1 2013-09-06 Richard Biener <rguenther@suse.de>
3 * basic-block.h (class control_dependences): New.
4 * tree-ssa-dce.c (control_dependence_map): Remove.
6 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
8 find_pdom, find_control_dependence, find_all_control_dependences):
10 (mark_control_dependent_edges_necessary, find_obviously_necessary_stmts,
11 propagate_necessity, tree_dce_init, tree_dce_done,
12 perform_tree_ssa_dce): Adjust.
13 * cfganal.c (set_control_dependence_map_bit,
14 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
15 find_all_control_dependences): Move from tree-ssa-dce.c and
16 implement as methods of control_dependences class.
17 (control_dependences::control_dependences): New.
18 (control_dependences::~control_dependences): Likewise.
19 (control_dependences::get_edges_dependent_on): Likewise.
20 (control_dependences::get_edge): Likewise.
22 2013-09-04 Jan Hubicka <jh@suse.cz>
24 * tree.c (types_same_for_odr): Drop overactive check.
25 * ipa-devirt.c (hash_type_name): Likewise.
27 2013-09-04 Jan Hubicka <jh@suse.cz>
29 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
30 (analyze_functions): ... here.
32 2013-09-04 Jan Hubicka <jh@suse.cz>
35 * cgraphunit.c (analyze_functions): Clear AUX fields
36 after processing; initialize assembler name has.
38 2013-09-05 Jeff Law <law@redhat.com>
40 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
41 from thread_around_empty_block. Record threading path into PATH.
42 Recurse if threading through the initial block is successful.
43 (thread_across_edge): Corresponding changes to slightly simplify.
45 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
47 * config/aarch64/aarch64.md
48 (type): Remove frecpe, frecps, frecpx.
49 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
50 fix to be a TARGET_SIMD instruction.
51 (aarch64_frecps): Remove.
52 * config/aarch64/aarch64-simd.md
53 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
54 (aarch64_frecps<mode>): Handle all float/vector of float modes.
56 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
57 Sofiane Naci <sofiane.naci@arm.com>
59 * config/arm/types.md (define_attr "type"):
61 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
63 into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm", "alu_ext",
64 "alu_reg", "alus_ext", "alus_reg", "bfm", "csel", "logic_reg",
67 into "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
69 Expand "arlo_shift_reg"
70 into "alu_shift_reg", "alus_shift_reg", "logic_shift_reg",
72 Expand "clz" into "clz, "rbit".
73 Rename "shift" to "shift_imm".
74 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
76 Update for attribute changes all occurrences of arlo_* and
78 * config/arm/arm-fixed.md: Update for attribute changes
79 all occurrences of arlo_* types.
80 * config/arm/thumb2.md: Update for attribute changes all occurrences
82 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
83 (cortexa7_older_only): Likewise.
84 (cortexa7_younger): Likewise.
85 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
86 (1020alu_shift_op): Likewise.
87 (1020alu_shift_reg_op): Likewise.
88 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
89 (alu_shift_op): Likewise.
90 (alu_shift_reg_op): Likewise.
91 * config/arm/arm1136jfs.md (11_alu_op): Update for
93 (11_alu_shift_op): Likewise.
94 (11_alu_shift_reg_op): Likewise.
95 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
96 (9_alu_shift_reg_op): Likewise.
97 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
99 (cortex_a15_alu_shift): Likewise.
100 (cortex_a15_alu_shift_reg): Likewise.
101 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
103 (cortex_a5_alu_shift): Likewise.
104 * config/arm/cortex-a53.md
105 (cortex_a53_alu): Update for attribute changes.
106 (cortex_a53_alu_shift): Likewise.
107 * config/arm/cortex-a7.md
108 (cortex_a7_alu_imm): Update for attribute changes.
109 (cortex_a7_alu_reg): Likewise.
110 (cortex_a7_alu_shift): Likewise.
111 * config/arm/cortex-a8.md
112 (cortex_a8_alu): Update for attribute changes.
113 (cortex_a8_alu_shift): Likewise.
114 (cortex_a8_alu_shift_reg): Likewise.
115 * config/arm/cortex-a9.md
116 (cortex_a9_dp): Update for attribute changes.
117 (cortex_a9_dp_shift): Likewise.
118 * config/arm/cortex-m4.md
119 (cortex_m4_alu): Update for attribute changes.
120 * config/arm/cortex-r4.md
121 (cortex_r4_alu): Update for attribute changes.
122 (cortex_r4_mov): Likewise.
123 (cortex_r4_alu_shift_reg): Likewise.
124 * config/arm/fa526.md
125 (526_alu_op): Update for attribute changes.
126 (526_alu_shift_op): Likewise.
127 * config/arm/fa606te.md
128 (606te_alu_op): Update for attribute changes.
129 * config/arm/fa626te.md
130 (626te_alu_op): Update for attribute changes.
131 (626te_alu_shift_op): Likewise.
132 * config/arm/fa726te.md
133 (726te_alu_op): Update for attribute changes.
134 (726te_alu_shift_op): Likewise.
135 (726te_alu_shift_reg_op): Likewise.
136 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
137 (mp626_alu_shift_op): Likewise.
138 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
139 (pj4_alu_conds): Likewise.
140 (pj4_shift): Likewise.
141 (pj4_shift_conds): Likewise.
142 (pj4_alu_shift): Likewise.
143 (pj4_alu_shift_conds): Likewise.
144 * config/aarch64/aarch64.md: Update for attribute change
145 all occurrences of arlo_* and shift* types.
147 2013-09-05 Mike Stump <mikestump@comcast.net>
149 * tree.h: Move documentation for tree_function_decl to tree-core.h
150 with the declaration.
152 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
155 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
156 looking for widest mode.
158 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
160 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
162 2013-09-05 Richard Biener <rguenther@suse.de>
164 PR tree-optimization/58137
165 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
166 Do not create vectors of pointers.
167 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
168 types for the components of the vector initializer.
169 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
170 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
172 2013-09-05 Martin Jambor <mjambor@suse.cz>
174 * ipa-prop.c (remove_described_reference): Accept missing references,
175 return false if that hppens, otherwise return true.
176 (cgraph_node_for_jfunc): New function.
177 (try_decrement_rdesc_refcount): Likewise.
178 (try_make_edge_direct_simple_call): Use them.
179 (ipa_edge_removal_hook): Remove references from rdescs.
180 (ipa_edge_duplication_hook): Clone rdescs and their references
181 when the new edge has the same caller as the old one.
182 * cgraph.c (cgraph_resolve_speculation): Remove speculative
183 reference before removing any edges.
185 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
187 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
189 * thumb2.md (thumb2_storewb_parisi): New pattern.
191 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
193 * config/aarch64/aarch64-option-extensions.def: Add
194 AARCH64_OPT_EXTENSION of 'crc'.
195 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
196 (AARCH64_ISA_CRC): Ditto.
197 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
198 description of the CRC extension.
200 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
202 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
203 * config/rs6000/linux.h: Ditto.
204 * alpha/linux.h: Ditto.
205 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
206 no_c99_libc_has_function.
207 * config/c6x/uclinux-elf.h: Ditto.
208 * config/lm32/uclinux-elf.h: Ditto.
209 * config/m68k/uclinux.h: Ditto.
210 * config/moxie/uclinux.h: Ditto.
211 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
212 (crisv32-*-linux*, cris-*-linux*): Ditto.
213 * config/bfin/bfin.c: Include "tm_p.h".
215 2013-09-05 Richard Biener <rguenther@suse.de>
217 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
218 check for a definition without a basic-block.
220 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
221 Sofiane Naci <sofiane.naci@arm.com>
223 * config/aarch64/aarch64.md
224 (*movti_aarch64): Rename r_2_f and f_2_r.
225 (*movsf_aarch64): Likewise.
226 (*movdf_aarch64): Likewise.
227 (*movtf_aarch64): Likewise.
228 (aarch64_movdi_<mode>low): Likewise.
229 (aarch64_movdi_<mode>high): Likewise.
230 (aarch64_mov<mode>high_di): Likewise.
231 (aarch64_mov<mode>low_di): Likewise.
232 (aarch64_movtilow_tilow): Likewise.
233 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
234 values to config/arm/types.md
235 (attribute "conds"): Update for attribute change.
236 (anddi3_insn): Likewise.
237 (iordi3_insn): Likewise.
238 (xordi3_insn): Likewise.
239 (one_cmpldi2): Likewise.
240 * config/arm/types.md (type): Add Neon types.
241 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
242 use "type" attribute.
243 (movmisalign<mode>_neon_store): Likewise.
244 (movmisalign<mode>_neon_load): Likewise.
245 (vec_set<mode>_internal): Likewise.
246 (vec_setv2di_internal): Likewise.
247 (vec_extract<mode>): Likewise.
248 (vec_extractv2di): Likewise.
249 (add<mode>3_neon): Likewise.
250 (adddi3_neon): Likewise.
251 (sub<mode>3_neon): Likewise.
252 (subdi3_neon): Likewise.
253 (mul<mode>3_neon): Likewise.
254 (mul<mode>3add<mode>_neon): Likewise.
255 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
256 (fma<VCVTF:mode>4)): Likewise.
257 (fma<VCVTF:mode>4_intrinsic): Likewise.
258 (fmsub<VCVTF:mode>4)): Likewise.
259 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
260 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
261 (ior<mode>3): Likewise.
262 (and<mode>3): Likewise.
263 (anddi3_neon): Likewise.
264 (orn<mode>3_neon): Likewise.
265 (orndi3_neon): Likewise.
266 (bic<mode>3_neon): Likewise.
267 (bicdi3_neon): Likewise.
268 (xor<mode>3): Likewise.
269 (one_cmpl<mode>2): Likewise.
270 (abs<mode>2): Likewise.
271 (neg<mode>2): Likewise.
272 (umin<mode>3_neon): Likewise.
273 (umax<mode>3_neon): Likewise.
274 (smin<mode>3_neon): Likewise.
275 (smax<mode>3_neon): Likewise.
276 (vashl<mode>3): Likewise.
277 (vashr<mode>3_imm): Likewise.
278 (vlshr<mode>3_imm): Likewise.
279 (ashl<mode>3_signed): Likewise.
280 (ashl<mode>3_unsigned): Likewise.
281 (neon_load_count): Likewise.
282 (ashldi3_neon_noclobber): Likewise.
283 (signed_shift_di3_neon): Likewise.
284 (unsigned_shift_di3_neon): Likewise.
285 (ashrdi3_neon_imm_noclobber): Likewise.
286 (lshrdi3_neon_imm_noclobber): Likewise.
287 (widen_ssum<mode>3): Likewise.
288 (widen_usum<mode>3): Likewise.
289 (quad_halves_<code>v4si): Likewise.
290 (quad_halves_<code>v4sf): Likewise.
291 (quad_halves_<code>v8hi): Likewise.
292 (quad_halves_<code>v16qi): Likewise.
293 (reduc_splus_v2di): Likewise.
294 (neon_vpadd_internal<mode>): Likewise.
295 (neon_vpsmin<mode>): Likewise.
296 (neon_vpsmax<mode>): Likewise.
297 (neon_vpumin<mode>): Likewise.
298 (neon_vpumax<mode>): Likewise.
299 (ss_add<mode>_neon): Likewise.
300 (us_add<mode>_neon): Likewise.
301 (ss_sub<mode>_neon): Likewise.
302 (us_sub<mode>_neon): Likewise.
303 (neon_vadd<mode>_unspec): Likewise.
304 (neon_vaddl<mode>): Likewise.
305 (neon_vaddw<mode>): Likewise.
306 (neon_vhadd<mode>): Likewise.
307 (neon_vqadd<mode>): Likewise.
308 (neon_vaddhn<mode>): Likewise.
309 (neon_vmul<mode>): Likewise.
310 (neon_vmla<mode>): Likewise.
311 (neon_vmlal<mode>): Likewise.
312 (neon_vmls<mode>): Likewise.
313 (neon_vmlsl<mode>): Likewise.
314 (neon_vqdmulh<mode>): Likewise.
315 (neon_vqdmlal<mode>): Likewise.
316 (neon_vqdmlsl<mode>): Likewise.
317 (neon_vmull<mode>): Likewise.
318 (neon_vqdmull<mode>): Likewise.
319 (neon_vsub<mode>_unspec): Likewise.
320 (neon_vsubl<mode>): Likewise.
321 (neon_vsubw<mode>): Likewise.
322 (neon_vqsub<mode>): Likewise.
323 (neon_vhsub<mode>): Likewise.
324 (neon_vsubhn<mode>): Likewise.
325 (neon_vceq<mode>): Likewise.
326 (neon_vcge<mode>): Likewise.
327 (neon_vcgeu<mode>): Likewise.
328 (neon_vcgt<mode>): Likewise.
329 (neon_vcgtu<mode>): Likewise.
330 (neon_vcle<mode>): Likewise.
331 (neon_vclt<mode>): Likewise.
332 (neon_vcage<mode>): Likewise.
333 (neon_vcagt<mode>): Likewise.
334 (neon_vtst<mode>): Likewise.
335 (neon_vabd<mode>): Likewise.
336 (neon_vabdl<mode>): Likewise.
337 (neon_vaba<mode>): Likewise.
338 (neon_vabal<mode>): Likewise.
339 (neon_vmax<mode>): Likewise.
340 (neon_vmin<mode>): Likewise.
341 (neon_vpaddl<mode>): Likewise.
342 (neon_vpadal<mode>): Likewise.
343 (neon_vpmax<mode>): Likewise.
344 (neon_vpmin<mode>): Likewise.
345 (neon_vrecps<mode>): Likewise.
346 (neon_vrsqrts<mode>): Likewise.
347 (neon_vqabs<mode>): Likewise.
348 (neon_vqneg<mode>): Likewise.
349 (neon_vcls<mode>): Likewise.
350 (clz<mode>2): Likewise.
351 (popcount<mode>2): Likewise.
352 (neon_vrecpe): Likewise.
353 (neon_vrsqrte): Likewise.
354 (neon_vget_lane<mode>_sext_internal): Likewise.
355 (neon_vget_lane<mode>_zext_internal): Likewise.
356 (neon_vdup_n<mode>): Likewise.
357 (neon_vdup_nv2di): Likewise.
358 (neon_vdpu_lane<mode>_internal): Likewise.
359 (neon_vswp<mode>): Likewise.
360 (float<mode><V_cvtto>2): Likewise.
361 (floatuns<mode><V_cvtto>2): Likewise.
362 (fix_trunc<mode><V_cvtto>)2): Likewise
363 (fixuns_trunc<mode><V_cvtto)2): Likewise.
364 (neon_vcvt<mode>): Likewise.
365 (neon_vcvtv4sfv4hf): Likewise.
366 (neon_vcvtv4hfv4sf): Likewise.
367 (neon_vcvt_n<mode>): Likewise.
368 (neon_vmovn<mode>): Likewise.
369 (neon_vqmovn<mode>): Likewise.
370 (neon_vqmovun<mode>): Likewise.
371 (neon_vmovl<mode>): Likewise.
372 (neon_vmul_lane<mode>): Likewise.
373 (neon_vmull_lane<mode>): Likewise.
374 (neon_vqdmull_lane<mode>): Likewise.
375 (neon_vqdmulh_lane<mode>): Likewise.
376 (neon_vmla_lane<mode>): Likewise.
377 (neon_vmlal_lane<mode>): Likewise.
378 (neon_vqdmlal_lane<mode>): Likewise.
379 (neon_vmls_lane<mode>): Likewise.
380 (neon_vmlsl_lane<mode>): Likewise.
381 (neon_vqdmlsl_lane<mode>): Likewise.
382 (neon_vext<mode>): Likewise.
383 (neon_vrev64<mode>): Likewise.
384 (neon_vrev32<mode>): Likewise.
385 (neon_vrev16<mode>): Likewise.
386 (neon_vbsl<mode>_internal): Likewise.
387 (neon_vshl<mode>): Likewise.
388 (neon_vqshl<mode>): Likewise.
389 (neon_vshr_n<mode>): Likewise.
390 (neon_vshrn_n<mode>): Likewise.
391 (neon_vqshrn_n<mode>): Likewise.
392 (neon_vqshrun_n<mode>): Likewise.
393 (neon_vshl_n<mode>): Likewise.
394 (neon_vqshl_n<mode>): Likewise.
395 (neon_vqshlu_n<mode>): Likewise.
396 (neon_vshll_n<mode>): Likewise.
397 (neon_vsra_n<mode>): Likewise.
398 (neon_vsri_n<mode>): Likewise.
399 (neon_vsli_n<mode>): Likewise.
400 (neon_vtbl1v8qi): Likewise.
401 (neon_vtbl2v8qi): Likewise.
402 (neon_vtbl3v8qi): Likewise.
403 (neon_vtbl4v8qi): Likewise.
404 (neon_vtbx1v8qi): Likewise.
405 (neon_vtbx2v8qi): Likewise.
406 (neon_vtbx3v8qi): Likewise.
407 (neon_vtbx4v8qi): Likewise.
408 (neon_vtrn<mode>_internal): Likewise.
409 (neon_vzip<mode>_internal): Likewise.
410 (neon_vuzp<mode>_internal): Likewise.
411 (neon_vld1<mode>): Likewise.
412 (neon_vld1_lane<mode>): Likewise.
413 (neon_vld1_dup<mode>): Likewise.
414 (neon_vld1_dupv2di): Likewise.
415 (neon_vst1<mode>): Likewise.
416 (neon_vst1_lane<mode>): Likewise.
417 (neon_vld2<mode>): Likewise.
418 (neon_vld2_lane<mode>): Likewise.
419 (neon_vld2_dup<mode>): Likewise.
420 (neon_vst2<mode>): Likewise.
421 (neon_vst2_lane<mode>): Likewise.
422 (neon_vld3<mode>): Likewise.
423 (neon_vld3qa<mode>): Likewise.
424 (neon_vld3qb<mode>): Likewise.
425 (neon_vld3_lane<mode>): Likewise.
426 (neon_vld3_dup<mode>): Likewise.
427 (neon_vst3<mode>): Likewise.
428 (neon_vst3qa<mode>): Likewise.
429 (neon_vst3qb<mode>): Likewise.
430 (neon_vst3_lane<mode>): Likewise.
431 (neon_vld4<mode>): Likewise.
432 (neon_vld4qa<mode>): Likewise.
433 (neon_vld4qb<mode>): Likewise.
434 (neon_vld4_lane<mode>): Likewise.
435 (neon_vld4_dup<mode>): Likewise.
436 (neon_vst4<mode>): Likewise.
437 (neon_vst4qa<mode>): Likewise.
438 (neon_vst4qb<mode>): Likewise.
439 (neon_vst4_lane<mode>): Likewise.
440 (neon_vec_unpack<US>_lo_<mode>): Likewise.
441 (neon_vec_unpack<US>_hi_<mode>): Likewise.
442 (neon_vec_<US>mult_lo_<mode>): Likewise.
443 (neon_vec_<US>mult_hi_<mode>): Likewise.
444 (neon_vec_<US>shiftl_<mode>): Likewise.
445 (neon_unpack<US>_<mode>): Likewise.
446 (neon_vec_<US>mult_<mode>): Likewise.
447 (vec_pack_trunc_<mode>): Likewise.
448 (neon_vec_pack_trunk_<mode>): Likewise.
449 (neon_vabd<mode>_2): Likewise.
450 (neon_vabd<mode>_3): Likewise.
451 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
452 (thumb2_movsi_vfp): Likewise.
453 (movdi_vfp): Likewise.
454 (movdi_vfp_cortexa8): Likewise.
455 (movhf_vfp_neon): Likewise.
456 (movhf_vfp): Likewiwse.
457 (movsf_vfp): Likewiwse.
458 (thumb2_movsf_vfp): Likewiwse.
459 (movdf_vfp): Likewise.
460 (thumb2_movdf_vfp): Likewise.
461 (movsfcc_vfp): Likewise.
462 (thumb2_movsfcc_vfp): Likewise.
463 (movdfcc_vfp): Likewise.
464 (thumb2_movdfcc_vfp): Likewise.
465 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
466 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
468 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
470 (cortex_a15_neon_int_2): Likewise.
471 (cortex_a15_neon_int_3): Likewise.
472 (cortex_a15_neon_int_4): Likewise.
473 (cortex_a15_neon_int_5): Likewise.
474 (cortex_a15_neon_vqneg_vqabs): Likewise.
475 (cortex_a15_neon_vmov): Likewise.
476 (cortex_a15_neon_vaba): Likewise.
477 (cortex_a15_neon_vaba_qqq): Likewise.
478 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
479 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
480 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
481 scalar_64_32_long_scalar): Likewise.
482 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
483 (cortex_a15_neon_mla_qqq_8_16): Likewise.
484 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
485 lotype_qdd_64_32_long): Likewise.
486 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
487 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
488 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
489 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
490 (cortex_a15_neon_shift_1): Likewise.
491 (cortex_a15_neon_shift_2): Likewise.
492 (cortex_a15_neon_shift_3): Likewise.
493 (cortex_a15_neon_vshl_ddd): Likewise.
494 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
495 (cortex_a15_neon_vsra_vrsra): Likewise.
496 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
497 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
498 (cortex_a15_neon_fp_vmul_ddd): Likewise.
499 (cortex_a15_neon_fp_vmul_qqd): Likewise.
500 (cortex_a15_neon_fp_vmla_ddd): Likewise.
501 (cortex_a15_neon_fp_vmla_qqq): Likewise.
502 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
503 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
504 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
505 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
506 (cortex_a15_neon_bp_simple): Likewise.
507 (cortex_a15_neon_bp_2cycle): Likewise.
508 (cortex_a15_neon_bp_3cycle): Likewise.
509 (cortex_a15_neon_vld1_1_2_regs): Likewise.
510 (cortex_a15_neon_vld1_3_4_regs): Likewise.
511 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
512 (cortex_a15_neon_vld2_4_regs): Likewise.
513 (cortex_a15_neon_vld3_vld4): Likewise.
514 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
515 (cortex_a15_neon_vst1_3_4_regs): Likewise.
516 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
517 (cortex_a15_neon_vst3_vst4): Likewise.
518 (cortex_a15_neon_vld1_vld2_lane): Likewise.
519 (cortex_a15_neon_vld3_vld4_lane" 10
520 (cortex_a15_neon_vst1_vst2_lane): Likewise.
521 (cortex_a15_neon_vst3_vst4_lane): Likewise.
522 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
523 (cortex_a15_neon_ldm_2): Likewise.0
524 (cortex_a15_neon_stm_2): Likewise.
525 (cortex_a15_neon_mcr): Likewise.
526 (cortex_a15_neon_mcr_2_mcrr): Likewise.
527 (cortex_a15_neon_mrc): Likewise.
528 (cortex_a15_neon_mrrc): Likewise.
529 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
531 (cortex_a15_alu_shift): Likewise.
532 (cortex_a15_alu_shift_reg): Likewise.
533 (cortex_a15_mult32): Likewise.
534 (cortex_a15_mult64): Likewise.
535 (cortex_a15_block): Likewise.
536 (cortex_a15_branch): Likewise.
537 (cortex_a15_load1): Likewise.
538 (cortex_a15_load3): Likewise.
539 (cortex_a15_store1): Likewise.
540 (cortex_a15_store3): Likewise.
541 (cortex_a15_call): Likewise.
542 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute
544 (cortex_a5_f2r): Likewise.
545 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
547 (cortex_a53_f2r): Likewise.
548 * config/arm/cortex-a7.md
549 (cortex_a7_branch): Update for attribute change.
550 (cortex_a7_call): Likewise.
551 (cortex_a7_alu_imm): Likewise.
552 (cortex_a7_alu_reg): Likewise.
553 (cortex_a7_alu_shift): Likewise.
554 (cortex_a7_mul): Likewise.
555 (cortex_a7_load1): Likewise.
556 (cortex_a7_store1): Likewise.
557 (cortex_a7_load2): Likewise.
558 (cortex_a7_store2): Likewise.
559 (cortex_a7_load3): Likewise.
560 (cortex_a7_store3): Likewise.
561 (cortex_a7_load4): Likewise.
562 (cortex_a7_store4): Likewise.
563 (cortex_a7_fpalu): Likewise.
564 (cortex_a7_fconst): Likewise.
565 (cortex_a7_fpmuls): Likewise.
566 (cortex_a7_neon_mul): Likewise.
567 (cortex_a7_fpmacs): Likewise.
568 (cortex_a7_neon_mla: Likewise.
569 (cortex_a7_fpmuld: Likewise.
570 (cortex_a7_fpmacd: Likewise.
571 (cortex_a7_fpfmad: Likewise.
572 (cortex_a7_fdivs: Likewise.
573 (cortex_a7_fdivd: Likewise.
574 (cortex_a7_r2f: Likewise.
575 (cortex_a7_f2r: Likewise.
576 (cortex_a7_f_flags: Likewise.
577 (cortex_a7_f_loads: Likewise.
578 (cortex_a7_f_loadd: Likewise.
579 (cortex_a7_f_stores: Likewise.
580 (cortex_a7_f_stored: Likewise.
581 (cortex_a7_neon): Likewise.
582 * config/arm/cortex-a8-neon.md
583 (cortex_a8_neon_mrc): Update for attribute change.
584 (cortex_a8_neon_mrrc): Likewise.
585 (cortex_a8_neon_int_1): Likewise.
586 (cortex_a8_neon_int_2): Likewise.
587 (cortex_a8_neon_int_3): Likewise.
588 (cortex_a8_neon_int_4): Likewise.
589 (cortex_a8_neon_int_5): Likewise.
590 (cortex_a8_neon_vqneg_vqabs): Likewise.
591 (cortex_a8_neon_vmov): Likewise.
592 (cortex_a8_neon_vaba): Likewise.
593 (cortex_a8_neon_vaba_qqq): Likewise.
594 (cortex_a8_neon_vsma): Likewise.
595 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
596 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
597 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
599 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
600 (cortex_a8_neon_mla_qqq_8_16): Likewise.
601 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
602 long_scalar_qdd_64_32_long): Likewise.
603 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
604 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
605 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
606 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
607 (cortex_a8_neon_shift_1): Likewise.
608 (cortex_a8_neon_shift_2): Likewise.
609 (cortex_a8_neon_shift_3): Likewise.
610 (cortex_a8_neon_vshl_ddd): Likewise.
611 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
612 (cortex_a8_neon_vsra_vrsra): Likewise.
613 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
614 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
615 (cortex_a8_neon_fp_vsum): Likewise.
616 (cortex_a8_neon_fp_vmul_ddd): Likewise.
617 (cortex_a8_neon_fp_vmul_qqd): Likewise.
618 (cortex_a8_neon_fp_vmla_ddd): Likewise.
619 (cortex_a8_neon_fp_vmla_qqq): Likewise.
620 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
621 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
622 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
623 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
624 (cortex_a8_neon_bp_simple): Likewise.
625 (cortex_a8_neon_bp_2cycle): Likewise.
626 (cortex_a8_neon_bp_3cycle): Likewise.
627 (cortex_a8_neon_ldr): Likewise.
628 (cortex_a8_neon_str): Likewise.
629 (cortex_a8_neon_vld1_1_2_regs): Likewise.
630 (cortex_a8_neon_vld1_3_4_regs): Likewise.
631 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
632 (cortex_a8_neon_vld2_4_regs): Likewise.
633 (cortex_a8_neon_vld3_vld4): Likewise.
634 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
635 (cortex_a8_neon_vst1_3_4_regs): Likewise.
636 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
637 (cortex_a8_neon_vst3_vst4): Likewise.
638 (cortex_a8_neon_vld1_vld2_lane): Likewise.
639 (cortex_a8_neon_vld3_vld4_lane): Likewise.
640 (cortex_a8_neon_vst1_vst2_lane): Likewise.
641 (cortex_a8_neon_vst3_vst4_lane): Likewise.
642 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
643 (cortex_a8_neon_mcr): Likewise.
644 (cortex_a8_neon_mcr_2_mcrr): Likewise.
645 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
647 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
649 (ca9_neon_mrrc): Likewise.
650 (cortex_a9_neon_int_1): Likewise.
651 (cortex_a9_neon_int_2): Likewise.
652 (cortex_a9_neon_int_3): Likewise.
653 (cortex_a9_neon_int_4): Likewise.
654 (cortex_a9_neon_int_5): Likewise.
655 (cortex_a9_neon_vqneg_vqabs): Likewise.
656 (cortex_a9_neon_vmov): Likewise.
657 (cortex_a9_neon_vaba): Likewise.
658 (cortex_a9_neon_vaba_qqq): Likewise.
659 (cortex_a9_neon_vsma): Likewise.
660 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
661 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
662 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
664 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
665 (cortex_a9_neon_mla_qqq_8_16): Likewise.
666 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
667 long_scalar_qdd_64_32_long): Likewise.
668 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
669 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
670 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
671 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
672 (cortex_a9_neon_shift_1): Likewise.
673 (cortex_a9_neon_shift_2): Likewise.
674 (cortex_a9_neon_shift_3): Likewise.
675 (cortex_a9_neon_vshl_ddd): Likewise.
676 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
677 (cortex_a9_neon_vsra_vrsra): Likewise.
678 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
679 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
680 (cortex_a9_neon_fp_vsum): Likewise.
681 (cortex_a9_neon_fp_vmul_ddd): Likewise.
682 (cortex_a9_neon_fp_vmul_qqd): Likewise.
683 (cortex_a9_neon_fp_vmla_ddd): Likewise.
684 (cortex_a9_neon_fp_vmla_qqq): Likewise.
685 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
686 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
687 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
688 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
689 (cortex_a9_neon_bp_simple): Likewise.
690 (cortex_a9_neon_bp_2cycle): Likewise.
691 (cortex_a9_neon_bp_3cycle): Likewise.
692 (cortex_a9_neon_ldr): Likewise.
693 (cortex_a9_neon_str): Likewise.
694 (cortex_a9_neon_vld1_1_2_regs): Likewise.
695 (cortex_a9_neon_vld1_3_4_regs): Likewise.
696 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
697 (cortex_a9_neon_vld2_4_regs): Likewise.
698 (cortex_a9_neon_vld3_vld4): Likewise.
699 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
700 (cortex_a9_neon_vst1_3_4_regs): Likewise.
701 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
702 (cortex_a9_neon_vst3_vst4): Likewise.
703 (cortex_a9_neon_vld1_vld2_lane): Likewise.
704 (cortex_a9_neon_vld3_vld4_lane): Likewise.
705 (cortex_a9_neon_vst1_vst2_lane): Likewise.
706 (cortex_a9_neon_vst3_vst4_lane): Likewise.
707 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
708 (cortex_a9_neon_mcr): Likewise.
709 (cortex_a9_neon_mcr_2_mcrr): Likewise.
710 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
711 (cortex_a9_fps): Likewise.
712 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
714 (cortex_m4_fmuls): Likewise.
715 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
717 (cortex_r4_mrc): Likewise.
718 * config/arm/iterators.md: Update comment referring to neon_type.
719 * config/arm/iwmmxt.md
720 (iwmmxt_arm_movdi): Update for attribute change.
721 (iwmmxt_movsi_insn): Likewise.
722 * config/arm/marvell-pj4.md
723 (pj4_vfp_to_core): Update for attribute change.
724 (pj4_core_to_vfp): Likewise.
725 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
727 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
728 (vfp_fstore): Likewise.
729 * doc/md.texi: Change references to neon_type to refer to type.
731 2013-09-04 Dodji Seketeli <dodji@redhat.com>
733 * tree.h (DECL_BUILT_IN): Fix typo in comment.
735 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
737 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
740 2013-09-04 Easwaran Raman <eraman@google.com>
743 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
744 (build_and_add_sum): Use it.
745 (appears_later_in_bb): Simplify code.
747 2013-09-04 Teresa Johnson <tejohnson@google.com>
749 * dumpfile.c (dump_finish): Don't close stderr/stdout.
751 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
753 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
755 2013-09-04 Jan Hubicka <jh@suse.cz>
757 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
758 * ipa-devirt.c: Include diganostic.h
759 (odr_type_d): Add types and types_set.
760 (hash_type_name): Work for types with vtables during LTO.
761 (odr_hasher::remove): Fix comment; destroy types_set.
762 (add_type_duplicate): New function,
763 (get_odr_type): Use it.
764 (dump_type_inheritance_graph): Dump type duplicates.
765 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
767 * tree.c (types_same_for_odr): Give exact answers on types with
770 2013-09-04 Dodji Seketeli <dodji@redhat.com>
772 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
773 explaining their differences.
775 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
777 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
779 2013-09-03 Jeff Law <law@redhat.com>
781 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
782 when not threading through a joiner block. Pass joiner/no joiner
783 state to register_jump_thread.
784 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
785 state from argument rather than implying on path length.
786 Dump the entire jump thread path into debugging dump.
787 * tree-flow.h (register_jump_thread): Update prototype.
789 2013-08-29 Xinliang David Li <davidxl@google.com>
791 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
792 Remove a trivial gcc_assert.
794 2013-08-29 Xinliang David Li <davidxl@google.com>
796 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
797 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
798 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
799 Delay base decl alignment adjustment.
800 * tree-vectorizer.c (vect_destroy_datarefs): New function.
801 * tree-vectorizer.h: New data structure.
802 (set_dr_misalignment): New function.
803 (dr_misalignment): Ditto.
804 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
805 (vectorizable_load): Ditto.
806 (ensure_base_align): New function.
807 (vectorize_loops): Add dbg_cnt support.
808 (execute_vect_slp): Ditto.
809 * dbgcnt.def: New debug counter.
810 * Makefile: New dependency.
812 2013-09-03 Meador Inge <meadori@codesourcery.com>
816 2013-08-30 Meador Inge <meadori@codesourcery.com>
818 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
820 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
822 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
825 2013-09-03 Richard Biener <rguenther@suse.de>
827 * tree-affine.c (add_elt_to_tree): Fix association issue,
828 avoid useless converts and make sure to always return a
829 properly typed result.
831 2013-09-03 Richard Biener <rguenther@suse.de>
834 * fold-const.c (negate_expr_p): Fix division case.
835 (negate_expr): Likewise.
837 2013-09-03 Richard Biener <rguenther@suse.de>
840 * tree-streamer-out.c: Include tm.h.
841 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
843 2013-09-03 Jan Hubicka <jh@suse.cz>
845 * tree-profile.c (tree_profiling): Cleanup CFG when done.
847 2013-09-03 Alan Modra <amodra@gmail.com>
849 * config.gcc (powerpc*-*-linux*): Add support for little-endian
850 multilibs to big-endian target and vice versa.
851 * config/rs6000/t-linux64: Use := assignment on all vars.
852 (MULTILIB_EXTRA_OPTS): Remove fPIC.
853 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
854 * config/rs6000/t-linux64le: New file.
855 * config/rs6000/t-linux64bele: New file.
856 * config/rs6000/t-linux64lebe: New file.
858 2013-09-02 Jan Hubicka <jh@suse.cz>
860 * ipa-inline-transform.c (inline_transform): Do not
861 optimize_inline_calls when not optimizing.
863 2013-09-02 Jan Hubicka <jh@suse.cz>
865 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
866 duplicated nodes for assembler names.
867 * symtab.c (symtab_unregister_node): Do not attempt to unlink
868 hard registers from assembler name hash.
870 2013-09-02 Jan Hubicka <jh@suse.cz>
872 * ipa-split.c (execute_split_functions): Split externally visible
873 functions called once.
875 2013-09-02 Martin Jambor <mjambor@suse.cz>
878 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
879 linked list. When finding the correct duplicate, also consider also
880 the caller in additon to its inlined_to node.
882 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
884 * config/aarch64/aarch64-simd-builtins.def
885 (dup_lane_scalar): Remove.
886 * config/aarch64/aarch64-simd.md
887 (aarch64_simd_dup): Add 'w->w' alternative.
888 (aarch64_dup_lane<mode>): Allow for VALL.
889 (aarch64_dup_lane_scalar<mode>): Remove.
890 (aarch64_dup_lane_<vswap_width_name><mode>): New.
891 (aarch64_get_lane_signed<mode>): Add w->w altenative.
892 (aarch64_get_lane_unsigned<mode>): Likewise.
893 (aarch64_get_lane<mode>): Likewise.
894 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
895 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
896 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
897 (VCON): Change container of V2SF.
898 (vswap_width_name): Likewise.
899 * config/aarch64/arm_neon.h
900 (__aarch64_vdup_lane_any): New.
901 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
902 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
903 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
905 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
908 * expr.c (emit_move_complex): Do not move complex FP values as parts if
909 the source or the destination is a single hard register.
911 2013-09-02 Richard Biener <rguenther@suse.de>
914 * tree-scalar-evolution.c (instantiate_scev_name): Allow
917 2013-09-02 Richard Biener <rguenther@suse.de>
919 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
920 arithmetic to sizetype.
922 2013-09-02 Bin Cheng <bin.cheng@arm.com>
924 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
925 Find auto-increment use both before and after candidate.
927 2013-09-02 Marek Polacek <polacek@redhat.com>
929 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
931 2013-09-01 Jan Hubicka <jh@suse.cz>
933 * Makefile.in: Add ipa-profile.o
934 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
935 * cgraph.c (struct cgraph_propagate_frequency_data,
936 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
937 ipa-profile.c; replace cgraph_ by ipa_ prefix.
938 * cgraph.h (cgraph_propagate_frequency): Remove.
939 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
940 (inline_update_callee_summaries): Update.
941 * ipa-profile.c: New file.
942 * ipa-utils.h (ipa_propagate_frequency): Declare.
943 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
944 data-streamer.h, value-prof.h
945 (symtab_remove_unreachable_nodes): Update profile.
946 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
947 account_time_size, cmp_counts, dump_histogram,
948 ipa_profile_generate_summary, ipa_profile_write_summary,
949 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
950 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
951 Move to ipa-profile.c
953 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
955 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
957 2013-09-01 Jan Hubicka <jh@suse.cz>
959 * common.opt (fdevirtualize-speculatively): New function.
960 * invoke.texi (fdevirtualize-speculatively): Document.
961 * ipa-devirt.c: Include ipa-inline.h
962 (likely_target_p): New function.
963 (ipa_devirt): New function.
964 (gate_ipa_devirt): New function.
965 (pass_data_ipa_devirt): New static var.
966 (pass_ipa_devirt): Likewise.
967 (make_pass_ipa_devirt): New function.
968 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
969 (common_handle_option): Disable devirtualization when
970 value range profiling is available.
971 * passes.def (pass_ipa_devirt): Add.
972 * timever.def (TV_IPA_DEVIRT): New timevar.
973 * tree-pass.h (make_pass_ipa_devirt):
975 2013-09-01 Iain Sandoe <iain@codesourcery.com>
977 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
978 include sanitize(undefined).
980 2013-08-31 Diego Novillo <dnovillo@google.com>
982 * Makefile.in (TREE_CORE_H): Define.
984 (GTFILES): Add tree-core.h.
985 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
987 * tree-core.h: New file.
988 Move all data structures, enum, typedefs, global
989 declarations and constants from ...
992 2013-08-31 Jan Hubicka <jh@suse.cz>
994 * bulitins.c (expand_builtin): Do not early exit for gcov
995 instrumented functions.
997 2013-08-31 Marek Polacek <polacek@redhat.com>
999 * ubsan.c: Include tm_p.h.
1001 2013-08-31 Jan Hubicka <jh@suse.cz>
1003 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
1006 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
1007 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
1008 * tree-cfg.c (verify_gimple_label): ... here.
1009 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
1010 (ipa_merge_profiles): New function.
1011 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
1012 (lto_input_function_body): Likewise.
1013 * ipa-utils.h (ipa_merge_profiles): Declare.
1014 * lto-streamer.h (lto_input_function_body): Update prototype.
1015 (emit_label_in_global_context_p): Remove.
1016 * lto-symtab.c: Include ipa-utils.h
1017 (lto_cgraph_replace_node): Use ipa_merge_profiles.
1019 2013-08-31 Jan Hubicka <jh@suse.cz>
1021 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
1023 2013-08-31 Jan Hubicka <jh@suse.cz>
1025 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
1027 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
1029 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
1030 "cmp" RTX before signed_comparison_operator check to account
1033 2013-08-30 Jan Hubicka <jh@suse.cz>
1035 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
1036 (detect_type_change_1): Rename to ...
1037 (detect_type_change): ... this one; early return on non-polymorphic
1039 (detect_type_change_ssa): Add comp_type parameter; update
1040 use of detect_type_change.
1041 (compute_complex_assign_jump_func): Add param_type parameter;
1042 update use of detect_type_change_ssa.
1043 (compute_complex_ancestor_jump_func): Likewise.
1044 (ipa_get_callee_param_type): New function.
1045 (ipa_compute_jump_functions_for_edge): Compute parameter type;
1046 update calls to the jump function computation functions.
1048 2013-08-30 Teresa Johnson <tejohnson@google.com>
1049 Steven Bosscher <steven@gcc.gnu.org>
1051 * cfgrtl.c (fixup_new_cold_bb): New routine.
1052 (commit_edge_insertions): Invoke fixup_partitions.
1053 (find_partition_fixes): New routine.
1054 (fixup_partitions): Ditto.
1055 (verify_hot_cold_block_grouping): Update comments.
1056 (rtl_verify_edges): Invoke find_partition_fixes.
1057 (rtl_verify_bb_pointers): Update comments.
1058 (rtl_verify_bb_layout): Ditto.
1059 * basic-block.h (probably_never_executed_edge_p): Declare.
1060 (fixup_partitions): Ditto.
1061 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
1062 * bb-reorder.c (sanitize_hot_paths): New function.
1063 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
1065 * predict.c (probably_never_executed_edge_p): New routine.
1066 * cfg.c (check_bb_profile): Add partition insanity warnings.
1068 2013-08-30 Meador Inge <meadori@codesourcery.com>
1070 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1072 2013-08-30 Marek Polacek <polacek@redhat.com>
1074 * Makefile.in (ubsan.o): Add.
1075 (c-family/c-ubsan.o): Add.
1076 (builtins.o): Add ubsan.h dependency.
1077 * ubsan.h: New file.
1078 * ubsan.c: New file.
1079 * common.opt: Add -fsanitize=undefined option.
1080 (flag_sanitize): Add variable.
1081 (fsanitize=): Add option. Add Driver.
1082 (fsanitize=thread): Remove option.
1083 (fsanitize=address): Likewise.
1084 (static-libubsan): New option.
1085 * doc/invoke.texi: Document the new flag and -static-libubsan.
1086 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
1087 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
1088 * builtin-attrs.def (ATTR_COLD): Define.
1089 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1090 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
1091 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
1092 * flag-types.h (sanitize_code): New enum.
1093 * opts.c (common_handle_option): Parse command line arguments
1094 of -fsanitize=. Add -fsanitize=unreachable option.
1095 * varasm.c (get_variable_section): Adjust.
1096 (assemble_noswitch_variable): Likewise.
1097 (assemble_variable): Likewise.
1098 (output_constant_def_contents): Likewise.
1099 (categorize_decl_for_section): Likewise.
1100 (place_block_symbol): Likewise.
1101 (output_object_block): Likewise.
1102 * builtins.def: Likewise.
1103 * toplev.c (compile_file): Likewise.
1104 (process_options): Likewise.
1105 * cppbuiltin.c: Likewise.
1106 * tsan.c (tsan_pass): Likewise.
1107 (tsan_gate): Likewise.
1108 (tsan_gate_O0): Likewise.
1109 * cfgexpand.c (partition_stack_vars): Likewise.
1110 (expand_stack_vars): Likewise.
1111 (defer_stack_allocation): Likewise.
1112 (expand_used_vars): Likewise.
1113 * cfgcleanup.c (old_insns_match_p): Likewise.
1114 * asan.c (asan_finish_file): Likewise.
1115 (asan_instrument): Likewise.
1116 (gate_asan): Likewise.
1117 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
1118 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1119 (asan_global_struct): Use pointer_sized_int_node instead
1120 calling build_nonstandard_integer_type.
1121 (initialize_sanitizer_builtins): Likewise.
1122 (asan_finish_file): Likewise.
1123 * gcc.c: Document %{%:function(args):X}.
1124 (static_spec_functions): Add sanitize.
1125 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
1126 store funcval != NULL there.
1127 (do_spec_1): Adjust handle_spec_function caller.
1128 (handle_braces): Allow %:function(args) as condition.
1129 (sanitize_spec_function): New function.
1130 (ADD_STATIC_LIBUBSAN_LIBS): Define.
1131 (LIBUBSAN_SPEC): Likewise.
1132 (LIBUBSAN_EARLY_SPEC): Likewise.
1133 (SANITIZER_SPEC): Handle libubsan.
1134 (SANITIZER_EARLY_SPEC): Likewise.
1135 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
1136 instead of fsanitize=address.
1137 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
1138 instead of fsanitize=address*.
1139 * builtins.c: Include ubsan.h.
1140 (fold_builtin_0): Instrument __builtin_unreachable.
1141 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
1142 instead of flag_asan.
1143 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
1144 (pointer_sized_int_node): Define.
1145 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
1147 2013-08-30 Mike Stump <mikestump@comcast.net>
1149 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
1152 2013-08-29 Jan Hubicka <jh@suse.cz>
1154 * cgraph.c (cgraph_function_body_availability): Handle weakref
1156 * passes.def: Remove pass_fixup_cfg.
1157 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
1158 track when we need to remove functions.
1159 (gate_ipa_inline): Execute inlining always; add comment why.
1160 (pass_data_ipa_inline): Remove TODO_remove_functions.
1161 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
1162 do not produce summaries.
1163 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
1164 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
1165 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
1166 constant pool and vtable.
1168 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
1170 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
1171 New arm_neon.h's internal macros to specify 64-bit constants.
1172 Avoid using stdint.h's macros.
1174 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
1176 * recog.c (verify_changes): Verify that changes[i].old is non-zero
1177 before applying REG_P.
1179 2013-08-30 Jakub Jelinek <jakub@redhat.com>
1181 PR tree-optimization/58277
1182 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
1183 after seeing too many stmts with vdef in between dombb and current
1184 bb, invalidate everything.
1186 2013-08-30 Richard Biener <rguenther@suse.de>
1188 * fold-const.c (fold_single_bit_test): Fix overflow test.
1190 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
1192 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
1193 and which can live in a register, always retrieve the value on entry.
1194 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
1195 passed by invisible reference specially.
1196 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
1197 (vt_add_function_parameter): Correctly deal with a parameter passed by
1198 invisible reference.
1200 2013-08-30 Jan Hubicka <jh@suse.cz>
1202 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
1204 2013-08-30 Richard Biener <rguenther@suse.de>
1206 PR tree-optimization/58228
1207 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
1208 allow invariant loads in nested loop vectorization.
1210 2013-08-30 Richard Biener <rguenther@suse.de>
1212 PR tree-optimization/58223
1213 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
1214 (has_anti_or_output_dependence): ... this and adjust to also
1215 look for output dependences.
