_node_for_versioning): Update.
14488 (cgraph_materialize_clone): Update.
14489 (cgraph_materialize_all_clones): Update.
14490 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
14491 propagate_pure_const, propagate_nothrow): Update.
14492 * lto-streamer-out.c (lto_output, write_symbol): Update.
14493 * ipa-utils.c (ipa_reverse_postorder): Update.
14494 * ipa-inline.c (can_inline_edge_p): Update.
14495 (update_caller_keys, ipa_inline): Update.
14496 * dwarf2out.c (reference_to_unused,
14497 premark_types_used_by_global_vars_helper): Update.
14498 * tree-eh.c (tree_could_trap_p): Update.
14499 * ipa-split.c (consider_split, execute_split_functions): Update.
14500 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
14501 has_addr_references_p): Update; move ahead in file for better
14503 (process_references): Simplify.
14504 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
14505 bodies are removed.
14506 (cgraph_comdat_can_be_unshared_p): Make static.
14507 (cgraph_externally_visible_p): Update.
14508 (varpool_externally_visible_p): Update.
14509 (function_and_variable_visibility): Update.
14510 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
14511 ipa_tm_mark_force_output_node): Update.
14512 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
14513 estimate_edge_devirt_benefit, inline_generate_summary,
14514 inline_write_summary): Update.
14515 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
14516 * ipa-prop.c (ipa_compute_jump_functions): Update.
14517 (ipa_print_node_params, ipa_prop_read_section,
14518 ipa_update_after_lto_read, read_replacements_section): Update.
14519 * varasm.c (mark_decl_referenced): Update.
14520 (assemble_alias, dump_tm_clone_pairs): Update.
14521 * tree-inline.c (copy_bb): Update.
14522 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
14524 * symtab.c (dump_symtab_base): Print new flags.
14525 (verify_symtab_base): Verify new flags.
14526 (symtab_alias_ultimate_target): New function.
14527 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
14528 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
14530 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
14532 * i386.c (ix86_get_function_versions_dispatcher,
14533 ix86_generate_version_dispatcher_body): Update.
14534 (fold_builtin_cpu): Use varpool_add_new_variable.
14535 * varpool.c (varpool_remove_initializer): Break out from ...
14536 (varpool_remove_node): ... this one.
14537 (dump_varpool_node, varpool_node_for_asm,
14538 cgraph_variable_initializer_availability, varpool_analyze_node,
14539 varpool_assemble_decl, varpool_remove_unreferenced_decls,
14540 varpool_finalize_named_section_flags, varpool_create_variable_alias):
14543 2013-05-29 Jan Hubicka <jh@suse.cz>
14545 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
14547 2013-05-29 Easwaran Raman <eraman@google.com>
14549 PR tree-optimization/57442
14550 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
14551 when control exits the main loop.
14553 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
14555 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
14557 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
14558 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
14559 * rx/t-rx: Add rx100 under multi library matches option for nofpu
14562 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14564 PR tree-optimization/57441
14565 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
14566 Don't limit size of incr_vec to number of candidates.
14568 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
14570 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
14571 and mips16 directories.
14572 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
14573 (MULTILIB_DIRNAMES): Ditto.
14574 (MULTILIB_EXCEPTIONS): Add new exceptions.
14575 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
14576 (MULTILIB_DIRNAMES): Ditto.
14577 (MULTILIB_EXCEPTIONS): Add new exceptions.
14579 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14580 Marcus Shawcroft <marcus.shawcroft@arm.com>
14582 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
14583 SYMBOL_TINY_ABSOLUTE.
14584 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
14585 SYMBOL_TINY_ABSOLUTE.
14586 (aarch64_expand_mov_immediate): Likewise.
14587 (aarch64_classify_symbol): Likewise.
14588 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
14589 Permit SYMBOL_TINY_ABSOLUTE.
14590 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
14592 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14593 Marcus Shawcroft <marcus.shawcroft@arm.com>
14595 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
14596 Refactor if/switch. Replace gcc_assert with if.
14598 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14600 * config/i386/i386.c (initial_ix86_tune_features): Enable
14601 FP Reassociation for AMD bdver1 and bdver2.
14603 2013-05-29 Martin Jambor <mjambor@suse.cz>
14605 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
14606 and IMAGPART_EXPR do not occur within other handled_components.
14608 2013-05-29 Richard Biener <rguenther@suse.de>
14610 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
14611 access on whether the use is in the BB we currently try to
14613 (vect_bb_vectorization_profitable_p): Pass the BB we currently
14614 vectorize to vect_bb_slp_scalar_cost.
14616 2013-05-29 Richard Biener <rguenther@suse.de>
14618 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
14619 computing scalar cost offsetted by stmts that are kept live
14621 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
14622 for computation of scalar cost.
14624 2013-05-28 Steve Ellcey <sellcey@mips.com>
14626 * config/mips/mips-cpus.def (mips32r2): Change processor type.
14628 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
14630 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
14631 array notation built-in reduction functions.
14632 * doc/passes.texi (Passes): Added documentation about changes done
14634 * doc/invoke.texi (C Dialect Options): Added documentation about
14635 the -fcilkplus flag.
14636 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
14637 (BUILTINS_DEF): Depend on cilkplus.def.
14638 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
14639 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
14640 * cilkplus.def: New file.
14642 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
14644 PR rtl-optimization/57439
14645 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
14647 2013-05-28 Easwaran Raman <eraman@google.com>
14649 PR tree-optimization/57337
14650 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
14651 (find_insert_point): Correctly identify the insertion point
14652 when two statements with the same UID is compared.
14654 2013-05-28 Richard Biener <rguenther@suse.de>
14656 PR tree-optimization/56787
14657 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
14658 from the list of data references.
14659 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
14661 (vect_analyze_loop_operations): Likewise.
14662 (vect_transform_loop): Remove clobbers.
14664 2013-05-28 Martin Jambor <mjambor@suse.cz>
14666 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
14667 and REALPART_EXPRs have scalar type.
14669 2013-05-28 Richard Biener <rguenther@suse.de>
14671 PR tree-optimization/57411
14672 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
14674 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
14675 virtual operand propagation.
14677 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
14679 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
14680 destination register for bmasksi_vis.
14681 (vector_init_bshuffle): Likewise.
14682 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
14684 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
14686 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
14687 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
14688 mode if the instruction isn't available in the original mode.
14689 * config/sparc/sparc.opt (mfix-ut699): New option.
14690 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
14691 (divdf3): Turn into expander.
14692 (divdf3_nofix): New insn.
14693 (divdf3_fix): Likewise.
14694 (divsf3): Disable if -mfix-ut699.
14695 (sqrtdf2): Turn into expander.
14696 (sqrtdf2_nofix): New insn.
14697 (sqrtdf2_fix): Likewise.
14698 (sqrtsf2): Disable if -mfix-ut699.
14700 2013-05-27 Richard Biener <rguenther@suse.de>
14702 PR middle-end/57412
14703 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
14704 block for the new loop.
14706 2013-05-27 Richard Biener <rguenther@suse.de>
14708 PR tree-optimization/57343
14709 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
14710 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
14711 (number_of_iterations_cond): Do not build the folded tree.
14713 2013-05-27 Richard Biener <rguenther@suse.de>
14716 PR middle-end/57381
14717 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
14718 OEP_CONSTANT_ADDRESS_OF retained.
14720 PR tree-optimization/57417
14721 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
14722 for unchanged base.
14723 (set_ssa_val_to): Compare addresses using
14724 get_addr_base_and_unit_offset.
14726 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
14728 PR rtl-optimization/56833
14729 * postreload.c (move2add_record_mode): New function.
14730 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
14731 (move2add_use_add2_insn): Use move2add_record_sym_value.
14732 (move2add_use_add3_insn): Likewise.
14733 (reload_cse_move2add): Use move2add_valid_value_p and
14734 move2add_record_mode. Invalidate call-clobbered and REG_INC
14735 affected regs by setting reg_mode to VOIDmode.
14736 (move2add_note_store): Don't pretend the inside of a SUBREG is
14737 the actual destination. Invalidate single/leading registers by
14738 setting reg_mode to VOIDmode.
14739 Use move2add_record_sym_value, move2add_valid_value_p and
14740 move2add_record_mode.
14742 2013-05-27 Richard Biener <rguenther@suse.de>
14744 PR tree-optimization/57396
14745 * tree-affine.c (double_int_constant_multiple_p): Properly
14746 return false for val == 0 and div != 0.
14748 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14750 * config/mips/mips.h: Use #elif in preprocessor conditions.
14752 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14755 * config/mips/constraints.md (kl): New constraint.
14756 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
14757 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
14758 constraint for operand 0. Split after CSE for MIPS16. Emit a move
14759 from LO for MIPS16.
14760 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
14762 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
14765 * config/mips/mips.c (mips_can_inline_p): New function.
14766 (TARGET_CAN_INLINE_P): Define.
14768 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
14770 * sched-int.h (ds_t, dw_t): Make unsigned int.
14771 Fix documentation that describes how all the ds_t bits are used.
14772 Reserve the last bit for delayed-branch scheduling.
14773 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
14774 (BITS_PER_DEP_WEAK): Fix definition and documentation.
14775 (gen_dep_weak_1): Remove prototype.
14776 * sched-deps.c (get_dep_weak_1): Make static.
14777 * target.def (speculate_insn, needs_block_p, gen_spec_check,
14778 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
14779 * doc/tm.texi: Regenerate.
14780 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
14782 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
14785 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
14787 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
14788 Sandra Loosemore <sandra@codesourcery.com>
14790 * config.gcc (powerpc-*): Allow native for with-cpu.
14792 2013-05-24 Jeff Law <law@redhat.com>
14794 PR tree-optimization/57124
14795 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
14796 conversion feeding a condition if the range has an overflow
14797 if -fstrict-overflow. Add warnings for when we do make the
14800 2013-05-24 Dehao Chen <dehao@google.com>
14802 * tree-cfg.c (locus_discrim_map): Fix the typo.
14803 (locus_discrim_hasher): Likewise.
14804 (locus_discrim_hasher::hash): Likewise.
14805 (locus_discrim_hasher::equal): Likewise.
14807 2013-05-24 Martin Jambor <mjambor@suse.cz>
14809 PR tree-optimization/57294
14810 * cgraph.h (ipa_record_stmt_references): Declare.
14811 * cgraphbuild.c (ipa_record_stmt_references): New function.
14812 (build_cgraph_edges): Use ipa_record_stmt_references.
14813 (rebuild_cgraph_edges): Likewise.
14814 (cgraph_rebuild_references): Likewise.
14815 * ipa-prop.c (ipa_modify_call_arguments): Discard references
14816 associated with the old statement and build references from the
14817 newly built statements.
14818 * ipa-ref.c (ipa_remove_stmt_references): New function.
14819 * ipa-ref.h (ipa_remove_stmt_references): Declare.
14821 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
14823 * lra-constraints.c (emit_spill_move): Use smaller mode for
14825 (check_and_process_move): Consider mem-reg moves for secondary
14827 (curr_insn_transform): Don't lose insns emitted before for
14828 secondary memory moves.
14829 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
14830 reg set up in the current insn.
14832 2013-05-24 Dehao Chen <dehao@google.com>
14834 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
14836 (locus_descrim_hasher::equal): Likewise.
14837 (build_gimple_cfg): New discriminator assignment algorithm.
14838 (make_edges): Likewise.
14839 (next_discriminator_for_locus): Likewise.
14840 (same_line_p): Likewise.
14841 (assign_discriminators): Likewise.
14842 (make_cond_expr_edges): Likewise.
14843 (make_gimple_switch_edges): Likewise.
14844 (make_goto_expr_edges): Likewise.
14845 (make_gimple_asm_edges): Likewise.
14847 2013-05-24 Ian Bolton <ian.bolton@arm.com>
14849 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
14850 X format specifier to only display bottom 16 bits.
14851 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
14852 immediate to match for operand 2, since it will be masked.
14854 2013-05-24 Richard Biener <rguenther@suse.de>
14856 PR tree-optimization/57287
14857 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
14858 all SSA names that occur in abnormal PHIs.
14860 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
14863 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
14864 that index is not negative.
14866 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
14868 PR rtl-optimization/55177
14869 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
14870 (simplify_byte_swapping_operation): New.
14871 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
14872 (simplify_relational_operation_1): Deal with BSWAP.
14874 2013-05-23 Richard Henderson <rth@redhat.com>
14877 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
14878 (ix86_reorg): Call it.
14880 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
14883 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
14884 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
14885 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
14887 2013-05-23 Christian Bruel <christian.bruel@st.com>
14890 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
14892 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14893 Marcus Shawcroft <marcus.shawcroft@arm.com>
14895 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
14896 * config/aarch64/constraints.md (Usa): Remove.
14897 * doc/md.texi (AArch64 Usa): Remove.
14899 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14900 Marcus Shawcroft <marcus.shawcroft@arm.com>
14902 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
14903 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
14904 * config/aarch64/predicates.md (aarch64_const_address): Remove.
14905 (aarch64_mov_operand): Use aarch64_mov_operand_p.
14907 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
14909 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
14910 instruction (AdvSIMD).
14911 * config/aarch64/aarch64-builtins.c
14912 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
14913 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
14915 2013-05-23 Martin Jambor <mjambor@suse.cz>
14917 PR middle-end/57347
14918 * tree.h (contains_bitfld_component_ref_p): Declare.
14919 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
14920 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
14922 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
14923 not access a bit-field. Assert all final offsets are byte-aligned.
14925 2013-05-23 Richard Biener <rguenther@suse.de>
14927 PR tree-optimization/57380
14928 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
14929 least one invariant or re-used load.
14930 * passes.c (init_optimization_passes): Move pass_phiprop before
14933 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
14935 * config/aarch64/aarch64-simd.md
14936 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
14938 2013-05-23 Richard Biener <rguenther@suse.de>
14940 PR middle-end/57381
14941 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
14942 OEP_CONSTANT_ADDRESS_OF retained.
14944 2013-05-23 Jakub Jelinek <jakub@redhat.com>
14946 PR middle-end/57344
14947 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
14948 don't lower unit. Handle unit not being always BITS_PER_WORD.
14950 2013-05-23 Richard Biener <rguenther@suse.de>
14952 PR rtl-optimization/57341
14953 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
14954 instead of true_dependence.
14956 2013-05-22 David Malcolm <dmalcolm@redhat.com>
14958 * bb-reorder.c (branch_threshold): Make const.
14959 (exec_threshold): Ditto.
14961 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14962 Pat Haugen <pthaugen@us.ibm.com>
14963 Peter Bergner <bergner@vnet.ibm.com>
14965 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
14966 documentation for the power8 crypto builtins.
14968 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
14970 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
14971 macros for defining power8 builtin functions.
14972 (BU_P8V_AV_2): Likewise.
14973 (BU_P8V_AV_P): Likewise.
14974 (BU_P8V_VSX_1): Likewise.
14975 (BU_P8V_OVERLOAD_1): Likewise.
14976 (BU_P8V_OVERLOAD_2): Likewise.
14977 (BU_CRYPTO_1): Likewise.
14978 (BU_CRYPTO_2): Likewise.
14979 (BU_CRYPTO_3): Likewise.
14980 (BU_CRYPTO_OVERLOAD_1): Likewise.
14981 (BU_CRYPTO_OVERLOAD_2): Likewise.
14982 (XSCVSPDP): Fix typo, point to the correct instruction.
14983 (VCIPHER): Add power8 crypto builtins.
14984 (VCIPHERLAST): Likewise.
14985 (VNCIPHER): Likewise.
14986 (VNCIPHERLAST): Likewise.
14987 (VPMSUMB): Likewise.
14988 (VPMSUMH): Likewise.
14989 (VPMSUMW): Likewise.
14990 (VPERMXOR_V2DI): Likewise.
14991 (VPERMXOR_V4SI: Likewise.
14992 (VPERMXOR_V8HI: Likewise.
14993 (VPERMXOR_V16QI: Likewise.
14994 (VSHASIGMAW): Likewise.
14995 (VSHASIGMAD): Likewise.
14996 (VPMSUM): Likewise.
14997 (VPERMXOR): Likewise.
14998 (VSHASIGMA): Likewise.
15000 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15001 __CRYPTO__ if the crypto instructions are available.
15002 (altivec_overloaded_builtins): Add support for overloaded power8
15005 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
15006 support for power8 crypto builtins.
15007 (builtin_function_type): Likewise.
15008 (altivec_init_builtins): Add support for builtins that take vector
15009 long long (V2DI) arguments.
15011 * config/rs6000/crypto.md: New file, define power8 crypto
15014 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
15015 Pat Haugen <pthaugen@us.ibm.com>
15016 Peter Bergner <bergner@vnet.ibm.com>
15018 * doc/invoke.texi (Option Summary): Add power8 options.
15019 (RS/6000 and PowerPC Options): Likewise.
15021 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
15022 constraints.md instead of rs6000.h. Reorder w* constraints. Add
15023 wm, wn, wr documentation.
15025 * config/rs6000/constraints.md (wm): New constraint for VSX
15026 registers if direct move instructions are enabled.
15027 (wn): New constraint for no registers.
15028 (wq): New constraint for quad word even GPR registers.
15029 (wr): New constraint if 64-bit instructions are enabled.
15030 (wv): New constraint if power8 vector instructions are enabled.
15031 (wQ): New constraint for quad word memory locations.
15033 * config/rs6000/predicates.md (const_0_to_15_operand): New
15034 constraint for 0..15 for crypto instructions.
15035 (gpc_reg_operand): If VSX allow registers in VSX registers as well
15036 as GPR and floating point registers.
15037 (int_reg_operand): New predicate to match only GPR registers.
15038 (base_reg_operand): New predicate to match base registers.
15039 (quad_int_reg_operand): New predicate to match even GPR registers
15040 for quad memory operations.
15041 (vsx_reg_or_cint_operand): New predicate to allow vector logical
15042 operations in both GPR and VSX registers.
15043 (quad_memory_operand): New predicate for quad memory operations.
15044 (reg_or_indexed_operand): New predicate for direct move support.
15046 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
15047 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
15048 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
15049 (POWERPC_MASKS): Add power8 options.
15050 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
15053 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
15054 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
15056 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
15057 (-mpower8-fusion): New power8 options.
15058 (-mpower8-fusion-sign): Likewise.
15059 (-mpower8-vector): Likewise.
15060 (-mcrypto): Likewise.
15061 (-mdirect-move): Likewise.
15062 (-mquad-memory): Likewise.
15064 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
15065 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
15067 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
15068 (rs6000_debug_vector_unit): Add p8_vector.
15069 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
15070 definitions. Also print fusion state.
15071 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
15072 (rs6000_builtin_mask_calculate): Add power8 builtin support.
15073 (rs6000_option_override_internal): Add support for power8.
15074 (rs6000_common_init_builtins): Add debugging for skipped builtins
15075 if -mdebug=builtin.
15076 (rs6000_adjust_cost): Add power8 support.
15077 (rs6000_issue_rate): Likewise.
15078 (insn_must_be_first_in_group): Likewise.
15079 (insn_must_be_last_in_group): Likewise.
15080 (force_new_group): Likewise.
15081 (rs6000_register_move_cost): Likewise.
15082 (rs6000_opt_masks): Likewise.
15084 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
15085 power8 capable assembler, default to power7 options.
15086 (TARGET_DIRECT_MOVE): Likewise.
15087 (TARGET_CRYPTO): Likewise.
15088 (TARGET_P8_VECTOR): Likewise.
15089 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
15090 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
15091 (VECTOR_MEM_P8_VECTOR_P): Likewise.
15092 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
15093 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
15094 (TARGET_XSCVDPSPN): Likewise.
15095 (TARGET_XSCVSPDPN): Likewsie.
15096 (TARGET_SYNC_HI_QI): Likewise.
15097 (TARGET_SYNC_TI): Likewise.
15098 (MASK_CRYPTO): Likewise.
15099 (MASK_DIRECT_MOVE): Likewise.
15100 (MASK_P8_FUSION): Likewise.
15101 (MASK_P8_VECTOR): Likewise.
15102 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
15103 temporary used by some of the direct move instructions to get two FP
15104 temporary registers does not force creation of a stack frame.
15105 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
15106 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
15107 that any VSX registers are tieable, even if they are also an
15108 Altivec vector mode.
15109 (r6000_reg_class_enum): Add wm, wr, wv constraints.
15110 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
15111 (RS6000_BTM_CRYPTO): Likewise.
15112 (RS6000_BTM_COMMON): Likewise.
15114 * config/rs6000/rs6000.md (cpu attribute): Add power8.
15115 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
15116 (enum rs6000_vector): Add power8 vector support.
15118 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15122 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
15123 (any_sibcall_could_use_r3): this and handle indirect calls.
15124 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
15126 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15128 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
15130 2013-05-22 Richard Biener <rguenther@suse.de>
15132 PR middle-end/57349
15133 * profile.c (branch_prob): Do not split blocks that are
15134 abnormally receiving from ECF_RETURNS_TWICE functions.
15136 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15138 * recog.c (offsettable_address_addr_space_p): Fix calculation of
15139 address mode. Move pointer mode initialization to the same place.
15141 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
15143 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
15144 while it has any effect.
15146 2013-05-21 Easwaran Raman <eraman@google.com>
15148 PR tree-optimization/57322
15149 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
15150 UID of the statement added to the BB to be 1.
15152 2013-05-21 Jakub Jelinek <jakub@redhat.com>
15154 PR tree-optimization/57331
15155 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
15156 of conversion from pointer type to integral type with integer.
15158 2013-05-21 Martin Jambor <mjambor@suse.cz>
15161 * ipa-prop.c (ipa_read_node_info): Process param_used and
15162 controlled_uses in the same order as when writing.
15164 2013-05-21 Magnus Granberg <baldrick@free.fr>
15167 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
15169 2013-05-21 Richard Biener <rguenther@suse.de>
15171 PR tree-optimization/57318
15172 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
15173 estimate stmts with side-effects as likely eliminated.
15175 2013-05-21 Richard Biener <rguenther@suse.de>
15177 PR tree-optimization/57330
15178 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
15179 preserve the call stmts fntype.
15181 2013-05-21 Richard Biener <rguenther@suse.de>
15183 PR tree-optimization/57303
15184 * tree-ssa-sink.c (statement_sink_location): Improve killing
15185 stmt detection and properly handle self-assignments.
15187 2013-05-21 Christian Bruel <christian.bruel@st.com>
15189 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
15190 spanning registers. LEAF_REG_REMAP is supported only for contiguous
15191 registers. Set register size out of the PARALLEL loop.
15193 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
15196 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
15197 (*fmasf4, *fmasf4_media): New insns.
15199 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
15201 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
15202 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
15203 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
15204 (mips_idiv_insns): Update the comments to say that the returned
15205 instruction counts are in units of BASE_INSN_LENGTH.
15206 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
15207 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
15208 using 2 rather than 4 as the length of indirect MIPS16 and
15209 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
15210 length of a NOP. Don't divide MIPS16 lengths by 2.
15211 (mips16_split_long_branches): Assume a branch is long if the
15212 length is greater than 4 rather than 8.
15213 * config/mips/mips.md (length): Give MIPS16 lengths directly,
15214 rather than multiplying them by 2. Multiply instruction counts
15215 by BASE_INSN_LENGTH rather than 4.
15216 (*jump_mips16, tls_get_tp_mips16_<mode>)
15217 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
15219 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
15221 * config/mips/mips.md (extended_mips16): Remove branch case.
15222 (length): Remove duplicated extended_mips16 test.
15224 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
15226 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
15228 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
15230 * recog.h (Recog_data): Rename to...
15231 (recog_data_d): ...this.
15232 (recog_data): Update accordingly.
15233 * recog.c (recog_data): Likewise.
15234 * reload.c (save_recog_data): Likewise.
15235 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
15236 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
15238 2013-05-17 Julian Brown <julian@codesourcery.com>
15240 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
15241 found in a REG_EQUAL note, invalidate it.
15243 2013-05-17 Easwaran Raman <eraman@google.com>
15245 * tree-ssa-reassoc.c (find_insert_point): New function.
15246 (insert_stmt_after): Likewise.
15247 (get_def_stmt): Likewise.
15248 (ensure_ops_are_available): Likewise.
15249 (not_dominated_by): Likewise.
15250 (rewrite_expr_tree): Do not move statements beyond what is
15251 necessary. Remove call to swap_ops_for_binary_stmt...
15252 (reassociate_bb): ... and move it here.
15253 (build_and_add_sum): Assign UIDs for new statements.
15254 (linearize_expr): Likewise.
15255 (do_reassoc): Renumber gimple statement UIDs.
15257 2013-05-17 Jan Hubicka <jh@suse.cz>
15259 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
15261 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
15262 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
15264 (output_weakrefs): Update.
15266 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
15267 Martin Jambor <mjambor@suse.cz>
15269 PR middle-end/57276
15270 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
15271 value that corresponds to the given aggval is found in values vector.
15273 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
15275 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
15276 sse, sse2, sse3, ssse3 and sse4a flags to options.
15278 2013-05-17 David Malcolm <dmalcolm@redhat.com>
15280 * gengtype-state.c: (s_expr_writer): New class, to handle
15281 prettifying of output layout of s-expressions.
15282 (state_writer): New class, to write out gtype.state.
15283 (state_written_type_count): Move this variable into member data of
15285 (s_expr_writer::s_expr_writer): New code: constructor for new class
15286 (state_writer::state_writer(): ditto
15287 (s_expr_writer::write_new_line): New function
15288 (s_expr_writer::write_any_indent): ditto
15289 (s_expr_writer::begin_s_expr): ditto
15290 (s_expr_writer::end_s_expr): ditto
15291 (write_state_fileloc): convert to method of state_writer...
15292 (state_writer:: write_state_fileloc): ...and use methods of
15293 s_expr_writer to write indentation into the gtype.state output file
15294 to visually represent the hierarchical structure of the list
15296 (write_state_fields): ditto, renaming to...
15297 (state_writer::write_state_fields)
15298 (write_state_a_string): ditto, renaming to...
15299 (state_writer::write_state_a_string)
15300 (write_state_string_option): ditto, renaming to...
15301 (state_writer::write_state_string_option)
15302 (write_state_type_option): ditto, renaming to...
15303 (state_writer::write_state_type_option)
15304 (write_state_nested_option): ditto, renaming to...
15305 (state_writer::write_state_nested_option)
15306 (write_state_option): ditto, renaming to...
15307 (state_writer::write_state_option)
15308 (write_state_options): ditto, renaming to...
15309 (state_writer::write_state_options)
15310 (write_state_lang_bitmap): ditto, renaming to...
15311 (state_writer::write_state_lang_bitmap)
15312 (write_state_version): ditto, renaming to...
15313 (state_writer::write_state_version)
15314 (write_state_scalar_type): ditto, renaming to...
15315 (state_writer::write_state_scalar_type)
15316 (write_state_string_type): ditto, renaming to...
15317 (state_writer::write_state_string_type)
15318 (write_state_undefined_type): ditto, renaming to...
15319 (state_writer::write_state_undefined_type)
15320 (write_state_struct_union_type): ditto, renaming to...
15321 (state_writer::write_state_struct_union_type)
15322 (write_state_struct_type): ditto, renaming to...
15323 (state_writer::write_state_struct_type)
15324 (write_state_user_struct_type): ditto, renaming to...
15325 (state_writer::write_state_user_struct_type)
15326 (write_state_lang_struct_type): ditto, renaming to...
15327 (state_writer::write_state_lang_struct_type)
15328 (write_state_param_struct_type): ditto, renaming to...
15329 (state_writer::write_state_param_struct_type)
15330 (write_state_pointer_type): ditto, renaming to...
15331 (state_writer::write_state_pointer_type)
15332 (write_state_array_type): ditto, renaming to...
15333 (state_writer::write_state_array_type)
15334 (write_state_gc_used): ditto, renaming to...
15335 (state_writer::write_state_gc_used)
15336 (write_state_common_type_content): ditto, renaming to...
15337 (state_writer::write_state_common_type_content)
15338 (write_state_type): ditto, renaming to...
15339 (state_writer::write_state_type)
15340 (write_state_pair_list): ditto, renaming to...
15341 (state_writer::write_state_pair_list)
15342 (write_state_pair): ditto, renaming to...
15343 (state_writer::write_state_pair)
15344 (write_state_typedefs): ditto, renaming to...
15345 (state_writer::write_state_typedefs)
15346 (write_state_structures): ditto, renaming to...
15347 (state_writer::write_state_structures)
15348 (write_state_param_structs): ditto, renaming to...
15349 (state_writer::write_state_param_structs)
15350 (write_state_variables): ditto, renaming to...
15351 (state_writer::write_state_variables)
15352 (write_state_srcdir): ditto, renaming to...
15353 (state_writer::write_state_srcdir)
15354 (write_state_files_list): ditto, renaming to...
15355 (state_writer::write_state_files_list)
15356 (write_state_languages): ditto, renaming to...
15357 (state_writer::write_state_languages)
15358 (write_state): create a state_writer instance and use it when
15359 writing out the state file
15361 2013-05-17 Mike Stump <mikestump@comcast.net>
15363 PR rtl-optimization/57304
15364 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
15365 accessing DF_REF_REAL_LOC.
15367 2013-05-17 Jakub Jelinek <jakub@redhat.com>
15369 PR rtl-optimization/57281
15370 PR rtl-optimization/57300
15371 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
15372 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
15373 what the other splitter did if the registers are dead.
15375 2013-05-17 Richard Biener <rguenther@suse.de>
15377 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
15380 2013-05-17 Jakub Jelinek <jakub@redhat.com>
15382 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
15385 2013-05-17 Marek Polacek <polacek@redhat.com>
15387 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
15388 length when doing non-zero store of storing '\0' to '\0'.
15390 2013-05-17 Jakub Jelinek <jakub@redhat.com>
15392 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
15393 vect_external_def oprnd1 with loop_vinfo, try to emit
15394 optional cast, negation and and stmts on the loop preheader
15395 edge instead of into the pattern def seq.
15397 PR tree-optimization/57051
15398 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
15399 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
15401 2013-05-16 Nick Clifton <nickc@redhat.com>
15403 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
15404 (rl78_is_naked_func): New function.
15405 (rl78_expand_prologue): Skip prologue generation for naked functions.
15406 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
15407 * doc/extend.texi (naked): Add RL78 to the list of processors
15408 that supports this attribute.
15410 2013-05-16 Jeff Law <law@redhat.com>
15412 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
15414 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
15416 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
15417 cache parameters using detect_caches_amd also for CYRIX,
15418 NSC and TM2 signatures.
15420 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
15421 Dzianis Kahanovich <mahatma@eu.by>
15425 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
15426 VIA/Centaur processors and determine their cache parameters
15427 using detect_caches_amd.
15429 2013-05-16 Teresa Johnson <tejohnson@google.com>
15431 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
15432 (rtl_verify_edges): New function.
15433 (rtl_verify_bb_insns): Ditto.
15434 (rtl_verify_bb_pointers): Ditto.
15435 (rtl_verify_bb_insn_chain): Ditto.
15436 (rtl_verify_fallthru): Ditto.
15437 (rtl_verify_bb_layout): Ditto.
15438 (rtl_verify_flow_info_1): Outline checks into new functions.
15439 (rtl_verify_flow_info): Ditto.
15441 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
15443 * cfghooks.c (copy_bbs): Add update_dominance argument.
15444 * cfghooks.h (copy_bbs): Update prototype.
15445 * tree-cfg.c (gimple_duplicate_sese_region):
15446 Add update_dominance argument.
15447 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
15448 * tree-ssa-loop-ch.c (copy_loop_headers): Update
15449 gimple_duplicate_sese_region call.
15450 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
15451 Update copy_bbs call.
15452 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
15453 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
15455 2013-05-16 Jakub Jelinek <jakub@redhat.com>
15457 * tree-vectorizer.h (NUM_PATTERNS): Increment.
15458 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
15459 vect_recog_rotate_pattern.
15460 (vect_recog_rotate_pattern): New function.
15462 2013-05-16 Jason Merrill <jason@redhat.com>
15464 * Makefile.in (LLINKER): New variable.
15465 (mostlyclean): Remove link mutex.
15466 * configure.ac: Handle --enable-link-mutex.
15467 * lock-and-run.sh: New script.
15469 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15472 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
15475 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15477 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
15479 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
15481 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
15482 * config/arm/arm.c (next_consecutive_mem): New function.
15483 (gen_movmem_ldrd_strd): Likewise.
15484 * config/arm/arm.md (movmemqi): Update condition and code.
15485 (unaligned_loaddi, unaligned_storedi): New patterns.
15487 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15489 * config.gcc: Obsolete *-*-solaris2.9*.
15490 * doc/install.texi (Specific, *-*-solaris2*): Document it.
15492 2013-05-16 Richard Biener <rguenther@suse.de>
15494 * passes.c (init_optimization_passes): Move pass_parallelize_loops
15495 earlier, after GRAPHITE transforms and IV canonicalization.
15497 2013-05-16 Jakub Jelinek <jakub@redhat.com>
15499 * omp-low.c (extract_omp_for_data): For collapsed loops,
15500 if at least one of the loops is known at compile time to
15501 iterate zero times, set count to 0.
15502 (expand_omp_regimplify_p): New function.
15503 (expand_omp_for_generic): For collapsed loops, if at least
15504 one of the loops isn't known to iterate at least once,
15505 add runtime check with setting count to 0.
15506 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
15507 For unsigned types if it isn't known at compile time that
15508 the loop will iterate at least once, add runtime check to bypass
15509 the whole loop if initial condition isn't true.
15511 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
15513 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
15515 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
15517 PR middle-end/57286
15518 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
15519 transformations to avoid an infinite loop.
15521 2013-05-16 Marek Polacek <polacek@redhat.com>
15523 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
15525 2013-05-15 Leif Ekblad <leif@rdos.net>
15527 * config/i386/i386.c (ix86_decompose_address): Use
15528 DEFAULT_TLS_SEG_REG to access TLS segment register.
15529 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
15530 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
15531 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
15533 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
15536 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
15537 sibling calls to functions that would normally be lazily bound,
15538 unless $gp is call-clobbered.
15540 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
15542 * config/i386/i386.c (ix86_option_override_internal): Update
15543 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
15544 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
15545 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
15546 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
15548 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
15550 2013-05-15 Andreas Schwab <schwab@suse.de>
15552 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
15553 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
15556 2013-05-15 Teresa Johnson <tejohnson@google.com>
15558 * loop-unroll.c (report_unroll_peel): Check decision before
15559 emitting unroll/peel message.
15561 2013-05-15 Teresa Johnson <tejohnson@google.com>
15563 * function.h (has_bb_partition): New rtl_data flag.
15564 (bb_reorder_complete): Ditto.
15565 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
15566 instead of flag_reorder_blocks_and_partition.
15567 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
15568 with some enhancements.
15569 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
15570 * bb-reorder.c (connect_traces): Check for has_bb_partition
15571 instead of flag_reorder_blocks_and_partition.
15572 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
15573 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
15574 verify_hot_cold_block_grouping.
15575 (partition_hot_cold_basic_blocks): Set has_bb_partition.
15577 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15580 * config/arm/predicates.md (call_insn_operand): New predicate.
15581 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
15582 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
15583 if insn is not a tail call.
15584 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
15586 * config/arm/arm.h (enum reg_class): New caller save register class.
15587 (REG_CLASS_NAMES): Likewise.
15588 (REG_CLASS_CONTENTS): Likewise.
15589 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
15592 2013-05-15 Richard Biener <rguenther@suse.de>
15594 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
15595 of MSG_OPTIMIZED_LOCATIONS.
15596 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
15597 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
15599 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
15600 of MSG_OPTIMIZED_LOCATIONS.
15601 (execute_vect_slp): Likewise.
15602 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
15603 (vect_create_cond_for_alias_checks): Likewise.
15604 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
15605 (vect_recog_widen_mult_pattern): Likewise.
15606 (vect_recog_widen_sum_pattern): Likewise.
15607 (vect_recog_over_widening_pattern): Likewise.
15608 (vect_recog_widen_shift_pattern): Likewise.
15609 (vect_recog_vector_vector_shift_pattern): Likewise.
15610 (vect_recog_divmod_pattern): Likewise.
15611 (vect_recog_mixed_size_cond_pattern): Likewise.
15612 (vect_recog_bool_pattern): Likewise.
15613 (vect_pattern_recog_1): Likewise.
15615 2013-05-15 Martin Jambor <mjambor@suse.cz>
15617 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
15618 non-functions to builtin_unreachable.
15619 * ipa-inline-transform.c (inline_call): Do not assert estimates were
15620 correct when new direct edges were discovered.
15622 2013-05-15 Martin Jambor <mjambor@suse.cz>
15624 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
15625 header, print symbol order instead of node uid, print more information
15626 about indirect edge targets.
15627 (ipa_make_edge_direct_to_target): Print symbol order instead of node
15629 (ipa_make_edge_direct_to_target): Likewise.
15630 (remove_described_reference): Likewise.
15631 (propagate_controlled_uses): Likewise.
15632 (ipa_print_node_params): Also print symbol order.
15633 (ipcp_transform_function): Print symbol order instead of node uids.
15634 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
15635 (cgraph_get_create_real_symbol_node): Likewise.
15636 * ipa-cp.c (print_lattice): Likewise.
15637 (print_all_lattices): Likewise.
15638 (determine_versionability): Likewise.
15639 (initialize_node_lattices): Likewise.
15640 (estimate_local_effects): Likewise.
15641 (update_profiling_info): Likewise.
15642 (create_specialized_node): Likewise.
15643 (perhaps_add_new_callers): Likewise.
15644 (decide_about_value): Likewise.
15645 (decide_whether_version_node): Likewise.
15646 (identify_dead_nodes): Likewise.
15647 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
15648 (dump_inline_summary): Likewise.
15649 (estimate_node_size_and_time): Likewise.
15650 (inline_analyze_function): Likewise.
15651 * ipa-inline.c (report_inline_failed_reason): Likewise.
15652 (want_early_inline_function_p): Likewise.
15653 (edge_badness): Likewise.
15654 (update_edge_key): Likewise.
15655 (inline_small_functions): Likewise. Add dumping of order to two other
15657 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
15658 instead of node uids.
15659 (propagate_pure_const): Likewise.
15660 (propagate_pure_const): Likewise.
15661 * ipa-utils.c (dump_cgraph_node_set): Likewise.
15662 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
15663 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
15665 * tree-pretty-print.c (dump_function_header): Likewise.
15666 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
15667 Print symbol order instead of node uids.
15669 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15671 * config/s390/s390.c (s390_register_move_cost): Don't impose the
15672 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
15674 2013-05-15 Richard Biener <rguenther@suse.de>
15676 PR tree-optimization/57275
15677 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
15678 return value for fail to do runtime alias checks for gather loads.
15680 2013-05-15 Jan Hubicka <jh@suse.cz>
15684 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
15685 weakrefs are not external.
15686 (lto_symtab_merge_decls): Fix thinko when dealing with
15687 non-lto_symtab decls.
15688 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
15689 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
15690 * varpool.c (dump_varpool_node): Dump more flags.
15692 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15694 * config/i386/i386.c (processor_alias_table): Add instruction
15695 FSGSBASE for AMD bdver3 architecture.
15697 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15699 * tree.c (warn_deprecated_use): Print file:line using locus color.
15700 * diagnostic.c (diagnostic_report_current_module): Print file:line
15701 and file:line:column using locus color.
15703 2013-05-14 Mike Stump <mikestump@comcast.net>
15705 * gdbinit.in: Add __null.
15707 2013-05-14 Mike Stump <mikestump@comcast.net>
15709 * recog.h: Rename struct recog_data to Recog_data.
15710 * recog.c: Likewise.
15711 * reload.c (can_reload_into): Likewise.
15712 * config/picochip/picochip.c: Likewise.
15714 2013-05-14 Mike Stump <mikestump@comcast.net>
15716 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
15718 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
15720 * resource.h (struct resources): Remove unch_memory member.
15721 (CLEAR_RESOURCE): Don't clear unch_memory.
15722 * resource.c (mark_referenced_resources): Don't set it.
15723 (mark_set_resources): Likewise.
15724 (mark_target_live_regs): Don't clear it.
15725 (init_resource_info): Likewise.
15726 * reorg.c (resource_conflicts_p): Don't compare it.
15727 (redundant_insn): Don't set it.
15729 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
15731 * emit-rtl.c (next_label): Remove unused function.
15732 (skip_consecutive_labels, link_cc0_insns): Move to ...
15733 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
15734 only place where these functions are used, and make them static.
15736 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
15738 * fold-const.c (fold_negate_expr): Handle vectors.
15739 (fold_truth_not_expr): Make it static.
15740 (fold_invert_truthvalue): New static function.
15741 (invert_truthvalue_loc): Handle vectors. Do not call
15742 fold_truth_not_expr directly.
15743 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
15744 <TRUTH_NOT_EXPR>: Do not cast to boolean.
15745 (fold_comparison): Handle vector constants.
15746 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
15747 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
15748 * tree.h (fold_truth_not_expr): Remove declaration.
15750 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
15752 * config/aarch64/aarch64-simd.md
15753 (aarch64_vcond_internal<mode>): Rename to...
15754 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
15755 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
15756 float modes. Clarify all iterator modes.
15757 (vcond<mode><mode>): Use new name for vcond expanders.
15758 (vcond<v_cmp_result><mode>): Likewise.
15759 (vcondu<mode><mode>: Likewise.
15760 * config/aarch64/iterators.md (VDQF_COND): New.
15762 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
15765 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
15766 variable for the shift amount. Check that we shift by non-negative
15769 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
15772 * config/arm/arm.h (EPILOGUE_USES): Only return true
15773 for LR_REGNUM after epilogue_completed.
15775 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
15777 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
15778 is error_mark_node.
15780 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15783 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
15784 and Solaris 11+/x86 with gld.
15785 * configure: Regenerate.
15787 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15789 * expmed.c (expand_shift_1): Canonicalize rotates by
15790 constant bitsize / 2 to bitsize - 1.
15791 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
15792 case ROTATERT>: Likewise.
15795 2013-05-10 Jakub Jelinek <jakub@redhat.com>
15797 * config/i386/i386.md (rotateinv): New code attr.
15798 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
15799 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
15800 roll $31, %eax, etc.
15802 2013-05-14 Richard Biener <rguenther@suse.de>
15804 PR middle-end/57235
15805 * tree-eh.c (sink_clobbers): Give up for successors with
15806 multiple predecessors and no virtual uses.
15808 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
15810 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
15811 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
15813 2013-05-14 Jakub Jelinek <jakub@redhat.com>
15815 PR middle-end/57251
15816 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
15817 the case when both op0 and op1 have VOIDmode.
15819 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15821 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
15822 in multiply-accumulate mode.
15824 2013-05-13 Guozhi Wei <carrot@google.com>
15826 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
15828 2013-05-13 Kai Tietz <ktietz@redhat.com>
15831 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
15832 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
15833 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
15834 * config/i386/i386.c (ix86_option_override_internal): Likewise.
15835 (ix86_expand_prologue): Likewise.
15836 (ix86_expand_split_stack_prologue): Likewise.
15837 (legitimate_pic_address_disp_p): Likewise.
15838 (legitimize_pic_address): Likewise.
15839 (legitimize_tls_address): Likewise.
15840 (legitimize_pe_coff_symbol): Likewise.
15841 (output_pic_addr_const): Likewise.
15842 (construct_plt_address): Likewise.
15843 (ix86_expand_call): Likewise.
15844 (x86_output_mi_thunk): Likewise.
15845 (x86_function_profiler): Likewise.
15847 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
15849 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
15850 similar switch cases.
15851 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
15852 (aarch64_simd_mov_to_<mode>low): Delete.
15853 (aarch64_simd_mov_to_<mode>high): Delete.
15854 (move_lo_quad_<mode>): Add w<-r alternative.
15855 (aarch64_simd_move_hi_quad_<mode>): Likewise.
15856 (aarch64_simd_mov_from_*): Update type attribute.
15857 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
15860 2013-05-13 Jan Hubicka <jh@suse.cz>
15862 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
15863 * config/i386/i386.c (ix86_compute_frame_layout,
15864 ix86_expand_epilogue, emit_i387_cw_initialization,
15865 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
15866 ix86_local_alignment): Fix use of size/speed predicates.
15868 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15870 PR tree-optimization/45216
15871 PR tree-optimization/57157
15872 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
15873 the (-Y) & (B - 1) variant if OP is |.
15874 * expmed.c (expand_shift_1): For rotations by const0_rtx just
15875 return shifted. Use (-op1) & (prec - 1) as other_amount
15876 instead of prec - op1.
15878 2013-05-13 Martin Jambor <mjambor@suse.cz>
15880 PR middle-end/42371
15881 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
15882 (ipa_constant_data): New type.
15883 (ipa_jump_func): Use ipa_constant_data to hold information about
15884 constant jump functions.
15885 (ipa_get_jf_constant): Adjust to jump function type changes.
15886 (ipa_get_jf_constant_rdesc): New function.
15887 (ipa_param_descriptor): New field controlled_uses.
15888 (ipa_get_controlled_uses): New function.
15889 (ipa_set_controlled_uses): Likewise.
15890 * ipa-ref.h (ipa_find_reference): Declare.
15891 * ipa-prop.c (ipa_cst_ref_desc): New type.
15892 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
15894 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
15895 New parameter cs. Adjust all callers.
15896 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
15897 (remove_described_reference): New function.
15898 (jfunc_rdesc_usable): Likewise.
15899 (try_make_edge_direct_simple_call): Decrement controlled use count,
15900 attempt to remove reference if it hits zero.
15901 (combine_controlled_uses_counters): New function.
15902 (propagate_controlled_uses): Likewise.
15903 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
15904 (ipa_edge_duplication_hook): Duplicate reference descriptions.
15905 (ipa_print_node_params): Print described use counter.
15906 (ipa_write_jump_function): Adjust to jump function type changes.
15907 (ipa_read_jump_function): New parameter CS, pass it to
15908 ipa_set_jf_constant. Adjust caller.
15909 (ipa_write_node_info): Stream controlled use count
15910 (ipa_read_node_info): Likewise.
15911 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
15913 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
15914 count. Remove cloning-added reference if it reaches zero.
15915 * ipa-ref.c (ipa_find_reference): New function.
15917 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15919 * config/i386/i386.c (processor_target_table): Modified default
15920 alignment values for AMD BD and BT architectures.
15922 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
15924 * tree-vect-generic.c (uniform_vector_p): Move ...
15925 * tree.c (uniform_vector_p): ... here.
15926 * tree.h (uniform_vector_p): Declare it.
15927 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
15930 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15932 PR tree-optimization/57230
15933 * tree-ssa-strlen.c (handle_char_store): Record length for
15934 array store from STRING_CST.
15936 PR tree-optimization/57230
15937 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
15940 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
15942 * config/epiphany/epiphany.c (epiphany_init): Check size of
15943 NUM_MODES_FOR_MODE_SWITCHING.
15944 (epiphany_expand_prologue):
15945 Remove CONFIG_REGNUM initial value handling code.
15946 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
15947 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
15948 (emit_set_fp_mode, epiphany_mode_after): Likewise.
15949 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
15950 Don't return 1 for FP_MODE_NONE.
15951 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
15952 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
15953 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
15954 * config/epiphany/epiphany.md (save_config): New pattern.
15956 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
15958 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
15960 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
15962 * config/i386/i386.md (memory): Handle sseishft1.
15963 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
15964 (*vec_extractv2di_1): Ditto.
15966 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
15968 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
15971 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
15973 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
15974 Add mthumb/march=armv7-a multilib.
15975 Add mthumb/march=armv7-r multilib.
15976 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
15978 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
15980 * config/v850/t-rtems: Add more multilibs.
15982 2013-05-10 Richard Biener <rguenther@suse.de>
15984 PR tree-optimization/57214
15985 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
15986 not propagate from SSA names that occur in abnormal PHI nodes.
15988 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
15990 * stor-layout.c (element_precision): New function.
15991 * machmode.h (element_precision): Declare it.
15992 * tree.c (build_minus_one_cst): New function.
15993 (element_precision): Likewise.
15994 * tree.h (build_minus_one_cst): Declare new function.
15995 (element_precision): Likewise.
15996 * fold-const.c (operand_equal_p): Use element_precision.
15997 (fold_binary_loc): Handle vector types.
15998 * convert.c (convert_to_integer): Use element_precision.
15999 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
16002 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
16004 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
16005 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
16006 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
16007 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
16008 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
16009 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
16010 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
16011 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
16012 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
16013 (Uuw8): New constraints.
16014 (Usb4): Move into alphabetical order.
16015 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
16016 (sd8_operand, ub8_operand, uw8_operand): New predicates.
16017 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
16018 previously unnamed patterns.
16019 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
16020 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
16021 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
16022 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
16023 of set_attr_alternative/if_then_else. Use extended_mips16 instead
16024 of specific lengths.
16026 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16028 * config/i386/i386.md (rotateinv): New code attr.
16029 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16030 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16031 roll $31, %eax, etc.
16033 PR tree-optimization/45216
16034 PR tree-optimization/57157
16035 * tree-ssa-forwprop.c (simplify_rotate): New function.
16036 (ssa_forward_propagate_and_combine): Call it.
16038 2013-05-10 Richard Biener <rguenther@suse.de>
16040 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
16041 disable peeling when we version for aliasing.
16042 (vector_alignment_reachable_p): Honor explicit user alignment.
16043 (vect_supportable_dr_alignment): Likewise.
16044 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
16045 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
16046 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
16047 then peeling to arrange for the cost-model check to come first.
16049 2013-05-10 Alan Modra <amodra@gmail.com>
16051 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
16052 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
16053 * configure: Regenerate.
16055 2013-05-10 Alan Modra <amodra@gmail.com>
16058 * varasm.c (default_elf_select_section): Move !DECL_P check..
16059 (get_named_section): ..to here before calling get_section_name.
16061 (default_section_type_flags): Add DECL_P check.
16062 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
16063 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
16065 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
16067 * config/epiphany/epiphany.c (epiphany_expand_prologue):
16068 When using gen_stack_adjust_str with a register offset, add a
16069 REG_FRAME_RELATED_EXPR note.
16071 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
16073 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
16074 (*vec_extractv4si_zext_mem): Ditto.
16075 (*vec_extractv2di): Add 0->x and x->x alternatives.
16076 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
16077 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
16079 2013-05-09 Jason Merrill <jason@redhat.com>
16081 N3639 C++1y VLA support
16082 * gimplify.c (gimplify_vla_decl): Don't touch an existing
16085 * tree.c (build_constructor_va): New.
16086 * tree.h: Declare it.
16088 2013-05-09 Martin Jambor <mjambor@suse.cz>
16091 * gimple-fold.c (canonicalize_constructor_val): Call
16092 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
16094 2013-05-09 Jan Hubicka <jh@suse.cz>
16095 Richard Biener <rguenther@suse.de>
16098 * symtab.c (symtab_make_decl_local): Do not add private names.
16100 2013-05-09 Jan Hubicka <jh@suse.cz>
16103 * symtab.c (insert_to_assembler_name_hash): Handle clones.
16104 (unlink_from_assembler_name_hash): Likewise.
16105 (symtab_prevail_in_asm_name_hash, symtab_register_node,
16106 symtab_unregister_node, symtab_initialize_asm_name_hash,
16107 change_decl_assembler_name): Update.
16109 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
16111 * config/aarch64/aarch64.md: New movtf split.
16112 (*movtf_aarch64): Update.
16113 (aarch64_movdi_tilow): Handle TF modes and rename to
16114 aarch64_movdi_<mode>low.
16115 (aarch64_movdi_tihigh): Handle TF modes and rename to
16116 aarch64_movdi_<mode>high
16117 (aarch64_movtihigh_di): Handle TF modes and rename to
16118 aarch64_mov<mode>high_di
16119 (aarch64_movtilow_di): Handle TF modes and rename to
16120 aarch64_mov<mode>low_di
16121 (aarch64_movtilow_tilow): Remove spurious whitespace.
16122 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
16124 (aarch64_print_operand): Update.
16126 2013-05-09 Alan Modra <amodra@gmail.com>
16128 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
16130 * configure: Regenerate.
16132 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
16134 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
16135 splitter preparation statements.
16136 * config/i386/sse.md (*vec_extract* splitters): Ditto.
16137 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
16140 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16142 * gimple-ssa-strength-reduction.c (count_candidates): Change
16143 return value to int.
16144 (analyze_candidates_and_replace): Change type of length to int.
16146 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
16148 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
16149 (*vec_extract<mode>): Use VI12_128 mode iterator.
16150 (*vec_extract<mode>_mem): Ditto.
16151 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
16154 2013-05-08 Diego Novillo <dnovillo@google.com>
16159 2012-08-17 Diego Novillo <dnovillo@google.com>
16162 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
16163 * config.in: Regenerate.
16164 * configure: Regenerate.
16165 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
16167 2013-05-08 Jan Hubicka <jh@suse.cz>
16170 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
16171 * cgraph.h (symtab_node_base): Add unique_name.
16172 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
16173 input_overwrite_node, input_varpool_node): Stream unique_name.
16174 * cgraphclones.c (cgraph_create_virtual_clone,
16175 cgraph_function_versioning): Set unique_name.
16176 * ipa.c (function_and_variable_visibility): Set unique_name.
16178 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16180 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
16181 (alloc_cand_and_find_basis): Restrict conditional candidate
16182 processing to CAND_MULTs.
16184 2013-05-08 Jan Hubicka <jh@suse.cz>
16187 lto-symtab.c (lto_symtab_symbol_p): New function.
16188 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
16189 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
16190 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
16191 Skip static symbols.
16193 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
16195 PR tree-optimization/57200
16196 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
16197 Only call inform if the preceding warning_at returns true.
16199 2013-05-07 Han Shen <shenhan@google.com>
16201 * cfgexpand.c (record_or_union_type_has_array_p): New function.
16202 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
16203 * common.opt (fstack-protector-strong): New option.
16204 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
16205 * doc/invoke.texi (Optimization Options): Document
16206 "-fstack-protector-strong".
16207 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
16209 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
16211 * config/mips/mips.c (mips_machine_reorg2): Return 0.
16213 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
16215 * ira.c (update_equiv_regs): Add insn having equiv memory even if
16216 it is not lhs of the insn.
16217 (setup_reg_equiv): Remove insn having equiv memory which it is not
16219 * lra-constraints.c (process_address): Try to improve generation
16220 code for address base + disp.
16221 (lra_constraints): Make correct the code for checking insn setting
16222 up backward equivalence. Remove insn only if it is in the init
16224 * lra-eliminations.c (update_reg_eliminate): Change return value.
16225 (lra_eliminate): Use the result.
16227 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
16229 * config/i386/sse.md (ssescalarnummask): New mode attribute.
16230 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
16231 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
16232 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
16233 register target operands.
16234 (*vec_extractv8hi_sse2): New pattern.
16235 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
16236 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
16237 (*vec_extract<mode>_mem): New insn and split pattern.
16239 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
16241 * config/arm/arm.c (arm_asan_shadow_offset): New function.
16242 (TARGET_ASAN_SHADOW_OFFSET): Define.
16243 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
16244 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
16246 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16248 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
16249 (incr_vec_index): Return -1 if increment not found.
16250 (create_add_on_incoming_edge): Assert if increment not found.
16251 (record_increment): Limit number of increments recorded.
16252 (all_phi_incrs_profitable): Return false if an increment not found.
16253 (replace_profitable_candidates): Don't process increments that were
16255 (analyze_candidates_and_replace): Limit size of incr_vec.
16257 2013-05-07 Richard Biener <rguenther@suse.de>
16259 * calls.c (special_function_p): setjmp-like functions are leaf.
16260 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
16261 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
16263 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
16265 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
16266 (aarch64_simd_mov<mode>): New expander.
16267 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
16268 (aarch64_simd_mov_to_<mode>high): Likewise.
16269 (aarch64_simd_mov_from_<mode>low): Likewise.
16270 (aarch64_simd_mov_from_<mode>high): Likewise.
16271 (aarch64_dup_lane<mode>): Update.
16272 (aarch64_dup_lanedi): New instruction pattern.
16273 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
16274 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
16276 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16278 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
16279 (replace_mult_candidate): Remove unnecessary argument; remove
16280 unnecessary parameter from call to introduce_cast_before_cand.
16281 (replace_unconditional_candidate): Remove unnecessary parameter
16282 from call to replace_mult_candidate.
16283 (replace_conditional_candidate): Likewise.
16284 (insert_initializers): Use make_temp_ssa_name.
16285 (introduce_cast_before_cand): Remove unnecessary argument; use
16286 make_temp_ssa_name.
16287 (replace_one_candidate): Remove unnecessary argument; remove
16288 unnecessary parameter from calls to introduce_cast_before_cand.
16289 (replace_profitable_candidates): Remove unnecessary parameters
16290 from calls to replace_one_candidate.
16292 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16294 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
16295 phi def as possibly hiding a basis for a CAND_ADD whose operands
16296 have been commuted in the analysis.
16297 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
16299 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16301 * config/aarch64/aarch64.md
16302 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
16303 shift value between 0-4.
16305 2013-05-07 Richard Biener <rguenther@suse.de>
16307 * double-int.h (rshift): New overload.
16308 * double-int.c (rshift): New function.
16309 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
16310 (create_reference_ops_from_ref): Remove.
16311 (vn_reference_insert): Use shared ops for constructing the
16312 reference and copy it.
16314 2013-05-07 Richard Biener <rguenther@suse.de>
16316 PR middle-end/57190
16317 * tree-eh.c (sink_clobbers): Properly propagate
16318 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
16320 2013-05-07 Jakub Jelinek <jakub@redhat.com>
16322 PR tree-optimization/57149
16323 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
16324 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
16325 collect_phi_def_edges, execute_late_warn_uninitialized): Use
16326 uninit_undefined_value_p instead of ssa_undefined_value_p.
16329 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
16330 for modifier == EXPAND_INITIALIZER.
16332 2013-05-07 Anton Blanchard <anton@samba.org>
16334 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
16335 for powerpc64 little endian.
16336 * configure: Regenerate.
16338 2013-05-06 Graham Stott <grahams@btinternet.com>
16340 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
16341 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
16342 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
16343 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
16345 2013-05-06 Graham Stott <grahams@btinternet.com>
16347 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
16348 codes which allow non-lvalues.
16350 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
16352 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
16353 components are all 1s.
16354 (integer_minus_onep): New function.
16355 * tree.h (integer_minus_onep): Declare it.
16356 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
16357 integer_minus_onep instead of integer_all_onesp.
16359 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
16362 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
16363 variations of these patterns.
16365 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
16367 * config/i386/i386.md (isa): Add x64_sse4 member.
16368 (enabled): Handle x64_sse4.
16369 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
16370 instruction for 64bit SSE4_1 targets. Update insn attributes.
16371 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
16372 instruction for SSE4_1 targets. Update insn attributes.
16373 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
16374 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
16375 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
16377 (*vec_extractv4si): Rename from *sse4_1_pextrd.
16378 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
16379 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
16381 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
16384 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
16386 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
16388 * final.c (do_assembler_dialects): Don't handle curly braces and
16389 vertical bar escaped by % as dialect delimiters.
16390 (output_asm_insn): Print curly braces and vertical bar if escaped
16391 by % and ASSEMBLER_DIALECT defined.
16392 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
16393 * doc/tm.texi: Regenerated.
16395 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
16397 * config/mips/mips.c: Include tree-pass.h.
16398 (mips_reorg): Split in pre- and post-dbr_schedule parts.
16399 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
16400 (pass_mips_machine_reorg2): New machine specific pass.
16401 (insert_pass_mips_machine_reorg2): New pass plugin definition.
16402 (mips_option_override): Register the new pass.
16403 * rtl.h (cleanup_barriers): Remove prototype.
16404 (dbr_schedule): Likewise.
16405 * jump.c (cleanup_barriers): Make static.
16406 * reorg.c (dbr_schedule): Likewise.
16408 2013-05-06 Richard Biener <rguenther@suse.de>
16410 PR tree-optimization/57185
16411 * tree-parloops.c (add_field_for_reduction): Handle anonymous
16412 SSA names properly.
16414 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
16417 * config/i386/i386.c (add_parameter_dependencies): Add dependence
16418 between "first_arg" and "insn", not "last" and "insn".
16420 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16422 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
16423 (find_candidates_in_block): Re-enable slsr_process_phi.
16424 (create_phi_basis): Fix double counting of candidate adjustment.
16426 2013-05-06 Richard Biener <rguenther@suse.de>
16428 PR middle-end/57147
16429 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
16430 the edge is also fallthru, preserve it and just clear the
16432 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
16433 also complex, preserve that and just clear the fallthru flag.
16434 * tree-inline.c (update_ssa_across_abnormal_edges): Also
16435 update virtual operands.
16437 2013-05-06 Alan Modra <amodra@gmail.com>
16439 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
16440 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
16441 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
16442 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
16443 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
16444 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
16446 2013-05-06 Alan Modra <amodra@gmail.com>
16448 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
16449 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
16450 (DEFAULT_ASM_ENDIAN): Define.
16451 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
16452 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
16453 Update -K PIC clause from sysv4.h.
16454 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
16455 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
16457 2013-05-06 Alan Modra <amodra@gmail.com>
16459 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
16460 twice for little-endian.
16461 (ashrdi3_no_power, ashrdi3): Support little-endian.
16463 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
16466 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
16467 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
16469 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
16470 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
16471 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
16474 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
16475 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16477 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
16478 * config/pa/pa.opt: Make mbig-switch a no-op.
16479 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
16480 (CASE_VECTOR_MODE): Always return SImode.
16481 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
16482 for the !TARGET_BIG_SWITCH case.
16483 * config/pa/pa-linux.h: Likewise.
16484 * config/pa/pa-openbsd.h: Likewise.
16485 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
16486 * config/pa/pa.md (short_jump): Remove define_insn.
16487 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
16488 (casesi0): Remove define_insn.
16489 (type): Remove btable_branch.
16490 (pa_combine_type): Likewise.
16491 (in_nullified_branch_delay): Likewise.
16492 (in_call_delay): Likewise.
16493 (define_delay): Likewise.
16494 (define_insn_reservation "Z3"): Likewise.
16495 (define_insn_reservation "Z4"): Likewise.
16496 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
16497 (pa_adjust_insn_length): Remove adjustment for btable branches.
16498 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
16501 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
16503 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
16504 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
16505 Add m->r,x alternatives.
16506 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
16507 splitters using SWI48x mode iterator.
16508 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
16509 TARGET_64BIT. Add m->x alternative.
16510 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
16511 Add o->x alternative. Enable for TARGET_SSE.
16512 (sse_storeq): Remove expander.
16513 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
16514 with memory input operand.
16515 (*vec_extractv2di_1 splitter): New.
16516 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
16517 * config/i386/i386.md (ssevecmodelower): New mode attribute.
16519 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
16521 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
16522 (INT_LOWPART): Delete.
16523 (extract_MB): Adjust.
16524 (extract_ME): Adjust.
16525 (print_operand): Adjust.
16527 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
16529 * config/rs6000/predicates.md (reg_or_add_cint_operand,
16530 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
16531 (reg_or_logical_cint_operand, easy_fp_constant,
16532 logical_const_operand): Delete "CONST_DOUBLE" case.
16533 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
16534 "HOST_BITS_PER_WIDE_INT == 64" test.
16535 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
16536 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
16537 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
16538 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
16540 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
16541 CONST_DOUBLE DImode/VOIDmode case.
16542 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
16543 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
16544 CONST_DOUBLE VOIDmode case.
16545 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
16546 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
16547 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
16548 Delete CONST_DOUBLE case.
16549 (splitters for mov FMOVE64 const_double): Delete
16550 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
16551 "HOST_BITS_PER_WIDE_INT >= 64" test.
16552 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
16554 (mov DI const_double): Delete.
16556 2013-05-04 Jakub Jelinek <jakub@redhat.com>
16558 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
16559 on op shows all bits zero in mode of a lowpart subreg, return zero.
16561 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
16564 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
16565 to save TFmode registers and DImode to save TImode registers for
16566 caller save operations.
16567 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
16568 mark being partially clobbered since they only use the first
16571 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
16572 and TDmode only use the upper 64-bits of each VSX register.
16574 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16576 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
16577 (find_candidates_in_block): Disable slsr_process_phi.
16579 2013-05-03 Guozhi Wei <carrot@google.com>
16581 * coverage.c (coverage_obj_init): Move the construction of gcov
16583 (build_init_ctor): ... here.
16585 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16587 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
16588 (slsr_cand_d): Redefine def_phi.
16589 (stride_status, phi_adjust_status, count_phis_status): New enums.
16590 (find_phi_def): New.
16591 (find_basis_for_base_expr): New.
16592 (find_basis_for_candidate): Handle hidden bases.
16593 (alloc_cand_and_find_basis): Handle phi candidates.
16594 (slsr_process_phi): New.
16595 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
16596 (create_mul_imm_cand): Likewise.
16597 (create_add_ssa_cand): Exclude phi base candidates.
16598 (create_add_imm_cand): Likewise.
16599 (slsr_process_cast): Likewise.
16600 (slsr_process_copy): Likewise.
16601 (find_candidates_in_block): Handle phi candidates.
16602 (dump_candidate): Likewise.
16603 (unconditional_cands): Delete.
16604 (unconditional_cands_with_known_stride_p): Delete.
16605 (phi_dependent_cand_p): New.
16606 (cand_increment): Handle phi-dependent candidates.
16607 (replace_dependent): Delete.
16608 (replace_mult_candidate): New.
16609 (replace_unconditional_candidate): New.
16610 (incr_vec_index): Move to avoid forward reference.
16611 (create_add_on_incoming_edge): New.
16612 (create_phi_basis): New.
16613 (replace_dependents): Delete.
16614 (replace_conditional_candidate): New.
16615 (phi_add_costs): New.
16616 (replace_uncond_cands_and_profitable_phis): New.
16617 (record_increment): Handle phi adjustments.
16618 (record_phi_increments): New.
16619 (record_increments): Handle phi adjustments.
16620 (phi_incr_cost): New.
16621 (lowest_cost_path): Handle phis.
16622 (total_savings): Likewise.
16623 (analyze_increments): Likewise.
16624 (ncd_with_phi): New.
16625 (ncd_of_cand_and_phis): New.
16626 (nearest_common_dominator_for_cands): Handle phi increments.
16627 (all_phi_incrs_profitable): New.
16628 (replace_profitable_candidates): Handle phi-dependent candidates.
16629 (analyze_candidates_and_replace): Likewise.
16631 2013-05-03 Teresa Johnson <tejohnson@google.com>
16634 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
16635 do not exceed REG_BR_PROB_BASE.
16637 2013-05-03 Jeff Law <law@redhat.com>
16639 PR tree-optimization/57144
16640 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
16641 operand of the condition will bit into the new type when eliminating
16642 a cast feeding a condition.
16644 2013-05-03 Jakub Jelinek <jakub@redhat.com>
16646 PR rtl-optimization/57130
16647 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
16648 of COMPARE as in_code to the recursive call if needed.
16650 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
16652 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
16653 (enabled): Handle new members.
16654 * config/i386/sse.md (*vec_concatv2si): Merge from
16655 *vec_concatv2si_sse2 and vec_concatv2si_sse.
16656 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
16658 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
16660 PR tree-optimization/57027
16661 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
16662 for fnms opportunity, check we got the prerequisite kind
16663 of tree / gimple before using accessor functions.
16665 2013-05-03 Richard Biener <rguenther@suse.de>
16667 * double-int.h (lshift): New overload without precision
16668 and arith argument.
16669 (operator *=, operator +=, operator -=): Move ...
16670 * double-int.c (operator *=, operator +=, operator -=): ... here
16671 and implement more efficiently.
16672 (mul_double_with_sign): Remove.
16673 (lshift_double): Adjust to take unsinged shift argument, push
16674 dispatching code to callers.
16675 (mul_double_wide_with_sign): Add early out for callers that
16676 are not interested in high parts or overflow.
16677 (lshift): New function.
16678 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
16679 dispatch code here.
16680 (lrotate, rrotate): Use logical shifts.
16681 * expr.c (get_inner_reference): Use lshift.
16682 * fixed-value.c (do_fixed_divide): Likewise.
16683 * tree-dfa.c (get_ref_base_and_extent): Likewise.
16684 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
16685 (indirect_refs_may_alias_p): Likewise.
16686 (stmt_kills_ref_p_1): Likewise.
16688 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
16690 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
16692 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
16694 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
16695 scalar form of FABD instruction.
16697 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
16699 * lra-constraints.c (process_alt_operands): Add checking alt
16700 number to choose the best alternative.
16702 2013-05-02 Richard Biener <rguenther@suse.de>
16704 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
16705 bitmap and its handling.
16706 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
16708 2013-05-02 Richard Biener <rguenther@suse.de>
16710 PR middle-end/57140
16711 * tree-inline.c (copy_loops): Properly handle removed loops.
16712 (copy_cfg_body): Mark destination loops for fixup if source
16713 loops needed fixup.
16715 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
16718 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
16719 generating simple_return for naked functions.
16721 2013-05-02 Martin Jambor <mjambor@suse.cz>
16723 PR middle-end/56988
16724 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
16725 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
16727 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
16728 ipa_agg_replacement_value structures.
16729 (known_aggs_to_agg_replacement_list): Likewise.
16730 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
16731 (read_agg_replacement_chain): Likewise.
16732 (ipcp_transform_function): Also check that by_ref flags match.
16734 2013-05-02 Richard Biener <rguenther@suse.de>
16736 * graphds.h (struct graph): Add obstack member.
16737 * graphds.c (new_graph): Initialize obstack and allocate
16739 (add_edge): Allocate edge from the obstack.
16740 (free_graph): Free the obstack instead of all edges and vertices.
16742 2013-05-02 Teresa Johnson <tejohnson@google.com>
16744 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
16746 * cfg.c (update_bb_profile_for_threading): Ditto.
16747 * tree-inline.c (copy_bb): Ditto.
16748 (copy_edges_for_bb): Ditto.
16749 (initialize_cfun): Ditto.
16750 (copy_cfg_body): Ditto.
16751 (expand_call_inline): Ditto.
16752 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
16753 (estimate_node_size_and_time): Ditto.
16754 (inline_merge_summary): Ditto.
16755 * cgraphclones.c (cgraph_clone_edge): Ditto.
16756 (cgraph_clone_node): Ditto.
16757 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16758 (compute_trg_info): Ditto.
16760 2013-05-02 Ian Bolton <ian.bolton@arm.com>
16762 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
16763 S reg when fp attribute set.
16764 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
16766 2013-05-02 Ian Bolton <ian.bolton@arm.com>
16768 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
16770 (*and_one_cmplsi3_compare0_uxtw): Likewise.
16771 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
16772 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
16774 2013-05-02 Richard Biener <rguenther@suse.de>
16776 * tree-scalar-evolution.c (scev_info_hasher): Remove.
16777 (struct instantiate_cache_entry): New type.
16778 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
16779 (struct instantiate_cache_type): New type.
16780 (set_instantiated_value, get_instantiated_value): Remove.
16781 (get_instantiated_value_entry): New function.
16782 (instantiate_scev_name): Use the new cache and adjust.
16783 (instantiate_scev_poly): Adjust.
16784 (instantiate_scev_binary): Likewise.
16785 (instantiate_array_ref): Likewise.
16786 (instantiate_scev_convert): Likewise.
16787 (instantiate_scev_not): Likewise.
16788 (instantiate_scev_3): Likewise.
16789 (instantiate_scev_2): Likewise.
16790 (instantiate_scev_r): Likewise.
16791 (instantiate_scev): Likewise.
16792 (resolve_mixers): Likewise.
16794 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
16797 * lra-constraints.c (best_small_class_operands_num): Remove.
16798 (process_alt_operands): Remove small_class_operands_num. Take
16799 small classes operands into losers and only if the operand is not
16800 matched. Modify debugging output.
16801 (curr_insn_transform): Remove best_small_class_operands_num.
16804 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16806 * config/aarch64/aarch64-builtins.c
16807 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
16808 * config/aarch64/aarch64-simd-builtins.def
16809 (reduc_splus_): Add new modes.
16810 (reduc_uplus_): New.
16811 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
16812 (reduc_uplus_v4sf): Likewise.
16813 (reduc_splus_v4sf): Likewise.
16814 (aarch64_addv<mode>): Likewise.
16815 (reduc_uplus_<mode>): Likewise.
16816 (reduc_splus_<mode>): Likewise.
16817 (aarch64_addvv2di): Likewise.
16818 (reduc_uplus_v2di): Likewise.
16819 (reduc_splus_v2di): Likewise.
16820 (aarch64_addvv2si): Likewise.
16821 (reduc_uplus_v2si): Likewise.
16822 (reduc_splus_v2si): Likewise.
16823 (reduc_<sur>plus_<mode>): New.
16824 (reduc_<sur>plus_v2di): Likewise.
16825 (reduc_<sur>plus_v2si): Likewise.
16826 (reduc_<sur>plus_v4sf): Likewise.
16827 (aarch64_addpv4sf): Likewise.
16828 * config/aarch64/arm_neon.h
16829 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
16830 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
16831 add UNSPEC_SADDV, UNSPEC_UADDV.
16833 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
16835 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16837 * config/aarch64/arm_neon.h
16838 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
16840 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16842 * config/aarch64/aarch64-builtins
16843 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
16845 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16847 * config/aarch64/aarch64-simd-builtins.def
16848 (reduc_smax_): New.
16849 (reduc_smin_): Likewise.
16850 (reduc_umax_): Likewise.
16851 (reduc_umin_): Likewise.
16852 (reduc_smax_nan_): Likewise.
16853 (reduc_smin_nan_): Likewise.
16856 (smax): Update for V2SF, V4SF and V2DF modes.
16859 (smin_nan): Likewise.
16860 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
16861 (<su><maxmin><mode>3): ...This, refactor.
16862 (s<maxmin><mode>3): New.
16863 (<maxmin_uns><mode>3): Likewise.
16864 (reduc_<maxmin_uns>_<mode>): Refactor.
16865 (reduc_<maxmin_uns>_v4sf): Likewise.
16866 (reduc_<maxmin_uns>_v2si): Likewise.
16867 (aarch64_<fmaxmin><mode>: Remove.
16868 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
16870 (vmin<q>_f<32,64>): Likewise.
16871 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16873 (su): Add mappings for smax, smin, umax, umin.
16875 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16876 (FMAXMIN): Rename as...
16877 (FMAXMIN_UNS): ...This.
16879 (fmaxminv): Likewise.
16880 (fmaxmin): Likewise.
16882 (maxmin_uns_op): Likewise.
16884 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16886 * config/aarch64/arm_neon.h
16887 (vac<ge, gt><sd>_f<32, 64>): Rename to...
16888 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
16889 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
16891 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16893 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
16894 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
16896 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16898 * config/aarch64/aarch64-simd.md
16899 (vcond<mode>_internal): Handle special cases for constant masks.
16900 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
16901 (vcondu<mode><mode>): Likewise.
16902 (vcond<v_cmp_result><mode>): New.
16904 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16906 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
16907 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
16908 * config/aarch64/aarch64-simd-builtins.def
16909 (cmeq): Update to BUILTIN_VALLDI.
16914 * config/aarch64/arm_neon.h
16915 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
16916 to builtins or C as appropriate.
16918 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16920 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
16922 (cmhi): Rename to...
16924 * config/aarch64/aarch64-simd.md
16925 (simd_mode): Add SF.
16926 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
16927 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
16928 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
16929 (cstore<mode>_neg): ...This.
16930 * config/aarch64/iterators.md
16932 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
16933 (COMPARISONS): New.
16934 (UCOMPARISONS): Likewise.
16935 (optab): Add missing comparisons.
16941 (VCMP_S): Likewise.
16942 (VCMP_U): Likewise.
16943 (V_cmp_result): Add DF, SF modes.
16944 (v_cmp_result): Likewise.
16946 (vmtype): Likewise.
16947 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
16949 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
16951 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
16952 define_insn to define_insn_and_split.
16953 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
16954 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
16955 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
16956 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
16957 (thumb2_negscc): Likewise.
16959 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16961 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
16963 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16965 * config/arm/thumb2.md: Remove trailing whitespaces.
16967 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16969 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
16970 Use gen_int_mode rather than GEN_INT.
16972 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
16974 * value-prof.c (stream_in_histogram_value): Remove the strayed
16977 2013-04-30 Richard Biener <rguenther@suse.de>
16979 PR middle-end/57122
16980 * cfghooks.c (split_edge): Properly check for the loop latch edge.
16982 2013-04-30 Richard Biener <rguenther@suse.de>
16984 PR middle-end/57107
16985 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
16987 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
16989 PR rtl-optimization/56957
16990 PR rtl-optimization/57105
16991 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
16992 variable. Use just INSN_UID for determining whether an insn
16993 should be only disconnected from the insn stream.
16994 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
16996 2013-04-30 Jakub Jelinek <jakub@redhat.com>
16998 PR tree-optimization/57104
16999 * tsan.c (instrument_expr): Don't instrument accesses to
17000 DECL_HARD_REGISTER VAR_DECLs.
17002 2013-04-30 Richard Biener <rguenther@suse.de>
17004 * function.h (loops_for_fn): New inline function.
17005 (set_loops_for_fn): Likewise.
17006 * cfgloop.h (place_new_loop): Add struct function parameter.
17007 (get_loop): Likewise.
17008 (get_loops): Likewise.
17009 (number_of_loops): Likewise.
17010 (fel_next): Adjust.
17011 (fel_init): Likewise.
17012 * cfg.c (get_loop_copy): Adjust.
17013 * cfgloop.c (flow_loops_dump): Likewise.
17014 (record_loop_exits): Likewise.
17015 (verify_loop_structure): Likewise.
17016 * cfgloopanal.c (mark_irreducible_loops): Likewise.
17017 (estimate_reg_pressure_cost): Likewise.
17018 (mark_loop_exit_edges): Likewise.
17019 * cfgloopmanip.c (place_new_loop): Likewise.
17020 (add_loop): Likewise.
17021 (duplicate_loop): Likewise.
17022 * graph.c (draw_cfg_nodes): Likewise.
17023 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
17024 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
17025 (extract_affine_chrec): Likewise.
17026 (build_scop_iteration_domain): Likewise.
17027 * graphite.c (graphite_initialize): Likewise.
17028 * ira-build.c (create_loop_tree_nodes): Likewise.
17029 (more_one_region_p): Likewise.
17030 (rebuild_regno_allocno_maps): Likewise.
17031 (mark_loops_for_removal): Likewise.
17032 (mark_all_loops_for_removal): Likewise.
17033 (remove_unnecessary_regions): Likewise.
17034 (ira_build): Likewise.
17035 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
17036 * loop-init.c (fix_loop_structure): Likewise.
17037 (gate_rtl_move_loop_invariants): Likewise.
17038 (gate_rtl_unswitch): Likewise.
17039 (gate_rtl_unroll_and_peel_loops): Likewise.
17040 (rtl_doloop): Likewise.
17041 * lto-streamer-in.c (input_cfg): Likewise.
17042 * lto-streamer-out.c (output_cfg): Likewise.
17043 * modulo-sched.c (sms_schedule): Likewise.
17044 * predict.c (tree_estimate_probability): Likewise.
17045 (tree_estimate_probability_driver): Likewise.
17046 (estimate_loops): Likewise.
17047 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
17048 (move_sese_region_to_fn): Likewise.
17049 (debug_loop_num): Likewise.
17050 * tree-chrec.c (chrec_evaluate): Likewise.
17051 (hide_evolution_in_other_loops_than_loop): Likewise.
17052 (chrec_component_in_loop_num): Likewise.
17053 (reset_evolution_in_loop): Likewise.
17054 (evolution_function_is_invariant_rec_p): Likewise.
17055 * tree-if-conv.c (main_tree_if_conversion): Likewise.
17056 * tree-inline.c (copy_loops): Likewise.
17057 (copy_cfg_body): Likewise.
17058 (tree_function_versioning): Likewise.
17059 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
17060 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
17062 (add_to_evolution_1): Likewise.
17063 (scev_const_prop): Likewise.
17064 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
17065 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
17066 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
17067 (tree_ssa_lim_initialize): Likewise.
17068 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
17069 (verify_loop_closed_ssa): Likewise.
17070 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
17071 (tree_ssa_loop_im): Likewise.
17072 (tree_ssa_loop_unswitch): Likewise.
17073 (tree_vectorize): Likewise.
17074 (check_data_deps): Likewise.
17075 (tree_ssa_loop_ivcanon): Likewise.
17076 (tree_ssa_loop_bounds): Likewise.
17077 (tree_complete_unroll): Likewise.
17078 (tree_complete_unroll_inner): Likewise.
17079 (tree_parallelize_loops): Likewise.
17080 (tree_ssa_loop_prefetch): Likewise.
17081 (tree_ssa_loop_ivopts): Likewise.
17082 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
17083 * tree-vectorizer.c (vectorize_loops): Likewise.
17085 2013-04-29 Mike Frysinger <vapier@gentoo.org>
17087 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
17088 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
17089 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
17090 with EABI_LINK_SPEC.
17092 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17095 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
17098 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
17101 * lra-constraints.c (process_alt_operands): Discourage a bit more
17102 using memory for pseudos. Print cost dump for alternatives.
17103 Modify cost values for conflicts with early clobbers.
17104 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
17106 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17109 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
17111 2013-04-29 Ian Bolton <ian.bolton@arm.com>
17113 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
17114 from/to S register.
17115 (movdi_aarch64): Support LDR/STR from/to D register.
17117 2013-04-29 Ian Bolton <ian.bolton@arm.com>
17119 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
17120 or higher by default.
17122 2013-04-29 Richard Biener <rguenther@suse.de>
17124 PR middle-end/57075
17125 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
17126 even if not adding abnormal edges for calls that can make
17129 2013-04-29 Richard Biener <rguenther@suse.de>
17131 PR middle-end/57103
17132 * tree-cfg.c (move_stmt_op): Fix condition under which to update
17134 (move_stmt_r): Remove redundant checking.
17136 2013-04-29 Teresa Johnson <tejohnson@google.com>
17139 * basic-block.h (apply_scale): New function.
17140 (apply_probability): Use apply_scale.
17141 * gimple-streamer-in.c (input_bb): Ditto.
17142 * lto-streamer-in.c (input_cfg): Ditto.
17143 * lto-cgraph.c (merge_profile_summaries): Ditto.
17144 * tree-optimize.c (execute_fixup_cfg): Ditto.
17145 * tree-inline.c (copy_bb): Update comment to use apply_scale.
17146 (copy_edges_for_bb): Ditto.
17147 (copy_cfg_body): Ditto.
17149 2013-04-29 Tom de Vries <tom@codesourcery.com>
17151 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
17152 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
17153 (tail_merge_optimize): Handle current_loops == NULL.
17155 2013-04-26 Jeff Law <law@redhat.com>
17157 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
17158 (simplify_cond_using_ranges): Generalize code to simplify
17159 COND_EXPRs where one argument is a constant and the other
17160 is an SSA_NAME created by an integral type conversion.
17162 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17164 * config/arm/arm.md (store_minmaxsi): Use only when
17165 optimize_insn_for_size_p.
17167 2013-04-29 Christian Bruel <christian.bruel@st.com>
17170 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
17172 2013-04-29 Richard Biener <rguenther@suse.de>
17174 PR middle-end/57089
17175 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
17176 loop tree make sure to schedule a fixup for the child as well.
17177 (expand_omp_for_generic): Properly add loops.
17178 (expand_omp_for_static_nochunk): Likewise.
17179 (expand_omp_for_static_chunk): Likewise.
17180 (expand_omp_for): For the degenerate case fixup loops.
17181 (expand_omp_sections): Fix default bb placement in loops.
17182 (expand_omp_atomic_pipeline): Properly add loops.
17184 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17186 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
17188 2013-04-29 Tom de Vries <tom@codesourcery.com>
17190 * tree-ssa-tail-merge.c: Update header comment.
17192 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17194 * config/aarch64/arm_neon.h
17195 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
17196 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
17197 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
17198 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
17199 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
17200 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
17201 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
17202 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
17204 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17206 * config/aarch64/aarch64-simd.md
17207 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
17208 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
17209 fix_trunc, fixuns_trunc.
17210 (ftrunc<VDQF:mode>2): New.
17211 * config/aarch64/iterators.md (optab): Add fix, fixuns.
17212 (fix_trunc_optab): New.
17214 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17216 * config/aarch64/aarch64-builtins.c
17217 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
17218 iceilf, lround, iroundf.
17220 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17223 * config/i386/i386.h (enum ix86_tune_indices)
17224 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
17225 New, split from X86_TUNE_INTER_UNIT_MOVES.
17226 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
17227 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
17228 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
17229 (TARGET_INTER_UNIT_MOVES): Remove.
17230 * config/i386/i386.c (initial_ix86_tune_features): Update.
17231 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
17232 (ix86_expand_convert_uns_didf_sse): Use
17233 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
17234 (ix86_expand_vector_init_one_nonzero): Ditto.
17235 (ix86_expand_vector_init_interleave): Ditto.
17236 (inline_secondary_memory_needed): Return true for moves from SSE class
17237 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
17238 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
17239 * config/i386/constraints.md (Yi, Ym): Depend on
17240 TARGET_INTER_UNIT_MOVES_TO_VEC.
17241 (Yj, Yn): New constraints.
17242 * config/i386/i386.md (*movdi_internal): Change constraints of
17243 operand 1 from Yi to Yj and from Ym to Yn.
17244 (*movsi_internal): Ditto.
17245 (*movdf_internal): Ditto.
17246 (*movsf_internal): Ditto.
17247 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
17248 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
17249 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
17250 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
17251 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
17252 * config/i386/sse.md (movdi_to_sse): Ditto.
17253 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
17254 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
17255 TARGET_INTER_UNIT_MOVES.
17256 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
17257 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
17258 instead of TARGET_INTER_UNIT_MOVES.