1216 (mark_nodes_having_upstream_mem_writes): Adjust.
1217 (rdg_flag_uses): Likewise.
1219 2013-08-30 Richard Biener <rguenther@suse.de>
1221 PR tree-optimization/58010
1222 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
1223 assert that we have a loop-closed PHI.
1225 2013-08-29 Jan Hubicka <jh@suse.cz>
1227 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
1228 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
1229 * lto-section-in.c (lto_free_function_in_decl_state): New function.
1230 (lto_free_function_in_decl_state_for_node): New function.
1232 2013-08-29 Xinliang David Li <davidxl@google.com>
1234 * loop-unroll.c (report_unroll_peel): Minor message change.
1235 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
1236 Emit alignment peeling message with default -fopt-info.
1237 (vect_loop_versioning): Emit loop version info message.
1238 * tree-vectorizer.c (vectorize_loops): Minor message change.
1239 (execute_vect_slp): Ditto.
1241 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
1243 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
1244 of the clone from the DECL_NAME of the original function.
1246 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
1248 * passes.c (register_pass): Add overload.
1249 * tree-pass.h (register_pass): Forward declare it. Add comment.
1251 2013-08-29 Jan Hubicka <jh@suse.cz>
1253 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
1254 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
1255 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
1256 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
1257 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
1258 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
1259 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
1260 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
1262 2013-08-29 Teresa Johnson <tejohnson@google.com>
1264 * dumpfile.c (dump_loc): Output column number.
1265 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
1266 * doc/invoke.texi: Document optall -fopt-info flag.
1267 * profile.c (read_profile_edge_counts): Use new dump framework.
1268 (compute_branch_probabilities): Ditto.
1269 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
1270 when pass not in any opt group.
1271 * pass_manager.h (pass_manager::get_pass_profile): New method.
1272 * value-prof.c (check_counter): Use new dump framework.
1273 (check_ic_target): Ditto.
1274 * coverage.c (get_coverage_counts): Ditto.
1275 (coverage_init): Setup new dump framework.
1277 2013-08-29 Richard Biener <rguenther@suse.de>
1279 PR tree-optimization/58246
1280 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
1281 handle the dominance check inside a basic-block.
1283 2013-08-29 Richard Biener <rguenther@suse.de>
1286 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
1287 of default defs that appear in abnormal PHI nodes.
1289 2013-08-29 Richard Biener <rguenther@suse.de>
1291 PR tree-optimization/57685
1292 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
1293 single-use operands to avoid exponential complexity.
1295 2013-08-28 Dehao Chen <dehao@google.com>
1297 * ipa-inline.c (edge_badness): Fix integer underflow.
1299 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
1301 * gtm-builtins.def (_ITM_free): Declare leaf.
1303 2013-08-28 Jakub Jelinek <jakub@redhat.com>
1306 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
1307 (*tls_local_dynamic_base_64_largepic): Likewise.
1308 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
1309 Remove predicate from call operand.
1310 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
1311 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
1313 2013-08-28 Jeff Law <law@redhat.com>
1315 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
1316 checks for the number of predecessors and successors allowed.
1317 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
1318 which require copying a joiner block if there is a request which
1319 is a subpath that requires no joiner block copying.
1321 2013-08-28 Jan Hubicka <jh@suse.cz>
1323 * lto-streamer-out.c (DFS_write_tree_body): Drop
1324 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
1325 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
1326 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
1327 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
1328 Do not read DECL_ERROR_ISSUED.
1329 (unpack_ts_decl_with_vis_value_fields): Do not read
1331 (lto_input_ts_binfo_tree_pointers): Do not read
1332 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
1333 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
1334 write DECL_ERROR_ISSUED..
1335 (pack_ts_decl_with_vis_value_fields): Do not write
1337 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
1338 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
1339 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
1340 * tree.h (tree_decl_common): Update comment.
1341 (DECL_ERROR_ISSUED): Remove.
1343 2013-08-28 Jakub Jelinek <jakub@redhat.com>
1346 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
1348 2013-08-28 Jan Hubicka <jh@suse.cz>
1350 * builtins.def (free): Declare leaf.
1352 2013-08-27 David Malcolm <dmalcolm@redhat.com>
1355 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
1356 * configure: Regenerate.
1358 2013-08-27 Martin Jambor <mjambor@suse.cz>
1360 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
1361 (ipa_ancestor_jf_data): Likewise.
1362 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
1363 (ipa_get_jf_pass_through_type_preserved): New function.
1364 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
1365 (ipa_get_jf_ancestor_type_preserved): New function.
1366 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
1367 (ipa_get_jf_ancestor_result): Likewise.
1368 (propagate_vals_accross_pass_through): Use
1369 ipa_get_jf_pass_through_result to do all the value mappings.
1370 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
1371 type_preserved flag.
1372 (ipa_set_jf_cst_copy): New function.
1373 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
1374 (ipa_set_jf_arith_pass_through): Likewise.
1375 (ipa_set_ancestor_jf): Likewise.
1376 (compute_complex_assign_jump_func): Set type_preserved instead of
1378 (ipa_compute_jump_functions_for_edge): Likewise.
1379 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
1380 (update_jump_functions_after_inlining): Update type_preserved.
1381 Explicitely create jump functions when combining one with pass_through.
1382 (ipa_write_jump_function): Stream the type_preserved flags.
1383 (ipa_read_jump_function): Likewise.
1385 2013-08-27 Jakub Jelinek <jakub@redhat.com>
1386 Aldy Hernandez <aldyh@redhat.com>
1388 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
1389 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
1390 * function.h (struct function): Add has_force_vect_loops and
1392 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
1393 * gimple.c (gimple_build_omp_critical): Add KIND argument and
1395 * gimple.def: Update CLAUSES comments.
1396 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
1397 (gimple_build_omp_for): Add argument to prototype.
1398 (gimple_omp_for_kind): New.
1399 (gimple_omp_for_set_kind): New.
1400 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
1401 GOVD_DATA_SHARE_CLASS.
1402 (enum omp_region_type): Add ORT_SIMD.
1403 (gimple_add_tmp_var): Handle ORT_SIMD.
1404 (gimplify_var_or_parm_decl): Same.
1405 (is_gimple_stmt): Same.
1406 (omp_firstprivatize_variable): Same.
1407 (omp_add_variable): Only use splay_tree_insert if lookup failed.
1408 (omp_notice_variable): Handle ORT_SIMD.
1409 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
1410 (omp_check_private): Handle ORT_SIMD.
1411 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1413 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
1414 Handle OMP_CLAUSE_LASTPRIVATE.
1415 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1417 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
1418 (gimplify_expr): Handle OMP_SIMD.
1419 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
1420 (expand_GOMP_SIMD_VF): New.
1421 (expand_GOMP_SIMD_LAST_LANE): New.
1422 * internal-fn.def (GOMP_SIMD_LANE): New.
1423 (GOMP_SIMD_VF): New.
1424 (GOMP_SIMD_LAST_LANE): New.
1425 * omp-low.c: Include target.h.
1426 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
1428 (check_omp_nesting_restrictions): Same.
1430 (lower_rec_simd_input_clauses): New.
1431 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
1433 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
1434 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
1435 (expand_omp_build_assign): New.
1436 (expand_omp_for_init_counts): New.
1437 (expand_omp_for_init_vars): New.
1438 (extract_omp_for_update_vars): New.
1439 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
1440 and rewrite accordingly.
1441 (expand_omp_simd): New.
1442 (expand_omp_for): Use expand_omp_simd.
1443 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
1444 (lower_omp_for): Do not lower the body.
1445 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
1447 * tree-flow.h (find_omp_clause): Remove prototype.
1448 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
1449 forcing vectorization of the loop, or if flag_tree_vectorize.
1450 (gate_tree_if_conversion): Similarly.
1451 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
1452 gimple_build_omp_for.
1453 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
1454 * tree-parloops (create_parallel_loop): Pass kind argument to
1455 gimple_build_omp_for.
1456 * tree-pretty-print.c (dump_omp_clause): Add cases for
1457 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
1458 OMP_CLAUSE__SIMDUID_.
1459 (dump_generic_node): Handle OMP_SIMD.
1460 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
1461 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
1462 unroll OMP_SIMD loops here.
1463 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
1464 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
1466 (vect_analyze_data_refs): Handle simd loops.
1467 * tree-vect-loop.c (vectorizable_live_operation): Handle
1469 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
1470 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
1471 (vectorizable_load): Same.
1472 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
1473 (struct simduid_to_vf): New.
1474 (simduid_to_vf::hash): New.
1475 (simduid_to-vf::equal): New.
1476 (struct simd_array_to_simduid): New.
1477 (simd_array_to_simduid::hash): New.
1478 (simd_array_to_simduid::equal): New.
1479 (adjust_simduid_builtins): New.
1480 (struct note_simd_array_uses_struct): New.
1481 (note_simd_array_uses_cb): New.
1482 (note_simd_array_uses): New.
1483 (vectorize_loops): Handle simd hints and adjust simd builtins
1485 * tree-vectorizer.h (struct _stmt_vec_info): Add
1486 simd_lane_access_p field.
1487 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
1488 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
1489 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
1490 (omp_clause_code_name): Same.
1491 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
1492 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
1493 * tree.def (OMP_SIMD): New entry.
1494 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
1495 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
1496 (OMP_CLAUSE_DECL): Adjust range for new clauses.
1497 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
1498 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
1499 (OMP_CLAUSE_LINEAR_STEP): New.
1500 (OMP_CLAUSE_SAFELEN_EXPR): New.
1501 (OMP_CLAUSE__SIMDUID__DECL): New.
1502 (find_omp_clause): New prototype.
1504 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
1506 * config/i386/driver-i386.c (host_detect_local_cpu): Update
1507 Haswell processor detection.
1509 2013-08-27 Christian Widmer <shadow@umbrox.de>
1512 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
1513 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
1514 AVX2 capable processors.
1516 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
1518 * config/aarch64/arm_neon.h: Replace all inline asm implementations
1519 of vget_low_* with implementations in terms of other intrinsics.
1521 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
1524 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
1525 values to -1, 0 and 1.
1527 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
1529 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
1530 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
1531 (<optab><mode>3_insn): Remove.
1532 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
1533 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
1534 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
1535 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
1536 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
1537 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
1538 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
1539 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
1540 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
1541 (ror<mode>3_insn): Likewise.
1542 * config/aarch64/predicates.md (aarch64_simd_register): New.
1544 2013-08-27 Richard Biener <rguenther@suse.de>
1546 PR tree-optimization/57521
1547 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
1548 one edge is non-critical.
1549 (find_phi_replacement_condition): Make sure to use a non-critical
1550 edge. Cleanup and remove old bug workarounds.
1551 (bb_postdominates_preds): Remove.
1552 (if_convertible_loop_p_1): Do not compute post-dominators.
1553 (combine_blocks): Do not free post-dominators.
1554 (main_tree_if_conversion): Likewise.
1555 (pass_data_if_conversion): Add TODO_verify_ssa.
1557 2013-08-27 DJ Delorie <dj@redhat.com>
1559 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
1561 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
1563 * function.c (assign_parm_find_data_types): Set passed_mode and
1564 nominal_mode to the TYPE_MODE of nominal_type for the built
1565 pointer type in case of the struct-pass-by-reference.
1567 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1569 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
1570 (UINT16_TYPE): Change default to "unsigned int".
1572 * config/avr/avr.opt (mfract-convert-truncate): New option.
1573 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
1574 is set, round negative fractional integers according to n1169
1575 when converting to integer types.
1577 2013-08-26 Jan Hubicka <jh@suse.cz>
1579 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
1580 methods can not be called indirectly when their address is not taken.
1582 2013-08-26 Jan Hubicka <jh@suse.cz>
1584 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
1587 2013-08-26 Jan Hubicka <jh@suse.cz>
1589 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
1592 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1594 * reload.c (find_valid_class): Allow classes that do not include
1595 FIRST_PSEUDO_REGISTER - 1.
1597 2013-08-26 Jan Hubicka <jh@suse.cz>
1599 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
1600 fix edge count/frequency when speculation failed; fix type check
1601 for the direct call.
1603 2013-08-26 Jan Hubicka <jh@suse.cz>
1605 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
1607 2013-08-26 Jan Hubicka <jh@suse.cz>
1609 * ipa-inline-transform.c (inline_transform): Be ready for basic block
1610 to be changed by edge redirection.
1612 2013-08-26 Jan Hubicka <jh@suse.cz>
1614 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
1615 formating; add sanity check.
1616 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
1617 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
1618 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
1619 (resolve_noninline_speculation): Update callee keys, too.
1621 2013-08-26 Jan Hubicka <jh@suse.cz>
1623 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
1624 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
1626 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1628 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
1631 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
1633 * config/i386/i386.c (ix86_debug_options): Remove prototype.
1634 (x86_64_elf_select_section): Ditto.
1635 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
1637 (ix86_pass_by_reference): Ditto.
1638 (output_set_got): Ditto.
1639 (ix86_unary_operator_ok): Ditto.
1640 (ix86_expand_builtin): Ditto.
1642 2013-08-23 Jan Hubicka <jh@suse.cz>
1644 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
1646 2013-08-23 Jan Hubicka <jh@suse.cz>
1648 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
1649 (tree_decl_with_vis): Add FINAL field.
1651 2013-08-23 Jeff Law <law@redhat.com>
1653 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
1654 the debugging dump when the expression is fully redundant.
1656 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
1658 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
1659 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
1660 * pretty-print.c (pp_formatted_text_data): Likewise.
1661 (pp_write_text_to_stream): Likewise.
1662 (pp_write_text_as_dot_label_to_stream): Likewise.
1663 (pp_append_r): Likewise.
1664 (pp_format): Likewise.
1665 (pp_flush): Likewise.
1666 (pp_clear_output_area): Likewise.
1667 (pp_append_text): Likewise.
1668 (pp_formatted_text): Likewise.
1669 (pp_remaining_character_count_for_line): Likewise.
1670 (pp_newline): Likewise.
1671 (pp_character): Likewise.
1672 (output_buffer::~output_buffer): Define.
1673 (pretty_printer::~pretty_printer): Destruct output buffer.
1674 * pretty-print.h (output_buffer::~output_buffer): Declare.
1675 (pretty_printer::~pretty_printer): Declare virtual.
1677 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
1680 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
1681 HOST_WIDE_INT_M1U): New macros.
1682 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
1683 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
1684 unsigned -1 for lshift.
1685 * cse.c (cse_insn): Likewise.
1686 * double-int.c (rshift_double, lshift_double): Likewise.
1687 * builtins.c (fold_builtin_bitop): Likewise.
1688 * combine.c (force_to_mode): Likewise.
1689 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
1690 * simplify-rtx.c (simplify_const_unary_operation,
1691 simplify_const_binary_operation): Likewise.
1692 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
1693 check_va_list_escapes): Likewise.
1694 * rtlanal.c (nonzero_bits1): Likewise.
1695 * expmed.c (expand_smod_pow2): Likewise.
1696 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
1698 2013-08-23 Jan Hubicka <jh@suse.cz>
1700 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
1701 as having address taken.
1703 2013-08-23 Jan Hubicka <jh@suse.cz>
1705 * ipa-utils.h (method_class_type): Declare.
1706 * ipa-devirt.c (method_class_type): Export.
1708 * cgraphunit.c (analyze_functions): Do basic devirtualization;
1709 do not walk base classes of anonymous types.
1711 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
1713 PR rtl-optimization/58220
1715 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
1716 handle SEQUENCE insns properly.
1718 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
1720 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
1722 (pp_newline_and_indent): Likewise.
1723 (pp_separate_with): Likewise.
1724 * pretty-print.c (pp_newline_and_flush): Define.
1725 (pp_newline_and_indent): Likewise.
1726 (pp_separate_with): Likewise.
1728 2013-08-23 Jakub Jelinek <jakub@redhat.com>
1731 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
1732 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
1734 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
1736 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
1737 * gcc/config/i386/i386.md (*movti_internal): Use
1738 predicate to determine if EVEX is needed.
1739 (*movsi_internal): Ditto.
1740 (*movdf_internal): Ditto.
1741 (*movsf_internal): Ditto.
1742 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
1744 2013-08-23 Jakub Jelinek <jakub@redhat.com>
1746 PR tree-optimization/58209
1747 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
1748 (find_tail_calls): Give up for pointer result types if m is non-NULL.
1749 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
1750 emit POINTER_PLUS_EXPR.
1751 (create_tailcall_accumulator): For pointer result type accumulate in
1754 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
1756 * configure.ac: Add backslashes missing from the last change.
1757 * configure: Regenerate.
1759 2013-08-22 Jan Hubicka <jh@susue.cz>
1761 * ipa.c (function_and_variable_visibility): First remember function
1762 was global and then make it local.
1764 2013-08-22 Julian Brown <julian@codesourcery.com>
1766 * configure.ac: Add aarch64 to list of arches which use "nop" in
1768 * configure: Regenerate.
1770 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1772 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
1773 gnu_libc_has_function.
1774 * config/s390/tpf.h: Likewise.
1776 2013-08-22 Jan Hubicka <jh@susue.cz>
1778 * timevar.c (validate_phases): Add cast.
1780 2013-08-22 Jan Hubicka <jh@susue.cz>
1782 * timevar.c (validate_phases): Use size_t for memory.
1783 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
1785 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
1787 * pretty-print.h (output_buffer::output_buffer): Declare.
1788 (pretty_printer::pretty_printer): Likewise.
1789 (pp_construct): Remove.
1790 * pretty-print.c (output_buffer::output_buffer): Define.
1791 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
1792 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
1793 (print_gimple_expr): Likewise.
1794 (print_gimple_seq): Likewise.
1795 (gimple_dump_bb): Likewise.
1796 * sched-vis.c (dump_value_slim): Likewise.
1797 (dump_insn_slim): Likewise.
1798 (dump_rtl_slim): Likewise.
1799 (str_pattern_slim): Likewise.
1800 * tree-mudflap.c (mf_varname_tree): Likewise.
1801 * graph.c (print_graph_cfg): Likewise.
1802 (start_graph_dump): Likewise.
1803 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
1805 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
1806 pretty printer initialization.
1807 * coretypes.h (diagnostic_context): Remove superflous type alias
1809 (pretty_printer): Likewise. Declare directly as a class.
1810 (pretty_print_info): Remove declaration as class.
1811 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
1812 and pp_clear_output_area.
1813 (asan_add_global): Likewise.
1815 2013-08-22 Jan Hubicka <jh@suse.cz>
1817 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
1818 * ipa-utils.h (update_type_inheritance_graph): Declare.
1819 (possible_polymorphic_call_target_p): Declare.
1820 (possible_polymorphic_call_target_p): New.
1821 * ipa-devirt.c: Update toplevel comments.
1822 (cached_polymorphic_call_targets): Move up.
1823 (odr_type_d): Move ID down.
1824 (polymorphic_type_binfo_p): Update comment.
1825 (odr_hasher::remove): Likewise;
1826 (get_odr_type): Set anonymous_namespace.
1827 (dump_odr_type): Dump it.
1828 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
1829 (maybe_record_node): Record node in cached_polymorphic_call_targets.
1830 (record_binfo): Add comment.
1831 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
1833 (devirt_node_removal_hook): Do not iCE when cache is freed.
1834 (possible_polymorphic_call_target_p): New predicate.
1835 (update_type_inheritance_graph): New function.
1837 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
1838 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1839 Sergey Lega <sergey.s.lega@intel.com>
1840 Anna Tikhonova <anna.tikhonova@intel.com>
1841 Ilya Tocar <ilya.tocar@intel.com>
1842 Andrey Turetskiy <andrey.turetskiy@intel.com>
1843 Ilya Verbin <ilya.verbin@intel.com>
1844 Kirill Yukhin <kirill.yukhin@intel.com>
1845 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1847 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
1848 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
1849 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
1850 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
1851 (OPTION_MASK_ISA_AVX2_UNSET): Update.
1852 (OPTION_MASK_ISA_AVX512F_UNSET): New.
1853 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
1854 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
1855 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
1856 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
1857 OPT_mavx512pf, OPT_mavx512er cases.
1858 * config/i386/constraints.md (v): New constraint.
1859 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
1860 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
1861 (bit_AVX512CD): New.
1862 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
1863 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
1864 * config/i386/i386-c.c (ix86_target_macros_internal):
1865 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
1867 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
1868 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
1869 * config/i386/i386.c (regclass_map, dbx_register_map)
1870 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
1871 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
1872 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
1874 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
1875 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
1876 -mavx512pf options. Fix formatting.
1877 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
1878 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
1879 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
1880 -mavx512cd, -mavx512pf options.
1881 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
1882 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
1883 (ix86_preferred_output_reload_class): Replace SSE_REGS with
1885 (ix86_hard_regno_mode_ok): Support 512-bit registers.
1886 (ix86_set_reg_reg_cost): Ditto.
1887 (x86_order_regs_for_local_alloc): Ditto.
1888 (MAX_VECT_LEN): Extend to 64-byte.
1889 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
1890 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
1891 (TARGET_AVX512ER, TARGET_AVX512CD): New.
1892 (BIGGEST_ALIGNMENT): Extend to 512-bits.
1893 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
1894 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
1895 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
1896 (SSE_REG_MODE_P): Support new modes.
1897 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
1898 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
1899 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
1900 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
1901 (REG_CLASS_CONTENTS): Add new registers.
1902 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
1903 (EXT_REX_SSE_REGNO_P): New.
1904 (HI_REGISTER_NAMES): Add new registers.
1905 * config/i386/i386.md: Define constants for new registers.
1906 (mode): Add new 512-bit modes.
1907 (prefix): Support evex prefix.
1908 (isa): Support avx512f, noavx512f, fma_avx512f.
1909 (ssemodesuffix): Add new 512-bit modes.
1911 (*movxi_internal_avx512f): Ditto.
1912 (*movdi_internal): Replace constraint "x" with the new constraint "v".
1914 (*movsi_internal): Likewise.
1915 (*movdf_internal): Likewise.
1916 (*movsf_internal): Likewise.
1917 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
1918 (<code><mode>3): Likewise.
1919 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
1921 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
1922 with the new constraint "v".
1923 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
1925 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
1926 with the new constraint "v".
1927 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
1928 (<sse2>_storedqu<avxsizesuffix>): Likewise.
1929 (*<plusminus_insn><mode>3): Likewise.
1930 (<sse>_vm<plusminus_insn><mode>3): Likewise.
1931 (*mul<mode>3): Likewise.
1932 (<sse>_vmmul<mode>3): Likewise.
1933 (<sse>_div<mode>3): Likewise.
1934 (<sse>_vmdiv<mode>3): Likewise.
1935 (<sse>_sqrt<mode>2): Likewise.
1936 (<sse>_vmsqrt<mode>2): Likewise.
1937 (*<code><mode>3_finite): Likewise.
1938 (*<code><mode>3) <smaxmin>: Likewise.
1939 (<sse>_vm<code><mode>3): Likewise.
1940 (*<code><mode>3) <any_logic>: Likewise.
1941 (*fma_fmadd_<mode>): Likewise.
1942 (*fma_fmsub_<mode>): Likewise.
1943 (*fma_fnmadd_<mode>): Likewise.
1944 (*fma_fnmsub_<mode>): Likewise.
1945 (*fma_fmaddsub_<mode>): Likewise.
1946 (*fma_fmsubadd_<mode>): Likewise.
1947 (*fmai_fmadd_<mode>): Likewise.
1948 (*fmai_fmsub_<mode>): Likewise.
1949 (*fmai_fnmadd_<mode>): Likewise.
1950 (*fmai_fnmsub_<mode>): Likewise.
1951 (sse_cvtsi2ss): Likewise.
1952 (sse_cvtsi2ssq): Likewise.
1953 (sse_cvtss2si): Likewise.
1954 (sse_cvtss2si_2): Likewise.
1955 (sse_cvtss2siq): Likewise.
1956 (sse_cvtss2siq_2): Likewise.
1957 (sse_cvttss2si): Likewise.
1958 (sse_cvtss2siq_2): Likewise.
1959 (float<sseintvecmodelower><mode>2): Likewise.
1960 (sse2_cvtsd2si_2): Likewise.
1961 (sse2_cvtsd2siq_2): Likewise.
1962 (*<plusminus_insn><mode>3): Likewise.
1963 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
1964 (*<sse4_1_avx2>_mul<mode>3): Likewise.
1965 (ashr<mode>3): Likewise.
1966 (<shift_insn><mode>3): Likewise.
1967 (avx2_<code><mode>3): Likewise.
1968 (*avx2_<code><mode>3): Likewise.
1969 (*andnot<mode>3): Likewise.
1970 (*<code><mode>3) <any_logic>: Likewise.
1971 (abs<mode>2): Likewise.
1972 (avx2_permvar<mode>): Likewise.
1973 (avx2_perm<mode>_1): Likewise.
1974 (*avx_vpermilp<mode>): Likewise.
1975 (avx_vpermilvar<mode>3): Likewise.
1976 (avx2_ashrv<mode>): Likewise.
1977 (avx2_<shift_insn>v<mode>): Likewise.
1978 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
1980 * doc/rtl.texi: Document XImode.
1982 2013-08-21 Jeff Law <law@redhat.com>
1984 * tree-flow.h (register_jump_thread): Pass vector of edges
1985 instead of each important edge.
1986 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
1987 thread path into a vector and pass that to register_jump_thread.
1988 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
1989 passed in edge vector to the current 3-edge form.
1992 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1994 * dce.c (fini_dce): Call df_analyze again just in case
1995 delete_unmarked_insns removed anything.
1997 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
1999 * reload.h (struct reg_equivs): Rename to ..
2000 (struct reg_equivs_s): .. this.
2002 2013-08-20 Martin Liska <marxin.liska@gmail.com>
2004 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
2006 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2008 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
2010 2013-08-21 Jeff Law <law@redhat.com>
2012 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
2013 simplify assignments too. If the RHS collapses to a singleton
2014 range, then return the value for the range.
2016 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
2018 * config/i386/sse.md (V16): Rename to...
2020 (mov<mode>): Update iterator name.
2021 (*mov<mode>_internal): Ditto.
2022 (push<mode>1): Ditto.
2023 (movmisalign<mode>): Ditto.
2025 2013-08-20 Jan Hubicka <jh@suse.cz>
2028 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
2029 entry for direct edges.
2030 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
2032 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2034 Revert my last two changes, r201865 and r201864:
2037 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2039 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2040 instances can own GC refs.
2042 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2043 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2044 (gcc::context::gt_pch_nx): Likewise.
2045 (gcc::context::gt_pch_nx): Likewise.
2046 * ggc.h (gt_ggc_mx <T>): New.
2047 (gt_pch_nx_with_op <T>): New.
2048 (gt_pch_nx <T>): New.
2049 * passes.c (opt_pass::gt_ggc_mx): New.
2050 (opt_pass::gt_pch_nx): New.
2051 (opt_pass::gt_pch_nx_with_op): New.
2052 (pass_manager::gt_ggc_mx): New.
2053 (pass_manager::gt_pch_nx): New.
2054 (pass_manager::gt_pch_nx_with_op): New.
2055 (pass_manager::operator new): Use
2056 ggc_internal_cleared_alloc_stat rather than xcalloc.
2057 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2058 (pass_manager::gt_ggc_mx): New.
2059 (pass_manager::gt_pch_nx): New.
2060 (pass_manager::gt_pch_nx_with_op): New.
2061 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2062 (opt_pass::operator new): New.
2063 (opt_pass::gt_ggc_mx): New.
2064 (opt_pass::gt_pch_nx): New.
2065 (opt_pass::gt_pch_nx_with_op): New.
2068 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2070 * Makefile.in (GTFILES): Add context.h.
2071 * context.c (gcc::context::operator new): New.
2072 (gcc::context::gt_ggc_mx): New.
2073 (gcc::context::gt_pch_nx): New.
2074 (gcc::context::gt_pch_nx): New.
2075 * context.h (gcc::context): Add GTY((user)) marking.
2076 (gcc::context::operator new): New.
2077 (gcc::context::gt_ggc_mx): New.
2078 (gcc::context::gt_pch_nx): New.
2079 (gcc::context::gt_pch_nx): New.
2080 (g): Add GTY marking.
2081 (gt_ggc_mx (gcc::context *)): New.
2082 (gt_pch_nx (gcc::context *)): New.
2083 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2084 void *cookie)): New.
2085 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2087 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2089 * dce.c (fini_dce): Call df_analyze again just in case
2090 delete_unmarked_insns removed anything.
2092 2013-08-20 Teresa Johnson <tejohnson@google.com>
2094 PR rtl-optimizations/57451
2095 * final.c (reemit_insn_block_notes): Prevent lexical blocks
2096 from crossing split section boundaries.
2098 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
2100 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
2101 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
2102 with MULTLIB_DEFAULTS.
2104 2013-08-20 Nick Clifton <nickc@redhat.com>
2106 * target.def (narrow_volatile_bitfield): Note that the default
2107 value is false, not !TARGET_STRICT_ALIGN.
2108 * doc/tm.texi: Regenerate.
2110 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
2112 Fix LIB_SPEC for systems without libpthread.
2114 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
2115 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
2117 * config/i386/linux-common.h: Likewise.
2118 * config/mips/linux-common.h: Likewise.
2120 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2122 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
2125 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2127 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2128 instances can own GC refs.
2130 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2131 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2132 (gcc::context::gt_pch_nx): Likewise.
2133 (gcc::context::gt_pch_nx): Likewise.
2134 * ggc.h (gt_ggc_mx <T>): New.
2135 (gt_pch_nx_with_op <T>): New.
2136 (gt_pch_nx <T>): New.
2137 * passes.c (opt_pass::gt_ggc_mx): New.
2138 (opt_pass::gt_pch_nx): New.
2139 (opt_pass::gt_pch_nx_with_op): New.
2140 (pass_manager::gt_ggc_mx): New.
2141 (pass_manager::gt_pch_nx): New.
2142 (pass_manager::gt_pch_nx_with_op): New.
2143 (pass_manager::operator new): Use
2144 ggc_internal_cleared_alloc_stat rather than xcalloc.
2145 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2146 (pass_manager::gt_ggc_mx): New.
2147 (pass_manager::gt_pch_nx): New.
2148 (pass_manager::gt_pch_nx_with_op): New.
2149 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2150 (opt_pass::operator new): New.
2151 (opt_pass::gt_ggc_mx): New.
2152 (opt_pass::gt_pch_nx): New.
2153 (opt_pass::gt_pch_nx_with_op): New.
2155 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2157 * Makefile.in (GTFILES): Add context.h.
2158 * context.c (gcc::context::operator new): New.
2159 (gcc::context::gt_ggc_mx): New.
2160 (gcc::context::gt_pch_nx): New.
2161 (gcc::context::gt_pch_nx): New.
2162 * context.h (gcc::context): Add GTY((user)) marking.
2163 (gcc::context::operator new): New.
2164 (gcc::context::gt_ggc_mx): New.
2165 (gcc::context::gt_pch_nx): New.
2166 (gcc::context::gt_pch_nx): New.
2167 (g): Add GTY marking.
2168 (gt_ggc_mx (gcc::context *)): New.
2169 (gt_pch_nx (gcc::context *)): New.
2170 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2171 void *cookie)): New.
2172 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2174 2013-08-20 Alan Modra <amodra@gmail.com>
2177 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
2178 (rs6000_emit_epilogue): Likewise.
2180 2013-08-19 Dehao Chen <dehao@google.com>
2182 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
2184 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
2185 Jakub Jelinek <jakub@redhat.com>
2187 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
2188 (BUILT_IN_FABSD64): Likewise.
2189 (BUILT_IN_FABSD128): Likewise.
2190 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
2191 (fold_builtin_1): Likewise.
2192 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
2193 destination and source operands.
2194 (*abstd2_fpr): Likewise.
2195 (*nabstd2_fpr): Likewise.
2197 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
2199 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
2202 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
2204 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
2207 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
2209 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
2210 * builtins.c (default_libc_has_function): New.
2211 (gnu_libc_has_function): Ditto.
2212 (no_c99_libc_has_function): Ditto.
2213 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
2214 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
2215 (fold_builtin_sincos): Likewise.
2216 (fold_builtin_cexp): Likewise.
2217 * builtins.def (DEF_C94_BUILTIN): Likewise.
2218 (DEF_C99_BUILTIN): Likewise.
2219 (DEF_C99_C90RES_BUILTIN): Likewise.
2220 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
2221 definitions to using this define.
2222 * config/darwin-protos.h (darwin_libc_has_function): New.
2223 * config/darwin.c (darwin_libc_has_function): Ditto.
2224 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
2225 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
2226 * config/darwin.h: Ditto.
2227 * config/elfos.h: Ditto.
2228 * config/freebsd.h: Ditto.
2229 * config/i386/cygming.h: Ditto.
2230 * config/i386/djgpp.h: Ditto.
2231 * config/i386/i386-interix.h: Ditto.
2232 * config/microblaze/microblaze.h: Ditto.
2233 * config/mmix/mmix.h: Ditto.
2234 * config/gnu-user.h: Ditto.
2235 * config/ia64/hpux.h: Ditto.
2236 * config/pa/pa-hpux.h: Ditto.
2237 * config/pdp11/pdp11.h: Ditto.
2238 * config/picochip/picochip.h: Ditto.
2239 * config/linux.h: Ditto.
2240 * config/netbsd.h: Ditto.
2241 * config/openbsd.h: Ditto.
2242 * config/rs6000/aix43.h: Ditto.
2243 * config/rs6000/aix51.h: Ditto.
2244 * config/rs6000/aix52.h: Ditto.
2245 * config/rs6000/aix53.h: Ditto.
2246 * config/rs6000/aix61.h: Ditto.
2247 * config/rs6000/darwin.h: Ditto.
2248 * config/rs6000/linux.h: Ditto.
2249 * config/rs6000/linux64.h: Ditto.
2250 * config/s390/tpf.h: Ditto.
2251 * config/sol2-10.h: Ditto.
2252 * config/sol2.h: Ditto.
2253 * config/vms/vms.h: Ditto.
2254 * config/vxworks.h: Ditto.
2255 * config/linux-android.c (linux_android_libc_has_function):
2256 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
2257 * config/linux-protos.h (linux_android_libc_has_function):
2259 * config/i386/i386.c (ix86_libc_has_function): New.
2260 * config/i386/i386-protos.h
2261 (ix86_libc_has_function): New declaration.
2262 * config/i386/i386.md
2263 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
2264 ("isinf<mode>2): Likewise.
2265 * convert.c (convert_to_integer): Using new target hook
2266 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
2267 TARGET_C99_FUNCTIONS.
2268 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
2269 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
2270 * coretypes.h (function_class): New enum for different
2271 classes of functions.
2272 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
2273 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
2274 (TARGET_HAS_SINCOS): Likewise.
2275 (TARGET_LIBC_HAS_FUNCTION): New.
2276 * doc/tm.texi: Regenerated.
2277 * targhooks.h (default_libc_has_function): New declaration.
2278 (no_c99_libc_has_function): Ditto.
2279 (gnu_libc_has_function): Ditto.
2280 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
2281 and TARGET_HAS_SINCOS.
2283 2013-08-18 Jan Hubicka <jh@suse.cz>
2285 * Makeifle-in (ipa-devirt.o): New.
2286 (GTFILES): Add ipa-utils.h and ipa-devirt.c
2287 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
2288 (analyze_functions): Look into possible targets of polymorphic call.
2289 * dumpfile.c (dump_files): Add type-inheritance dump.
2290 * dumpfile.h (TDI_inheritance): New.
2291 * ipa-devirt.c: New file.
2292 * ipa-utils.h (odr_type_d): Forward declare.
2293 (odr_type): New type.
2294 (build_type_inheritance_graph): Declare.
2295 (possible_polymorphic_call_targets): Declare and introduce inline
2296 variant when only edge is pased.
2297 (dump_possible_polymorphic_call_targets): Likewise.
2298 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
2299 * tree.c (type_in_anonymous_namespace_p): Break out from ...
2300 (types_same_for_odr): ... here.
2301 * tree.h (type_in_anonymous_namespace_p): Declare.
2303 2013-08-18 Jakub Jelinek <jakub@redhat.com>
2305 PR tree-optimization/58006
2306 * tree-parloops.c (take_address_of): Don't ICE if get_name
2308 (eliminate_local_variables_stmt): Remove clobber stmts.
2310 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
2312 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
2313 error message is issued for an alias to undefined symbol.
2315 2013-08-18 Jan Hubicka <jh@suse.cz>
2317 * cgraph.c (cgraph_create_indirect_edge): Discover
2318 polymorphic calls and record basic info into indirect_info.
2319 * gimple-fold.c (gimple_fold_call): When doing BINFO based
2320 devirtualization, ignore objc function calls.
2321 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
2322 call with no parm index info.
2323 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
2324 * tree.c (virtual_method_call_p): New function.
2325 * tree.h (virtual_method_call_p): Declare.
2327 2013-08-16 Jan Hubicka <jh@suse.cz>
2330 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
2332 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
2334 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
2337 2013-08-16 David Malcolm <dmalcolm@redhat.com>
2339 * gengtype.c (type_for_name): Add special-case support for
2340 locating types within the "gcc::" namespace.
2341 (open_base_files): Emit a "using namespace gcc" directive.
2343 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2346 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
2347 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
2349 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
2350 array instead of each individual operand as a separate argument.
2351 (emit_fusion_gpr_load): Likewise.
2352 (expand_fusion_gpr_load): Add new function declaration.
2354 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
2355 signature to have the operands passed as an array, instead of as
2356 separate arguments. Allow ZERO_EXTEND to be in the memory
2357 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
2358 depend on the register live/dead flags when peepholes are run.
2359 (expand_fusion_gpr_load): New function to be called from the
2360 peephole2 pass, to change the register that addis sets to be the
2362 (emit_fusion_gpr_load): Change the calling signature to have the
2363 operands passed as an array, instead of as separate arguments.
2364 Allow ZERO_EXTEND to be in the memory address, and also
2365 SIGN_EXTEND if -mpower8-fusion-sign.
2367 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
2369 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
2370 adjust the register addis loads up in the peephole2 pass. Do not
2371 depend on the register live/dead state when the peephole pass is done.
2373 2013-08-16 David Malcolm <dmalcolm@redhat.com>
2375 * gengtype.c (create_user_defined_type): Ensure that the kind
2376 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
2377 declaration is seen before the GTY((user)) marking.
2379 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
2382 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
2384 2013-08-16 Jan Hubicka <jh@suse.cz>
2386 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
2387 arugment expected_type.
2388 (gimple_fold_call): Use it.
2389 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
2390 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
2391 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
2392 (try_make_edge_direct_virtual_call): Likewise.
2393 * tree.c (obj_type_ref_class): New.
2394 * tree.h (obj_type_ref_class): Use it.
2396 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
2398 * sched-vis.c (rtl_slim_pp_initialized): Remove.
2399 (rtl_slim_pp): Likewise.
2400 (init_rtl_slim_pretty_print): Likewise.
2401 (dump_value_slim): Don't call it. Use local pretty printer.
2402 (dump_insn_slim): Likewise.
2403 (dump_rtl_slim): Likewise.
2404 (str_pattern_slim): Likewise.
2405 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
2408 2013-08-16 Jakub Jelinek <jakub@redhat.com>
2410 PR tree-optimization/58164
2411 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
2412 walk gimple_goto_dest of GIMPLE_GOTO.
2414 PR tree-optimization/58165
2415 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
2416 bi_call must be the last stmt in a bb, don't split_block, instead
2417 use fallthru edge from it and give up if there is none.
2418 Release conds vector when returning early.
2420 2013-08-14 Xinliang David Li <davidxl@google.com>
2422 * config/i386/i386.c (ix86_option_override_internal):
2423 Remove unused variable and field.
2425 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2428 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
2429 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
2430 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
2431 and Linux, correct BLKmode alignment when 128-bit alignment is
2432 required and compatibility flag is not set.
2433 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
2434 for zero-size arguments when compatibility flag is not set.
2436 2013-08-14 Jakub Jelinek <jakub@redhat.com>
2438 PR tree-optimization/58145
2439 * tree-sra.c (build_ref_for_offset): If prev_base has
2440 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
2442 2013-08-14 Xinliang David Li <davidxl@google.com>
2444 * config/i386/i386.c (ix86_option_override_internal):
2445 Fix uninitialized variable error.
2447 2013-08-14 Xinliang David Li <davidxl@google.com>
2449 * config/i386/i386.opt: Define two new options.
2450 * config/i386/x86-tune.def: Add arch selector field in macros.
2451 * config/i386/i386.h: Adjust macro definition.
2452 * config/i386/i386.c (ix86_option_override_internal):
2454 (parse_mtune_ctrl_str): New function.
2455 (set_ix86_tune_features): New function.
2456 (ix86_function_specific_restore): Call the new helper function.
2458 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
2460 PR rtl-optimization/57662
2461 * sel-sched.c (code_motion_process_successors): When the current insn
2462 is removed after the recursive traversal, break from the loop.
2463 Add comments and debug printouts.
2465 2013-08-14 Jakub Jelinek <jakub@redhat.com>
2466 Alexandre Oliva <aoliva@redhat.com>
2469 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
2470 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
2471 there also UNSPEC_PLTOFF.
2473 2013-08-14 Marek Polacek <polacek@redhat.com>
2475 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
2476 NUM_CONDITIONS bit positions.
2478 2013-08-13 Cary Coutant <ccoutant@google.com>
2480 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
2481 (attr_checksum): Hash vector contents instead of pointer.
2482 (attr_checksum_ordered): Likewise.
2484 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
2486 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
2487 when Pmode != word_mode. Add length_address attribute.
2488 (sse3_monitor_<mode>): Merge from sse3_monitor and
2489 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
2490 Pmode != word_mode. Update insn length attribute.
2491 * config/i386/i386.c (ix86_option_override_internal): Update
2492 ix86_gen_monitor selection for merged sse3_monitor insn.
2494 2013-08-13 Julian Brown <julian@codesourcery.com>
2496 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
2497 perform invalid legitimization on greater-than-word-size modes for
2500 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
2502 * ira.c (setup_class_translate_array): Use aclass instead of cl
2503 for classes not fully covered by allocno classes.
2505 2013-08-13 Jakub Jelinek <jakub@redhat.com>
2507 PR tree-optimization/57661
2508 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
2509 * tree-inline.c (tree_function_versioning): Initialize it.
2510 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
2511 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
2512 that is not being copied.
2515 * asan.c (instrument_strlen_call): Fix typo in comment.
2516 Use char * type even for the lhs of POINTER_PLUS_EXPR.
2518 2013-08-13 Steve Ellcey <sellcey@mips.com>
2520 * config/mips/mips.md (prefetch): Use lw instead of ld on
2521 loongson in 32bit mode.
2523 2013-08-13 Nick Clifton <nickc@redhat.com>
2525 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
2527 2013-08-13 Jan Hubicka <jh@suse.cz>
2529 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
2530 introduced edge; fix typo in sanity check.
2531 (cgraph_resolve_speculation): Export; improve diagnostic.
2532 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
2533 speculation at type mismatch.
2534 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
2535 (cgraph_resolve_speculation): Declare.
2536 (symtab_can_be_discarded): New function.
2537 * value-prof.c (gimple_ic_transform): Remove actual transform code.
2538 * ipa-inline-transform.c (speculation_removed): New global var.
2539 (clone_inlined_nodes): See if speculation can be removed.
2540 (inline_call): If speculations was removed, we growths may not match.
2541 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
2542 (speculation_useful_p): New function.
2543 (resolve_noninline_speculation): New function.
2544 (inline_small_functions): Resolve useless speculations.
2545 * ipa-inline.h (speculation_useful_p): Declare
2546 * ipa.c (can_replace_by_local_alias): Simplify.
2547 (ipa_profile): Produce speculative calls in non-lto, too;
2548 add simple cost model; produce local aliases.
2550 2013-08-13 David Malcolm <dmalcolm@redhat.com>
2552 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
2555 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
2557 * config/i386/i386.c (ix86_function_versions): Use error + inform.
2559 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
2561 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
2562 iterator instead of X87MODEF.
2564 2013-08-12 Perez Read <netfirewall@gmail.com>
2567 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
2568 operand 0 for intel asm alternative.
2569 (*movabs<mode>_2): Ditto for operand 1.
2571 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
2573 * config/aarch64/arm_none.h
2574 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
2576 2013-08-12 Nick Clifton <nickc@redhat.com>
2578 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
2580 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2582 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
2583 expand for QI/HImode operand to produce more effictive code for
2584 unsigned char(short) --> float(double) conversion.
2586 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
2588 * doc/invoke.texi: Mention that -ftls-model does not force the final
2591 2013-08-12 Marek Polacek <polacek@redhat.com>
2592 Marc Glisse <marc.glisse@inria.fr>
2594 PR tree-optimization/57980
2595 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
2596 when creating -1 constant.
2598 2013-08-10 Jan Hubicka <jh@suse.cz>
2600 Workaround binutils PR14342.
2601 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
2602 (gimple_init_edge_profiler): Likewise.
2603 (gimple_gen_ic_func_profiler): Likewise.
2605 2013-08-09 Jan Hubicka <jh@suse.cz>
2607 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
2609 2013-08-09 Xinliang David Li <davidxl@google.com>
2611 * config/i386/stringop.def: New file.
2612 * config/i386/stringop.opt: New file.
2613 * config/i386/i386-opts.h: Include stringopt.def.
2614 * config/i386/i386.opt: Include stringopt.opt.
2615 * config/i386/i386.c (ix86_option_override_internal):
2616 Override default size based stringop inline strategies with options.
2617 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
2620 2013-08-09 Jan Hubicka <jh@suse.cz>
2622 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
2624 2013-08-09 Jan Hubicka <jh@suse.cz>
2626 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
2628 (dump_cgraph_node): Dump profile-id.
2629 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
2630 and common_target_probability.
2631 * lto-cgraph.c (lto_output_edge): Stream common targets.
2632 (lto_output_node): Stream profile ids.
2633 (input_node): Stream profile ids.
2634 (input_edge): Stream common targets.
2635 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
2636 * ipa.c: Include value-prof.h
2637 (ipa_profile_generate_summary): Turn indirect call statement histograms
2638 into common targets.
2639 (ipa_profile): Turn common targets into speculative edges.
2641 2013-08-09 Jan Hubicka <jh@suse.cz>
2643 * cgraph.h (cgraph_node): Add profile_id.
2644 * value-prof.c (cgraph_node_map): Turn into pointer_map.
2645 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2646 (del_node_map): Update.
2647 (find_func_by_funcdef_no): Replace by ...
2648 (find_func_by_profile_id): ... this one.
2649 (gimple_ic_transform): Do not remove useful histograms when
2650 speculation is not done; dump info when indirect call removal
2652 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
2653 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
2654 (__gcov_indirect_call_profiler_v2): .. this one.