17259 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
17260 operand 1 from Yi to Yj and from Ym to Yn.
17262 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17264 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
17265 (float_truncate_hi_): Likewise.
17266 (float_extend_lo_): Likewise.
17267 (float_truncate_lo_): Likewise.
17268 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
17269 (aarch64_float_extend_lo_v2df): Likewise.
17270 (vec_unpacks_hi_v4sf): Likewise.
17271 (aarch64_float_truncate_lo_v2sf): Likewise.
17272 (aarch64_float_truncate_hi_v4sf): Likewise.
17273 (vec_pack_trunc_v2df): Likewise.
17274 (vec_pack_trunc_df): Likewise.
17276 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17278 * config/aarch64/aarch64-builtins.c
17279 (aarch64_fold_builtin): Fold float conversions.
17280 * config/aarch64/aarch64-simd-builtins.def
17281 (floatv2si, floatv4si, floatv2di): New.
17282 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
17283 * config/aarch64/aarch64-simd.md
17284 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
17285 * config/aarch64/iterators.md (FLOATUORS): New.
17286 (optab): Add float, floatuns.
17287 (su_optab): Likewise.
17289 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17291 * config/aarch64/aarch64-builtins.c
17292 (aarch64_builtin_vectorized_function): Use new names for
17294 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
17295 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
17296 (fcvtzu): Split as...
17297 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
17298 (fcvtas): Split as...
17299 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
17300 (fcvtau): Split as...
17301 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
17302 (fcvtps): Split as...
17303 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
17304 (fcvtpu): Split as...
17305 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
17306 (fcvtms): Split as...
17307 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
17308 (fcvtmu): Split as...
17309 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
17310 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
17311 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
17312 (lfrintnusf, lfrintnudf): Likewise.
17313 * config/aarch64/aarch64-simd.md
17314 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
17316 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
17317 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
17318 (fcvt_pattern): Likewise.
17320 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17322 * config/aarch64/aarch64-simd.md
17323 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
17324 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
17326 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17328 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
17329 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
17330 (vrnd<a,m,n,p>_f32): Implement using builtins.
17331 (vrnd<i,x><q>_f<32, 64>): New.
17333 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
17335 * config/aarch64/aarch64-builtins.c
17336 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
17337 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
17338 (frintz): Rename to...
17340 (frintp): Rename to...
17342 (frintm): Rename to...
17344 (frinti): Rename to...
17345 (nearbyint): ...this.
17346 (frintx): Rename to...
17348 (frinta): Rename to...
17350 * config/aarch64/aarch64-simd.md
17351 (aarch64_frint<frint_suffix><mode>): Delete.
17352 (<frint_pattern><mode>2): Convert to insn.
17353 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
17354 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
17355 (frint_pattern): Likewise.
17356 (frint_suffix): Likewise.
17358 2013-04-29 Richard Biener <rguenther@suse.de>
17360 PR tree-optimization/57081
17361 * loop-init.c: Include tree-flow.h.
17362 (loop_optimizer_finalize): Free number of iteration estimates.
17363 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
17365 2013-04-29 Jakub Jelinek <jakub@redhat.com>
17367 PR tree-optimization/57083
17368 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
17369 non-singleton shift count range, zero extend low_bound for uns case.
17371 * config/i386/predicates.md (general_vector_operand): New predicate.
17372 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
17373 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
17374 if they aren't nonimmediate operands. If their original values
17375 satisfy const_vector_equal_evenodd_p, don't shift them.
17376 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
17377 predicates. For the SSE4.1 case force operands[{1,2}] into registers
17378 if not nonimmediate_operand.
17379 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
17380 instead of register_operand.
17381 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
17383 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
17385 * stor-layout.c (finalize_size_functions): Allocate a structure and
17386 reset cfun before dumping the functions.
17388 2013-04-27 Jakub Jelinek <jakub@redhat.com>
17390 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
17393 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
17394 use xop_pmacsdqh if uns_p.
17395 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
17396 the immediate rotate count.
17398 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
17400 * rtl.h (struct rtx_def): Add comment for field jump.
17401 (LRA_SUBREG_P): New macro.
17402 * recog.c (register_operand): Check LRA_SUBREG_P.
17403 * lra.c (lra): Add note at the end of RTL code. Align non-empty
17405 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
17406 (lra_final_code_change): Skip subreg change for operators.
17407 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
17408 if there are no operand changes.
17409 * lra-constraints.c (curr_insn_set): New.
17410 (match_reload): Set LRA_SUBREG_P.
17411 (emit_spill_move): Ditto.
17412 (check_and_process_move): Use curr_insn_set. Process only single
17413 set insns. Don't initialize sec_mem_p and change_p.
17414 (simplify_operand_subreg): Use LRA_SUBREG_P.
17415 (reg_in_class_p): New function.
17416 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
17417 of #ifdef. Add code to remove cycling.
17418 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
17419 non-null disp. Reload inner instead of disp when base and index
17420 are null. Try to put lo_sum into register.
17421 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
17422 (check_and_process_move): Move code for move cost check to
17423 simple_move_p. Remove equiv_substitution.
17424 (simple_move_p): New function.
17425 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
17426 curr_insn_set. Call check_and_process_move only for single set
17427 insns. Use the new function. Move call of check_and_process_move
17428 after operand equiv substitution and address process.
17430 2013-04-26 Jakub Jelinek <jakub@redhat.com>
17433 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
17434 with nonlocal goto receivers or returns twice calls, ignore
17435 unininitialized values from abnormal edges to nl goto receiver
17436 or returns twice call.
17438 2013-04-26 Jakub Jelinek <jakub@redhat.com>
17440 PR tree-optimization/57051
17441 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
17442 and VEC_RSHIFT_EXPR if shift count is a multiple of element
17445 2013-04-26 Richard Biener <rguenther@suse.de>
17447 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
17448 (expand_omp_taskreg): Likewise. Mark loops for fixup.
17449 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
17450 (fixup_loop_arrays_after_move): New function.
17451 (move_sese_region_to_fn): Properly outline the loop tree parts
17452 of the SESE region.
17454 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
17456 * config/i386/i386.md (type, unit): Fix long lines.
17458 2013-04-26 Richard Biener <rguenther@suse.de>
17460 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
17461 (lto-streamer-out.o): Likewise.
17462 * cfgloop.c (init_loops_structure): Export, add struct function
17463 argument and adjust.
17464 (flow_loops_find): Adjust.
17465 * cfgloop.h (enum loop_estimation): Add EST_LAST.
17466 (init_loops_structure): Declare.
17467 * lto-streamer-in.c: Include cfgloop.h.
17468 (input_cfg): Input the loop tree.
17469 * lto-streamer-out.c: Include cfgloop.h.
17470 (output_cfg): Output the loop tree.
17471 (output_struct_function_base): Do not drop PROP_loops.
17473 2013-03-26 Richard Biener <rguenther@suse.de>
17475 * tree-cfg.c (execute_build_cfg): Build the loop tree.
17476 (pass_build_cfg): Provide PROP_loops.
17477 (move_sese_region_to_fn): Remove loops that are outlined into fn
17479 * tree-inline.c: Include cfgloop.h.
17480 (initialize_cfun): Do not drop PROP_loops.
17481 (copy_loops): New function.
17482 (copy_cfg_body): Copy loop structure.
17483 (tree_function_versioning): Initialize destination loop tree.
17484 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
17485 (pass_parallelize_loops): Do IL verification.
17486 * loop-init.c (loop_optimizer_init): Fixup loops if required.
17487 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
17488 the CFG make sure we fixup loops as well.
17489 * tree-ssa-tail-merge.c: Include cfgloop.h.
17490 (replace_block_by): When merging loop latches mark loops for fixup.
17491 * lto-streamer-out.c (output_struct_function_base): Drop
17492 PROP_loops for now.
17493 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
17494 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
17495 * ipa-split.c: Include cfgloop.h.
17496 (split_function): Add the new return block to the loop tree root.
17497 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
17498 whether we have removed the forwarder block.
17499 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
17500 * cfgloop.h (place_new_loop): Declare.
17501 * cfgloopmanip.c (place_new_loop): Export.
17502 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
17503 (tree-switch-conversion.o): Likewise.
17504 (tree-complex.o): Likewise.
17505 (tree-inline.o): Likewise.
17506 (tree-ssa-tailmerge.o): Likewise.
17507 (ipa-split.o): Likewise.
17508 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
17509 (tree-ssa-copy.o): Likewise.
17510 * tree-switch-conversion.c: Include cfgloop.h
17511 (process_switch): If we emit a bit-test cascade, schedule loops
17513 * tree-complex.c: Include cfgloop.h.
17514 (expand_complex_div_wide): Properly add new basic-blocks to loops.
17515 * asan.c: Include cfgloop.h.
17516 (create_cond_insert_point): Properly add new basic-blocks to
17517 loops, schedule loop fixup.
17518 * cfgloop.c (verify_loop_structure): Check that looks are not
17520 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
17522 (expand_omp_for_generic): Likewise.
17523 (expand_omp_sections): Likewise.
17524 (expand_omp_atomic_pipeline): Schedule loops for fixup.
17525 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
17526 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
17527 is initialized, not when loops are present.
17528 * tree-parloops.c (parallelize_loops): Remove checking here.
17529 * passes.c (init_optimization_passes): Schedule a copy-propagation
17530 pass before complete unrolling of inner loops.
17532 2013-04-26 Jakub Jelinek <jakub@redhat.com>
17534 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
17535 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
17536 (colorize_init): Add argument to _WIN32 version.
17537 * toplev.c: Include diagnostic-color.h.
17538 (process_options): Default to -fdiagnostics-color=auto if
17539 GCC_COLORS env var is in the environment.
17540 * common.opt (fdiagnostics-color=): Add Var and Init.
17541 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
17542 env var is in the environment, the default is auto rather than never.
17544 * diagnostic.h (file_name_as_prefix): Add context argument.
17545 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
17546 the string as locus.
17547 * langhooks.c (lhd_print_error_function): Adjust caller.
17549 2013-04-25 Lawrence Crowl <crowl@google.com>
17551 * var-tracking.c (shared_hash_def::htab):
17552 Change type to hash_table. Update dependent calls and types.
17554 2013-04-25 Lawrence Crowl <crowl@google.com>
17556 * Makefile.in: Update as needed below.
17558 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
17559 Move declaration to after the type's method definitons.
17561 * attribs.c (htab_t scoped_attributes::attribute_hash):
17562 Change type to hash_table. Update dependent calls and types.
17564 * bitmap.c (htab_t bitmap_desc_hash):
17565 Change type to hash_table. Update dependent calls and types.
17567 * cselib.c (htab_t cselib_hash_table):
17568 Change type to hash_table. Update dependent calls and types.
17570 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
17571 (hash_string_slot_node): Move implementation into lto-streamer.h
17572 struct string_slot_hasher.
17573 (eq_string_slot_node): Likewise.
17575 * data-streamer-out.c: Update output_block::string_hash_table
17576 dependent calls and types.
17578 * dwarf2cfi.c (htab_t trace_index):
17579 Change type to hash_table. Update dependent calls and types.
17581 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
17582 Change type to hash_table. Update dependent calls and types.
17583 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
17584 (htab_t optimize_external_refs::map): Likewise.
17585 (htab_t output_comp_unit::extern_map): Likewise.
17586 (htab_t output_comdat_type_unit::extern_map): Likewise.
17587 (htab_t output_macinfo::macinfo_htab): Likewise.
17588 (htab_t optimize_location_lists::htab): Likewise.
17589 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
17591 * except.c (htab_t ehspec_hash_type):
17592 Change type to hash_table. Update dependent calls and types.
17593 (assign_filter_values::ttypes): Likewise.
17594 (assign_filter_values::ehspec): Likewise.
17595 (sjlj_assign_call_site_values::ar_hash): Likewise.
17596 (convert_to_eh_region_ranges::ar_hash): Likewise.
17598 * gcse.c (htab_t pre_ldst_table):
17599 Change type to hash_table. Update dependent calls and types.
17601 * ggc-common.c (htab_t saving_htab):
17602 Change type to hash_table. Update dependent calls and types.
17603 (htab_t loc_hash): Likewise.
17604 (htab_t ptr_hash): Likewise.
17605 (call_count): Rename ggc_call_count.
17606 (call_alloc): Rename ggc_call_alloc.
17607 (loc_descriptor): Rename make_loc_descriptor.
17608 (add_statistics): Rename ggc_add_statistics.
17610 * ggc-common.c (saving_htab):
17611 Change type to hash_table. Update dependent calls and types.
17613 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
17614 (push_gimplify_context): Likewise.
17615 (pop_gimplify_context): Likewise.
17616 (struct gimple_temp_hash_elt): Added.
17617 (struct gimplify_hasher): Likewise.
17618 (struct gimplify_ctx.temp_htab):
17619 Change type to hash_table. Update dependent calls and types.
17621 * gimple-fold.c: Include gimplify-ctx.h.
17623 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
17624 Change type to hash_table. Update dependent calls and types.
17625 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
17626 avoid potential global name collision.
17628 * gimplify.c: Include gimplify-ctx.h.
17629 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
17630 (htab_t gimplify_ctx::temp_htab):
17631 Update dependent calls and types for new type hash_table.
17632 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
17633 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
17635 * gimplify-ctx.h: New.
17636 (struct gimple_temp_hash_elt): Move from gimplify.c.
17637 (class gimplify_hasher): New.
17638 (struct gimplify_ctx): Move from gimple.h.
17639 (htab_t gimplify_ctx::temp_htab):
17640 Change type to hash_table. Update dependent calls and types.
17642 * graphite-clast-to-gimple.c: Include graphite-htab.h.
17643 (htab_t ivs_params::newivs_index):
17644 Change type to hash_table. Update dependent calls and types.
17645 (htab_t ivs_params::params_index): Likewise.
17646 (htab_t print_generated_program::params_index): Likewise.
17647 (htab_t gloog::newivs_index): Likewise.
17648 (htab_t gloog::params_index): Likewise.
17650 * graphite.c: Include graphite-htab.h.
17651 4htab_t graphite_transform_loops::bb_pbb_mapping):
17652 Change type to hash_table. Update dependent calls and types.
17654 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
17655 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
17656 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
17658 * graphite-dependences.c: Include graphite-htab.h.
17659 (loop_is_parallel_p): Change hash table type of parameter.
17661 * graphite-htab.h: New.
17662 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
17663 (extern find_pbb_via_hash): Move from graphite-poly.h.
17664 (extern loop_is_parallel_p): Move from graphite-poly.h.
17665 (extern get_loop_body_pbbs): Move from graphite-poly.h.
17667 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
17668 (extern loop_is_parallel_p): Move to graphite-htab.h.
17669 (extern get_loop_body_pbbs): Move to graphite-htab.h.
17671 * haifa-sched.c (htab_t delay_htab):
17672 Change type to hash_table. Update dependent calls and types.
17673 (htab_t delay_htab_i2): Likewise.
17675 * ira-color.c (htab_t allocno_hard_regs_htab):
17676 Change type to hash_table. Update dependent calls and types.
17678 * ira-costs.c (htab_t cost_classes_htab):
17679 Change type to hash_table. Update dependent calls and types.
17681 * loop-invariant.c (htab_t merge_identical_invariants::eq):
17682 Change type to hash_table. Update dependent calls and types.
17684 * loop-iv.c (htab_t bivs):
17685 Change type to hash_table. Update dependent calls and types.
17687 * loop-unroll.c (htab_t opt_info::insns_to_split):
17688 Change type to hash_table. Update dependent calls and types.
17689 (htab_t opt_info::insns_with_var_to_expand): Likewise.
17691 * lto-streamer.h (struct string_slot): Move from data-streamer.h
17692 (struct string_slot_hasher): New.
17693 (htab_t output_block::string_hash_table):
17694 Change type to hash_table. Update dependent calls and types.
17696 * lto-streamer-in.c (freeing_string_slot_hasher): New.
17697 (htab_t file_name_hash_table):
17698 Change type to hash_table. Update dependent calls and types.
17700 * lto-streamer-out.c: Update output_block::string_hash_table dependent
17703 * lto-streamer.c (htab_t tree_htab):
17704 Change type to hash_table. Update dependent calls and types.
17706 * omp-low.c: Include gimplify-ctx.h.
17708 * passes.c (htab_t name_to_pass_map):
17709 Change type to hash_table. Update dependent calls and types.
17710 (pass_traverse): Rename to passes_pass_traverse.
17712 * plugin.c (htab_t event_tab):
17713 Change type to hash_table. Update dependent calls and types.
17715 * postreload-gcse.c (htab_t expr_table):
17716 Change type to hash_table. Update dependent calls and types.
17717 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
17719 * sese.c (debug_rename_map_1): Make extern.
17720 (htab_t copy_bb_and_scalar_dependences::rename_map):
17721 Change type to hash_table. Update dependent calls and types.
17723 * sese.h (extern debug_rename_map): Move to .c file.
17725 * store-motion.c (htab_t store_motion_mems_table):
17726 Change type to hash_table. Update dependent calls and types.
17728 * trans-mem.c (htab_t tm_new_mem_hash):
17729 Change type to hash_table. Update dependent calls and types.
17731 * tree-browser.c (htab_t TB_up_ht):
17732 Change type to hash_table. Update dependent calls and types.
17734 * tree-cfg.c (htab_t discriminator_per_locus):
17735 Change type to hash_table. Update dependent calls and types.
17737 * tree-complex.c: Include tree-hasher.h
17738 (htab_t complex_variable_components):
17739 Change type to hash_table. Update dependent calls and types.
17741 * tree-eh.c (htab_t finally_tree):
17742 Change type to hash_table. Update dependent calls and types.
17744 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
17745 struct int_tree_hasher.
17746 (extern int_tree_map_eq): Likewise.
17747 (uid_decl_map_hash): Removed.
17748 (extern decl_tree_map_eq): Likewise.
17750 * tree-hasher.h: New.
17751 (struct int_tree_hasher): New.
17752 (typedef int_tree_htab_type): New.
17754 * tree-inline.c: Include gimplify-ctx.h.
17756 * tree-mudflap.c: Include gimplify-ctx.h.
17758 * tree-parloops.c: Include tree-hasher.h.
17759 (htab_t eliminate_local_variables_stmt::decl_address):
17760 Change type to hash_table. Update dependent calls and types.
17761 (htab_t separate_decls_in_region::decl_copies): Likewise.
17763 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
17764 Change type to hash_table. Update dependent calls and types.
17766 * tree-sra.c (candidates):
17767 Change type to hash_table. Update dependent calls and types.
17769 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
17771 (int_tree_map_hash): Likewise.
17773 * tree-ssa-dom.c (htab_t avail_exprs):
17774 Change type to hash_table. Update dependent calls and types.
17776 * tree-ssa-live.c (var_map_base_init::tree_to_index):
17777 Change type to hash_table. Update dependent calls and types.
17779 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
17780 Change type to hash_table. Update dependent calls and types.
17782 * tree-ssa-phiopt.c (seen_ssa_names):
17783 Change type to hash_table. Update dependent calls and types.
17785 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
17786 Change type to hash_table. Update dependent calls and types.
17788 * tree-ssa-uncprop.c (equiv):
17789 Change type to hash_table. Update dependent calls and types.
17791 2013-04-25 Jakub Jelinek <jakub@redhat.com>
17793 PR rtl-optimization/57003
17794 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
17795 call note_stores with kill_clobbered_value callback again after
17796 killing regs_invalidated_by_call.
17798 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17800 * config/aarch64/aarch64-simd.md
17801 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
17802 (aarch64_simd_bsl<mode>): Likewise.
17803 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
17805 2013-04-25 Marek Polacek <polacek@redhat.com>
17807 PR tree-optimization/57066
17808 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
17810 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
17812 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
17814 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17816 * config/aarch64/aarch64-builtins.c
17817 (aarch64_fold_builtin): New.
17818 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
17819 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
17820 * config/aarch64/aarch64-simd-builtins.def (abs): New.
17821 * config/aarch64/arm_neon.h
17822 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
17824 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17825 Tejas Belagod <tejas.belagod@arm.com>
17827 * config/aarch64/aarch64-builtins.c
17828 (aarch64_gimple_fold_builtin): New.
17829 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
17830 * config/aarch64/aarch64-simd-builtins.def (addv): New.
17831 * config/aarch64/aarch64-simd.md (addpv4sf): New.
17832 (addvv4sf): Update.
17833 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
17835 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17837 * config/aarch64/aarch64.md
17838 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
17840 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17842 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
17843 (*ngcsi_uxtw): New pattern.
17845 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17846 Julian Brown <julian@codesourcery.com>
17848 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
17849 (TB_DREG): Add T_V4HF.
17850 (v4hf_UP): New macro.
17851 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17852 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17853 Handle initialisation of V4HF. Adjust initialisation of reinterpret
17855 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17856 (arm_vector_mode_supported_p): Handle V4HF.
17857 (arm_mangle_map): Handle V4HFmode.
17858 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
17859 * config/arm/arm_neon_builtins.def: Add entries for
17860 vcvtv4hfv4sf, vcvtv4sfv4hf.
17861 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
17862 (neon_vcvtv4hfv4sf): Likewise.
17863 * config/arm/neon-gen.ml: Handle half-precision floating point
17865 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
17866 * config/arm/arm_neon.h: Regenerate.
17867 * config/arm/neon.ml (type elts): Add F16.
17868 (type vectype): Add T_float16x4, T_floatHF.
17869 (type vecmode): Add V4HF.
17870 (type features): Add Requires_FP_bit feature.
17871 (elt_width): Handle F16.
17872 (elt_class): Likewise.
17873 (elt_of_class_width): Likewise.
17874 (mode_of_elt): Refactor.
17875 (type_for_elt): Handle F16, fix error messages.
17876 (vectype_size): Handle T_float16x4.
17877 (vcvt_sh): New function.
17878 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
17879 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
17880 (string_of_mode): Handle V4HF.
17881 * doc/arm-neon-intrinsics.texi: Regenerate.
17883 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17885 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
17886 format specifier in 'X' case.
17888 2013-04-25 Alan Modra <amodra@gmail.com>
17891 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
17892 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
17893 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
17894 Repeat for many other rotate/shift and mask patterns using subregs.
17895 Name lshiftrt insns.
17896 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
17897 on WORDS_BIG_ENDIAN.
17899 2013-04-25 Alan Modra <amodra@gmail.com>
17901 * config.gcc: Support little-endian powerpc-linux targets.
17902 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
17903 (LINK_OS_LINUX_SPEC): Define.
17904 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
17905 Preserve MASK_LITTLE_ENDIAN.
17906 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
17907 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
17908 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
17909 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
17910 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
17911 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
17912 Correct fp word order for little-endian. Don't shift toc entries
17913 smaller than a word for little-endian.
17914 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
17915 (bswapdi2 splits): Correct low-part subreg for little-endian.
17916 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
17917 low/high where such is correct only for be.
17918 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
17919 little-endian for -mcall-aixdesc.
17921 2013-04-25 Alan Modra <amodra@gmail.com>
17923 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
17924 replace_equiv_address_nv.
17926 2013-04-25 Alan Modra <amodra@gmail.com>
17928 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
17930 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17933 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17934 * rtl.h (struct rtx_def): ...
17936 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17938 PR rtl-optimizations/57046
17939 * lra-constraints (split_reg): Set up lra_risky_transformations_p
17940 for multi-reg splits.
17942 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
17944 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
17946 2013-04-24 Sterling Augustine <saugustine@google.com>
17948 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
17949 (comp_dir_string, debug_str_dwo_section): New.
17950 (DEBUG_STR_DWO_SECTION): Rename to ...
17951 (DEBUG_DWO_STR_SECTION): ... this.
17952 (DEBUG_NORM_STR_SECTION): Delete.
17953 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
17954 (DEBUG_STR_DWO_SECTION_FLAGS): New.
17955 (find_AT_string): Move most logic to ...
17956 (find_AT_string_in_table): ... here. New.
17957 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
17958 add_skeleton_AT_string. Delete logic.
17959 (output_skeleton_debug_sections): Remove call to
17960 add_top_level_skeleton_die_attrs.
17961 (add_comp_dir_attribute): Move logic to comp_dir_string.
17962 (dwarf2out_init): Initialize debug_str_dwo_section.
17963 (output_indirect_string): Call find_string_form.
17964 (output_indirect_strings): Rewrite.
17965 (prune_unused_types): Empty skeleton_debug_str_hash.
17966 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
17967 (dwarf2out_finish): Call output_indirect_strings.
17969 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17971 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
17973 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17975 * rtl.h (struct rtx_def): Add comment for field jump.
17976 (LRA_SUBREG_P): New macro.
17977 * recog.c (register_operand): Check LRA_SUBREG_P.
17978 * lra.c (lra): Add note at the end of RTL code. Align non-empty
17980 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
17981 (lra_final_code_change): Skip subreg change for operators.
17982 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
17983 if there are no operand changes.
17984 * lra-constraints.c (curr_insn_set): New.
17985 (match_reload): Set LRA_SUBREG_P.
17986 (emit_spill_move): Ditto.
17987 (check_and_process_move): Use curr_insn_set. Process only single
17988 set insns. Don't initialize sec_mem_p and change_p.
17989 (simplify_operand_subreg): Use LRA_SUBREG_P.
17990 (reg_in_class_p): New function.
17991 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
17992 of #ifdef. Add code to remove cycling.
17993 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
17994 non-null disp. Reload inner instead of disp when base and index
17995 are null. Try to put lo_sum into register.
17996 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
17997 (check_and_process_move): Move code for move cost check to
17998 simple_move_p. Remove equiv_substitution.
17999 (simple_move_p): New function.
18000 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18001 curr_insn_set. Call check_and_process_move only for single set
18002 insns. Use the new function. Move call of check_and_process_move
18003 after operand equiv substitution and address process.
18005 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
18007 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
18008 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
18009 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
18011 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
18013 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
18015 2013-04-24 Marek Polacek <polacek@redhat.com>
18017 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
18018 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
18019 (select_loops_exit_conditions): Likewise.
18020 (number_of_iterations_for_all_loops): Likewise.
18021 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
18022 (scev_analysis): Likewise.
18024 2013-04-02 Catherine Moore <clm@codesourcery.com>
18025 Chao-ying Fu <fu@mips.com>
18027 * config/mips/micromips.md (jraddiusp): New pattern.
18028 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
18029 instruction if possible.
18031 2013-04-24 Alan Modra <amodra@gmail.com>
18033 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
18035 2013-04-24 Julian Brown <julian@codesourcery.com>
18036 Chung-Lin Tang <cltang@codesourcery.com>
18038 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
18039 dependency behavior in enumeration type DIE generation. Add TODO note
18040 to comments about future DW_FORM_sdata/udata re-work of related code.
18042 2013-04-23 Lawrence Crowl <crowl@google.com>
18044 * Makefile.in: Update as needed below.
18046 * hash-table.h (class hash_table):
18047 Correct many methods with parameter types compare_type to the correct
18048 value_type. (Correct code was unlikely to notice the change.)
18049 (hash_table::elements_with_deleted) New.
18050 (class hashtable::iterator): New.
18051 (hashtable::begin()): New.
18052 (hashtable::end()): New.
18053 (FOR_EACH_HASH_TABLE_ELEMENT): New.
18055 * statistics.c (statistics_hashes):
18056 Change type to hash_table. Update dependent calls and types.
18058 * tree-into-ssa.c (var_infos):
18059 Change type to hash_table. Update dependent calls and types.
18061 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
18062 Change type to hash_table. Update dependent calls and types.
18064 * tree-ssa-loop-im.c (struct mem_ref.refs):
18065 Change type to hash_table. Update dependent calls and types.
18067 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
18068 Change type to hash_table. Update dependent calls and types.
18070 * tree-ssa-sccvn.c (vn_tables_s::nary):
18071 Change type to hash_table. Update dependent calls and types.
18072 (vn_tables_s::phis): Likewise.
18073 (vn_tables_s::references): Likewise.
18075 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
18076 (vn_reference_eq): Update parameter and return types.
18078 * tree-ssa-structalias.c (pointer_equiv_class_table):
18079 Change type to hash_table. Update dependent calls and types.
18080 (location_equiv_class_table): Likewise.
18082 * tree-vect-data-refs.c: Consequential changes for making
18083 peeling a hash_table.
18085 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
18086 (destroy_loop_vec_info): Dependent hash_table update.
18088 * tree-vectorizer.h (peeling_htab):
18089 Change type to hash_table. Update dependent calls and types.
18091 2013-04-23 Shiva Chen <shiva0217@gmail.com>
18093 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
18094 to check the register content is equal or not.
18095 * lra-constraints.c (match_reload): Use lra_assign_reg_val
18096 to assign register content record.
18097 * lra-eliminations.c (update_reg_eliminate): Use
18098 lra_update_reg_val_offset to update register content offset.
18099 * lra-int.h (struct lra_reg): Add offset member.
18100 (lra_reg_val_equal_p): New static inline function.
18101 (lra_update_reg_val_offset): New static inline function.
18102 (lra_assign_reg_val): New static inline function.
18103 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
18104 to assign register content record.
18105 (initialize_lra_reg_info_element): Initial offset to zero.
18107 2013-04-23 Catherine Moore <clm@codesourcery.com>
18109 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
18110 operands. Record compression.
18112 2013-04-23 Xinliang David Li <davidxl@google.com>
18114 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
18116 2013-04-23 Richard Biener <rguenther@suse.de>
18118 PR middle-end/57036
18119 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
18120 parameter, only add abnormal goto edges from the copied body
18121 if the call could perform abnormal gotos.
18122 (copy_cfg_body): Adjust.
18124 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
18126 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
18128 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
18130 * coretypes.h (gimple_stmt_iterator): Add struct to make
18133 2013-04-23 Richard Biener <rguenther@suse.de>
18135 PR tree-optimization/57026
18136 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
18137 from SSA names occuring in abnormal PHI nodes.
18139 2013-04-22 Andi Kleen <ak@linux.intel.com>
18141 * lto/lto.c (print_lto_report_1): Fix LTO report names.
18143 2013-04-22 Andi Kleen <ak@linux.intel.com>
18145 * lto/lto.c (print_lto_report_1): Declare early.
18146 (read_cgraph_and_symbols): Call print_lto_report_1 early.
18148 2013-04-22 Andi Kleen <ak@linux.intel.com>
18150 * common.opt (-flto-report-wpa): Add.
18151 * doc/invoke.texi (-flto-report-wpa): Add.
18152 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
18155 2013-04-22 Xinliang David Li <davidxl@google.com>
18157 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
18158 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
18159 * Makefile.in: New dependency
18161 David Daney <ddaney.cavm@gmail.com>
18163 * configure.ac (gcc_cv_as_micromips_support): Use the
18164 --fatal-warnings option.
18165 * configure: Regenerate.
18167 2013-04-22 Marek Polacek <polacek@redhat.com>
18170 * tsan.c (instrument_expr): Don't instrument expression
18171 in case its size is zero.
18173 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
18177 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18179 * config/alpha/alpha.c (TARGET_LRA_P): New define.
18181 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
18183 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
18184 (gimple_stmt_iterator): New typedef.
18185 * gimple.h (gimple_stmt_iterator): Rename to...
18186 (gimple_stmt_iterator_d): ... This.
18187 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
18188 trees be valid for GIMPLE and GENERIC.
18189 (TARGET_GIMPLE_FOLD_BUILTIN): New.
18190 * gimple-fold.c (gimple_fold_call): Call target hook
18191 gimple_fold_builtin.
18192 * hooks.c (hook_bool_gsiptr_false): New.
18193 * hooks.h (hook_bool_gsiptr_false): New.
18194 * target.def (fold_stmt): New.
18195 * doc/tm.texi: Regenerate.
18197 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
18200 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
18201 a set sp if no stack realignment.
18203 2013-04-22 Nick Clifton <nickc@redhat.com>
18205 * config.gcc (tilegx-linux): Extend extra_objs rather than
18207 (tilepro-linux): Likewise.
18209 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
18211 * config/aarch64/aarch64-builtins.c
18213 (CF0, CF1, CF2, CF3, CF4, CF10): New.
18214 (VAR<1-12>): Add MAP parameter.
18215 (BUILTIN_*): Likewise.
18216 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
18217 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
18218 (aarch64_ushl_n<mode>): Likewise.
18219 (aarch64_sshr_n<mode>): Likewise.
18220 (aarch64_ushr_n<mode>): Likewise.
18221 (aarch64_<maxmin><mode>): Likewise.
18222 (aarch64_sqrt<mode>): Likewise.
18223 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
18224 (vshr<q>_n_*): Likewise.
18226 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
18228 * config/aarch64/aarch64-builtins.c
18229 (aarch64_simd_builtin_type_mode): Handle SF types.
18231 (BUILTIN_GPF): Define.
18232 (aarch64_init_simd_builtins): Handle SF types.
18233 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
18234 (frecps): Likewise.
18235 (frecpx): Likewise.
18236 * config/aarch64/aarch64-simd.md
18237 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
18238 (aarch64_frecpe<mode>): New.
18239 (aarch64_frecps<mode>): Likewise.
18240 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
18241 (v8type): Add frecp<esx>.
18242 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
18243 (aarch64_frecps<mode>): Likewise.
18244 * config/aarch64/iterators.md (FRECP): New.
18245 (frecp_suffix): Likewise.
18246 * config/aarch64/arm_neon.h
18247 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
18249 2013-04-22 Christian Bruel <christian.bruel@st.com>
18252 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
18253 (REG_CLASS_NAMES): Idem.
18254 (REG_CLASS_CONTENTS): Idem.
18255 (REGCLASS_HAS_FP_REG): Idem.
18256 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
18257 (sh_conditional_register_usage): Idem.
18259 2013-04-21 Jeff Law <law@redhat.com>
18261 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
18262 (ssa_forward_propagate_and_combine): Use it.
18264 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
18266 * lra.c: Update the flow chart diagram.
18268 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
18270 PR rtl-optimization/56847
18271 * lra-constraints.c (process_alt_operands): Discourage alternative
18272 with non-matche doffsettable memory constraint fro memory with
18275 2013-04-19 Richard Biener <rguenther@suse.de>
18277 PR tree-optimization/56982
18278 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
18280 * gimplify.c (gimplify_call_expr): Notice special calls.
18281 (gimplify_modify_expr): Likewise.
18282 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
18283 abnormal control flow receivers.
18284 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
18285 in the same way as cfun->has_nonlocal_labels.
18286 (gimple_purge_dead_abnormal_call_edges): Likewise.
18287 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
18288 receivers start a basic-block.
18290 2013-04-19 Richard Biener <rguenther@suse.de>
18292 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
18294 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
18295 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
18296 (SLP_TREE_LOAD_PERMUTATION): Add.
18297 (vect_transform_slp_perm_load): Adjust prototype.
18298 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
18299 (vect_free_slp_instance): Likewise.
18300 (vect_create_new_slp_node): Likewise.
18301 (vect_supported_slp_permutation_p): Remove.
18302 (vect_slp_rearrange_stmts): Adjust.
18303 (vect_supported_load_permutation_p): Likewise. Inline
18304 vect_supported_slp_permutation_p here.
18305 (vect_analyze_slp_instance): Compute load permutations per
18306 slp node instead of per instance.
18307 (vect_get_slp_defs): Adjust.
18308 (vect_transform_slp_perm_load): Likewise.
18309 (vect_schedule_slp_instance): Remove redundant code.
18310 (vect_schedule_slp): Remove hack for PR56270, add it ...
18311 * tree-vect-stmts.c (vectorizable_load): ... here, do not
18312 CSE loads for SLP. Adjust.
18314 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
18316 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
18317 spelling in two comments.
18319 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
18322 * config/arm/arm.c (load_multiple_sequence): Require SP
18323 as base register for loads if SP is in the register list.
18325 2013-04-19 Martin Jambor <mjambor@suse.cz>
18327 PR tree-optimization/56718
18328 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
18329 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
18330 and made public. Adjusted all callers.
18331 (ipa_intraprocedural_devirtualization): New function.
18332 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
18333 (ipa_intraprocedural_devirtualization): Likewise.
18334 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
18336 2013-04-19 Richard Biener <rguenther@suse.de>
18338 PR tree-optimization/57000
18339 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
18341 2013-04-19 Terry Guo <terry.guo@arm.com>
18343 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
18345 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
18346 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
18347 (cortex_m4_fmacs): Use new reservations.
18348 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
18350 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
18352 PR rtl-optimization/56999
18353 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
18355 (lra_coalesce): Remove split_origin_bitmap and related code.
18356 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
18357 ranges if necessary.
18359 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
18361 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
18363 (ix86_expand_call): Remove clobbered_registers array and use
18364 x86_64_ms_sysv_extra_clobbered_registers instead.
18365 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
18367 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
18369 * config/i386/i386.md (*call_rex64_ms_sysv): Use
18370 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
18371 (*call_value_rex64_ms_sysv): Ditto.
18373 2013-04-18 Cary Coutant <ccoutant@google.com>
18375 * dwarf2out.c (output_pubnames): Check die_perennial_p of
18376 parent instead of die_mark.
18378 2013-04-18 Diego Novillo <dnovillo@google.com>
18380 * gimple.c (create_gimple_tmp): New.
18381 (get_expr_type): New.
18382 (build_assign): New.
18383 (build_type_cast): New.
18384 * gimple.h (enum ssa_mode): Define.
18385 (gimple_seq_set_location): New.
18386 * asan.c (build_check_stmt): Change some gimple_build_* calls
18387 to use build_assign and build_type_cast.
18389 2013-04-18 Richard Biener <rguenther@suse.de>
18391 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
18392 handle negative step. Remove redundant checks.
18393 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
18394 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
18395 for negative step and grouped loads fail to vectorize.
18397 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
18399 * emit-rtl.c (reset_insn_used_flags): New function.
18400 (reset_all_used_flags): Use it.
18401 (verify_insn_sharing): New function.
18402 (verify_rtl_sharing): Fix verification for SEQUENCEs.
18404 2013-04-18 Jakub Jelinek <jakub@redhat.com>
18406 PR tree-optimization/56984
18407 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
18408 and (x >> M) >= N don't register any assertion if N << M is the
18411 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
18413 * lower-subreg.c (resolve_simple_move): If called self-recursive,
18414 do not delete_insn insns that have not yet been emitted, only
18415 unlink them with remove_insn.
18416 * df-scan.c (df_insn_delete): Revert r197492.
18418 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
18420 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
18421 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
18423 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
18425 * config/arm/arm.md (movsicc_insn): Convert define_insn into
18426 define_insn_and_split.
18427 (and_scc,ior_scc,negscc): Likewise.
18428 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
18430 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
18432 * config/arm/arm.c (use_return_insn): Return 0 for targets that
18433 can benefit from using a sequence of LDRD instructions in epilogue
18434 instead of a single LDM instruction.
18436 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
18439 * doc/extend.texi: Fix typo.
18441 2013-04-17 Richard Biener <rguenther@suse.de>
18443 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
18444 (vect_build_slp_tree): ... here.
18445 (vect_build_slp_tree_1): Compute which stmts of the SLP group
18446 match. Remove special-casing of mismatched complex loads.
18447 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
18448 re-try the match with swapped commutative operands.
18449 (vect_supported_load_permutation_p): Remove special-casing of
18450 mismatched complex loads.
18451 (vect_analyze_slp_instance): Adjust.
18453 2013-04-17 Richard Biener <rguenther@suse.de>
18455 PR rtl-optimization/56921
18456 * cfgloop.h (struct loop): Add simple_loop_desc member.
18457 (struct niter_desc): Mark with GTY(()).
18458 (simple_loop_desc): Do not use aux field but simple_loop_desc.
18459 * loop-iv.c (get_simple_loop_desc): Likewise.
18460 (free_simple_loop_desc): Likewise.
18463 2013-04-16 Richard Biener <rguenther@suse.de>
18465 PR rtl-optimization/56921
18466 * loop-init.c (pass_rtl_move_loop_invariants): Add
18467 TODO_do_not_ggc_collect to todo_flags_finish.
18468 (pass_rtl_unswitch): Same.
18469 (pass_rtl_unroll_and_peel_loops): Same.
18470 (pass_rtl_doloop): Same.
18472 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
18474 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
18475 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
18476 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
18478 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
18479 * tree-streamer.c (record_common_node): Adjust reference in comment.
18481 2013-04-17 Terry Guo <terry.guo@arm.com>
18483 * config/arm/cortex-m4.md: Add a new bypass.
18485 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18487 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
18489 (*subs_<optab><mode>_multp2): New pattern.
18490 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
18491 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
18493 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18495 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
18496 (*subs_mul_imm_<mode>): New pattern.
18498 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
18501 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
18502 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
18503 (vsx_movti_32bit): Change j->wa to O->wa.
18505 2013-04-16 Richard Biener <rguenther@suse.de>
18507 PR rtl-optimization/56921
18508 * loop-init.c (pass_rtl_move_loop_invariants): Add
18509 TODO_do_not_ggc_collect to todo_flags_finish.
18510 (pass_rtl_unswitch): Same.
18511 (pass_rtl_unroll_and_peel_loops): Same.
18512 (pass_rtl_doloop): Same.
18514 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
18516 * config/arm/arm.c (emit_multi_reg_push): New declaration
18517 for an existing function.
18518 (arm_emit_strd_push): New function.
18519 (arm_expand_prologue): Used here.
18520 (arm_emit_ldrd_pop): New function.
18521 (arm_expand_epilogue): Used here.
18522 (arm_get_frame_offsets): Update condition.
18523 (arm_emit_multi_reg_pop): Add a special case for load of a single
18524 register with writeback.
18526 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
18528 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
18531 2013-04-16 Richard Biener <rguenther@suse.de>
18533 PR tree-optimization/56756
18534 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
18535 (first_mem_ref_loc): New.
18536 (execute_sm): Place the load temporarily before a previous
18537 access instead of in the latch edge to ensure its SSA dependencies
18538 are defined at points dominating the load.
18540 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
18542 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
18543 correct fix by moving header and footer insn to the footer of
18544 the merged basic block. Clear BB_END of the merged-away block.
18546 PR middle-end/43631
18547 * emit-rtl.c (make_note_raw): New function.
18548 (link_insn_into_chain): New static inline function.
18549 (add_insn): Use it.
18550 (add_insn_before, add_insn_after): Factor insn chain linking code...
18551 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
18552 using link_insn_into_chain.
18553 (note_outside_basic_block_p): New helper function for emit_note_after
18554 and emit_note_before.
18555 (emit_note_after): Use nobb variant of add_insn_after if the note
18556 should not be contained in a basic block.
18557 (emit_note_before): Use nobb variant of add_insn_before if the note
18558 should not be contained in a basic block.
18559 (emit_note_copy): Use make_note_raw.
18560 (emit_note): Likewise.
18561 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
18562 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
18563 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
18564 the moved barrier the tail of the basic block it follows.
18565 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
18567 2013-04-15 Jakub Jelinek <jakub@redhat.com>
18569 PR tree-optimization/56962
18570 * gimple-ssa-strength-reduction.c (record_increment): Only set
18571 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
18572 either rhs1 or rhs2 is equal to c->base_expr.
18574 2013-04-15 Richard Biener <rguenther@suse.de>
18576 PR tree-optimization/56933
18577 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
18579 (GROUP_READ_WRITE_DEPENDENCE): Remove.
18580 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
18581 * tree-vect-data-refs.c (vect_analyze_group_access): Move
18582 dependence check ...
18583 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
18585 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
18586 GROUP_READ_WRITE_DEPENDENCE.
18588 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18590 * emit-rtl.c (reset_all_used_flags): New function.
18591 (verify_rtl_sharing): Call reset_all_used_flags before and after
18592 performing the checks.
18594 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18596 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
18597 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
18598 * config/arm/constraints.md (De): New constraint.
18599 * config/arm/neon.md (anddi3_neon): Delete.
18600 (neon_vand<mode>): Expand to standard anddi3 pattern.
18601 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
18602 Move earlier in the file.
18603 (neon_inv_logic_op2): Likewise.
18604 (arm_anddi_operand_neon): New predicate.
18606 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18608 * configure.ac (gcc_cv_ld_as_needed): Set
18609 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
18610 Use -z ignore, -z record on *-*-solaris2*.
18611 (HAVE_LD_AS_NEEDED): Update comment.
18612 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
18613 * configure: Regenerate.
18614 * config.in: Regenerate.
18615 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
18616 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
18617 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
18618 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
18619 equivalents. Fix markup.
18620 * doc/tm.texi: Regenerate.
18622 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
18624 * config/i386/i386.opt: New option mstack-protector-guard=.
18625 * config/i386/i386-opts.h: Add enum stack_protector_guard.
18626 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
18627 TARGET_SSP_TLS_GUARD.
18628 * config/i386/i386.c (ix86_option_override_internal): Set
18629 ix86_stack_protector_guard.
18630 * config/i386/i386.md (stack_protect_set): Enable for
18631 TARGET_SSP_TLS_GUARD only.
18632 (stack_protect_set_<mode>): Ditto.
18633 (stack_protect_test): Ditto.
18634 (stack_protect_test_<mode>): Ditto.
18635 * doc/invoke.texi (i386 Option): Document.
18637 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
18640 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
18641 (S_MODES): Set H_MODE bit.
18642 (SF_MODES): Set only S_MODE and SF_MODE bits.
18643 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
18644 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
18645 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
18646 <MODE_FLOAT>: Likewise.
18648 2013-04-15 Joey Ye <joey.ye@arm.com>
18650 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
18652 2013-04-15 Joey Ye <joey.ye@arm.com>
18654 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
18656 (thumb_far_jump_used_p): Count instruction size and set
18659 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
18661 * reorg.c (fill_simple_delay_slots): Reindent block of code.
18662 * resource.c (mark_target_live_regs): Reformat conditional block.
18664 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
18666 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
18667 notes, they are emitted only just before final.
18668 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
18670 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
18672 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
18673 * cfgrtl.c (delete_insn): Call it here instead.
18674 * lra-spills.c (lra_final_code_change): Use delete_insn.
18675 * haifa-sched.c (sched_remove_insn): Likewise.
18676 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
18677 returning to the nop pool.
18678 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
18679 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
18681 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
18683 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
18684 * doc/tm.texi: Regenerated.
18686 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
18688 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
18691 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
18693 * df-core.c (df_find_def): Compare register numbers.
18694 (df_find_use): Likewise.
18696 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
18699 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
18700 lra_in_progress for return.
18702 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
18704 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
18705 define_insn into define_insn_and_split and emit movsicc patterns.
18707 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
18709 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
18711 2013-04-12 Richard Biener <rguenther@suse.de>
18713 * tree-pass.h (TODO_do_not_ggc_collect): New.
18714 * passes.c (execute_one_ipa_transform_pass): Honor
18715 TODO_do_not_ggc_collect.
18716 (execute_one_pass): Likewise.
18719 2013-04-10 Richard Biener <rguenther@suse.de>
18721 * passes.c (init_optimization_passes): Remove reload pass.
18722 * ira.c (do_reload): Merge into ...
18724 (rest_of_handle_reload): Remove.
18725 (pass_reload): Likewise.
18726 * config/i386/i386.c (ix86_option_override): Refer to ira instead
18727 of reload for vzeroupper pass placement.
18729 2013-04-12 Jakub Jelinek <jakub@redhat.com>
18731 PR tree-optimization/56918
18732 PR tree-optimization/56920
18733 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
18734 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
18735 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
18736 use wide_mul_with_sign method.
18738 2013-04-12 Richard Biener <rguenther@suse.de>
18740 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
18741 not be considered a gimple constant.
18743 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
18745 * fold-const.c (const_binop): Handle vector shifts by a scalar.
18746 (fold_binary_loc): Call const_binop also for mixed vector-scalar
18749 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
18750 Jakub Jelinek <jakub@redhat.com>
18752 * opts.c: Include diagnostic-color.h.
18753 (common_handle_option): Handle OPT_fdiagnostics_color_.
18754 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
18755 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
18756 (diagnostic-color.o): New.
18757 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
18758 (diagnostic_color_rule): New enum.
18759 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
18760 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
18761 the location string.
18762 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
18763 either NULL, or color kind.
18764 * diagnostic-color.c: New file.
18765 * diagnostic-color.h: New file.
18766 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
18768 * doc/invoke.texi (-fdiagnostics-color): Document.
18769 * pretty-print.h (pp_show_color): Define.
18770 (struct pretty_print_info): Add show_color field.
18771 * diagnostic.c: Include diagnostic-color.h.
18772 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
18773 macros. Colorize error:, warning: etc. strings and also the location
18775 (diagnostic_show_locus): Colorize the caret line.
18776 * pretty-print.c: Include diagnostic-color.h.
18777 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
18778 inside of %< %> quotes or quoted through q format modifier.
18780 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18782 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
18784 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18786 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
18787 code in CC_NZ mode.
18788 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
18791 2013-04-11 Marek Polacek <polacek@redhat.com>
18793 PR tree-optimization/48184
18794 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
18796 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
18798 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
18799 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
18800 (skip_simple_arithmetic): Tidy up.
18801 * tree.h (skip_simple_constant_arithmetic): Declare.
18803 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18805 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
18807 2013-04-11 Richard Biener <rguenther@suse.de>
18809 * tree-vect-loop.c (get_initial_def_for_induction): Properly
18810 generate vector constants.
18812 2013-04-11 Richard Biener <rguenther@suse.de>
18814 PR tree-optimization/56878
18815 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
18816 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
18818 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
18819 Prefer to align the DR with the most invariant base address.
18821 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18823 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
18826 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
18828 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
18829 floating-point vector comparisons against 0.
18831 2013-04-11 Jakub Jelinek <jakub@redhat.com>
18833 PR tree-optimization/56899
18834 * fold-const.c (extract_muldiv_1): Apply distributive law
18835 only if TYPE_OVERFLOW_WRAPS (ctype).
18837 2013-04-11 Bin Cheng <bin.cheng@arm.com>
18840 * ira-costs.c (scan_one_insn): Check whether the source rtx of
18841 loading has side effect.
18843 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
18845 * config/sparc/sparc.c: Include tree-pass.h.
18846 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
18847 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
18848 head of file. Change return type. Split off gate function.
18849 (sparc_gate_work_around_errata): New function.
18850 (pass_work_around_errata): New pass definition.
18851 (insert_pass_work_around_errata) New pass insert definition to
18852 insert pass_work_around_errata just after delayed-branch scheduling.
18853 (sparc_option_override): Insert the pass.
18854 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
18856 2013-04-10 David S. Miller <davem@davemloft.net>
18858 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
18859 or -mcpu=hypersparc.
18861 * target.def (cstore_mode): New hook.
18862 * target.h: Include insn-codes.h
18863 * targhooks.c: Likewise.
18864 (default_cstore_mode): New function.
18865 * targhooks.h: Declare it.
18866 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
18867 * doc/tm.texi: Rebuild.
18868 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
18869 target hook, rather than inspecting the insn_data.
18870 * config/sparc/sparc.c (sparc_cstore_mode): New function.
18871 (TARGET_CSTORE_MODE): Redefine.
18872 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
18874 * config/sparc/predicates.md (cstore_result_operand): New special
18876 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
18877 Use it for operand 0.
18878 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
18879 (*snesi_special): Likewise.
18880 (*snesi_zero): Likewise.
18881 (*seqsi_zero): Likewise.
18882 (*sltu_insn): Likewise.
18883 (*sgeu_insn): Likewise.
18884 (*seqdi_special): Make operand 0 and comparison operation be of
18886 (*snedi_special): Likewise.
18887 (*snedi_special_vis3): Likewise.
18888 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
18889 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
18890 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
18891 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
18892 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
18893 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
18894 (*sltu_extend_sp64): Likewise.
18895 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
18896 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
18897 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
18898 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
18899 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
18901 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
18903 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
18904 (aarch64_start_file): Use the new function.
18906 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18907 Jason Merrill <jason@redhat.com>
18909 * common.opt: Add -gdwarf.
18910 * opts.c (common_handle_option): Handle it.
18911 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
18913 2013-04-10 Richard Biener <rguenther@suse.de>
18915 * passes.c (execute_todo): Do not call ggc_collect conditional here.
18916 (execute_one_ipa_transform_pass): But unconditionally here.
18917 (execute_one_pass): And here.
18918 (init_optimization_passes): Remove reload pass.
18919 * tree-pass.h (TODO_ggc_collect): Remove.
18920 (pass_reload): Likewise.
18921 * ira.c (do_reload): Merge into ...
18923 (rest_of_handle_reload): Remove.
18924 (pass_reload): Likewise.
18925 * config/i386/i386.c (ix86_option_override): Refer to ira instead
18926 of reload for vzeroupper pass placement.
18927 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
18928 and todo_flags_finish of all passes.
18930 2013-04-10 Richard Biener <rguenther@suse.de>
18932 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
18933 first_const_oprnd field, rename first_def_type to first_op_type.
18934 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
18935 (vect_get_and_check_slp_defs): Always use the type of the
18936 operand. Allow mixed vect_external_def, vect_constant_def types.
18937 (vect_get_constant_vectors): Handle mixed vect_external_def,
18938 vect_constant_def types.
18940 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
18942 PR tree-optimization/55524
18943 * tree-ssa-math-opts.c
18944 (convert_mult_to_fma): Don't use an fms construct
18945 when we don't have an fms operation, but fnma, and it looks
18946 likely that we'll be able to use the latter.
18948 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
18950 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
18952 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
18953 inline fail caused by overwritable functions.
18955 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
18957 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
18958 unnecessary bits in the constant power of two case.
18960 2013-04-10 Richard Biener <rguenther@suse.de>
18962 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
18963 broken code swapping operands.
18964 (vect_build_slp_tree): Do not compute load permutations here.
18965 (vect_analyze_slp_instance): Compute load permutations here,
18966 after building the SLP tree.
18968 2013-04-09 Christian Bruel <christian.bruel@st.com>
18970 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
18971 of next/prev_real_insn.
18973 2013-04-09 Jan Hubicka <jh@suse.cz>
18975 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18976 Drop aliased parameter.
18977 (function_and_variable_visibility): Do not handle alias pairs.
18978 * cgraph.c (varpool_externally_visible_p): Update prototype.
18979 * varpool.c (varpool_add_new_variable): Update.
18981 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18983 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
18985 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
18987 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
18989 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
18991 2013-04-09 Marek Polacek <polacek@redhat.com>
18993 PR tree-optimization/48762
18994 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
18996 2013-04-09 Richard Biener <rguenther@suse.de>
18998 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
19000 (vect_build_slp_tree): Likewise.
19001 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
19002 calculating the cost of a SLP instance.
19003 (vect_analyze_slp_instance): Use it from here, after building
19006 2013-04-09 Jakub Jelinek <jakub@redhat.com>
19008 PR middle-end/56883
19009 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
19010 expand_omp_for_static_chunk): Use simple_p = true in
19011 force_gimple_operand_gsi calls when assigning to addressable decls.
19013 2013-04-09 Jeff Law <law@redhat.com>
19015 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
19016 when the boolean was created by converting a wider object which
19017 had a boolean range.
19019 2013-04-09 Richard Biener <rguenther@suse.de>
19021 * tree-vectorizer.h (slp_void_p): Remove.
19022 (slp_tree): Typedef before _slp_tree declaration.
19023 (struct _slp_tree): Use a vector of slp_tree as children.
19024 (vect_get_place_in_interleaving_chain): Remove.
19025 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
19027 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
19029 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
19030 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
19031 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
19032 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
19033 Use slp_node instead of slp_void_p and adjust.
19035 2013-04-09 Richard Biener <rguenther@suse.de>
19037 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
19038 work that is not necessary.
19040 2013-04-09 Jakub Jelinek <jakub@redhat.com>
19042 PR tree-optimization/56854
19043 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
19044 forward into clobber stmts if it would change MEM_REF lhs into
19047 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
19049 * tree.c (type_hash_lookup, type_hash_add): Make static.
19050 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
19052 2013-04-09 Richard Biener <rguenther@suse.de>
19054 * tree.h (unsave_expr_now): Remove.
19055 * tree-inline.c (mark_local_for_remap_r): Remove.
19056 (unsave_expr_1): Likewise.
19057 (unsave_r): Likewise.
19058 (unsave_expr_now): Likewise.
19059 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
19060 (propagate_tree_value): Likewise.
19062 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
19064 * doc/rtl.texi (sequence): Rewrite documentation to match the
19065 current use of SEQUENCE rtl objects.
19066 * rtl.def (SEQUENCE): Likewise.
19068 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
19069 Update documentation.
19070 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
19071 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
19073 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
19075 2013-04-08 Teresa Johnson <tejohnson@google.com>
19077 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
19078 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
19080 (estimate_edge_size_and_time): Add comment to suggest using rounding
19082 (estimate_node_size_and_time): Ditto.
19083 (remap_edge_change_prob): Use helper rounding divide methods.
19084 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
19085 (gimple_mod_pow2_value_transform): Ditto.
19086 (gimple_mod_subtract_transform): Ditto.
19087 (gimple_ic_transform): Ditto.
19088 (gimple_stringops_transform): Ditto.
19089 * stmt.c (conditional_probability): Ditto.
19090 (emit_case_dispatch_table): Ditto.
19091 * lto-cgraph.c (merge_profile_summaries): Ditto.
19092 * tree-optimize.c (execute_fixup_cfg): Ditto.
19093 * cfgcleanup.c (try_forward_edges): Ditto.
19094 * cfgloopmanip.c (scale_loop_profile): Ditto.
19096 (duplicate_loop_to_header_edge): Ditto.
19097 (lv_adjust_loop_entry_edge): Ditto.
19098 * tree-vect-loop.c (vect_transform_loop): Ditto.
19099 * profile.c (compute_branch_probabilities): Ditto.
19100 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
19101 * lto-streamer-in.c (input_cfg): Ditto.
19102 * gimple-streamer-in.c (input_bb): Ditto.
19103 * ipa-cp.c (update_profiling_info): Ditto.
19104 (update_specialized_profile): Ditto.
19105 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
19106 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
19108 * sched-rgn.c (compute_dom_prob_ps): Ditto.
19109 (compute_trg_info): Ditto.
19110 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
19111 (purge_dead_edges): Ditto.
19112 * loop-unswitch.c (unswitch_loop): Ditto.
19113 * cgraphclones.c (cgraph_clone_edge): Ditto.
19114 (cgraph_clone_node): Ditto.
19115 * tree-inline.c (copy_bb): Ditto.
19116 (copy_edges_for_bb): Ditto.
19117 (initialize_cfun): Ditto.
19118 (copy_cfg_body): Ditto.
19119 (expand_call_inline): Ditto.
19121 2013-04-08 Kai Tietz <ktietz@redhat.com>
19123 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
19124 TARGET_CYGWIN64 by TARGET_64BIT.
19126 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
19128 * config/epiphany/epiphany.md (GPR_1): New constant.
19129 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
19130 * config/epiphany/epiphany.c (gen_compare_reg):
19131 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
19132 is already in place.
19133 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
19134 Don't require being called during rtl expansion; If y operlaps r0,
19136 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
19137 (epiphany_expand_epilogue): Likewise.
19139 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
19140 Don't use CC_FPmode for ORDERED / UNORDERED.
19141 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
19143 * config/epiphany/constraints.md (CnL): New constraint.
19144 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
19145 * config/epiphany/predicates.md (add_operand): Allow 1024.
19147 * config/epiphany/epiphany.md (logical_op): New code iterator.
19148 (op_mnc): New code attribute.
19149 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
19150 (mov_f+1, mov_f+2): New peephole2 patterns.
19152 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
19153 (cstoresi4): Also allow re-use of zero result when doing a NE
19154 comparison to a non-zero operand.
19155 Use (clobber (scratch)) for first insn if the gpr output is not needed.
19157 * config/epiphany/epiphany.md (<insn_opname>v2si3):
19158 Use gen_addsi3_i / gen_subsi3_i.
19160 2013-04-08 Jakub Jelinek <jakub@redhat.com>
19164 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
19165 contain anything but clobbers, at most one __builtin_stack_restore,
19166 optionally debug stmts and final resx, and if it has at least one
19167 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
19168 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
19169 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
19170 which isn't defaut definition, remove them.
19171 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
19172 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
19173 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
19174 with MEM_REF LHS with SSA_NAME address.
19176 2013-04-08 Jeff Law <law@redhat.com>
19178 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
19180 2013-04-08 Richard Biener <rguenther@suse.de>
19182 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
19184 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
19185 determined vector type.
19186 (vect_analyze_data_refs): Likewise.
19187 (vect_get_new_vect_var): Adjust.
19188 (vect_create_destination_var): Preserve SSA name versions.
19189 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
19190 not dump anything here.
19192 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
19194 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
19195 Add member lr_slot_known.
19196 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
19198 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
19199 Remove code that sets lr_slot_offset according to what a previous
19200 version of epiphany_emit_save_restore used to do.
19201 (epiphany_emit_save_restore): When doing an lr save or restore,
19202 set/verify lr_slot_known and lr_slot_offset.
19204 2013-04-08 Xinyu Qi <xyqi@marvell.com>
19207 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
19210 2013-04-08 Richard Biener <rguenther@suse.de>
19212 * alias.c (find_base_term): Fix thinko in previous change.
19214 2013-04-08 Jakub Jelinek <jakub@redhat.com>
19216 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
19217 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
19218 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
19219 if possible to compute val.
19220 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
19221 For QImode integers don't require anything about precision. Use
19222 const_with_all_bytes_same to find out if the constant doesn't have
19223 repeated bytes in it.
19225 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19227 * config/s390/s390.c (s390_expand_insv): Only accept insertions
19230 2013-04-08 Marek Polacek <polacek@redhat.com>
19232 PR rtl-optimization/48182
19233 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
19236 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19239 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
19240 nuses, make sure we have a label.
19242 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19245 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
19246 (rs6000_emit_swdiv_low_precision): Remove.
19247 (rs6000_emit_swdiv): Rewrite to handle between one and four
19248 iterations of Newton-Raphson generally; modify required number of
19249 iterations for some cases.
19250 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
19252 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
19254 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
19255 set-but-unused variable.
19257 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
19258 basic blocks of released function bodies garbage-collectable.
19260 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
19261 (struct rtl_opt_pass): Add TODO_df_finish.
19263 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
19265 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19267 * config/arm/constraints.md (q): New constraint.
19268 * config/arm/ldrdstrd.md: New file.
19269 * config/arm/arm.md (ldrdstrd.md) New include.
19270 (arm_movdi): Use "q" instead of "r" constraint
19271 for double-word memory access.
19272 (movdf_soft_insn): Likewise.
19273 * config/arm/vfp.md (movdi_vfp): Likewise.
19274 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
19275 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
19276 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
19277 (mem_ok_for_ldrd_strd): Likewise.
19278 (output_move_double): Update assertion.
19280 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19282 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
19284 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19286 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
19287 define_insn_and_split.
19288 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
19290 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19292 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
19293 define_insn_and_split.
19294 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
19295 (shiftsi3_compare): New pattern.
19296 (rrx): New pattern.
19297 * config/arm/unspecs.md (UNSPEC_RRX): New.
19299 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19301 * config/arm/arm.md (negdi_extendsidi): New pattern.
19302 (negdi_zero_extendsidi): Likewise.
19304 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19306 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
19307 define_insn_and_split.
19308 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
19309 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
19311 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19313 * config/arm/arm.md (arm_subdi3): Convert define_insn into
19314 define_insn_and_split.
19315 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
19316 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
19318 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19320 * config/arm/arm.md (subsi3_carryin): New pattern.
19321 (subsi3_carryin_const): Likewise.
19322 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
19323 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
19325 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19327 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
19329 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
19331 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
19332 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
19334 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19336 * config/arm/arm.c (arm_expand_builtin): Change fcode
19337 type to unsigned int.
19339 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19341 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
19343 2013-04-04 Ian Lance Taylor <iant@google.com>
19345 * doc/standards.texi (Standards): The Go frontend supports the Go 1
19348 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
19350 PR middle-end/56729
19351 * df-scan.c (df_insn_delete): Disable failing assert.
19353 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19355 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
19356 New function prototype.
19357 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
19358 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
19359 (arm_builtin_vectorized_function): New function.
19361 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19363 * config/arm/arm_neon_builtins.def: New file.
19364 * config/arm/arm.c (neon_builtin_data): Move contents to
19365 arm_neon_builtins.def.
19366 (enum arm_builtins): Include neon builtin definitions.
19367 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
19368 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
19370 2013-04-04 Marek Polacek <polacek@redhat.com>
19372 PR tree-optimization/48186
19373 * predict.c (maybe_hot_frequency_p): Return false if
19374 HOT_BB_FREQUENCY_FRACTION is 0.
19375 (cgraph_maybe_hot_edge_p): Likewise.
19377 2013-04-04 Richard Biener <rguenther@suse.de>
19379 PR tree-optimization/56826
19380 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
19383 2013-04-04 Richard Biener <rguenther@suse.de>
19385 PR tree-optimization/56213
19386 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
19387 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
19389 2013-04-04 Richard Biener <rguenther@suse.de>
19391 PR tree-optimization/56837
19392 * tree-loop-distribution.c (classify_partition): For non-zero
19393 values require that the value has the same precision as its
19394 mode to be useful as memset value.
19396 2013-04-03 Nick Clifton <nickc@redhat.com>
19398 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
19399 (fmssf4): Use fmsf.s on E3V5 architectures.
19400 (fnmasf4): Use fnmaf.s on E3V5 architectures.
19401 (fnmssf4): Use fnmsf.s on E3V5 architectures.
19403 2013-04-03 Jeff Law <law@redhat.com>
19405 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
19406 (lra-eliminations.o): Likewise.
19408 2013-04-03 Teresa Johnson <tejohnson@google.com>
19410 * gcov-io.c (compute_working_sets): Moved most of body of old
19411 compute_working_sets here from profile.c.
19412 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
19413 (gcov_working_set_t): Moved typedef here from basic-block.h
19414 (compute_working_set): Declare.
19415 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
19416 (get_working_sets): Renamed from compute_working_set,
19417 replace most of body with call to new compute_working_sets.
19418 (get_exec_counts): Replace call to compute_working_sets
19419 to get_working_sets.
19420 * profile.h (get_working_sets): Renamed from compute_working_set.
19421 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
19422 to get_working_sets.
19423 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
19424 * gcov-dump.c (dump_working_sets): New function.
19426 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
19428 * hwint.c (sext_hwi, zext_hwi): New functions.
19429 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
19430 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
19431 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
19432 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
19433 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
19434 (sext_hwi, zext_hwi): New functions.
19436 2013-04-03 Jeff Law <law@redhat.com>
19438 PR tree-optimization/56799
19439 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
19440 back test for widening conversion erroneously dropped in prior change.
19442 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19445 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
19446 instead of next_real_insn.
19448 2013-04-03 Marek Polacek <polacek@redhat.com>
19451 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
19453 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19456 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
19458 (thumb1_output_casesi): Likewise.
19459 (thumb2_output_casesi): Likewise.
19461 2013-04-03 Richard Biener <rguenther@suse.de>
19463 PR tree-optimization/56817
19464 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
19466 (tree_unroll_loops_completely_1): ... new function to manually
19467 walk the loop tree, properly defering outer loops of unrolled
19468 loops to later iterations.
19470 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19472 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
19473 (vectorizable_load): Likewise.
19474 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
19475 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
19477 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19479 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
19482 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
19484 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
19486 2013-04-03 Bin Cheng <bin.cheng@arm.com>
19488 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
19490 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19492 PR tree-optimization/56790
19493 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
19496 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
19498 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
19500 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
19501 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
19504 2013-04-03 Jakub Jelinek <jakub@redhat.com>
19507 * tree.h (force_folding_builtin_constant_p): New decl.
19508 * builtins.c (force_folding_builtin_constant_p): New variable.
19509 (fold_builtin_constant_p): Fold immediately also if
19510 force_folding_builtin_constant_p.
19512 2013-04-03 Richard Biener <rguenther@suse.de>
19514 PR tree-optimization/56812
19515 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
19516 DRs of the same interleaving chain are independent.
19518 2013-04-02 Jason Merrill <jason@redhat.com>
19520 * gdbinit.in (pbb): Use debug fn.
19522 2013-04-02 Lawrence Crowl <crowl@google.com>
19524 * sese.h (struct ivtype_map_elt_s): Remove unused.
19525 (extern debug_ivtype_map): Remove unused.
19526 (extern eq_ivtype_map_elts): Remove unused.
19527 * sese.c (debug_ivtype_map): Removed unused.
19528 (debug_ivtype_map_1): Removed unused.
19529 (debug_ivtype_elt): Remove unused.
19530 (eq_ivtype_map_elts): Remove unused.
19533 2013-04-02 Kai Tietz <ktietz@redhat.com>
19536 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
19537 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
19538 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
19540 (legitimize_pe_coff_symbol): Likewise.
19541 (is_imported_p): New helper-function.
19542 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
19543 for Windows x64 targets.
19544 (ix86_expand_prologue): Optimize for pe-coff targets.
19545 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
19546 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
19547 medium/large code-model.
19548 (legitimize_pic_address): Likewise.
19549 (legitimize_tls_address): Likewise.
19550 (ix86_expand_call): Likewise.
19551 (x86_output_mi_thunk): Likewise.
19552 (get_dllimport_decl): Add new beimport argument.
19553 (construct_plt_address): Don't assert for x64 pe-coff targets.
19554 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
19556 (SYMBOL_FLAG_STUBVAR): New macro.
19557 (SYMBOL_REF_STUBVAR_P): Likewise.
19558 * config/i386/winnt.c (stub_list): New structure.
19559 (stub_head): New local variable.
19560 (i386_pe_record_stub): New function.
19561 (i386_pe_file_end): Emit refptr-stubs.
19563 2013-04-02 Jakub Jelinek <jakub@redhat.com>
19565 PR rtl-optimization/56745
19566 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
19567 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
19570 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
19571 and both of them are MEM_REFs, just compare first argument for
19572 equality and attempt to deal even with differing offsets.
19575 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
19576 of gimple_clobber_p to be MEM_REF.
19577 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
19578 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
19579 after gimplification.
19580 * asan.c (get_mem_ref_of_assignment): Don't instrument
19581 gimple_clobber_p stmts.
19582 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
19583 gimple_clobber_p stmt if they have MEM_REF lhs and
19584 are dead because of another gimple_clobber_p stmt.
19585 * tree-ssa-live.c (clear_unused_block_pointer): Treat
19586 gimple_clobber_p stmts like debug stmts.
19587 (remove_unused_locals): Remove clobbers with MEM_REF lhs
19588 that refer to unused VAR_DECLs or uninitialized values.
19589 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
19590 gimple_clobber_p stmts if they refer to removed parameters.
19591 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
19594 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
19596 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
19597 using SWI48 mode attribute.
19599 2013-04-02 Wei Mi <wmi@google.com>
19601 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
19602 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
19603 *<rotate_insn><mode>3_mask in i386.md.
19605 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
19607 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
19609 2013-04-02 Richard Biener <rguenther@suse.de>
19611 PR tree-optimization/56778
19612 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
19613 Runtime alias tests are not supported for gather loads.
19614 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
19615 stmts referenced from SSA operands before updating SSA form.
19617 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
19618 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19620 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
19621 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
19622 * config/arm/cortex-a53.md: New file.
19623 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
19624 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
19625 * config/arm/arm.c (arm_issue_rate): Likewise.
19626 * config/arm/arm-tune.md: Regenerate
19627 * config/arm/arm-tables.opt: Regenerate.
19628 * config/arm/arm-cores.def: Add cortex-a53.
19630 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
19632 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
19635 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
19637 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
19638 scalar load/store operations using B/H registers.
19639 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
19641 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
19643 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
19645 * config/aarch64/aarch64.c
19646 (aarch64_simd_scalar_immediate_valid_for_move): New.
19647 * config/aarch64/aarch64-protos.h
19648 (aarch64_simd_scalar_immediate_valid_for_move): New.
19649 * config/aarch64/constraints.md (Dh, Dq): New.
19650 * config/aarch64/iterators.md (hq): New.
19652 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
19654 * reorg.c (get_branch_condition): Deal with conditional returns.
19655 (fill_simple_delay_slots): Remove dead code dealing with jumps.
19657 2013-04-01 Wei Mi <wmi@google.com>
19659 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
19660 Truncate operand 2 using %b asm operand modifier.
19661 (*<shift_insn><mode>3_mask): Ditto.
19662 (*<rotate_insn><mode>3_mask): Ditto.
19664 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
19666 PR middle-end/56798
19667 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
19669 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
19671 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
19673 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
19675 2013-03-30 Lawrence Crowl <crowl@google.com>
19677 * dse.c (clear_alias_sets): Remove never set.
19678 (disqualified_clear_alias_sets): Remove never set.
19679 (clear_alias_mode_pool): Remove never set.
19680 (dse_step0): Remove condition that is never true.
19681 (canon_address): Remove condition that is never true.
19682 (dse_step7): Remove condition that is never true.
19683 (rest_of_handle_dse): Remove condition that is never true.
19684 (rest_of_handle_dse::did_global): Remove never read from above.
19685 (dse_step2_spill): Remove never called from above.
19686 (dse_step5_spill): Remove never called from above.
19688 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
19690 * doc/md.texi (Standard Names) <casesi>: Update documentation for
19691 JUMP_TABLE_DATA changes.
19692 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
19693 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
19694 (Insns) <jump_table_data>: New entry.
19695 * doc/tm.texi: Regenerate.
19697 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
19699 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
19700 for table jump at the end of a basic block using tablejump_p.
19701 * targhooks.c (default_invalid_within_doloop): Likewise.
19702 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
19703 target hook implementation that is identical to the default hook.
19704 (rs6000_invalid_within_doloop): Remove.
19706 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
19707 unused variable from tablejump_p call.
19709 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
19710 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
19711 (INSN_DELETED_P): Likewise.
19712 (emit_jump_table_data): New prototype.
19713 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
19714 after 4th as unused.
19715 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
19716 * sched-vis.c (print_insn): Likewise.
19717 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
19718 insn for compatibility with back ends that use next_active_insn to
19719 identify jump table data.
19720 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
19721 (remove_insn): Likewise.
19722 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
19724 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
19725 (emit_jump_table_data): New function.
19727 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
19728 basic block, a JUMP_TABLE_DATA never is.
19729 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
19730 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
19731 off from code handling real insns.
19732 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
19733 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
19734 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
19735 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
19736 is not a NONDEBUG_INSN_P.
19737 * ira-costs.c (scan_one_insn): Likewise.
19738 * jump.c (mark_all_labels): Likewise.
19739 (mark_jump_label_1): Likewise.
19740 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
19741 * lra.c (get_insn_freq): Expect all insns reaching here to be in
19743 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
19744 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
19745 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
19746 JUMP_TABLE_DATA_P insns.
19747 (calculate_elim_costs_all_insns): Likewise.
19748 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
19749 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
19750 (delete_output_reload): Code style fixups.
19751 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
19752 insn flags on this non-insn.
19753 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
19754 as scheduling barriers, for pre-change compatibility.
19755 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
19756 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
19758 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
19759 redundant JUMP_TABLE_DATA_P test.
19760 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
19761 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
19762 (frv_for_each_packet): Likewise.
19763 * config/i386/i386.c (min_insn_size): Likewise.
19764 (ix86_avoid_jump_mispredicts): Likewise.
19765 * config/m32r/m32r.c (m32r_is_insn): Likewise.
19766 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
19767 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
19768 (mips16_insn_length): Robustify.
19769 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
19770 (mips16_split_long_branches): Likewise.
19771 * config/pa/pa.c (pa_combine_instructions): Likewise.
19772 * config/rs6000/rs6000.c (get_next_active_insn): Treat
19773 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
19774 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
19775 as contributing to pool range lengths.
19776 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
19777 Remove redundant JUMP_TABLE_DATA_P test.
19778 (sh_loop_align): Likewise.
19779 (split_branches): Likewise.
19780 (sh_insn_length_adjustment): Likewise.
19781 * config/spu/spu.c (get_branch_target): Likewise.
19783 2013-03-29 Jan Hubicka <jh@suse.cz>
19785 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
19786 gcov streaming; stream hot bb threshold to ltrans.
19787 * predict.c (get_hot_bb_threshold): Break out from ....
19788 (maybe_hot_count_p): ... here.
19789 (set_hot_bb_threshold): New function.
19790 * lto-section-in.c (lto_section_name): Add profile.
19791 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
19792 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
19793 and data-streamer.h
19794 (histogram_entry): New structure.
19795 (histogram, histogram_pool): New global vars.
19796 (histogram_hash): New structure.
19797 (histogram_hash::hash): New method.
19798 (histogram_hash::equal): Likewise.
19799 (account_time_size): New function.
19800 (cmp_counts): New function.
19801 (dump_histogram): New function.
19802 (ipa_profile_generate_summary): New function.
19803 (ipa_profile_write_summary): New function.
19804 (ipa_profile_read_summary): New function.
19805 (ipa_profile): Decide on threshold.
19806 (pass_ipa_profile): Add ipa_profile_write_summary and
19807 ipa_profile_read_summary.
19808 * Makefile.in (ipa.o): Update dependencies.
19809 * lto-streamer.h (LTO_section_ipa_profile): New section.
19811 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
19813 * tree.h (VAR_P): New.
19815 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
19818 * doc/invoke.texi ([-fwhole-program]): Fix typo.
19820 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
19822 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
19823 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19824 (control_flow_insn_p): Likewise.
19825 * cfgrtl.c (duplicate_insn_chain): Likewise.
19826 * final.c (get_attr_length_1): Likewise.
19827 (shorten_branches): Likewise.
19828 (final_scan_insn): Likewise.
19829 * function.c (instantiate_virtual_regs): Likewise.
19830 * gcse.c (insert_insn_end_basic_block): Likewise.
19831 * ira-costs.c (scan_one_insn): Likewise.
19832 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
19833 * lra.c (check_rtl): Likewise.
19834 * reload1.c (elimination_costs_in_insn): Likewise.
19835 * reorg.c (follow_jumps): Likewise.
19837 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
19838 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19839 (thumb_far_jump_used_p): Likewise.
19840 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
19841 (workaround_speculation): Likewise.
19842 (add_sched_insns_for_speculation): Likewise.
19843 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
19844 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
19845 (frv_for_each_packet): Likewise.
19846 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
19847 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
19848 (final_emit_insn_group_barriers): Likewise.
19849 * config/m32r/m32r.c (m32r_is_insn): Likewise.
19850 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
19851 (mips16_insn_length): Likewise.
19852 * config/pa/pa.c (pa_reorg): Likewise.
19853 (pa_combine_instructions): Likewise.
19854 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
19855 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
19856 (sh_reorg): Likewise.
19857 (split_branches): Likewise.
19858 * config/spu/spu.c (get_branch_target): Likewise.
19860 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
19863 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
19865 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
19866 Fix declaration name.
19868 2013-03-28 Lawrence Crowl <crowl@google.com>
19870 * graphds.h (struct graph.indicies): Remove unused.
19871 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
19872 (SCOP_ORIGINAL_PDDRS): Remove unused.
19873 * sese.h (extern insert_loop_close_phis): Removed unused.
19874 (extern insert_guard_phis): Removed unused.
19875 (extern ivtype_map_elt_info): Removed unused.
19876 (new_ivtype_map_elt): Removed unused.
19877 * sese.c (ivtype_map_elt_info): Removed unused.
19879 2013-03-28 Lawrence Crowl <crowl@google.com>
19881 * Makefile.in: Add several missing include dependences.
19883 (test-dump.o): New. This object is not added to any executable,
19884 but is present for ad-hoc testing.
19886 (debug (const bitmap_head_def &)): New.
19887 (debug (const bitmap_head_def *)): New.
19889 (extern debug (const bitmap_head_def &)): New.
19890 (extern debug (const bitmap_head_def *)): New.
19892 (debug (edge_def &)): New.
19893 (debug (edge_def *)): New.
19895 (debug (basic_block_def &)): New.
19896 (debug (basic_block_def *)): New.
19898 (dump_node (const_tree, int, FILE *)): Correct source file.
19900 (debug (die_struct &)): New.
19901 (debug (die_struct *)): New.
19903 (extern debug (die_struct &)): New.
19904 (extern debug (die_struct *)): New.
19905 * gimple-pretty-print.c
19906 (debug (gimple_statement_d &)): New.
19907 (debug (gimple_statement_d *)): New.
19908 * gimple-pretty-print.h
19909 (extern debug (gimple_statement_d &)): New.
19910 (extern debug (gimple_statement_d *)): New.
19912 (debug (ira_allocno_copy &)): New.
19913 (debug (ira_allocno_copy *)): New.
19914 (debug (ira_allocno &)): New.
19915 (debug (ira_allocno *)): New.
19917 (extern debug (ira_allocno_copy &)): New.
19918 (extern debug (ira_allocno_copy *)): New.
19919 (extern debug (ira_allocno &)): New.
19920 (extern debug (ira_allocno *)): New.
19922 (debug (live_range &)): New.
19923 (debug (live_range *)): New.
19925 (debug (lra_live_range &)): New.
19926 (debug (lra_live_range *)): New.
19928 (debug (lra_live_range &)): New.
19929 (debug (lra_live_range *)): New.
19931 (debug (omega_pb_d &)): New.
19932 (debug (omega_pb_d *)): New.
19934 (extern debug (omega_pb_d &)): New.
19935 (extern debug (omega_pb_d *)): New.
19937 (debug (const rtx_def &)): New.
19938 (debug (const rtx_def *)): New.
19940 (debug_tree (tree): Move within file.
19941 (debug_raw (const tree_node &)): New.
19942 (debug_raw (const tree_node *)): New.
19943 (dump_tree_via_hooks (const tree_node *, int)): New.
19944 (debug (const tree_node &)): New.
19945 (debug (const tree_node *)): New.
19946 (debug_verbose (const tree_node &)): New.
19947 (debug_verbose (const tree_node *)): New.
19948 (debug_head (const tree_node &)): New.
19949 (debug_head (const tree_node *)): New.
19950 (debug_body (const tree_node &)): New.
19951 (debug_body (const tree_node *)): New.
19952 (debug_vec_tree (tree): Move and reimplement in terms of dump.
19953 (debug (vec<tree, va_gc> &)): New.
19954 (debug (vec<tree, va_gc> *)): New.
19956 (extern debug (const rtx_def &)): New.
19957 (extern debug (const rtx_def *)): New.
19959 (debug_raw (simple_bitmap_def &)): New.
19960 (debug_raw (simple_bitmap_def *)): New.
19961 (debug (simple_bitmap_def &)): New.
19962 (debug (simple_bitmap_def *)): New.
19964 (extern debug (simple_bitmap_def &)): New.
19965 (extern debug (simple_bitmap_def *)): New.
19966 (extern debug_raw (simple_bitmap_def &)): New.
19967 (extern debug_raw (simple_bitmap_def *)): New.
19969 (debug (vinsn_def &)): New.
19970 (debug (vinsn_def *)): New.
19971 (debug_verbose (vinsn_def &)): New.
19972 (debug_verbose (vinsn_def *)): New.
19973 (debug (expr_def &)): New.
19974 (debug (expr_def *)): New.
19975 (debug_verbose (expr_def &)): New.
19976 (debug_verbose (expr_def *)): New.
19977 (debug (vec<rtx> &)): New.
19978 (debug (vec<rtx> *)): New.
19980 (extern debug (vinsn_def &)): New.
19981 (extern debug (vinsn_def *)): New.
19982 (extern debug_verbose (vinsn_def &)): New.
19983 (extern debug_verbose (vinsn_def *)): New.
19984 (extern debug (expr_def &)): New.
19985 (extern debug (expr_def *)): New.
19986 (extern debug_verbose (expr_def &)): New.
19987 (extern debug_verbose (expr_def *)): New.
19988 (extern debug (vec<rtx> &)): New.
19989 (extern debug (vec<rtx> *)): New.
19991 (_list_iter_cond_expr): Make inline instead of static.
19993 (debug (sreal &)): New.
19994 (debug (sreal *)): New.
19996 (extern debug (sreal &)): New.
19997 (extern debug (sreal *)): New.
19999 (extern debug_raw (const tree_node &)): New.
20000 (extern debug_raw (const tree_node *)): New.
20001 (extern debug (const tree_node &)): New.
20002 (extern debug (const tree_node *)): New.
20003 (extern debug_verbose (const tree_node &)): New.
20004 (extern debug_verbose (const tree_node *)): New.
20005 (extern debug_head (const tree_node &)): New.
20006 (extern debug_head (const tree_node *)): New.
20007 (extern debug_body (const tree_node &)): New.
20008 (extern debug_body (const tree_node *)): New.
20009 (extern debug (vec<tree, va_gc> &)): New.
20010 (extern debug (vec<tree, va_gc> *)): New.
20012 (debug (struct loop &)): New.
20013 (debug (struct loop *)): New.
20014 (debug_verbose (struct loop &)): New.
20015 (debug_verbose (struct loop *)): New.
20016 * tree-dump.c: Add header dependence.
20018 (extern debug (struct loop &)): New.
20019 (extern debug (struct loop *)): New.
20020 (extern debug_verbose (struct loop &)): New.
20021 (extern debug_verbose (struct loop *)): New.
20023 (debug (data_reference &)): New.
20024 (debug (data_reference *)): New.
20025 (debug (vec<data_reference_p> &)): New.
20026 (debug (vec<data_reference_p> *)): New.
20027 (debug (vec<ddr_p> &)): New.
20028 (debug (vec<ddr_p> *)): New.
20030 (extern debug (data_reference &)): New.
20031 (extern debug (data_reference *)): New.
20032 (extern debug (vec<data_reference_p> &)): New.
20033 (extern debug (vec<data_reference_p> *)): New.
20034 (extern debug (vec<ddr_p> &)): New.
20035 (extern debug (vec<ddr_p> *)): New.
20037 (debug (pt_solution &)): New.
20038 (debug (pt_solution *)): New.
20040 (extern debug (pt_solution &)): New.
20041 (extern debug (pt_solution *)): New.