2655 * profile.h (init_node_map): Update.
2656 * coverage.c (coverage_compute_profile_id): New function.
2657 * coverage.h (coverage_compute_profile_id): Declare.
2658 * tree-profile.c (init_ic_make_global_vars): Make
2659 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
2660 (gimple_init_edge_profiler): Update prototype of
2661 __gcov_indirect_call_profiler.
2662 (gimple_gen_ic_func_profiler): Simplify.
2663 (tree_profiling): Use init_node_map
2665 2013-08-09 Jan Hubicka <jh@suse.cz>
2667 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
2668 non-speculative refs.
2669 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
2670 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
2671 (cgraph_set_call_stmt): Likewise.
2672 (cgraph_create_edge_1): Fix release checking compilatoin;
2674 (cgraph_free_edge): Free indirect info.
2675 (cgraph_turn_edge_to_speculative): New function.
2676 (cgraph_speculative_call_info): New function.
2677 (cgraph_make_edge_direct): Return direct edge; handle speculation.
2678 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
2679 (dump_cgraph_node): Dump speculation.
2680 (verify_edge_count_and_frequency): Accept speculative edges.
2681 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
2682 (verify_cgraph_node): Handle speculation.
2683 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
2684 (cgraph_set_call_stmt): Update prototype.
2685 (cgraph_make_edge_direct): Update prototype.
2686 (cgraph_speculative_call_info): Declare.
2687 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
2688 to change; update call of ipa_find_references.
2689 * ipa-ref.c (ipa_record_reference): Fix return value; clear
2690 lto_stmt_uid and speculative flags.
2691 (ipa_dump_references): Dump speculation.
2692 (ipa_clone_references): Clone speculative flag.
2693 (ipa_clone_referring): Likewise.
2694 (ipa_clone_ref): New function.
2695 (ipa_find_reference): Look into lto_stmt_uids
2696 (ipa_clear_stmts_in_references): Do not clear speculative calls.
2697 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
2698 (ipa_find_reference): Update declaration.
2699 (ipa_clone_ref): Declare.
2700 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
2701 stream speculative flag.
2702 (lto_output_ref): Stream statements uids and speculation.
2703 (input_ref): Likewise.
2704 (input_edge): Stream speuclation.
2705 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
2706 (cgraph_set_call_stmt_including_clones): Handle speculation.
2707 * ipa-inline.c (heap_edge_removal_hook): New function.
2708 (inline_small_functions): Register it.
2709 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
2710 also initialize refs.
2711 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
2713 (try_make_edge_direct_simple_call): Likewise.
2714 (try_make_edge_direct_simple_call): Likewise.
2715 (update_indirect_edges_after_inlining): Likewise.
2716 (remove_described_reference): Look proper lto_stmt_uid.
2717 (propagate_controlled_uses): Likewise.
2718 (propagate_controlled_uses): Liekwise.
2719 * tree-inline.c (copy_bb): Copy speculative edges.
2720 (redirect_all_calls): New function.
2721 (copy_cfg_body): Do redirection after loop info is updated.
2722 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
2724 2013-08-09 Jan Hubicka <jh@suse.cz>
2726 * lto-streamer-out.c (output_function): Renumber PHIs.
2727 * lto-streamer-in.c (input_function): Likewise.
2729 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
2731 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
2732 (get_lane_unsigned): Likewise.
2733 (dup_lane_scalar): Likewise.
2734 (get_lane): enable for VALL.
2735 * config/aarch64/aarch64-simd.md
2736 (aarch64_dup_lane_scalar<mode>): Remove.
2737 (aarch64_get_lane_signed<mode>): Likewise.
2738 (aarch64_get_lane_unsigned<mode>): Likewise.
2739 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
2740 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
2741 (aarch64_get_lane<mode>): Enable for all vector modes.
2742 (aarch64_get_lanedi): Remove misleading constraints.
2743 * config/aarch64/arm_neon.h
2744 (__aarch64_vget_lane_any): Define.
2745 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
2746 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
2747 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
2748 * config/aarch64/iterators.md (VDQQH): New.
2752 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
2754 * configure.ac: Add GAS check for LEON instructions on SPARC.
2755 * configure: Regenerate.
2756 * config.in: Likewise.
2757 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
2759 * config/sparc/sparc.opt (LEON, LEON3): New masks.
2760 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
2762 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
2763 (AS_LEON_FLAG): New macro.
2764 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
2765 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
2766 Deal with LEON and LEON3 for the memory model.
2767 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
2768 (atomic_compare_and_swap<mode>_1): Likewise.
2769 (*atomic_compare_and_swap<mode>_1): Likewise.
2771 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2773 * config/arm/neon.md (vcond): Fix floating-point vector
2774 comparisons against 0.
2776 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
2778 * lra-constraints.c (emit_spill_move): Remove assert.
2779 (process_alt_operands): Add more debugging
2780 output. Increase reject for spilling into memory. Decrease
2781 reject for reloading scratch.
2782 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
2784 2013-08-08 Steve Ellcey <sellcey@mips.com>
2786 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
2787 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
2788 micromips incompatible. Add nan2008.
2789 (MULTILIB_DIRNAMES): Add nan2008.
2790 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
2791 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
2792 and micromips incompatible. Add nan2008.
2793 (MULTILIB_DIRNAMES): Add nan2008.
2794 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
2796 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
2798 PR rtl-optimization/58079
2799 * combine.c (combine_simplify_rtx): Avoid using SUBST if
2800 simplify_comparison has widened a comparison with an integer.
2802 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2804 * config/arm/neon.md (movmisalign<mode>): Disable when we
2805 don't allow unaligned accesses.
2806 (*movmisalign<mode>_neon_store): Likewise.
2807 (*movmisalign<mode>_neon_load): Likewise.
2808 (*movmisalign<mode>_neon_store): Likewise.
2809 (*movmisalign<mode>_neon_load): Likewise.
2811 2013-08-08 Jan Hubicka <jh@suse.cz>
2813 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
2814 (make_pass_rebuild_cgraph_edges): Also clear references.
2815 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
2816 * ipa-inline-transform.c (inline_transform): Remove all references
2818 * cgraphunit.c (expand_function): Remove all references after
2820 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
2821 (ipa_find_reference): Rewrite to iterator.
2822 (remove_stmt_references): Likewise.
2823 (ipa_clear_stmts_in_references): New function.
2824 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
2825 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
2827 * ipa-split.c (split_function): Remove references in split function.
2829 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2832 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
2833 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
2835 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2838 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
2839 suggested mode for the assignment isn't compatible with the
2842 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2845 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
2847 2013-08-07 Xinliang David Li <davidxl@google.com>
2849 * config/i386/i386.opt: New option -mtune-ctrl=.
2850 * config/i386/x86-tune.def: New file.
2851 * config/i386/i386.h: include x86-tune.def.
2852 * config/i386/i386.c (ix86_option_override_internal):
2853 Parsing -mtune-ctrl= option and set tune features.
2855 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
2858 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
2859 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
2862 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2864 * rtl.h (update_alignments): Declare.
2865 * final.c (grow_label_align): New function extracted from...
2866 (shorten_branches): ...here. Call it.
2867 (update_alignments): New function.
2868 * reorg.c (sibling_labels): New variable.
2869 (get_label_before): Add SIBLING parameter. If it is non-zero, push
2870 the new label along with it onto the sibling_labels vector.
2871 (fill_simple_delay_slots): Adjust call to get_label_before.
2872 (fill_slots_from_thread): Likewise.
2873 (relax_delay_slots): Likewise.
2874 (make_return_insns): Likewise.
2875 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
2877 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2879 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
2880 document its semantics.
2881 (diagnostic_report_diagnostic): Adjust accordingly.
2883 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2885 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
2886 (sparc_option_override): ...and port to new C++ pass API.
2887 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
2889 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
2891 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
2893 2013-08-06 Caroline Tice <cmtice@google.com>
2895 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
2896 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
2897 * tree-pass.h: Add pass_vtable_verify.
2898 * varasm.c (assemble_variable): Add code to properly set the comdat
2899 section and name for the .vtable_map_vars section.
2900 (assemble_vtyv_preinit_initializer): New function.
2901 (default_sectin_type_flags): Make sure .vtable_map_vars section has
2903 * output.h: Add function decl for assemble_vtv_preinit_initializer.
2904 * vtable-verify.c: New file.
2905 * vtable-verify.h: New file.
2906 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
2907 initialiation levels.
2908 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
2909 * passes.def: Insert pass_vtable_verify.
2910 * aclocal.m4: Reorder includes.
2911 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
2912 -fvtv-counts options.
2913 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
2914 as appropriate, if -fvtable-verify=... is used.
2915 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
2916 -fvtable-verify=... is used.
2917 * Makefile.in (OBJS): Add vtable-verify.o to list.
2918 (vtable-verify.o): Add new build rule.
2919 (GTFILES): Add vtable-verify.c to list.
2920 * common.opt (fvtable-verify=): New flag.
2921 (vtv_priority): Values for fvtable-verify= flag.
2922 (fvtv-counts): New flag.
2923 (fvtv-debug): New flag.
2924 * tree.h (save_vtable_map_decl): New extern function decl.
2926 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2928 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
2929 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
2930 (pass_data_rl78_devirt): ...new pass_data instance and...
2931 (make_pass_rl78_devirt): ...new function.
2932 (rl78_asm_file_start): Port pass registration to new C++ API.
2934 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2936 * coretypes.h (rtl_opt_pass): Add.
2937 (gcc::context): Add.
2938 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
2939 (epiphany_init): Port to new C++ pass API.
2940 (epiphany_optimize_mode_switching): Likewise.
2941 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
2942 (pass_manager::get_pass_mode_switching): New.
2943 (pass_manager::get_pass_peephole2): New.
2944 * mode-switching.c (pass_mode_switching): Add clone method.
2945 * recog.c (pass_peephole2): Add clone method.
2946 (pass_split_all_insns): Add clone method.
2948 2013-08-06 David Malcolm <dmalcolm@redhat.com>
2950 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
2951 (mips_option_override): ...here, porting to new C++ API for passes.
2953 2013-08-06 Jan Hubicka <jh@suse.cz>
2955 * cgraph.c (cgraph_get_body): New function based on lto.c
2957 * cgraph.h (cgraph_get_body): Declare.
2958 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
2960 * cgraphunit.c (expand_function): Get body prior expanding.
2961 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
2962 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
2964 * passes.c (do_per_function_toporder): Get body.
2965 * tree-inline.c (expand_call_inline): Get body prior inlining it.
2966 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
2968 2013-08-06 Martin Jambor <mjambor@suse.cz>
2971 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
2972 is not re-finalized. Rename second parameter to no_collect.
2974 2013-08-06 Martin Jambor <mjambor@suse.cz>
2977 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
2978 MEM_REF has proper alignment information.
2980 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
2983 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
2985 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
2986 rtx (*) (rtx, ...) with insn_gen_fn.
2987 * genoutput.c (output_insn_data): Cast gen_? function pointers to
2988 insn_gen_fn::stored_funcptr. Add initializer braces.
2990 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2992 Rewrite how instances of passes are cloned to remove assumptions
2993 about their sizes (thus allowing pass subclasses to have
2994 additional data fields, albeit non-GC-managed ones at this point).
2996 * passes.c (make_pass_instance): Now that passes have clone
2997 methods, rewrite this function to eliminate XNEW and memcpy
2998 calls that used hardcoded sizes. Since this function no longer
2999 creates pass instances, rename it to...
3000 (add_pass_instance): ...this. Document the old way that passes were
3001 numbered and flagged, and rework this function to continue using it.
3002 (next_pass_1): Add an initial_pass argument for use by
3004 (position_pass): When adding multiple instances of a pass, use
3005 the pass's clone method, rather than relying on the XNEW/memcpy
3006 within the former make_pass_instance (now add_pass_instance).
3007 (pass_manager::pass_manager): When invoking next_pass_1, also supply
3008 the initial instance of the current pass within the pass manager.
3010 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3012 This is the automated part of the conversion of passes from C
3013 structs to C++ classes.
3015 Patch autogenerated by refactor_passes.py from
3016 https://github.com/davidmalcolm/gcc-refactoring-scripts
3017 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
3019 * asan.c (pass_asan): Convert from a global struct to a subclass of
3020 gimple_opt_pass along with...
3021 (pass_data_asan): ...new pass_data instance and...
3022 (make_pass_asan): ...new function.
3023 (pass_asan_O0): Convert from a global struct to a subclass of
3024 gimple_opt_pass along with...
3025 (pass_data_asan_O0): ...new pass_data instance and...
3026 (make_pass_asan_O0): ...new function.
3027 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
3028 subclass of rtl_opt_pass along with...
3029 (pass_data_inc_dec): ...new pass_data instance and...
3030 (make_pass_inc_dec): ...new function.
3031 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
3032 a subclass of rtl_opt_pass along with...
3033 (pass_data_reorder_blocks): ...new pass_data instance and...
3034 (make_pass_reorder_blocks): ...new function.
3035 (pass_duplicate_computed_gotos): Convert from a global struct to a
3036 subclass of rtl_opt_pass along with...
3037 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
3038 (make_pass_duplicate_computed_gotos): ...new function.
3039 (pass_partition_blocks): Convert from a global struct to a subclass of
3040 rtl_opt_pass along with...
3041 (pass_data_partition_blocks): ...new pass_data instance and...
3042 (make_pass_partition_blocks): ...new function.
3043 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
3044 struct to a subclass of rtl_opt_pass along with...
3045 (pass_data_branch_target_load_optimize1): ...new pass_data instance
3047 (make_pass_branch_target_load_optimize1): ...new function.
3048 (pass_branch_target_load_optimize2): Convert from a global struct to a
3049 subclass of rtl_opt_pass along with...
3050 (pass_data_branch_target_load_optimize2): ...new pass_data instance
3052 (make_pass_branch_target_load_optimize2): ...new function.
3053 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
3054 of rtl_opt_pass along with...
3055 (pass_data_jump): ...new pass_data instance and...
3056 (make_pass_jump): ...new function.
3057 (pass_jump2): Convert from a global struct to a subclass of
3058 rtl_opt_pass along with...
3059 (pass_data_jump2): ...new pass_data instance and...
3060 (make_pass_jump2): ...new function.
3061 * cfgexpand.c (pass_expand): Convert from a global struct to a
3062 subclass of rtl_opt_pass along with...
3063 (pass_data_expand): ...new pass_data instance and...
3064 (make_pass_expand): ...new function.
3065 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
3066 of rtl_opt_pass along with...
3067 (pass_data_free_cfg): ...new pass_data instance and...
3068 (make_pass_free_cfg): ...new function.
3069 (pass_into_cfg_layout_mode): Convert from a global struct to a
3070 subclass of rtl_opt_pass along with...
3071 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
3072 (make_pass_into_cfg_layout_mode): ...new function.
3073 (pass_outof_cfg_layout_mode): Convert from a global struct to a
3074 subclass of rtl_opt_pass along with...
3075 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
3076 (make_pass_outof_cfg_layout_mode): ...new function.
3077 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
3078 struct to a subclass of gimple_opt_pass along with...
3079 (pass_data_build_cgraph_edges): ...new pass_data instance and...
3080 (make_pass_build_cgraph_edges): ...new function.
3081 (pass_rebuild_cgraph_edges): Convert from a global struct to a
3082 subclass of gimple_opt_pass along with...
3083 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
3084 (make_pass_rebuild_cgraph_edges): ...new function.
3085 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
3086 subclass of gimple_opt_pass along with...
3087 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
3089 (make_pass_remove_cgraph_callee_edges): ...new function.
3090 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
3091 struct to a subclass of rtl_opt_pass along with...
3092 (pass_data_stack_adjustments): ...new pass_data instance and...
3093 (make_pass_stack_adjustments): ...new function.
3094 * combine.c (pass_combine): Convert from a global struct to a subclass
3095 of rtl_opt_pass along with...
3096 (pass_data_combine): ...new pass_data instance and...
3097 (make_pass_combine): ...new function.
3098 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
3099 global struct to a subclass of rtl_opt_pass along with...
3100 (pass_data_compare_elim_after_reload): ...new pass_data instance
3102 (make_pass_compare_elim_after_reload): ...new function.
3103 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
3104 of rtl_opt_pass along with...
3105 (pass_data_rtl_cprop): ...new pass_data instance and...
3106 (make_pass_rtl_cprop): ...new function.
3107 * cse.c (pass_cse): Convert from a global struct to a subclass of
3108 rtl_opt_pass along with...
3109 (pass_data_cse): ...new pass_data instance and...
3110 (make_pass_cse): ...new function.
3111 (pass_cse2): Convert from a global struct to a subclass of
3112 rtl_opt_pass along with...
3113 (pass_data_cse2): ...new pass_data instance and...
3114 (make_pass_cse2): ...new function.
3115 (pass_cse_after_global_opts): Convert from a global struct to a
3116 subclass of rtl_opt_pass along with...
3117 (pass_data_cse_after_global_opts): ...new pass_data instance and...
3118 (make_pass_cse_after_global_opts): ...new function.
3119 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
3120 of rtl_opt_pass along with...
3121 (pass_data_ud_rtl_dce): ...new pass_data instance and...
3122 (make_pass_ud_rtl_dce): ...new function.
3123 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
3124 rtl_opt_pass along with...
3125 (pass_data_fast_rtl_dce): ...new pass_data instance and...
3126 (make_pass_fast_rtl_dce): ...new function.
3127 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
3128 a subclass of rtl_opt_pass along with...
3129 (pass_data_df_initialize_opt): ...new pass_data instance and...
3130 (make_pass_df_initialize_opt): ...new function.
3131 (pass_df_initialize_no_opt): Convert from a global struct to a
3132 subclass of rtl_opt_pass along with...
3133 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
3134 (make_pass_df_initialize_no_opt): ...new function.
3135 (pass_df_finish): Convert from a global struct to a subclass of
3136 rtl_opt_pass along with...
3137 (pass_data_df_finish): ...new pass_data instance and...
3138 (make_pass_df_finish): ...new function.
3139 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
3140 rtl_opt_pass along with...
3141 (pass_data_rtl_dse1): ...new pass_data instance and...
3142 (make_pass_rtl_dse1): ...new function.
3143 (pass_rtl_dse2): Convert from a global struct to a subclass of
3144 rtl_opt_pass along with...
3145 (pass_data_rtl_dse2): ...new pass_data instance and...
3146 (make_pass_rtl_dse2): ...new function.
3147 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
3148 subclass of rtl_opt_pass along with...
3149 (pass_data_dwarf2_frame): ...new pass_data instance and...
3150 (make_pass_dwarf2_frame): ...new function.
3151 * except.c (pass_set_nothrow_function_flags): Convert from a global
3152 struct to a subclass of rtl_opt_pass along with...
3153 (pass_data_set_nothrow_function_flags): ...new pass_data instance
3155 (make_pass_set_nothrow_function_flags): ...new function.
3156 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
3157 subclass of rtl_opt_pass along with...
3158 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
3160 (make_pass_convert_to_eh_region_ranges): ...new function.
3161 * final.c (pass_compute_alignments): Convert from a global struct to a
3162 subclass of rtl_opt_pass along with...
3163 (pass_data_compute_alignments): ...new pass_data instance and...
3164 (make_pass_compute_alignments): ...new function.
3165 (pass_final): Convert from a global struct to a subclass of
3166 rtl_opt_pass along with...
3167 (pass_data_final): ...new pass_data instance and...
3168 (make_pass_final): ...new function.
3169 (pass_shorten_branches): Convert from a global struct to a subclass of
3170 rtl_opt_pass along with...
3171 (pass_data_shorten_branches): ...new pass_data instance and...
3172 (make_pass_shorten_branches): ...new function.
3173 (pass_clean_state): Convert from a global struct to a subclass of
3174 rtl_opt_pass along with...
3175 (pass_data_clean_state): ...new pass_data instance and...
3176 (make_pass_clean_state): ...new function.
3177 * function.c (pass_instantiate_virtual_regs): Convert from a global
3178 struct to a subclass of rtl_opt_pass along with...
3179 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
3180 (make_pass_instantiate_virtual_regs): ...new function.
3181 (pass_leaf_regs): Convert from a global struct to a subclass of
3182 rtl_opt_pass along with...
3183 (pass_data_leaf_regs): ...new pass_data instance and...
3184 (make_pass_leaf_regs): ...new function.
3185 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
3186 subclass of rtl_opt_pass along with...
3187 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
3189 (make_pass_thread_prologue_and_epilogue): ...new function.
3190 (pass_match_asm_constraints): Convert from a global struct to a
3191 subclass of rtl_opt_pass along with...
3192 (pass_data_match_asm_constraints): ...new pass_data instance and...
3193 (make_pass_match_asm_constraints): ...new function.
3194 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
3195 subclass of rtl_opt_pass along with...
3196 (pass_data_rtl_fwprop): ...new pass_data instance and...
3197 (make_pass_rtl_fwprop): ...new function.
3198 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
3199 rtl_opt_pass along with...
3200 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
3201 (make_pass_rtl_fwprop_addr): ...new function.
3202 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
3203 rtl_opt_pass along with...
3204 (pass_data_rtl_pre): ...new pass_data instance and...
3205 (make_pass_rtl_pre): ...new function.
3206 (pass_rtl_hoist): Convert from a global struct to a subclass of
3207 rtl_opt_pass along with...
3208 (pass_data_rtl_hoist): ...new pass_data instance and...
3209 (make_pass_rtl_hoist): ...new function.
3210 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
3211 subclass of gimple_opt_pass along with...
3212 (pass_data_lower_cf): ...new pass_data instance and...
3213 (make_pass_lower_cf): ...new function.
3214 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
3215 from a global struct to a subclass of gimple_opt_pass along with...
3216 (pass_data_strength_reduction): ...new pass_data instance and...
3217 (make_pass_strength_reduction): ...new function.
3218 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
3219 of rtl_opt_pass along with...
3220 (pass_data_rtl_ifcvt): ...new pass_data instance and...
3221 (make_pass_rtl_ifcvt): ...new function.
3222 (pass_if_after_combine): Convert from a global struct to a subclass of
3223 rtl_opt_pass along with...
3224 (pass_data_if_after_combine): ...new pass_data instance and...
3225 (make_pass_if_after_combine): ...new function.
3226 (pass_if_after_reload): Convert from a global struct to a subclass of
3227 rtl_opt_pass along with...
3228 (pass_data_if_after_reload): ...new pass_data instance and...
3229 (make_pass_if_after_reload): ...new function.
3230 * init-regs.c (pass_initialize_regs): Convert from a global struct to
3231 a subclass of rtl_opt_pass along with...
3232 (pass_data_initialize_regs): ...new pass_data instance and...
3233 (make_pass_initialize_regs): ...new function.
3234 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
3235 of ipa_opt_pass_d along with...
3236 (pass_data_ipa_cp): ...new pass_data instance and...
3237 (make_pass_ipa_cp): ...new function.
3238 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
3239 global struct to a subclass of gimple_opt_pass along with...
3240 (pass_data_inline_parameters): ...new pass_data instance and...
3241 (make_pass_inline_parameters): ...new function.
3242 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
3243 subclass of gimple_opt_pass along with...
3244 (pass_data_early_inline): ...new pass_data instance and...
3245 (make_pass_early_inline): ...new function.
3246 (pass_ipa_inline): Convert from a global struct to a subclass of
3247 ipa_opt_pass_d along with...
3248 (pass_data_ipa_inline): ...new pass_data instance and...
3249 (make_pass_ipa_inline): ...new function.
3250 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
3251 struct to a subclass of gimple_opt_pass along with...
3252 (pass_data_local_pure_const): ...new pass_data instance and...
3253 (make_pass_local_pure_const): ...new function.
3254 (pass_ipa_pure_const): Convert from a global struct to a subclass of
3255 ipa_opt_pass_d along with...
3256 (pass_data_ipa_pure_const): ...new pass_data instance and...
3257 (make_pass_ipa_pure_const): ...new function.
3258 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
3259 to a subclass of ipa_opt_pass_d along with...
3260 (pass_data_ipa_reference): ...new pass_data instance and...
3261 (make_pass_ipa_reference): ...new function.
3262 * ipa-split.c (pass_split_functions): Convert from a global struct to
3263 a subclass of gimple_opt_pass along with...
3264 (pass_data_split_functions): ...new pass_data instance and...
3265 (make_pass_split_functions): ...new function.
3266 (pass_feedback_split_functions): Convert from a global struct to a
3267 subclass of gimple_opt_pass along with...
3268 (pass_data_feedback_split_functions): ...new pass_data instance and...
3269 (make_pass_feedback_split_functions): ...new function.
3270 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
3271 global struct to a subclass of simple_ipa_opt_pass along with...
3272 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
3274 (make_pass_ipa_function_and_variable_visibility): ...new function.
3275 (pass_ipa_free_inline_summary): Convert from a global struct to a
3276 subclass of simple_ipa_opt_pass along with...
3277 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
3278 (make_pass_ipa_free_inline_summary): ...new function.
3279 (pass_ipa_whole_program_visibility): Convert from a global struct to a
3280 subclass of ipa_opt_pass_d along with...
3281 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
3283 (make_pass_ipa_whole_program_visibility): ...new function.
3284 (pass_ipa_profile): Convert from a global struct to a subclass of
3285 ipa_opt_pass_d along with...
3286 (pass_data_ipa_profile): ...new pass_data instance and...
3287 (make_pass_ipa_profile): ...new function.
3288 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
3289 ipa_opt_pass_d along with...
3290 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
3291 (make_pass_ipa_cdtor_merge): ...new function.
3292 * ira.c (pass_ira): Convert from a global struct to a subclass of
3293 rtl_opt_pass along with...
3294 (pass_data_ira): ...new pass_data instance and...
3295 (make_pass_ira): ...new function.
3296 (pass_reload): Convert from a global struct to a subclass of
3297 rtl_opt_pass along with...
3298 (pass_data_reload): ...new pass_data instance and...
3299 (make_pass_reload): ...new function.
3300 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
3301 subclass of rtl_opt_pass along with...
3302 (pass_data_cleanup_barriers): ...new pass_data instance and...
3303 (make_pass_cleanup_barriers): ...new function.
3304 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
3305 of rtl_opt_pass along with...
3306 (pass_data_loop2): ...new pass_data instance and...
3307 (make_pass_loop2): ...new function.
3308 (pass_rtl_loop_init): Convert from a global struct to a subclass of
3309 rtl_opt_pass along with...
3310 (pass_data_rtl_loop_init): ...new pass_data instance and...
3311 (make_pass_rtl_loop_init): ...new function.
3312 (pass_rtl_loop_done): Convert from a global struct to a subclass of
3313 rtl_opt_pass along with...
3314 (pass_data_rtl_loop_done): ...new pass_data instance and...
3315 (make_pass_rtl_loop_done): ...new function.
3316 (pass_rtl_move_loop_invariants): Convert from a global struct to a
3317 subclass of rtl_opt_pass along with...
3318 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
3319 (make_pass_rtl_move_loop_invariants): ...new function.
3320 (pass_rtl_unswitch): Convert from a global struct to a subclass of
3321 rtl_opt_pass along with...
3322 (pass_data_rtl_unswitch): ...new pass_data instance and...
3323 (make_pass_rtl_unswitch): ...new function.
3324 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
3325 subclass of rtl_opt_pass along with...
3326 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
3328 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3329 (pass_rtl_doloop): Convert from a global struct to a subclass of
3330 rtl_opt_pass along with...
3331 (pass_data_rtl_doloop): ...new pass_data instance and...
3332 (make_pass_rtl_doloop): ...new function.
3333 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
3334 a subclass of rtl_opt_pass along with...
3335 (pass_data_lower_subreg): ...new pass_data instance and...
3336 (make_pass_lower_subreg): ...new function.
3337 (pass_lower_subreg2): Convert from a global struct to a subclass of
3338 rtl_opt_pass along with...
3339 (pass_data_lower_subreg2): ...new pass_data instance and...
3340 (make_pass_lower_subreg2): ...new function.
3341 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
3342 struct to a subclass of ipa_opt_pass_d along with...
3343 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
3344 (make_pass_ipa_lto_gimple_out): ...new function.
3345 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
3346 of ipa_opt_pass_d along with...
3347 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
3348 (make_pass_ipa_lto_finish_out): ...new function.
3349 * mode-switching.c (pass_mode_switching): Convert from a global struct
3350 to a subclass of rtl_opt_pass along with...
3351 (pass_data_mode_switching): ...new pass_data instance and...
3352 (make_pass_mode_switching): ...new function.
3353 * modulo-sched.c (pass_sms): Convert from a global struct to a
3354 subclass of rtl_opt_pass along with...
3355 (pass_data_sms): ...new pass_data instance and...
3356 (make_pass_sms): ...new function.
3357 * omp-low.c (pass_expand_omp): Convert from a global struct to a
3358 subclass of gimple_opt_pass along with...
3359 (pass_data_expand_omp): ...new pass_data instance and...
3360 (make_pass_expand_omp): ...new function.
3361 (pass_lower_omp): Convert from a global struct to a subclass of
3362 gimple_opt_pass along with...
3363 (pass_data_lower_omp): ...new pass_data instance and...
3364 (make_pass_lower_omp): ...new function.
3365 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
3366 of gimple_opt_pass along with...
3367 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
3368 (make_pass_diagnose_omp_blocks): ...new function.
3369 * passes.c (pass_early_local_passes): Convert from a global struct to
3370 a subclass of simple_ipa_opt_pass along with...
3371 (pass_data_early_local_passes): ...new pass_data instance and...
3372 (make_pass_early_local_passes): ...new function.
3373 (pass_all_early_optimizations): Convert from a global struct to a
3374 subclass of gimple_opt_pass along with...
3375 (pass_data_all_early_optimizations): ...new pass_data instance and...
3376 (make_pass_all_early_optimizations): ...new function.
3377 (pass_all_optimizations): Convert from a global struct to a subclass
3378 of gimple_opt_pass along with...
3379 (pass_data_all_optimizations): ...new pass_data instance and...
3380 (make_pass_all_optimizations): ...new function.
3381 (pass_all_optimizations_g): Convert from a global struct to a subclass
3382 of gimple_opt_pass along with...
3383 (pass_data_all_optimizations_g): ...new pass_data instance and...
3384 (make_pass_all_optimizations_g): ...new function.
3385 (pass_rest_of_compilation): Convert from a global struct to a subclass
3386 of rtl_opt_pass along with...
3387 (pass_data_rest_of_compilation): ...new pass_data instance and...
3388 (make_pass_rest_of_compilation): ...new function.
3389 (pass_postreload): Convert from a global struct to a subclass of
3390 rtl_opt_pass along with...
3391 (pass_data_postreload): ...new pass_data instance and...
3392 (make_pass_postreload): ...new function.
3393 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
3394 subclass of rtl_opt_pass along with...
3395 (pass_data_gcse2): ...new pass_data instance and...
3396 (make_pass_gcse2): ...new function.
3397 * postreload.c (pass_postreload_cse): Convert from a global struct to
3398 a subclass of rtl_opt_pass along with...
3399 (pass_data_postreload_cse): ...new pass_data instance and...
3400 (make_pass_postreload_cse): ...new function.
3401 * predict.c (pass_profile): Convert from a global struct to a subclass
3402 of gimple_opt_pass along with...
3403 (pass_data_profile): ...new pass_data instance and...
3404 (make_pass_profile): ...new function.
3405 (pass_strip_predict_hints): Convert from a global struct to a subclass
3406 of gimple_opt_pass along with...
3407 (pass_data_strip_predict_hints): ...new pass_data instance and...
3408 (make_pass_strip_predict_hints): ...new function.
3409 * recog.c (pass_peephole2): Convert from a global struct to a subclass
3410 of rtl_opt_pass along with...
3411 (pass_data_peephole2): ...new pass_data instance and...
3412 (make_pass_peephole2): ...new function.
3413 (pass_split_all_insns): Convert from a global struct to a subclass of
3414 rtl_opt_pass along with...
3415 (pass_data_split_all_insns): ...new pass_data instance and...
3416 (make_pass_split_all_insns): ...new function.
3417 (pass_split_after_reload): Convert from a global struct to a subclass
3418 of rtl_opt_pass along with...
3419 (pass_data_split_after_reload): ...new pass_data instance and...
3420 (make_pass_split_after_reload): ...new function.
3421 (pass_split_before_regstack): Convert from a global struct to a
3422 subclass of rtl_opt_pass along with...
3423 (pass_data_split_before_regstack): ...new pass_data instance and...
3424 (make_pass_split_before_regstack): ...new function.
3425 (pass_split_before_sched2): Convert from a global struct to a subclass
3426 of rtl_opt_pass along with...
3427 (pass_data_split_before_sched2): ...new pass_data instance and...
3428 (make_pass_split_before_sched2): ...new function.
3429 (pass_split_for_shorten_branches): Convert from a global struct to a
3430 subclass of rtl_opt_pass along with...
3431 (pass_data_split_for_shorten_branches): ...new pass_data instance
3433 (make_pass_split_for_shorten_branches): ...new function.
3434 * ree.c (pass_ree): Convert from a global struct to a subclass of
3435 rtl_opt_pass along with...
3436 (pass_data_ree): ...new pass_data instance and...
3437 (make_pass_ree): ...new function.
3438 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
3439 subclass of rtl_opt_pass along with...
3440 (pass_data_stack_regs): ...new pass_data instance and...
3441 (make_pass_stack_regs): ...new function.
3442 (pass_stack_regs_run): Convert from a global struct to a subclass of
3443 rtl_opt_pass along with...
3444 (pass_data_stack_regs_run): ...new pass_data instance and...
3445 (make_pass_stack_regs_run): ...new function.
3446 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
3447 subclass of rtl_opt_pass along with...
3448 (pass_data_cprop_hardreg): ...new pass_data instance and...
3449 (make_pass_cprop_hardreg): ...new function.
3450 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
3451 subclass of rtl_opt_pass along with...
3452 (pass_data_reginfo_init): ...new pass_data instance and...
3453 (make_pass_reginfo_init): ...new function.
3454 * regmove.c (pass_regmove): Convert from a global struct to a subclass
3455 of rtl_opt_pass along with...
3456 (pass_data_regmove): ...new pass_data instance and...
3457 (make_pass_regmove): ...new function.
3458 * regrename.c (pass_regrename): Convert from a global struct to a
3459 subclass of rtl_opt_pass along with...
3460 (pass_data_regrename): ...new pass_data instance and...
3461 (make_pass_regrename): ...new function.
3462 * reorg.c (pass_delay_slots): Convert from a global struct to a
3463 subclass of rtl_opt_pass along with...
3464 (pass_data_delay_slots): ...new pass_data instance and...
3465 (make_pass_delay_slots): ...new function.
3466 (pass_machine_reorg): Convert from a global struct to a subclass of
3467 rtl_opt_pass along with...
3468 (pass_data_machine_reorg): ...new pass_data instance and...
3469 (make_pass_machine_reorg): ...new function.
3470 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
3471 of rtl_opt_pass along with...
3472 (pass_data_sched): ...new pass_data instance and...
3473 (make_pass_sched): ...new function.
3474 (pass_sched2): Convert from a global struct to a subclass of
3475 rtl_opt_pass along with...
3476 (pass_data_sched2): ...new pass_data instance and...
3477 (make_pass_sched2): ...new function.
3478 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
3479 to a subclass of rtl_opt_pass along with...
3480 (pass_data_stack_ptr_mod): ...new pass_data instance and...
3481 (make_pass_stack_ptr_mod): ...new function.
3482 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
3483 to a subclass of rtl_opt_pass along with...
3484 (pass_data_rtl_store_motion): ...new pass_data instance and...
3485 (make_pass_rtl_store_motion): ...new function.
3486 * tracer.c (pass_tracer): Convert from a global struct to a subclass
3487 of gimple_opt_pass along with...
3488 (pass_data_tracer): ...new pass_data instance and...
3489 (make_pass_tracer): ...new function.
3490 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
3491 to a subclass of gimple_opt_pass along with...
3492 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
3493 (make_pass_diagnose_tm_blocks): ...new function.
3494 (pass_lower_tm): Convert from a global struct to a subclass of
3495 gimple_opt_pass along with...
3496 (pass_data_lower_tm): ...new pass_data instance and...
3497 (make_pass_lower_tm): ...new function.
3498 (pass_tm_init): Convert from a global struct to a subclass of
3499 gimple_opt_pass along with...
3500 (pass_data_tm_init): ...new pass_data instance and...
3501 (make_pass_tm_init): ...new function.
3502 (pass_tm_mark): Convert from a global struct to a subclass of
3503 gimple_opt_pass along with...
3504 (pass_data_tm_mark): ...new pass_data instance and...
3505 (make_pass_tm_mark): ...new function.
3506 (pass_tm_edges): Convert from a global struct to a subclass of
3507 gimple_opt_pass along with...
3508 (pass_data_tm_edges): ...new pass_data instance and...
3509 (make_pass_tm_edges): ...new function.
3510 (pass_tm_memopt): Convert from a global struct to a subclass of
3511 gimple_opt_pass along with...
3512 (pass_data_tm_memopt): ...new pass_data instance and...
3513 (make_pass_tm_memopt): ...new function.
3514 (pass_ipa_tm): Convert from a global struct to a subclass of
3515 simple_ipa_opt_pass along with...
3516 (pass_data_ipa_tm): ...new pass_data instance and...
3517 (make_pass_ipa_tm): ...new function.
3518 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
3519 subclass of gimple_opt_pass along with...
3520 (pass_data_call_cdce): ...new pass_data instance and...
3521 (make_pass_call_cdce): ...new function.
3522 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
3523 subclass of gimple_opt_pass along with...
3524 (pass_data_build_cfg): ...new pass_data instance and...
3525 (make_pass_build_cfg): ...new function.
3526 (pass_split_crit_edges): Convert from a global struct to a subclass of
3527 gimple_opt_pass along with...
3528 (pass_data_split_crit_edges): ...new pass_data instance and...
3529 (make_pass_split_crit_edges): ...new function.
3530 (pass_warn_function_return): Convert from a global struct to a
3531 subclass of gimple_opt_pass along with...
3532 (pass_data_warn_function_return): ...new pass_data instance and...
3533 (make_pass_warn_function_return): ...new function.
3534 (pass_warn_function_noreturn): Convert from a global struct to a
3535 subclass of gimple_opt_pass along with...
3536 (pass_data_warn_function_noreturn): ...new pass_data instance and...
3537 (make_pass_warn_function_noreturn): ...new function.
3538 (pass_warn_unused_result): Convert from a global struct to a subclass
3539 of gimple_opt_pass along with...
3540 (pass_data_warn_unused_result): ...new pass_data instance and...
3541 (make_pass_warn_unused_result): ...new function.
3542 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
3543 a subclass of gimple_opt_pass along with...
3544 (pass_data_merge_phi): ...new pass_data instance and...
3545 (make_pass_merge_phi): ...new function.
3546 * tree-complex.c (pass_lower_complex): Convert from a global struct to
3547 a subclass of gimple_opt_pass along with...
3548 (pass_data_lower_complex): ...new pass_data instance and...
3549 (make_pass_lower_complex): ...new function.
3550 (pass_lower_complex_O0): Convert from a global struct to a subclass of
3551 gimple_opt_pass along with...
3552 (pass_data_lower_complex_O0): ...new pass_data instance and...
3553 (make_pass_lower_complex_O0): ...new function.
3554 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
3555 subclass of gimple_opt_pass along with...
3556 (pass_data_lower_eh): ...new pass_data instance and...
3557 (make_pass_lower_eh): ...new function.
3558 (pass_refactor_eh): Convert from a global struct to a subclass of
3559 gimple_opt_pass along with...
3560 (pass_data_refactor_eh): ...new pass_data instance and...
3561 (make_pass_refactor_eh): ...new function.
3562 (pass_lower_resx): Convert from a global struct to a subclass of
3563 gimple_opt_pass along with...
3564 (pass_data_lower_resx): ...new pass_data instance and...
3565 (make_pass_lower_resx): ...new function.
3566 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
3567 of gimple_opt_pass along with...
3568 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
3569 (make_pass_lower_eh_dispatch): ...new function.
3570 (pass_cleanup_eh): Convert from a global struct to a subclass of
3571 gimple_opt_pass along with...
3572 (pass_data_cleanup_eh): ...new pass_data instance and...
3573 (make_pass_cleanup_eh): ...new function.
3574 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
3575 to a subclass of simple_ipa_opt_pass along with...
3576 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
3577 (make_pass_ipa_lower_emutls): ...new function.
3578 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
3579 a subclass of gimple_opt_pass along with...
3580 (pass_data_if_conversion): ...new pass_data instance and...
3581 (make_pass_if_conversion): ...new function.
3582 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
3583 subclass of gimple_opt_pass along with...
3584 (pass_data_build_ssa): ...new pass_data instance and...
3585 (make_pass_build_ssa): ...new function.
3586 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
3587 global struct to a subclass of gimple_opt_pass along with...
3588 (pass_data_loop_distribution): ...new pass_data instance and...
3589 (make_pass_loop_distribution): ...new function.
3590 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
3591 subclass of gimple_opt_pass along with...
3592 (pass_data_mudflap_1): ...new pass_data instance and...
3593 (make_pass_mudflap_1): ...new function.
3594 (pass_mudflap_2): Convert from a global struct to a subclass of
3595 gimple_opt_pass along with...
3596 (pass_data_mudflap_2): ...new pass_data instance and...
3597 (make_pass_mudflap_2): ...new function.
3598 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
3599 subclass of gimple_opt_pass along with...
3600 (pass_data_mudflap_1): ...new pass_data instance and...
3601 (make_pass_mudflap_1): ...new function.
3602 (pass_mudflap_2): Convert from a global struct to a subclass of
3603 gimple_opt_pass along with...
3604 (pass_data_mudflap_2): ...new pass_data instance and...
3605 (make_pass_mudflap_2): ...new function.
3606 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
3607 gimple_opt_pass along with...
3608 (pass_data_nrv): ...new pass_data instance and...
3609 (make_pass_nrv): ...new function.
3610 (pass_return_slot): Convert from a global struct to a subclass of
3611 gimple_opt_pass along with...
3612 (pass_data_return_slot): ...new pass_data instance and...
3613 (make_pass_return_slot): ...new function.
3614 * tree-object-size.c (pass_object_sizes): Convert from a global struct
3615 to a subclass of gimple_opt_pass along with...
3616 (pass_data_object_sizes): ...new pass_data instance and...
3617 (make_pass_object_sizes): ...new function.
3618 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
3619 global struct to a subclass of gimple_opt_pass along with...
3620 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
3622 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3623 (pass_fixup_cfg): Convert from a global struct to a subclass of
3624 gimple_opt_pass along with...
3625 (pass_data_fixup_cfg): ...new pass_data instance and...
3626 (make_pass_fixup_cfg): ...new function.
3627 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
3628 (make_pass_mudflap_1): ...new function.
3629 (pass_mudflap_2): Replace declaration with that of...
3630 (make_pass_mudflap_2): ...new function.
3631 (pass_asan): Replace declaration with that of...
3632 (make_pass_asan): ...new function.
3633 (pass_asan_O0): Replace declaration with that of...
3634 (make_pass_asan_O0): ...new function.
3635 (pass_tsan): Replace declaration with that of...
3636 (make_pass_tsan): ...new function.
3637 (pass_tsan_O0): Replace declaration with that of...
3638 (make_pass_tsan_O0): ...new function.
3639 (pass_lower_cf): Replace declaration with that of...
3640 (make_pass_lower_cf): ...new function.
3641 (pass_refactor_eh): Replace declaration with that of...
3642 (make_pass_refactor_eh): ...new function.
3643 (pass_lower_eh): Replace declaration with that of...
3644 (make_pass_lower_eh): ...new function.
3645 (pass_lower_eh_dispatch): Replace declaration with that of...
3646 (make_pass_lower_eh_dispatch): ...new function.
3647 (pass_lower_resx): Replace declaration with that of...
3648 (make_pass_lower_resx): ...new function.
3649 (pass_build_cfg): Replace declaration with that of...
3650 (make_pass_build_cfg): ...new function.
3651 (pass_early_tree_profile): Replace declaration with that of...
3652 (make_pass_early_tree_profile): ...new function.
3653 (pass_cleanup_eh): Replace declaration with that of...
3654 (make_pass_cleanup_eh): ...new function.
3655 (pass_sra): Replace declaration with that of...
3656 (make_pass_sra): ...new function.
3657 (pass_sra_early): Replace declaration with that of...
3658 (make_pass_sra_early): ...new function.
3659 (pass_early_ipa_sra): Replace declaration with that of...
3660 (make_pass_early_ipa_sra): ...new function.
3661 (pass_tail_recursion): Replace declaration with that of...
3662 (make_pass_tail_recursion): ...new function.
3663 (pass_tail_calls): Replace declaration with that of...
3664 (make_pass_tail_calls): ...new function.
3665 (pass_tree_loop): Replace declaration with that of...
3666 (make_pass_tree_loop): ...new function.
3667 (pass_tree_loop_init): Replace declaration with that of...
3668 (make_pass_tree_loop_init): ...new function.
3669 (pass_lim): Replace declaration with that of...
3670 (make_pass_lim): ...new function.
3671 (pass_tree_unswitch): Replace declaration with that of...
3672 (make_pass_tree_unswitch): ...new function.
3673 (pass_predcom): Replace declaration with that of...
3674 (make_pass_predcom): ...new function.
3675 (pass_iv_canon): Replace declaration with that of...
3676 (make_pass_iv_canon): ...new function.
3677 (pass_scev_cprop): Replace declaration with that of...
3678 (make_pass_scev_cprop): ...new function.
3679 (pass_empty_loop): Replace declaration with that of...
3680 (make_pass_empty_loop): ...new function.
3681 (pass_record_bounds): Replace declaration with that of...
3682 (make_pass_record_bounds): ...new function.
3683 (pass_graphite): Replace declaration with that of...
3684 (make_pass_graphite): ...new function.
3685 (pass_graphite_transforms): Replace declaration with that of...
3686 (make_pass_graphite_transforms): ...new function.
3687 (pass_if_conversion): Replace declaration with that of...
3688 (make_pass_if_conversion): ...new function.
3689 (pass_loop_distribution): Replace declaration with that of...
3690 (make_pass_loop_distribution): ...new function.
3691 (pass_vectorize): Replace declaration with that of...
3692 (make_pass_vectorize): ...new function.
3693 (pass_slp_vectorize): Replace declaration with that of...
3694 (make_pass_slp_vectorize): ...new function.
3695 (pass_complete_unroll): Replace declaration with that of...
3696 (make_pass_complete_unroll): ...new function.
3697 (pass_complete_unrolli): Replace declaration with that of...
3698 (make_pass_complete_unrolli): ...new function.
3699 (pass_parallelize_loops): Replace declaration with that of...
3700 (make_pass_parallelize_loops): ...new function.
3701 (pass_loop_prefetch): Replace declaration with that of...