20043 (debug (_var_map &)): New.
20044 (debug (_var_map *)): New.
20045 (debug (tree_live_info_d &)): New.
20046 (debug (tree_live_info_d *)): New.
20048 (extern debug (_var_map &)): New.
20049 (extern debug (_var_map *)): New.
20050 (extern debug (tree_live_info_d &)): New.
20051 (extern debug (tree_live_info_d *)): New.
20053 2013-03-28 Jan Hubicka <jh@suse.cz>
20055 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
20057 2013-03-28 Ian Bolton <ian.bolton@arm.com>
20059 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
20060 record only when desired or required.
20062 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
20064 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
20065 *vec_extractv2di_1_rex64. Use x64 isa attribute.
20067 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20069 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
20070 (*andsi3_compare0_uxtw): New pattern.
20071 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
20072 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
20074 2013-03-28 Jan Hubicka <jh@suse.cz>
20076 * data-streamer-in.c (streamer_read_gcov_count): New function.
20077 * gimple-streamer-out.c: Include value-prof.h.
20078 (output_gimple_stmt): Output histogram.
20079 (output_bb): Use streamer_write_gcov_count.
20080 * value-prof.c: Include data-streamer.h
20081 (dump_histogram_value): Add HIST_TYPE_MAX.
20082 (stream_out_histogram_value): New function.
20083 (stream_in_histogram_value): New function.
20084 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
20085 (stream_out_histogram_value, stream_in_histogram_value): Declare.
20086 * data-streamer-out.c (streamer_write_gcov_count): New function.
20087 (streamer_write_gcov_count_stream): New function.
20088 * lto-cgraph.c (lto_output_edge): Update counter streaming.
20089 (lto_output_node): Likewise.
20090 (input_node, input_edge): Likewise.
20091 * lto-streamer-out.c (output_cfg): Update streaming.
20092 * lto-streamer-in.c (input_cfg): Likewise.
20093 * data-streamer.h (streamer_write_gcov_count,
20094 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
20095 * gimple-streamer-in.c: Include value-prof.h
20096 (input_gimple_stmt): Input histograms.
20097 (input_bb): Update profile streaming.
20099 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
20101 * genmodes.c (emit_max_int): New function.
20102 (emit_insn_modes_h): Added call to emit_max_function.
20103 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
20105 * machmode.def: Fixed comment.
20107 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
20109 * combine.c (try_combine): Removed useless assert.
20110 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
20112 2013-03-28 Marek Polacek <polacek@redhat.com>
20113 Richard Biener <rguenther@suse.de>
20115 PR tree-optimization/56695
20116 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
20117 build signed result of a vector comparison.
20118 * tree-cfg.c (verify_gimple_comparison): Check that a result
20119 of a vector comparison has signed type.
20121 2013-03-28 Richard Biener <rguenther@suse.de>
20123 PR tree-optimization/37021
20124 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
20125 do not restrict gaps between groups.
20126 * tree-vect-stmts.c (vectorizable_load): Properly account for
20127 a gap between groups.
20129 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
20131 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
20132 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
20135 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
20137 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
20138 * doc/extend.texi (Named Address Spaces): Ditto.
20139 (Variable Attributes): Ditto.
20141 2013-03-27 Kai Tietz <ktietz@redhat.com>
20143 * config.build: Add support for cygwin x64 target.
20144 * config.gcc: Likewise.
20145 * config.host: Likewise.
20146 * configure.ac: Likewise
20147 * configure: Regenerated.
20149 2013-03-27 Kai Tietz <ktietz@redhat.com>
20151 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
20152 * config/i386/t-cygwin-w64: New file.
20153 * config/i386/cygwin-w64.h: New file.
20154 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
20155 and add support for x64-cygwin target.
20156 (CPP_SPEC): Likewise.
20157 (CXX_WRAP_SPEC_LIST): Undefine before define.
20158 (LIBGCJ_SONAME): Use 15 as version.
20160 2013-03-27 Richard Biener <rguenther@suse.de>
20162 PR tree-optimization/56716
20163 * tree-ssa-structalias.c (perform_var_substitution): Adjust
20164 dumping for ref nodes.
20166 2013-03-27 Martin Jambor <mjambor@suse.cz>
20168 PR tree-optimization/55334
20169 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
20170 restricted pointers to arrays.
20172 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
20174 * Makefile.in (.SUFFIXES): Add .cc.
20175 (.c.o): Apply same recipe for implicit rule .cc.o.
20177 2013-03-27 Richard Biener <rguenther@suse.de>
20179 PR tree-optimization/37021
20180 * tree-vect-data-refs.c (vect_check_strided_load): Allow
20181 REALPART/IMAGPART_EXPRs around the supported refs.
20182 * tree-ssa-structalias.c (find_func_aliases): Assume that
20183 floating-point values are not used to transfer pointers.
20185 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
20187 * target.def (TARGET_HAS_IFUNC_P): New target hook.
20188 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
20189 * doc/tm.texi: Regenerate.
20190 * targhooks.h (default_has_ifunc_p): New.
20191 * targhooks.c (default_has_ifunc_p): Ditto.
20192 * config/linux-protos.h: New file.
20193 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
20194 hook for linux which disables support of indirect functions in android.
20195 * config/linux-android.c: New file.
20196 * config/t-linux-android.c: Ditto.
20197 * config.gcc: Added new object file linux-android.o.
20198 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
20199 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
20200 * varasm.c (do_assemble_alias): Likewise.
20201 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
20202 doesn't support indirect functions.
20203 * configure: Regenerate.
20205 2013-03-27 Bin Cheng <bin.cheng@arm.com>
20208 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
20209 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
20212 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20214 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
20216 2013-03-27 Terry Guo <terry.guo@arm.com>
20218 * config/arm/arm-cores.def: Added core cortex-r7.
20219 * config/arm/arm-tune.md: Regenerated.
20220 * config/arm/arm-tables.opt: Regenerated.
20221 * doc/invoke.texi: Added entry for core cortex-r7.
20223 2013-03-27 Walter Lee <walt@tilera.com>
20225 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
20226 double-decrement of next_scratch_regno.
20228 2013-03-27 Walter Lee <walt@tilera.com>
20230 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
20232 (insn_v1mulus): Ditto.
20233 (insn_v2muls): Ditto.
20235 2013-03-27 Walter Lee <walt@tilera.com>
20237 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
20238 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
20240 2013-03-27 Walter Lee <walt@tilera.com>
20242 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
20243 (*sibcall_value): Ditto.
20245 2013-03-27 Walter Lee <walt@tilera.com>
20247 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
20248 (insn_mnz_v8qi): ... this ...
20249 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
20251 (insn_v<n>mnz): Replaced by ...
20252 (insn_v1mnz): ... this ...
20253 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
20255 (insn_mz_<mode>): Replaced by ...
20256 (insn_mz_v8qi): ... this ...
20257 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
20259 (insn_v<n>mz): Replaced by ...
20260 (insn_v1mz): ... this ...
20261 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
20264 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
20266 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
20268 2013-03-26 Roland McGrath <mcgrathr@google.com>
20270 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
20271 than fprintf with a non-constant, non-format string.
20273 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
20275 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
20276 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
20277 operand 0 predicate.
20278 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
20279 attribute. Use general_x64nomem_operand as operand 1 predicate.
20280 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
20281 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
20282 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
20283 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
20284 (mov<mode>_insv_1): Remove expander. Merge insn with
20285 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
20286 Use general_x64nomem_operand as operand 1 predicate.
20287 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
20288 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
20289 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
20290 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
20291 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
20292 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
20293 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
20294 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
20295 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
20296 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
20297 (general_x64nomem_operand): Ditto.
20299 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
20301 * config/rtems.opt: Add -pthread option.
20303 2013-03-26 Richard Biener <rguenther@suse.de>
20305 * alias.c (find_base_term): Avoid redundant and not used recursion.
20306 (base_alias_check): Get the initial base term from the caller.
20307 (true_dependence_1): Compute and pass base terms to base_alias_check.
20308 (write_dependence_p): Likewise.
20309 (may_alias_p): Likewise.
20311 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
20313 * config/aarch64/aarch64.c (aarch64_classify_address): Support
20314 PC-relative load in SI modes and above only.
20316 2013-03-26 Xinyu Qi <xyqi@marvell.com>
20318 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
20319 * config/arm/iwmmxt.md (WCGR0): Update.
20320 (WCGR1, WCGR2, WCGR3): Likewise.
20322 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
20324 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
20325 Use x64 and nox64 isa attributes.
20327 2013-03-26 Richard Biener <rguenther@suse.de>
20329 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
20330 alignment computations and rely on get_object_alignment_1
20331 for the !TYPE_P case.
20332 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
20334 2013-03-26 Walter Lee <walt@tilera.com>
20336 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
20337 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
20339 2013-03-25 Jeff Law <law@redhat.com>
20341 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
20342 check for INTEGRAL_TYPE_P that was missing due to checking in
20343 wrong version of prior patch.
20345 2013-03-25 Walter Lee <walt@tilera.com>
20347 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
20348 TILEGX_INSN_SHUFFLEBYTES1.
20349 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
20351 (tilegx_builtins): Ditto.
20352 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
20354 2013-03-25 Walter Lee <walt@tilera.com>
20356 * config/tilegx/tilegx.md (floatsisf2): New pattern.
20357 (floatunssisf2): New pattern.
20358 (floatsidf2): New pattern.
20359 (floatunssidf2): New pattern.
20361 2013-03-25 Walter Lee <walt@tilera.com>
20363 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
20364 tests for constraint J, K, N, P.
20366 2013-03-25 Walter Lee <walt@tilera.com>
20368 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
20369 Use indirect/pcrel encoding.
20370 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
20373 2013-03-25 Steve Ellcey <sellcey@mips.com>
20375 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
20376 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
20377 * config/mips/mips.c (mips_option_override): Set IMADD default.
20378 * config/mips/mips.h (PTF_AVOID_IMADD): New.
20379 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
20380 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
20381 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
20382 * doc/invoke.texi (-mimadd/-mno-imadd): New.
20384 2013-03-25 Jeff Law <law@redhat.com>
20386 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
20387 slightly to avoid creating and folding useless trees. Simplify
20388 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
20390 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
20392 * config/i386/i386.md (*zero_extendsidi2): Merge with
20393 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
20394 * config/i386/predicates.md (x86_64_zext_operand): Rename from
20395 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
20396 targets. Clarify comment.
20398 2013-03-25 Martin Jambor <mjambor@suse.cz>
20400 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
20401 pass-through jump functions differently.
20402 (ipa_read_jump_function): Likewise. Also use setter functions to set
20405 2013-03-25 Martin Jambor <mjambor@suse.cz>
20407 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
20408 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
20410 (ipa_get_indirect_edge_target): New function.
20411 (devirtualization_time_bonus): New parameter known_aggs, pass it to
20412 ipa_get_indirect_edge_target. Update all callers.
20413 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
20414 ipa_get_indirect_edge_target_1 instead of calling
20415 ipa_get_indirect_edge_target.
20416 (create_specialized_node): Pass aggvlas to
20417 ipcp_discover_new_direct_edges.
20419 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20421 * config/arm/arm.md (f_sels, f_seld): New types.
20422 (*cmov<mode>): New pattern.
20423 * config/arm/predicates.md (arm_vsel_comparison_operator): New
20426 2013-03-25 Kai Tietz <ktietz@redhat.com>
20428 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
20429 POSIX-printf for mingw-hosted builds.
20431 2013-03-25 Richard Biener <rguenther@suse.de>
20433 PR middle-end/56694
20434 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
20435 must-not-throw stmt location.
20437 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20439 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
20440 Emit load-acquire versions when acq is true.
20441 (arm_emit_store_exclusive): Add rel parameter.
20442 Emit store-release versions when rel is true.
20443 (arm_split_compare_and_swap): Use acquire-release instructions
20445 of barriers when appropriate.
20446 (arm_split_atomic_op): Likewise.
20447 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
20448 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
20449 (VUNSPEC_SLX): Likewise.
20450 (VUNSPEC_LDA): Likewise.
20451 (VUNSPEC_STL): Likewise.
20452 * config/arm/sync.md (atomic_load<mode>): New pattern.
20453 (atomic_store<mode>): Likewise.
20454 (arm_load_acquire_exclusive<mode>): Likewise.
20455 (arm_load_acquire_exclusivesi): Likewise.
20456 (arm_load_acquire_exclusivedi): Likewise.
20457 (arm_store_release_exclusive<mode>): Likewise.
20459 2013-03-25 Catherine Moore <clm@codesourcery.com>
20461 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
20462 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
20463 * config/mip/predicates.md (lwsp_swsp_operand,
20464 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
20465 sb16_operand, db4_operand, db7_operand, ib3_operand,
20466 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
20467 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
20468 andi16_operand): New predicates.
20469 * config/mips/mips.md (compression): New attribute.
20470 (enabled): New attribute.
20471 (length): Consider compression in computing length.
20472 (shift_compression): New code attribute.
20473 (*add<mode>3): New operands. Record compression.
20474 (sub<mode>3): Likewise.
20475 (one_cmpl<mode>2): Likewise.
20476 (*and<mode>3): Likewise.
20477 (*ior<mode>3): Likewise.
20478 (unnamed pattern for xor): Likewise.
20479 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
20480 (*<optab><mode>3): Likewise.
20481 (*mov<mode>_internal: Likewise.
20482 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
20483 (mips_unsigned_immediate_p): New.
20484 (umips_lwsp_swsp_address_p): New.
20485 (m16_based_address_p): New.
20486 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
20487 (mips_unsigned_immediate_p): New prototype.
20488 (lwsp_swsp_address_p): New prototype.
20489 (m16_based_address_p): New prototype.
20490 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
20491 (mips_signed_immediate_p): New function.
20492 (m16_based_address_p): New function.
20493 (lwsp_swsp_address_p): New function.
20494 (mips_print_operand_punctuation): Recognize short delay slot insns
20495 for microMIPS.add<mode>3"
20497 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20500 * config/arm/iterators.md (v_cmp_result): New mode attribute.
20501 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
20503 2013-03-25 Richard Biener <rguenther@suse.de>
20505 PR tree-optimization/56689
20506 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
20509 2013-03-25 Richard Biener <rguenther@suse.de>
20511 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
20513 (memory_references): Likewise.
20514 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
20515 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
20516 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
20517 (gather_mem_refs_in_loops): Fold into ...
20518 (analyze_memory_references): ... this. Move initialization
20519 to tree_ssa_lim_initialize.
20520 (fill_always_executed_in): Rename to ...
20521 (fill_always_executed_in_1): ... this.
20522 (fill_always_executed_in): Move contains_call computation to
20523 this new function from ...
20524 (tree_ssa_lim_initialize): ... here.
20525 (tree_ssa_lim): Call fill_always_executed_in.
20527 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
20529 * postreload.c (reload_combine): Fix code detecting returns.
20531 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
20533 * function.c (emit_use_return_register_into_block): On cc0 targets,
20534 do not emit the sequence between cc0 setter and user.
20536 2013-03-25 Kai Tietz <ktietz@redhat.com>
20538 * config/i386/predicates.md (local_symbolic_operand): Interpret
20539 dll-imported symbols as none-local.
20541 2013-03-25 Richard Biener <rguenther@suse.de>
20543 * tree-ssa-loop-im.c (struct depend): Remove.
20544 (struct lim_aux_data): Make depends a vec of gimples.
20545 (free_lim_aux_data): Adjust.
20546 (add_dependency): Likewise.
20547 (set_level): Likewise.
20549 2013-03-25 Richard Biener <rguenther@suse.de>
20551 PR middle-end/56434
20552 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
20553 the pointer returned by calls with ECF_MALLOC set.
20555 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
20557 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
20559 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
20561 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
20562 using MMXMODE mode iterator.
20563 (*move<mode>_internal): Merge with *movv2sf_internal and
20564 *movv2sf_internal_rex64 using MMXMODE mode iterator.
20566 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
20568 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
20569 (record_last_mem_set_info): Likewise.
20571 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
20572 of XNEWVEC followed by memset.
20573 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
20575 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
20577 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
20578 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
20579 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
20580 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
20581 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
20582 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
20583 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
20584 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
20585 BARRIER_P instead of GET_CODE.
20587 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20589 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
20590 inaccuracy in the probing code.
20592 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
20593 (ctrapdi4): Likewise.
20595 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20597 * calls.c (expand_call): Add missing guard to code handling return
20598 of non-BLKmode structures in MSB.
20599 * function.c (expand_function_end): Likewise.
20601 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
20603 * combine.c (try_combine): Adjust comment. Do not add the set of
20604 insn #0 if the destination indirectly is set or dies in insn #2.
20605 Tidy up code to distribute a new note.
20607 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
20609 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
20610 also for alternatives 16 and 17.
20612 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
20614 * config/i386/sse.md (*mov<mode>_internal): Merge with
20615 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
20616 Emit insn template depending on type attribute. Use
20617 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
20618 movd instead of movq mnemonic for interunit moves. Rewrite mode
20619 attribute calculation. Remove unit attribute calculation.
20620 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
20621 Set prefix_data16 attribute for DImode ssemov types.
20622 Use Ym instead of y for SSE-MMX conversion alternatives.
20623 Reorder operand constraints.
20625 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
20627 * df.h (df_insn_delete): Adjust prototype.
20628 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
20629 and let it decide whether mark the basic block dirty.
20630 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
20631 * df-scan.c (df_insn_info_delete): New helper function, split
20632 off from df_insn_delete.
20633 (df_scan_free_bb_info): Use it.
20634 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
20636 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
20637 that the insn is actually an insn and it has a non-NULL basic block.
20638 Do not mark basic block dirty if only deleting a DEBUG_INSN.
20640 2013-03-22 Richard Biener <rguenther@suse.de>
20642 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
20644 (mem_ref_alloc): Do not allocate them.
20645 (refs_independent_p): Do not query or maintain a cache.
20647 2013-03-22 Richard Biener <rguenther@suse.de>
20649 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
20650 (gather_mem_refs_in_loops): Do not compute it.
20651 (analyze_memory_references): Do not allocate it.
20652 (tree_ssa_lim_finalize): Do not free it.
20653 (for_all_locs_in_loop): Do not query all_refs_in_loop.
20655 2013-03-22 Richard Biener <rguenther@suse.de>
20657 * is-a.h (as_a): Use gcc_checking_assert.
20659 2013-03-22 Ian Bolton <ian.bolton@arm.com>
20661 * config/aarch64/aarch64.c (aarch64_print_operand): New
20662 format specifier for printing a constant in hex.
20663 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
20664 format specifier for printing second operand.
20666 2013-03-22 Richard Biener <rguenther@suse.de>
20668 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
20670 (gather_mem_refs_in_loops): Perform store accumulation here.
20671 (create_vop_ref_mapping_loop): Remove.
20672 (create_vop_ref_mapping): Likewise.
20673 (analyze_memory_references): Initialize refs_stored_in_loop.
20674 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
20675 (record_indep_loop): Remove.
20676 (record_dep_loop): New function.
20677 (ref_indep_loop_p_1): Adjust to only walk over references
20678 in the loop, not its subloops.
20679 (ref_indep_loop_p): Rename to ...
20680 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
20681 maintaining a more fine-grained cache.
20682 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
20683 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
20685 2013-03-22 Richard Biener <rguenther@suse.de>
20687 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
20688 (struct mem_ref): Make accesses_in_loop a vec of a vec of
20689 aggregate mem_ref_loc.
20690 (free_mem_ref_locs): Inline into ...
20691 (memref_free): ... this and adjust.
20692 (mem_ref_alloc): Adjust.
20693 (mem_ref_locs_alloc): Remove.
20694 (record_mem_ref_loc): Adjust.
20695 (get_all_locs_in_loop): Rewrite into ...
20696 (for_all_locs_in_loop): ... this iterator.
20697 (rewrite_mem_ref_loc): New functor.
20698 (rewrite_mem_refs): Use for_all_locs_in_loop.
20699 (sm_set_flag_if_changed): New functor.
20700 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
20701 (ref_always_accessed): New functor.
20702 (ref_always_accessed_p): Use for_all_locs_in_loop.
20704 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20706 * tree-pass.h (PROP_gimple_lvec): New.
20707 * passes.c (dump_properties): Handle PROP_gimple_lvec.
20708 (init_optimization_passes): Move pass_lower_vector.
20709 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
20711 (pass_lower_vector): Provide PROP_gimple_lvec.
20712 (pass_lower_vector_ssa): Likewise.
20713 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
20715 2013-03-21 Mark Wielaard <mjw@redhat.com>
20717 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
20719 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
20721 * config/i386/i386.md (*movdi_internal): Disparage slightly
20722 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
20723 conversion alternatives.
20725 2013-03-21 Jakub Jelinek <jakub@redhat.com>
20727 PR middle-end/48087
20728 * diagnostic.def (DK_WERROR): New kind.
20729 * diagnostic.h (werrorcount): Define.
20730 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
20731 promoted to DK_ERROR, increment DK_WERROR counter instead of
20733 * toplev.c (toplev_main): Call print_ignored_options even if
20734 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
20735 even if just werrorcount is non-zero.
20738 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
20740 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
20741 (string_cst_pool_decl): New function.
20742 (optimize_one_addr_into_implicit_ptr): New function.
20743 (resolve_addr_in_expr): Optimize DWARF location expression
20744 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
20745 which doesn't live in memory, but has DW_AT_location or
20746 DW_AT_const_value, or refers to a string literal, into
20747 DW_OP_GNU_implicit_pointer.
20748 (optimize_location_into_implicit_ptr): New function.
20749 (resolve_addr): If removing DW_AT_location of a variable because
20750 it was DW_OP_addr of address of the variable, but the variable doesn't
20751 live in memory, try to emit const value attribute for the initializer.
20753 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20755 * tree.h (VECTOR_TYPE_P): New macro.
20756 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
20757 TYPE_MODE): Use it.
20758 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
20759 VEC_COND_EXPR cannot be lvalues.
20760 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
20762 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
20764 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
20765 Restrict the transformation to equal modes.
20767 2013-03-21 Richard Biener <rguenther@suse.de>
20769 PR tree-optimization/39326
20770 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
20771 (MEM_ANALYZABLE): Adjust.
20772 (record_mem_ref_loc): Move bitmap ops ...
20773 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
20774 unanalyzable refs, do not record locations for it.
20775 (analyze_memory_references): Allocate ref zero as shared
20777 (refs_independent_p): Do not test for unanalyzed mems here.
20778 (ref_indep_loop_p_1): Special-case disambiguation against
20779 the unanalyzed ref.
20780 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
20782 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
20784 * config/arm/arm-protos.h (tune_params): Add
20785 prefer_neon_for_64bits field.
20786 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
20787 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
20788 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
20789 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
20790 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
20791 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
20792 (arm_option_override): Handle -mneon-for-64bits new option.
20793 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
20794 (prefer_neon_for_64bits): Declare new variable.
20795 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
20796 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
20797 (arch_enabled): Handle new arch types. Remove support for onlya8
20799 (one_cmpldi2): Use new arch names.
20800 (zero_extend<mode>di2, extend<mode>di2): Ditto.
20801 * config/arm/arm.opt (mneon-for-64bits): Add option.
20802 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
20803 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
20804 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
20806 * doc/invoke.texi (-mneon-for-64bits): Document.
20808 2013-03-21 Richard Biener <rguenther@suse.de>
20810 PR tree-optimization/39326
20811 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
20812 (sort_bbs_in_loop_postorder_cmp): New function.
20813 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
20815 2013-03-21 Richard Biener <rguenther@suse.de>
20817 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
20818 (vect_insert_into_interleaving_chain): Likewise.
20819 (vect_drs_dependent_in_basic_block): Inline ...
20820 (vect_slp_analyze_data_ref_dependence): ... here. New function,
20822 (vect_analyze_data_ref_dependence): ... here. Simplify.
20823 (vect_check_interleaving): Simplify.
20824 (vect_analyze_data_ref_dependences): Likewise. Split out ...
20825 (vect_slp_analyze_data_ref_dependences): ... this new function.
20826 (dr_group_sort_cmp): New function.
20827 (vect_analyze_data_ref_accesses): Compute data-reference groups
20828 here instead of in vect_analyze_data_ref_dependence. Use
20829 a more efficient algorithm.
20830 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
20831 vect_slp_analyze_data_ref_dependences. Call
20832 vect_analyze_data_ref_accesses earlier.
20833 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
20834 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
20835 (vect_slp_analyze_data_ref_dependences): New prototype.
20837 2013-03-21 Richard Biener <rguenther@suse.de>
20839 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
20840 ref is stored in the loop.
20841 (find_refs_for_sm): Walk only over all stores.
20842 (store_motion_loop): Allocate from lim_bitmap_obstack.
20843 (store_motion): Likewise.
20845 2013-03-21 Richard Biener <rguenther@suse.de>
20847 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
20848 Update virtual SSA form.
20850 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20852 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
20853 * configure: Regenerate.
20854 * config.in: Regenerate.
20855 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
20856 if !HAVE_LD_EH_FRAME_CIEV3.
20858 2013-03-21 Richard Biener <rguenther@suse.de>
20860 * tree-cfg.c (verify_expr_no_block): New function.
20861 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
20862 nor DECL_VALUE_EXPR have locations with associated blocks.
20863 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
20864 (clear_unused_block_pointer): Remove code dealing with
20865 blocks in DECL_DEBUG_EXPR locations.
20867 2013-03-21 Richard Biener <rguenther@suse.de>
20869 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
20870 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
20871 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
20872 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
20873 instead of DECL_DEBUG_EXPR_IS_FROM.
20874 * gimplify.c (gimplify_modify_expr): Likewise.
20875 * tree-cfg.c (verify_expr_location_1): Likewise.
20876 * tree-complex.c (create_one_component_var): Likewise.
20877 * tree-sra.c (create_access_replacement): Likewise.
20878 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
20879 (clear_unused_block_pointer): Likewise.
20880 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
20881 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
20882 * var-tracking.c (var_debug_decl): Likewise.
20883 (track_expr_p): Likewise.
20884 * tree-inline.c (add_local_variables): Likewise. Set
20885 DECL_HAS_DEBUG_EXPR_P after copying it.
20886 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
20887 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
20889 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
20892 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
20893 * configure: Regenerate.
20894 * config.in: Regenerate.
20895 * config/i386/i386.md (*movdf_internal): Use
20896 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
20897 movd instead of movq mnemonic for interunit moves.
20898 (*movdi_internal): Ditto.
20900 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20902 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
20903 (abd<mode>_3): New pattern.
20904 (aba<mode>_3): New pattern.
20905 (fabd<mode>_3): New pattern.
20907 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20909 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
20910 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
20911 occurrence of REGISTER_PREFIX as its empty string.
20913 2013-03-20 Jeff Law <law@redhat.com>
20915 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
20916 addititional equivalences for equality comparisons between an SSA_NAME
20917 and a constant where the SSA_NAME was set from a widening conversion.
20919 2013-03-20 Walter Lee <walt@tilera.com>
20921 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
20923 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20925 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
20926 depending on type attribute.
20927 (*movti_internal): Ditto.
20928 (*movtf_internal): Ditto.
20929 (*movxf_internal): Ditto.
20930 (*movdf_internal): Ditto.
20931 (*movsf_internal): Ditto.
20933 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20935 * config/i386/i386.md (*movti_internal): Set prefix attribute to
20936 maybe_vex for sselog1 and ssemov types.
20937 (*movdi_internal): Reorder operand constraints.
20938 (*movsi_internal): Ditto. Set prefix attribute to
20939 maybe_vex for sselog1 and ssemov types.
20940 (*movtf_internal): Set prefix attribute to maybe_vex
20941 for sselog1 and ssemov types.
20942 (*movdf_internal): Ditto. Set prefix_data16 attribute for
20943 DImode ssemov types. Reorder operand constraints.
20944 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
20945 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
20946 attribute for SImode ssemov types. Reorder operand constraints.
20948 2013-03-20 Martin Jambor <mjambor@suse.cz>
20950 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
20951 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
20953 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
20955 * config/rs6000/predicates.md (indexed_address, update_address_mem
20956 update_indexed_address_mem): New predicates.
20957 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
20958 attribute for load/store instructions.
20959 * config/rs6000/dfp.md (movsd_store): Likewise.
20960 (movsd_load): Likewise.
20961 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
20962 (unnamed HI->DI extend define_insn): Likewise.
20963 (unnamed SI->DI extend define_insn): Likewise.
20964 (unnamed QI->SI extend define_insn): Likewise.
20965 (unnamed QI->HI extend define_insn): Likewise.
20966 (unnamed HI->SI extend define_insn): Likewise.
20967 (unnamed HI->SI extend define_insn): Likewise.
20968 (extendsfdf2_fpr): Likewise.
20969 (movsi_internal1): Likewise.
20970 (movsi_internal1_single): Likewise.
20971 (movhi_internal): Likewise.
20972 (movqi_internal): Likewise.
20973 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
20974 attribute for load/store instructions.
20975 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
20977 (mov<mode>_softfloat): Likewise.
20978 (mov<mode>_hardfloat32): Likewise.
20979 (mov<mode>_hardfloat64): Likewise.
20980 (mov<mode>_softfloat64): Likewise.
20981 (movdi_internal32): Likewise.
20982 (movdi_internal64): Likewise.
20983 (probe_stack_<mode>): Likewise.
20985 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
20987 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
20988 floating point, and decimal floating point to reload iterator.
20990 * config/rs6000/constraints.md (wl constraint): New constraints to
20991 return FLOAT_REGS if certain options are used to reduce the number
20992 of separate patterns that exist in the file.
20993 (wx constraint): Likewise.
20994 (wz constraint): Likewise.
20996 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20997 -mdebug=reg, print wg, wl, wx, and wz constraints.
20998 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20999 Initialize the reload functions for 64-bit binary/decimal floating
21001 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
21002 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
21003 create the buffer on the stack to overcome not having a 32-bit
21005 (rs6000_emit_move): Likewise.
21006 (rs6000_secondary_memory_needed_rtx): Likewise.
21007 (rs6000_alloc_sdmode_stack_slot): Likewise.
21008 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
21009 via xxlxor, just like DFmode 0.0.
21011 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
21012 define as 1 if we are running on a power7 or newer.
21013 (enum r6000_reg_class_enum): Add new constraints.
21015 * config/rs6000/dfp.md (movsd): Delete, combine with binary
21016 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
21017 with other moves by using conditional constraits (wg). Use LFIWZX
21018 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
21019 (movsd splitter): Likewise.
21020 (movsd_hardfloat): Likewise.
21021 (movsd_softfloat): Likewise.
21023 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
21024 binary and decimal floating point moves.
21025 (fmove_ok): New attributes to combine binary and decimal floating
21026 point moves, and to combine power6x (mfpgpr) moves along normal
21028 (real_value_to_target): Likewise.
21029 (f32_lr): Likewise.
21030 (f32_lm): Likewise.
21031 (f32_li): Likewise.
21032 (f32_sr): Likewise.
21033 (f32_sm): Likewise.
21034 (f32_si): Likewise.
21035 (movsf): Combine binary and decimal floating point moves. Combine
21036 power6x (mfpgpr) moves with other moves by using conditional
21037 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
21038 (mov<mode> for SFmode/SDmode); Likewise.
21039 (SFmode/SDmode splitters): Likewise.
21040 (movsf_hardfloat): Likewise.
21041 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
21042 (movsf_softfloat): Likewise.
21043 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
21045 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
21046 wx and wz constraints.
21048 * config/rs6000/constraints.md (wg constraint): New constraint to
21049 return FLOAT_REGS if -mmfpgpr (power6x) was used.
21051 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
21054 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
21055 -mdebug=reg, print wg, wl, wx, and wz constraints.
21056 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
21057 Initialize the reload functions for 64-bit binary/decimal floating
21059 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
21060 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
21061 create the buffer on the stack to overcome not having a 32-bit
21063 (rs6000_emit_move): Likewise.
21064 (rs6000_secondary_memory_needed_rtx): Likewise.
21065 (rs6000_alloc_sdmode_stack_slot): Likewise.
21066 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
21067 via xxlxor, just like DFmode 0.0.
21069 * config/rs6000/dfp.md (movdd): Delete, combine with binary
21070 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
21071 with other moves by using conditional constraits (wg). Use LFIWZX
21072 and STFIWX for loading SDmode on power7.
21073 (movdd splitters): Likewise.
21074 (movdd_hardfloat32): Likewise.
21075 (movdd_softfloat32): Likewise.
21076 (movdd_hardfloat64_mfpgpr): Likewise.
21077 (movdd_hardfloat64): Likewise.
21078 (movdd_softfloat64): Likewise.
21080 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
21081 64-bit binary and decimal floating point moves.
21082 (FMOVE64X): Likewise.
21083 (movdf): Combine 64-bit binary and decimal floating point moves.
21084 Combine power6x (mfpgpr) moves with other moves by using
21085 conditional constraits (wg).
21086 (mov<mode> for DFmode/DDmode): Likewise.
21087 (DFmode/DDmode splitters): Likewise.
21088 (movdf_hardfloat32): Likewise.
21089 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
21090 (movdf_softfloat32): Likewise.
21091 (movdf_hardfloat64_mfpgpr): Likewise.
21092 (movdf_hardfloat64): Likewise.
21093 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
21094 (movdf_softfloat64): Likewise.
21095 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
21096 (reload_<mode>_load): Move to later in the file so they aren't in
21097 the middle of the floating point move insns.
21098 (reload_<mode>_store): Likewise.
21100 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
21103 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
21104 constraint if -mdebug=reg.
21105 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
21106 Enable using dd reload support if needed.
21108 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
21109 binary and decimal floating point moves in rs6000.md.
21110 (movtd_internal): Likewise.
21112 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
21113 decimal floating point moves.
21115 (movtf_internal): Likewise.
21116 (mov<mode>_internal, TDmode/TFmode): Likewise.
21117 (movtf_softfloat): Likewise.
21118 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
21120 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
21121 movdi_internal64, using wg constraint for move direct operations.
21122 (movdi_internal64): Likewise.
21124 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
21125 MODES_TIEABLE_P for selected modes. Print the numerical value of
21126 the various virtual registers. Use GPR/FPR first/last values,
21127 instead of hard coding the register numbers. Print which modes
21128 have reload functions registered.
21129 (rs6000_option_override_internal): If -mdebug=reg, trace the options
21130 settings before/after setting cpu, target and subtarget settings.
21131 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
21132 and for secondary reload failures in rs6000_secondary_reload_inner.
21133 (rs6000_secondary_reload_fail): Likewise.
21134 (rs6000_secondary_reload_inner): Likewise.
21136 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
21137 macros for first/last GPR and FPR registers.
21138 (LAST_GPR_REGNO): Likewise.
21139 (FIRST_FPR_REGNO): Likewise.
21140 (LAST_FPR_REGNO): Likewise.
21142 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
21143 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
21144 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
21145 (vcond<mode><mode>): Likewise.
21146 (vcondu<mode><mode>): Likewise.
21147 (vector_gtu<mode>): Likewise.
21148 (vector_gte<mode>): Likewise.
21149 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
21150 to prevent the compiler from converting DImode operations to TImode.
21151 (ior<mode>3): Likewise.
21152 (and<mode>3): Likewise.
21153 (one_cmpl<mode>2): Likewise.
21154 (nor<mode>3): Likewise.
21155 (andc<mode>3): Likewise.
21157 * config/rs6000/constraints.md (wt constraint): New constraint
21158 that returns VSX_REGS if TImode is allowed in VSX registers.
21160 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
21161 constant under VSX.
21163 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
21164 similar to TImode, but it is restricted to being in the GPRs.
21166 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
21167 TImode to occupy a single VSX register.
21169 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
21170 -mvsx-timode for power7/power8.
21171 (power7 cpu): Likewise.
21172 (power8 cpu): Likewise.
21174 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
21175 sure that TFmode/TDmode take up two registers if they are ever
21176 allowed in the upper VSX registers.
21177 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
21179 (rs6000_init_hard_regno_mode_ok): Likewise.
21180 (rs6000_debug_reg_global): Add debugging for PTImode and wt
21181 constraint. Print if LRA is turned on.
21182 (rs6000_option_override_internal): Give an error if -mvsx-timode
21183 and VSX is not enabled.
21184 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
21185 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
21186 to reg+offset addressing. Use PTImode when checking offset
21187 addresses for validity.
21188 (reg_offset_addressing_ok_p): Likewise.
21189 (rs6000_legitimate_offset_address_p): Likewise.
21190 (rs6000_legitimize_address): Likewise.
21191 (rs6000_legitimize_reload_address): Likewise.
21192 (rs6000_legitimate_address_p): Likewise.
21193 (rs6000_eliminate_indexed_memrefs): Likewise.
21194 (rs6000_emit_move): Likewise.
21195 (rs6000_secondary_reload): Likewise.
21196 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
21197 reloads to fpr registers to continue to use reg+offset addressing,
21198 but 64-bit reloads to altivec registers need reg+reg addressing.
21199 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
21200 it. Treat LO_SUM like a PLUS operation.
21201 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
21202 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
21203 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
21204 registers to share a register with a smaller sized type, since VSX
21205 puts scalars in the upper 64-bits.
21206 (print_operand): Add support for PTImode.
21207 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
21208 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
21209 registers, but don't have arithmetic support.
21210 (rs6000_memory_move_cost): Add test for VSX.
21211 (rs6000_opt_masks): Add -mvsx-timode.
21213 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
21216 (VSr): Use wt constraint for TImode.
21217 (VSv): Drop TImode support.
21218 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
21219 (vsx_movti_64bit): Likewise.
21220 (vsx_movti_32bit): Likewise.
21221 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
21222 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
21223 one '?' on the appropriate output constraint. Do not allow TImode
21224 logical operations on 32-bit systems.
21225 (vsx_ior<mode>3): Likewise.
21226 (vsx_xor<mode>3): Likewise.
21227 (vsx_one_cmpl<mode>2): Likewise.
21228 (vsx_nor<mode>3): Likewise.
21229 (vsx_andc<mode>3): Likewise.
21230 (vsx_concat_<mode>): Likewise.
21231 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
21233 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
21234 OPTION_MASK_VSX_TIMODE.
21235 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
21236 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
21238 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
21239 (TI2 iterator): New iterator for TImode, PTImode.
21240 (wd mode attribute): Add values for vector types.
21241 (movti_string): Replace TI move operations with operations for TImode
21242 and PTImode. Add support for TImode being allowed in VSX registers.
21243 (mov<mode>_string, TImode/PTImode): Likewise.
21244 (movti_ppc64): Likewise.
21245 (mov<mode>_ppc64, TImode/PTImode): Likewise.
21246 (TI mode splitters): Likewise.
21248 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
21251 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
21253 PR tree-optimization/56355
21254 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
21255 Also handle integers with undefined overflow.
21257 2013-03-20 Catherine Moore <clm@codesourcery.com>
21258 Maciej W. Rozycki <macro@codesourcery.com>
21259 Tom de Vries <tom@codesourcery.com>
21260 Nathan Sidwell <nathan@codesourcery.com>
21261 Iain Sandoe <iain@codesourcery.com>
21262 Nathan Froyd <froydnj@codesourcery.com>
21263 Chao-ying Fu <fu@mips.com>
21265 * doc/extend.texi: (micromips, nomicromips, nocompression):
21266 Document new function attributes.
21267 * doc/invoke.texi (minterlink-compressed, mmicromips,
21268 m14k, m14ke, m14kec): Document new options.
21269 (minterlink-mips16): Update documentation.
21270 * doc/md.texi (ZC, ZD): Document new constraints.
21271 * configure.ac (gcc_cv_as_micromips): Check if linker
21272 supports the .set micromips directive.
21273 * configure: Regenerate.
21274 * config.in: Regenerate.
21275 * config/mips/mips-tables.opt: Regenerate.
21276 * config/mips/micromips.md: New file.
21277 * constraints.md (ZC, ZD): New constraints.
21278 * config/mips/predicates.md (movep_src_register): New predicate.
21279 (movep_src_operand): New predicate.
21280 (non_volatile_mem_operand): New predicate.
21281 * config/mips/mips.md (multimem): New type.
21282 (length): Differentiate between 17-bit and 18-bit branch offsets.
21283 (MOVEP1, MOVEP2): New mode iterator.
21284 (mov_<load>l): Use ZC constraint.
21285 (mov_<load>r): Likewise.
21286 (mov_<store>l): Likewise.
21287 (mov_<store>r): Likewise.
21288 (*branch_equality<mode>_inverted): Add microMIPS support.
21289 (*branch_equality<mode>): Likewise.
21290 (*jump_absolute): Likewise.
21291 (indirect_jump_<mode>): Likewise.
21292 (tablejump_<mode>): Likewise.
21293 (<optab>_internal): Likewise.
21294 (sibcall_internal): Likewise.
21295 (sibcall_value_internal): Likewise.
21296 (prefetch): Use constraint ZD.
21297 * config/mips/mips.opt (minterlink-compressed): New option.
21298 (minterlink-mips16): Now an alias for minterlink-compressed.
21299 (mmicromips): New option.
21300 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21301 (compare_and_swap_12): Likewise.
21302 (sync_add<mode>): Likewise.
21303 (sync_<optab>_12): Likewise.
21304 (sync_old_<optab>_12): Likewise.
21305 (sync_new_<optab>_12): Likewise.
21306 (sync_nand_12): Likewise.
21307 (sync_old_nand_12): Likewise.
21308 (sync_new_nand_12): Likewise.
21309 (sync_sub<mode>): Likewise.
21310 (sync_old_add<mode>): Likewise.
21311 (sync_old_sub<mode>): Likewise.
21312 (sync_new_add<mode>): Likewise.
21313 (sync_new_sub<mode>): Likewise.
21314 (sync_<optab><mode>): Likewise.
21315 (sync_old_<optab><mode>): Likewise.
21316 (sync_new_<optab><mode>): Likewise.
21317 (sync_nand<mode>): Likewise.
21318 (sync_old_nand<mode>): Likewise.
21319 (sync_new_nand<mode>): Likewise.
21320 (sync_lock_test_and_set<mode>): Likewise.
21321 (test_and_set_12): Likewise.
21322 (atomic_compare_and_swap<mode>): Likewise.
21323 (atomic_exchange<mode>_llsc): Likewise.
21324 (atomic_fetch_add<mode>_llsc): Likewise.
21325 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21326 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21327 (umips_save_restore_pattern_p): Likewise.
21328 (umips_load_store_pair_p): Likewise.
21329 (umips_output_load_store_pair): Likewise.
21330 (umips_movep_target_p): Likewise.
21331 (umips_12bit_offset_address_p): Likewise.
21332 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21333 (mips_base_mips16): Rename this...
21334 (mips_base_compression_flags): ...to this. Update all uses.
21335 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21336 (mips_mips16_decl_p): Delete.
21337 (mips_nomips16_decl_p): Delete.
21338 (mips_get_compress_on_flags): New function.
21339 (mips_get_compress_off_flags): New function.
21340 (mips_get_compress_mode): New function.
21341 (mips_get_compress_on_name): New function.
21342 (mips_get_compress_off_name): New function.
21343 (mips_insert_attributes): Support multiple compression types.
21344 (mips_merge_decl_attributes): Likewise.
21345 (umips_12bit_offset_address_p): New function.
21346 (mips_start_function_definition): Emit .set micromips directive.
21347 (mips_call_may_need_jalx_p): New function.
21348 (mips_function_ok_for_sibcall): Add microMIPS support.
21349 (mips_print_operand_punctuation): Support short delay slots and
21351 (umips_swm_mask, umips_swm_encoding): New.
21352 (umips_build_save_restore): New function.
21353 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21354 (was_mips16_p): Remove.
21355 (old_compression_mode): New.
21356 (mips_set_compression_mode): New function.
21357 (mips_set_current_function): Add microMIPS support.
21358 (mips_option_override): Likewise.
21359 (umips_save_restore_pattern_p): New function.
21360 (umips_output_save_restore): New function.
21361 (umips_load_store_pair_p_1): New function.
21362 (umips_load_store_pair_p): New function.
21363 (umips_output_load_store_pair_1): New function.
21364 (umips_output_load_store_pair): New function.
21365 (umips_movep_target_p) New function.
21366 (mips_prepare_pch_save): Add microMIPS support.
21367 * config/mips/mips.h (TARGET_COMPRESSION): New.
21368 (TARGET_CPU_CPP_BUILTINS): Update macro
21369 to use new compression flags and to support microMIPS.
21370 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21371 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21372 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21373 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21374 (ASM_SPEC): Support mmicromips and mno-micromips.
21375 (M16STORE_REG_P): New macro.
21376 (MIPS_CALL): Support TARGET_MICROMIPS.
21377 (MICROMIPS_J): New macro.
21378 (mips_base_mips16): Rename this...
21379 (mips_base_compression_flags): ...to this.
21380 (UMIPS_12BIT_OFFSET_P): New macro.
21381 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21382 (MULTILIB_DIRNAMES): Likewise.
21383 2013-03-20 Richard Biener <rguenther@suse.de>
21385 PR tree-optimization/56661
21386 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
21387 the result does not have to be distinct.
21389 2013-03-20 Richard Biener <rguenther@suse.de>
21391 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
21394 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21395 Steven Bosscher <steven@gcc.gnu.org>
21397 PR rtl-optimization/56605
21398 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
21400 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
21403 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
21404 that require movd instead of movq.
21406 2013-03-20 Richard Biener <rguenther@suse.de>
21408 * tree-ssa-structalias.c (struct variable_info): Add pointer
21409 to the first field of an aggregate with sub-vars. Make
21410 this and the pointer to the next subfield its ID.
21411 (vi_next): New function.
21412 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
21413 storedanything_id, integer_id): Increment by one.
21414 (new_var_info, get_call_vi, lookup_call_clobber_vi,
21415 get_call_clobber_vi): Adjust.
21416 (solution_set_expand): Simplify and speedup.
21417 (solution_set_add): Inline into ...
21418 (set_union_with_increment): ... this. Adjust accordingly.
21419 (do_sd_constraint): Likewise.
21420 (do_ds_constraint): Likewise.
21421 (do_complex_constraint): Simplify.
21422 (build_pred_graph): Adjust.
21423 (solve_graph): Likewise. Simplify and speedup.
21424 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
21425 get_constraint_for_component_ref, get_constraint_for_1,
21426 first_vi_for_offset, first_or_preceding_vi_for_offset,
21427 create_function_info_for, create_variable_info_for_1,
21428 create_variable_info_for, intra_create_variable_infos): Adjust.
21429 (init_base_vars): Push NULL for ID zero.
21430 (compute_points_to_sets): Adjust.
21432 2013-03-20 Richard Biener <rguenther@suse.de>
21434 * cfgloop.c (verify_loop_structure): Streamline and avoid
21435 ICEing on corrupt loop tree.
21436 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
21439 2013-03-20 Richard Biener <rguenther@suse.de>
21441 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
21442 check whether an SSA update is needed.
21444 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
21446 * config/mips/constraints.md (T): Rename to...
21450 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
21451 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
21453 2013-03-19 Ian Bolton <ian.bolton@arm.com>
21455 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
21456 (*subsi3_carryin_uxtw): Likewise.
21458 2013-03-19 Ian Bolton <ian.bolton@arm.com>
21460 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
21461 (*rorsi3_insn_uxtw): Likewise.
21463 2013-03-19 Ian Bolton <ian.bolton@arm.com>
21465 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
21466 (*extrsi5_insn_uxtw): Likewise.
21468 2013-03-19 Richard Biener <rguenther@suse.de>
21470 PR tree-optimization/56273
21471 * passes.c (init_optimization_passes): Move second VRP after DOM.
21473 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
21475 * config/i386/i386.md (*movti_internal): Merge from
21476 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
21477 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
21478 nox64 isa attributes.
21480 2013-03-18 Richard Biener <rguenther@suse.de>
21482 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
21484 (merge_node_constraints): Likewise.
21485 (build_succ_graph): Likewise.
21486 (valid_graph_edge): Inline into single caller.
21487 (unify_nodes): Likewise. Use bitmap_set_bit return value
21489 (scc_visit): Fix formatting and variable use.
21490 (do_sd_constraint): Use gcc_checking_assert.
21491 (do_ds_constraint): Likewise.
21492 (do_complex_constraint): Likewise.
21493 (condense_visit): Likewise. Cleanup.
21494 (dump_pred_graph): New function.
21495 (perform_var_substitution): Dump the pred-graph before
21496 variable substitution.
21497 (find_equivalent_node): Use gcc_checking_assert.
21498 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
21500 2013-03-18 Richard Biener <rguenther@suse.de>
21502 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
21503 Remove cond_expr_stmt_list argument and do not gimplify the
21505 (vect_loop_versioning): Adjust.
21506 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
21507 Cleanup to use less temporaries.
21508 (vect_create_data_ref_ptr): Cleanup.
21510 2013-03-18 Jakub Jelinek <jakub@redhat.com>
21512 PR tree-optimization/56635
21513 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
21514 require types_compatible_p types.
21516 2013-03-18 Nick Clifton <nickc@redhat.com>
21518 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
21519 spurious backslash.
21521 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
21522 Add missing line to comment describing function.
21524 2013-03-18 Richard Biener <rguenther@suse.de>
21526 PR tree-optimization/56210
21527 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21528 Handle string / character search functions.
21529 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
21531 2013-03-18 Richard Biener <rguenther@suse.de>
21533 PR middle-end/56483
21534 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
21535 and implement properly.
21536 * gimple.h (gimple_cond_single_var_p): Remove.
21538 2013-03-18 Richard Biener <rguenther@suse.de>
21540 * tree-data-ref.h (find_data_references_in_loop): Declare.
21541 * tree-data-ref.c (get_references_in_stmt): Use a stack
21542 vector pre-allocated in the callers.
21543 (find_data_references_in_stmt): Adjust.
21544 (graphite_find_data_references_in_stmt): Likewise.
21545 (create_rdg_vertices): Likewise.
21546 (find_data_references_in_loop): Export.
21547 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
21548 Compute dependences here...
21549 (vect_analyze_data_refs): ...not here. When we encounter
21550 a non-vectorizable data reference in basic-block vectorization
21551 truncate the data reference vector. Do not bother to
21552 fixup data-dependence information for gather loads.
21553 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
21554 of data references, as reported.
21556 2013-03-18 Richard Biener <rguenther@suse.de>
21558 PR tree-optimization/3713
21559 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
21560 has_constants and expr.
21561 (stmt_has_constants): Properly valueize SSA names when deciding
21562 whether the stmt has constants.
21564 2013-03-18 Richard Biener <rguenther@suse.de>
21566 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
21567 whole function when there is nothing to do.
21568 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
21569 * tree-vectorizer.c (vectorize_loops): Update virtual and
21570 loop-closed SSA once.
21571 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
21573 2013-03-18 Richard Biener <rguenther@suse.de>
21575 PR middle-end/56113
21576 * domwalk.c (bb_postorder): New global static.
21577 (cmp_bb_postorder): New function.
21578 (walk_dominator_tree): Replace scheme imposing an order for
21579 visiting dominator sons by one sorting them at the time they
21580 are pushed on the stack.
21582 2013-03-18 Richard Biener <rguenther@suse.de>
21584 PR tree-optimization/39326
21585 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
21586 (struct mem_ref): Replace mem member with ao_ref typed member.
21587 (MEM_ANALYZABLE): Adjust.
21588 (memref_eq): Likewise.
21589 (mem_ref_alloc): Likewise.
21590 (gather_mem_refs_stmt): Likewise.
21591 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
21592 (execute_sm_if_changed_flag_set): Adjust.
21593 (execute_sm): Likewise.
21594 (ref_always_accessed_p): Likewise.
21595 (refs_independent_p): Likewise.
21596 (can_sm_ref_p): Likewise.
21598 2013-03-18 Jakub Jelinek <jakub@redhat.com>
21601 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
21602 return 1 even for !unsignedp.
21604 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21606 * config/i386/i386.md (isa): Add x64 and nox64.
21607 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
21608 (*pushtf): Enable *roF alternative for x64 isa only.
21609 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
21610 mode attribute of integer alternatives to DImode for TARGET_64BIT.
21611 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
21612 (*movtf_internal): Merge from *movtf_internal_rex64 and
21613 *movtf_internal_sse. Use x64 and nox64 isa attributes.
21614 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
21615 nox64 isa attributes.
21616 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
21617 nox64 isa attributes.
21618 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
21620 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
21622 * config/alpha/alpha.c (TARGET_LRA_P): New define.
21624 2013-03-17 Jakub Jelinek <jakub@redhat.com>
21627 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
21628 class names. Remove trailing comma after "ALL_REGS".
21630 2013-03-16 Jan Hubicka <jh@suse.cz>
21632 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
21633 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
21634 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
21635 of cgraph_get_create_node.
21636 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
21638 2013-03-16 Jason Merrill <jason@redhat.com>
21641 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
21642 with DW_AT_default_value.
21644 2013-03-16 Jakub Jelinek <jakub@redhat.com>
21646 * BASE-VER: Set to 4.9.0.
21648 2013-03-14 Andi Kleen <ak@linux.intel.com>
21651 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
21652 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
21653 Document _x* TSX intrinsics.
21655 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
21656 David Holsgrove <david.holsgrove@xilinx.com>
21658 * configure.ac: Add MicroBlaze TLS support detection.
21659 * configure: Regenerate.
21660 * config/microblaze/microblaze-protos.h
21661 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
21662 symbol_mentioned_p, label_mentioned_p): Add prototypes.
21663 * config/microblaze/microblaze.c (microblaze_address_type): Add
21664 ADDRESS_TLS and tls_reloc address types.
21665 (microblaze_address_info): Add tls_reloc.
21666 (TARGET_HAVE_TLS): Define.
21667 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
21668 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
21669 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
21670 load_tls_operand, microblaze_call_tls_get_addr,
21671 microblaze_legitimize_tls_address): New functions.
21672 (microblaze_classify_unspec): Handle UNSPEC_TLS.
21673 (get_base_reg): Use microblaze_tls_symbol_p.
21674 (microblaze_classify_address): Handle TLS.
21675 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
21676 label_mentioned_p and microblaze_tls_referenced_p.
21677 (microblaze_legitimize_address): Handle TLS.
21678 (microblaze_address_insns): Handle ADDRESS_TLS.
21679 (pic_address_needs_scratch): Handle TLS.
21680 (print_operand_address): Handle TLS.
21681 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
21682 (microblaze_expand_move): Handle TLS.
21683 (microblaze_legitimate_constant_p): Check
21684 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
21685 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
21686 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
21687 (PIC_OFFSET_TABLE_REGNUM): Set.
21688 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
21689 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
21690 (addsi3, movsi_internal2, movdf_internal): Update constraints
21691 * config/microblaze/predicates.md (arith_plus_operand): Define
21692 (move_operand): Redefine as move_src_operand,
21693 check microblaze_tls_referenced_p.
21695 2013-03-14 Ian Bolton <ian.bolton@arm.com>
21697 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
21698 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
21700 2013-03-14 Ian Bolton <ian.bolton@arm.com>
21702 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
21705 2013-03-14 Jakub Jelinek <jakub@redhat.com>
21707 PR tree-optimization/53265
21708 * common.opt (Waggressive-loop-optimizations): New option.
21709 * tree-ssa-loop-niter.c: Include tree-pass.h.
21710 (do_warn_aggressive_loop_optimizations): New function.
21711 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
21712 if number_of_latch_executions returned constant.
21713 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
21714 early. If number_of_latch_executions returned constant, set
21715 nb_iterations_upper_bound back to it.
21716 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
21718 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
21719 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
21721 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
21722 (MULTILIB_OSDIRNAMES): Set.
21723 * genmultilib: If defaultosdirname doesn't start with :: , set
21724 defaultosdirname2 instead, clear it and emit two . multilib_raw
21725 entries instead of just one.
21727 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
21729 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
21730 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
21731 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
21732 (SUBTARGET_OVERRIDE_OPTIONS): New.
21734 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
21737 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
21738 (musermode): Convert to Var(TARGET_USERMODE).
21739 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
21740 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
21741 * config/sh/sh.c (sh_option_override): Use
21742 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
21743 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
21745 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
21747 (udivsi3_i4_single, divsi3_i4_single): Use
21748 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
21750 2013-03-13 Dave Korn <dave.korn.cygwin@....>
21752 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
21755 2013-03-13 Richard Biener <rguenther@suse.de>
21757 PR tree-optimization/56608
21758 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
21759 calls when vectorizing basic-blocks.
21761 2013-03-13 Jakub Jelinek <jakub@redhat.com>
21764 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
21767 2013-03-12 Jakub Jelinek <jakub@redhat.com>
21769 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
21771 2013-03-11 Jan Hubicka <jh@suse.cz>
21774 * lto-streamer-out.c (output_symbol_p): Skip references from
21775 constructors of external variables.
21777 2013-03-11 Jan Hubicka <jh@suse.cz>
21779 PR middle-end/56571
21780 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
21782 * emit-rtl.c (verify_rtx_sharing): Likewise.
21783 (copy_insn_1): Likewise.
21784 * rtl.c (copy_rtx): Likewise.
21786 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
21789 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
21790 output_operand_lossage message.
21792 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
21795 * arm.c (shift_op): Validate RTL pattern on the fly.
21796 (arm_print_operand, case 'S'): Don't use shift_operator to validate
21799 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21802 * config/pa/pa.md (call_value): Check for calls to powf and direct to
21803 new call patterns that clobber %fr12.
21804 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
21805 split and postreload patterns.
21806 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
21807 registers %fr12 and %fr12R as call used.
21809 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
21811 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
21812 (canon_address, record_store, replace_read, check_mem_read_rtx,
21813 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
21814 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
21815 rest_of_handle_dse): Likewise.
21817 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
21819 PR middle-end/56524
21820 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
21822 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
21823 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
21824 (save_optabs_if_changed): Replace with...
21825 (init_tree_optimization_optabs): ...this.
21826 * optabs.c (save_optabs_if_changed): Rename to...
21827 (init_tree_optimization_optabs): ...this. Take the optimization node
21828 as argument. Do nothing if the base optabs are already correct.
21829 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
21830 to recompute optabs.
21831 * function.h (function): Remove optabs field.
21832 * function.c (invoke_set_current_function_hook): Call
21833 init_tree_optimization_optabs. Use the result to initialize
21836 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
21838 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
21839 if GTMA_HAS_NO_INSTRUMENTATION.
21840 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
21841 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
21842 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
21843 * gimple-pretty-print.c (dump_gimple_transaction): Handle
21844 GTMA_HAS_NO_INSTRUMENTATION.
21846 2013-03-08 Jakub Jelinek <jakub@redhat.com>
21848 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
21851 2013-03-08 Marek Polacek <polacek@redhat.com>
21852 Jakub Jelinek <jakub@redhat.com>
21854 PR tree-optimization/56478
21855 * predict.c (is_comparison_with_loop_invariant_p): Change the
21856 type of loop_step to tree.
21857 (predict_loops): Adjust.
21858 (predict_iv_comparison): Perform the computations on double_ints.
21860 2013-03-08 Richard Biener <rguenther@suse.de>
21862 PR tree-optimization/56570
21863 * tree-cfg.c (verify_expr_location_1): Verify locations for
21865 * tree-sra.c (create_access_replacement): Strip locations
21866 from DECL_DEBUG_EXPRs.
21868 2013-03-08 Richard Biener <rguenther@suse.de>
21870 * tree-inline.c (expand_call_inline): Do not associate
21871 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
21872 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
21874 2013-03-08 Richard Biener <rguenther@suse.de>
21876 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
21877 or block changes with -Og. Fix for location / block encoding
21878 changes and PHI arguments with locations.
21880 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
21882 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
21884 (struct output_info): Likewise.
21885 (register_overhead): Remove bad gcc_assert.
21886 (bitmap_find_bit): If there is only a single bitmap element, do not
21887 count a miss as a search.
21888 (print_statistics): Update for counter type changes.
21889 (dump_bitmap_statistics): Likewise. Print headers such that they
21890 are properly lined up with the printed counters.
21892 2013-03-07 Jakub Jelinek <jakub@redhat.com>
21894 PR tree-optimization/56559
21895 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
21896 check that it has only a single use.
21898 2013-03-07 Richard Biener <rguenther@suse.de>
21900 * doc/invoke.texi (fwhole-program): Discourage use in combination
21903 2013-03-06 Jakub Jelinek <jakub@redhat.com>
21905 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
21907 PR tree-optimization/56539
21908 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
21909 instead of GSI_CONTINUE_LINKING as last argument to
21910 force_gimple_operand_gsi. Adjust function comment.
21912 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
21915 PR middle-end/56548
21916 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
21917 promoted mode, convert the result back to the original mode.
21919 2013-03-06 Richard Biener <rguenther@suse.de>
21921 PR middle-end/56294
21922 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
21923 (insert_updated_phi_nodes_compare_uids): New function.
21924 (update_ssa): Sort symbols_to_rename after UID before
21925 traversing it to insert PHI nodes.
21927 2013-03-06 Richard Biener <rguenther@suse.de>
21929 PR middle-end/50494
21930 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
21931 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
21934 2013-02-13 Richard Biener <rguenther@suse.de>
21937 * varasm.c (output_constant_def_1): Get the decl representing
21938 the constant as argument.
21939 (output_constant_def): Wrap output_constant_def_1.
21940 (make_decl_rtl): Use output_constant_def_1 with the decl
21941 representing the constant.
21942 (build_constant_desc): Optionally re-use a decl already
21943 representing the constant.
21944 (tree_output_constant_def): Adjust.
21946 2013-03-06 Joey Ye <joey.ye@arm.com>
21949 * gcc.c (convert_white_space): New function.
21950 (main): Handles white space in function name.
21952 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
21955 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
21956 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
21957 to SH_DIV_CALL_TABLE for TARGET_SH2.
21958 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
21960 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
21961 call-table options.
21963 2013-03-05 Sterling Augustine <saugustine@google.com>
21964 Cary Coutant <ccoutant@google.com>
21967 * dwarf2out.c (resolve_addr): Don't call
21968 remove_loc_list_addr_table_entries a second time for the same
21971 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21974 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
21975 (avoid_complex_debug_insns): New function.
21976 (expand_debug_locations): Call it.
21978 PR rtl-optimization/56484
21979 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
21980 lifetimes of hard registers on small register class machines.
21982 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21984 * config/microblaze/microblaze-protos.h: Rename
21985 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
21986 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
21988 (microblaze_fast_interrupt_function_p): New function.
21989 (microblaze_is_interrupt_handler): Rename to
21990 microblaze_is_interrupt_variant and add fast_interrupt check.
21991 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
21992 (save_restore_insns): Likewise.
21993 (compute_frame_size): Likewise.
21994 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
21995 (microblaze_globalize_label): Likewise.
21996 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
21997 * config/microblaze/microblaze.md: Use wrapper
21998 microblaze_is_interrupt_variant.
22000 2013-03-05 Kai Tietz <ktietz@redhat.com>
22002 * sdbout.c (sdbout_one_type): Switch to current function's section
22003 supporting cold/hot.
22005 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
22007 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
22010 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22012 PR middle-end/56461
22013 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
22014 if VALGRIND_GET_VBITS is defined, temporarily make object
22015 memory all defined, and restore previous valgrind addressability
22016 and definability afterwards. Free this_object at the end.
22018 PR middle-end/56461
22019 * lra.c (lra): Call lra_clear_live_ranges if live_p,
22020 right before calling lra_create_live_ranges, also call it
22021 when clearing live_p. Only call lra_clear_live_ranges
22022 at the end if live_p.
22024 PR middle-end/56461
22025 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
22027 2013-03-05 Richard Biener <rguenther@suse.de>
22029 PR tree-optimization/56521
22030 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
22033 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
22036 * except.h (remove_unreachable_eh_regions): New prototype.
22037 * except.c (remove_eh_handler_splicer): New function, split out
22038 of remove_eh_handler.
22039 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
22040 warning about running it on many EH regions one at a time.
22041 (remove_unreachable_eh_regions_worker): New function, walk the
22042 EH tree in depth-first order and remove non-marked regions.
22043 (remove_unreachable_eh_regions): New function.
22044 * tree-eh.c (mark_reachable_handlers): New function, split out
22045 from remove_unreachable_handlers.
22046 (remove_unreachable_handlers): Use mark_reachable_handlers and
22047 remove_unreachable_eh_regions.
22048 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
22049 and remove_unreachable_eh_regions.
22051 2013-03-05 Richard Biener <rguenther@suse.de>
22053 PR middle-end/56525
22054 * loop-init.c (fix_loop_structure): Remove loops in two stages,
22055 not freeing them until the end.
22057 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22059 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
22061 2013-03-05 Richard Biener <rguenther@suse.de>
22063 PR tree-optimization/56270
22064 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
22065 of loads after scheduling an SLP instance.
22067 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22069 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
22071 (check_gcc_parallelize): Run guality.exp as a separate job from
22072 vect.exp with unsorted.exp and $(dg_target_exps) separately from
22073 struct-layout-1.exp with stackalign.exp.
22075 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
22077 PR middle-end/56461
22078 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
22079 load_index sbitmap even if some bit in it isn't set.
22081 PR middle-end/56461
22082 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
22083 (discover_iteration_bound_by_body_walk): Change queues to
22084 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
22085 spelling in comment. Call safe_push on queues[bound_index] directly.
22086 Release queues[queue_index] in every iteration unconditionally.
22087 Release bounds vector.
22089 PR middle-end/56461
22090 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
22091 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
22092 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
22095 2013-03-05 Richard Biener <rguenther@suse.de>
22098 * tree-inline.c (remap_blocks_to_null): New function.
22099 (expand_call_inline): When expanding a call stmt without
22100 an associated block inline remap all callee blocks to NULL.
22102 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22104 PR rtl-optimization/56494
22105 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
22106 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
22107 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
22109 PR middle-end/56461
22110 * sel-sched-ir.c (free_sched_pools): Release
22111 succs_info_pool.stack[succs_info_pool.max_top] vectors too
22112 if succs_info_pool.max_top isn't -1.
22115 * opts.c (opts_obstack, opts_concat): Moved to...
22116 * opts-common.c (opts_obstack, opts_concat): ... here.
22118 2013-03-04 Jakub Jelinek <jakub@redhat.com>
22120 PR middle-end/56461
22121 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
22123 2013-03-04 Martin Jambor <mjambor@suse.cz>
22125 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
22126 all appropriate places.
22128 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
22130 PR tree-optimization/56424
22131 * ipa-split.c (split_function): Do not set the RSO flag if result is
22132 not by reference and its type is a register type.
22134 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
22136 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
22137 (microblaze_legitimate_pic_operand): Likewise
22138 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
22139 new function microblaze_legitimate_pic_operand
22140 * config/microblaze/microblaze-protos.h
22141 (microblaze_legitimate_pic_operand): Declare.
22143 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22145 * config/microblaze/predicates.md (call_insn_simple_operand):
22146 New predicate for supported rtx code types.
22147 * config/microblaze/microblaze.md (call_internal1): Use
22148 call_insn_simple_operand predicate.
22150 2013-03-04 Jakub Jelinek <jakub@redhat.com>
22152 PR middle-end/56461
22153 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
22154 partitions.ordered_remove.
22156 PR middle-end/56461
22157 * tree-vect-stmts.c (vectorizable_conversion): Don't call
22158 vec_oprnds0.create (1) for modifier == NONE.
22160 PR middle-end/56461
22161 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
22162 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
22163 vec_oprnds1 right before pushing anything to it for
22166 PR middle-end/56461
22167 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
22168 set nbbs to 0 instead of having separate code path.
22169 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
22170 instead of false as last argument if returning NULL.
22172 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
22174 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
22175 the attribute is now called "target" instead of "option".
22176 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
22177 * doc/tm.texi.in (Target Attributes): Likewise document the correct
22178 attribute/pragma name for TARGET_OPTION_VALID_P and
22179 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
22180 * doc/tm.texi: Regenerated.
22182 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
22184 * config/microblaze/microblaze.c:
22185 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
22186 * config/microblaze/microblaze.h: Add -mxl-reorder to
22188 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
22189 instructions emitted if TARGET_REORDER.
22190 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
22191 or 0 for -m/-mno case, but initialises as 2 to detect default use case
22194 2013-03-01 Xinliang David Li <davidxl@google.com>
22196 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
22199 2013-03-01 Jakub Jelinek <jakub@redhat.com>
22201 PR middle-end/56461
22202 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
22203 vector even when returning true. Fix up function comment formatting.
22205 PR middle-end/56461
22206 * ira-build.c (ira_loop_nodes_count): New variable.
22207 (create_loop_tree_nodes): Initialize it.
22208 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
22210 PR middle-end/56461
22211 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
22212 method on dr_chain and result_chain.
22213 * tree-vect-stmts.c (vectorizable_store): Only call
22214 result_chain.create if j == 0.
22216 PR middle-end/56461
22217 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
22218 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
22219 before overwriting it.
22221 2013-03-01 Tobias Burnus <burnus@net-b.de>
22223 * doc/extended.texi (C Extensions): Change order in @menu
22225 (Other MIPS Built-in Functions): Move last MIPS entry before
22226 "picoChip Built-in Functions".
22227 (SH Built-in Functions): Move after RX Built-in Functions.
22228 * doc/gcc.texi (Introduction): Change order in @menu
22230 * doc/md.texi (Constraints): Ditto.
22231 * gty.texi (Type Information): Ditto.
22232 (User-provided marking routines for template types): Make
22234 * doc/invoke.texi (AArch64 Options): Move before
22235 "Adapteva Epiphany Options".
22237 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
22238 Jakub Jelinek <jakub@redhat.com>
22241 * asan.c (gate_asan): Lookup no_sanitize_address instead of
22242 no_address_safety_analysis attribute.
22243 * doc/extend.texi (no_address_safety_attribute): Rename to
22244 no_sanitize_address attribute, mention no_address_safety_analysis
22245 attribute as deprecated alias.
22247 2013-02-28 Jakub Jelinek <jakub@redhat.com>
22249 PR middle-end/56461
22250 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
22251 type to vec<vec<tree> > *.
22252 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
22253 to be vec<tree> instead of vec<tree> *, set vec_defs
22254 to vNULL and call vec_defs.create (number_of_vects), adjust other
22256 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
22257 vectorizable_condition): Adjust vect_get_slp_defs callers.
22259 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22261 * config/aarch64/aarch64.c
22262 (aarch64_float_const_representable): Remove unused variable.
22264 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22266 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
22268 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22270 * config/aarch64/aarch64-builtins.c
22271 (aarch64_init_simd_builtins): Make static.
22273 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
22275 * config/aarch64/aarch64.c
22276 (aarch64_simd_make_constant): Make static.
22278 2013-02-28 Martin Jambor <mjambor@suse.cz>
22280 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
22281 with no initialization to the RHS of debug statements.
22283 2013-02-28 Martin Jambor <mjambor@suse.cz>
22285 PR tree-optimization/56294
22286 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
22288 (get_access_replacement): Do not call create_access_replacement.
22289 Assert a replacement exists.
22290 (get_repl_default_def_ssa_name): Create the replacement declaration
22293 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22295 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
22296 final_end_function.
22298 2013-02-28 Marek Polacek <polacek@redhat.com>
22300 PR rtl-optimization/56466
22301 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
22302 if we're changing a loop.
22303 (peel_loops_completely): Likewise.
22305 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
22308 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
22310 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
22313 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
22314 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
22315 INTX_FTYPE_FX, FX_FTYPE_INTX.
22316 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
22318 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
22320 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
22321 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
22322 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
22323 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
22324 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
22325 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
22326 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
22327 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
22328 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
22329 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
22330 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
22331 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
22332 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
22333 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
22334 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
22335 (avrxmega6): Increase max flash segments from 5 to 6.
22336 * config/avr/t-multilib: Regenerate.
22337 * config/avr/avr-tables.opt: Regenerate.
22338 * doc/avr-mmcu.texi: Regenerate.
22340 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
22342 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
22343 (avr_device_to_arch): Rename to avr_device_to_ld.
22344 (avr_device_to_as): New prototype.
22345 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
22346 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
22347 * config/avr/driver-avr.c (avr_device_to_as): New.
22348 (avr_device_to_arch): Rename to avr_device_to_ld.
22350 2013-02-27 Jakub Jelinek <jakub@redhat.com>
22352 PR middle-end/56461
22353 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
22354 method on dr_chain and result_chain.
22356 PR middle-end/56461
22357 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
22358 pointer_set_destroy on not_executed_last_iteration.
22360 PR middle-end/56461
22361 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
22363 PR middle-end/56461
22364 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
22365 FOR_EACH_DEFINED_FUNCTION when freeing state.
22367 PR middle-end/56461
22368 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
22370 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
22373 PR middle-end/56461
22374 * ipa-cp.c (decide_whether_version_node): Call vec_free on
22375 known_aggs[i].items and release known_aggs vector.
22377 PR middle-end/56461
22378 * ipa-reference.c (propagate): Free node_info even for alias nodes.
22380 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22382 * config/microblaze/microblaze.c (microblaze_emit_compare):
22383 Use xor for EQ/NE comparisions.
22384 * config/microblaze/microblaze.md (cstoresf4): Add constraints
22385 (cbranchsf4): Adjust operator to comparison_operator.
22387 2013-02-27 Jakub Jelinek <jakub@redhat.com>
22389 PR middle-end/56461
22390 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
22392 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
22393 vec_safe_push, always update *slot.
22394 (redirect_edge_var_map_clear): Use vec_free.
22395 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
22396 (free_var_map_entry): Use vec_free.
22397 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
22398 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
22400 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
22402 PR middle-end/45472
22403 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
22404 when the may_trap_p bit of the exprs being merged differs.
22405 Reorder tests for speculativeness in the logical and operator.
22407 2013-02-27 Jakub Jelinek <jakub@redhat.com>
22409 * incpath.c (add_standard_paths): Use reconcat instead of concat
22410 where appropriate and avoid leaking memory.
22412 * opts.h: Include obstack.h.
22413 (opts_concat): New prototype.
22414 (opts_obstack): New declaration.
22415 * opts.c (opts_concat): New function.
22416 (opts_obstack): New variable.
22417 (init_options_struct): Call gcc_init_obstack on opts_obstack.
22418 (finish_options): Use opts_concat instead of concat
22419 and XOBNEWVEC instead of XNEWVEC.
22420 * opts-common.c (generate_canonical_option, decode_cmdline_option,
22421 generate_option): Likewise.
22422 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
22423 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
22426 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
22427 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
22429 2013-02-26 Jakub Jelinek <jakub@redhat.com>
22431 PR middle-end/56461
22432 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
22434 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
22436 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
22437 (arm_block_move_unaligned_straight): Likewise.
22438 (arm_adjust_block_mem): Likewise.
22440 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
22443 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
22444 temp, cond and label.
22445 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
22448 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
22449 * config/c6x/c6x.h (dbx_register_map): Update declaration.
22452 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
22453 of prologue/epilogue functions.
22456 * config/tilegx/tilegx.c (tilegx_expand_prologue):
22457 Remove unused variable cfa_offset.
22458 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
22461 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
22462 type promotion to unsigned.
22465 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
22466 for HOST_WIDE_INT of 32 bit / same size as int.
22467 (arm_block_move_unaligned_straight): Likewise.
22468 (arm_adjust_block_mem): Likewise.
22471 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
22474 2013-02-26 Marek Polacek <polacek@redhat.com>
22476 PR tree-optimization/56426
22477 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
22479 2013-02-26 Richard Biener <rguenther@suse.de>
22482 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
22483 unused variable loops.
22485 2013-02-26 Jakub Jelinek <jakub@redhat.com>
22487 PR tree-optimization/56448
22488 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
22489 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
22490 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
22491 later operands of the references, or even first operand for
22492 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
22494 PR tree-optimization/56443
22495 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
22496 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
22497 to type_for_mode langhook.
22499 2013-02-25 Matt Turner <mattst88@gmail.com>
22501 * doc/invoke.texi: Document r4700.
22503 2013-02-25 Richard Biener <rguenther@suse.de>
22505 PR tree-optimization/56175
22506 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
22508 (simplify_bitwise_binary): ... here. Also guard the conversion
22509 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
22511 2013-02-25 Catherine Moore <clm@codesourcery.com>
22514 2013-02-24 Catherine Moore <clm@codesourcery.com>
22515 Maciej W. Rozycki <macro@codesourcery.com>
22516 Tom de Vries <tom@codesourcery.com>
22517 Nathan Sidwell <nathan@codesourcery.com>
22518 Iain Sandoe <iain@codesourcery.com>
22519 Nathan Froyd <froydnj@codesourcery.com>
22520 Chao-ying Fu <fu@mips.com>
22522 * doc/extend.texi: (micromips, nomicromips, nocompression):
22523 Document new function attributes.
22524 * doc/invoke.texi (minterlink-compressed, mmicromips,
22525 m14k, m14ke, m14kec): Document new options.
22526 (minterlink-mips16): Update documentation.
22527 * doc/md.texi (ZC, ZD): Document new constraints.
22528 * configure.ac (gcc_cv_as_micromips): Check if linker
22529 supports the .set micromips directive.
22530 * configure: Regenerate.
22531 * config.in: Regenerate.
22532 * config/mips/mips-tables.opt: Regenerate.
22533 * config/mips/micromips.md: New file.
22534 * constraints.md (ZC, AD): New constraints.
22535 * config/mips/predicates.md (movep_src_register): New predicate.
22536 (movep_src_operand): New predicate.
22537 (non_volatile_mem_operand): New predicate.
22538 * config/mips/mips.md (multimem): New type.
22539 (length): Differentiate between 17-bit and 18-bit branch offsets.
22540 (MOVEP1, MOVEP2): New mode iterator.
22541 (mov_<load>l): Use ZC constraint.
22542 (mov_<load>r): Likewise.
22543 (mov_<store>l): Likewise.
22544 (mov_<store>r): Likewise.
22545 (*branch_equality<mode>_inverted): Add microMIPS support.
22546 (*branch_equality<mode>): Likewise.
22547 (*jump_absolute): Likewise.
22548 (indirect_jump_<mode>): Likewise.
22549 (tablejump_<mode>): Likewise.
22550 (<optab>_internal): Likewise.
22551 (sibcall_internal): Likewise.
22552 (sibcall_value_internal): Likewise.
22553 (prefetch): Use constraint ZD.
22554 * config/mips/mips.opt (minterlink-compressed): New option.
22555 (minterlink-mips16): Now an alias for minterlink-compressed.
22556 (mmicromips): New option.
22557 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22558 (compare_and_swap_12): Likewise.
22559 (sync_add<mode>): Likewise.
22560 (sync_<optab>_12): Likewise.
22561 (sync_old_<optab>_12): Likewise.
22562 (sync_new_<optab>_12): Likewise.
22563 (sync_nand_12): Likewise.
22564 (sync_old_nand_12): Likewise.
22565 (sync_new_nand_12): Likewise.
22566 (sync_sub<mode>): Likewise.
22567 (sync_old_add<mode>): Likewise.
22568 (sync_old_sub<mode>): Likewise.
22569 (sync_new_add<mode>): Likewise.
22570 (sync_new_sub<mode>): Likewise.
22571 (sync_<optab><mode>): Likewise.
22572 (sync_old_<optab><mode>): Likewise.
22573 (sync_new_<optab><mode>): Likewise.
22574 (sync_nand<mode>): Likewise.
22575 (sync_old_nand<mode>): Likewise.
22576 (sync_new_nand<mode>): Likewise.
22577 (sync_lock_test_and_set<mode>): Likewise.
22578 (test_and_set_12): Likewise.
22579 (atomic_compare_and_swap<mode>): Likewise.
22580 (atomic_exchange<mode>_llsc): Likewise.
22581 (atomic_fetch_add<mode>_llsc): Likewise.
22582 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22583 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22584 (umips_save_restore_pattern_p): Likewise.
22585 (umips_load_store_pair_p): Likewise.
22586 (umips_output_load_store_pair): Likewise.
22587 (umips_movep_target_p): Likewise.
22588 (umips_12bit_offset_address_p): Likewise.
22589 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22590 (mips_base_mips16): Rename this...
22591 (mips_base_compression_flags): ...to this. Update all uses.
22592 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22593 (mips_mips16_decl_p): Delete.
22594 (mips_nomips16_decl_p): Delete.
22595 (mips_get_compress_on_flags): New function.
22596 (mips_get_compress_off_flags): New function.
22597 (mips_get_compress_mode): New function.
22598 (mips_get_compress_on_name): New function.
22599 (mips_get_compress_off_name): New function.
22600 (mips_insert_attributes): Support multiple compression types.
22601 (mips_merge_decl_attributes): Likewise.
22602 (umips_12bit_offset_address_p): New function.
22603 (mips_start_function_definition): Emit .set micromips directive.
22604 (mips_call_may_need_jalx_p): New function.
22605 (mips_function_ok_for_sibcall): Add microMIPS support.
22606 (mips_print_operand_punctuation): Support short delay slots and
22608 (umips_swm_mask, umips_swm_encoding): New.
22609 (umips_build_save_restore): New function.
22610 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22611 (was_mips16_p): Remove.
22612 (old_compression_mode): New.
22613 (mips_set_compression_mode): New function.
22614 (mips_set_current_function): Add microMIPS support.
22615 (mips_option_override): Likewise.
22616 (umips_save_restore_pattern_p): New function.
22617 (umips_output_save_restore): New function.
22618 (umips_load_store_pair_p_1): New function.
22619 (umips_load_store_pair_p): New function.
22620 (umips_output_load_store_pair_1): New function.
22621 (umips_output_load_store_pair): New function.
22622 (umips_movep_target_p) New function.
22623 (mips_prepare_pch_save): Add microMIPS support.
22624 * config/mips/mips.h (TARGET_COMPRESSION): New.
22625 (TARGET_CPU_CPP_BUILTINS): Update macro
22626 to use new compression flags and to support microMIPS.
22627 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22628 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22629 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22630 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22631 (ASM_SPEC): Support mmicromips and mno-micromips.
22632 (M16STORE_REG_P): New macro.
22633 (MIPS_CALL): Support TARGET_MICROMIPS.
22634 (MICROMIPS_J): New macro.
22635 (mips_base_mips16): Rename this...
22636 (mips_base_compression_flags): ...to this.
22637 (UMIPS_12BIT_OFFSET_P): New macro.
22638 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22639 (MULTILIB_DIRNAMES): Likewise.
22641 2013-02-25 Tom de Vries <tom@codesourcery.com>
22643 PR rtl-optimization/56131
22644 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
22645 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
22646 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
22648 2013-02-25 Tobias Burnus <burnus@net-b.de>
22650 * doc/invoke.texi (-fsanitize=): Move from optimization
22651 to debugging options.
22653 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
22655 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
22657 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
22658 Alexander Monakov <amonakov@ispras.ru>
22660 PR middle-end/56077
22661 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
22662 flush pending lists also on non-jumps. Adjust comment.
22664 2013-02-24 Catherine Moore <clm@codesourcery.com>
22665 Maciej W. Rozycki <macro@codesourcery.com>
22666 Tom de Vries <tom@codesourcery.com>
22667 Nathan Sidwell <nathan@codesourcery.com>
22668 Iain Sandoe <iain@codesourcery.com>
22669 Nathan Froyd <froydnj@codesourcery.com>
22670 Chao-ying Fu <fu@mips.com>
22672 * doc/extend.texi: (micromips, nomicromips, nocompression):
22673 Document new function attributes.
22674 * doc/invoke.texi (minterlink-compressed, mmicromips,
22675 m14k, m14ke, m14kec): Document new options.
22676 (minterlink-mips16): Update documentation.
22677 * doc/md.texi (ZC, ZD): Document new constraints.
22678 * configure.ac (gcc_cv_as_micromips): Check if linker
22679 supports the .set micromips directive.
22680 * configure: Regenerate.
22681 * config.in: Regenerate.
22682 * config/mips/mips-tables.opt: Regenerate.
22683 * config/mips/micromips.md: New file.
22684 * constraints.md (ZC, AD): New constraints.
22685 * config/mips/predicates.md (movep_src_register): New predicate.
22686 (movep_src_operand): New predicate.
22687 (non_volatile_mem_operand): New predicate.
22688 * config/mips/mips.md (multimem): New type.
22689 (length): Differentiate between 17-bit and 18-bit branch offsets.
22690 (MOVEP1, MOVEP2): New mode iterator.
22691 (mov_<load>l): Use ZC constraint.
22692 (mov_<load>r): Likewise.
22693 (mov_<store>l): Likewise.
22694 (mov_<store>r): Likewise.
22695 (*branch_equality<mode>_inverted): Add microMIPS support.
22696 (*branch_equality<mode>): Likewise.
22697 (*jump_absolute): Likewise.
22698 (indirect_jump_<mode>): Likewise.
22699 (tablejump_<mode>): Likewise.
22700 (<optab>_internal): Likewise.
22701 (sibcall_internal): Likewise.
22702 (sibcall_value_internal): Likewise.
22703 (prefetch): Use constraint ZD.
22704 * config/mips/mips.opt (minterlink-compressed): New option.
22705 (minterlink-mips16): Now an alias for minterlink-compressed.
22706 (mmicromips): New option.
22707 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22708 (compare_and_swap_12): Likewise.
22709 (sync_add<mode>): Likewise.
22710 (sync_<optab>_12): Likewise.
22711 (sync_old_<optab>_12): Likewise.
22712 (sync_new_<optab>_12): Likewise.
22713 (sync_nand_12): Likewise.
22714 (sync_old_nand_12): Likewise.
22715 (sync_new_nand_12): Likewise.
22716 (sync_sub<mode>): Likewise.
22717 (sync_old_add<mode>): Likewise.
22718 (sync_old_sub<mode>): Likewise.
22719 (sync_new_add<mode>): Likewise.
22720 (sync_new_sub<mode>): Likewise.
22721 (sync_<optab><mode>): Likewise.
22722 (sync_old_<optab><mode>): Likewise.
22723 (sync_new_<optab><mode>): Likewise.