3702 (make_pass_loop_prefetch): ...new function.
3703 (pass_iv_optimize): Replace declaration with that of...
3704 (make_pass_iv_optimize): ...new function.
3705 (pass_tree_loop_done): Replace declaration with that of...
3706 (make_pass_tree_loop_done): ...new function.
3707 (pass_ch): Replace declaration with that of...
3708 (make_pass_ch): ...new function.
3709 (pass_ccp): Replace declaration with that of...
3710 (make_pass_ccp): ...new function.
3711 (pass_phi_only_cprop): Replace declaration with that of...
3712 (make_pass_phi_only_cprop): ...new function.
3713 (pass_build_ssa): Replace declaration with that of...
3714 (make_pass_build_ssa): ...new function.
3715 (pass_build_alias): Replace declaration with that of...
3716 (make_pass_build_alias): ...new function.
3717 (pass_build_ealias): Replace declaration with that of...
3718 (make_pass_build_ealias): ...new function.
3719 (pass_dominator): Replace declaration with that of...
3720 (make_pass_dominator): ...new function.
3721 (pass_dce): Replace declaration with that of...
3722 (make_pass_dce): ...new function.
3723 (pass_dce_loop): Replace declaration with that of...
3724 (make_pass_dce_loop): ...new function.
3725 (pass_cd_dce): Replace declaration with that of...
3726 (make_pass_cd_dce): ...new function.
3727 (pass_call_cdce): Replace declaration with that of...
3728 (make_pass_call_cdce): ...new function.
3729 (pass_merge_phi): Replace declaration with that of...
3730 (make_pass_merge_phi): ...new function.
3731 (pass_split_crit_edges): Replace declaration with that of...
3732 (make_pass_split_crit_edges): ...new function.
3733 (pass_pre): Replace declaration with that of...
3734 (make_pass_pre): ...new function.
3735 (pass_profile): Replace declaration with that of...
3736 (make_pass_profile): ...new function.
3737 (pass_strip_predict_hints): Replace declaration with that of...
3738 (make_pass_strip_predict_hints): ...new function.
3739 (pass_lower_complex_O0): Replace declaration with that of...
3740 (make_pass_lower_complex_O0): ...new function.
3741 (pass_lower_complex): Replace declaration with that of...
3742 (make_pass_lower_complex): ...new function.
3743 (pass_lower_vector): Replace declaration with that of...
3744 (make_pass_lower_vector): ...new function.
3745 (pass_lower_vector_ssa): Replace declaration with that of...
3746 (make_pass_lower_vector_ssa): ...new function.
3747 (pass_lower_omp): Replace declaration with that of...
3748 (make_pass_lower_omp): ...new function.
3749 (pass_diagnose_omp_blocks): Replace declaration with that of...
3750 (make_pass_diagnose_omp_blocks): ...new function.
3751 (pass_expand_omp): Replace declaration with that of...
3752 (make_pass_expand_omp): ...new function.
3753 (pass_expand_omp_ssa): Replace declaration with that of...
3754 (make_pass_expand_omp_ssa): ...new function.
3755 (pass_object_sizes): Replace declaration with that of...
3756 (make_pass_object_sizes): ...new function.
3757 (pass_strlen): Replace declaration with that of...
3758 (make_pass_strlen): ...new function.
3759 (pass_fold_builtins): Replace declaration with that of...
3760 (make_pass_fold_builtins): ...new function.
3761 (pass_stdarg): Replace declaration with that of...
3762 (make_pass_stdarg): ...new function.
3763 (pass_early_warn_uninitialized): Replace declaration with that of...
3764 (make_pass_early_warn_uninitialized): ...new function.
3765 (pass_late_warn_uninitialized): Replace declaration with that of...
3766 (make_pass_late_warn_uninitialized): ...new function.
3767 (pass_cse_reciprocals): Replace declaration with that of...
3768 (make_pass_cse_reciprocals): ...new function.
3769 (pass_cse_sincos): Replace declaration with that of...
3770 (make_pass_cse_sincos): ...new function.
3771 (pass_optimize_bswap): Replace declaration with that of...
3772 (make_pass_optimize_bswap): ...new function.
3773 (pass_optimize_widening_mul): Replace declaration with that of...
3774 (make_pass_optimize_widening_mul): ...new function.
3775 (pass_warn_function_return): Replace declaration with that of...
3776 (make_pass_warn_function_return): ...new function.
3777 (pass_warn_function_noreturn): Replace declaration with that of...
3778 (make_pass_warn_function_noreturn): ...new function.
3779 (pass_cselim): Replace declaration with that of...
3780 (make_pass_cselim): ...new function.
3781 (pass_phiopt): Replace declaration with that of...
3782 (make_pass_phiopt): ...new function.
3783 (pass_forwprop): Replace declaration with that of...
3784 (make_pass_forwprop): ...new function.
3785 (pass_phiprop): Replace declaration with that of...
3786 (make_pass_phiprop): ...new function.
3787 (pass_tree_ifcombine): Replace declaration with that of...
3788 (make_pass_tree_ifcombine): ...new function.
3789 (pass_dse): Replace declaration with that of...
3790 (make_pass_dse): ...new function.
3791 (pass_nrv): Replace declaration with that of...
3792 (make_pass_nrv): ...new function.
3793 (pass_rename_ssa_copies): Replace declaration with that of...
3794 (make_pass_rename_ssa_copies): ...new function.
3795 (pass_sink_code): Replace declaration with that of...
3796 (make_pass_sink_code): ...new function.
3797 (pass_fre): Replace declaration with that of...
3798 (make_pass_fre): ...new function.
3799 (pass_check_data_deps): Replace declaration with that of...
3800 (make_pass_check_data_deps): ...new function.
3801 (pass_copy_prop): Replace declaration with that of...
3802 (make_pass_copy_prop): ...new function.
3803 (pass_vrp): Replace declaration with that of...
3804 (make_pass_vrp): ...new function.
3805 (pass_uncprop): Replace declaration with that of...
3806 (make_pass_uncprop): ...new function.
3807 (pass_return_slot): Replace declaration with that of...
3808 (make_pass_return_slot): ...new function.
3809 (pass_reassoc): Replace declaration with that of...
3810 (make_pass_reassoc): ...new function.
3811 (pass_rebuild_cgraph_edges): Replace declaration with that of...
3812 (make_pass_rebuild_cgraph_edges): ...new function.
3813 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
3814 (make_pass_remove_cgraph_callee_edges): ...new function.
3815 (pass_build_cgraph_edges): Replace declaration with that of...
3816 (make_pass_build_cgraph_edges): ...new function.
3817 (pass_local_pure_const): Replace declaration with that of...
3818 (make_pass_local_pure_const): ...new function.
3819 (pass_tracer): Replace declaration with that of...
3820 (make_pass_tracer): ...new function.
3821 (pass_warn_unused_result): Replace declaration with that of...
3822 (make_pass_warn_unused_result): ...new function.
3823 (pass_diagnose_tm_blocks): Replace declaration with that of...
3824 (make_pass_diagnose_tm_blocks): ...new function.
3825 (pass_lower_tm): Replace declaration with that of...
3826 (make_pass_lower_tm): ...new function.
3827 (pass_tm_init): Replace declaration with that of...
3828 (make_pass_tm_init): ...new function.
3829 (pass_tm_mark): Replace declaration with that of...
3830 (make_pass_tm_mark): ...new function.
3831 (pass_tm_memopt): Replace declaration with that of...
3832 (make_pass_tm_memopt): ...new function.
3833 (pass_tm_edges): Replace declaration with that of...
3834 (make_pass_tm_edges): ...new function.
3835 (pass_split_functions): Replace declaration with that of...
3836 (make_pass_split_functions): ...new function.
3837 (pass_feedback_split_functions): Replace declaration with that of...
3838 (make_pass_feedback_split_functions): ...new function.
3839 (pass_strength_reduction): Replace declaration with that of...
3840 (make_pass_strength_reduction): ...new function.
3841 (pass_ipa_lower_emutls): Replace declaration with that of...
3842 (make_pass_ipa_lower_emutls): ...new function.
3843 (pass_ipa_function_and_variable_visibility): Replace declaration with
3845 (make_pass_ipa_function_and_variable_visibility): ...new function.
3846 (pass_ipa_tree_profile): Replace declaration with that of...
3847 (make_pass_ipa_tree_profile): ...new function.
3848 (pass_early_local_passes): Replace declaration with that of...
3849 (make_pass_early_local_passes): ...new function.
3850 (pass_ipa_whole_program_visibility): Replace declaration with that
3852 (make_pass_ipa_whole_program_visibility): ...new function.
3853 (pass_ipa_lto_gimple_out): Replace declaration with that of...
3854 (make_pass_ipa_lto_gimple_out): ...new function.
3855 (pass_ipa_increase_alignment): Replace declaration with that of...
3856 (make_pass_ipa_increase_alignment): ...new function.
3857 (pass_ipa_inline): Replace declaration with that of...
3858 (make_pass_ipa_inline): ...new function.
3859 (pass_ipa_free_lang_data): Replace declaration with that of...
3860 (make_pass_ipa_free_lang_data): ...new function.
3861 (pass_ipa_free_inline_summary): Replace declaration with that of...
3862 (make_pass_ipa_free_inline_summary): ...new function.
3863 (pass_ipa_cp): Replace declaration with that of...
3864 (make_pass_ipa_cp): ...new function.
3865 (pass_ipa_reference): Replace declaration with that of...
3866 (make_pass_ipa_reference): ...new function.
3867 (pass_ipa_pure_const): Replace declaration with that of...
3868 (make_pass_ipa_pure_const): ...new function.
3869 (pass_ipa_pta): Replace declaration with that of...
3870 (make_pass_ipa_pta): ...new function.
3871 (pass_ipa_lto_finish_out): Replace declaration with that of...
3872 (make_pass_ipa_lto_finish_out): ...new function.
3873 (pass_ipa_tm): Replace declaration with that of...
3874 (make_pass_ipa_tm): ...new function.
3875 (pass_ipa_profile): Replace declaration with that of...
3876 (make_pass_ipa_profile): ...new function.
3877 (pass_ipa_cdtor_merge): Replace declaration with that of...
3878 (make_pass_ipa_cdtor_merge): ...new function.
3879 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
3881 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3882 (pass_init_datastructures): Replace declaration with that of...
3883 (make_pass_init_datastructures): ...new function.
3884 (pass_fixup_cfg): Replace declaration with that of...
3885 (make_pass_fixup_cfg): ...new function.
3886 (pass_expand): Replace declaration with that of...
3887 (make_pass_expand): ...new function.
3888 (pass_instantiate_virtual_regs): Replace declaration with that of...
3889 (make_pass_instantiate_virtual_regs): ...new function.
3890 (pass_rtl_fwprop): Replace declaration with that of...
3891 (make_pass_rtl_fwprop): ...new function.
3892 (pass_rtl_fwprop_addr): Replace declaration with that of...
3893 (make_pass_rtl_fwprop_addr): ...new function.
3894 (pass_jump): Replace declaration with that of...
3895 (make_pass_jump): ...new function.
3896 (pass_jump2): Replace declaration with that of...
3897 (make_pass_jump2): ...new function.
3898 (pass_lower_subreg): Replace declaration with that of...
3899 (make_pass_lower_subreg): ...new function.
3900 (pass_cse): Replace declaration with that of...
3901 (make_pass_cse): ...new function.
3902 (pass_fast_rtl_dce): Replace declaration with that of...
3903 (make_pass_fast_rtl_dce): ...new function.
3904 (pass_ud_rtl_dce): Replace declaration with that of...
3905 (make_pass_ud_rtl_dce): ...new function.
3906 (pass_rtl_dce): Replace declaration with that of...
3907 (make_pass_rtl_dce): ...new function.
3908 (pass_rtl_dse1): Replace declaration with that of...
3909 (make_pass_rtl_dse1): ...new function.
3910 (pass_rtl_dse2): Replace declaration with that of...
3911 (make_pass_rtl_dse2): ...new function.
3912 (pass_rtl_dse3): Replace declaration with that of...
3913 (make_pass_rtl_dse3): ...new function.
3914 (pass_rtl_cprop): Replace declaration with that of...
3915 (make_pass_rtl_cprop): ...new function.
3916 (pass_rtl_pre): Replace declaration with that of...
3917 (make_pass_rtl_pre): ...new function.
3918 (pass_rtl_hoist): Replace declaration with that of...
3919 (make_pass_rtl_hoist): ...new function.
3920 (pass_rtl_store_motion): Replace declaration with that of...
3921 (make_pass_rtl_store_motion): ...new function.
3922 (pass_cse_after_global_opts): Replace declaration with that of...
3923 (make_pass_cse_after_global_opts): ...new function.
3924 (pass_rtl_ifcvt): Replace declaration with that of...
3925 (make_pass_rtl_ifcvt): ...new function.
3926 (pass_into_cfg_layout_mode): Replace declaration with that of...
3927 (make_pass_into_cfg_layout_mode): ...new function.
3928 (pass_outof_cfg_layout_mode): Replace declaration with that of...
3929 (make_pass_outof_cfg_layout_mode): ...new function.
3930 (pass_loop2): Replace declaration with that of...
3931 (make_pass_loop2): ...new function.
3932 (pass_rtl_loop_init): Replace declaration with that of...
3933 (make_pass_rtl_loop_init): ...new function.
3934 (pass_rtl_move_loop_invariants): Replace declaration with that of...
3935 (make_pass_rtl_move_loop_invariants): ...new function.
3936 (pass_rtl_unswitch): Replace declaration with that of...
3937 (make_pass_rtl_unswitch): ...new function.
3938 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
3939 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3940 (pass_rtl_doloop): Replace declaration with that of...
3941 (make_pass_rtl_doloop): ...new function.
3942 (pass_rtl_loop_done): Replace declaration with that of...
3943 (make_pass_rtl_loop_done): ...new function.
3944 (pass_web): Replace declaration with that of...
3945 (make_pass_web): ...new function.
3946 (pass_cse2): Replace declaration with that of...
3947 (make_pass_cse2): ...new function.
3948 (pass_df_initialize_opt): Replace declaration with that of...
3949 (make_pass_df_initialize_opt): ...new function.
3950 (pass_df_initialize_no_opt): Replace declaration with that of...
3951 (make_pass_df_initialize_no_opt): ...new function.
3952 (pass_reginfo_init): Replace declaration with that of...
3953 (make_pass_reginfo_init): ...new function.
3954 (pass_inc_dec): Replace declaration with that of...
3955 (make_pass_inc_dec): ...new function.
3956 (pass_stack_ptr_mod): Replace declaration with that of...
3957 (make_pass_stack_ptr_mod): ...new function.
3958 (pass_initialize_regs): Replace declaration with that of...
3959 (make_pass_initialize_regs): ...new function.
3960 (pass_combine): Replace declaration with that of...
3961 (make_pass_combine): ...new function.
3962 (pass_if_after_combine): Replace declaration with that of...
3963 (make_pass_if_after_combine): ...new function.
3964 (pass_ree): Replace declaration with that of...
3965 (make_pass_ree): ...new function.
3966 (pass_partition_blocks): Replace declaration with that of...
3967 (make_pass_partition_blocks): ...new function.
3968 (pass_match_asm_constraints): Replace declaration with that of...
3969 (make_pass_match_asm_constraints): ...new function.
3970 (pass_regmove): Replace declaration with that of...
3971 (make_pass_regmove): ...new function.
3972 (pass_split_all_insns): Replace declaration with that of...
3973 (make_pass_split_all_insns): ...new function.
3974 (pass_fast_rtl_byte_dce): Replace declaration with that of...
3975 (make_pass_fast_rtl_byte_dce): ...new function.
3976 (pass_lower_subreg2): Replace declaration with that of...
3977 (make_pass_lower_subreg2): ...new function.
3978 (pass_mode_switching): Replace declaration with that of...
3979 (make_pass_mode_switching): ...new function.
3980 (pass_sms): Replace declaration with that of...
3981 (make_pass_sms): ...new function.
3982 (pass_sched): Replace declaration with that of...
3983 (make_pass_sched): ...new function.
3984 (pass_ira): Replace declaration with that of...
3985 (make_pass_ira): ...new function.
3986 (pass_reload): Replace declaration with that of...
3987 (make_pass_reload): ...new function.
3988 (pass_clean_state): Replace declaration with that of...
3989 (make_pass_clean_state): ...new function.
3990 (pass_branch_prob): Replace declaration with that of...
3991 (make_pass_branch_prob): ...new function.
3992 (pass_value_profile_transformations): Replace declaration with that
3994 (make_pass_value_profile_transformations): ...new function.
3995 (pass_postreload_cse): Replace declaration with that of...
3996 (make_pass_postreload_cse): ...new function.
3997 (pass_gcse2): Replace declaration with that of...
3998 (make_pass_gcse2): ...new function.
3999 (pass_split_after_reload): Replace declaration with that of...
4000 (make_pass_split_after_reload): ...new function.
4001 (pass_branch_target_load_optimize1): Replace declaration with that
4003 (make_pass_branch_target_load_optimize1): ...new function.
4004 (pass_thread_prologue_and_epilogue): Replace declaration with that
4006 (make_pass_thread_prologue_and_epilogue): ...new function.
4007 (pass_stack_adjustments): Replace declaration with that of...
4008 (make_pass_stack_adjustments): ...new function.
4009 (pass_peephole2): Replace declaration with that of...
4010 (make_pass_peephole2): ...new function.
4011 (pass_if_after_reload): Replace declaration with that of...
4012 (make_pass_if_after_reload): ...new function.
4013 (pass_regrename): Replace declaration with that of...
4014 (make_pass_regrename): ...new function.
4015 (pass_cprop_hardreg): Replace declaration with that of...
4016 (make_pass_cprop_hardreg): ...new function.
4017 (pass_reorder_blocks): Replace declaration with that of...
4018 (make_pass_reorder_blocks): ...new function.
4019 (pass_branch_target_load_optimize2): Replace declaration with that
4021 (make_pass_branch_target_load_optimize2): ...new function.
4022 (pass_leaf_regs): Replace declaration with that of...
4023 (make_pass_leaf_regs): ...new function.
4024 (pass_split_before_sched2): Replace declaration with that of...
4025 (make_pass_split_before_sched2): ...new function.
4026 (pass_compare_elim_after_reload): Replace declaration with that of...
4027 (make_pass_compare_elim_after_reload): ...new function.
4028 (pass_sched2): Replace declaration with that of...
4029 (make_pass_sched2): ...new function.
4030 (pass_stack_regs): Replace declaration with that of...
4031 (make_pass_stack_regs): ...new function.
4032 (pass_stack_regs_run): Replace declaration with that of...
4033 (make_pass_stack_regs_run): ...new function.
4034 (pass_df_finish): Replace declaration with that of...
4035 (make_pass_df_finish): ...new function.
4036 (pass_compute_alignments): Replace declaration with that of...
4037 (make_pass_compute_alignments): ...new function.
4038 (pass_duplicate_computed_gotos): Replace declaration with that of...
4039 (make_pass_duplicate_computed_gotos): ...new function.
4040 (pass_variable_tracking): Replace declaration with that of...
4041 (make_pass_variable_tracking): ...new function.
4042 (pass_free_cfg): Replace declaration with that of...
4043 (make_pass_free_cfg): ...new function.
4044 (pass_machine_reorg): Replace declaration with that of...
4045 (make_pass_machine_reorg): ...new function.
4046 (pass_cleanup_barriers): Replace declaration with that of...
4047 (make_pass_cleanup_barriers): ...new function.
4048 (pass_delay_slots): Replace declaration with that of...
4049 (make_pass_delay_slots): ...new function.
4050 (pass_split_for_shorten_branches): Replace declaration with that of...
4051 (make_pass_split_for_shorten_branches): ...new function.
4052 (pass_split_before_regstack): Replace declaration with that of...
4053 (make_pass_split_before_regstack): ...new function.
4054 (pass_convert_to_eh_region_ranges): Replace declaration with that
4056 (make_pass_convert_to_eh_region_ranges): ...new function.
4057 (pass_shorten_branches): Replace declaration with that of...
4058 (make_pass_shorten_branches): ...new function.
4059 (pass_set_nothrow_function_flags): Replace declaration with that of...
4060 (make_pass_set_nothrow_function_flags): ...new function.
4061 (pass_dwarf2_frame): Replace declaration with that of...
4062 (make_pass_dwarf2_frame): ...new function.
4063 (pass_final): Replace declaration with that of...
4064 (make_pass_final): ...new function.
4065 (pass_rtl_seqabstr): Replace declaration with that of...
4066 (make_pass_rtl_seqabstr): ...new function.
4067 (pass_release_ssa_names): Replace declaration with that of...
4068 (make_pass_release_ssa_names): ...new function.
4069 (pass_early_inline): Replace declaration with that of...
4070 (make_pass_early_inline): ...new function.
4071 (pass_inline_parameters): Replace declaration with that of...
4072 (make_pass_inline_parameters): ...new function.
4073 (pass_update_address_taken): Replace declaration with that of...
4074 (make_pass_update_address_taken): ...new function.
4075 (pass_convert_switch): Replace declaration with that of...
4076 (make_pass_convert_switch): ...new function.
4077 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
4078 to a subclass of simple_ipa_opt_pass along with...
4079 (pass_data_ipa_tree_profile): ...new pass_data instance and...
4080 (make_pass_ipa_tree_profile): ...new function.
4081 * tree-sra.c (pass_sra_early): Convert from a global struct to a
4082 subclass of gimple_opt_pass along with...
4083 (pass_data_sra_early): ...new pass_data instance and...
4084 (make_pass_sra_early): ...new function.
4085 (pass_sra): Convert from a global struct to a subclass of
4086 gimple_opt_pass along with...
4087 (pass_data_sra): ...new pass_data instance and...
4088 (make_pass_sra): ...new function.
4089 (pass_early_ipa_sra): Convert from a global struct to a subclass of
4090 gimple_opt_pass along with...
4091 (pass_data_early_ipa_sra): ...new pass_data instance and...
4092 (make_pass_early_ipa_sra): ...new function.
4093 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
4094 subclass of gimple_opt_pass along with...
4095 (pass_data_ccp): ...new pass_data instance and...
4096 (make_pass_ccp): ...new function.
4097 (pass_fold_builtins): Convert from a global struct to a subclass of
4098 gimple_opt_pass along with...
4099 (pass_data_fold_builtins): ...new pass_data instance and...
4100 (make_pass_fold_builtins): ...new function.
4101 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
4102 subclass of gimple_opt_pass along with...
4103 (pass_data_copy_prop): ...new pass_data instance and...
4104 (make_pass_copy_prop): ...new function.
4105 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
4106 global struct to a subclass of gimple_opt_pass along with...
4107 (pass_data_rename_ssa_copies): ...new pass_data instance and...
4108 (make_pass_rename_ssa_copies): ...new function.
4109 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
4110 subclass of gimple_opt_pass along with...
4111 (pass_data_dce): ...new pass_data instance and...
4112 (make_pass_dce): ...new function.
4113 (pass_dce_loop): Convert from a global struct to a subclass of
4114 gimple_opt_pass along with...
4115 (pass_data_dce_loop): ...new pass_data instance and...
4116 (make_pass_dce_loop): ...new function.
4117 (pass_cd_dce): Convert from a global struct to a subclass of
4118 gimple_opt_pass along with...
4119 (pass_data_cd_dce): ...new pass_data instance and...
4120 (make_pass_cd_dce): ...new function.
4121 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
4122 subclass of gimple_opt_pass along with...
4123 (pass_data_dominator): ...new pass_data instance and...
4124 (make_pass_dominator): ...new function.
4125 (pass_phi_only_cprop): Convert from a global struct to a subclass of
4126 gimple_opt_pass along with...
4127 (pass_data_phi_only_cprop): ...new pass_data instance and...
4128 (make_pass_phi_only_cprop): ...new function.
4129 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
4130 subclass of gimple_opt_pass along with...
4131 (pass_data_dse): ...new pass_data instance and...
4132 (make_pass_dse): ...new function.
4133 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
4134 a subclass of gimple_opt_pass along with...
4135 (pass_data_forwprop): ...new pass_data instance and...
4136 (make_pass_forwprop): ...new function.
4137 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
4138 struct to a subclass of gimple_opt_pass along with...
4139 (pass_data_tree_ifcombine): ...new pass_data instance and...
4140 (make_pass_tree_ifcombine): ...new function.
4141 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
4142 subclass of gimple_opt_pass along with...
4143 (pass_data_ch): ...new pass_data instance and...
4144 (make_pass_ch): ...new function.
4145 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
4146 subclass of gimple_opt_pass along with...
4147 (pass_data_tree_loop): ...new pass_data instance and...
4148 (make_pass_tree_loop): ...new function.
4149 (pass_tree_loop_init): Convert from a global struct to a subclass of
4150 gimple_opt_pass along with...
4151 (pass_data_tree_loop_init): ...new pass_data instance and...
4152 (make_pass_tree_loop_init): ...new function.
4153 (pass_lim): Convert from a global struct to a subclass of
4154 gimple_opt_pass along with...
4155 (pass_data_lim): ...new pass_data instance and...
4156 (make_pass_lim): ...new function.
4157 (pass_tree_unswitch): Convert from a global struct to a subclass of
4158 gimple_opt_pass along with...
4159 (pass_data_tree_unswitch): ...new pass_data instance and...
4160 (make_pass_tree_unswitch): ...new function.
4161 (pass_predcom): Convert from a global struct to a subclass of
4162 gimple_opt_pass along with...
4163 (pass_data_predcom): ...new pass_data instance and...
4164 (make_pass_predcom): ...new function.
4165 (pass_vectorize): Convert from a global struct to a subclass of
4166 gimple_opt_pass along with...
4167 (pass_data_vectorize): ...new pass_data instance and...
4168 (make_pass_vectorize): ...new function.
4169 (pass_graphite): Convert from a global struct to a subclass of
4170 gimple_opt_pass along with...
4171 (pass_data_graphite): ...new pass_data instance and...
4172 (make_pass_graphite): ...new function.
4173 (pass_graphite_transforms): Convert from a global struct to a subclass
4174 of gimple_opt_pass along with...
4175 (pass_data_graphite_transforms): ...new pass_data instance and...
4176 (make_pass_graphite_transforms): ...new function.
4177 (pass_check_data_deps): Convert from a global struct to a subclass of
4178 gimple_opt_pass along with...
4179 (pass_data_check_data_deps): ...new pass_data instance and...
4180 (make_pass_check_data_deps): ...new function.
4181 (pass_iv_canon): Convert from a global struct to a subclass of
4182 gimple_opt_pass along with...
4183 (pass_data_iv_canon): ...new pass_data instance and...
4184 (make_pass_iv_canon): ...new function.
4185 (pass_scev_cprop): Convert from a global struct to a subclass of
4186 gimple_opt_pass along with...
4187 (pass_data_scev_cprop): ...new pass_data instance and...
4188 (make_pass_scev_cprop): ...new function.
4189 (pass_record_bounds): Convert from a global struct to a subclass of
4190 gimple_opt_pass along with...
4191 (pass_data_record_bounds): ...new pass_data instance and...
4192 (make_pass_record_bounds): ...new function.
4193 (pass_complete_unroll): Convert from a global struct to a subclass of
4194 gimple_opt_pass along with...
4195 (pass_data_complete_unroll): ...new pass_data instance and...
4196 (make_pass_complete_unroll): ...new function.
4197 (pass_complete_unrolli): Convert from a global struct to a subclass of
4198 gimple_opt_pass along with...
4199 (pass_data_complete_unrolli): ...new pass_data instance and...
4200 (make_pass_complete_unrolli): ...new function.
4201 (pass_parallelize_loops): Convert from a global struct to a subclass
4202 of gimple_opt_pass along with...
4203 (pass_data_parallelize_loops): ...new pass_data instance and...
4204 (make_pass_parallelize_loops): ...new function.
4205 (pass_loop_prefetch): Convert from a global struct to a subclass of
4206 gimple_opt_pass along with...
4207 (pass_data_loop_prefetch): ...new pass_data instance and...
4208 (make_pass_loop_prefetch): ...new function.
4209 (pass_iv_optimize): Convert from a global struct to a subclass of
4210 gimple_opt_pass along with...
4211 (pass_data_iv_optimize): ...new pass_data instance and...
4212 (make_pass_iv_optimize): ...new function.
4213 (pass_tree_loop_done): Convert from a global struct to a subclass of
4214 gimple_opt_pass along with...
4215 (pass_data_tree_loop_done): ...new pass_data instance and...
4216 (make_pass_tree_loop_done): ...new function.
4217 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
4218 struct to a subclass of gimple_opt_pass along with...
4219 (pass_data_cse_reciprocals): ...new pass_data instance and...
4220 (make_pass_cse_reciprocals): ...new function.
4221 (pass_cse_sincos): Convert from a global struct to a subclass of
4222 gimple_opt_pass along with...
4223 (pass_data_cse_sincos): ...new pass_data instance and...
4224 (make_pass_cse_sincos): ...new function.
4225 (pass_optimize_bswap): Convert from a global struct to a subclass of
4226 gimple_opt_pass along with...
4227 (pass_data_optimize_bswap): ...new pass_data instance and...
4228 (make_pass_optimize_bswap): ...new function.
4229 (pass_optimize_widening_mul): Convert from a global struct to a
4230 subclass of gimple_opt_pass along with...
4231 (pass_data_optimize_widening_mul): ...new pass_data instance and...
4232 (make_pass_optimize_widening_mul): ...new function.
4233 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
4234 subclass of gimple_opt_pass along with...
4235 (pass_data_phiopt): ...new pass_data instance and...
4236 (make_pass_phiopt): ...new function.
4237 (pass_cselim): Convert from a global struct to a subclass of
4238 gimple_opt_pass along with...
4239 (pass_data_cselim): ...new pass_data instance and...
4240 (make_pass_cselim): ...new function.
4241 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
4242 subclass of gimple_opt_pass along with...
4243 (pass_data_phiprop): ...new pass_data instance and...
4244 (make_pass_phiprop): ...new function.
4245 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
4246 subclass of gimple_opt_pass along with...
4247 (pass_data_pre): ...new pass_data instance and...
4248 (make_pass_pre): ...new function.
4249 (pass_fre): Convert from a global struct to a subclass of
4250 gimple_opt_pass along with...
4251 (pass_data_fre): ...new pass_data instance and...
4252 (make_pass_fre): ...new function.
4253 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
4254 subclass of gimple_opt_pass along with...
4255 (pass_data_reassoc): ...new pass_data instance and...
4256 (make_pass_reassoc): ...new function.
4257 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
4258 subclass of gimple_opt_pass along with...
4259 (pass_data_sink_code): ...new pass_data instance and...
4260 (make_pass_sink_code): ...new function.
4261 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
4262 subclass of gimple_opt_pass along with...
4263 (pass_data_strlen): ...new pass_data instance and...
4264 (make_pass_strlen): ...new function.
4265 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
4266 struct to a subclass of gimple_opt_pass along with...
4267 (pass_data_build_alias): ...new pass_data instance and...
4268 (make_pass_build_alias): ...new function.
4269 (pass_build_ealias): Convert from a global struct to a subclass of
4270 gimple_opt_pass along with...
4271 (pass_data_build_ealias): ...new pass_data instance and...
4272 (make_pass_build_ealias): ...new function.
4273 (pass_ipa_pta): Convert from a global struct to a subclass of
4274 simple_ipa_opt_pass along with...
4275 (pass_data_ipa_pta): ...new pass_data instance and...
4276 (make_pass_ipa_pta): ...new function.
4277 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
4278 subclass of gimple_opt_pass along with...
4279 (pass_data_uncprop): ...new pass_data instance and...
4280 (make_pass_uncprop): ...new function.
4281 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
4282 global struct to a subclass of gimple_opt_pass along with...
4283 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
4284 (make_pass_late_warn_uninitialized): ...new function.
4285 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
4286 to a subclass of gimple_opt_pass along with...
4287 (pass_data_init_datastructures): ...new pass_data instance and...
4288 (make_pass_init_datastructures): ...new function.
4289 (pass_early_warn_uninitialized): Convert from a global struct to a
4290 subclass of gimple_opt_pass along with...
4291 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
4292 (make_pass_early_warn_uninitialized): ...new function.
4293 (pass_update_address_taken): Convert from a global struct to a
4294 subclass of gimple_opt_pass along with...
4295 (pass_data_update_address_taken): ...new pass_data instance and...
4296 (make_pass_update_address_taken): ...new function.
4297 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
4298 struct to a subclass of gimple_opt_pass along with...
4299 (pass_data_release_ssa_names): ...new pass_data instance and...
4300 (make_pass_release_ssa_names): ...new function.
4301 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
4302 subclass of gimple_opt_pass along with...
4303 (pass_data_stdarg): ...new pass_data instance and...
4304 (make_pass_stdarg): ...new function.
4305 * tree-switch-conversion.c (pass_convert_switch): Convert from a
4306 global struct to a subclass of gimple_opt_pass along with...
4307 (pass_data_convert_switch): ...new pass_data instance and...
4308 (make_pass_convert_switch): ...new function.
4309 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
4310 to a subclass of gimple_opt_pass along with...
4311 (pass_data_tail_recursion): ...new pass_data instance and...
4312 (make_pass_tail_recursion): ...new function.
4313 (pass_tail_calls): Convert from a global struct to a subclass of
4314 gimple_opt_pass along with...
4315 (pass_data_tail_calls): ...new pass_data instance and...
4316 (make_pass_tail_calls): ...new function.
4317 * tree-vect-generic.c (pass_lower_vector): Convert from a global
4318 struct to a subclass of gimple_opt_pass along with...
4319 (pass_data_lower_vector): ...new pass_data instance and...
4320 (make_pass_lower_vector): ...new function.
4321 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
4322 gimple_opt_pass along with...
4323 (pass_data_lower_vector_ssa): ...new pass_data instance and...
4324 (make_pass_lower_vector_ssa): ...new function.
4325 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
4326 to a subclass of gimple_opt_pass along with...
4327 (pass_data_slp_vectorize): ...new pass_data instance and...
4328 (make_pass_slp_vectorize): ...new function.
4329 (pass_ipa_increase_alignment): Convert from a global struct to a
4330 subclass of simple_ipa_opt_pass along with...
4331 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
4332 (make_pass_ipa_increase_alignment): ...new function.
4333 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
4334 gimple_opt_pass along with...
4335 (pass_data_vrp): ...new pass_data instance and...
4336 (make_pass_vrp): ...new function.
4337 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
4338 subclass of simple_ipa_opt_pass along with...
4339 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
4340 (make_pass_ipa_free_lang_data): ...new function.
4341 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
4342 gimple_opt_pass along with...
4343 (pass_data_tsan): ...new pass_data instance and...
4344 (make_pass_tsan): ...new function.
4345 (pass_tsan_O0): Convert from a global struct to a subclass of
4346 gimple_opt_pass along with...
4347 (pass_data_tsan_O0): ...new pass_data instance and...
4348 (make_pass_tsan_O0): ...new function.
4349 * var-tracking.c (pass_variable_tracking): Convert from a global
4350 struct to a subclass of rtl_opt_pass along with...
4351 (pass_data_variable_tracking): ...new pass_data instance and...
4352 (make_pass_variable_tracking): ...new function.
4353 * web.c (pass_web): Convert from a global struct to a subclass of
4354 rtl_opt_pass along with...
4355 (pass_data_web): ...new pass_data instance and...
4356 (make_pass_web): ...new function.
4357 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
4358 declaration with that of...
4359 (make_pass_mode_switch_use): ...new function.
4360 (pass_resolve_sw_modes): Replace declaration with that of...
4361 (make_pass_resolve_sw_modes): ...new function.
4362 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
4363 from a global struct to a subclass of rtl_opt_pass along with...
4364 (pass_data_mode_switch_use): ...new pass_data instance and...
4365 (make_pass_mode_switch_use): ...new function.
4366 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
4367 from a global struct to a subclass of rtl_opt_pass along with...
4368 (pass_data_resolve_sw_modes): ...new pass_data instance and...
4369 (make_pass_resolve_sw_modes): ...new function.
4370 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
4371 struct to a subclass of rtl_opt_pass along with...
4372 (pass_data_insert_vzeroupper): ...new pass_data instance and...
4373 (make_pass_insert_vzeroupper): ...new function.
4374 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
4375 global struct to a subclass of rtl_opt_pass along with...
4376 (pass_data_work_around_errata): ...new pass_data instance and...
4377 (make_pass_work_around_errata): ...new function.
4378 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
4379 struct to a subclass of rtl_opt_pass along with...
4380 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
4381 (make_pass_mips_machine_reorg2): ...new function.
4383 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4385 * passes.c (pass_manager::operator new): New.
4387 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4389 Handwritten part of conversion of passes to C++ classes.
4391 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
4392 (toplev.o): Add dep on PASS_MANAGER_H.
4393 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
4394 of early local pases to reflect this moving from a global to a
4395 member of gcc::pass_manager.
4396 (cgraph_add_new_function): Likewise.
4397 * lto-cgraph.c (lto_output_node): Update for conversion of
4398 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
4399 * passes.c (opt_pass::clone): New.
4400 (opt_pass::gate): New.
4401 (opt_pass::execute): New.
4402 (opt_pass::opt_pass): New.
4403 (pass_manager::execute_early_local_passes): New.
4404 (pass_manager::execute_pass_mode_switching): new.
4405 (finish_optimization_passes): Convert to...
4406 (pass_manager::finish_optimization_passes): ...this.
4407 (finish_optimization_passes): Update for conversion of passes to
4409 (register_dump_files_1): Use has_gate since we cannot portably
4410 check a vtable entry against NULL.
4411 (dump_one_pass): Likewise.
4412 (ipa_write_summaries_2): Likewise.
4413 (ipa_write_optimization_summaries_1): Likewise.
4414 (ipa_read_summaries_1): Likewise.
4415 (ipa_read_optimization_summaries_1): Likewise.
4416 (execute_ipa_stmt_fixups): Likewise.
4417 (pass_manager::pass_manager): Rewrite pass-creation, invoking
4418 pass-creation functions rather than wiring up globals, and
4419 storing the results in fields of pass_manager generated using
4421 (pass_manager::dump_profile_report): Update for conversion of
4422 passes to C++ classes.
4423 (pass_manager::execute_ipa_summary_passes): Likewise.
4424 (execute_one_ipa_transform_pass): Likewise.
4425 (execute_one_pass): Use has_gate and has_execute since we cannot
4426 portably check a vtable entry against NULL.
4427 * pass_manager.h (pass_manager::finish_optimization_passes): New.
4428 (pass_manager): Use pass-instances.def to add fields for the
4429 various pass instances.
4430 * toplev.c (finalize): Update for move of
4431 finish_optimization_passes to a method of gcc::pass_manager.
4432 * toplev.h (finish_optimization_passes): Move to method of class
4434 * tree-pass.h (struct pass_data): New.
4435 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
4436 (opt_pass::gate): Convert to virtual function.
4437 (opt_pass::~opt_pass): New.
4438 (opt_pass::clone): New.
4439 (opt_pass::execute): Convert to virtual function.
4440 (opt_pass::opt_pass): New.
4441 (opt_pass::ctxt_): new.
4442 (gimple_opt_pass): Convert to subclass of opt_pass.
4443 (gimple_opt_pass::gimple_opt_pass): New.
4444 (rtl_opt_pass): Convert to subclass of opt_pass.
4445 (rtl_opt_pass::rtl_opt_pass): New.
4446 (ipa_opt_pass_d): Convert to subclass of opt_pass.
4447 (ipa_opt_pass_d::ipa_opt_pass_d): New.
4448 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
4449 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
4450 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
4451 invocation of pass_mode_switching to reflect this moving from a
4452 global to a member of gcc::pass_manager.
4453 (ix86_option_override): Rework how pass_insert_vzeroupper is
4454 added to the pass_manager to reflect autogenerated changes.
4455 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
4457 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
4459 PR rtl-optimization/57708
4460 * recog.c (peep2_find_free_register): Validate all regs in a
4463 2013-08-05 Jan Hubicka <jh@suse.cz>
4466 * cgraph.c (verify_cgraph_node): Accept local flags from other
4468 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
4469 (function_and_variable_visibility): Likewise.
4470 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
4472 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4474 * graph.c (init_graph_slim_pretty_print): Remove.
4475 (print_graph_cfg): Do not call it. Use local pretty printer.
4476 (start_graph_dump): Likewise.
4478 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4480 * gimple-pretty-print.c (buffer): Remove.
4481 (initialized): Likewise.
4482 (maybe_init_pretty_print): Likewise.
4483 (print_gimple_stmt): Do not call it. Use non-static local
4484 pretty_printer variable.
4485 (print_gimple_expr): Likewise.
4486 (print_gimple_seq): Likewise.
4487 (gimple_dump_bb): Likewise.
4489 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4491 * asan.c (asan_pp): Remove.
4492 (asan_pp_initialized): Likewise.
4493 (asan_pp_initialize): Likewise.
4494 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
4495 (asan_emit_stack_protection): Tidy. Use local pretty printer.
4496 (asan_add_global): Likewise.
4498 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
4500 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
4501 * diagnostic.h (diagnostic_flush_buffer): Adjust.
4502 * pretty-print.c (pp_formatted_text_data): Likewise.
4503 (pp_indent): Rename from pp_base_indent.
4504 (pp_format): Rename from pp_base_format.
4505 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
4506 (pp_format_verbatim): Rename from pp_base_format_verbatim.
4507 (pp_flush): Rename from pp_base_flush.
4508 (pp_set_line_maximum_length): Rename from
4509 pp_base_set_line_maximum_length.
4510 (pp_clear_output_area): Rename from pp_base_clear_output_area.
4511 (pp_set_prefix): Rename from pp_base_set_prefix.
4512 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
4513 (pp_emit_prefix): Rename from pp_base_emit_prefix.
4514 (pp_append_text): Rename from pp_base_append_text.
4515 (pp_formatted_text): Rename from pp_base_formatted_text.
4516 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
4517 (pp_remaining_character_count_for_line): Rename from
4518 pp_base_remaining_character_count_for_line.
4519 (pp_newline): Rename from pp_base_newline.
4520 (pp_character): Rename from pp_base_character.
4521 (pp_string): Rename from pp_base_string.
4522 (pp_maybe_space): Rename from pp_base_maybe_space.
4523 * asan.c (asan_pp_string): Adjust.
4524 (asan_emit_stack_protection): Likewise.
4525 (asan_add_global): Likewise.
4526 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
4527 * tree-mudflap.c (mf_varname_tree): Likewise.
4528 * tree-pretty-print.c (pp_tree_identifier): Rename from
4529 pp_base_tree_identifier.
4530 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
4531 Declare as function.
4533 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
4535 * pretty-print.h (pp_bar_bar): New.
4536 (pp_ampersand_ampersand): Likewise.
4537 (pp_less_equal): Likewise.
4538 (pp_greater_equal): Likewise.
4539 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
4540 printer functions instead of pp_string or operators and punctuators.
4541 (dump_gimple_call): Likewise.
4542 (dump_gimple_omp_for): Likewise.
4543 (dump_gimple_transaction): Likewise.
4544 (dump_gimple_phi): Likewise.
4545 (pp_gimple_stmt_1): Likewise.
4546 * sched-vis.c (print_insn): Likewise.
4547 * tree-mudflap.c (mf_varname_tree): Likewise.
4548 * tree-pretty-print.c (dump_block_node): Likewise.
4549 (dump_generic_node): Likewise.
4551 2013-08-02 Jan Hubicka <jh@suse.cz>
4553 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
4555 * lto-streamer-out.c (tree_is_indexable): Results decls and
4556 parm decls are not indexable.
4557 (DFS_write_tree_body): Do not follow args and results.
4558 (hash_tree): Likewise.
4559 (output_functions): Rearrange so struct function is needed
4560 only when real body is output; be able to also ouptut abstract
4561 functions; output DECL_ARGUMENTS and DECL_RESULT.
4562 (lto_output): When not in WPA, ale store abstract functions.
4563 (write_symbol): Do not care about RESULT_DECL.
4564 (output_symbol_p): Handle correctly sbtract decls.
4565 * lto-streamer-in.c (input_function): Rearrange so struct
4566 function can be NULL at entry; allow streaming of
4567 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
4568 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
4569 sanity check during LTO.
4570 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
4571 RESULT_DECl and DECL_ARGUMENTS.
4572 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
4575 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
4577 * pretty-print.h (pp_underscore): New.
4579 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
4580 printer functions instead of pp_character.
4581 (dump_binary_rhs): Likewise.
4582 (dump_ternary_rhs): Likewise.
4583 (dump_gimple_call_args): Likewise.
4584 (pp_points_to_solution): Likewise.
4585 (dump_gimple_call): Likewise.
4586 (dump_gimple_switch): Likewise.
4587 (dump_gimple_cond): Likewise.
4588 (dump_gimple_bind): Likewise.
4589 (dump_gimple_try): Likewise.
4590 (dump_gimple_omp_for): Likewise.
4591 (dump_gimple_omp_continue): Likewise.
4592 (dump_gimple_omp_single): Likewise.
4593 (dump_gimple_omp_sections): Likewise.
4594 (dump_gimple_omp_block): Likewise.
4595 (dump_gimple_omp_critical): Likewise.
4596 (dump_gimple_transaction): Likewise.
4597 (dump_gimple_asm): Likewise.
4598 (dump_gimple_phi): Likewise.
4599 (dump_gimple_omp_parallel): Likewise.
4600 (dump_gimple_omp_task): Likewise.
4601 (dump_gimple_omp_atomic_load): Likewise.
4602 (dump_gimple_omp_atomic_store): Likewise.
4603 (dump_gimple_mem_ops): Likewise.
4604 (pp_gimple_stmt_1): Likewise.
4605 (pp_cfg_jump): Likewise.
4606 (dump_implicit_edges): Likewise.
4607 (gimple_dump_bb_for_graph): Likewise.
4608 * graph.c (draw_cfg_node): Likewise.
4609 * langhooks.c (lhd_print_error_function): Likewise.
4610 * sched-vis.c (print_exp): Likewise.
4611 (print_value): Likewise.
4612 (print_pattern): Likewise.
4613 (print_insn): Likewise.
4614 (rtl_dump_bb_for_graph): Likewise.
4615 * tree-pretty-print.c (dump_function_declaration): Likewise.
4616 (dump_array_domain): Likewise.
4617 (dump_omp_clause): Likewise.
4618 (dump_location): Likewise.
4619 (dump_generic_node): Likewise.
4620 (print_struct_decl): Likewise.
4621 * diagnostic.c (diagnostic_show_locus): Use pp_space.
4623 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4625 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
4626 candidate table when replacing a candidate statement.
4627 (replace_rhs_if_not_dup): Likewise.
4628 (replace_one_candidate): Likewise.
4630 2013-08-02 Jan Hubicka <jh@suse.cz>
4631 Martin Liska <marxin.liska@gmail.com>
4633 * cgraphunit.c (add_new_function): Fix logic when adding from
4635 (assemble_thunk): Rename to ...
4636 (expand_thunk); .. this one; export; get it working with
4637 general functions; make produced gimple valid.
4638 * cgraph.h (expand_thunk): Declare.
4640 2013-08-02 Jan Hubicka <jh@suse.cz>
4642 * ipa-cp.c (gather_context_independent_values): Use
4643 ipa_get_param_move_cost.
4644 (get_replacement_map): Remove PARAM; move parameter folding
4646 (create_specialized_node): Update.
4647 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
4648 assert that we have gimple body; update move_cost.
4649 (count_formal_params): Assert that we have gimple body.
4650 (ipa_dump_param): New function.
4651 (ipa_alloc_node_params): Break out from ...
4652 (ipa_initialize_node_params): ... here.
4653 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
4654 (ipa_write_node_info): Stream move costs.
4655 (ipa_read_node_info): Read move costs.
4656 (ipa_update_after_lto_read): Do not recompute node params.
4657 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
4658 (ipa_get_param): Check we are not in WPA.
4659 (ipa_get_param_move_cost): New.
4660 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
4661 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
4662 parm numbers to be present.
4664 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
4666 PR rtl-optimization/58048
4667 * lra-constraints.c (process_alt_operands): Don't check asm
4668 operand on register.
4670 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
4672 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
4673 the implied StoreLoad barrier for atomic operations if before.
4675 2013-08-02 Jan Hubicka <jh@suse.cz>
4676 Martin Liska <marxin.liska@gmail.com>
4678 * cgraph.c (cgraph_function_body_availability): Do not check
4680 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
4681 symtab_node_availability): Declare.
4682 * ipa.c (can_replace_by_local_alias): New.
4683 (function_and_variable_visibility): Use it.
4684 * symtab.c (symtab_for_node_and_aliases,
4685 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
4687 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
4689 PR rtl-optimization/57963
4690 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
4691 (lra_constraints): Use them.
4693 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
4695 * config/arm/types.md (define_attr "type"): Add "load_acq"
4697 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
4699 (cortex_a53_store1): Likewise.
4701 2013-08-01 Jan Hubicka <jh@suse.cz>
4703 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
4704 partitions are not needed.
4706 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
4708 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
4709 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
4710 MAYBE_NON_Q_CLASS_P where appropriate.
4712 2013-08-01 Jan Hubicka <jh@suse.cz>
4714 * cgraph.h (release_function_body): Declare.
4715 * tree.c (free_lang_data_in_decl): Free, parameters and return values
4716 of unused delcarations.
4718 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4720 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
4721 RTL form when subtracting a constant.
4723 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4725 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
4726 Generate canonical plus rtx with negated immediate instead of minus
4728 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
4730 2013-08-01 Jan Hubicka <jh@suse.cz>
4732 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
4733 (cgraph_release_function_body): Likewise.
4734 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
4735 * cgraph.h (cgrpah_node): Rename abstract_and_needed
4736 to used_as_abstract_origin.
4737 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
4738 symbols used as abstract origins.
4739 * cgraphunit.c (analyze_functions): Update.
4740 * ipa.c (symtab_remove_unreachable_nodes): Recompute
4741 used_as_abstract_origin.
4742 * tree-inline.c (tree_function_versioning): Update
4743 used_as_abstract_origin; be ready for DECL_RESULT and
4744 DECL_ARGUMENTS to be NULL.
4746 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
4747 for abstract functions.
4748 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
4751 2013-08-01 Jan Hubicka <jh@suse.cz>
4753 * profile.c (compute_value_histograms): Fix thinko.
4755 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
4757 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
4758 aarch-common-protos.h to extra_headers.
4759 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
4760 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
4761 * config/aarch64/t-aarch64 (aarch-common.o): Define.
4763 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
4765 * config/aarch64/aarch64.md (define_attr "type"): Delete.
4766 Include "../arm/types.md". Define "type" attribute for all patterns.
4767 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
4770 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
4772 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
4773 to support power8 load fusion.
4774 (fusion_gpr_mem_load): Likewise.
4776 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
4778 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
4779 declarations for power8 load fusion.
4780 (emit_fusion_gpr_load): Likewise.
4782 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
4783 tuning for power8, turn on fusion mode by default. Turn on sign
4784 extending fusion mode if normal fusion mode is on, and we are at
4786 (fusion_gpr_load_p): New function, return true if we can fuse an
4787 addis instruction with a dependent load to a GPR.
4788 (emit_fusion_gpr_load): Emit the instructions for power8 load
4791 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
4792 (VSX load fusion peepholes): New peepholes to fuse together an
4793 addi instruction with a VSX load instruction.
4795 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
4796 peepholes to fuse an addis instruction with a load to a GPR base
4797 register. If we are supporting sign extending fusions, convert
4798 sign extending loads to zero extending loads and add an explicit
4801 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4803 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
4804 aarch-common-protos.h to extra_headers.
4805 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
4806 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
4807 (arm_early_store_addr_dep): Likewise.
4808 (arm_no_early_alu_shift_dep): Likewise.
4809 (arm_no_early_alu_shift_value_dep): Likewise.
4810 (arm_no_early_mul_dep): Likewise.
4811 (arm_no_early_store_addr_dep): Likewise.
4812 (arm_mac_accumulator_is_mul_result): Likewise.
4813 (arm_mac_accumulator_is_result): Likewise.
4814 * config/arm/aarch-common.c: ... here. New file.
4815 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
4817 (arm_early_store_addr_dep): Likewise.
4818 (arm_no_early_alu_shift_dep): Likewise.
4819 (arm_no_early_alu_shift_value_dep): Likewise.
4820 (arm_no_early_mul_dep): Likewise.
4821 (arm_no_early_store_addr_dep): Likewise.
4822 (arm_mac_accumulator_is_mul_result): Likewise.
4823 (arm_mac_accumulator_is_result): Likewise.
4824 * config/arm/aarch-common-protos.h: ... here. New file.
4825 * config/arm/t-arm (aarch-common.o): Define.
4827 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4829 * config/arm/arm.md: Include new file "types.md".
4830 (define_attr "type"): Move from here to ...
4831 (define_attr "mul32"): Likewise.
4832 (define_attr "mul64"): Likewise.
4833 * config/arm/types.md: ... here. New file.
4835 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
4837 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
4838 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
4840 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4842 * gen-pass-instances.awk: Fix offset of substr().
4844 2013-07-31 David Malcolm <dmalcolm@redhat.com>
4846 * Makefile.in (pass-instances.def): New.
4847 (passes.o): Replace dependency on passes.def with one on
4850 * gen-pass-instances.awk: New.
4852 * passes.c (pass_manager::pass_manager): Use pass-instances.def
4853 rather than passes.def, updating local definition of NEXT_PASS
4854 macro to add an extra NUM parameter (currently unused).
4856 2013-07-30 David Malcolm <dmalcolm@redhat.com>
4858 * Makefile.in (PASS_MANAGER_H): New.
4859 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
4860 (passes.o): Likewise.
4861 (statistics.o): Likewise.
4862 (cgraphunit.o): Likewise.
4863 (context.o): Depend on PASS_MANAGER_H.
4865 * pass_manager.h: New.
4867 * cgraphunit.c (cgraph_add_new_function): Update for moves
4868 of globals to fields of pass_manager.
4869 (analyze_function): Likewise.
4870 (expand_function): Likewise.
4871 (ipa_passes): Likewise.
4872 (compile): Likewise.
4874 * context.c (context::context): New.
4875 * context.h (context::context): New.
4876 (context::get_passes): New.
4877 (context::passes_): New.
4879 * lto-cgraph.c (input_node): Update for moves of globals to
4880 fields of pass_manager.
4882 * passes.c (all_passes): Remove, in favor of a field of the
4883 same name within the new class pass_manager.
4884 (all_small_ipa_passes): Likewise.
4885 (all_lowering_passes): Likewise.
4886 (all_regular_ipa_passes): Likewise.
4887 (all_late_ipa_passes): Likewise.
4888 (all_lto_gen_passes): Likewise.
4889 (passes_by_id): Likewise.
4890 (passes_by_id_size): Likewise.
4891 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4892 the new class pass_manager.
4893 (set_pass_for_id): Convert to...
4894 (pass_manager::set_pass_for_id): ...method.
4895 (get_pass_for_id): Convert to...
4896 (pass_manager::get_pass_for_id): ...method.
4897 (register_one_dump_file): Move body of implementation into...
4898 (pass_manager::register_one_dump_file): ...here.
4899 (register_dump_files_1): Convert to...
4900 (pass_manager::register_dump_files_1): ...method.
4901 (register_dump_files): Convert to...
4902 (pass_manager::register_dump_files): ...method.
4903 (create_pass_tab): Update for moves of globals to fields of
4905 (dump_passes): Move body of implementation into...
4906 (pass_manager::dump_passes): ...here.
4907 (register_pass): Move body of implementation into...
4908 (pass_manager::register_pass): ...here.
4909 (init_optimization_passes): Convert into...
4910 (pass_manager::pass_manager): ...constructor for new
4911 pass_manager class, and initialize the pass_lists array.
4912 (check_profile_consistency): Update for moves of globals to
4913 fields of pass_manager.
4914 (dump_profile_report): Move body of implementation into...
4915 (pass_manager::dump_profile_report): ...here.
4916 (ipa_write_summaries_1): Update for moves of pass lists from
4917 being globals to fields of pass_manager.
4918 (ipa_write_optimization_summaries): Likewise.
4919 (ipa_read_summaries): Likewise.
4920 (ipa_read_optimization_summaries): Likewise.
4921 (execute_all_ipa_stmt_fixups): Likewise.
4923 * statistics.c (statistics_fini): Update for moves of globals to
4924 fields of pass_manager.
4926 * toplev.c (general_init): Replace call to
4927 init_optimization_passes with construction of the pass_manager
4930 * tree-pass.h (all_passes): Remove, in favor of a field of the
4931 same name within the new class pass_manager.
4932 (all_small_ipa_passes): Likewise.
4933 (all_lowering_passes): Likewise.
4934 (all_regular_ipa_passes): Likewise.
4935 (all_lto_gen_passes): Likewise.
4936 (all_late_ipa_passes): Likewise.
4937 (passes_by_id): Likewise.
4938 (passes_by_id_size): Likewise.
4939 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4940 the new class pass_manager.
4941 (get_pass_for_id): Remove.
4943 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4945 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
4948 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4950 * arm.md (mulhi3): New expand pattern.
4952 2013-07-30 Jan Hubicka <jh@suse.cz>
4953 Martin Liska <marxin.liska@gmail.com>
4955 * profile.c (compute_value_histograms): Do not ICE when
4956 there is mismatch only on some counters.
4958 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4960 PR rtl-optimization/57637
4961 * function.c (move_insn_for_shrink_wrap): Also check the
4962 GEN set of the LIVE problem for the liveness analysis
4963 if it exists, otherwise give up.
4965 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4967 PR tree-optimization/57993
4968 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
4969 replaced statement in the candidate table.
4970 (phi_add_costs): Return infinite cost when the hidden basis does
4971 not dominate all phis on which the candidate is dependent.
4972 (replace_one_candidate): Record replaced statement in the
4975 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
4977 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
4978 (ashlv2si3): New expander.
4979 (*ashlv2si3_i): New define_insn_and_split.
4980 * predicates.md (float_operation): Allow patterns with three
4983 PR rtl-optimization/58021
4984 * mode-switching.c (create_pre_exit): Always split off preceding
4985 insns if we are not at the basic block head.
4987 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
4989 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
4990 (UCLIBC_DYNAMIC_LINKER): New macro.
4991 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
4993 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
4994 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
4996 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
4997 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
4998 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
4999 for SF and DF modes. Use ieee_quad_format for TF mode.
5000 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
5001 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
5002 (mips_option_override): Handle `-mnan=legacy'.
5003 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
5004 `-mabs=2008' and `-mnan=2008'.
5005 (OPTION_DEFAULT_SPECS): Add "nan" default.
5006 (ASM_SPEC): Handle `-mnan='.
5007 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
5008 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
5009 comment accordingly.
5010 (neg<mode>2): Likewise.
5011 * config/mips/mips.opt (mabs, mnan): New options.
5012 * doc/install.texi (Configuration): Document `--with-nan=' option.
5013 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
5015 (MIPS Options): Document them.
5016 * config.gcc <mips*-*-*>: Handle `--with-nan='.
5017 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
5018 * configure: Regenerate.
5019 * config.in: Regenerate.
5021 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
5023 * config/i386/i386.md (float post-reload splitters): Do not check
5024 for subregs of SSE registers.
5026 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
5027 H.J. Lu <hongjiu.lu@intel.com>
5031 * config/i386/i386.md (post-reload splitter
5032 to avoid partial SSE reg dependency stalls): New pattern.
5034 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
5036 * config/s390/s390.md ("movcc"): Swap load and store instructions.
5038 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
5040 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
5041 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
5043 2013-07-26 Cary Coutant <ccoutant@google.com>
5045 * dwarf2out.c (die_checksum_ordered): Don't include template
5046 instantiations in signature.
5047 (is_template_parameter): New function.
5048 (is_template_instantiation): New function.
5049 (generate_skeleton_bottom_up): Don't include template instantiations
5051 (generate_skeleton): Likewise.
5052 (break_out_comdat_types): Move recursive call to break out nested
5054 (prune_unused_types_mark_generic_parms_dies): Call
5055 is_template_parameter.
5057 2013-07-26 Ian Bolton <ian.bolton@arm.com>
5059 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
5060 uses vector registers.
5061 * config/aarch64/iterators.md: Add attributes rtn and vas.
5063 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5064 Richard Earnshaw <richard.earnshaw@arm.com>
5066 * combine.c (simplify_comparison): Re-canonicalize operands
5068 * config/arm/arm.md (movcond_addsi): New splitter.
5070 2013-07-25 Sterling Augustine <saugustine@google.com>
5072 * dwarf2out.c (size_of_pubnames): Move code to...
5073 (include_pubname_in_output): ...here. New.
5074 (want_pubnames): Rearrange.
5075 (output_pubnames): Call include_pubname_in_output. Move assertion.
5077 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5079 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
5081 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5084 * doc/extend.texi: Remove obsolete builtins. Fix
5085 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
5087 2013-07-25 Jan Hubicka <jh@suse.cz>
5089 * cgraph.c (release_function_body): Break out from ...
5090 (cgraph_release_function_body): ... this one; also release DECL_RESULT
5092 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
5093 old_tree in the map.
5094 (create_specialized_node): Update.
5095 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
5097 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
5098 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
5099 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
5100 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
5103 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5105 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
5106 addsi3_carryin_alt2_<optab>): Correct output template.
5108 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5110 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
5111 Adjust for arm_restrict_it.
5112 Remove trailing whitespace.
5114 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
5116 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
5117 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
5119 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
5121 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
5123 PR rtl-optimization/57960
5124 * lra-constraints.c (process_alt_operands): Use the right mode
5125 when checking strict_low.
5127 2013-07-25 Jan Hubicka <jh@suse.cz>
5129 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
5130 * cgraph.c (cgraph_remove_node): Do not release function body
5131 when in cgraph streaming.
5132 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
5133 in other partitions are not considered reachable; fix handling of
5136 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5138 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
5140 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5145 * config/arm/arm.md ("*sibcall_insn): Replace use of
5146 Ss with US. Adjust output for v5 and v4t.
5147 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
5149 * config/arm/constraints.md ("Ss"): Rename to US.
5151 2013-07-25 Terry Guo <terry.guo@arm.com>
5153 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
5154 shift_add/shift_sub0/shift_sub1 RTXs.
5156 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
5157 Anton Blanchard <anton@au1.ibm.com>
5159 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
5160 (altivec_vpks<VI_char>ss): Likewise.
5161 (altivec_vpks<VI_char>us): Likewise.
5162 (altivec_vpku<VI_char>us): Likewise.
5163 (altivec_vpku<VI_char>um): Likewise.
5165 2013-07-24 David Malcolm <dmalcolm@redhat.com>
5167 Introduce context class.
5169 * Makefile.in (CONTEXT_H): New.
5170 (OBJS): Add context.o.
5171 (toplev.o): Add CONTEXT_H to dependencies.
5174 * toplev.c (general_init): Create the singleton gcc::context instance.
5180 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
5182 PR rtl-optimization/57968
5183 * mode-switching.c (create_pre_exit): Allow instructions that
5184 don't set a return register to need a non-exit mode.
5186 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5187 Anton Blanchard <anton@au1.ibm.com>
5189 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
5190 operands to vperm for little endian.
5191 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
5192 of lvsl to create the control mask for a vperm for little endian.
5194 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5195 Anton Blanchard <anton@au1.ibm.com>
5197 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
5198 two operands for little-endian.
5200 2013-07-23 Steve Ellcey <sellcey@mips.com>
5202 * config/mips/mips.c (mips_case_values_threshold): New.
5203 (TARGET_CASE_VALUES_THRESHOLD): Define.
5205 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5206 Anton Blanchard <anton@au1.ibm.com>
5208 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
5209 selection of field for vector splat in little endian mode.
5211 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
5213 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
5214 expanders to rs6000.md.
5215 (ior<mode>3): Likewise.
5216 (and<mode>3): Likewise.
5217 (one_cmpl<mode>2): Likewise.
5218 (nor<mode>3): Likewise.
5219 (andc<mode>3): Likewise.
5220 (eqv<mode>3): Likewise.
5221 (nand<mode>3): Likewise.
5222 (orc<mode>3): Likewise.
5224 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
5227 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
5228 to split multi-word logical operations.
5229 (rs6000_split_logical_di): Likewise.
5230 (rs6000_split_logical): Likewise.
5232 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
5233 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
5234 and allow TImode operations in 32-bit.
5235 (vsx_and<mode>3_64bit): Likewise.
5236 (vsx_ior<mode>3_32bit): Likewise.
5237 (vsx_ior<mode>3_64bit): Likewise.
5238 (vsx_xor<mode>3_32bit): Likewise.
5239 (vsx_xor<mode>3_64bit): Likewise.
5240 (vsx_one_cmpl<mode>2_32bit): Likewise.
5241 (vsx_one_cmpl<mode>2_64bit): Likewise.
5242 (vsx_nor<mode>3_32bit): Likewise.
5243 (vsx_nor<mode>3_64bit): Likewise.
5244 (vsx_andc<mode>3_32bit): Likewise.
5245 (vsx_andc<mode>3_64bit): Likewise.
5246 (vsx_eqv<mode>3_32bit): Likewise.
5247 (vsx_eqv<mode>3_64bit): Likewise.
5248 (vsx_nand<mode>3_32bit): Likewise.
5249 (vsx_nand<mode>3_64bit): Likewise.
5250 (vsx_orc<mode>3_32bit): Likewise.
5251 (vsx_orc<mode>3_64bit): Likewise.
5253 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
5254 logical types in GPRs.
5256 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
5257 logical insns to rs6000.md, and allow TImode operations in
5259 (altivec_ior<mode>3): Likewise.
5260 (altivec_xor<mode>3): Likewise.
5261 (altivec_one_cmpl<mode>2): Likewise.
5262 (altivec_nor<mode>3): Likewise.
5263 (altivec_andc<mode>3): Likewise.
5265 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
5266 attributes for moving the 128-bit logical operations into
5268 (BOOL_REGS_OUTPUT): Likewise.
5269 (BOOL_REGS_OP1): Likewise.
5270 (BOOL_REGS_OP2): Likewise.
5271 (BOOL_REGS_UNARY): Likewise.
5272 (BOOL_REGS_AND_CR0): Likewise.
5273 (one_cmpl<mode>2): Add support for DI logical operations on
5274 32-bit, splitting the operations to 32-bit.
5278 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
5279 changes to combine the 32/64-bit code, allow logical operations on
5280 TI mode in 32-bit, and to use similar match_operator patterns like
5281 scalar mode uses. Combine the Altivec and VSX code for logical
5282 operations, and move it here.
5283 (ior<mode>3, 128-bit types): Likewise.
5284 (xor<mode>3, 128-bit types): Likewise.
5285 (one_cmpl<mode>3, 128-bit types): Likewise.
5286 (nor<mode>3, 128-bit types): Likewise.
5287 (andc<mode>3, 128-bit types): Likewise.
5288 (eqv<mode>3, 128-bit types): Likewise.
5289 (nand<mode>3, 128-bit types): Likewise.
5290 (orc<mode>3, 128-bit types): Likewise.
5291 (and<mode>3_internal): Likewise.
5292 (bool<mode>3_internal): Likewise.
5293 (boolc<mode>3_internal1): Likewise.
5294 (boolc<mode>3_internal2): Likewise.
5295 (boolcc<mode>3_internal1): Likewise.
5296 (boolcc<mode>3_internal2): Likewise.
5297 (eqv<mode>3_internal1): Likewise.
5298 (eqv<mode>3_internal2): Likewise.
5299 (one_cmpl1<mode>3_internal): Likewise.
5301 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
5303 * config/microblaze/microblaze.c (microblaze_expand_prologue):
5304 Rename flag_stack_usage to flag_stack_usage_info.
5306 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
5308 * config/microblaze/sync.md: New file.
5309 * config/microblaze/microblaze.md: Include sync.md
5310 * config/microblaze/microblaze.c: Add print_operand 'y'.
5311 * config/microblaze/constraints.md: Add memory_contraint
5312 'Q' which is a single register.
5314 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
5316 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
5318 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
5320 * reload.c (find_reloads): Exit loop once we find this operand
5321 cannot be reloaded somehow for this alternative.
5323 * reload.c (find_reloads): Exit loop once we find a hard register.
5325 * rtlanal.c (computed_jump_p): Exit loop once we find label
5328 * i386.c (ix86_pad_returns): Exit loop after setting replace.
5330 * cfgloopmanip.c (remove_path): Exit loop after setting
5333 * gensupport.c (subst_dup): Avoid loop if code is not
5334 MATCH_DUP nor MATCH_OP_DUP.
5336 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
5338 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
5340 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5342 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
5343 true for SP_REGNUM if mode == ptr_mode.
5344 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
5345 with value R0_REGNUM + 31.
5347 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5349 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
5350 pad pointer-typed argument downward.
5352 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5354 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
5355 and __ILP32__ when the ILP32 model is in use.
5357 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5359 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
5360 (aarch64_load_symref_appropriately): In the case of
5361 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
5362 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
5363 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
5364 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
5365 if 'mode' doesn't equal to 'ptr_mode'.
5366 (aarch64_output_mi_thunk): Add an assertion on the alignment of
5367 'vcall_offset'; change to call aarch64_emit_move differently depending
5368 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
5369 to calculate the upper bound of 'vcall_offset'.
5370 (aarch64_cannot_force_const_mem): Change to also return true if
5372 (aarch64_legitimize_reload_address): In the case of large
5373 displacements, add new local variable 'xmode' and an assertion
5374 based on it; change to use 'xmode' to generate the new rtx and
5376 (aarch64_asm_trampoline_template): Change to generate the template
5377 differently depending on TARGET_ILP32 or not; change to use
5378 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
5379 (aarch64_trampoline_size): Removed.
5380 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
5381 and replace immediate literals with it. Change to use 'ptr_mode'
5382 instead of 'DImode' and call convert_memory_address if the mode
5383 of 'fnaddr' doesn't equal to 'ptr_mode'.
5384 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
5386 (aarch64_elf_asm_destructor): Likewise.
5387 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
5388 on TARGET_ILP32 instead of aarch64_trampoline_size.
5389 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
5390 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
5391 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
5392 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
5393 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
5394 (storewb_pair<GPI:mode>_<P:mode>): ... this.
5395 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
5396 depending on the value of 'mode'.
5397 (add_losym_<mode>): New.
5398 (ldr_got_small_<mode>): New, based on ldr_got_small.
5399 (ldr_got_small): Remove.
5400 (ldr_got_small_sidi): New.
5401 * config/aarch64/iterators.md (P): New.
5402 (PTR): Change to 'ptr_mode' in the condition.
5404 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5406 * config.gcc (aarch64*-*-*): Support --with-abi.
5407 (aarch64*-*-elf): Support --with-multilib-list.
5408 (aarch64*-*-linux*): Likewise.
5409 (supported_defaults): Add abi to aarch64*-*-*.
5410 * configure.ac: Mention AArch64 for --with-multilib-list.
5411 * configure: Re-generated.
5412 * config/aarch64/biarchilp32.h: New file.
5413 * config/aarch64/biarchlp64.h: New file.
5414 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
5416 (MULTILIB_DEFAULTS): Ditto.
5417 (DRIVER_SELF_SPECS): Ditto.
5418 (ASM_SPEC): Update to also substitute -mabi.
5419 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
5420 file whose name depends on -mabi= and -mbig-endian.
5421 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
5423 (POINTER_SIZE): New define.
5424 (POINTERS_EXTEND_UNSIGNED): Ditto.
5425 (enum aarch64_abi_type): New enumeration tag.
5426 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
5427 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
5428 (TARGET_ILP32): New define.
5429 * config/aarch64/aarch64.opt (mabi): New.
5431 (ilp32, lp64): New values for -mabi.
5432 * config/aarch64/t-aarch64 (comma): New define.
5433 (MULTILIB_OPTIONS): Ditto.
5434 (MULTILIB_DIRNAMES): Ditto.
5435 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
5436 * doc/invoke.texi: Document -mabi for AArch64.
5438 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
5440 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
5442 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5443 Anton Blanchard <anton@au1.ibm.com>
5445 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
5446 endianness when selecting field to splat.
5448 2013-07-22 Eric Christopher <echristo@gmail.com>
5450 * dwarf2out.c (die_odr_checksum): New function to use
5451 CHECKSUM_ macros and ULEB128 for DIE tag.
5452 (generate_type_signature): Use.
5454 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
5456 * config.gcc (sparc*-*-*): Accept leon3 processor.
5457 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
5458 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
5459 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
5460 * config/sparc/sparc.opt (enum processor_type): Add leon3.
5461 (mfix-ut699): Adjust comment.
5462 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
5463 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
5464 (CPP_CPU_SPEC): Likewise.
5465 (ASM_CPU_SPEC): Likewise.
5466 * config/sparc/sparc.c (leon3_cost): New constant.
5467 (sparc_option_override): Add leon3 support.
5468 (mem_ref): New function.
5469 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
5470 (sparc_do_work_around_errata): Look into the instruction in the delay
5471 slot and adjust accordingly. Add fix for the data cache nullify issues
5472 of the UT699. Change insertion position for the NOP.
5473 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
5474 (leon3_load): New reservation.
5475 (leon_store): Bump latency to 2.
5476 (grfpu): New automaton.
5477 (grfpu_alu): New unit.
5478 (grfpu_ds): Likewise.
5479 (leon_fp_alu): Adjust.
5480 (leon_fp_mult): Delete.
5481 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
5482 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
5483 * config/sparc/sparc.md (cpu): Add leon3.
5484 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
5486 (atomic_test_and_set): Likewise.
5489 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
5491 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
5492 default for R5900 targets.
5493 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
5494 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
5495 * config/mips/mips.c (mips_option_override): Report an error for
5496 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
5497 for -march=r5900 -mhard-float.
5499 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
5501 * df-problems.c (can_move_insns_across): Exit loop once we
5502 find a non-fixed, non-global register.
5504 * ipa-pure-const.c (propagate_nothrow): Exit loop after
5507 * omega.c (omega_eliminate_red): Break after setting red_found.
5508 (omega_problem_has_red_equations): Similarly after setting found.
5509 (omega_query_variable): Similarly after setting coupled.
5511 2013-07-22 Marek Polacek <polacek@redhat.com>
5513 * gimplify.c: Don't include gimple.h twice.
5515 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5517 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
5518 instead of TARGET_THUMB1.
5519 (Pz): New constraint.
5520 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
5522 (compare_negsi_si): Likewise.
5523 (compare_addsi2_op0): Likewise.
5524 (compare_addsi2_op1): Likewise.
5525 (addsi3_carryin_<optab>): Likewise.
5526 (addsi3_carryin_alt2_<optab>): Likewise.
5527 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
5528 for arm_restrict_it.
5529 (subsi3_carryin): Likewise.
5530 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
5531 (minmax_arithsi): Disable for arm_restrict_it.
5532 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
5533 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
5534 (satsi_<SAT:code>_shift): Likewise.
5535 (arm_shiftsi3): Add alternative for 16-bit encoding.
5536 (arm32_movhf): Disable for arm_restrict_it.
5537 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
5538 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
5540 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
5542 * config/arm/arm.md (attribute "insn"): Delete.
5543 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
5544 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
5545 (not_shiftsi): Update for attribute change.
5546 (not_shiftsi_compare0): Likewise.
5547 (not_shiftsi_compare0_scratch): Likewise.
5548 (arm_one_cmplsi2): Likewise.
5549 (thumb1_one_cmplsi2): Likewise.
5550 (notsi_compare0): Likewise.
5551 (notsi_compare0_scratch): Likewise.
5552 (thumb1_movdi_insn): Likewise.
5553 (arm_movsi_insn): Likewise.
5554 (movhi_insn_arch4): Likewise.
5555 (movhi_bytes): Likewise.
5556 (arm_movqi_insn): Likewise.
5557 (thumb1_movqi_insn): Likewise.
5558 (arm32_movhf): Likewise.
5559 (thumb1_movhf): Likewise.
5560 (arm_movsf_soft_insn): Likewise.
5561 (thumb1_movsf_insn): Likewise.
5562 (thumb_movdf_insn): Likewise.
5563 (movsicc_insn): Likewise.
5564 (movsfcc_soft_insn): Likewise.
5565 (and_scc): Likewise.
5566 (cond_move): Likewise.
5567 (if_move_not): Likewise.
5568 (if_not_move): Likewise.
5569 (if_shift_move): Likewise.
5570 (if_move_shift): Likewise.
5571 (if_shift_shift): Likewise.
5572 (if_not_arith): Likewise.
5573 (if_arith_not): Likewise.
5574 (cond_move_not): Likewise.
5575 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
5576 (neon_mov<mode>): Likewise.
5577 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
5578 (thumb2_movsi_vfp): Likewise.
5579 (movsf_vfp): Likewise.
5580 (thumb2_movsf_vfp): Likewise.
5581 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5583 (cortexa7_older_only): Likewise.
5584 (cortexa7_younger): Likewise.
5585 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
5586 (1020alu_shift_op): Likewise.
5587 (1020alu_shift_reg_op): Likewise.
5588 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
5589 (alu_shift_op): Likewise.
5590 (alu_shift_reg_op): Likewise.
5591 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
5592 (11_alu_shift_op): Likewise.
5593 (11_alu_shift_reg_op): Likewise.
5594 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
5595 (9_alu_shift_reg_op): Likewise.
5596 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5598 (cortex_a15_alu_shift): Likewise.
5599 (cortex_a15_alu_shift_reg): Likewise.
5600 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
5601 (cortex_a5_alu_shift): Likewise.
5602 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
5604 (cortex_a53_alu_shift): Likewise.
5605 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5607 (cortex_a7_alu_reg): Likewise.
5608 (cortex_a7_alu_shift): Likewise.
5609 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5610 (cortex_a8_alu_shift): Likewise.
5611 (cortex_a8_alu_shift_reg): Likewise.
5612 (cortex_a8_mov): Likewise.
5613 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
5614 (cortex_a9_dp_shift): Likewise.
5615 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
5616 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
5617 (cortex_r4_mov): Likewise.
5618 (cortex_r4_alu_shift): Likewise.
5619 (cortex_r4_alu_shift_reg): Likewise.
5620 * config/arm/fa526.md (526_alu_op): Update for attribute change.
5621 (526_alu_shift_op): Likewise.
5622 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
5623 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
5624 (626te_alu_shift_op): Likewise.
5625 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
5626 (726te_alu_op): Likewise.
5627 (726te_alu_shift_op): Likewise.
5628 (726te_alu_shift_reg_op): Likewise.
5629 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
5630 (mp626_alu_shift_op): Likewise.
5631 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
5632 (pj4_alu_e1_conds): Likewise.
5633 (pj4_alu): Likewise.
5634 (pj4_alu_conds): Likewise.
5635 (pj4_shift): Likewise.
5636 (pj4_shift_conds): Likewise.
5637 (pj4_alu_shift): Likewise.
5638 (pj4_alu_shift_conds): Likewise.
5640 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5642 * config/arm/predicates.md (shiftable_operator_strict_it):
5644 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
5645 Disable cond_exec version for arm_restrict_it.
5646 (thumb2_smaxsi3): Convert to generate cond_exec.
5647 (thumb2_sminsi3): Likewise.
5648 (thumb32_umaxsi3): Likewise.
5649 (thumb2_uminsi3): Likewise.
5650 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
5651 (thumb2_neg_abssi2): Likewise.
5652 (thumb2_mov_scc): Add alternative for 16-bit encoding.
5653 (thumb2_movsicc_insn): Adjust alternatives.
5654 (thumb2_mov_negscc): Disable for arm_restrict_it.
5655 (thumb2_mov_negscc_strict_it): New pattern.
5656 (thumb2_mov_notscc_strict_it): New pattern.
5657 (thumb2_mov_notscc): Disable for arm_restrict_it.
5658 (thumb2_ior_scc): Likewise.
5659 (thumb2_ior_scc_strict_it): New pattern.
5660 (thumb2_cond_move): Adjust for arm_restrict_it.
5661 (thumb2_cond_arith): Disable for arm_restrict_it.
5662 (thumb2_cond_arith_strict_it): New pattern.
5663 (thumb2_cond_sub): Adjust for arm_restrict_it.
5664 (thumb2_movcond): Likewise.
5665 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
5666 (thumb2_zero_extendhisi2_v6): Likewise.
5667 (thumb2_zero_extendqisi2_v6): Likewise.
5668 (orsi_notsi_si): Likewise.
5669 (orsi_not_shiftsi_si): Likewise.
5671 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
5673 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
5674 instruction sequence is 1 byte shorter.
5676 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
5678 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
5679 it is not needed after split.
5681 2013-07-20 Iain Sandoe <iain@codesourcery.com>
5684 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
5685 second label for nonlocal goto receivers. Don't output pic base labels
5686 unless we're producing PIC; mark that action unreachable().
5687 (ix86_save_reg): If the function contains a nonlocal label, save the
5689 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
5690 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
5691 (update_pic_label_number_if_needed): New.
5692 (machopic_output_function_base_name): Adjust for nonlocal receiver
5694 (machopic_should_output_picbase_label): New.
5695 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
5696 (nonlocal_goto_receiver): New insn and split.
5698 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
5700 * config/aarch64/aarch64-builtins.c
5701 (aarch64_fold_builtin): Fold abs in all modes.
5702 * config/aarch64/aarch64-simd-builtins.def
5703 (abs): Enable for all modes.
5704 * config/aarch64/arm_neon.h
5705 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
5706 (vabs_f64): Add missing intrinsic.
5708 2013-07-19 Ian Bolton <ian.bolton@arm.com>
5710 * config/aarch64/arm_neon.h (vabs_s64): New function
5712 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
5715 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
5716 * config/avr/avr.md (adjust_len): Add `round'.
5717 * config/avr/avr-protos.h (avr_out_round): New prototype.
5718 (avr_out_plus): Add `out_label' argument.
5719 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
5720 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
5721 Handle the case where `insn' is just a pattern.
5722 (avr_out_bitop): Handle the case where `insn' is just a pattern.
5723 (avr_out_round): New function.
5724 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
5726 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
5728 * config/microblaze/microblaze.c (microblaze_expand_prologue):
5729 Add check for flag_stack_usage to handle -fstack-usage support
5731 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
5733 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
5734 interaction for new Power8 flags and VSX.
5736 2013-07-18 Sriraman Tallam <tmsriram@google.com>
5739 * tree-inline.c (expand_call_inline): Emit errors during
5740 early_inlining only if optimization is not turned on.
5742 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5746 * passes.c (init_optimization_passes): Move the construction of
5747 the pass hierarchy into a new passes.def file.
5749 * Makefile.in (passes.o): Add dependency on passes.def.
5751 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5753 * passes.c (init_optimization_passes): Introduce macros for
5754 constructing the tree of passes (INSERT_PASSES_AFTER,
5755 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
5756 TERMINATE_PASS_LIST).
5758 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
5759 Wei Mi <wmi@google.com>
5761 PR rtl-optimization/57878
5762 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
5764 (reload_pseudo_compare_func): Check nregs first for reload
5767 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5769 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
5771 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
5773 * read-rtl.c (validate_const_int): Once an invalid character is
5774 seen, quit the loop.
5776 * gengtype.c (write_roots): Similarly once we find the "deletable"
5777 or "if_marked" option.
5779 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
5781 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
5782 "xtab" and "sat". Move value "clz" from here to ...
5783 (attriubte "type"): ... here.
5784 (satsi_<SAT:code>): Delete "insn" attribute.
5785 (satsi_<SAT:code>_shift): Likewise.
5786 (arm_zero_extendqisi2addsi): Likewise.
5787 (arm_extendqisi2addsi): Likewise.
5788 (clzsi2): Update for attribute changes.
5789 (rbitsi2): Likewise.
5790 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
5792 (arm_usatsihi): Likewise.
5793 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5795 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
5797 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
5798 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
5799 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
5800 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
5801 in alphabetical order.
5802 (attribute "core_cycles"): Update for attribute changes.
5803 (arm_addsi3): Likewise.
5804 (addsi3_compare0): Likewise.
5805 (addsi3_compare0_scratch): Likewise.
5806 (addsi3_compare_op1): Likewise.
5807 (addsi3_compare_op2): Likewise.
5808 (compare_addsi2_op0): Likewise.
5809 (compare_addsi2_op1): Likewise.
5810 (addsi3_carryin_shift_<optab>): Likewise.
5811 (subsi3_carryin_shift): Likewise.
5812 (rsbsi3_carryin_shift): Likewise.
5813 (arm_subsi3_insn): Likewise.
5814 (subsi3_compare0): Likewise.
5815 (subsi3_compare): Likewise.
5816 (arm_andsi3_insn): Likewise.
5817 (thumb1_andsi3_insn): Likewise.
5818 (andsi3_compare0): Likewise.
5819 (andsi3_compare0_scratch): Likewise.
5820 (zeroextractsi_compare0_scratch
5821 (andsi_not_shiftsi_si): Likewise.
5822 (iorsi3_insn): Likewise.
5823 (iorsi3_compare0): Likewise.
5824 (iorsi3_compare0_scratch): Likewise.
5825 (arm_xorsi3): Likewise.
5826 (thumb1_xorsi3_insn): Likewise.
5827 (xorsi3_compare0): Likewise.
5828 (xorsi3_compare0_scratch): Likewise.
5829 (satsi_<SAT:code>_shift): Likewise.
5831 (arm_shiftsi3): Likewise.
5832 (shiftsi3_compare0): Likewise.
5833 (not_shiftsi): Likewise.
5834 (not_shiftsi_compare0): Likewise.
5835 (not_shiftsi_compare0_scratch): Likewise.
5836 (arm_one_cmplsi2): Likewise.
5837 (thumb_one_complsi2): Likewise.
5838 (notsi_compare0): Likewise.
5839 (notsi_compare0_scratch): Likewise.
5840 (thumb1_zero_extendhisi2): Likewise.
5841 (arm_zero_extendhisi2): Likewise.
5842 (arm_zero_extendhisi2_v6): Likewise.
5843 (arm_zero_extendhisi2addsi): Likewise.
5844 (thumb1_zero_extendqisi2): Likewise.
5845 (thumb1_zero_extendqisi2_v6): Likewise.
5846 (arm_zero_extendqisi2): Likewise.
5847 (arm_zero_extendqisi2_v6): Likewise.
5848 (arm_zero_extendqisi2addsi): Likewise.
5849 (thumb1_extendhisi2): Likewise.
5850 (arm_extendhisi2): Likewise.
5851 (arm_extendhisi2_v6): Likewise.
5852 (arm_extendqisi): Likewise.
5853 (arm_extendqisi_v6): Likewise.
5854 (arm_extendqisi2addsi): Likewise.
5855 (thumb1_extendqisi2): Likewise.
5856 (thumb1_movdi_insn): Likewise.
5857 (arm_movsi_insn): Likewise.
5858 (movsi_compare0): Likewise.
5859 (movhi_insn_arch4): Likewise.
5860 (movhi_bytes): Likewise.
5861 (arm_movqi_insn): Likewise.
5862 (thumb1_movqi_insn): Likewise.
5863 (arm32_movhf): Likewise.
5864 (thumb1_movhf): Likewise.
5865 (arm_movsf_soft_insn): Likewise.
5866 (thumb1_movsf_insn): Likewise.
5867 (movdf_soft_insn): Likewise.
5868 (thumb_movdf_insn): Likewise.
5869 (arm_cmpsi_insn): Likewise.
5870 (cmpsi_shiftsi): Likewise.
5871 (cmpsi_shiftsi_swp): Likewise.
5872 (arm_cmpsi_negshiftsi_si): Likewise.
5873 (movsicc_insn): Likewise.
5874 (movsfcc_soft_insn): Likewise.
5875 (arith_shiftsi): Likewise.
5876 (arith_shiftsi_compare0
5877 (arith_shiftsi_compare0_scratch
5878 (sub_shiftsi): Likewise.
5879 (sub_shiftsi_compare0
5880 (sub_shiftsi_compare0_scratch
5881 (and_scc): Likewise.
5882 (cond_move): Likewise.
5883 (if_plus_move): Likewise.
5884 (if_move_plus): Likewise.
5885 (if_move_not): Likewise.
5886 (if_not_move): Likewise.
5887 (if_shift_move): Likewise.
5888 (if_move_shift): Likewise.
5889 (if_shift_shift): Likewise.
5890 (if_not_arith): Likewise.
5891 (if_arith_not): Likewise.
5892 (cond_move_not): Likewise.
5893 (thumb1_ashlsi3): Set type attribute.
5894 (thumb1_ashrsi3): Likewise.
5895 (thumb1_lshrsi3): Likewise.
5896 (thumb1_rotrsi3): Likewise.
5897 (shiftsi3_compare0_scratch): Likewise.
5898 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
5899 (neon_mov<mode>): Likewise.
5900 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
5902 (thumb2_movsi_insn): Likewise.
5903 (thumb2_cmpsi_neg_shiftsi): Likewise.
5904 (thumb2_extendqisi_v6): Likewise.
5905 (thumb2_zero_extendhisi2_v6): Likewise.
5906 (thumb2_zero_extendqisi2_v6): Likewise.
5907 (thumb2_shiftsi3_short): Likewise.
5908 (thumb2_addsi3_compare0_scratch): Likewise.
5909 (orsi_not_shiftsi_si): Likewise.
5910 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5911 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
5913 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5914 (1020alu_shift_op): Likewise.
5915 (1020alu_shift_reg_op): Likewise.
5916 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5917 (alu_shift_op): Likewise.
5918 (alu_shift_reg_op): Likewise.
5919 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5920 (11_alu_shift_op): Likewise.
5921 (11_alu_shift_reg_op): Likewise.
5922 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5923 (9_alu_shift_reg_op): Likewise.
5924 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5926 (cortex_a15_alu_shift): Likewise.
5927 (cortex_a15_alu_shift_reg): Likewise.
5928 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
5930 (cortex_a5_alu_shift): Likewise.
5931 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
5933 (cortex_a53_alu_shift): Likewise.
5934 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5936 (cortex_a7_alu_reg): Likewise.
5937 (cortex_a7_alu_shift): Likewise.
5938 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
5940 (cortex_a8_alu_shift): Likewise.
5941 (cortex_a8_alu_shift_reg): Likewise.
5942 (cortex_a8_mov): Likewise.
5943 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5944 (cortex_a9_dp_shift): Likewise.
5945 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
5947 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
5949 (cortex_r4_mov): Likewise.
5950 (cortex_r4_alu_shift): Likewise.
5951 (cortex_r4_alu_shift_reg): Likewise.
5952 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5953 (526_alu_shift_op): Likewise.
5954 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5955 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5956 (626te_alu_shift_op): Likewise.
5957 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
5958 (726te_alu_op): Likewise.
5959 (726te_alu_shift_op): Likewise.
5960 (726te_alu_shift_reg_op): Likewise.
5961 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5962 (mp626_alu_shift_op): Likewise.
5963 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
5964 (pj4_alu_e1_conds): Likewise.
5965 (pj4_alu): Likewise.
5966 (pj4_alu_conds): Likewise.
5967 (pj4_shift): Likewise.
5968 (pj4_shift_conds): Likewise.
5969 (pj4_alu_shift): Likewise.
5970 (pj4_alu_shift_conds): Likewise.
5971 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5973 (cortexa7_older_only): Likewise.
5974 (cortexa7_younger): Likewise.
5976 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5978 * ipa-pure-const.c (generate_summary): Rename to...
5979 (pure_const_generate_summary): ... this.
5981 2013-07-17 Iain Sandoe <iain@codesourcery.com>
5983 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
5985 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
5988 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
5991 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5993 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
5994 enabled without -march=zEC12.
5995 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
5998 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6000 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
6001 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
6002 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
6003 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
6004 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
6005 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
6006 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
6008 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
6009 and ISA_HAS_NMADD3_NMSUB3.
6010 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
6011 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
6012 (nmsub4<mode>, nmsub3<mode>): Likewise.
6013 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
6015 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6017 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
6018 TARGET_MIPS5400 checking.
6020 2013-07-16 Jakub Jelinek <jakub@redhat.com>
6021 Peter Bergner <bergner@vnet.ibm.com>
6023 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
6024 registers in the comment.
6025 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
6026 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
6027 rather than FIRST_PSEUDO_REGISTERS.
6029 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
6031 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
6032 enable extra ISA flags with TARGET_HTM.
6034 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6036 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
6039 2013-07-15 Cong Hou <congh@google.com>
6041 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
6042 in compare function for sorting.
6044 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
6046 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
6047 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
6048 * config/rs6000/rs6000.opt: Add -mhtm option.
6049 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
6050 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
6051 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6052 __HTM__ if the HTM instructions are available.
6053 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
6054 htm_spr_reg_operand): New define_predicates.
6055 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
6056 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
6058 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
6059 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
6060 HTM builtin functions.
6061 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
6062 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
6063 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
6064 (rs6000_builtin_mask_calculate): Likewise.
6065 (rs6000_option_override_internal): Likewise.
6066 (bdesc_htm): Add new HTM builtin support.
6067 (htm_spr_num): New function.
6068 (htm_spr_regno): Likewise.
6069 (rs6000_htm_spr_icode): Likewise.
6070 (htm_expand_builtin): Likewise.
6071 (htm_init_builtins): Likewise.
6072 (rs6000_expand_builtin): Add support for HTM builtin functions.
6073 (rs6000_init_builtins): Likewise.
6074 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
6076 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
6077 (TARGET_HTM, MASK_HTM): Define macros.