22724 (sync_nand<mode>): Likewise.
22725 (sync_old_nand<mode>): Likewise.
22726 (sync_new_nand<mode>): Likewise.
22727 (sync_lock_test_and_set<mode>): Likewise.
22728 (test_and_set_12): Likewise.
22729 (atomic_compare_and_swap<mode>): Likewise.
22730 (atomic_exchange<mode>_llsc): Likewise.
22731 (atomic_fetch_add<mode>_llsc): Likewise.
22732 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22733 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22734 (umips_save_restore_pattern_p): Likewise.
22735 (umips_load_store_pair_p): Likewise.
22736 (umips_output_load_store_pair): Likewise.
22737 (umips_movep_target_p): Likewise.
22738 (umips_12bit_offset_address_p): Likewise.
22739 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22740 (mips_base_mips16): Rename this...
22741 (mips_base_compression_flags): ...to this. Update all uses.
22742 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22743 (mips_mips16_decl_p): Delete.
22744 (mips_nomips16_decl_p): Delete.
22745 (mips_get_compress_on_flags): New function.
22746 (mips_get_compress_off_flags): New function.
22747 (mips_get_compress_mode): New function.
22748 (mips_get_compress_on_name): New function.
22749 (mips_get_compress_off_name): New function.
22750 (mips_insert_attributes): Support multiple compression types.
22751 (mips_merge_decl_attributes): Likewise.
22752 (umips_12bit_offset_address_p): New function.
22753 (mips_start_function_definition): Emit .set micromips directive.
22754 (mips_call_may_need_jalx_p): New function.
22755 (mips_function_ok_for_sibcall): Add microMIPS support.
22756 (mips_print_operand_punctuation): Support short delay slots and
22758 (umips_swm_mask, umips_swm_encoding): New.
22759 (umips_build_save_restore): New function.
22760 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22761 (was_mips16_p): Remove.
22762 (old_compression_mode): New.
22763 (mips_set_compression_mode): New function.
22764 (mips_set_current_function): Add microMIPS support.
22765 (mips_option_override): Likewise.
22766 (umips_save_restore_pattern_p): New function.
22767 (umips_output_save_restore): New function.
22768 (umips_load_store_pair_p_1): New function.
22769 (umips_load_store_pair_p): New function.
22770 (umips_output_load_store_pair_1): New function.
22771 (umips_output_load_store_pair): New function.
22772 (umips_movep_target_p) New function.
22773 (mips_prepare_pch_save): Add microMIPS support.
22774 * config/mips/mips.h (TARGET_COMPRESSION): New.
22775 (TARGET_CPU_CPP_BUILTINS): Update macro
22776 to use new compression flags and to support microMIPS.
22777 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22778 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22779 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22780 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22781 (ASM_SPEC): Support mmicromips and mno-micromips.
22782 (M16STORE_REG_P): New macro.
22783 (MIPS_CALL): Support TARGET_MICROMIPS.
22784 (MICROMIPS_J): New macro.
22785 (mips_base_mips16): Rename this...
22786 (mips_base_compression_flags): ...to this.
22787 (UMIPS_12BIT_OFFSET_P): New macro.
22788 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22789 (MULTILIB_DIRNAMES): Likewise.
22791 2013-02-24 Jakub Jelinek <jakub@redhat.com>
22794 * target-globals.c (save_target_globals): For init_reg_sets and
22795 target_reinit remporarily set this_fn_optabs to this_target_optabs.
22797 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
22799 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
22800 * config/aarch64/t-aarch64
22801 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
22803 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
22805 PR inline-asm/56148
22806 * lra-constraints.c (process_alt_operands): Reload operand
22807 conflicting with earlier clobber only if no more other conflicting
22810 2013-02-22 Jakub Jelinek <jakub@redhat.com>
22813 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
22814 if not linking a shared library.
22816 2013-02-22 Seth LaForge <sethml@google.com>
22818 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
22820 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
22822 * config/arm/arm.md (split for extendsidi): Update condition.
22823 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
22824 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
22825 (qhs_zextenddi_cstr): Likewise.
22827 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22829 PR middle-end/56420
22830 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
22831 avoid signed wrapping.
22832 (expand_mult): Handle properly multiplication by
22833 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
22834 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
22835 in the compiler if coeff is HOST_WIDE_INT_MIN.
22836 (expand_divmod): Don't make ext_op1 static, change it's type to
22837 uhwi. Avoid undefined behavior in -INTVAL (op1).
22839 PR rtl-optimization/50339
22840 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
22842 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
22843 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
22844 into splitting_ashiftrt field.
22845 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
22847 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
22850 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
22852 PR middle-end/56108
22853 * trans-mem.c (execute_tm_mark): Do not expand transactions that
22854 are sure to go irrevocable.
22856 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
22858 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
22859 scalars are valid operands.
22861 2013-02-21 Martin Jambor <mjambor@suse.cz>
22863 PR tree-optimization/56310
22864 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
22865 only matching indices and non-negative final offsets.
22866 (intersect_aggregates_with_edge): Pass src_idx to
22867 agg_replacements_to_vector. Pass src_idx insstead of index to
22868 intersect_with_agg_replacements.
22870 2013-02-21 Martin Jambor <mjambor@suse.cz>
22872 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
22873 instead of hard-wired defaults.
22875 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
22877 * doc/invoke.texi (MIPS Options): Update documentation of the
22878 floating-point multiply-accumulate instruction restrictions.
22880 2013-02-21 Kostya Serebryany <kcc@google.com>
22882 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
22883 asan_shadow_offset on x86_64 linux.
22885 2013-02-21 Richard Biener <rguenther@suse.de>
22887 PR tree-optimization/56415
22889 2013-02-11 Richard Biener <rguenther@suse.de>
22891 PR tree-optimization/56273
22892 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22895 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22898 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
22901 PR inline-asm/56405
22902 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
22903 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
22905 2013-02-20 Jan Hubicka <jh@suse.cz>
22907 PR tree-optimization/56265
22908 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
22909 when target is referenced for first time.
22911 2013-02-20 Richard Biener <rguenther@suse.de>
22913 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
22914 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
22915 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
22916 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
22917 not return anything.
22918 (rename_ssa_copies): Do not remove unused locals.
22919 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
22920 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
22921 * passes.c (execute_function_todo): Do not schedule unused locals
22922 removal if cleanup_tree_cfg did something.
22923 * tree-ssa-live.c (remove_unused_locals): Dump statistics
22924 about the number of removed locals.
22926 2013-02-20 Richard Biener <rguenther@suse.de>
22928 PR tree-optimization/56398
22929 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
22931 2013-02-20 Martin Jambor <mjambor@suse.cz>
22933 PR tree-optimization/55334
22934 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
22935 restricted pointers to arrays.
22937 2013-02-20 Richard Biener <rguenther@suse.de>
22938 Jakub Jelinek <jakub@redhat.com>
22940 PR tree-optimization/56396
22941 * tree-ssa-ccp.c (n_const_val): New static variable.
22942 (get_value): Return NULL for SSA names we don't have a lattice
22944 (ccp_initialize): Initialize n_const_val.
22945 * tree-ssa-copy.c (n_copy_of): New static variable.
22946 (init_copy_prop): Initialize n_copy_of.
22947 (get_value): Return NULL_TREE for SSA names we don't have a
22950 2013-02-20 Martin Jambor <mjambor@suse.cz>
22952 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
22954 2013-02-20 Richard Biener <rguenther@suse.de>
22956 * genpreds.c (write_lookup_constraint): Do not compare first
22957 letter of the constraint again.
22959 2013-02-20 Richard Biener <rguenther@suse.de>
22961 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
22963 (get_use_iv_cost): Terminate hashtable walk when coming across
22966 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
22968 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
22969 reassociation for avx2 targets.
22971 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22973 * config/microblaze/microblaze.c: microblaze_has_clz = 0
22974 Add version check for v8.10.a to enable microblaze_has_clz
22975 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
22976 version and TARGET_PATTERN_COMPARE check
22977 * config/microblaze/microblaze.md: New clzsi2 instruction
22979 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22981 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
22982 function before branching.
22984 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22986 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
22988 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
22990 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22992 PR middle-end/55889
22993 * sel-sched.c: Include ira.h.
22994 (implicit_clobber_conflict_p): New function.
22995 (moveup_expr): Use it.
22996 * Makefile.in (sel-sched.o): Depend on ira.h.
22998 2013-02-19 Richard Biener <rguenther@suse.de>
23000 PR tree-optimization/56384
23001 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
23002 (vn_hash_type): Split out from ...
23003 (vn_hash_constant_with_type): ... here.
23004 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
23005 (vn_phi_eq): Compare types from vn_phi_s structure.
23006 (vn_phi_lookup): Populate vn_phi_s type.
23007 (vn_phi_insert): Likewise.
23009 2013-02-19 Jakub Jelinek <jakub@redhat.com>
23011 PR tree-optimization/56350
23012 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
23013 if haven't found reduction or nested cycle operand, rather than
23014 asserting we must find it.
23016 PR tree-optimization/56381
23017 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
23020 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
23021 Jakub Jelinek <jakub@redhat.com>
23024 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
23025 (swap_optab_enable): Same.
23026 (init_all_optabs): Use argument instead of global.
23027 * tree.h (struct tree_optimization_option): New field target_optabs.
23028 * expr.h (init_all_optabs): Add argument to prototype.
23029 (TREE_OPTIMIZATION_OPTABS): New.
23030 (save_optabs_if_changed): Protoize.
23031 * optabs.h: Declare this_fn_optabs.
23032 * optabs.c (save_optabs_if_changed): New.
23033 Declare this_fn_optabs.
23034 (init_optabs): Add argument to init_all_optabs() call.
23035 * function.c (invoke_set_current_function_hook): Handle per
23037 * function.h (struct function): New field optabs.
23038 * config/mips/mips.c (mips_set_mips16_mode): Handle when
23039 optimization_current_node has changed.
23040 * target-globals.h (save_target_globals_default_opts): Protoize.
23041 * target-globals.c (save_target_globals_default_opts): New.
23043 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23046 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
23047 registers %fr12 and %fr12R as call used.
23050 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
23051 and HImode, require all displacements to be an integer multiple of
23053 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
23054 only allow QImode and HImode when reload is in progress and strict is
23055 true. Likewise for symbolic addresses. Use base14_operand to check
23056 displacements in REG+BASE addresses.
23058 2013-02-18 Richard Biener <rguenther@suse.de>
23060 PR tree-optimization/56366
23061 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23062 handle sign-conversion of outer-loop initial induction value.
23064 2013-02-18 Richard Biener <rguenther@suse.de>
23066 PR middle-end/56349
23067 * cfghooks.c (merge_blocks): If we merge a latch into another
23068 block adjust references to it.
23069 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
23070 (verify_loop_structure): Verify that a recorded latch is in fact
23073 2013-02-18 Richard Biener <rguenther@suse.de>
23075 PR tree-optimization/56321
23076 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
23077 order SSA name release and virtual operand unlinking.
23079 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23081 * config/microblaze/microblaze.md (save_stack_block): Define.
23082 (restore_stack_block): Likewise.
23084 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23086 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
23087 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
23088 * config/microblaze/microblaze.c (microblaze_option_override):
23089 Bail out early for PIC modes when target does not support PIC.
23091 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23093 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
23094 Replace with a microblaze version.
23095 (microblaze_trampoline_init): Adapt for microblaze.
23096 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
23099 2013-02-16 Jakub Jelinek <jakub@redhat.com>
23100 Dodji Seketeli <dodji@redhat.com>
23103 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
23104 (instrument_mem_region_access): Do not forget to always put
23105 instrumentation of the of 'base' and 'base + len' in a "if (len !=
23106 0) statement, even for cases where either 'base' or 'base + len'
23107 are not instrumented -- because they have been previously
23108 instrumented. Simplify the logic by putting all the statements
23109 instrument 'base + len' inside a sequence, and then insert that
23110 sequence right before the current insertion point. Then, to
23111 instrument 'base + len', just get an iterator on that statement.
23112 And do not forget to update the pointer to iterator the function
23113 received as argument.
23115 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
23117 PR rtl-optimization/56348
23118 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
23120 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
23122 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
23123 (clean_graph_dump_file): Pass base to start_graph_dump.
23125 2013-02-14 Richard Henderson <rth@redhat.com>
23128 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
23130 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
23132 * collect2-aix.h: Define F_LOADONLY.
23134 2013-02-14 Richard Biener <rguenther@suse.de>
23137 * varasm.c (output_constant_def_1): Get the decl representing
23138 the constant as argument.
23139 (output_constant_def): Wrap output_constant_def_1.
23140 (make_decl_rtl): Use output_constant_def_1 with the decl
23141 representing the constant.
23142 (build_constant_desc): Optionally re-use a decl already
23143 representing the constant.
23144 (tree_output_constant_def): Adjust.
23146 2013-02-14 Dodji Seketeli <dodji@redhat.com>
23149 * asan.c (instrument_builtin_call): Really put the length of the
23150 second source argument into src1_len.
23152 2013-02-13 Jakub Jelinek <jakub@redhat.com>
23154 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
23155 argument. If it is false, don't create edge from then_bb to
23157 (insert_if_then_before_iter): Pass true to it.
23158 (build_check_stmt): Pass false to it.
23159 (transform_statements): Flush hash table only on extended basic
23160 block boundaries, rather than at the beginning of every bb.
23161 Don't flush hash table on nonfreeing_call_p calls.
23162 * tree-flow.h (nonfreeing_call_p): New prototype.
23163 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
23165 2013-02-13 David S. Miller <davem@davemloft.net>
23167 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
23169 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
23172 * ira.c (max_regno_before_ira): Move from ...
23174 (fix_reg_equiv_init): Use max_regno_before_ira instead of
23177 2013-02-13 Jakub Jelinek <jakub@redhat.com>
23179 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
23181 2013-02-13 Richard Biener <rguenther@suse.de>
23184 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
23185 globals in MEM_REFs.
23187 2013-02-13 Richard Biener <rguenther@suse.de>
23189 * loop-init.c (loop_optimizer_init): Clear loop state when
23190 re-initializing preserved loops.
23191 * loop-unswitch.c (unswitch_single_loop): Return whether
23192 we unswitched the loop. Do not verify loop state here.
23193 (unswitch_loops): When we unswitched a loop discover new loops.
23195 2013-02-13 Kostya Serebryany <kcc@google.com>
23197 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
23199 * sanitizer.def: Rename __asan_init to __asan_init_v1.
23201 2013-02-12 Dodji Seketeli <dodji@redhat.com>
23203 Avoid instrumenting duplicated memory access in the same basic block
23204 * Makefile.in (asan.o): Add new dependency on hash-table.h
23205 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
23206 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
23207 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
23208 (free_mem_ref_resources, has_mem_ref_been_instrumented)
23209 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
23210 (get_mem_ref_of_assignment): New functions.
23211 (get_mem_refs_of_builtin_call): Extract from
23212 instrument_builtin_call and tweak a little bit to make it fit with
23214 (instrument_builtin_call): Use the new
23215 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
23216 of is_gimple_builtin_call.
23217 (instrument_derefs, instrument_mem_region_access): Insert the
23218 instrumented memory reference into the hash table.
23219 (maybe_instrument_assignment): Renamed instrument_assignment into
23220 this, and change it to advance the iterator when instrumentation
23221 actually happened and return true in that case. This makes it
23222 homogeneous with maybe_instrument_assignment, and thus give a
23223 chance to callers to be more 'regular'.
23224 (transform_statements): Clear the memory reference hash table
23225 whenever we enter a new BB, when we cross a function call, or when
23226 we are done transforming statements. Use
23227 maybe_instrument_assignment instead of instrumentation. No more
23228 need to special case maybe_instrument_assignment and advance the
23229 iterator after calling it; it's now handled just like
23230 maybe_instrument_call. Update comment.
23232 2013-02-13 Richard Biener <rguenther@suse.de>
23234 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
23235 Fix loop discovery code.
23237 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
23239 PR inline-asm/56148
23240 * lra-constraints.c (process_alt_operands): Match early clobber
23241 operand with itself. Check conflicts with earlyclobber only if
23242 the operand is not reloaded. Prefer to reload conflicting operand
23243 if earlyclobber and matching operands are the same.
23245 2013-02-12 Richard Biener <rguenther@suse.de>
23248 * lto-streamer-out.c (write_symbol): Do not output symbols
23249 for hard register variables.
23251 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
23254 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
23255 (umulsidi3_insn, mulsidi3_insn): New insns.
23257 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
23259 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
23260 (struct tune_params): Add vec_costs field.
23261 * config/arm/arm.c (arm_builtin_vectorization_cost)
23262 (arm_add_stmt_cost): New functions.
23263 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
23264 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
23265 (arm_default_vec_cost): New struct of type cpu_vec_costs.
23266 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
23267 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
23268 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
23269 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
23271 2013-02-12 Richard Biener <rguenther@suse.de>
23274 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
23275 decls again if possible.
23277 2013-02-12 Richard Biener <rguenther@suse.de>
23279 PR middle-end/56288
23280 * tree-ssa.c (verify_ssa_name): Fix check, move
23281 SSA_NAME_IN_FREE_LIST check up.
23283 2013-02-12 Jakub Jelinek <jakub@redhat.com>
23284 Steven Bosscher <steven@gcc.gnu.org>
23286 PR rtl-optimization/56151
23287 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
23288 equal to op0 or op1, and last_insn pattern is CODE operation
23289 with MEM dest and one of the operands matches that MEM.
23291 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
23293 * doc/extend.texi: Document Function Multiversioning and "default"
23294 parameter string to target attribute.
23295 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
23296 target attribute parameter is "default".
23297 (ix86_compare_version_priority): Remove checks for target attribute.
23298 (ix86_mangle_function_version_assembler_name): Change error to sorry.
23299 Remove check for target attribute equal to NULL. Add assert.
23300 (ix86_generate_version_dispatcher_body): Change error to sorry.
23302 2013-02-11 Iain Sandoe <iain@codesourcery.com>
23303 Jack Howarth <howarth@bromo.med.uc.edu>
23304 Patrick Marlier <patrick.marlier@gmail.com>
23307 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
23308 define ENDFILE_SPEC as TM_DESTRUCTOR.
23309 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
23311 2013-02-11 Alexander Potapenko <glider@google.com>
23312 Jack Howarth <howarth@bromo.med.uc.edu>
23313 Jakub Jelinek <jakub@redhat.com>
23316 * config/darwin.c (cdtor_record): Rename ctor_record.
23317 (sort_cdtor_records): Rename sort_ctor_records.
23318 (finalize_dtors): New routine to sort destructors by
23319 priority before use in assemble_integer.
23320 (machopic_asm_out_destructor): Use finalize_dtors if needed.
23322 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
23324 PR rtl-optimization/56275
23325 * simplify-rtx.c (avoid_constant_pool_reference): Check that
23326 offset is non-negative and less than cmode size before
23327 calling simplify_subreg.
23329 2013-02-11 Richard Biener <rguenther@suse.de>
23331 PR tree-optimization/56264
23332 * cfgloop.h (fix_loop_structure): Adjust prototype.
23333 * loop-init.c (fix_loop_structure): Return the number of
23334 newly discovered loops.
23335 * tree-cfgcleanup.c (repair_loop_structures): When new loops
23336 are discovered, do a full loop-closed SSA rewrite.
23338 2013-02-11 Richard Biener <rguenther@suse.de>
23340 PR tree-optimization/56273
23341 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
23343 (check_array_ref): Fix missing newline in dumps.
23344 (search_for_addr_array): Likewise.
23346 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
23348 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
23350 2013-02-09 Jakub Jelinek <jakub@redhat.com>
23353 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
23355 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
23357 PR rtl-optimization/56246
23358 * lra-constraints.c (simplify_operand_subreg): Try to reuse
23360 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
23361 constraints are satisfied.
23363 2013-02-08 Jeff Law <law@redhat.com>
23366 * emit-rtl.c (reg_is_parm_p): New function.
23367 * regs.h (reg_is_parm_p): New prototype.
23368 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
23369 callee-clobbered registers.
23371 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
23374 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
23375 If there is no implicit builtin declaration, just return NULL.
23377 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
23379 * config/i386/sse.md (FMAMODEM): New mode iterator.
23380 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
23381 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
23383 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
23385 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
23386 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
23387 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
23389 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23391 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
23392 (microblaze*-*-elf): Likewise.
23393 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
23395 * config/microblaze/microblaze-c.c: Add builtin defines for
23396 _LITTLE_ENDIAN and _BIG_ENDIAN.
23397 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
23398 add to TARGET_DEFAULT flags.
23399 Expand ASM_SPEC and LINK_SPEC.
23400 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
23401 * config/microblaze/microblaze.md: Update extendsidi2 and
23402 movdi_internal instructions to use low-order / high-order reg
23404 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
23405 options and inversemask / mask of LITTLE_ENDIAN.
23406 * config/microblaze/t-microblaze: Expand multilib options to
23407 include mlittle-endian (le) and update exceptions patterns.
23409 2013-02-08 Jakub Jelinek <jakub@redhat.com>
23411 PR rtl-optimization/56195
23412 * lra-constraints.c (get_reload_reg): Don't reuse regs
23413 if they have smaller mode than requested, if they have
23414 wider mode than requested, try to return a SUBREG.
23416 PR tree-optimization/56250
23417 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
23418 if type is unsigned and code isn't MULT_EXPR.
23420 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
23422 PR tree-optimization/56064
23423 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
23424 bits according to mode.
23425 * fixed-value.h (fixed_from_double_int)
23426 (const_fixed_from_double_int): Adjust comments.
23428 2013-02-08 Richard Biener <rguenther@suse.de>
23431 * lto-streamer.h (struct data_in): Remove current_file, current_line
23432 and current_col members.
23433 * lto-streamer-out.c (lto_output_location): Stream changed bits
23434 en-block for efficiency.
23435 * lto-streamer-in.c (clear_line_info): Remove.
23436 (lto_input_location): Cache current file, line and column
23437 globally via local statics. Read changed bits en-block.
23438 (input_function): Do not call clear_line_info.
23439 (lto_read_body): Likewise.
23440 (lto_input_toplevel_asms): Likewise.
23442 2013-02-08 Michael Matz <matz@suse.de>
23444 PR tree-optimization/52448
23445 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
23446 (nt_call_phase): New static.
23447 (add_or_mark_expr): Only mark accesses with newer phase than any
23449 (nonfreeing_call_p): New.
23450 (nt_init_block): Update nt_call_phase, mark blocks as visited.
23451 (nt_fini_block): Keep blocks marked as visited.
23452 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
23454 2013-02-08 Richard Biener <rguenther@suse.de>
23456 * ira.c (ira): Free broken dominator information.
23458 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
23460 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
23462 2013-02-08 Marek Polacek <polacek@redhat.com>
23464 * cfgloop.c (verify_loop_structure): Add more checking of headers.
23466 2013-02-08 Richard Biener <rguenther@suse.de>
23468 PR middle-end/56181
23469 * cfgloop.h (flow_loops_find): Adjust.
23470 (bb_loop_header_p): Declare.
23471 * cfgloop.c (bb_loop_header_p): New function split out from ...
23472 (flow_loops_find): ... here. Adjust function signature,
23473 support incremental loop structure update.
23474 (verify_loop_structure): Cleanup. Verify a loop is a loop.
23475 * cfgloopmanip.c (fix_loop_structure): Move ...
23476 * loop-init.c (fix_loop_structure): ... here.
23477 (apply_loop_flags): Split out from ...
23478 (loop_optimizer_init): ... here.
23479 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
23480 in incremental mode, only remove dead loops here.
23482 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
23485 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
23486 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
23487 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
23488 (*round<mode>3.libgcc): New insns for fixed-modes.
23489 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
23490 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
23491 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
23492 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
23493 implementations. Define to __builtin_avr_absFX,
23494 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
23495 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
23496 __builtin_avr_countlsFX, respectively.
23497 * config/avr/avr-c.c (target.h): Include it.
23498 (enum avr_builtin_id): New enum.
23499 (avr_resolve_overloaded_builtin): New static function.
23500 (avr_register_target_pragmas): Use it to set
23501 targetm.resolve_overloaded_builtin.
23502 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
23503 tree nodes used by DEF_BUILTIN.
23504 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
23505 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
23506 <AVR_BUILTIN_xxBITS>: Same.
23508 2013-02-08 Richard Biener <rguenther@suse.de>
23510 * cfgloop.c (verify_loop_structure): Properly handle
23511 a loop exiting to another loop header.
23512 * ira-int.h (ira_loops): Remove.
23513 * ira.c (ira_loops): Remove.
23514 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
23515 (do_reload): Use loop_optimizer_finalize.
23516 * ira-build.c (create_loop_tree_nodes): Use get_loops and
23517 number_of_loops to access the loop tree.
23518 (more_one_region_p): Likewise.
23519 (finish_loop_tree_nodes): Likewise.
23520 (rebuild_regno_allocno_maps): Likewise.
23521 (mark_loops_for_removal): Likewise.
23522 (mark_all_loops_for_removal): Likewise.
23523 (remove_unnecessary_regions): Likewise.
23524 (ira_build): Likewise.
23525 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
23527 2013-02-08 Richard Biener <rguenther@suse.de>
23529 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
23530 * ipa-pure-const.c (analyze_function): Avoid calling
23531 mark_irreducible_loops twice.
23532 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
23534 2013-02-07 David S. Miller <davem@davemloft.net>
23536 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
23538 * var-tracking.c (vt_add_function_parameter): Test the presence of
23539 HAVE_window_save properly and do not remap argument registers when
23540 we have a leaf function.
23542 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
23545 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
23547 * config/i386/i386.c (ix86_print_operand): Ditto.
23549 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
23551 * lra-constraints.c (process_alt_operands): Fix recently added comment.
23553 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
23555 PR rtl-optimization/56225
23556 * lra-constraints.c (process_alt_operands): Check that reload hard
23557 reg can hold value for strict_low_part.
23559 2013-02-07 Jakub Jelinek <jakub@redhat.com>
23562 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
23563 dwarf2out_end_function.
23564 (in_first_function_p, maybe_at_text_label_p,
23565 first_loclabel_num_not_at_text_label): New variables.
23566 (dwarf2out_var_location): In the first function find out
23567 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
23568 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
23571 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
23573 PR rtl-optimization/56178
23574 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
23575 SUBREG of a register. Tidy up related block of code.
23576 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
23577 note if the source is a register or a SUBREG of a register.
23579 2013-02-07 Jakub Jelinek <jakub@redhat.com>
23582 * config/rs6000/rs6000.md (ptrm): New mode attr.
23583 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
23584 call_value_indirect_aix<pttrsize>,
23585 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
23588 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
23590 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
23591 if -bnortl. Convert to strcmp and strncmp.
23593 2013-02-07 Alan Modra <amodra@gmail.com>
23596 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
23597 addresses won't wrap when offsetting.
23598 (rs6000_secondary_reload): Provide secondary reloads needed for
23599 wrapping LO_SUM addresses.
23601 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
23603 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
23604 MACH, just __MACH__.
23606 2013-02-06 Richard Biener <rguenther@suse.de>
23608 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
23609 instead of calling fix_loop_structure.
23611 2013-02-06 Jakub Jelinek <jakub@redhat.com>
23613 PR middle-end/56217
23614 * omp-low.c (use_pointer_for_field): Return false if
23615 lower_send_shared_vars doesn't generate any copy-out code.
23617 2013-02-06 Tom de Vries <tom@codesourcery.com>
23619 PR rtl-optimization/56131
23620 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
23621 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
23622 of the label is NULL. Add comment.
23624 2013-02-05 Jakub Jelinek <jakub@redhat.com>
23626 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
23629 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
23630 (STATIC_LIBTSAN_LIBS): Likewise.
23631 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
23632 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
23633 is defined, don't add anything else beyond that.
23634 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
23635 (LINK_COMMAND_SPEC): Use them.
23637 PR tree-optimization/56205
23638 * tree-stdarg.c (check_all_va_list_escapes): Return true if
23639 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
23640 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
23642 2013-02-05 Richard Biener <rguenther@suse.de>
23644 PR tree-optimization/53342
23645 PR tree-optimization/53185
23646 * tree-vectorizer.h (vect_check_strided_load): Remove.
23647 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
23648 not disallow peeling for vectorized strided loads.
23649 (vect_check_strided_load): Make static and simplify.
23650 (vect_analyze_data_refs): Adjust.
23651 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
23652 correctly when vectorizing strided loads.
23654 2013-02-05 Richard Biener <rguenther@suse.de>
23656 * doc/install.texi: Refer to ISL, not PPL.
23658 2013-02-05 Jan Hubicka <jh@suse.cz>
23660 PR tree-optimization/55789
23661 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
23663 2013-02-05 Jan Hubicka <jh@suse.cz>
23665 PR tree-optimization/55789
23666 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
23667 the dead call anyway.
23669 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
23672 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
23674 2013-02-04 Alexander Potapenko <glider@google.com>
23675 Jack Howarth <howarth@bromo.med.uc.edu>
23676 Jakub Jelinek <jakub@redhat.com>
23679 * config/darwin.c (sort_ctor_records): Stabilized qsort
23680 on constructor priority by using original position.
23681 (finalize_ctors): New routine to sort constructors by
23682 priority before use in assemble_integer.
23683 (machopic_asm_out_constructor): Use finalize_ctors if needed.
23685 2013-02-04 Jakub Jelinek <jakub@redhat.com>
23688 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
23689 about visibility on artificial decls.
23690 * config/sol2.c (solaris_assemble_visibility): Likewise.
23692 2013-02-04 Kai Tietz <ktietz@redhat.com>
23695 * config/i386/i386.c (function_value_ms_64): Add additional valtype
23696 argument and improve checking of return-argument types for 16-byte
23698 (ix86_function_value_1): Add additional valtype argument on call
23699 of function_value_64.
23700 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
23701 handling infunction_value_64 function.
23703 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
23705 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
23707 2013-02-04 Richard Biener <rguenther@suse.de>
23709 PR tree-optimization/56188
23710 * tree-ssa-structalias.c (label_visit): Consider case with
23711 initially non-empty points-to set.
23712 (perform_var_substitution): Dump node mapping and clean up.
23714 2013-02-04 Richard Guenther <rguenther@suse.de>
23717 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
23718 node prevail as last resort.
23719 (lto_symtab_merge_decls): Remove guard on LTRANS here.
23720 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
23722 2013-02-04 Richard Biener <rguenther@suse.de>
23724 PR tree-optimization/56113
23725 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
23727 (equiv_class_lookup_or_add): ... this.
23728 (label_visit): Adjust and fix error in previous patch.
23729 (perform_var_substitution): Adjust.
23731 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
23733 * config/sh/divtab.c: Fix formatting and comments throughout the file.
23734 * config/sh/sh4-300.md: Likewise.
23735 * config/sh/sh4a.md: Likewise.
23736 * config/sh/constraints.md: Likewise.
23737 * config/sh/sh.md: Likewise.
23738 * config/sh/netbsd-elf.h: Likewise.
23739 * config/sh/predicates.md: Likewise.
23740 * config/sh/sh-protos.h: Likewise.
23741 * config/sh/ushmedia.h: Likewise.
23742 * config/sh/linux.h: Likewise.
23743 * config/sh/sh.c: Likewise.
23744 * config/sh/superh.h: Likewise.
23745 * config/sh/elf.h: Likewise.
23746 * config/sh/sh4.md: Likewise.
23747 * config/sh/sh.h: Likewise.
23749 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23751 * config/pa/constraints.md: Adjust unused letters. Change "T"
23752 constraint to match_test floating_point_store_memory_operand().
23753 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
23754 (base14_operand): New.
23755 (floating_point_store_memory_operand): New.
23756 (integer_store_memory_operand): Revise to use base14_operand and
23757 reg_plus_base_memory_operand.
23758 (move_dest_operand): Allow symbolic_memory_operands.
23759 (symbolic_memory_operand): Check for LO_SOM.
23760 (symbolic_operand): Change default case to break.
23761 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
23762 CONST_DOUBLE values to be reloaded by putting them into memory when
23763 the destination is a floating point register.
23764 (movdf): Remove code to handle CONST_DOUBLE.
23766 (reload_indf_r1): New.
23767 (reload_insf_r1): New.
23768 Consistently use "Q" and "T" constraints with integer and floating
23769 point move instructions, respectively.
23770 (movdi): Remove FAIL.
23771 Change predicate for source operand unamed DImode move from
23772 general_operand to move_src_operand.
23773 (umulsidi3): Change predicate for destination operand to
23775 Likewise for similar unamed patterns.
23776 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
23777 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
23778 (hppa_legitimize_address): Simplify mask calculation.
23779 (pa_emit_move_sequence): Revised handling of secondary reloads from
23780 REG+D addresses for floating point loads and stores. Directly handle
23781 loading CONST0_RTX (mode) to a floating point register.
23782 (pa_secondary_reload): Handle reloading DF and SFmode constant values
23783 to floating point registers. Don't restrict secondary reloads to
23784 floating point registers to integer modes. Revise some comments and
23786 (TARGET_LEGITIMATE_ADDRESS_P): Define.
23787 (pa_legitimate_address_p): New.
23788 (pa_legitimize_reload_address): New.
23789 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
23790 (STRICT_REG_OK_FOR_BASE_P): New.
23791 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
23792 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
23794 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
23795 Andrew Dixie <andrewd@gentrack.com>
23797 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
23800 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
23802 * expmed.c (extract_bit_field_1): Pass the full width of the
23803 structure to get_best_reg_extraction_insn.
23805 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
23808 * configure.ac (use_cxa_atexit): Add AIX.
23809 * configure: Regenerate.
23811 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
23813 2013-02-01 Jakub Jelinek <jakub@redhat.com>
23816 * final.c (need_profile_function): New variable.
23817 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
23818 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
23819 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
23820 notes, targetm.asm_out.function_prologue doesn't emit anything,
23821 HAVE_prologue and profiler should be emitted before prologue,
23822 set need_profile_function instead of emitting it.
23823 (final_scan_insn): If need_profile_function, emit
23824 profile_function on the first NOTE_INSN_BASIC_BLOCK or
23825 NOTE_INSN_FUNCTION_BEG note.
23827 2013-02-01 Richard Henderson <rth@redhat.com>
23829 * config/rs6000/rs6000.md (smulditi3): New.
23832 * config/alpha/alpha.md (umulditi3): New.
23834 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
23836 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
23837 (ASM_OUTPUT_ALIGNED_LOCAL): New.
23839 2013-02-01 Richard Biener <rguenther@suse.de>
23841 PR tree-optimization/56113
23842 * tree-ssa-structalias.c (label_visit): Reduce work for
23843 single-predecessor nodes.
23845 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
23847 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
23848 range isn't testing for zero.
23850 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
23852 PR middle-end/56113
23853 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
23855 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
23856 Nick Clifton <nickc@redhat.com>
23858 * config/v850/constraints.md (Q): Define as a memory constraint.
23859 * config/v850/predicates.md (label_ref_operand): New predicate.
23860 (e3v5_shift_operand): New predicate.
23861 (ior_operator): New predicate.
23862 * config/v850/t-v850: Add e3v5 multilib.
23863 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
23864 (v850_gen_movdi): Prototype.
23865 * config/v850/v850.c: Add support for e3v5 architecture.
23866 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
23868 (construct_save_jarl): Add e3v5 long JARL support.
23869 (v850_adjust_insn_length): New function. Adjust length of call
23870 insns when using e3v5 instructions.
23871 (v850_gen_movdi): New function: Generate instructions to move a
23873 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
23874 (CPP_SPEC): Define __v850e3v5__ as appropriate.
23875 (TARGET_USE_FPU): Enable for e3v5.
23876 (CONST_OK_FOR_W): New macro.
23877 (ADJUST_INSN_LENGTH): Define.
23878 * config/v850/v850.md (UNSPEC_LOOP): Define.
23879 (attr cpu): Add v850e3v5.
23880 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
23881 (movdi): New pattern.
23882 (movdi_internal): New pattern.
23883 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
23884 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
23885 (cstoresf4): Likewise.
23886 (cstoredf4): Likewise.
23887 (insv): New pattern.
23888 (rotlso3_a): New pattern.
23889 (rotlsi3_b): New pattern
23890 (rotlsi3_v850e3v5): New pattern.
23891 (doloop_begin): New pattern.
23892 (fix_loop_counter): New pattern.
23893 (doloop_end): New pattern.
23894 (branch_normal): Add e3v5 long branch support.
23895 (branch_invert): Likewise.
23896 (branch_z_normal): Likewise.
23897 (branch_z_invert): Likewise.
23898 (branch_nz_normal): Likewise.
23899 (branch_nz_invert): Likewise.
23900 (call_internal_short): Add e3v5 register-indirect JARL support.
23901 (call_internal_long): Likewise.
23902 (call_value_internal_short): Likewise.
23903 (call_value_internal_long): Likewise.
23904 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
23905 (mloop): New option.
23906 * config.gcc: Add support for configuring v840e3v5 target.
23907 * doc/invoke.texi: Document new v850 specific command line options.
23909 2013-01-31 Paul Koning <ni1d@arrl.net>
23913 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
23914 children if parent is a class.
23915 (prune_unused_types_prune): Don't add DW_AT_declaration.
23917 2013-01-31 Richard Biener <rguenther@suse.de>
23919 PR tree-optimization/56157
23920 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
23921 match up operand with SLP child.
23923 2013-01-31 Jason Merrill <jason@redhat.com>
23926 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
23927 parameters the first time.
23928 (gen_scheduled_generic_parms_dies): Check completeness here.
23930 2013-01-31 Richard Biener <rguenther@suse.de>
23932 PR middle-end/53073
23933 * common.opt (faggressive-loop-optimizations): New flag,
23934 enabled by default.
23935 * doc/invoke.texi (faggressive-loop-optimizations): Document.
23936 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
23937 infer_loop_bounds_from_undefined by it.
23939 2013-01-31 Richard Biener <rguenther@suse.de>
23941 PR tree-optimization/56150
23942 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
23943 visit virtual operands.
23944 (find_uses_to_rename_bb): Likewise.
23946 2013-01-31 Richard Biener <rguenther@suse.de>
23948 PR tree-optimization/56150
23949 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
23950 mixed store non-store stmts.
23952 2013-01-30 Jakub Jelinek <jakub@redhat.com>
23955 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
23956 LIBASAN_EARLY_SPEC is defined.
23957 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
23958 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
23960 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
23963 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
23964 invalid args instead of ICEing on it.
23965 (ix86_valid_target_attribute_tree): Return error_mark_node if
23966 ix86_valid_target_attribute_inner_p failed.
23967 (ix86_valid_target_attribute_p): Return false only if
23968 ix86_valid_target_attribute_tree returned error_mark_node. Allow
23969 target("default") attribute.
23970 (sorted_attr_string): Change argument from const char * to tree,
23971 merge in all target attribute arguments rather than just one.
23972 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
23973 instead of free. Avoid using strcat.
23974 (ix86_mangle_function_version_assembler_name): Mangle
23975 target("default") as if no target attribute is present. Adjust
23976 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
23977 instead of xmalloc and XDELETEVEC instead of free.
23978 (ix86_function_versions): Don't return true if one of the decls
23979 doesn't have target attribute. If they don't and one of the decls
23980 is DECL_FUNCTION_VERSIONED, report an error. Adjust
23981 sorted_attr_string caller. Use XDELETEVEC instead of free.
23982 (ix86_supports_function_versions): Remove.
23983 (make_name): Fix up formatting.
23984 (make_dispatcher_decl): Remove resolver_name and its initialization.
23985 Avoid leaking memory.
23986 (is_function_default_version): Return true if there is
23987 target("default") attribute rather than no target attribute at all.
23988 (make_resolver_func): Avoid leaking memory.
23989 (ix86_generate_version_dispatcher_body): Likewise.
23990 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
23991 * target.def (supports_function_versions): Remove.
23992 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
23993 * doc/tm.texi: Regenerated.
23995 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
23997 PR rtl-optimization/56144
23998 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
23999 for values with side effects.
24001 2013-01-30 Richard Biener <rguenther@suse.de>
24003 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
24004 (sparseset_pop): Likewise.
24005 * cfganal.c (compute_idf): Likewise. Increase work-stack size
24006 to be able to use quick_push in the worker loop.
24008 2013-01-30 Marek Polacek <polacek@redhat.com>
24010 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
24012 2013-01-30 Richard Biener <rguenther@suse.de>
24015 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
24017 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
24019 PR tree-optimization/56064
24020 * fixed-value.c (fixed_from_double_int): New function.
24021 * fixed-value.h (fixed_from_double_int): New prototype.
24022 (const_fixed_from_double_int): New static inline function.
24023 * fold-const.c (native_interpret_fixed): New static function.
24024 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
24025 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
24026 (native_encode_fixed): New static function.
24027 (native_encode_expr) <FIXED_CST>: Use it.
24028 (native_interpret_int): Move double_int worker code to...
24029 * double-int.c (double_int::from_buffer): ...this new static method.
24030 * double-int.h (double_int::from_buffer): Prototype it.
24032 2013-01-30 Richard Biener <rguenther@suse.de>
24034 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
24035 New pointer-map and obstack.
24036 (init_alias_vars): Allocate pointer-map and obstack.
24037 (delete_points_to_sets): Free them.
24038 (find_what_var_points_to): Cache result.
24039 (find_what_p_points_to): Adjust for changed interface of
24040 find_what_var_points_to.
24041 (compute_points_to_sets): Likewise.
24042 (ipa_pta_execute): Likewise.
24044 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24046 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
24047 * configure: Regenerate.
24048 * config.in: Regenerate.
24049 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
24050 #nobits/#progbits if supported.
24052 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
24055 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
24056 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
24057 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
24059 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24061 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
24062 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
24064 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24066 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
24067 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
24069 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24071 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
24073 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
24074 * config/arm/cortex-a7.md: New bypasses using
24075 arm_mac_accumulator_is_result.
24077 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24079 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
24080 (cortex_a7_neon_mla): Likewise.
24081 (cortex_a7_fpfmad): New reservation.
24082 (cortex_a7_fpmacs): Use ffmas and update required units.
24083 (cortex_a7_fpmuld): Update required units and latency.
24084 (cortex_a7_fpmacd): Likewise.
24085 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
24086 (cortex_a7_neon). Likewise.
24087 (bypass) Update participating units.
24089 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24091 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
24092 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
24094 * config/arm/vfp11.md (vfp_farith): Use ffmas.
24095 (vfp_fmul): Use ffmad.
24096 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
24097 (cortex_r4_fmacd): Use ffmad.
24098 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
24099 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
24100 (cortex_a9_fmacd): Use ffmad.
24101 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
24102 (cortex_a8_vfp_macd): Use ffmad.
24103 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
24104 (cortex_a5_fpmacd): Use ffmad.
24105 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
24106 (cortex_a15_vfp_macd): Use ffmad.
24107 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
24109 2013-01-29 Jason Merrill <jason@redhat.com>
24112 * varasm.c (default_assemble_visibility): Don't warn about
24113 visibility on artificial decls.