6078 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
6079 (FIXED_REGISTERS): Likewise.
6080 (CALL_USED_REGISTERS): Likewise.
6081 (CALL_REALLY_USED_REGISTERS): Likewise.
6082 (REG_ALLOC_ORDER): Likewise.
6083 (enum reg_class): Likewise.
6084 (REG_CLASS_NAMES): Likewise.
6085 (REG_CLASS_CONTENTS): Likewise.
6086 (REGISTER_NAMES): Likewise.
6087 (ADDITIONAL_REGISTER_NAMES): Likewise.
6088 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
6089 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
6090 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
6091 * config/rs6000/htm.md: New file.
6092 * config/rs6000/htmintrin.h: New file.
6093 * config/rs6000/htmxlintrin.h: New file.
6095 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
6097 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
6098 Define SYMBOL_TINY_GOT, update comment.
6099 * config/aarch64/aarch64.c
6100 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
6101 (aarch64_expand_mov_immediate): Likewise.
6102 (aarch64_print_operand): Likewise.
6103 (aarch64_classify_symbol): Likewise.
6104 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
6105 (ldr_got_tiny): Define.
6107 2013-07-13 Tobias Grosser <tobias@grosser.es>
6109 PR tree-optimization/54094
6110 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
6111 scheduling dimension for the parallelism check from the polyhedral
6112 information in the AST.
6113 * graphite-dependences.c (carries_deps): Do not assume the schedule is
6116 2013-07-13 Jason Merrill <jason@redhat.com>
6118 * print-tree.c (debug_vec_tree): Use debug_raw.
6119 (debug_raw (vec<tree, va_gc> &)): New.
6120 (debug_raw (vec<tree, va_gc> *)): New.
6121 * tree.h: Declare them.
6123 2013-07-13 Bin Cheng <bin.cheng@arm.com>
6125 * ifcvt.c (ifcvt_after_combine): New static variable.
6126 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
6128 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
6129 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
6130 rest_of_handle_if_after_reload): Pass new argument for if_convert.
6132 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
6134 * config/mips/mips.c (mips_expand_call): Remove empty statement.
6136 2013-07-12 Michael Matz <matz@suse.de>
6139 * convert.c (convert_to_real): Reject non-float inner types.
6141 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
6143 * config/aarch64/aarch64-protos.h
6144 (aarch64_simd_immediate_valid_for_move): Remove.
6145 * config/aarch64/aarch64.c (simd_immediate_info): New member.
6146 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
6148 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
6150 2013-07-11 Steve Ellcey <sellcey@mips.com>
6152 * config/mips/mips.c (mips_conditional_register_usage): Do not
6153 use t[0-7] registers in MIPS16 mode when optimizing for size.
6155 2013-07-11 Sriraman Tallam <tmsriram@google.com>
6157 * config/i386/i386.c (dispatch_function_versions): Fix array
6158 indexing of function_version_info to match actual_versions.
6160 2013-07-11 Teresa Johnson <tejohnson@google.com>
6162 * vec.h (struct va_gc): Move release out-of-line.
6163 (va_gc::release): Call ggc_free on released vec.
6165 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6167 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
6168 Require GOT register as additional operand in UNSPEC.
6169 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
6170 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6171 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6172 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
6173 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
6174 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
6175 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
6177 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
6180 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
6181 name seen by assembler/linker rather if available.
6183 2013-07-11 Andreas Schwab <schwab@suse.de>
6185 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
6187 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
6189 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
6191 2013-07-10 Joseph Myers <joseph@codesourcery.com>
6193 * doc/tm.texi.in: Move hook documentation to ....
6194 * target.def: ... here.
6196 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
6198 * doc/tm.texi: Regenerate.
6200 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
6203 * doc/invoke.texi: Document -Wconditionally-supported.
6205 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6208 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
6211 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6214 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
6215 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
6216 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
6217 Remove duplicate devices.
6218 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
6219 * config/avr/t-multilib: Regenerate.
6220 * config/avr/avr-tables.opt: Regenerate.
6221 * doc/avr-mmcu.texi: Regenerate.
6223 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
6226 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
6228 2013-07-10 Graham Stott <graham.stott@btinternet.com>
6230 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
6231 the cost of MULT when optimizing for size.
6233 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6235 * config/cr16/cr16-protos.h: Don't include target.h.
6237 2013-07-09 Joseph Myers <joseph@codesourcery.com>
6239 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
6240 adjust register size for TDmode and TFmode for VSX registers.
6242 2013-07-08 Kai Tietz <ktietz@redhat.com>
6245 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
6246 hook_bool_const_tree_true.
6248 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6250 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
6251 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
6252 * config/s390/s390.md: Define FPR*_REGNUM constants.
6253 Fix FPR2_REGNUM constant (18 -> 17).
6254 ("*trunc<BFP:mode><DFP_ALL:mode>2")
6255 ("*trunc<DFP_ALL:mode><BFP:mode>2")
6256 ("trunc<BFP:mode><DFP_ALL:mode>2")
6257 ("trunc<DFP_ALL:mode><BFP:mode>2")
6258 ("*extend<BFP:mode><DFP_ALL:mode>2")
6259 ("*extend<DFP_ALL:mode><BFP:mode>2")
6260 ("extend<BFP:mode><DFP_ALL:mode>2")
6261 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
6264 2013-07-08 Graham Stott <graham.stott@btinternet.com>
6266 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
6268 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6270 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
6271 and cfun_fpr_bit_p to cfun_fpr_save_p.
6272 (s390_frame_area, s390_register_info, s390_frame_info)
6273 (s390_emit_prologue, s390_emit_epilogue)
6274 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
6276 * config/s390/s390.h: Define *_REGNUM macros for floating point
6279 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
6281 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
6283 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
6285 PR rtl-optimization/57786
6286 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
6287 and break out of the loop when it is set to false.
6289 2013-07-08 Jakub Jelinek <jakub@redhat.com>
6292 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
6293 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
6294 (const_int 63)) 0)).
6295 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
6296 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
6297 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
6299 PR rtl-optimization/57829
6300 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
6301 mask bits outside of mode are just sign-extension from mode to HWI.
6303 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6305 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
6306 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
6307 adjust_address instead of change_address to keep info about alignment.
6308 (emit_strmov): Remove.
6309 (emit_memmov): New function.
6310 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
6311 (expand_movmem_epilogue): Likewise and return updated rtx for
6313 (expand_constant_movmem_prologue): Likewise and return updated rtx for
6314 destination and source.
6315 (decide_alignment): Refactor, handle vector_loop.
6316 (ix86_expand_movmem): Likewise.
6317 (ix86_expand_setmem): Likewise.
6318 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
6320 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
6322 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
6323 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
6325 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
6327 * config/i386/sse.md (sse_movlhps): Change alternative 3
6328 of operand 2 to "m".
6330 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
6333 * config/i386/sse.md (iptr): New mode attribute.
6334 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
6335 (<sse>_vm<plusminus_insn><mode>3): Ditto.
6336 (<sse>_vmmul<mode>3): Ditto.
6337 (<sse>_vmdiv<mode>3): Ditto.
6338 (sse_vmrcpv4sf2): Ditto.
6339 (<sse>_vmsqrt<mode>2): Ditto.
6340 (sse_vmrsqrtv4sf2): Ditto.
6341 (<sse>_vm<code><mode>3): Ditto.
6342 (avx_vmcmp<mode>3): Ditto.
6343 (<sse>_vmmaskcmp<mode>3): Ditto.
6344 (<sse>_comi): Ditto.
6345 (<sse>_ucomi): Ditto.
6346 (*xop_vmfrcz_<mode>): Ditto.
6347 (*fmai_fmadd_<mode>): Ditto.
6348 (*fmai_fmsub_<mode>): Ditto.
6349 (*fmai_fnmadd_<mode>): Ditto.
6350 (*fmai_fnmsub_<mode>): Ditto.
6351 (*fma4i_vmfmadd_<mode>): Ditto.
6352 (*fma4i_vmfmsub_<mode>): Ditto.
6353 (*fma4i_vmfnmadd_<mode>): Ditto.
6354 (*fma4i_vmfnmsub_<mode>): Ditto.
6355 (*xop_vmfrcz_<mode>): Ditto.
6356 (sse_cvtps2pi): Ditto.
6357 (sse_cvttps2pi): Ditto.
6358 (sse_cvtss2si): Ditto.
6359 (sse_cvtss2si_2): Ditto.
6360 (sse_cvtss2siq_2): Ditto.
6361 (sse_cvttss2si): Ditto.
6362 (sse_cvttss2siq): Ditto.
6363 (sse_cvtsd2si): Ditto.
6364 (sse_cvtsd2si_2): Ditto.
6365 (sse_cvtsd2siq_2): Ditto.
6366 (sse_cvttsd2si): Ditto.
6367 (sse_cvttsd2siq): Ditto.
6368 (sse_cvtsd2ss): Ditto.
6369 (sse_cvtss2sd): Ditto.
6370 (avx2_pbroadcast<mode>): Ditto.
6371 (avx2_pbroadcast<mode>_1): Ditto.
6372 (*avx_vperm_broadcast_v4sf): Ditto.
6374 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
6375 (sse_movlhps): Ditto.
6376 (sse_storehps): Ditto.
6377 (sse_loadhps): Ditto.
6378 (sse_storelps): Ditto.
6379 (sse_loadlps): Ditto.
6380 (*vec_concatv4sf): Ditto.
6381 (*vec_interleave_highv2df): Ditto.
6382 (*vec_interleave_lowv2df): Ditto.
6383 (*vec_extractv2df_1_sse): Ditto.
6384 (*vec_extractv2df_0_sse): Ditto.
6385 (sse2_storelpd): Ditto.
6386 (sse2_loadlpd): Ditto.
6387 (sse2_movsd): Ditto.
6388 (*vec_concatv4si): Ditto.
6389 (vec_concatv2di): Ditto.
6391 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
6392 for Intel asm dialect.
6393 (mmx_punpcklwd): Ditto.
6394 (mmx_punpckldq): Ditto.
6396 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
6397 for intel assembler dialect.
6399 2013-07-06 Jakub Jelinek <jakub@redhat.com>
6402 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
6403 for BUILT_IN_C{LZ,LRSB}*.
6404 * tree.h (CASE_INT_FN): Add FN##IMAX case.
6405 * tree-vrp.c (extract_range_basic): Handle
6406 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
6407 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
6408 fall thru to code calling set_value*.
6409 * builtins.c (expand_builtin): Remove *IMAX cases.
6410 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
6411 if width is bigger than 2*HWI.
6413 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
6415 PR rtl-optimization/55342
6416 * lra-int.h (lra_subreg_reload_pseudos): New.
6417 * lra.c: Add undoing optional reloads to the block diagram.
6418 (lra_subreg_reload_pseudos): New.
6419 (lra_optional_reload_pseudos): Change comments.
6420 (lra): Init and clear lra_subreg_reload_pseudos. Clear
6421 lra_optional_reload_pseudos after undo transformations.
6422 * lra-assigns.c (pseudo_prefix_title): New.
6423 (lra_setup_reg_renumber): Use it.
6424 (spill_for): Ditto. Check subreg reload pseudos too.
6425 (assign_by_spills): Consider subreg reload pseudos too.
6426 * lra-constraints.c (simplify_operand_subreg): Use
6427 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
6428 (curr_insn_transform): Recognize and do optional reloads.
6429 (undo_optional_reloads): New.
6430 (lra_undo_inheritance): Call undo_optional_reloads.
6432 2013-07-05 Thomas Quinot <quinot@adacore.com>
6434 * tree-complex.c (expand_complex_operations_1): Fix typo.
6436 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
6438 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
6439 (tune_params): New member 'const vec_costs'.
6440 * config/aarch64/aarch64.c (generic_vector_cost): New.
6441 (generic_tunings): New member 'generic_vector_cost'.
6442 (aarch64_builtin_vectorization_cost): New.
6443 (aarch64_add_stmt_cost): New.
6444 (TARGET_VECTORIZE_ADD_STMT_COST): New.
6445 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
6447 2013-07-03 Jakub Jelinek <jakub@redhat.com>
6450 * config/i386/predicates.md (vsib_address_operand): Disallow
6451 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
6453 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
6456 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
6457 expand_builtin_setjmp_receiver.
6458 (expand_label): Adjust, call expand_builtin_setjmp_receiver
6459 with NULL for the label parameter.
6460 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
6461 the frame-pointer. Adjust comments.
6462 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
6463 only if LABEL is non-NULL.
6465 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
6467 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
6468 (ARM_ABI_AAPCS64): Ditto.
6470 (ARM_DEFAULT_ABI): Ditto.
6472 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
6474 * config/aarch64/aarch64-builtins.c
6475 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
6476 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
6478 * config/aarch64/aarch64-simd.md
6479 (aarch64_ld1<VALL:mode>): New.
6480 (aarch64_st1<VALL:mode>): Likewise.
6481 * config/aarch64/arm_neon.h
6482 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
6484 2013-07-02 Sriraman Tallam <tmsriram@google.com>
6486 * config/i386/i386.c (gate_insert_vzeroupper): Check if
6488 (ix86_option_override_internal):Turn on all -mavx target flags by
6489 default as they are dependent on AVX anyway.
6491 2013-07-02 Cary Coutant <ccoutant@google.com>
6493 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
6495 (loc_checksum_ordered): Likewise.
6496 (hash_loc_operands): Remove inline keyword.
6498 2013-07-02 Jakub Jelinek <jakub@redhat.com>
6500 PR tree-optimization/57741
6501 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
6502 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
6503 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
6504 Allow REAL_CST step_exprs if flag_associative_math.
6505 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
6507 2013-07-02 Ian Bolton <ian.bolton@arm.com>
6509 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
6511 2013-07-02 Ian Bolton <ian.bolton@arm.com>
6513 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
6515 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6517 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
6519 (iorsi3_insn): Likewise.
6520 (arm_xorsi3): Likewise.
6522 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
6524 * arm.md (attribute "wtype"): Delete. Move attribute values from here
6526 (attribute "type"): ... here, and prefix with "wmmx_".
6527 (attribute "core_cycles"): Update for attribute changes.
6528 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
6529 (tbcstv4hi): Likewise.
6530 (tbcstv2si): Likewise.
6531 (iwmmxt_iordi3): Likewise.
6532 (iwmmxt_xordi3): Likewise.
6533 (iwmmxt_anddi3): Likewise.
6534 (iwmmxt_nanddi3): Likewise.
6535 (iwmmxt_arm_movdi): Likewise.
6536 (iwmmxt_movsi_insn): Likewise.
6537 (mov<mode>_internal): Likewise.
6538 (and<mode>3_iwmmxt): Likewise.
6539 (ior<mode>3_iwmmxt): Likewise.
6540 (xor<mode>3_iwmmxt): Likewise.
6541 (add<mode>3_iwmmxt): Likewise.
6542 (ssaddv8qi3): Likewise.
6543 (ssaddv4hi3): Likewise.
6544 (ssaddv2si3): Likewise.
6545 (usaddv8qi3): Likewise.
6546 (usaddv4hi3): Likewise.
6547 (usaddv2si3): Likewise.
6548 (sub<mode>3_iwmmxt): Likewise.
6549 (sssubv8qi3): Likewise.
6550 (sssubv4hi3): Likewise.
6551 (sssubv2si3): Likewise.
6552 (ussubv8qi3): Likewise.
6553 (ussubv4hi3): Likewise.
6554 (ussubv2si3): Likewise.
6555 (mulv4hi3_iwmmxt): Likewise.
6556 (smulv4hi3_highpart): Likewise.
6557 (umulv4hi3_highpart): Likewise.
6558 (iwmmxt_wmacs): Likewise.
6559 (iwmmxt_wmacsz): Likewise.
6560 (iwmmxt_wmacu): Likewise.
6561 (iwmmxt_wmacuz): Likewise.
6562 (iwmmxt_clrdi): Likewise.
6563 (iwmmxt_clrv8qi): Likewise.
6564 (iwmmxt_clr4hi): Likewise.
6565 (iwmmxt_clr2si): Likewise.
6566 (iwmmxt_uavgrndv8qi3): Likewise.
6567 (iwmmxt_uavgrndv4hi3): Likewise.
6568 (iwmmxt_uavgv8qi3): Likewise.
6569 (iwmmxt_uavgv4hi3): Likewise.
6570 (iwmmxt_tinsrb): Likewise.
6571 (iwmmxt_tinsrh): Likewise.
6572 (iwmmxt_tinsrw): Likewise.
6573 (iwmmxt_textrmub): Likewise.
6574 (iwmmxt_textrmsb): Likewise.
6575 (iwmmxt_textrmuh): Likewise.
6576 (iwmmxt_textrmsh): Likewise.
6577 (iwmmxt_textrmw): Likewise.
6578 (iwmxxt_wshufh): Likewise.
6579 (eqv8qi3): Likewise.
6580 (eqv4hi3): Likewise.
6581 (eqv2si3): Likewise.
6582 (gtuv8qi3): Likewise.
6583 (gtuv4hi3): Likewise.
6584 (gtuv2si3): Likewise.
6585 (gtv8qi3): Likewise.
6586 (gtv4hi3): Likewise.
6587 (gtv2si3): Likewise.
6588 (smax<mode>3_iwmmxt): Likewise.
6589 (umax<mode>3_iwmmxt): Likewise.
6590 (smin<mode>3_iwmmxt): Likewise.
6591 (umin<mode>3_iwmmxt): Likewise.
6592 (iwmmxt_wpackhss): Likewise.
6593 (iwmmxt_wpackwss): Likewise.
6594 (iwmmxt_wpackdss): Likewise.
6595 (iwmmxt_wpackhus): Likewise.
6596 (iwmmxt_wpackwus): Likewise.
6597 (iwmmxt_wpackdus): Likewise.
6598 (iwmmxt_wunpckihb): Likewise.
6599 (iwmmxt_wunpckihh): Likewise.
6600 (iwmmxt_wunpckihw): Likewise.
6601 (iwmmxt_wunpckilb): Likewise.
6602 (iwmmxt_wunpckilh): Likewise.
6603 (iwmmxt_wunpckilw): Likewise.
6604 (iwmmxt_wunpckehub): Likewise.
6605 (iwmmxt_wunpckehuh): Likewise.
6606 (iwmmxt_wunpckehuw): Likewise.
6607 (iwmmxt_wunpckehsb): Likewise.
6608 (iwmmxt_wunpckehsh): Likewise.
6609 (iwmmxt_wunpckehsw): Likewise.
6610 (iwmmxt_wunpckelub): Likewise.
6611 (iwmmxt_wunpckeluh): Likewise.
6612 (iwmmxt_wunpckeluw): Likewise.
6613 (iwmmxt_wunpckelsb): Likewise.
6614 (iwmmxt_wunpckelsh): Likewise.
6615 (iwmmxt_wunpckelsw): Likewise.
6616 (ror<mode>3): Likewise.
6617 (ashr<mode>3_iwmmxt): Likewise.
6618 (lshr<mode>3_iwmmxt): Likewise.
6619 (ashl<mode>3_iwmmxt): Likewise.
6620 (ror<mode>3_di): Likewise.
6621 (ashr<mode>3_di): Likewise.
6622 (lshr<mode>3_di): Likewise.
6623 (ashl<mode>3_di): Likewise.
6624 (iwmmxt_wmadds): Likewise.
6625 (iwmmxt_wmaddu): Likewise.
6626 (iwmmxt_tmia): Likewise.
6627 (iwmmxt_tmiaph): Likewise.
6628 (iwmmxt_tmiabb): Likewise.
6629 (iwmmxt_tmiatb): Likewise.
6630 (iwmmxt_tmiabt): Likewise.
6631 (iwmmxt_tmiatt): Likewise.
6632 (iwmmxt_tmovmskb): Likewise.
6633 (iwmmxt_tmovmskh): Likewise.
6634 (iwmmxt_tmovmskw): Likewise.
6635 (iwmmxt_waccb): Likewise.
6636 (iwmmxt_wacch): Likewise.
6637 (iwmmxt_waccw): Likewise.
6638 (iwmmxt_waligni): Likewise.
6639 (iwmmxt_walignr): Likewise.
6640 (iwmmxt_walignr0): Likewise.
6641 (iwmmxt_walignr1): Likewise.
6642 (iwmmxt_walignr2): Likewise.
6643 (iwmmxt_walignr3): Likewise.
6644 (iwmmxt_wsadb): Likewise.
6645 (iwmmxt_wsadh): Likewise.
6646 (iwmmxt_wsadbz): Likewise.
6647 (iwmmxt_wsadhz): Likewise.
6648 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
6649 (iwmmxt_wabsdiffb): Likewise.
6650 (iwmmxt_wabsdiffh): Likewise.
6651 (iwmmxt_wabsdiffw): Likewise.
6652 (iwmmxt_waddsubhx): Likewise
6653 (iwmmxt_wsubaddhx): Likewise.
6654 (addc<mode>3): Likewise.
6655 (iwmmxt_avg4): Likewise.
6656 (iwmmxt_avg4r): Likewise.
6657 (iwmmxt_wmaddsx): Likewise.
6658 (iwmmxt_wmaddux): Likewise.
6659 (iwmmxt_wmaddsn): Likewise.
6660 (iwmmxt_wmaddun): Likewise.
6661 (iwmmxt_wmulwsm): Likewise.
6662 (iwmmxt_wmulwum): Likewise.
6663 (iwmmxt_wmulsmr): Likewise.
6664 (iwmmxt_wmulumr): Likewise.
6665 (iwmmxt_wmulwsmr): Likewise.
6666 (iwmmxt_wmulwumr): Likewise.
6667 (iwmmxt_wmulwl): Likewise.
6668 (iwmmxt_wqmulm): Likewise.
6669 (iwmmxt_wqmulwm): Likewise.
6670 (iwmmxt_wqmulmr): Likewise.
6671 (iwmmxt_wqmulwmr): Likewise.
6672 (iwmmxt_waddbhusm): Likewise.
6673 (iwmmxt_waddbhusl): Likewise.
6674 (iwmmxt_wqmiabb): Likewise.
6675 (iwmmxt_wqmiabt): Likewise.
6676 (iwmmxt_wqmiatb): Likewise.
6677 (iwmmxt_wqmiatt): Likewise.
6678 (iwmmxt_wqmiabbn): Likewise.
6679 (iwmmxt_wqmiabtn): Likewise.
6680 (iwmmxt_wqmiatbn): Likewise.
6681 (iwmmxt_wqmiattn): Likewise.
6682 (iwmmxt_wmiabb): Likewise.
6683 (iwmmxt_wmiabt): Likewise.
6684 (iwmmxt_wmiatb): Likewise.
6685 (iwmmxt_wmiatt): Likewise.
6686 (iwmmxt_wmiabbn): Likewise.
6687 (iwmmxt_wmiabtn): Likewise.
6688 (iwmmxt_wmiatbn): Likewise.
6689 (iwmmxt_wmiattn): Likewise.
6690 (iwmmxt_wmiawbb): Likewise.
6691 (iwmmxt_wmiawbt): Likewise.
6692 (iwmmxt_wmiawtb): Likewise.
6693 (iwmmxt_wmiawtt): Likewise.
6694 (iwmmxt_wmiawbbn): Likewise.
6695 (iwmmxt_wmiawbtn): Likewise.
6696 (iwmmxt_wmiawtbn): Likewise.
6697 (iwmmxt_wmiawttn): Likewise.
6698 (iwmmxt_wmerge): Likewise.
6699 (iwmmxt_tandc<mode>3): Likewise.
6700 (iwmmxt_torc<mode>3): Likewise.
6701 (iwmmxt_torvsc<mode>3): Likewise.
6702 (iwmmxt_textrc<mode>3): Likewise.
6703 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
6704 (wmmxt_pack): Likewise.
6705 (wmmxt_mult_c1): Likewise.
6706 (wmmxt_mult_c2): Likewise.
6707 (wmmxt_alu_c1): Likewise.
6708 (wmmxt_alu_c2): Likewise.
6709 (wmmxt_alu_c3): Likewise.
6710 (wmmxt_transfer_c1): Likewise.
6711 (wmmxt_transfer_c2): Likewise.
6712 (wmmxt_transfer_c3): Likewise.
6713 (marvell_f_iwmmxt_wstr): Likewise.
6714 (marvell_f_iwmmxt_wldr): Likewise.
6716 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
6718 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
6720 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6723 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6724 * lra-constraints.c (need_for_split_p): Check call used hard regs
6725 living through calls.
6727 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
6728 call used regs for call insn.
6730 2013-06-28 Jakub Jelinek <jakub@redhat.com>
6733 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
6734 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
6736 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6738 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
6740 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6742 * lra-constraints.c (need_for_split_p): Check call used hard regs
6743 living through calls.
6745 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
6748 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
6749 to tie with any other modes. Eliminate Altivec vector mode tests,
6750 since these are a subset of ALTIVEC or VSX vector modes. Simplify
6751 code, to return 0 if testing MODE2 for a condition, if we've
6752 already tested MODE1 for the same condition.
6754 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6756 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
6759 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6761 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
6762 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
6764 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6766 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
6768 (aarch64_symbolic_constant_p): Remove.
6769 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
6770 static. Fix line length and white space.
6771 (aarch64_symbolic_constant_p): Remove.
6772 * config/aarch64/predicates.md (aarch64_valid_symref):
6773 Use aarch64_classify_symbol_expression.
6775 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6777 * config/arm/constraints.md (Ts): New constraint.
6778 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
6780 (compare_scc): Use "Ts" constraint for operand 0.
6781 (ior_scc_scc): Likewise.
6782 (and_scc_scc): Likewise.
6783 (and_scc_scc_nodom): Likewise.
6784 (ior_scc_scc_cmp): Likewise for operand 7.
6785 (and_scc_scc_cmp): Likewise.
6786 * config/arm/thumb2.md (thumb2_movsi_insn):
6787 Add alternatives for 16-bit encodings.
6788 (thumb2_movhi_insn): Likewise.
6789 (thumb2_movsicc_insn): Likewise.
6790 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
6791 (thumb2_negscc): Use "Ts" constraint.
6792 Move mvn instruction outside cond_exec block.
6793 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
6794 for 16-bit encodings.
6796 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6798 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
6800 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
6801 (mulsi3subsi): Likewise.
6802 (mulsidi3adddi): Likewise.
6803 (mulsidi3_v6): Likewise.
6804 (umulsidi3_v6): Likewise.
6805 (umulsidi3adddi_v6): Likewise.
6806 (smulsi3_highpart_v6): Likewise.
6807 (umulsi3_highpart_v6): Likewise.
6808 (mulhisi3tb): Likewise.
6809 (mulhisi3bt): Likewise.
6810 (mulhisi3tt): Likewise.
6811 (maddhisi4): Likewise.
6812 (maddhisi4tb): Likewise.
6813 (maddhisi4tt): Likewise.
6814 (maddhidi4): Likewise.
6815 (maddhidi4tb): Likewise.
6816 (maddhidi4tt): Likewise.
6817 (zeroextractsi_compare0_scratch): Likewise.
6818 (insv_zero): Likewise.
6819 (insv_t2): Likewise.
6820 (anddi_notzesidi_di): Likewise.
6821 (anddi_notsesidi_di): Likewise.
6822 (andsi_notsi_si): Likewise.
6823 (iordi_zesidi_di): Likewise.
6824 (xordi_zesidi_di): Likewise.
6825 (andsi_iorsi3_notsi): Likewise.
6827 (smax_m1): Likewise.
6829 (not_shiftsi): Likewise.
6830 (unaligned_loadsi): Likewise.
6831 (unaligned_loadhis): Likewise.
6832 (unaligned_loadhiu): Likewise.
6833 (unaligned_storesi): Likewise.
6834 (unaligned_storehi): Likewise.
6835 (extv_reg): Likewise.
6836 (extzv_t2): Likewise.
6838 (udivsi3): Likewise.
6839 (arm_zero_extendhisi2addsi): Likewise.
6840 (arm_zero_extendqisi2addsi): Likewise.
6841 (compareqi_eq0): Likewise.
6842 (arm_extendhisi2_v6): Likewise.
6843 (arm_extendqisi2addsi): Likewise.
6844 (arm_movt): Likewise.
6845 (thumb2_ldrd): Likewise.
6846 (thumb2_ldrd_base): Likewise.
6847 (thumb2_ldrd_base_neg): Likewise.
6848 (thumb2_strd): Likewise.
6849 (thumb2_strd_base): Likewise.
6850 (thumb2_strd_base_neg): Likewise.
6851 (arm_negsi2): Add alternative for 16-bit encoding.
6852 (arm_one_cmplsi2): Likewise.
6854 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6856 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
6857 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
6858 (movdfcc): Likewise.
6859 * config/arm/vfp.md (*thumb2_movsf_vfp):
6860 Disable predication for arm_restrict_it.
6861 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
6862 (*thumb2_movdfcc_vfp): Likewise.
6863 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
6864 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
6865 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
6866 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
6867 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
6868 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
6869 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
6870 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
6871 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
6872 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
6873 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
6874 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
6875 Disable predication for arm_restrict_it.
6877 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
6879 * config/i386/bmiintrin.h (_bextr_u32): New.
6880 (_bextr_u64): Ditto.
6882 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
6884 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
6885 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
6886 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
6887 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
6888 * config/mips/n32-elf.h: ...this new file.
6890 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
6893 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
6894 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
6896 2013-06-27 Catherine Moore <clm@codesourcery.com>
6898 * config/mips/mips-tables.opt: Regenerate.
6899 * config/mips/mips-cpus.def: Add m14ke and m14kec.
6900 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
6901 * doc/invoke.texi: Add -m14kc.
6903 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6906 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
6907 constraints of operand 1 and 2.
6910 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
6911 to match RTL canonicalization. Swap predicates and
6912 constraints of operand 1 and 2.
6914 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
6916 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
6917 Process OP_INOUT regs for splitting too.
6919 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6921 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
6922 decl before the loop, initialize to NULL.
6923 (vectorizable_load): Initialize ptr_incr to NULL.
6925 2013-06-27 Martin Jambor <mjambor@suse.cz>
6928 * ipa-ref.h (ipa_maybe_record_reference): Declare.
6929 * ipa-ref.c (ipa_maybe_record_reference): New function.
6930 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
6931 * ipa-cp.c (create_specialized_node): Record potential references from
6933 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
6935 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6937 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
6938 parameter 'mode' of type 'enum machine_mode mode'; change to pass
6939 'mode' to force_reg.
6940 (aarch64_add_offset): Update calls to aarch64_force_temporary.
6941 (aarch64_expand_mov_immediate): Likewise.
6943 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6945 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
6946 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
6948 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6950 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
6951 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6952 (struct machine_function): Add tbegin_p.
6953 (s390_canonicalize_comparison): Fold CC mode compares to
6954 conditional jump if possible.
6955 (s390_emit_jump): Return the emitted jump.
6956 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
6957 Handle CCRAWmode compares.
6958 (s390_option_override): Default to -mhtm if available.
6959 (s390_reg_clobbered_rtx): Handle floating point regs as well.
6960 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
6961 FPRs instead of df_regs_ever_live_p.
6962 (s390_optimize_nonescaping_tx): New function.
6963 (s390_init_frame_layout): Extend clobbered_regs array to cover
6965 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
6966 (s390_expand_tbegin): New function.
6967 (enum s390_builtin): New enum definition.
6968 (code_for_builtin): New array definition.
6969 (s390_init_builtins): New function.
6970 (s390_expand_builtin): New function.
6971 (TARGET_INIT_BUILTINS): Define.
6972 (TARGET_EXPAND_BUILTIN): Define.
6973 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
6974 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
6975 (s390_alc_comparison): Likewise.
6976 * config/s390/s390-modes.def: Add CCRAWmode.
6977 * config/s390/s390.h (processor_flags): Add PF_TX.
6978 (TARGET_CPU_HTM): Define macro.
6979 (TARGET_HTM): Define macro.
6980 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
6981 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
6982 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6983 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
6984 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
6985 (TBEGIN_MASK, TBEGINC_MASK): New constants.
6986 ("*cc_to_int"): Move up.
6987 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
6988 constants other than 0.
6989 ("*ccraw_to_int"): New insn and splitter definition.
6990 ("tbegin", "tbegin_nofloat", "tbegin_retry")
6991 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
6992 ("tx_assist"): New expander.
6993 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
6994 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
6995 * config/s390/s390.opt: Add -mhtm option.
6996 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
6997 * config/s390/htmxlintrin.h: New file.
6998 * config/s390/htmintrin.h: New file.
6999 * config/s390/s390intrin.h: New file.
7000 * doc/extend.texi: Document htm builtins.
7001 * config.gcc: Add the new header files to extra_headers.
7003 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
7005 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
7006 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
7008 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
7009 Pat Haugen <pthaugen@us.ibm.com>
7010 Peter Bergner <bergner@vnet.ibm.com>
7012 * config/rs6000/power8.md: New.
7013 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
7014 setting for power8 entry.
7015 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
7016 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
7017 test for Power4/Power5 only.
7018 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
7020 (force_new_group): Adjust comment.
7021 * config/rs6000/rs6000.md: Include power8.md.
7023 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
7025 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
7026 * config/arm/arm-protos.h (arm_max_conditional_execute): New
7028 (tune_params): Update comment.
7029 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
7030 (arm_max_conditional_execute): New function.
7031 (thumb2_final_prescan_insn): Use max_insn_skipped and
7032 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
7034 2013-06-25 Jakub Jelinek <jakub@redhat.com>
7036 PR tree-optimization/57705
7037 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
7038 SSA_NAME step, provided that it is not defined inside the loop.
7039 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
7040 (get_initial_def_for_induction): Handle SSA_NAME IV step.
7042 2013-06-25 Martin Jambor <mjambor@suse.cz>
7045 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
7046 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
7048 (ipa_note_param_call): Initialize member_ptr flag.
7049 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
7050 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
7051 (ipa_write_indirect_edge_info): Stream member_ptr flag.
7052 (ipa_read_indirect_edge_info): Likewise.
7054 2013-06-25 Richard Biener <rguenther@suse.de>
7057 * passes.c (init_optimization_passes): Move pass_fold_builtins
7058 and pass_dce earlier with -Og.
7060 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
7062 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
7063 <BIT_FIELD_REF>: Remove trailing TAB.
7064 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
7067 2013-06-24 Martin Jambor <mjambor@suse.cz>
7069 PR tree-optimization/57358
7070 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
7071 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
7072 (ipa_analyze_params_uses): Generate pessimistic info when true.
7074 2013-06-24 Martin Jambor <mjambor@suse.cz>
7076 PR tree-optimization/57539
7077 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
7078 global.inlined_to of the new node to it. All callers changed.
7079 * ipa-inline-transform.c (clone_inlined_nodes): New variable
7080 inlining_into, pass it to cgraph_clone_node.
7081 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
7082 ipa_free_edge_args_substructures.
7083 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
7084 rdesc linked list. Do not assert rdesc edges have inlined caller.
7085 Assert we have found an rdesc in the rdesc list.
7087 2013-06-24 Richard Biener <rguenther@suse.de>
7089 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7090 (pointer_set_lookup): Declare.
7091 (class pointer_map): New template class implementing a
7092 generic pointer to T map.
7093 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
7094 pointer_map<T>::contains, pointer_map<T>::insert,
7095 pointer_map<T>::traverse): New functions.
7096 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
7097 (pointer_set_lookup): New function.
7098 (pointer_set_contains): Use pointer_set_lookup.
7099 (pointer_set_insert): Likewise.
7100 (insert_aux): Remove.
7101 (struct pointer_map_t): Embed a pointer_set_t.
7102 (pointer_map_create): Adjust.
7103 (pointer_map_destroy): Likewise.
7104 (pointer_map_contains): Likewise.
7105 (pointer_map_insert): Likewise.
7106 (pointer_map_traverse): Likewise.
7107 * tree-streamer.h (struct streamer_tree_cache_d): Use a
7108 pointer_map<unsigned> instead of a pointer_map_t.
7109 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
7110 (streamer_tree_cache_lookup): Likewise.
7111 (streamer_tree_cache_create): Likewise.
7112 (streamer_tree_cache_delete): Likewise.
7113 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
7114 pointer_map<unsigned> instead of a pointer_map_t.
7115 (lto_init_tree_ref_encoder): Adjust.
7116 (lto_destroy_tree_ref_encoder): Likewise.
7117 * lto-section-out.c (lto_output_decl_index): Likewise.
7118 (lto_record_function_out_decl_state): Likewise.
7119 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
7121 2013-06-24 Richard Biener <rguenther@suse.de>
7123 PR tree-optimization/57488
7124 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
7126 2013-06-24 Alan Modra <amodra@gmail.com>
7128 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
7129 (gen_easy_altivec_constant): Likewise.
7130 * config/rs6000/predicates.md (easy_vector_constant_add_self,
7131 easy_vector_constant_msb): Likewise.
7133 2013-06-23 Jakub Jelinek <jakub@redhat.com>
7136 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
7137 add missing return true.
7139 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
7142 * config/sh/predicates.md (general_extend_operand): Invoke
7143 general_movsrc_operand for memory operands.
7144 (general_movsrc_operand): Allow reg+reg addressing, do not use
7145 general_operand for memory operands.
7147 2013-06-23 Sriraman Tallam <tmsriram@google.com>
7149 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
7150 when current target options does not apply.
7151 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
7152 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
7153 * config/i386/bmiintrin.h: Pass appropriate target
7154 attributes to header.
7155 * config/i386/mmintrin.h: Ditto.
7156 * config/i386/nmmintrin.h: Ditto.
7157 * config/i386/avx2intrin.h: Ditto.
7158 * config/i386/fxsrintrin.h: Ditto.
7159 * config/i386/tbmintrin.h: Ditto.
7160 * config/i386/xsaveintrin.h: Ditto.
7161 * config/i386/f16cintrin.h: Ditto.
7162 * config/i386/xtestintrin.h: Ditto.
7163 * config/i386/xsaveoptintrin.h: Ditto.
7164 * config/i386/bmi2intrin.h: Ditto.
7165 * config/i386/lzcntintrin.h: Ditto.
7166 * config/i386/smmintrin.h: Ditto.
7167 * config/i386/wmmintrin.h: Ditto.
7168 * config/i386/x86intrin.h: Remove all header include guards.
7169 * config/i386/prfchwintrin.h: Ditto.
7170 * config/i386/pmmintrin.h: Ditto.
7171 * config/i386/tmmintrin.h: Ditto.
7172 * config/i386/xmmintrin.h: Ditto.
7173 * config/i386/popcntintrin.h: Ditto.
7174 * config/i386/rdseedintrin.h: Ditto.
7175 * config/i386/ammintrin.h: Ditto.
7176 * config/i386/emmintrin.h: Ditto.
7177 * config/i386/immintrin.h: Remove all header include guards.
7178 * config/i386/fma4intrin.h: Ditto.
7179 * config/i386/lwpintrin.h: Ditto.
7180 * config/i386/xopintrin.h: Ditto.
7181 * config/i386/ia32intrin.h: Ditto.
7182 * config/i386/avxintrin.h: Ditto.
7183 * config/i386/rtmintrin.h: Ditto.
7184 * config/i386/fmaintrin.h: Ditto.
7185 * config/i386/mm3dnow.h: Ditto.
7187 2013-06-22 Sriraman Tallam <tmsriram@google.com>
7189 * common/config/i386/i386-common.c: Handle LZCNT.
7191 2013-06-22 Andi Kleen <ak@linux.intel.com>
7193 * doc/extend.texi: Use __atomic_store_n instead of
7194 __atomic_store in HLE example.
7196 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
7198 * config/sh/sh.c: Remove <cstdlib> workaround.
7200 2013-06-21 Andi Kleen <ak@linux.intel.com>
7202 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
7204 2013-06-21 Andi Kleen <ak@linux.intel.com>
7206 * doc/extend.texi: Document that __atomic_clear and
7207 __atomic_test_and_set should only be used with bool.
7209 2013-06-20 Jan Hubicka <jh@suse.cz>
7211 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
7213 * tree.c (decls_same_for_odr): New function.
7214 (same_for_edr): New function.
7215 (types_same_for_odr): New function.
7216 (get_binfo_at_offset): Use it.
7217 * tree.h (types_same_for_odr): Declare.
7219 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
7220 Jason Merrill <jason@redhat.com>
7222 * system.h: Include <cstdlib> as well as <stdlib.h>.
7224 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
7227 * config/i386/i386.c (construct_container): Report error if
7228 long double is used with disabled x87 float returns.
7230 2013-06-20 Jan Hubicka <jh@suse.cz>
7232 * lto-cgraph.c (input_symtab): Do not set cgraph state.
7234 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
7236 PR rtl-optimization/57425
7237 PR rtl-optimization/57569
7238 * alias.c (write_dependence_p): Remove parameters mem_mode and
7239 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
7240 Changed all callers.
7241 (canon_anti_dependence): Get comments and semantics in sync.
7242 Add parameter mem_canonicalized. Changed all callers.
7243 * rtl.h (canon_anti_dependence): Update prototype.
7245 2013-06-20 Richard Biener <rguenther@suse.de>
7247 * data-streamer-in.c (streamer_read_uhwi): Optimize single
7248 byte case, inline streamer_read_uchar and defer section
7251 2013-06-20 Richard Biener <rguenther@suse.de>
7253 PR tree-optimization/57584
7254 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
7255 SSA names into the expanded expression that take part in
7256 abnormal coalescing.
7258 2013-06-19 Sharad Singhai <singhai@google.com>
7260 * gcov.c (print_usage): Handle new option.
7261 (process_args): Ditto.
7262 (get_gcov_intermediate_filename): New function.
7263 (output_intermediate_file): New function.
7264 (output_gcov_file): New function
7265 (generate_results): Handle new option.
7266 (release_function): Relase demangled name.
7267 (read_graph_file): Handle demangled name.
7268 (output_lines): Ditto.
7269 * doc/gcov.texi: Document gcov intermediate format.
7271 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
7274 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
7275 (lra_emit_add): Use the functions. Add comment about Y as an
7278 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
7281 * collect2.c (collect_atexit): New.
7282 (collect_exit): Delete.
7283 (main): Register collect_atexit with atexit.
7284 (collect_wait): Change collect_exit to exit.
7286 * collect2.h (collect_exit): Delete.
7287 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
7289 2013-06-19 Wei Mi <wmi@google.com>
7291 PR rtl-optimization/57518
7292 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
7293 if regno is used in paradoxical subreg.
7294 (update_equiv_regs): Check pdx_subregs[regno] before
7295 set a reg to be equivalent with a mem.
7297 2013-06-19 Matthias Klose <doko@ubuntu.com>
7300 * file-find.h (find_a_file): Add a mode parameter.
7301 * file-find.c (find_a_file): Likewise.
7302 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
7303 with X_OK for the executables.
7304 * collect2.c (main): Call find_a_file with X_OK.
7306 2013-06-19 Steve Ellcey <sellcey@mips.com>
7309 * config/mips/mips.md (casesi_internal_mips16_<mode>):
7310 Use NEXT_INSN instead of next_real_insn.
7312 2013-06-19 Jan Hubicka <jh@suse.cz>
7314 * cgraph.h (const_value_known_p): Replace by ...
7315 (ctor_for_folding): .. this one.
7316 * cgraphunit.c (process_function_and_variable_attributes): Use it.
7317 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
7318 * expr.c (expand_expr_real_1): Likewise.
7319 (string_constant): Likewise.
7320 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
7321 * ipa.c (process_references): Likewise.
7322 (symtab_remove_unreachable_nodes): Likewise.
7323 * ipa-inline-analysis.c (param_change_prob): Likewise.
7324 * gimple-fold.c (canonicalize_constructor_val): Likewise.
7325 (get_base_constructor): Likwise.
7326 * varpool.c (varpool_remove_node): Likewise.
7327 (varpool_remove_initializer): LIkewise.
7328 (dump_varpool_node): LIkwise.
7329 (const_value_known_p): Rewrite to ...
7330 (ctor_for_folding): ... this one.
7332 2013-06-19 Jakub Jelinek <jakub@redhat.com>
7335 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
7336 PERSONALITY in $PATH derived prefixes.
7338 2013-06-19 Jeff Law <law@redhat.com>
7340 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
7343 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
7344 (simplify_bitwise_binary): Use it to simpify certain binary ops on
7347 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
7349 * config/arm/vfp.md: Move VFP instruction classification documentation
7351 * config/arm/arm.md: ... here. Update instruction classification
7354 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
7356 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
7357 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
7358 pattern. Use more efficient sequences on ARMv5 and Thumb2.
7360 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
7363 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7364 with NEXT_INSN. Use tablejump_p to check for jump table data
7367 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
7370 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
7371 that now in C++ the value is correct per the C++ standards.
7373 2013-06-19 Richard Biener <rguenther@suse.de>
7375 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
7378 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7381 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7384 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7385 with next_active_insn.
7387 2013-06-18 Sriraman Tallam <tmsriram@google.com>
7389 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
7390 functions are inlined during failures to flag an error.
7391 * tree-inline.c (expand_call_inline): Allow the error to be flagged
7392 in early inline pass.
7394 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
7396 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
7399 2013-06-18 Julian Brown <julian@codesourcery.com>
7401 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
7402 Permit virtual register pre-reload if !strict.
7403 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
7405 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
7407 * config/arm/neon.md (movmisalign<mode>): Use
7408 neon_perm_struct_or_reg_operand instead of
7409 neon_struct_or_register_operand.
7410 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
7411 neon_permissive_struct_operand instead of neon_struct_operand.
7412 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
7413 neon_vector_mem_operand.
7414 * config/arm/predicates.md (neon_struct_operand): Adjust call to
7415 neon_vector_mem_operand.
7416 (neon_permissive_struct_operand): New.
7417 (neon_struct_or_register_operand): Rename to...
7418 (neon_perm_struct_or_reg_operand): This. Adjust call to
7419 neon_vector_mem_operand.
7421 2013-06-18 Richard Biener <rguenther@suse.de>
7423 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
7424 * lto-streamer.h: Include pointer-set.h.
7425 (struct lto_decl_slot): Remove.
7426 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
7427 Remove next_index entry.
7428 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7429 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7430 (lto_init_tree_ref_encoder): Adjust.
7431 (lto_destroy_tree_ref_encoder): Likewise.
7432 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7433 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7434 (lto_output_decl_index): Adjust.
7435 (lto_new_out_decl_state): Likewise.
7436 (lto_record_function_out_decl_state): Likewise.
7437 * lto-streamer-out.c (copy_function): Likewise.
7439 2013-06-18 Richard Biener <rguenther@suse.de>
7441 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
7442 * cgraphunit.c: Include cfgloop.h.
7443 (init_lowered_empty_function): Initialize the loop tree.
7444 (assemble_thunk): Insert new BBs into loops.
7446 2013-06-18 Richard Biener <rguenther@suse.de>
7448 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
7449 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
7450 the map from cache entry to cache index optional.
7451 (streamer_tree_cache_replace_tree): Adjust accordingly.
7452 (streamer_tree_cache_append): Likewise.
7453 (streamer_tree_cache_delete): Likewise.
7454 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
7455 streamer cache map from cache entry to cache index.
7456 * lto-streamer-out.c (create_output_block): Adjust.
7458 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
7460 * config/arm/arm.md (attribute "insn"): Move multiplication and
7461 division attributes to...
7462 (attribute "type"): ... here. Remove mult.
7463 (attribute "mul32"): New attribute.
7464 (attribute "mul64"): Add umaal.
7465 (*arm_mulsi3): Update attributes.
7466 (*arm_mulsi3_v6): Likewise.
7467 (*thumb_mulsi3): Likewise.
7468 (*thumb_mulsi3_v6): Likewise.
7469 (*mulsi3_compare0): Likewise.
7470 (*mulsi3_compare0_v6): Likewise.
7471 (*mulsi_compare0_scratch): Likewise.
7472 (*mulsi_compare0_scratch_v6): Likewise.
7473 (*mulsi3addsi): Likewise.
7474 (*mulsi3addsi_v6): Likewise.
7475 (*mulsi3addsi_compare0): Likewise.
7476 (*mulsi3addsi_compare0_v6): Likewise.
7477 (*mulsi3addsi_compare0_scratch): Likewise.
7478 (*mulsi3addsi_compare0_scratch_v6): Likewise.
7479 (*mulsi3subsi): Likewise.
7480 (*mulsidi3adddi): Likewise.
7481 (*mulsi3addsi_v6): Likewise.
7482 (*mulsidi3adddi_v6): Likewise.
7483 (*mulsidi3_nov6): Likewise.
7484 (*mulsidi3_v6): Likewise.
7485 (*umulsidi3_nov6): Likewise.
7486 (*umulsidi3_v6): Likewise.
7487 (*umulsidi3adddi): Likewise.
7488 (*umulsidi3adddi_v6): Likewise.
7489 (*smulsi3_highpart_nov6): Likewise.
7490 (*smulsi3_highpart_v6): Likewise.
7491 (*umulsi3_highpart_nov6): Likewise.
7492 (*umulsi3_highpart_v6): Likewise.
7493 (mulhisi3): Likewise.
7494 (*mulhisi3tb): Likewise.
7495 (*mulhisi3bt): Likewise.
7496 (*mulhisi3tt): Likewise.
7497 (maddhisi4): Likewise.
7498 (*maddhisi4tb): Likewise.
7499 (*maddhisi4tt): Likewise.
7500 (maddhidi4): Likewise.
7501 (*maddhidi4tb): Likewise.
7502 (*maddhidi4tt): Likewise.
7504 (udivsi3): Likewise.
7505 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
7506 (thumb2_mulsi_short_compare0): Likewise.
7507 (thumb2_mulsi_short_compare0_scratch): Likewise.
7508 * config/arm/arm1020e.md (1020mult1): Update attribute change.
7509 (1020mult2): Likewise.
7510 (1020mult3): Likewise.
7511 (1020mult4): Likewise.
7512 (1020mult5): Likewise.
7513 (1020mult6): Likewise.
7514 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
7516 (cortex_a15_mult64): Likewise.
7517 (cortex_a15_sdiv): Likewise.
7518 (cortex_a15_udiv): Likewise.
7519 * config/arm/arm1026ejs.md (mult1): Update attribute change.
7525 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
7526 (pj4_ir_div): Likewise.
7527 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
7528 (11_mult2): Likewise.
7529 (11_mult3): Likewise.
7530 (11_mult4): Likewise.
7531 (11_mult5): Likewise.
7532 (11_mult6): Likewise.
7533 (11_mult7): Likewise.
7534 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
7535 (cortex_a8_mla): Likewise.
7536 (cortex_a8_mull): Likewise.
7537 (cortex_a8_smulwy): Likewise.
7538 (cortex_a8_smlald): Likewise.
7539 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
7540 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
7541 (cortex_r4_mul_3): Likewise.
7542 (cortex_r4_mla_4): Likewise.
7543 (cortex_r4_mla_3): Likewise.
7544 (cortex_r4_smlald): Likewise.
7545 (cortex_r4_mull): Likewise.
7546 (cortex_r4_sdiv): Likewise.
7547 (cortex_r4_udiv): Likewise.
7548 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
7549 (cortex_a7_idiv): Likewise.
7550 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
7551 (9_mult2): Likewise.
7552 (9_mult3): Likewise.
7553 (9_mult4): Likewise.
7554 (9_mult5): Likewise.
7555 (9_mult6): Likewise.
7556 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
7557 (cortex_a53_sdiv): Likewise.
7558 (cortex_a53_udiv): Likewise.
7559 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
7560 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
7561 (mp626_mult2): Likewise.
7562 (mp626_mult3): Likewise.
7563 (mp626_mult4): Likewise.
7564 * config/arm/fa526.md (526_mult1): Update attribute change.
7565 (526_mult2): Likewise.
7566 * config/arm/arm-generic.md (mult): Update attribute change.
7567 (mult_ldsched_strongarm): Likewise.
7568 (mult_ldsched): Likewise.
7569 (multi_cycle): Likewise.
7570 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
7571 * config/arm/fa606te.md (606te_mult1): Update attribute change.
7572 (606te_mult2): Likewise.
7573 (606te_mult3): Likewise.
7574 (606te_mult4): Likewise.
7575 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
7576 (cortex_a9_mac16): Likewise.
7577 (cortex_a9_multiply): Likewise.
7578 (cortex_a9_mac): Likewise.
7579 (cortex_a9_multiply_long): Likewise.
7580 * config/arm/fa626te.md (626te_mult1): Update attribute change.
7581 (626te_mult2): Likewise.
7582 (626te_mult3): Likewise.
7583 (626te_mult4): Likewise.
7585 2013-06-18 Richard Biener <rguenther@suse.de>
7588 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
7590 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7593 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7594 with next_active_insn.
7596 2013-06-18 Alan Modra <amodra@gmail.com>
7598 * config/rs6000/rs6000.h (enum data_align): New.
7599 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
7600 (DATA_ABI_ALIGNMENT): Define.
7601 (CONSTANT_ALIGNMENT): Correct comment.
7602 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
7603 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
7605 2013-06-17 David Malcolm <dmalcolm@redhat.com>
7607 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
7608 ATTRIBUTE_UNUSED marking.
7610 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
7612 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
7613 alternative and update.
7614 (aarch64_dup_lanedi): Delete.
7615 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
7616 * config/aarch64/aarch64-simd-builtins.def: Update.
7618 2013-06-17 Richard Biener <rguenther@suse.de>
7620 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
7621 (lto_input_scc): Declare.
7622 (lto_input_tree_1): Likewise.
7623 (struct lto_stats_d): Add num_tree_bodies_output and
7624 num_pickle_refs_output.
7625 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
7626 (lto_read_tree_1): Split out from ...
7627 (lto_read_tree): ... this.
7628 (lto_input_scc): New function.
7629 (lto_input_tree_1): Split out from ...
7630 (lto_input_tree): ... this. Handle LTO_tree_scc.
7631 (lto_data_in_create): Create the streamer cache without hashes.
7632 * lto-streamer-out.c (create_output_block): Create the streamer
7633 cache with hashes when not doing WPA.
7634 (lto_write_tree_1): Split out from ...
7635 (lto_write_tree): ... this.
7636 (get_symbol_initial_value): New function.
7637 (lto_output_tree_1): Split out from ...
7638 (lto_output_tree): ... this. Write trees as series of SCCs
7639 using a DFS walk via DFS_write_tree.
7640 (struct sccs, struct scc_entry): New types.
7641 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
7642 (DFS_write_tree_body): New function.
7643 (DFS_write_tree): Likewise.
7644 (hash_tree): Likewise.
7645 (scc_entry_compare): Likewise.
7646 (hash_scc): Likewise.
7647 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
7648 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
7649 TREE_CHAIN as regular reference.
7650 (streamer_read_integer_cst): Remove.
7651 (streamer_get_pickled_tree): Adjust.
7652 * tree-streamer-out.c (streamer_write_chain): Disable streaming
7653 of DECL_EXTERNALs in BLOCK_VARS for now.
7654 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
7656 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
7657 Add hash value argument and record that if hashes are recorded
7659 (streamer_tree_cache_insert_1): Adjust.
7660 (streamer_tree_cache_insert): Likewise.
7661 (streamer_tree_cache_insert_at): Rename to ...
7662 (streamer_tree_cache_replace_tree): ... this and adjust.
7663 (streamer_tree_cache_append): Adjust.
7664 (record_common_node): Likewise.
7665 (streamer_tree_cache_create): Add argument whether to
7666 record hash values together with trees.
7667 (streamer_tree_cache_delete): Adjust.
7668 * tree-streamer.h (struct streamer_tree_cache_d): Add
7670 (streamer_read_integer_cst): Remove.
7671 (streamer_tree_cache_insert): Adjust.
7672 (streamer_tree_cache_append): Likewise.
7673 (streamer_tree_cache_insert_at): Rename to ...
7674 (streamer_tree_cache_replace_tree): ... this and adjust.
7675 (streamer_tree_cache_create): Add argument whether to record hashes.
7676 (streamer_tree_cache_get): Rename to ...
7677 (streamer_tree_cache_get_tree): ... this.
7678 (streamer_tree_cache_get_hash): New function.
7679 * tree.c (cache_integer_cst): New function.
7680 * tree.h (cache_integer_cst): Declare.
7681 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
7682 * lto-symtab.c (lto_varpool_replace_node): Only release
7683 DECL_INITIAL of non-prevailing decls.
7684 * varpool.c (varpool_remove_initializer): Do not release
7685 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
7687 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
7689 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
7690 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
7691 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
7692 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
7693 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
7694 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
7695 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
7696 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
7697 instead of TARGET_64BIT.
7698 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
7699 Require ISA_HAS_<D>DIV.
7701 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
7703 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
7704 (mips*-*-linux*): Move default with_llsc setting to where other
7706 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
7708 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
7709 Likewise. Remove default with_tune setting. Move default float
7710 setting to its own block. Handle with_llsc in the same block as above.
7712 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
7714 PR rtl-optimization/57425
7715 PR rtl-optimization/57569
7716 * alias.c (write_dependence_p): Add new parameters mem_mode,
7717 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
7718 Changed all callers.
7719 (canon_anti_dependence): New function.
7720 * cse.c (check_dependence): Use canon_anti_dependence.
7721 * cselib.c (cselib_invalidate_mem): Likewise.
7722 * rtl.h (canon_anti_dependence): Declare.
7724 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
7726 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
7727 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
7728 ".set mips3" for 64-bit targets.
7730 2013-06-15 Dehao Chen <dehao@google.com>
7732 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
7733 * gimple-low.c (gimple_check_call_matching_types): Likewise.
7734 (gimple_check_call_args): Likewise.
7735 * value-prof.c (check_ic_target): Likewise.
7736 * ipa-inline.c (early_inliner): Likewise.
7737 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
7738 * cgraph.c (cgraph_create_edge_1): Likewise.
7739 (cgraph_make_edge_direct): Likewise.
7741 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
7744 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
7745 rs6000_output_move_128bit to handle emitting quad memory
7746 operations. Set attribute length to 8 bytes.
7748 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
7750 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
7752 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
7753 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
7754 (aarch64_<su>mlsl<mode>): Likewise.
7756 2013-06-14 Mike Stump <mikestump@comcast.net>
7758 * Makefile.in (TARGET_H): Add insn-codes.h.
7760 2013-06-14 Alan Modra <amodra@gmail.com>
7764 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
7765 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
7766 bitfield expansion when EXPAND_MEMORY.
7767 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
7769 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
7771 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
7772 test for clearing quad memory on 32-bit later.
7774 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
7776 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
7777 (fold_negate_expr): Likewise.
7778 (fold_real_zero_addition_p): Handle vectors.
7779 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
7781 2013-06-14 Alan Modra <amodra@gmail.com>
7783 * varasm.c (force_const_mem): Revert 2013-06-07 change.
7785 2013-06-13 Jan Hubicka <jh@suse.cz>
7787 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
7788 Local comdats are not externally visible.
7789 * symtab.c (dump_symtab_base): Dump externally visible.
7790 (verify_symtab_base): Verify back links in the symtab hash.
7792 2013-06-13 Bin Cheng <bin.cheng@arm.com>
7794 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
7795 CONVERT_EXPR as equal nodes.
7797 2013-06-13 Bin Cheng <bin.cheng@arm.com>
7799 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
7801 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
7803 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
7804 Generalize to complex and vector.
7805 * tree.c (build_all_ones_cst): New function.
7806 * tree.h (build_all_ones_cst): Declare it.
7808 2013-06-13 Alan Modra <amodra@gmail.com>
7810 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
7811 * config/rs6000/rs6000.md (signbittf2): New insn.
7812 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
7813 (abstf2_internal, cmptf_internal2): Likewise.
7814 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
7816 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7817 Pat Haugen <pthaugen@us.ibm.com>
7818 Peter Bergner <bergner@vnet.ibm.com>
7820 * config/rs6000/rs6000.c (emit_load_locked): Add support for
7821 power8 byte, half-word, and quad-word atomic instructions.
7822 (emit_store_conditional): Likewise.
7823 (rs6000_expand_atomic_compare_and_swap): Likewise.
7824 (rs6000_expand_atomic_op): Likewise.
7826 * config/rs6000/sync.md (larx): Add new modes for power8.
7828 (AINT): New mode iterator to include TImode as well as normal
7829 integer modes on power8.
7830 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
7831 that VSX registers are not considered. Use AINT mode iterator
7832 instead of INT1 to allow inclusion of quad word atomic operations
7834 (load_locked<mode>): Likewise.
7835 (store_conditional<mode>): Likewise.
7836 (atomic_compare_and_swap<mode>): Likewise.
7837 (atomic_exchange<mode>): Likewise.
7838 (atomic_nand<mode>): Likewise.
7839 (atomic_fetch_<fetchop_name><mode>): Likewise.
7840 (atomic_nand_fetch<mode>): Likewise.
7841 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
7843 (ATOMIC): On power8, add QImode, HImode modes.
7844 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
7845 modes that promote to SImode.
7846 (load_lockedti): Convert TImode arguments to PTImode, so that we
7847 get a guaranteed even/odd register pair.
7848 (load_lockedpti): Likewise.
7849 (store_conditionalti): Likewise.
7850 (store_conditionalpti): Likewise.
7852 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
7853 atomic load/store instructions.
7856 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
7858 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
7860 (insn_count): New attribute, with most cases extracted from...
7861 (length): ...here. Redefine most cases in terms of insn_count.
7862 (single_insn): Delete.
7863 (can_delay): Use insn_count to check for single instructions.
7864 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
7865 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
7866 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
7867 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
7868 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
7869 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
7870 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
7871 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
7872 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
7873 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
7874 rather than "length".
7875 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
7876 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
7877 Use "insn_count" rather than "length".
7878 * config/mips/mips-dsp.md
7879 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
7880 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
7883 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
7885 PR tree-optimization/57361
7886 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
7888 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
7890 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
7892 (aarch64_simd_combine<mode>): New instruction expansion.
7893 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
7895 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
7896 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
7898 2013-06-12 Jan Hubicka <jh@suse.cz>
7900 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
7901 decl has when in streaming stage.
7902 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
7903 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
7905 2013-06-12 Roland Stigge <stigge@antcom.de>
7908 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
7910 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7912 PR tree-optimization/57537
7913 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
7914 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
7916 2013-06-12 Richard Biener <rguenther@suse.de>
7918 * data-streamer.h (streamer_write_char_stream): CSE
7919 obs->current_pointer.
7920 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
7921 streamer_write_char_stream manually and optimize the resulting loop.
7922 (streamer_write_hwi_stream): Likewise.
7924 2013-06-12 Jan Hubicka <jh@suse.cz>
7926 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
7927 * cgraph.h (varpool_create_empty_node): Declare.
7928 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
7930 * symtab.c (symtab_unregister_node): Be lax about missin entries
7932 (symtab_get_node): Update comment.
7933 * varpool.c (varpool_create_empty_node): Break out from ...
7934 (varpool_node_for_decl): ... here.
7935 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
7937 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
7939 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
7940 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
7941 part. Use straight-line flow at the end.
7942 <COMPONENT_REF>: Remove superfluous else.
7943 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
7945 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7948 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
7949 target hook even for !TREE_PUBLIC decls. If no resolution info
7950 is available, return false for common and external decls.
7952 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7954 * config/rl78/constraints.md (U): New constraint.
7955 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
7958 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7961 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
7962 to allow returning address to AT_PLATFORM name.
7964 2013-06-11 Jan Hubicka <jh@suse.cz>
7966 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
7967 * cgraph.h (symtab_node_base): Add weakref flag.
7968 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
7969 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
7970 (output_weakrefs): Use weakref flag.
7971 * fold-const.c (simple_operand_p): Handle WEAK.
7972 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
7973 * ipa.c (varpool_externally_visible_p): Drop weakref.
7974 (function_and_variable_visibility): Update comment; fix weakref
7975 sanity checks; do not clear DECL_WEAK on them.
7976 * lto-cgraph.c (lto_output_node): update.
7977 (lto_output_varpool_node): Update.
7978 (input_overwrite_node): Update.
7979 (input_node): Update.
7980 (input_varpool_node): Update.
7981 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
7982 (lto_symtab_merge_symbols): Add sanity check.
7983 (lto_symtab_prevailing_decl): Do not special case weakrefs.
7984 * passes.c (rest_of_decl_compilation): Set static flag, too.
7985 * symtab.c (dump_symtab_base): Dump weakref.
7986 (verify_symtab_base): Sanity check weakrefs.
7987 (symtab_make_decl_local): Remove duplicated code.
7988 (symtab_alias_ultimate_target): Simplify.
7989 * varpool.c (varpool_create_variable_alias): Set weakref flag.
7991 2013-06-11 DJ Delorie <dj@redhat.com>
7993 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
7994 (rl78_unwind_word_mode): New.
7996 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7998 * final.c (debug_prefix_maps): Make static.
8000 2013-06-11 David Malcolm <dmalcolm@redhat.com>
8002 * function.c (initial_trampoline): Remove stray copy.
8004 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
8006 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
8008 2013-06-11 Martin Jambor <mjambor@suse.cz>
8010 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
8011 within bounds at the beginning of the function.
8013 2013-06-11 Alan Modra <amodra@gmail.com>
8015 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
8017 (get_named_section): Don't NULL !DECL_P decl.
8019 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
8021 * doc/invoke.texi (core-avx2): Document.
8023 (atom): Updated with MOVBE.
8025 2013-06-11 Richard Biener <rguenther@suse.de>
8027 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
8029 2013-06-11 Anton Blanchard <anton@samba.org>
8031 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
8032 correct shift value in little-endian mode.
8034 2013-06-11 Jakub Jelinek <jakub@redhat.com>
8037 * varasm.c (get_variable_align): Move #endif to the right place.
8039 2013-06-10 Cary Coutant <ccoutant@google.com>
8041 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
8042 for hash so that hash table traversal order is deterministic.
8044 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
8045 Pat Haugen <pthaugen@us.ibm.com>
8046 Peter Bergner <bergner@vnet.ibm.com>
8048 * config/rs6000/vector.md (GPR move splitter): Do not split moves
8049 of vectors in GPRS if they are direct moves or quad word load or
8052 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
8054 (direct_move_p): Likewise.
8055 (quad_load_store_p): Likewise.
8057 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
8058 classes into bins based on the physical register type.
8059 (reg_class_to_reg_type): Likewise.
8060 (IS_STD_REG_TYPE): Likewise.
8061 (IS_FP_VECT_REG_TYPE): Likewise.
8062 (reload_fpr_gpr): Arrays to determine what insn to use if we can
8063 use direct move instructions.
8064 (reload_gpr_vsx): Likewise.
8065 (reload_vsx_gpr): Likewise.
8066 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
8067 information that is a simplification of register classes. Also
8068 precalculate direct move reload helpers.
8069 (direct_move_p): New function to return true if the operation can
8070 be done as a direct move instruciton.
8071 (quad_load_store_p): New function to return true if the operation
8072 is a quad memory operation.
8073 (rs6000_legitimize_address): If quad memory, only allow register
8074 indirect for TImode addresses.
8075 (rs6000_legitimate_address_p): Likewise.
8076 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
8077 (rs6000_reload_register_type): Likewise.
8078 (register_to_reg_type): Return register type.
8079 (rs6000_secondary_reload_simple_move): New helper function for
8080 secondary reload and secondary memory needed to identify anything
8081 that is a simple move, and does not need reloading.
8082 (rs6000_secondary_reload_direct_move): New helper function for
8083 secondary reload to identify cases that can be done with several
8084 instructions via the direct move instructions.
8085 (rs6000_secondary_reload_move): New helper function for secondary
8086 reload to identify moves between register types that can be done.
8087 (rs6000_secondary_reload): Add support for quad memory operations
8088 and for direct move.
8089 (rs6000_secondary_memory_needed): Likewise.
8090 (rs6000_debug_secondary_memory_needed): Change argument names.
8091 (rs6000_output_move_128bit): New function to return the move to
8092 use for 128-bit moves, including knowing about the various
8093 limitations of quad memory operations.
8095 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
8096 memory operations. call rs6000_output_move_128bit for the actual
8097 instruciton(s) to generate.
8098 (vsx_movti_64bit): Likewise.
8100 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
8101 (UNSPEC_P8V_MTVSRWZ): Likewise.
8102 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
8103 (UNSPEC_P8V_MTVSRD): Likewise.
8104 (UNSPEC_P8V_XXPERMDI): Likewise.
8105 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
8106 (UNSPEC_FUSION_GPR): Likewise.
8107 (FMOVE128_GPR): New iterator for direct move.
8108 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
8111 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
8112 loads and direct move instructions.
8113 (zero_extendsidi2_lfiwzx): Likewise.
8114 (extendsidi2_lfiwax): Likewise.
8115 (extendsidi2_nocell): Likewise.
8116 (floatsi<mode>2_lfiwax): Likewise.
8118 (floatunssi<mode>2_lfiwzx): Likewise.
8120 (fix_trunc<mode>_stfiwx): Likewise.
8121 (fixuns_trunc<mode>_stfiwx): Likewise.
8122 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
8123 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
8124 (parity<mode>2_cmpb): Set length/type attr.
8125 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
8126 for 'mr.' to fast_compare.
8127 (bpermd_<mode>): Change type attr to popcnt.
8128 (p8_fmrgow_<mode>): New insns for power8 direct move support.
8129 (p8_mtvsrwz_1): Likewise.
8130 (p8_mtvsrwz_2): Likewise.
8131 (reload_fpr_from_gpr<mode>): Likewise.
8132 (p8_mtvsrd_1): Likewise.
8133 (p8_mtvsrd_2): Likewise.
8134 (p8_xxpermdi_<mode>): Likewise.
8135 (reload_vsx_from_gpr<mode>): Likewise.
8136 (reload_vsx_from_gprsf): Likewise.
8137 (p8_mfvsrd_3_<mode>): LIkewise.
8138 (reload_gpr_from_vsx<mode>): Likewise.
8139 (reload_gpr_from_vsxsf): Likewise.
8140 (p8_mfvsrd_4_disf): Likewise.
8141 (multi-word GPR splits): Do not split direct moves or quad memory
8144 2013-06-10 David Malcolm <dmalcolm@redhat.com>
8146 * tree-into-ssa.c (interesting_blocks): Make static.
8148 2013-06-10 Jakub Jelinek <jakub@redhat.com>
8151 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
8152 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
8153 Use DATA_ABI_ALIGNMENT for that case instead if defined.
8154 (get_variable_align): New function.
8155 (get_variable_section, emit_bss, emit_common,
8156 assemble_variable_contents, place_block_symbol): Use
8157 get_variable_align instead of DECL_ALIGN.
8158 (assemble_noswitch_variable): Add align argument, use it
8159 instead of DECL_ALIGN.
8160 (assemble_variable): Adjust caller. Use get_variable_align
8161 instead of DECL_ALIGN.
8162 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
8164 (DATA_ABI_ALIGNMENT): Define.
8165 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
8166 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
8167 opt is false, only return the psABI mandated alignment increase.
8168 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
8169 (DATA_ABI_ALIGNMENT): ... this.
8170 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
8171 (DATA_ABI_ALIGNMENT): ... this.
8172 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
8173 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
8174 (DATA_ABI_ALIGNMENT): ... this.
8175 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
8176 * doc/tm.texi: Regenerated.
8178 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
8180 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
8181 cmp_code to construct REG_EQUAL note.
8183 2013-06-09 Jakub Jelinek <jakub@redhat.com>
8186 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
8187 that operands[2] doesn't overlap with operands[0].
8189 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
8190 Jan Hubicka <jh@suse.cz>
8192 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
8193 hack to mark symbols as used.
8195 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
8197 PR rtl-optimization/57559
8198 * lra-constraints.c (process_alt_operands): Don't discourage
8199 memory with known offset for offsetable memory constraint.
8200 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
8202 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
8204 * varasm.c (struct oc_local_state): Reorder fields.
8205 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
8206 and adjust accordingly.
8207 (output_constructor): Reorder initialization code and adjust call to
8208 output_constructor_bitfield.
8210 2013-06-07 Jan Hubicka <jh@suse.cz>
8212 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
8214 2013-06-07 David Malcolm <dmalcolm@redhat.com>
8216 * tree-object-size.c (unknown): Make const.
8218 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8220 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
8221 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
8222 for last alternative in the cpu_facility attribute.
8224 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8227 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
8228 (xordi3): Change operand 2 constraint to arm_xordi_operand.
8229 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
8230 * config/arm/constraints.md (Dg): New constraint.
8231 * config/arm/neon.md (xordi3_neon): Remove.
8232 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
8233 * config/arm/predicates.md (arm_xordi_operand): New predicate.
8235 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8237 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
8238 Clean up alternatives.
8240 2013-06-07 Alan Modra <amodra@gmail.com>
8242 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
8245 2013-06-07 Alan Modra <amodra@gmail.com>
8247 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
8249 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8251 * config/arm/constraints.md (Df): New constraint.
8252 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
8253 Correct length attribute for last two alternatives.
8255 2013-06-07 Alan Modra <amodra@gmail.com>
8257 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8258 override user -mfp-in-toc.
8259 (offsettable_ok_by_alignment): Consider just the current access
8260 rather than the whole object, unless BLKmode. Handle
8261 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
8262 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
8263 for -mcmodel=medium.
8264 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
8265 override user -mfp-in-toc or -msum-in-toc. Default to
8266 -mno-fp-in-toc for -mcmodel=medium.
8268 2013-06-06 DJ Delorie <dj@redhat.com>
8270 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
8271 TARGET_VALID_POINTER_MODE.
8273 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
8274 Pat Haugen <pthaugen@us.ibm.com>
8275 Peter Bergner <bergner@vnet.ibm.com>
8277 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8278 Document new power8 builtins.
8280 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
8281 condition code register, to allow 128-bit logical operations to be
8282 done in the VSX or GPR registers.
8283 (nor<mode>3): Use the canonical form for nor.
8284 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
8285 vclz*, and vpopcnt* vector instructions.
8286 (nand<mode>3): Likewise.
8287 (orc<mode>3): Likewise.
8288 (clz<mode>2): LIkewise.
8289 (popcount<mode>2): Likewise.
8291 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
8292 that only the GPRs are recognized.
8294 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8295 support for new power8 builtins.
8297 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
8299 (xscvdpspn): Likewise.
8305 (vpopcnt): Likewise.
8306 (vpopcntb): Likewise.
8307 (vpopcnth): Likewise.
8308 (vpopcntw): Likewise.
8309 (vpopcntd): Likewise.
8314 (eqv_v16qi3): Likewise.
8315 (eqv_v8hi3): Likewise.
8316 (eqv_v4si3): Likewise.
8317 (eqv_v2di3): Likewise.
8318 (eqv_v4sf3): Likewise.
8319 (eqv_v2df3): Likewise.
8321 (nand_v16qi3): Likewise.
8322 (nand_v8hi3): Likewise.
8323 (nand_v4si3): Likewise.
8324 (nand_v2di3): Likewise.
8325 (nand_v4sf3): Likewise.
8326 (nand_v2df3): Likewise.
8328 (orc_v16qi3): Likewise.
8329 (orc_v8hi3): Likewise.
8330 (orc_v4si3): Likewise.
8331 (orc_v2di3): Likewise.
8332 (orc_v4sf3): Likewise.
8333 (orc_v2df3): Likewise.
8335 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
8336 allow power8 quad mode in 64-bit.
8337 (rs6000_builtin_vectorized_function): Add support to vectorize
8338 ISA 2.07 count leading zeros, population count builtins.
8339 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
8340 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
8341 (builtin_function_type): Add vgbbd builtin function which takes an
8343 (altivec_expand_vec_perm_const): Add support for new power8 merge
8346 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
8347 that does not include TImdoe for use with 32-bit.
8348 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
8350 (UNSPEC_VSX_CVDPSPN): Likewise.
8351 (vsx_xscvdpspn): Likewise.
8352 (vsx_xscvspdpn): Likewise.
8353 (vsx_xscvdpspn_scalar): Likewise.
8354 (vsx_xscvspdpn_directmove): Likewise.
8355 (vsx_and<mode>3): Split logical operations into 32-bit and
8356 64-bit. Add support to do logical operations on TImode as well as
8357 VSX vector types. Allow logical operations to be done in either
8358 VSX registers or in general purpose registers in 64-bit mode. Add
8359 splitters if GPRs were used. For AND, add clobber of CCmode to
8360 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
8362 (vsx_and<mode>3_32bit): Likewise.
8363 (vsx_and<mode>3_64bit): Likewise.
8364 (vsx_ior<mode>3): Likewise.
8365 (vsx_ior<mode>3_32bit): Likewise.
8366 (vsx_ior<mode>3_64bit): Likewise.
8367 (vsx_xor<mode>3): Likewise.
8368 (vsx_xor<mode>3_32bit): Likewise.
8369 (vsx_xor<mode>3_64bit): Likewise.
8370 (vsx_one_cmpl<mode>2): Likewise.
8371 (vsx_one_cmpl<mode>2_32bit): Likewise.
8372 (vsx_one_cmpl<mode>2_64bit): Likewise.
8373 (vsx_nor<mode>3): Likewise.
8374 (vsx_nor<mode>3_32bit): Likewise.
8375 (vsx_nor<mode>3_64bit): Likewise.
8376 (vsx_andc<mode>3): Likewise.
8377 (vsx_andc<mode>3_32bit): Likewise.
8378 (vsx_andc<mode>3_64bit): Likewise.
8379 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
8380 and xxlorc instructions.
8381 (vsx_eqv<mode>3_64bit): Likewise.
8382 (vsx_nand<mode>3_32bit): Likewise.
8383 (vsx_nand<mode>3_64bit): Likewise.
8384 (vsx_orc<mode>3_32bit): Likewise.
8385 (vsx_orc<mode>3_64bit): Likewise.
8387 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
8389 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
8391 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
8392 (p8_vmrgow): Likewise.
8393 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
8394 GPRs to be split under VSX.
8395 (p8v_clz<mode>2): Add power8 count leading zero support.
8396 (p8v_popcount<mode>2): Add power8 population count support.
8397 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
8400 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
8403 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
8405 (vec_nand): Likewise.
8406 (vec_vclz): Likewise.
8407 (vec_vclzb): Likewise.
8408 (vec_vclzd): Likewise.
8409 (vec_vclzh): Likewise.
8410 (vec_vclzw): Likewise.
8411 (vec_vgbbd): Likewise.
8412 (vec_vmrgew): Likewise.
8413 (vec_vmrgow): Likewise.
8414 (vec_vpopcnt): Likewise.
8415 (vec_vpopcntb): Likewise.
8416 (vec_vpopcntd): Likewise.
8417 (vec_vpopcnth): Likewise.
8418 (vec_vpopcntw): Likewise.
8420 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8422 PR rtl-optimization/57468
8423 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
8426 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8428 PR rtl-optimization/57459
8429 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
8430 type when setting live regs.
8432 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8434 * config/s390/s390.opt (mlra): New option.
8435 * config/s390/s390.c (s390_decompose_address): Check displacement
8436 for all registers for LRA.
8437 (s390_secondary_reload): Don't used secondary reloads for LRA.
8438 (s390_lra_p): New function.
8439 (TARGET_LRA_P): Define.
8440 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
8441 of attribute cpu_facility to zarch for the last alternative.
8442 (*cmpmem_short): Ditto.
8444 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
8446 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
8447 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
8448 (arm_expand_prologue): Likewise.
8450 2013-06-06 Teresa Johnson <tejohnson@google.com>
8453 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
8454 as this is now done by redirect_edge_and_branch_force.
8455 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
8456 barriers, and fix interaction with splitting.
8457 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
8458 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
8459 reflect changes made in the routine.
8460 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
8461 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
8462 since this is called in cfglayout mode, and replace partition fixup
8463 with assert as that is now done by force_nonfallthru_and_redirect.
8464 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
8465 already be marked with region crossing note.
8466 (insert_section_boundary_note): Make non-static, gate on flag
8467 has_bb_partition, rewrite to also check for multiple partitions.
8468 (rest_of_handle_reorder_blocks): Remove call to
8469 insert_section_boundary_note, now done later during free_cfg.
8470 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
8471 * bb-reorder.h (insert_section_boundary_note): Declare.
8472 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
8473 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
8474 invoke insert_section_boundary_note.
8475 (try_redirect_by_replacing_jump): Remove unnecessary
8476 check for region crossing note.
8477 (fixup_partition_crossing): New function.
8478 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
8479 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
8480 in non-cfglayout mode.
8481 (force_nonfallthru_and_redirect): Fixup partition boundaries,
8482 remove old code that tried to do this. Emit barrier correctly
8483 when we are in cfglayout mode.
8484 (last_bb_in_partition): New function.
8485 (rtl_split_edge): Correctly fixup partition boundaries.
8486 (commit_one_edge_insertion): Remove old code that tried to
8487 fixup region crossing edge since this is now handled in
8488 split_block, and set up insertion point correctly since
8489 block may now end in a jump.
8490 (verify_hot_cold_block_grouping): Guard against checking when not in
8491 linearized RTL mode.
8492 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
8494 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
8495 rtl_verify_flow_info, so not called in cfglayout mode.
8496 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
8497 (fixup_reorder_chain): Remove old code that attempted to fixup region
8498 crossing note as this is now handled in force_nonfallthru_and_redirect.
8499 (duplicate_insn_chain): Don't duplicate switch section notes.
8500 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
8502 * basic-block.h (emit_barrier_after_bb): Declare.
8504 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8506 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
8507 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
8508 arm_usatsihi): Adjust alternatives for arm_restrict_it.
8510 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8512 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
8514 * config/arm/ldmstm.md: Regenerate.
8516 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8518 * config/arm/sync.md (atomic_loaddi_1):
8519 Disable predication for arm_restrict_it.
8520 (arm_load_exclusive<mode>): Likewise.
8521 (arm_load_exclusivesi): Likewise.
8522 (arm_load_exclusivedi): Likewise.
8523 (arm_load_acquire_exclusive<mode>): Likewise.
8524 (arm_load_acquire_exclusivesi): Likewise.
8525 (arm_load_acquire_exclusivedi): Likewise.
8526 (arm_store_exclusive<mode>): Likewise.
8527 (arm_store_exclusive<mode>): Likewise.
8528 (arm_store_release_exclusivedi): Likewise.
8529 (arm_store_release_exclusive<mode>): Likewise.
8531 2013-06-06 Richard Biener <rguenther@suse.de>
8533 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
8535 (lto_tag_is_tree_code_p): Adjust.
8536 (lto_tag_is_gimple_code_p): Likewise.
8537 (lto_gimple_code_to_tag): Likewise.
8538 (lto_tag_to_gimple_code): Likewise.
8539 (lto_tree_code_to_tag): Likewise.
8540 (lto_tag_to_tree_code): Likewise.
8541 * data-streamer.h (streamer_write_hwi_in_range): Use
8542 uhwi streaming to stream the normalized range.
8543 (streamer_read_hwi_in_range): Likewise.
8545 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8547 * config/arm/arm.md (enabled_for_depr_it): New attribute.
8548 (predicable_short_it): Likewise.
8549 (predicated): Likewise.
8550 (enabled): Handle above.
8551 (define_cond_exec): Set predicated attribute to yes.
8553 2013-06-05 Mike Stump <mikestump@comcast.net>
8555 * gdbinit.in (__FUNCTION__): Add.
8557 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
8559 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
8560 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
8562 2013-06-05 Jan Hubicka <jh@suse.cz>
8564 * varasm.c (mark_decl_referenced): Revert the removal until targets
8567 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
8569 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
8570 instead of mark_decl_referenced.
8572 2013-06-05 Jan Hubicka <jh@suse.cz>
8574 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
8575 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
8576 and symtab_used_from_object_file_p.
8577 (cgraph_make_node_local_1): Clear forced_by_abi.
8578 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
8579 * cgraph.h (symtab_node_base): Add forced_by_abi.
8580 (decide_is_variable_needed): Remove.
8581 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
8582 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
8583 (decide_is_symbol_needed): ... this one; handle symbols in general;
8584 always analyze virtuals; honnor forced_by_abi.
8585 (cgraph_finalize_function): Update.
8586 (varpool_finalize_decl): Update.
8587 (symbol_defined_and_needed): Remove.
8588 (analyze_functions): Update.
8589 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8590 output_refs, input_overwrite_node): Handle forced_by_abi.
8591 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
8592 (address_taken_from_non_vtable_p): ... this one.
8593 (comdat_can_be_unshared_p_1): New function.
8594 (cgraph_comdat_can_be_unshared_p): Rename to ...
8595 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
8596 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
8597 (function_and_variable_visibility): Clear forced_by_abi as needed.
8598 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
8599 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
8600 * symtab.c (dump_symtab_base): Dump forced_by_abi.
8601 * varpool.c (decide_is_variable_needed): Remove.
8603 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8605 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
8606 (arm_option_override): Override arm_restrict_it where appropriate.
8607 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
8608 * config/arm/arm.opt (mrestrict-it): New command-line option.
8609 * doc/invoke.texi: Document -mrestrict-it.
8611 2013-06-05 David Malcolm <dmalcolm@redhat.com>
8613 * tsan.c (tsan_atomic_table): Make const.
8615 2013-06-05 Richard Biener <rguenther@suse.de>
8617 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
8618 index associated with the tree we are supposed to replace.
8619 * tree-streamer-out.c (pack_ts_base_value_fields): Output
8620 TREE_ASM_WRITTEN as zero for everything but SSA names.
8622 2013-06-05 David Malcolm <dmalcolm@redhat.com>
8624 * tree-ssa-structalias.c (call_stmt_vars): Make static.
8626 2013-06-04 Jan Hubicka <jh@suse.cz>
8628 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
8629 (input_node, input_varpool_node): Handle correctly external same
8631 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
8632 nodes at ltrans stage.
8634 2013-06-04 Jan Hubicka <jh@suse.cz>
8636 * ipa-inline.c (update_caller_keys): Fix availability test.
8637 (update_callee_keys): Likewise.
8638 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
8639 to follow ELF standard.
8641 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
8643 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
8644 (mips64r5900el-*-elf*): New configurations.
8645 * config/mips/mips-cpus.def (r5900): New processor.
8646 * config/mips/mips-tables.opt: Regenerate.
8647 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
8648 (mips_issue_rate): Handle PROCESSOR_R5900.
8649 (mips_reorg_process_insns): Force reorder mode for the R5900.
8650 * config/mips/mips.h (TARGET_MIPS5900): Define.
8651 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
8653 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
8655 * config/mips/mips.md (processor): Add r5900.
8656 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
8658 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8660 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
8661 into function to generate MOVI instruction.
8662 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
8663 (aarch64_preferred_simd_mode): Turn into wrapper.
8664 (aarch64_output_scalar_simd_mov_immediate): New function.
8665 * config/aarch64/aarch64-protos.h: Add prototype for above.
8667 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8669 * config/aarch64/aarch64.c (simd_immediate_info): Remove
8670 element_char member.
8671 (sizetochar): Return signed char.
8672 (aarch64_simd_valid_immediate): Remove elchar and other
8673 unnecessary variables.
8674 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
8675 Calculate element_char as required.
8676 * config/aarch64/aarch64-protos.h: Update and move prototype
8677 for aarch64_output_simd_mov_immediate.
8678 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
8681 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8683 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
8684 information completed by aarch64_simd_valid_immediate.
8685 (aarch64_legitimate_constant_p): Update arguments.
8686 (aarch64_simd_valid_immediate): Work with struct rather than many
8688 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
8689 (aarch64_simd_make_constant): Update arguments.
8690 (aarch64_output_simd_mov_immediate): Work with struct rather than
8691 many pointers. Output immediate directly rather than as operand.
8692 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
8694 * config/aarch64/constraints.md (Dn): Update arguments.
8696 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8698 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
8700 (aarch64_simd_immediate_valid_for_move): Remove.
8701 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
8702 (aarch64_simd_make_constant): Update call.
8703 (aarch64_output_simd_mov_immediate): Update call.
8704 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
8706 * config/aarch64/constraints.md (Dn): Update call.
8708 2013-06-04 Ian Bolton <ian.bolton@arm.com>
8710 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
8711 return type to bool for prototype.
8712 (aarch64_legitimate_constant_p): Check for true instead of not -1.
8713 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
8714 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
8716 2013-06-04 Catherine Moore <clm@codesourcery.com>
8718 * config/mips/mips.opt (meva): New.
8719 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
8720 (ASM_SPEC): Handle -meva.
8721 * doc/invoke.texi (meva): Document.
8723 2013-06-04 Alan Modra <amodra@gmail.com>
8725 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
8728 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8730 * rtl.def: Add extra fourth optional field to define_cond_exec.
8731 * gensupport.c (process_one_cond_exec): Process attributes from
8733 * doc/md.texi: Document fourth field in define_cond_exec.
8735 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
8737 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
8738 out the processing order as in store_bit_field_1.
8740 2013-06-04 Jan Hubicka <jh@suse.cz>
8743 * cgraphunit.c (cgraph_process_same_body_aliases): Create
8744 non-VAR_DECL node if it does not exist yet.
8746 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
8748 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
8749 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
8750 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
8751 target_cpu_default setting.
8753 2013-06-03 Teresa Johnson <tejohnson@google.com>
8755 * dumpfile.c (opt_info_switch_p): Change -fopt-info
8756 default to -fopt-info=optimized instead of all.
8757 * doc/invoke.texi: Ditto.
8758 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
8759 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
8760 (execute_vect_slp): Emit BB vectorization success under
8761 MSG_OPTIMIZED_LOCATIONS.
8762 * tree-vect-slp.c (vect_slp_transform_bb): Change
8763 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
8764 * tree-vect-loop.c (vect_transform_loop): Ditto.
8766 2013-06-03 Jason Merrill <jason@redhat.com>
8769 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8770 Use TARGET_EXPR for C++.
8772 2013-06-03 Jakub Jelinek <jakub@redhat.com>
8774 PR rtl-optimization/57268
8775 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
8776 if DEBUG_INSN_P (insn).
8779 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8781 PR rtl-optimization/57268
8782 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8783 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8785 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8787 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
8788 (ix86_avoid_lea_for_addr): Likewise.
8789 (exact_dependency_1): Likewise.
8790 (ix86_adjust_cost): Likewise.
8791 (swap_top_of_ready_list): Fix formatting and !reload_completed check
8793 (do_reorder_for_imul): Fix typo, formatting and
8794 !reload_completed check removed.
8795 (ix86_sched_reorder): Fix typo and formatting.
8796 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
8799 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
8801 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
8803 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
8805 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
8806 <STRING_CST>: Likewise.
8807 <VECTOR_CST>: Likewise.
8809 2013-06-01 Janus Weil <janus@gcc.gnu.org>
8810 Mikael Morin <mikael@gcc.gnu.org>
8812 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
8813 * config.in: Regenerated.
8814 * configure: Regenerated.
8816 2013-06-01 Jan Hubicka <jh@suse.cz>
8819 * cgraphunit.c (compile): When weakref is not supported,
8820 set up transparent aliases before final output pass.
8821 * varasm.c (assemble_alias): Do not try to do it here.
8823 2013-06-01 Jan Hubicka <jh@suse.cz>
8826 * passes.c (for_per_function): Skip unanalyzed functions.
8828 2013-06-01 Jan Hubicka <jh@suse.cz>
8830 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
8831 (lto_symtab_merge_symbols_1): ... this one.
8832 (lto_symtab_merge_cgraph_nodes): Rename to ...
8833 (lto_symtab_merge_symbols): ... this one; simplify.
8834 * cgraph.c (same_body_aliases_done): Rename to ...
8835 (cpp_implicit_aliases_done): ... this one.
8836 (cgraph_create_function_alias): Update.
8837 (cgraph_same_body_alias): Update.
8838 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
8839 (verify_edge_corresponds_to_fndecl): Simplify.
8840 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
8841 (cgraph_node): Remove same_body_alias.
8842 (varpool_node): Remove alias_of and extra_name_alias.
8843 (same_body_aliases_done): Rename to ..
8844 (cpp_implicit_aliases_done): ... this one.
8845 (symtab_alias_ultimate_target): Add default parameter.
8846 (symtab_resolve_alias): New function.
8847 (fixup_same_cpp_alias_visibility): Declare.
8848 (cgraph_function_node): Add default parameter.
8849 (cgraph_node_asm_name): Likewise.
8850 (cgraph_function_or_thunk_node): Add default parameter; do
8851 not ICE when it is NULL.
8852 (varpool_variable_node): Likewise.
8853 * tree-emutls.c (create_emultls_var): Update.
8854 (ipa_lower_emutls): Update.
8855 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
8856 (cgraph_reset_node): Reset alias info.
8857 (cgraph_finalize_function): Update.
8858 (fixup_same_cpp_alias_visibility): Move to symtab.c.
8859 (analyze_function): Simplify.
8860 (cgraph_process_same_body_aliases): Simplify.
8861 (analyze_functions): Fixup same body aliases.
8862 (handle_alias_pairs): Simplify.
8863 (assemble_thunk): Update.
8864 (assemble_thunks_and_aliases): Update.
8865 (output_weakrefs): Rewrite.
8866 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
8867 (lto_output_varpool_node): Likewise.
8868 (compute_ltrans_boundary): Remve assert.
8869 (get_alias_symbol): New functoin.
8870 (input_node): Rewrite alias handling.
8871 (input_varpool_node): Likewise.
8872 * ipa-pure-const.c (propagate_pure_const): Fix formating.
8873 * ipa.c (process_references): Handle weakrefs correctly.
8874 (symtab_remove_unreachable_nodes): Likewise.
8875 * trans-mem.c (get_cg_data): Update.