24115 2013-01-29 Richard Biener <rguenther@suse.de>
24117 PR tree-optimization/56113
24118 * tree-ssa-structalias.c (equiv_class_lookup): Also return
24120 (label_visit): Free duplicate bitmaps and record the leader instead.
24121 (perform_var_substitution): Adjust.
24123 2013-01-29 Richard Biener <rguenther@suse.de>
24125 PR tree-optimization/55270
24126 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
24127 the CFG, schedule loops for fixup.
24129 2013-01-29 Nick Clifton <nickc@redhat.com>
24131 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
24134 2013-01-28 Leif Ekblad <leif@rdos.net>
24136 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
24137 * config/i386/i386.h (TARGET_RDOS): New macro.
24138 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
24139 * config/i386/i386.c (ix86_option_override_internal): For 64bit
24140 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
24141 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
24142 DEFAULT_LARGE_SECTION_THRESHOLD.
24143 * config/i386/i386.md (R14_REG, R15_REG): New constants.
24144 * config/i386/rdos.h: New file.
24145 * config/i386/rdos64.h: New file.
24147 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
24150 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
24153 2013-01-28 Jakub Jelinek <jakub@redhat.com>
24155 PR rtl-optimization/56117
24156 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
24157 call cselib_lookup_from_insn on the MEM before calling
24158 add_insn_mem_dependence.
24160 2013-01-28 Richard Biener <rguenther@suse.de>
24162 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
24163 to a stmt that didn't have one.
24164 (copy_phis_for_bb): Likewise for PHI arguments.
24165 (copy_debug_stmt): Likewise for debug stmts.
24167 2013-01-28 Richard Biener <rguenther@suse.de>
24169 PR tree-optimization/56034
24170 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
24171 (partition_builtin_p): Adjust.
24172 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
24173 it is the last partition.
24174 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
24175 up the vertex for the definition.
24176 (classify_partition): Classify whether a partition is a
24177 PKIND_REDUCTION, thus has uses outside of the loop.
24178 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
24179 Merge all PKIND_REDUCTION partitions into the last partition.
24180 (tree_loop_distribution): Seed partitions from reductions as well.
24182 2013-01-28 Jakub Jelinek <jakub@redhat.com>
24184 PR tree-optimization/56125
24185 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
24186 pow(x,c) into sqrt(x) * powi(x, n/2) or
24187 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
24188 optimizing for size.
24189 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
24190 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
24193 PR tree-optimization/56094
24194 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
24195 to UNKNOWN_LOCATION while gimplifying expr.
24197 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
24200 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
24201 operand 0 in movabs insn template for -masm=intel asm alternative.
24202 (*movabs<mode>_2): Ditto for operand 1.
24204 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
24207 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
24210 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24212 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
24215 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24217 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
24218 Undef to avoid warning.
24220 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24222 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
24223 * configure: Regenerate.
24225 2013-01-25 Jakub Jelinek <jakub@redhat.com>
24227 PR tree-optimization/56098
24228 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
24229 for stmts with volatile ops.
24230 (cond_store_replacement): Don't optimize if assign has volatile ops.
24231 (cond_if_else_store_replacement_1): Don't optimize if either
24232 then_assign or else_assign have volatile ops.
24233 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
24236 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
24238 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
24240 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
24242 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
24243 missing ':' in asm example.
24245 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
24247 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
24248 entries into lane and laneq entries.
24249 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
24250 Remove AdvSIMD scalar modes.
24251 (aarch64_sq<r>dmulh_laneq<mode>): New.
24252 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
24254 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
24255 builtin implementations to relfect changes in RTL in aarch64-simd.md.
24256 * config/aarch64/iterators.md (VCOND): New.
24259 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
24262 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
24263 Add NULL LIBNAME argument to existing definitions.
24264 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
24265 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
24266 * config/avr/avr.c (DEF_BUILTIN): Same.
24267 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
24268 (avr_expand_builtin): Expand to a vanilla call if a libgcc
24269 implementation is available (DECL_ASSEMBLER_NAME is set).
24270 (avr_fold_absfx): New static function.
24271 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
24272 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
24273 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
24274 AVR_BUILTIN_ABSLLK.
24275 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
24276 (abshk, absk, abslk, absllk): Provide as static inline functions.
24278 2013-01-25 Marek Polacek <polacek@redhat.com>
24280 PR tree-optimization/56035
24281 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
24283 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
24285 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
24286 (*movtf_internal_rex64): Add (!o,C) alternative
24287 (*movxf_internal_rex64): Ditto.
24288 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
24290 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
24292 * doc/invoke.texi: fix typo.
24293 * doc/objc.texi: fix typo.
24295 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
24297 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
24298 for the first two alternatives.
24300 2013-01-24 Diego Novillo <dnovillo@google.com>
24302 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
24303 (ggc-zone.o): Remove.
24304 * configure.ac: Remove option --with-gc.
24305 * configure: Re-generate.
24306 * doc/install.texi: Remove documentation for --with-gc.
24307 * gengtype.c (write_enum_defn): Remove. Update all users.
24308 (write_Types_process_field): Remove generation of gt_e_* argument.
24309 (output_type_enum): Remove. Update all users.
24310 (write_enum_defn): Remove. Update all users.
24311 (enum alloc_zone): Remove. Update all users.
24312 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
24313 * ggc-common.c (ggc_splay_alloc): Remove first argument.
24314 Update all callers.
24315 (struct ptr_data): Remove field TYPE. Update all users.
24316 (gt_pch_note_object): Remove argument TYPE. Update all users.
24317 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
24319 * ggc-none.c (ggc_alloc_typed_stat): Remove.
24320 (struct alloc_zone): Remove.
24321 (ggc_internal_alloc_zone_stat): Remove.
24322 (ggc_internal_cleared_alloc_zone_stat): Remove.
24323 * ggc-page.c (ggc_alloc_typed_stat): Remove.
24324 (ggc_pch_count_object): Remove last argument. Update all users.
24325 (ggc_pch_alloc_object): Remove last argument. Update all users.
24326 (struct alloc_zone): Remove.
24327 * ggc-zone.c: Remove.
24328 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
24329 (struct alloc_zone): Remove.
24330 (ggc_alloc_typed_stat): Remove.
24331 (ggc_alloc_typed): Remove.
24332 (ggc_splay_alloc): Remove first argument.
24333 (rtl_zone): Remove. Update all users.
24334 (tree_zone): Remove. Update all users.
24335 (tree_id_zone): Remove. Update all users.
24336 (ggc_internal_zone_alloc_stat): Remove. Update all users.
24337 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
24338 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
24339 * tree-ssanames.c: Remove references to zone allocator in comments.
24341 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
24343 * config/avr/avr.c (avr_out_fract): Make register numbers that
24344 might be outside of source operand signed.
24346 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
24348 * config/i386/constraints.md (Yf): New constraint.
24349 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
24350 of f constraint to conditionaly disable x87 register preferences.
24351 (*movdf_internal): Ditto.
24352 (*movsf_internal): Ditto.
24354 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
24356 PR inline-asm/55934
24357 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
24358 that have operands with impossible constraints.
24359 Add a FIXME for a speed-up opportunity.
24360 * lra-constraints.c (process_alt_operands): Verify that a class
24361 selected from constraints on asms is valid for the operand mode.
24362 (curr_insn_transform): Remove incorrect comment.
24364 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
24366 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
24367 TOC operand is a valid symbol ref in the constant pool.
24369 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24371 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
24373 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
24376 * config/avr/stdfix.h: New file.
24377 * t-avr (stdfix-gcc.h): New rule to build it.
24378 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
24380 2013-01-23 Kostya Serebryany <kcc@google.com>
24382 * config/darwin.h: remove dependency on
24383 CoreFoundation (asan on Mac OS).
24385 2013-01-23 Jakub Jelinek <jakub@redhat.com>
24388 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
24389 instead of cmpdi_operand for first comparison operand.
24390 Don't assert that comparison operands aren't both constants.
24392 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
24394 * doc/install.texi (Downloading the Source): Update references to
24395 downloading separate components.
24397 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
24399 * doc/extend.texi (__int128): Improve grammar.
24401 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
24404 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
24405 alternative to (o,r).
24406 (*movdi_internal_rex64): Remove (!o,n) alternative.
24407 (DImode immediate->memory splitter): Remove.
24408 (DImode immediate->memory peephole2): Remove.
24409 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
24410 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
24411 alternative to (!o,*r).
24412 (*movtf_internal_sse): New pattern.
24413 (*movxf_internal_rex64): New pattern.
24414 (*movxf_internal): Disable for TARGET_64BIT.
24415 (*movdf_internal_rex64): Remove (!o,F) alternative.
24417 2013-01-22 Jakub Jelinek <jakub@redhat.com>
24419 PR middle-end/56074
24420 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
24421 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
24422 * tree-vect-loop-manip.c (find_loop_location): Also ignore
24423 stmt locations where LOCATION_LOCUS of the stmt location is
24424 UNKNOWN_LOCATION or BUILTINS_LOCATION.
24427 * config/i386/i386.md (UNSPEC_STOS): New.
24428 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
24429 *strsetqi_1): Add UNSPEC_STOS.
24431 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
24434 * doc/invoke.texi: Remove left over -Wsynth example.
24436 2013-01-21 Jakub Jelinek <jakub@redhat.com>
24438 PR tree-optimization/56051
24439 * fold-const.c (fold_binary_loc): Don't fold
24440 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
24441 a narrowing conversion, or widening conversion from signed
24444 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
24446 PR rtl-optimization/56023
24447 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
24448 dependent on debug instruction.
24450 2013-01-21 Martin Jambor <mjambor@suse.cz>
24452 PR middle-end/56022
24453 * function.c (allocate_struct_function): Call
24454 invoke_set_current_function_hook earlier.
24456 2013-01-21 Jakub Jelinek <jakub@redhat.com>
24458 * reload1.c (init_reload): Only initialize reload_obstack
24459 during the first call.
24461 2013-01-21 Marek Polacek <polacek@redhat.com>
24463 * cfgloop.c (verify_loop_structure): Fix up grammar.
24465 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
24467 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
24468 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
24470 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24473 * config/arm/marvell-pj4.md: Update copyright year.
24474 Fix up use of alu to alu_reg and simple_alu_imm.
24476 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
24478 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
24480 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
24483 * lra-constraints.c (curr_insn_transform): Don't reuse original
24484 insn for secondary memory move when memory mode should be different.
24486 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24488 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
24489 atomic_storedi_1): New patterns.
24491 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
24493 btver2 pipeline descriptions.
24494 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
24496 * config/i386/i386.md (btver2_decode): New type attributes.
24497 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
24499 * config/i386/btver2.md: New file describing btver2 pipelines.
24501 2013-01-19 Andrew Pinski <apinski@cavium.com>
24503 PR tree-optimization/52631
24504 * tree-ssa-sccvn (visit_use): Before looking up the original
24505 statement, try looking up the simplified expression.
24507 2013-01-19 Anthony Green <green@moxielogic.com>
24509 * config/moxie/moxie.c (moxie_expand_prologue): Set
24510 current_function_static_stack_size.
24512 2013-01-18 Jakub Jelinek <jakub@redhat.com>
24514 PR tree-optimization/56029
24515 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
24516 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
24518 2013-01-18 Sharad Singhai <singhai@google.com>
24520 PR tree-optimization/55995
24521 * dumpfile.c (dump_loc): Print location only if available.
24522 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
24524 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
24527 * lra-constraints.c (curr_insn_transform): Reuse original insn for
24528 secondary memory move.
24529 (inherit_reload_reg): Use rclass instead of cl for
24530 check_secondary_memory_needed_p.
24532 2013-01-18 Jakub Jelinek <jakub@redhat.com>
24534 PR middle-end/56015
24535 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
24536 the case where writing real complex part of target modifies op1.
24538 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
24540 * config/aarch64/aarch64-simd.md
24541 (aarch64_vcond_internal<mode>): Handle unordered cases.
24542 * config/aarch64/iterators.md (v_cmp_result): New.
24544 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
24545 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24547 * config/arm/marvell-pj4.md: New file.
24548 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
24549 * config/arm/arm.md (generic_sched): Add marvell_pj4.
24550 (generic_vfp): Likewise.
24551 * config/arm/arm-cores.def: Add marvell-pj4.
24552 * config/arm/arm-tune.md: Regenerate.
24553 * config/arm/arm-tables.opt: Regenerate.
24554 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
24555 * doc/invoke.texi: Document marvell-pj4.
24557 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
24559 * config/aarch64/arm_neon.h: Map scalar types to standard types.
24561 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24566 * var-tracking.c (negative_power_of_two_p): New.
24567 (global_get_addr_cache, local_get_addr_cache): New.
24568 (get_addr_from_global_cache, get_addr_from_local_cache): New.
24569 (vt_canonicalize_addr): Rewrite using the above. Adjust the
24571 (vt_stack_offset_p): Remove.
24572 (vt_canon_true_dep): Always canonicalize loc's address.
24573 (clobber_overlapping_mems): Make sure we have a MEM.
24574 (local_get_addr_clear_given_value): New.
24575 (val_reset): Clear local cached entries.
24576 (compute_bb_dataflow): Create and release the local cache.
24577 Disable duplicate MEMs clobbering.
24578 (emit_notes_in_bb): Clobber MEMs likewise.
24579 (vt_emit_notes): Create and release the local cache.
24580 (vt_initialize, vt_finalize): Create and release the global
24581 cache, respectively.
24582 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
24584 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24586 PR libmudflap/53359
24587 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
24588 not found in the symtab.
24590 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
24593 PR rtl-optimization/55547
24594 PR rtl-optimization/53827
24597 * alias.c (offset_overlap_p): New, factored out of...
24598 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
24599 the conservative special case for symbolic constants. Don't
24600 adjust zero sizes on alignment.
24602 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
24604 PR rtl-optimization/52573
24605 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
24606 REG_UNUSED for the same register.
24608 2013-01-17 Richard Biener <rguenther@suse.de>
24609 Marek Polacek <polacek@redhat.com>
24611 PR rtl-optimization/55833
24612 * loop-unswitch.c (unswitch_loops): Move loop verification...
24613 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
24614 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
24615 Set it to true when we're removing a loop from hierarchy tree in
24616 an irreducible region.
24617 (fix_bb_placements): Adjust caller.
24618 (fix_loop_placements): Likewise.
24620 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
24622 * config/avr/builtins.def (DEF_BUILTIN): Factor out
24623 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
24624 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
24625 Remove ID. Adjust comments.
24626 * config/avr/avr-c.c (avr_builtin_name): Remove.
24627 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
24628 * config/avr/avr.c (avr_tolower): New static function.
24629 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
24630 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
24631 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
24634 2013-01-17 Jan Hubicka <jh@suse.cz>
24636 PR tree-optimization/55273
24637 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
24639 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
24642 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
24643 store through atomic_store<mode>_1.
24644 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
24646 2013-01-17 Martin Jambor <mjambor@suse.cz>
24648 PR tree-optimizations/55264
24649 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
24650 for virtual methods.
24651 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
24652 virtual methods before inlining is over.
24653 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
24655 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
24658 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
24660 PR rtl-optimization/56005
24661 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
24662 pending reads for prefetch.
24664 2013-01-16 Ian Bolton <ian.bolton@arm.com>
24666 * config/aarch64/aarch64.md
24667 (*cstoresi_neg_uxtw): New pattern.
24668 (*cmovsi_insn_uxtw): New pattern.
24669 (*<optab>si3_uxtw): New pattern.
24670 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
24671 (*<optab>si3_insn_uxtw): New pattern.
24672 (*bswapsi2_uxtw): New pattern.
24674 2013-01-16 Richard Biener <rguenther@suse.de>
24676 * tree-inline.c (tree_function_versioning): Remove set but
24677 never used variable.
24679 2013-01-16 Richard Biener <rguenther@suse.de>
24681 PR tree-optimization/55964
24682 * tree-flow.h (rename_variables_in_loop): Remove.
24683 (rename_variables_in_bb): Likewise.
24684 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
24685 (copy_loop_before): Adjust and delete update-ssa status.
24686 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
24687 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
24688 (rename_variables_in_loop): Remove.
24689 (slpeel_update_phis_for_duplicate_loop): Likewise.
24690 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
24691 use available cfg machinery instead of duplicating it.
24692 Update PHI nodes and perform poor-mans SSA update here.
24693 (slpeel_tree_peel_loop_to_edge): Adjust.
24695 2013-01-16 Richard Biener <rguenther@suse.de>
24697 PR tree-optimization/54767
24698 PR tree-optimization/53465
24699 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
24700 (vrp_visit_phi_node): For PHI arguments coming via backedges
24701 drop all symbolical range information.
24702 (execute_vrp): Compute backedges.
24704 2013-01-16 Richard Biener <rguenther@suse.de>
24706 * doc/install.texi: Update CLooG and ISL requirements to
24709 2013-01-16 Christian Bruel <christian.bruel@st.com>
24712 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
24713 (broken_move): Handle UNSPECV_SP_SWITCH_B.
24714 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
24716 2013-01-16 DJ Delorie <dj@redhat.com>
24718 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
24719 (UNSPECV_SP_SWITCH_E): New.
24720 (sp_switch_1): Change to an unspec.
24721 (sp_switch_2): Change to an unspec. Don't use post-inc when we
24724 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
24726 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
24727 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
24728 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
24729 (expand_mem_thread_fence): Ditto.
24730 (expand_mem_signal_fence): Ditto.
24731 (expand_atomic_load): Ditto.
24732 (expand_atomic_store): Ditto.
24734 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
24736 PR rtl-optimization/55547
24737 PR rtl-optimization/53827
24740 * alias.c (memrefs_conflict_p): Set sizes to negative after
24743 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24746 * function.c (thread_prologue_and_epilogue_insns): Always
24747 add crtl->drap_reg to set_up_by_prologue.set, even if
24748 stack_realign_drap is false.
24750 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
24752 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
24753 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
24754 *call): Fix indention.
24756 2013-01-15 Tom de Vries <tom@codesourcery.com>
24759 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
24762 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
24764 PR rtl-optimization/55153
24765 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
24767 2013-01-15 Martin Jambor <mjambor@suse.cz>
24769 PR tree-optimization/55920
24770 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
24771 accesses as grp_to_be_debug_replaced.
24773 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24775 PR tree-optimization/55920
24776 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
24777 there is non-useless type conversion needed from debug rhs to lhs,
24778 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
24780 2013-01-15 Joseph Myers <joseph@codesourcery.com>
24781 Mikael Pettersson <mikpe@it.uu.se>
24784 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
24786 (ASM_OUTPUT_CASE_LABEL): Remove.
24787 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
24788 * final.c (shorten_branches): Update alignment of labels before
24789 jump tables if CASE_VECTOR_SHORTEN_MODE.
24791 2013-01-15 Richard Biener <rguenther@suse.de>
24794 * system.h: Do not include gmp.h for building host tools.
24796 2013-01-15 Richard Biener <rguenther@suse.de>
24798 PR middle-end/55882
24799 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
24800 account for bitpos when computing alignment.
24802 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
24804 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
24805 (ix86_target_macros_internal): Likewise.
24807 * config/i386/i386.c (m_CORE2I7): Removed.
24808 (m_CORE_HASWELL): New macro.
24809 (m_CORE_ALL): Likewise.
24810 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
24811 (initial_ix86_arch_features): Likewise.
24812 (processor_target_table): Initializations for Core avx2.
24813 (cpu_names): New names "core-avx2".
24814 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
24815 PROCESSOR_CORE_HASWELL.
24816 (ix86_issue_rate): New case.
24817 (ia32_multipass_dfa_lookahead): Likewise.
24818 (ix86_sched_init_global): Likewise.
24820 * config/i386/i386.h (TARGET_HASWELL): New macro.
24821 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
24822 (processor_type): New PROCESSOR_HASWELL.
24824 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24826 PR tree-optimization/55955
24827 * tree-vect-loop.c (vectorizable_reduction): Give up early on
24828 *SHIFT_EXPR and *ROTATE_EXPR codes.
24830 PR tree-optimization/48766
24831 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
24832 -ftrapv disable -fwrapv.
24834 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24837 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
24838 etc. to 1 and not to __flash.
24839 Use LL suffix for __INT24_MAX__ with -mint8.
24840 Use ULL suffix for __UINT24_MAX__ with -mint8.
24842 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24844 * config/avr/avr-arch.h
24845 (struct base_arch_s): Use typedef avr_arch_t instead.
24846 (struct arch_info_s): Use typedef avr_arch_info_t instead.
24847 (struct mcu_type_s): Use typedef avr_mcu_t instead.
24848 * config/avr/avr.c: Same.
24849 * config/avr/avr-devices.c: Same.
24850 * config/avr/driver-avr.c: Same.
24851 * config/avr/gen-avr-mmcu-texi.c: Same.
24852 * config/avr/avr-mcus.def: Adjust comment.
24854 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
24856 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
24857 * config/aarch64/iterators.md (VALLDI): New.
24859 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
24860 Andi Kleen <ak@linux.intel.com>
24863 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
24864 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
24867 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24869 * config/avr/avr-stdint.h: Remove trailing blanks.
24870 * config/avr/avr-log.h: Same.
24871 * config/avr/avr-arch.h: Same.
24872 * config/avr/avr-devices.c: Same.
24873 * config/avr/avr-dimode.md: Same.
24874 * config/avr/predicates.md: Same.
24875 * config/avr/avr-c.c: Same. And fix typo.
24877 * config/avr/avr-protos.h: Same. And:
24878 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
24879 (init_cumulative_args): Rename to avr_init_cumulative_args.
24880 (expand_prologue): Rename to avr_expand_prologue.
24881 (expand_epilogue): Rename to avr_expand_epilogue.
24882 (adjust_insn_length): Rename to avr_adjust_insn_length.
24883 (notice_update_cc): Rename to avr_notice_update_cc.
24884 (final_prescan_insn): Rename to avr_final_prescan_insn.
24885 * config/avr/avr.c: Same.
24886 * config/avr/avr.h: Same.
24887 * config/avr/avr.md: Remove trailing blanks.
24888 (prologue): Use avr_expand_prologue.
24889 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
24891 2013-01-14 Richard Biener <rguenther@suse.de>
24893 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
24894 verify_location, collect_subblocks): New functions.
24895 (verify_gimple_in_cfg): Verify that locations only reference
24896 BLOCKs in the functions BLOCK tree.
24898 2013-01-14 Richard Biener <rguenther@suse.de>
24900 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
24902 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
24904 (insert_out_of_ssa_copy_on_edge): Likewise.
24905 (rewrite_close_phi_out_of_ssa): Likewise.
24906 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
24908 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
24909 propagated constants.
24910 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
24913 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24915 * config/avr/avr-modes.def: Add GPL copyright notice.
24917 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
24919 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
24920 MEMMODEL_MASK to determine memory model.
24921 (atomic_store<mode>): Ditto from operands[2].
24922 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
24924 2013-01-13 Jakub Jelinek <jakub@redhat.com>
24927 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
24928 (fold_gimple_assign): Don't call unshare_expr here.
24929 (fold_ctor_reference): Call unshare_expr.
24931 2013-01-13 Terry Guo <terry.guo@arm.com>
24933 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
24934 * doc/fragments.texi: Document MULTILIB_REUSE.
24935 * gcc.c (multilib_reuse): New internal spec.
24936 (set_multilib_dir): Also search multilib from multilib_reuse.
24937 * genmultilib (tmpmultilib3): Refactor code.
24938 (tmpmultilib4): Ditto.
24939 (multilib_reuse): New multilib argument.
24941 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
24943 * Makefile.in: Update copyright.
24945 2013-01-12 Tom de Vries <tom@codesourcery.com>
24947 PR middle-end/55890
24948 * calls.c (expand_call): Check if arg_nr is valid.
24950 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
24952 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
24953 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
24954 documentation. Add missing '__' in front of
24955 __builtin_ia32_packssdw256.
24957 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24960 * config/s390/s390.c (s390_preferred_reload_class): Do not return
24961 NO_REGS for larl operands.
24962 (s390_reload_larl_operand): Use s390_load_address instead of
24965 2013-01-11 Richard Biener <rguenther@suse.de>
24967 * tree-cfg.c (verify_node_sharing_1): Split out from ...
24968 (verify_node_sharing): ... here.
24969 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
24971 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
24973 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
24974 Substitute TREECHECKING.
24975 * configure: Regenerate.
24976 * Makefile.in (TREECHECKING): New.
24978 2013-01-11 Richard Guenther <rguenther@suse.de>
24980 PR tree-optimization/44061
24981 * tree-vrp.c (extract_range_basic): Compute zero as
24982 value-range for __builtin_constant_p of function parameters.
24984 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
24986 Update copyright years.
24988 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
24990 PR rtl-optimization/55672
24991 * lra-eliminations.c (mark_not_eliminable): Permit addition with
24992 const to be eliminable.
24994 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
24996 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
24997 * configure: Regenerate.
24999 2013-01-10 Richard Biener <rguenther@suse.de>
25001 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
25003 2013-01-10 Richard Biener <rguenther@suse.de>
25006 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
25007 locations for virtual PHI arguments.
25008 (rewrite_update_phi_arguments): Likewise.
25010 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
25012 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
25015 2013-01-10 Jakub Jelinek <jakub@redhat.com>
25017 PR tree-optimization/55921
25018 * tree-complex.c (expand_complex_asm): New function.
25019 (expand_complex_operations_1): Call it for GIMPLE_ASM.
25021 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25024 * config/s390/s390.c (s390_symref_operand_p)
25025 (s390_loadrelative_operand_p): Merge the two functions.
25026 (s390_check_qrst_address, print_operand_address): Add parameters
25027 to s390_loadrelative_operand_p invokation.
25028 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
25029 (s390_reload_larl_operand, s390_secondary_reload): Use
25030 s390_loadrelative_operand_p instead of s390_symref_operand_p.
25031 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
25033 2013-01-09 Mike Stump <mikestump@comcast.net>
25035 * dse.c (record_store): Remove unnecessary assert.
25037 2013-01-09 Jan Hubicka <jh@suse.cz>
25039 PR tree-optimization/55569
25040 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
25041 * cfgloop.h (scale_loop_profile): Likewise.
25043 2013-01-09 Jan Hubicka <jh@suse.cz>
25046 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
25048 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
25050 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
25052 PR middle-end/55114
25053 * expr.h (maybe_emit_group_store): Declare.
25054 * expr.c (maybe_emit_group_store): New function.
25055 * builtins.c (expand_builtin_int_roundingfn): Call it.
25056 (expand_builtin_int_roundingfn_2): Likewise.
25058 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
25060 PR rtl-optimization/55829
25061 * lra-constraints.c (match_reload): Add code for absent output.
25062 (curr_insn_transform): Add code for reloads of matched inputs
25065 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
25067 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
25068 attribute of movddup insn to DF.
25069 (*vec_interleave_lowv2df): Ditto.
25070 (vec_dupv2df): Ditto.
25072 2013-01-09 Jan Hubicka <jh@suse.cz>
25074 PR tree-optimiation/55875
25075 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
25076 EVERY_ITERATION parameter.
25077 (number_of_iterations_exit): Check if exit is executed every iteration.
25078 (idx_infer_loop_bounds): Similarly here.
25079 (n_of_executions_at_most): Simplify
25080 to only test for cases where statement is dominated by the
25081 particular bound; handle correctly the "postdominance" test.
25082 (scev_probably_wraps_p): Use max loop iterations info
25083 as a global bound first.
25085 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
25086 Nick Clifton <nickc@redhat.com>
25088 * config/v850/v850.md (cbranchsf4): New pattern.
25089 (cstoresf4): New pattern.
25090 (cbranchdf4): New pattern.
25091 (cstoredf4): New pattern.
25092 (movsicc): Disallow floating point comparisons.
25093 (cmpsf_le_insn): Fix order of operators.
25094 (cmpsf_lt_insn): Likewise.
25095 (cmpsf_eq_insn): Likewise.
25096 (cmpdf_le_insn): Likewise.
25097 (cmpdf_lt_insn): Likewise.
25098 (cmpdf_eq_insn): Likewise.
25099 (cmpsf_ge_insn): Use LE comparison.
25100 (cmpdf_ge_insn): Likewise.
25101 (cmpsf_gt_insn): Use LT comparison.
25102 (cmpdf_gt_insn): Likewise.
25103 (cmpsf_ne_insn): Delete pattern.
25104 (cmpdf_ne_insn): Delete pattern.
25105 * config/v850/v850.c (v850_gen_float_compare): Use
25106 gen_cmpdf_eq_insn for NE comparison.
25107 (v850_float_z_comparison_operator)
25108 (v850_float_nz_comparison_operator): Move from here ...
25109 * config/v850/predicates.md: ... to here. Move GT and GE
25110 comparisons into v850_float_z_comparison_operator.
25111 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
25113 (v850_float_nz_comparison_operator): Likewise.
25115 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25117 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
25118 with calls to gen_insvsi/gen_insvdi.
25120 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25122 * config/i386/i386.c (initial_ix86_tune_features): Set up
25123 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
25125 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
25126 Jakub Jelinek <jakub@redhat.com>
25128 PR tree-optimization/48189
25129 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
25130 If nitercst is 0, don't predict the exit edge.
25132 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25134 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
25135 in asm_fprintf with reg_names.
25136 (aarch64_print_operand_address): Likewise.
25137 (aarch64_return_addr): Likewise.
25138 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
25140 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25142 * config/pa/pa.h (VAL_U6_BITS_P): Define.
25143 (INT_U6_BITS): Likewise.
25144 * config/pa/predicates.md (uint6_operand): New predicate.
25145 (shift5_operand, shift6_operand): Likewise.
25146 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
25148 (lshrdi3): Use shift6_operand.
25149 (shrpsi4, shrpdi4): New insn patterns.
25150 (extzv): Delete expander.
25151 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
25152 predicates in unamed zero extract patterns. Tighten common constraint.
25153 (extv): Delete expander.
25154 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
25155 predicates in unamed sign extract patterns. Tighten common constraint.
25156 (insv): Delete expander.
25157 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
25158 predicates in unamed insert patterns. Tighten common constraint.
25159 Change uint32_operand predicate to uint6_operand predicate in unamed
25160 DImode pattern to insert constant values of type 1...1xxxx.
25162 2013-01-04 Jan Hubicka <jh@suse.cz>
25164 PR tree-optimization/55823
25165 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
25168 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25169 Uros Bizjak <ubizjak@gmail.com>
25171 PR rtl-optimization/55845
25172 * df-problems.c (can_move_insns_across): Stop scanning at
25173 volatile_insn_p source instruction or give up if
25174 across_from .. across_to range contains any volatile_insn_p
25177 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
25179 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
25180 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
25182 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
25183 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
25185 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25188 * asan.c (asan_clear_shadow): New function.
25189 (asan_emit_stack_protection): Use it.
25191 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
25193 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
25194 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
25195 with tab instead of space.
25197 2013-01-08 Nick Clifton <nickc@redhat.com>
25199 * config/rl78/rl78.c (rl78_expand_prologue): Always select
25200 register bank 0 at the start of an interrupt handler.
25201 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
25204 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
25206 * config/aarch64/aarch64-simd.md
25207 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
25208 (aarch64_simd_bsl): Likewise.
25209 (aarch64_vcond_internal<mode>): Likewise.
25210 (vcond<mode><mode>): Likewise.
25211 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
25212 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
25214 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
25216 * config/aarch64/aarch64-builtins.c
25217 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
25219 2013-01-08 Martin Jambor <mjambor@suse.cz>
25222 * tree-sra.c (analyze_access_subtree): Return true also after
25223 potentially creating a debug-only replacement.
25225 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25227 PR middle-end/55890
25228 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
25230 PR tree-optimization/54120
25231 * tree-vrp.c (range_fits_type_p): Don't allow
25232 src_precision < precision from signed vr to unsigned_p
25233 if vr->min or vr->max is negative.
25234 (simplify_float_conversion_using_ranges): Test can_float_p
25235 against CODE_FOR_nothing.
25237 2013-01-08 Jakub Jelinek <jakub@redhat.com>
25238 Richard Biener <rguenther@suse.de>
25240 PR middle-end/55851
25241 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
25242 types instead of just INTEGER_TYPE types.
25244 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
25246 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
25249 2013-01-07 Steve Ellcey <sellcey@mips.com>
25252 * config/mips/mips.opt: Change mad to mmad to match documentation.
25254 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25257 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
25258 .progmemx.data now.
25260 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25263 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
25264 (avr_addrspace_t): Add .section_name field.
25265 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
25267 (avr_addrspace): Same. Initialize .section_name. Remove last
25268 NULL entry. Put __memx into .progmemx.data.
25269 (progmem_section_prefix): Remove.
25270 (avr_asm_init_sections): No need to initialize progmem_section.
25271 (avr_asm_named_section): Use avr_addrspace[].section_name to get
25272 section name prefix.
25273 (avr_asm_select_section): Ditto. And use get_unnamed_section to
25274 retrieve the progmem section.
25275 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
25276 boundary to run over avr_addrspace[].
25277 (avr_register_target_pragmas): Ditto.
25279 2013-01-06 Jakub Jelinek <jakub@redhat.com>
25281 * varasm.c (output_constant_def_contents): For asan_protect_global
25282 protected strings, adjust DECL_ALIGN if needed, before testing for
25284 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
25285 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
25287 (output_object_block): For asan protected decls, emit asan padding
25288 after their contents.
25289 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
25290 (asan_finish_file): Test it here instead.
25292 2013-01-07 Nick Clifton <nickc@redhat.com>
25293 Matthias Klose <doko@debian.org>
25294 Doug Kwan <dougkwan@google.com>
25295 H.J. Lu <hongjiu.lu@intel.com>
25298 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
25300 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
25302 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
25304 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
25306 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
25308 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25311 * doc/install.texi (Cross-Compiler-Specific Options): Document
25314 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
25316 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
25317 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
25318 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
25319 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
25320 vqmovun_high_s64): Fix source operand number and update copyright.
25322 2013-01-07 Richard Biener <rguenther@suse.de>
25324 PR middle-end/55890
25325 * gimple.h (gimple_call_builtin_p): New overload.
25326 * gimple.c (validate_call): New function.
25327 (gimple_call_builtin_p): Likewise.
25328 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25329 Use gimple_call_builtin_p.
25330 (find_func_clobbers): Likewise.
25331 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
25332 (strlen_optimize_stmt): Likewise.
25334 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
25336 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
25337 (vld1q_dup_*): Likewise.
25338 (vld1_*): Likewise.
25339 (vld1q_*): Likewise.
25340 (vld1_lane_*): Likewise.
25341 (vld1q_lane_*): Likewise.
25343 2013-01-07 Richard Biener <rguenther@suse.de>
25345 * lto-streamer.h (LTO_minor_version): Bump to 2.
25347 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
25349 * config/aarch64/aarch64-protos.h
25350 (aarch64_const_double_zero_rtx_p): Rename to...
25351 (aarch64_float_const_zero_rtx_p): ...this.
25352 (aarch64_float_const_representable_p): New.
25353 (aarch64_output_simd_mov_immediate): Likewise.
25354 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
25355 move immediate case.
25356 * config/aarch64/aarch64.c
25357 (aarch64_const_double_zero_rtx_p): Rename to...
25358 (aarch64_float_const_zero_rtx_p): ...this.
25359 (aarch64_print_operand): Allow printing of new constants.
25360 (aarch64_valid_floating_const): New.
25361 (aarch64_legitimate_constant_p): Check for valid floating-point
25363 (aarch64_simd_valid_immediate): Likewise.
25364 (aarch64_vect_float_const_representable_p): New.
25365 (aarch64_float_const_representable_p): Likewise.
25366 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
25367 (aarch64_output_simd_mov_immediate): New.
25368 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
25369 (*movdf_aarch64): Likewise.
25370 * config/aarch64/constraints.md (Ufc): New.
25371 (Y): call aarch64_float_const_zero_rtx.
25372 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
25374 2013-01-07 Richard Biener <rguenther@suse.de>
25376 PR tree-optimization/55888
25377 PR tree-optimization/55862
25378 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
25379 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
25380 not if it is contained therein.
25382 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25384 * config/avr/t-avr: Typo.
25386 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
25389 * config/avr/t-avr: Don't automatically rebuild
25390 $(srcdir)/config/avr/t-multilib
25391 $(srcdir)/config/avr/avr-tables.opt
25392 $(srcdir)/doc/avr-mmcu.texi
25393 (avr-mcus): New phony target to build them on request.
25394 (s-avr-mlib, s-avr-mmcu-texi): Remove.
25395 * avr/avr-mcus.def: Adjust comments.
25397 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
25399 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
25401 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
25403 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
25405 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
25407 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
25409 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
25411 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
25412 to generate profiling.
25413 * config/rs6000/aix64.h (LIB_SPEC): Same.
25415 2013-01-04 Andrew Pinski <apinski@cavium.com>
25417 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
25419 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
25421 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
25423 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
25425 (ix86_expand_move): Ditto.
25426 (ix86_zero_extend_to_Pmode): Ditto.
25427 (ix86_expand_call): Ditto.
25428 (ix86_expand_special_args_builtin): Ditto.
25429 (ix86_expand_builtin): Ditto.
25431 2013-01-04 Richard Biener <rguenther@suse.de>
25433 PR tree-optimization/55862
25434 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
25435 translating them through PHI nodes.
25437 2013-01-04 Martin Jambor <mjambor@suse.cz>
25439 PR tree-optimization/55755
25440 * tree-sra.c (sra_modify_assign): Do not check that an access has no
25441 children when trying to avoid producing a VIEW_CONVERT_EXPR.
25443 2013-01-04 Marek Polacek <polacek@redhat.com>
25445 PR middle-end/55859
25446 * opts.c (default_options_optimization): Clarify error message.
25448 2013-01-04 Richard Biener <rguenther@suse.de>
25450 PR middle-end/55863
25451 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
25454 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25457 * config/pa/pa.md (movsi): Revert previous change.
25458 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
25461 2013-01-03 Richard Henderson <rth@redhat.com>
25463 * config/i386/i386.c (ix86_expand_move): Always assign to op1
25464 after eliminating TLS symbols.
25466 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
25469 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
25470 * graphite-poly.c (debug_gmp_value): Likewise.
25472 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
25475 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
25476 selected code model, define __code_mode_small__, __code_model_medium__,
25477 __code_model_large__, __code_model_32__ or __code_model_kernel__.
25478 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
25479 xchg temporary register with %k. Declare temporary register as
25481 [__x86_64__]: For medium and large code models, preserve %rbx register.
25483 2013-01-03 Richard Biener <rguenther@suse.de>
25485 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
25486 (dump_subscript): Adjust.
25487 (finalize_ddr_dependent): Do not dump redundant info.
25488 (analyze_siv_subscript): Adjust.
25489 (subscript_dependence_tester): Likewise.
25490 (compute_affine_dependence): Likewise.
25492 2013-01-03 Richard Biener <rguenther@suse.de>
25495 2013-01-03 Richard Biener <rguenther@suse.de>
25497 PR tree-optimization/55857
25498 * tree-vect-stmts.c (vectorizable_load): Do not setup
25499 re-alignment for invariant loads.
25501 2013-01-02 Richard Biener <rguenther@suse.de>
25503 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
25504 invariant load do not generate a vector load from the scalar location.
25506 2013-01-03 Richard Biener <rguenther@suse.de>
25508 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
25509 for not vectorizing.
25510 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
25511 not build INDIRECT_REFs, call get_name once only.
25512 (vect_create_data_ref_ptr): Likewise. Dump base object kind
25513 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
25515 2013-01-03 Richard Biener <rguenther@suse.de>
25517 PR tree-optimization/55857
25518 * tree-vect-stmts.c (vectorizable_load): Do not setup
25519 re-alignment for invariant loads.
25521 2013-01-03 Richard Biener <rguenther@suse.de>
25524 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
25525 prefer a built-in decl.
25527 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25529 * gcc.c (process_command): Update copyright notice dates.
25530 * gcov.c (print_version): Likewise.
25531 * gcov-dump.c (print_version): Likewise.
25533 PR rtl-optimization/55838
25534 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
25535 iv0.step, iv1.step and step.
25537 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25538 Marc Glisse <marc.glisse@inria.fr>
25540 PR tree-optimization/55832
25541 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
25542 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
25543 integer_{one,zero}_node.
25545 2013-01-03 Jakub Jelinek <jakub@redhat.com>
25548 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
25549 * var-tracking.c (reverse_op): Don't add reverse ops to
25550 VALUEs that have already
25551 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
25553 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
25555 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
25557 2013-01-02 Teresa Johnson <tejohnson@google.com>
25559 * dumpfile.c (dump_loc): Print filename with location.
25560 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
25561 new location_t parameter to emit complete unroll message with
25562 new dump framework.
25563 (canonicalize_loop_induction_variables): Compute loops location
25564 and pass to try_unroll_loop_completely.
25565 * loop-unroll.c (report_unroll_peel): New function.
25566 (peel_loops_completely): Use new dump format with location
25567 for main dumpfile message, and invoke report_unroll_peel on success.
25568 (decide_unrolling_and_peeling): Ditto.
25569 (decide_peel_once_rolling): Remove old dumpfile message subsumed
25570 by report_unroll_peel.
25571 (decide_peel_completely): Ditto.
25572 (decide_unroll_constant_iterations): Ditto.
25573 (decide_unroll_runtime_iterations): Ditto.
25574 (decide_peel_simple): Ditto.
25575 (decide_unroll_stupid): Ditto.
25576 * cfgloop.c (get_loop_location): New function.
25577 * cfgloop.h (get_loop_location): Declare.
25579 2013-01-02 Sriraman Tallam <tmsriram@google.com>
25581 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
25584 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25586 PR middle-end/55198
25587 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
25588 BLKmode objects when EXPAND_MEMORY is specified.
25590 2013-01-02 Sriraman Tallam <tmsriram@google.com>
25592 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
25594 (fold_builtin_cpu): Do not share cpu model decls across statements.
25596 2013-01-02 Jason Merrill <jason@redhat.com>
25599 * tree.c (build_array_type_1): Revert earlier change.
25601 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
25603 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
25605 * config/aarch64/aarch64-tune.md: Re-generate.
25607 2013-01-02 Richard Biener <rguenther@suse.de>
25609 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
25610 invariant load do not generate a vector load from the scalar location.
25612 2013-01-02 Richard Biener <rguenther@suse.de>
25615 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
25616 * configure: Regenerate.
25618 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
25620 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
25621 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
25622 (expand_builtin_int_roundingfn_2): Keep the original target around
25623 for the fallback case.
25625 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
25627 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
25628 to be clear for sign changes.
25630 2013-01-01 Jan Hubicka <jh@suse.cz>
25632 * ipa-inline-analysis.c: Fix formatting.
25634 2013-01-01 Jakub Jelinek <jakub@redhat.com>
25636 PR tree-optimization/55831
25637 * tree-vect-loop.c (get_initial_def_for_induction): Use
25638 gsi_after_labels instead of gsi_start_bb.
25640 Copyright (C) 2013 Free Software Foundation, Inc.
25642 Copying and distribution of this file, with or without modification,
25643 are permitted in any medium without royalty provided the copyright
25644 notice and this notice are preserved.