8876 (ipa_tm_create_version_alias): Update.
8877 (ipa_tm_execute): Update.
8878 * symtab.c (dump_symtab_base): Dump aliases.
8879 (verify_symtab_base): Verify aliases.
8880 (symtab_node_availability): New function.
8881 (symtab_alias_ultimate_target): Simplify.
8882 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
8883 handle all the fixup cases.
8884 (symtab_resolve_alias): New function.
8885 * passes.c (ipa_write_summaries): Handle weakrefs.
8886 * varpool.c (varpool_analyze_node): Simplify.
8887 (assemble_aliases): Update.
8888 (varpool_create_variable_alias): Simplify.
8889 (varpool_extra_name_alias): Simplify.
8890 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
8891 (lto_symtab_merge_symbols): ... this one.
8893 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
8896 PR rtl-optimization/57268
8897 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8898 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8900 2013-06-01 Tobias Burnus <burnus@net-b.de>
8903 2013-05-31 Tobias Burnus <burnus@net-b.de>
8906 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
8909 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8911 PR rtl-optimization/57268
8912 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8913 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8915 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8917 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
8918 unordered comparison operators when -fno-trapping-math is in effect
8920 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
8921 and implement unordered comparison operators properly on the e500.
8923 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8925 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
8926 for constant scalar integers.
8927 (simplify_relational_operation_1): Likewise.
8929 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
8931 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
8932 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
8935 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8936 Igor Zamyatin <igor.zamyatin@intel.com>
8938 Silvermont (SLM) architecture performance tuning.
8939 * config/i386/i386.h (enum ix86_tune_indices): Add
8940 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
8941 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
8943 * config/i386/i386.c (initial_ix86_tune_features)
8944 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
8945 (ix86_lea_outperforms): Handle Silvermont tuning.
8946 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
8948 (ix86_use_lea_for_mov): Likewise.
8949 (ix86_avoid_lea_for_addr): Likewise.
8950 (ix86_lea_for_add_ok): Likewise.
8951 (exact_dependency_1): New function.
8952 (exact_store_load_dependency): Likewise.
8953 (ix86_adjust_cost): Handle Silvermont tuning.
8954 (do_reoder_for_imul): Likewise.
8955 (swap_top_of_ready_list): New function.
8956 (ix86_sched_reorder): Changed to handle Silvermont tuning.
8958 * config/i386/i386.md (peepholes that split memory operand in fp
8961 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8963 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
8964 Remove un-necessary braces.
8966 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8968 * config/aarch64/aarch64.c (aarch64_classify_symbol):
8969 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
8971 2013-05-31 Tobias Burnus <burnus@net-b.de>
8974 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
8976 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8979 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
8980 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
8981 * config/arm/neon.md (iordi3_neon): Remove.
8982 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
8983 * config/arm/predicates.md (imm_for_neon_logic_operand):
8984 Move to earlier in the file.
8985 (neon_logic_op2): Likewise.
8986 (arm_iordi_operand_neon): New predicate.
8988 2013-05-31 Richard Biener <rguenther@suse.de>
8990 PR tree-optimization/57478
8991 PR tree-optimization/57453
8992 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
8995 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8997 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
8998 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
9000 2013-05-30 Tobias Burnus <burnus@net-b.de>
9001 Thomas Koenig <tkoenig@gcc.gnu.org>
9004 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
9005 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
9007 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
9009 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
9011 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
9013 * target.def (register_usage_leveling_p): New hook.
9014 * targhooks.c (default_register_usage_leveling_p): New.
9015 * targhooks.h (default_register_usage_leveling_p): New prototype.
9016 * lra-assigns.c (register_usage_leveling_p): Use the hook.
9017 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
9018 * doc/tm.texi: Update.
9019 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
9021 2013-05-30 Ian Bolton <ian.bolton@arm.com>
9023 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
9024 (*insv_reg<mode>): New define_insn.
9026 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
9028 PR rtl-optimization/57439
9029 * postreload.c (move2add_valid_value_p): Check that we have
9030 a zero subreg_regno_offset when accessing the register in
9033 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9034 Igor Zamyatin <igor.zamyatin@intel.com>
9036 Silvermont (SLM) architecture pipeline model, tuning and
9038 * config.gcc: Add slm config options and target.
9040 * config/i386/slm.md: New.
9042 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
9044 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
9046 (ix86_target_macros_internal): Likewise.
9048 * gcc/config/i386/i386.c (slm_cost): New cost.
9049 (m_SLM): New macro flag.
9050 (initial_ix86_tune_features): Set m_SLM.
9051 (x86_accumulate_outgoing_args): Likewise.
9052 (x86_arch_always_fancy_math_387): Likewise.
9053 (processor_target_table): Add slm cost.
9054 (cpu_names): Add slm cpu name.
9055 (x86_option_override_internal): Set SLM ISA.
9056 (ix86_issue_rate): New case PROCESSOR_SLM.
9057 (ia32_multipass_dfa_lookahead): Likewise.
9058 (fold_builtin_cpu): Add slm.
9060 * config/i386/i386.h (TARGET_SLM): New target macro.
9061 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
9062 (processor_type): Add PROCESSOR_SLM.
9064 * config/i386/i386.md (cpu): Add new value "slm".
9065 (slm.md): Include slm.md.
9067 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
9068 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9070 * config/arm/arm-protos.h: Add and update function protos.
9071 * config/arm/arm.c (use_simple_return_p): New added.
9072 (thumb2_expand_return): Check simple_return flag.
9073 * config/arm/arm.md: Add simple_return and conditional simple_return.
9074 * config/arm/iterators.md: Add iterator for return and simple_return.
9076 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9078 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
9079 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
9080 (arm_emit_vfp_multi_reg_pop): Likewise.
9081 (thumb2_emit_ldrd_pop): Likewise.
9082 (arm_expand_epilogue): Add misc REG_CFA notes.
9083 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
9085 2013-05-29 Lawrence Crowl <crowl@google.com>
9087 * config/arm/t-arm: Update for below.
9089 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
9090 Change type to hash_table. Update dependent calls and types.
9092 * config/i386/t-cygming: Update for below.
9094 * config/i386/t-interix: Update for below.
9096 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
9097 Change type to hash_table. Update dependent calls and types.
9098 (i386_find_on_wrapper_list::wrappers): Likewise.
9100 * config/ia64/t-ia64: Update for below.
9102 * config/ia64/ia64.c (bundle_state_table):
9103 Change type to hash_table. Update dependent calls and types.
9105 * config/mips/mips.c (mips_reorg_process_insns::htab):
9106 Change type to hash_table. Update dependent calls and types.
9108 * config/sol2.c (solaris_comdat_htab):
9109 Change type to hash_table. Update dependent calls and types.
9111 * config/t-sol2: Update for above.
9113 2013-05-29 Teresa Johnson <tejohnson@google.com>
9115 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
9116 functions are not yet marked as defined.
9118 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
9119 Pat Haugen <pthaugen@us.ibm.com>
9120 Peter Bergner <bergner@vnet.ibm.com>
9122 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
9126 (vrotl<mode>3): Likewise.
9127 (vashl<mode>3): Likewise.
9128 (vlshr<mode>3): Likewise.
9129 (vashr<mode>3): Likewise.
9131 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9132 support for power8 V2DI builtins.
9134 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
9135 power8 V2DI builtins.
9136 (vupkhsw): Likewise.
9137 (vupklsw): Likewise.
9138 (vaddudm): Likewise.
9143 (vpkudum): Likewise.
9144 (vpksdss): Likewise.
9145 (vpkudus): Likewise.
9146 (vpksdus): Likewise.
9151 (vsubudm): Likewise.
9152 (vcmpequd): Likewise.
9153 (vcmpgtsd): Likewise.
9154 (vcmpgtud): Likewise.
9155 (vcmpequd_p): Likewise.
9156 (vcmpgtsd_p): Likewise.
9157 (vcmpgtud_p): Likewise.
9158 (vupkhsw): Likewise.
9159 (vupklsw): Likewise.
9160 (vaddudm): Likewise.
9165 (vpksdss): Likewise.
9166 (vpksdus): Likewise.
9167 (vpkudum): Likewise.
9168 (vpkudus): Likewise.
9173 (vsubudm): Likewise.
9175 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
9176 support for power8 V2DI instructions.
9178 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
9179 power8 V2DI instructions. Combine pack and unpack insns to use an
9180 iterator for each mode. Check whether a particular mode supports
9181 Altivec instructions instead of just checking TARGET_ALTIVEC.
9182 (UNSPEC_VPKUWUM): Likewise.
9183 (UNSPEC_VPKSHSS): Likewise.
9184 (UNSPEC_VPKSWSS): Likewise.
9185 (UNSPEC_VPKUHUS): Likewise.
9186 (UNSPEC_VPKSHUS): Likewise.
9187 (UNSPEC_VPKUWUS): Likewise.
9188 (UNSPEC_VPKSWUS): Likewise.
9189 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
9190 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
9191 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
9192 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
9193 (UNSPEC_VUPKHSB): Likewise.
9194 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
9195 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
9196 (UNSPEC_VUPKHSH): Likewise.
9197 (UNSPEC_VUPKLSB): Likewise.
9198 (UNSPEC_VUPKLSH): Likewise.
9200 (VI_char): Likewise.
9201 (VI_scalar): Likewise.
9202 (VI_unit): Likewise.
9204 (VP_small): Likewise.
9205 (VP_small_lc): Likewise.
9206 (VU_char): Likewise.
9207 (add<mode>3): Likewise.
9208 (altivec_vaddcuw): Likewise.
9209 (altivec_vaddu<VI_char>s): Likewise.
9210 (altivec_vadds<VI_char>s): Likewise.
9211 (sub<mode>3): Likewise.
9212 (altivec_vsubcuw): Likewise.
9213 (altivec_vsubu<VI_char>s): Likewise.
9214 (altivec_vsubs<VI_char>s): Likewise.
9215 (altivec_vavgs<VI_char>): Likewise.
9216 (altivec_vcmpbfp): Likewise.
9217 (altivec_eq<mode>): Likewise.
9218 (altivec_gt<mode>): Likewise.
9219 (altivec_gtu<mode>): Likewise.
9220 (umax<mode>3): Likewise.
9221 (smax<mode>3): Likewise.
9222 (umin<mode>3): Likewise.
9223 (smin<mode>3): Likewise.
9224 (altivec_vpkuhum): Likewise.
9225 (altivec_vpkuwum): Likewise.
9226 (altivec_vpkshss): Likewise.
9227 (altivec_vpkswss): Likewise.
9228 (altivec_vpkuhus): Likewise.
9229 (altivec_vpkshus): Likewise.
9230 (altivec_vpkuwus): Likewise.
9231 (altivec_vpkswus): Likewise.
9232 (altivec_vpks<VI_char>ss): Likewise.
9233 (altivec_vpks<VI_char>us): Likewise.
9234 (altivec_vpku<VI_char>us): Likewise.
9235 (altivec_vpku<VI_char>um): Likewise.
9236 (altivec_vrl<VI_char>): Likewise.
9237 (altivec_vsl<VI_char>): Likewise.
9238 (altivec_vsr<VI_char>): Likewise.
9239 (altivec_vsra<VI_char>): Likewise.
9240 (altivec_vsldoi_<mode>): Likewise.
9241 (altivec_vupkhsb): Likewise.
9242 (altivec_vupkhs<VU_char>): Likewise.
9243 (altivec_vupkls<VU_char>): Likewise.
9244 (altivec_vupkhsh): Likewise.
9245 (altivec_vupklsb): Likewise.
9246 (altivec_vupklsh): Likewise.
9247 (altivec_vcmpequ<VI_char>_p): Likewise.
9248 (altivec_vcmpgts<VI_char>_p): Likewise.
9249 (altivec_vcmpgtu<VI_char>_p): Likewise.
9250 (abs<mode>2): Likewise.
9251 (vec_unpacks_hi_v16qi): Likewise.
9252 (vec_unpacks_hi_v8hi): Likewise.
9253 (vec_unpacks_lo_v16qi): Likewise.
9254 (vec_unpacks_hi_<VP_small_lc>): Likewise.
9255 (vec_unpacks_lo_v8hi): Likewise.
9256 (vec_unpacks_lo_<VP_small_lc>): Likewise.
9257 (vec_pack_trunc_v8h): Likewise.
9258 (vec_pack_trunc_v4si): Likewise.
9259 (vec_pack_trunc_<mode>): Likewise.
9261 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
9263 (vec_vmaxsd): Likewise.
9264 (vec_vmaxud): Likewise.
9265 (vec_vminsd): Likewise.
9266 (vec_vminud): Likewise.
9267 (vec_vpksdss): Likewise.
9268 (vec_vpksdus): Likewise.
9269 (vec_vpkudum): Likewise.
9270 (vec_vpkudus): Likewise.
9271 (vec_vrld): Likewise.
9272 (vec_vsld): Likewise.
9273 (vec_vsrad): Likewise.
9274 (vec_vsrd): Likewise.
9275 (vec_vsubudm): Likewise.
9276 (vec_vupkhsw): Likewise.
9277 (vec_vupklsw): Likewise.
9279 2013-05-29 Jan Hubicka <jh@suse.cz>
9281 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
9282 flags; reorder rest of fields in more consistent way.
9283 (varpool_node): Remove analyzed, finalized and alias.
9284 (cgraph_ndoe): Likewise.
9285 (symtab_alias_ultimate_target): New function.
9286 (cgraph_function_node): Move offline.
9287 (cgraph_reset_node): Declare.
9288 (cgraph_comdat_can_be_unshared_p): Remove.
9289 (varpool_remove_initializer): Declare.
9290 (varpool_first_defined_variable, varpool_next_defined_variable
9291 cgraph_first_defined_function, cgraph_next_defined_function): Update.
9292 (cgraph_function_with_gimple_body_p): Update.
9293 (varpool_all_refs_explicit_p): Update.
9294 (symtab_alias_target): New function.
9295 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
9296 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
9297 (cgraph_function_or_thunk_node): Simplify using
9298 symtab_alias_ultimate_target.
9299 (varpool_variable_node): Likewise.
9300 * cgraph.c (cgraph_create_function_alias): Update.
9301 (cgraph_add_thunk): Update.
9302 (cgraph_remove_node): Update.
9303 (dump_cgraph_node): Do not dump removed flags.
9304 (cgraph_function_body_availability): Update.
9305 (cgraph_propagate_frequency): Update.
9306 (verify_cgraph_node): Check sanity of local flag.
9307 (cgraph_function_node): Move here from cgraph.h; revamp for
9308 cgraph_function_or_thunk_node.
9309 * lto-symtab.c (lto_varpool_replace_node): Update.
9310 (lto_symtab_resolve_can_prevail_p): Update.
9311 (lto_symtab_merge_cgraph_nodes): Update.
9312 * ipa-cp.c (determine_versionability, initialize_node_lattices,
9313 propagate_constants_accross_call, devirtualization_time_bonus,
9314 ipcp_propagate_stage): Update.
9315 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
9316 * ipa-inline-transform.c (clone_inlined_nodes,
9317 preserve_function_body_p): Update.
9318 * ipa-reference.c (propagate): Update.
9319 (write_node_summary_p): Update.
9320 * toplev.c (wrapup_global_declaration_2): Update.
9321 * cgraphunit.c (cgraph_analyze_function): Rename to ...
9322 (analyze_function) ... this one.
9323 (cgraph_process_new_functions): Update.
9324 (cgraph_reset_node): Export.
9325 (cgraph_finalize_function): Update.
9326 (cgraph_add_new_function): Update.
9327 (process_function_and_variable_attributes): Update.
9328 (varpool_finalize_decl): Update.
9329 (symbol_finalized): Remove.
9330 (symbol_finalized_and_needed): Rename to ...
9331 (symbol_defined_and_needed): ... update.
9332 (cgraph_analyze_functions): Update.
9333 (handle_alias_pairs): Update.
9334 (mark_functions_to_output): Update.
9335 (assemble_thunk): Update.
9336 (output_in_order): Update.
9337 (output_weakrefs): Update.
9338 (finalize_compilation_unit): Update.
9339 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
9340 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
9341 input_node, input_varpool_node): Update.
9342 * dbxout.c (dbxout_expand_expr): Update.
9343 * cgraphclones.c (cgraph_clone_node): Update.
9344 (cgraph_copy_node_for_versioning): Update.
9345 (cgraph_materialize_clone): Update.
9346 (cgraph_materialize_all_clones): Update.
9347 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
9348 propagate_pure_const, propagate_nothrow): Update.
9349 * lto-streamer-out.c (lto_output, write_symbol): Update.
9350 * ipa-utils.c (ipa_reverse_postorder): Update.
9351 * ipa-inline.c (can_inline_edge_p): Update.
9352 (update_caller_keys, ipa_inline): Update.
9353 * dwarf2out.c (reference_to_unused,
9354 premark_types_used_by_global_vars_helper): Update.
9355 * tree-eh.c (tree_could_trap_p): Update.
9356 * ipa-split.c (consider_split, execute_split_functions): Update.
9357 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
9358 has_addr_references_p): Update; move ahead in file for better
9360 (process_references): Simplify.
9361 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
9363 (cgraph_comdat_can_be_unshared_p): Make static.
9364 (cgraph_externally_visible_p): Update.
9365 (varpool_externally_visible_p): Update.
9366 (function_and_variable_visibility): Update.
9367 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
9368 ipa_tm_mark_force_output_node): Update.
9369 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
9370 estimate_edge_devirt_benefit, inline_generate_summary,
9371 inline_write_summary): Update.
9372 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
9373 * ipa-prop.c (ipa_compute_jump_functions): Update.
9374 (ipa_print_node_params, ipa_prop_read_section,
9375 ipa_update_after_lto_read, read_replacements_section): Update.
9376 * varasm.c (mark_decl_referenced): Update.
9377 (assemble_alias, dump_tm_clone_pairs): Update.
9378 * tree-inline.c (copy_bb): Update.
9379 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
9381 * symtab.c (dump_symtab_base): Print new flags.
9382 (verify_symtab_base): Verify new flags.
9383 (symtab_alias_ultimate_target): New function.
9384 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
9385 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
9387 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
9389 * i386.c (ix86_get_function_versions_dispatcher,
9390 ix86_generate_version_dispatcher_body): Update.
9391 (fold_builtin_cpu): Use varpool_add_new_variable.
9392 * varpool.c (varpool_remove_initializer): Break out from ...
9393 (varpool_remove_node): ... this one.
9394 (dump_varpool_node, varpool_node_for_asm,
9395 cgraph_variable_initializer_availability, varpool_analyze_node,
9396 varpool_assemble_decl, varpool_remove_unreferenced_decls,
9397 varpool_finalize_named_section_flags, varpool_create_variable_alias):
9400 2013-05-29 Jan Hubicka <jh@suse.cz>
9402 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
9404 2013-05-29 Easwaran Raman <eraman@google.com>
9406 PR tree-optimization/57442
9407 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
9408 when control exits the main loop.
9410 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
9412 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
9414 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
9415 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
9416 * rx/t-rx: Add rx100 under multi library matches option for nofpu
9419 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9421 PR tree-optimization/57441
9422 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
9423 Don't limit size of incr_vec to number of candidates.
9425 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
9427 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
9428 and mips16 directories.
9429 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
9430 (MULTILIB_DIRNAMES): Ditto.
9431 (MULTILIB_EXCEPTIONS): Add new exceptions.
9432 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
9433 (MULTILIB_DIRNAMES): Ditto.
9434 (MULTILIB_EXCEPTIONS): Add new exceptions.
9436 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9437 Marcus Shawcroft <marcus.shawcroft@arm.com>
9439 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
9440 SYMBOL_TINY_ABSOLUTE.
9441 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
9442 SYMBOL_TINY_ABSOLUTE.
9443 (aarch64_expand_mov_immediate): Likewise.
9444 (aarch64_classify_symbol): Likewise.
9445 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
9446 Permit SYMBOL_TINY_ABSOLUTE.
9447 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
9449 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9450 Marcus Shawcroft <marcus.shawcroft@arm.com>
9452 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
9453 Refactor if/switch. Replace gcc_assert with if.
9455 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9457 * config/i386/i386.c (initial_ix86_tune_features): Enable
9458 FP Reassociation for AMD bdver1 and bdver2.
9460 2013-05-29 Martin Jambor <mjambor@suse.cz>
9462 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
9463 and IMAGPART_EXPR do not occur within other handled_components.
9465 2013-05-29 Richard Biener <rguenther@suse.de>
9467 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
9468 access on whether the use is in the BB we currently try to
9470 (vect_bb_vectorization_profitable_p): Pass the BB we currently
9471 vectorize to vect_bb_slp_scalar_cost.
9473 2013-05-29 Richard Biener <rguenther@suse.de>
9475 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
9476 computing scalar cost offsetted by stmts that are kept live
9478 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
9479 for computation of scalar cost.
9481 2013-05-28 Steve Ellcey <sellcey@mips.com>
9483 * config/mips/mips-cpus.def (mips32r2): Change processor type.
9485 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9487 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
9488 array notation built-in reduction functions.
9489 * doc/passes.texi (Passes): Added documentation about changes done
9491 * doc/invoke.texi (C Dialect Options): Added documentation about
9492 the -fcilkplus flag.
9493 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
9494 (BUILTINS_DEF): Depend on cilkplus.def.
9495 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
9496 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
9497 * cilkplus.def: New file.
9499 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
9501 PR rtl-optimization/57439
9502 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
9504 2013-05-28 Easwaran Raman <eraman@google.com>
9506 PR tree-optimization/57337
9507 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
9508 (find_insert_point): Correctly identify the insertion point
9509 when two statements with the same UID is compared.
9511 2013-05-28 Richard Biener <rguenther@suse.de>
9513 PR tree-optimization/56787
9514 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
9515 from the list of data references.
9516 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
9518 (vect_analyze_loop_operations): Likewise.
9519 (vect_transform_loop): Remove clobbers.
9521 2013-05-28 Martin Jambor <mjambor@suse.cz>
9523 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
9524 and REALPART_EXPRs have scalar type.
9526 2013-05-28 Richard Biener <rguenther@suse.de>
9528 PR tree-optimization/57411
9529 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
9531 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
9532 virtual operand propagation.
9534 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
9536 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
9537 destination register for bmasksi_vis.
9538 (vector_init_bshuffle): Likewise.
9539 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
9541 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
9543 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
9544 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
9545 mode if the instruction isn't available in the original mode.
9546 * config/sparc/sparc.opt (mfix-ut699): New option.
9547 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
9548 (divdf3): Turn into expander.
9549 (divdf3_nofix): New insn.
9550 (divdf3_fix): Likewise.
9551 (divsf3): Disable if -mfix-ut699.
9552 (sqrtdf2): Turn into expander.
9553 (sqrtdf2_nofix): New insn.
9554 (sqrtdf2_fix): Likewise.
9555 (sqrtsf2): Disable if -mfix-ut699.
9557 2013-05-27 Richard Biener <rguenther@suse.de>
9560 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
9561 block for the new loop.
9563 2013-05-27 Richard Biener <rguenther@suse.de>
9565 PR tree-optimization/57343
9566 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
9567 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
9568 (number_of_iterations_cond): Do not build the folded tree.
9570 2013-05-27 Richard Biener <rguenther@suse.de>
9574 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
9575 OEP_CONSTANT_ADDRESS_OF retained.
9577 PR tree-optimization/57417
9578 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
9580 (set_ssa_val_to): Compare addresses using
9581 get_addr_base_and_unit_offset.
9583 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
9585 PR rtl-optimization/56833
9586 * postreload.c (move2add_record_mode): New function.
9587 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
9588 (move2add_use_add2_insn): Use move2add_record_sym_value.
9589 (move2add_use_add3_insn): Likewise.
9590 (reload_cse_move2add): Use move2add_valid_value_p and
9591 move2add_record_mode. Invalidate call-clobbered and REG_INC
9592 affected regs by setting reg_mode to VOIDmode.
9593 (move2add_note_store): Don't pretend the inside of a SUBREG is
9594 the actual destination. Invalidate single/leading registers by
9595 setting reg_mode to VOIDmode.
9596 Use move2add_record_sym_value, move2add_valid_value_p and
9597 move2add_record_mode.
9599 2013-05-27 Richard Biener <rguenther@suse.de>
9601 PR tree-optimization/57396
9602 * tree-affine.c (double_int_constant_multiple_p): Properly
9603 return false for val == 0 and div != 0.
9605 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9607 * config/mips/mips.h: Use #elif in preprocessor conditions.
9609 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9612 * config/mips/constraints.md (kl): New constraint.
9613 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
9614 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
9615 constraint for operand 0. Split after CSE for MIPS16. Emit a move
9617 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
9619 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9622 * config/mips/mips.c (mips_can_inline_p): New function.
9623 (TARGET_CAN_INLINE_P): Define.
9625 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
9627 * sched-int.h (ds_t, dw_t): Make unsigned int.
9628 Fix documentation that describes how all the ds_t bits are used.
9629 Reserve the last bit for delayed-branch scheduling.
9630 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
9631 (BITS_PER_DEP_WEAK): Fix definition and documentation.
9632 (gen_dep_weak_1): Remove prototype.
9633 * sched-deps.c (get_dep_weak_1): Make static.
9634 * target.def (speculate_insn, needs_block_p, gen_spec_check,
9635 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
9636 * doc/tm.texi: Regenerate.
9637 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
9639 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
9642 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
9644 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
9645 Sandra Loosemore <sandra@codesourcery.com>
9647 * config.gcc (powerpc-*): Allow native for with-cpu.
9649 2013-05-24 Jeff Law <law@redhat.com>
9651 PR tree-optimization/57124
9652 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
9653 conversion feeding a condition if the range has an overflow
9654 if -fstrict-overflow. Add warnings for when we do make the
9657 2013-05-24 Dehao Chen <dehao@google.com>
9659 * tree-cfg.c (locus_discrim_map): Fix the typo.
9660 (locus_discrim_hasher): Likewise.
9661 (locus_discrim_hasher::hash): Likewise.
9662 (locus_discrim_hasher::equal): Likewise.
9664 2013-05-24 Martin Jambor <mjambor@suse.cz>
9666 PR tree-optimization/57294
9667 * cgraph.h (ipa_record_stmt_references): Declare.
9668 * cgraphbuild.c (ipa_record_stmt_references): New function.
9669 (build_cgraph_edges): Use ipa_record_stmt_references.
9670 (rebuild_cgraph_edges): Likewise.
9671 (cgraph_rebuild_references): Likewise.
9672 * ipa-prop.c (ipa_modify_call_arguments): Discard references
9673 associated with the old statement and build references from the
9674 newly built statements.
9675 * ipa-ref.c (ipa_remove_stmt_references): New function.
9676 * ipa-ref.h (ipa_remove_stmt_references): Declare.
9678 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
9680 * lra-constraints.c (emit_spill_move): Use smaller mode for
9682 (check_and_process_move): Consider mem-reg moves for secondary
9684 (curr_insn_transform): Don't lose insns emitted before for
9685 secondary memory moves.
9686 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
9687 reg set up in the current insn.
9689 2013-05-24 Dehao Chen <dehao@google.com>
9691 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
9693 (locus_descrim_hasher::equal): Likewise.
9694 (build_gimple_cfg): New discriminator assignment algorithm.
9695 (make_edges): Likewise.
9696 (next_discriminator_for_locus): Likewise.
9697 (same_line_p): Likewise.
9698 (assign_discriminators): Likewise.
9699 (make_cond_expr_edges): Likewise.
9700 (make_gimple_switch_edges): Likewise.
9701 (make_goto_expr_edges): Likewise.
9702 (make_gimple_asm_edges): Likewise.
9704 2013-05-24 Ian Bolton <ian.bolton@arm.com>
9706 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
9707 X format specifier to only display bottom 16 bits.
9708 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
9709 immediate to match for operand 2, since it will be masked.
9711 2013-05-24 Richard Biener <rguenther@suse.de>
9713 PR tree-optimization/57287
9714 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
9715 all SSA names that occur in abnormal PHIs.
9717 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
9720 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
9721 that index is not negative.
9723 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
9725 PR rtl-optimization/55177
9726 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
9727 (simplify_byte_swapping_operation): New.
9728 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
9729 (simplify_relational_operation_1): Deal with BSWAP.
9731 2013-05-23 Richard Henderson <rth@redhat.com>
9734 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
9735 (ix86_reorg): Call it.
9737 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
9740 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
9741 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
9742 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
9744 2013-05-23 Christian Bruel <christian.bruel@st.com>
9747 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
9749 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9750 Marcus Shawcroft <marcus.shawcroft@arm.com>
9752 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
9753 * config/aarch64/constraints.md (Usa): Remove.
9754 * doc/md.texi (AArch64 Usa): Remove.
9756 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
9757 Marcus Shawcroft <marcus.shawcroft@arm.com>
9759 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
9760 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
9761 * config/aarch64/predicates.md (aarch64_const_address): Remove.
9762 (aarch64_mov_operand): Use aarch64_mov_operand_p.
9764 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
9766 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
9767 instruction (AdvSIMD).
9768 * config/aarch64/aarch64-builtins.c
9769 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
9770 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
9772 2013-05-23 Martin Jambor <mjambor@suse.cz>
9775 * tree.h (contains_bitfld_component_ref_p): Declare.
9776 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
9777 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
9779 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
9780 not access a bit-field. Assert all final offsets are byte-aligned.
9782 2013-05-23 Richard Biener <rguenther@suse.de>
9784 PR tree-optimization/57380
9785 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
9786 least one invariant or re-used load.
9787 * passes.c (init_optimization_passes): Move pass_phiprop before
9790 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
9792 * config/aarch64/aarch64-simd.md
9793 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
9795 2013-05-23 Richard Biener <rguenther@suse.de>
9798 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
9799 OEP_CONSTANT_ADDRESS_OF retained.
9801 2013-05-23 Jakub Jelinek <jakub@redhat.com>
9804 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
9805 don't lower unit. Handle unit not being always BITS_PER_WORD.
9807 2013-05-23 Richard Biener <rguenther@suse.de>
9809 PR rtl-optimization/57341
9810 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
9811 instead of true_dependence.
9813 2013-05-22 David Malcolm <dmalcolm@redhat.com>
9815 * bb-reorder.c (branch_threshold): Make const.
9816 (exec_threshold): Ditto.
9818 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9819 Pat Haugen <pthaugen@us.ibm.com>
9820 Peter Bergner <bergner@vnet.ibm.com>
9822 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
9823 documentation for the power8 crypto builtins.
9825 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
9827 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
9828 macros for defining power8 builtin functions.
9829 (BU_P8V_AV_2): Likewise.
9830 (BU_P8V_AV_P): Likewise.
9831 (BU_P8V_VSX_1): Likewise.
9832 (BU_P8V_OVERLOAD_1): Likewise.
9833 (BU_P8V_OVERLOAD_2): Likewise.
9834 (BU_CRYPTO_1): Likewise.
9835 (BU_CRYPTO_2): Likewise.
9836 (BU_CRYPTO_3): Likewise.
9837 (BU_CRYPTO_OVERLOAD_1): Likewise.
9838 (BU_CRYPTO_OVERLOAD_2): Likewise.
9839 (XSCVSPDP): Fix typo, point to the correct instruction.
9840 (VCIPHER): Add power8 crypto builtins.
9841 (VCIPHERLAST): Likewise.
9842 (VNCIPHER): Likewise.
9843 (VNCIPHERLAST): Likewise.
9844 (VPMSUMB): Likewise.
9845 (VPMSUMH): Likewise.
9846 (VPMSUMW): Likewise.
9847 (VPERMXOR_V2DI): Likewise.
9848 (VPERMXOR_V4SI: Likewise.
9849 (VPERMXOR_V8HI: Likewise.
9850 (VPERMXOR_V16QI: Likewise.
9851 (VSHASIGMAW): Likewise.
9852 (VSHASIGMAD): Likewise.
9854 (VPERMXOR): Likewise.
9855 (VSHASIGMA): Likewise.
9857 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9858 __CRYPTO__ if the crypto instructions are available.
9859 (altivec_overloaded_builtins): Add support for overloaded power8
9862 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
9863 support for power8 crypto builtins.
9864 (builtin_function_type): Likewise.
9865 (altivec_init_builtins): Add support for builtins that take vector
9866 long long (V2DI) arguments.
9868 * config/rs6000/crypto.md: New file, define power8 crypto
9871 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9872 Pat Haugen <pthaugen@us.ibm.com>
9873 Peter Bergner <bergner@vnet.ibm.com>
9875 * doc/invoke.texi (Option Summary): Add power8 options.
9876 (RS/6000 and PowerPC Options): Likewise.
9878 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
9879 constraints.md instead of rs6000.h. Reorder w* constraints. Add
9880 wm, wn, wr documentation.
9882 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
9883 registers if direct move instructions are enabled.
9884 (wn): New constraint for no registers.
9885 (wq): New constraint for quad word even GPR registers.
9886 (wr): New constraint if 64-bit instructions are enabled.
9887 (wv): New constraint if power8 vector instructions are enabled.
9888 (wQ): New constraint for quad word memory locations.
9890 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
9891 constraint for 0..15 for crypto instructions.
9892 (gpc_reg_operand): If VSX allow registers in VSX registers as well
9893 as GPR and floating point registers.
9894 (int_reg_operand): New predicate to match only GPR registers.
9895 (base_reg_operand): New predicate to match base registers.
9896 (quad_int_reg_operand): New predicate to match even GPR registers
9897 for quad memory operations.
9898 (vsx_reg_or_cint_operand): New predicate to allow vector logical
9899 operations in both GPR and VSX registers.
9900 (quad_memory_operand): New predicate for quad memory operations.
9901 (reg_or_indexed_operand): New predicate for direct move support.
9903 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
9904 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
9905 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
9906 (POWERPC_MASKS): Add power8 options.
9907 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
9910 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9911 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
9913 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
9914 (-mpower8-fusion): New power8 options.
9915 (-mpower8-fusion-sign): Likewise.
9916 (-mpower8-vector): Likewise.
9917 (-mcrypto): Likewise.
9918 (-mdirect-move): Likewise.
9919 (-mquad-memory): Likewise.
9921 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
9923 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
9925 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
9926 (rs6000_debug_vector_unit): Add p8_vector.
9927 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
9928 definitions. Also print fusion state.
9929 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
9930 (rs6000_builtin_mask_calculate): Add power8 builtin support.
9931 (rs6000_option_override_internal): Add support for power8.
9932 (rs6000_common_init_builtins): Add debugging for skipped builtins
9934 (rs6000_adjust_cost): Add power8 support.
9935 (rs6000_issue_rate): Likewise.
9936 (insn_must_be_first_in_group): Likewise.
9937 (insn_must_be_last_in_group): Likewise.
9938 (force_new_group): Likewise.
9939 (rs6000_register_move_cost): Likewise.
9940 (rs6000_opt_masks): Likewise.
9942 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
9943 power8 capable assembler, default to power7 options.
9944 (TARGET_DIRECT_MOVE): Likewise.
9945 (TARGET_CRYPTO): Likewise.
9946 (TARGET_P8_VECTOR): Likewise.
9947 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
9948 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
9949 (VECTOR_MEM_P8_VECTOR_P): Likewise.
9950 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
9951 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
9952 (TARGET_XSCVDPSPN): Likewise.
9953 (TARGET_XSCVSPDPN): Likewsie.
9954 (TARGET_SYNC_HI_QI): Likewise.
9955 (TARGET_SYNC_TI): Likewise.
9956 (MASK_CRYPTO): Likewise.
9957 (MASK_DIRECT_MOVE): Likewise.
9958 (MASK_P8_FUSION): Likewise.
9959 (MASK_P8_VECTOR): Likewise.
9960 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
9961 temporary used by some of the direct move instructions to get two FP
9962 temporary registers does not force creation of a stack frame.
9963 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
9964 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
9965 that any VSX registers are tieable, even if they are also an
9966 Altivec vector mode.
9967 (r6000_reg_class_enum): Add wm, wr, wv constraints.
9968 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
9969 (RS6000_BTM_CRYPTO): Likewise.
9970 (RS6000_BTM_COMMON): Likewise.
9972 * config/rs6000/rs6000.md (cpu attribute): Add power8.
9973 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
9974 (enum rs6000_vector): Add power8 vector support.
9976 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9980 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
9981 (any_sibcall_could_use_r3): this and handle indirect calls.
9982 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
9984 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9986 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
9988 2013-05-22 Richard Biener <rguenther@suse.de>
9991 * profile.c (branch_prob): Do not split blocks that are
9992 abnormally receiving from ECF_RETURNS_TWICE functions.
9994 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9996 * recog.c (offsettable_address_addr_space_p): Fix calculation of
9997 address mode. Move pointer mode initialization to the same place.
9999 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10001 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
10002 while it has any effect.
10004 2013-05-21 Easwaran Raman <eraman@google.com>
10006 PR tree-optimization/57322
10007 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
10008 UID of the statement added to the BB to be 1.
10010 2013-05-21 Jakub Jelinek <jakub@redhat.com>
10012 PR tree-optimization/57331
10013 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
10014 of conversion from pointer type to integral type with integer.
10016 2013-05-21 Martin Jambor <mjambor@suse.cz>
10019 * ipa-prop.c (ipa_read_node_info): Process param_used and
10020 controlled_uses in the same order as when writing.
10022 2013-05-21 Magnus Granberg <baldrick@free.fr>
10025 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
10027 2013-05-21 Richard Biener <rguenther@suse.de>
10029 PR tree-optimization/57318
10030 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
10031 estimate stmts with side-effects as likely eliminated.
10033 2013-05-21 Richard Biener <rguenther@suse.de>
10035 PR tree-optimization/57330
10036 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
10037 preserve the call stmts fntype.
10039 2013-05-21 Richard Biener <rguenther@suse.de>
10041 PR tree-optimization/57303
10042 * tree-ssa-sink.c (statement_sink_location): Improve killing
10043 stmt detection and properly handle self-assignments.
10045 2013-05-21 Christian Bruel <christian.bruel@st.com>
10047 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
10048 spanning registers. LEAF_REG_REMAP is supported only for contiguous
10049 registers. Set register size out of the PARALLEL loop.
10051 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
10054 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
10055 (*fmasf4, *fmasf4_media): New insns.
10057 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10059 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
10060 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
10061 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
10062 (mips_idiv_insns): Update the comments to say that the returned
10063 instruction counts are in units of BASE_INSN_LENGTH.
10064 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
10065 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
10066 using 2 rather than 4 as the length of indirect MIPS16 and
10067 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
10068 length of a NOP. Don't divide MIPS16 lengths by 2.
10069 (mips16_split_long_branches): Assume a branch is long if the
10070 length is greater than 4 rather than 8.
10071 * config/mips/mips.md (length): Give MIPS16 lengths directly,
10072 rather than multiplying them by 2. Multiply instruction counts
10073 by BASE_INSN_LENGTH rather than 4.
10074 (*jump_mips16, tls_get_tp_mips16_<mode>)
10075 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
10077 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10079 * config/mips/mips.md (extended_mips16): Remove branch case.
10080 (length): Remove duplicated extended_mips16 test.
10082 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10084 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
10086 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
10088 * recog.h (Recog_data): Rename to...
10089 (recog_data_d): ...this.
10090 (recog_data): Update accordingly.
10091 * recog.c (recog_data): Likewise.
10092 * reload.c (save_recog_data): Likewise.
10093 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
10094 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
10096 2013-05-17 Julian Brown <julian@codesourcery.com>
10098 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
10099 found in a REG_EQUAL note, invalidate it.
10101 2013-05-17 Easwaran Raman <eraman@google.com>
10103 * tree-ssa-reassoc.c (find_insert_point): New function.
10104 (insert_stmt_after): Likewise.
10105 (get_def_stmt): Likewise.
10106 (ensure_ops_are_available): Likewise.
10107 (not_dominated_by): Likewise.
10108 (rewrite_expr_tree): Do not move statements beyond what is
10109 necessary. Remove call to swap_ops_for_binary_stmt...
10110 (reassociate_bb): ... and move it here.
10111 (build_and_add_sum): Assign UIDs for new statements.
10112 (linearize_expr): Likewise.
10113 (do_reassoc): Renumber gimple statement UIDs.
10115 2013-05-17 Jan Hubicka <jh@suse.cz>
10117 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
10119 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
10120 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
10122 (output_weakrefs): Update.
10124 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
10125 Martin Jambor <mjambor@suse.cz>
10127 PR middle-end/57276
10128 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
10129 value that corresponds to the given aggval is found in values vector.
10131 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
10133 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
10134 sse, sse2, sse3, ssse3 and sse4a flags to options.
10136 2013-05-17 David Malcolm <dmalcolm@redhat.com>
10138 * gengtype-state.c: (s_expr_writer): New class, to handle
10139 prettifying of output layout of s-expressions.
10140 (state_writer): New class, to write out gtype.state.
10141 (state_written_type_count): Move this variable into member data of
10143 (s_expr_writer::s_expr_writer): New code: constructor for new class
10144 (state_writer::state_writer(): ditto
10145 (s_expr_writer::write_new_line): New function
10146 (s_expr_writer::write_any_indent): ditto
10147 (s_expr_writer::begin_s_expr): ditto
10148 (s_expr_writer::end_s_expr): ditto
10149 (write_state_fileloc): convert to method of state_writer...
10150 (state_writer:: write_state_fileloc): ...and use methods of
10151 s_expr_writer to write indentation into the gtype.state output file
10152 to visually represent the hierarchical structure of the list
10154 (write_state_fields): ditto, renaming to...
10155 (state_writer::write_state_fields)
10156 (write_state_a_string): ditto, renaming to...
10157 (state_writer::write_state_a_string)
10158 (write_state_string_option): ditto, renaming to...
10159 (state_writer::write_state_string_option)
10160 (write_state_type_option): ditto, renaming to...
10161 (state_writer::write_state_type_option)
10162 (write_state_nested_option): ditto, renaming to...
10163 (state_writer::write_state_nested_option)
10164 (write_state_option): ditto, renaming to...
10165 (state_writer::write_state_option)
10166 (write_state_options): ditto, renaming to...
10167 (state_writer::write_state_options)
10168 (write_state_lang_bitmap): ditto, renaming to...
10169 (state_writer::write_state_lang_bitmap)
10170 (write_state_version): ditto, renaming to...
10171 (state_writer::write_state_version)
10172 (write_state_scalar_type): ditto, renaming to...
10173 (state_writer::write_state_scalar_type)
10174 (write_state_string_type): ditto, renaming to...
10175 (state_writer::write_state_string_type)
10176 (write_state_undefined_type): ditto, renaming to...
10177 (state_writer::write_state_undefined_type)
10178 (write_state_struct_union_type): ditto, renaming to...
10179 (state_writer::write_state_struct_union_type)
10180 (write_state_struct_type): ditto, renaming to...
10181 (state_writer::write_state_struct_type)
10182 (write_state_user_struct_type): ditto, renaming to...
10183 (state_writer::write_state_user_struct_type)
10184 (write_state_lang_struct_type): ditto, renaming to...
10185 (state_writer::write_state_lang_struct_type)
10186 (write_state_param_struct_type): ditto, renaming to...
10187 (state_writer::write_state_param_struct_type)
10188 (write_state_pointer_type): ditto, renaming to...
10189 (state_writer::write_state_pointer_type)
10190 (write_state_array_type): ditto, renaming to...
10191 (state_writer::write_state_array_type)
10192 (write_state_gc_used): ditto, renaming to...
10193 (state_writer::write_state_gc_used)
10194 (write_state_common_type_content): ditto, renaming to...
10195 (state_writer::write_state_common_type_content)
10196 (write_state_type): ditto, renaming to...
10197 (state_writer::write_state_type)
10198 (write_state_pair_list): ditto, renaming to...
10199 (state_writer::write_state_pair_list)
10200 (write_state_pair): ditto, renaming to...
10201 (state_writer::write_state_pair)
10202 (write_state_typedefs): ditto, renaming to...
10203 (state_writer::write_state_typedefs)
10204 (write_state_structures): ditto, renaming to...
10205 (state_writer::write_state_structures)
10206 (write_state_param_structs): ditto, renaming to...
10207 (state_writer::write_state_param_structs)
10208 (write_state_variables): ditto, renaming to...
10209 (state_writer::write_state_variables)
10210 (write_state_srcdir): ditto, renaming to...
10211 (state_writer::write_state_srcdir)
10212 (write_state_files_list): ditto, renaming to...
10213 (state_writer::write_state_files_list)
10214 (write_state_languages): ditto, renaming to...
10215 (state_writer::write_state_languages)
10216 (write_state): create a state_writer instance and use it when
10217 writing out the state file
10219 2013-05-17 Mike Stump <mikestump@comcast.net>
10221 PR rtl-optimization/57304
10222 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
10223 accessing DF_REF_REAL_LOC.
10225 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10227 PR rtl-optimization/57281
10228 PR rtl-optimization/57300
10229 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
10230 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
10231 what the other splitter did if the registers are dead.
10233 2013-05-17 Richard Biener <rguenther@suse.de>
10235 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
10238 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10240 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
10243 2013-05-17 Marek Polacek <polacek@redhat.com>
10245 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
10246 length when doing non-zero store of storing '\0' to '\0'.
10248 2013-05-17 Jakub Jelinek <jakub@redhat.com>
10250 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
10251 vect_external_def oprnd1 with loop_vinfo, try to emit
10252 optional cast, negation and and stmts on the loop preheader
10253 edge instead of into the pattern def seq.
10255 PR tree-optimization/57051
10256 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
10257 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
10259 2013-05-16 Nick Clifton <nickc@redhat.com>
10261 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
10262 (rl78_is_naked_func): New function.
10263 (rl78_expand_prologue): Skip prologue generation for naked functions.
10264 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
10265 * doc/extend.texi (naked): Add RL78 to the list of processors
10266 that supports this attribute.
10268 2013-05-16 Jeff Law <law@redhat.com>
10270 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
10272 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
10274 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
10275 cache parameters using detect_caches_amd also for CYRIX,
10276 NSC and TM2 signatures.
10278 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
10279 Dzianis Kahanovich <mahatma@eu.by>
10283 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10284 VIA/Centaur processors and determine their cache parameters
10285 using detect_caches_amd.
10287 2013-05-16 Teresa Johnson <tejohnson@google.com>
10289 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
10290 (rtl_verify_edges): New function.
10291 (rtl_verify_bb_insns): Ditto.
10292 (rtl_verify_bb_pointers): Ditto.
10293 (rtl_verify_bb_insn_chain): Ditto.
10294 (rtl_verify_fallthru): Ditto.
10295 (rtl_verify_bb_layout): Ditto.
10296 (rtl_verify_flow_info_1): Outline checks into new functions.
10297 (rtl_verify_flow_info): Ditto.
10299 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
10301 * cfghooks.c (copy_bbs): Add update_dominance argument.
10302 * cfghooks.h (copy_bbs): Update prototype.
10303 * tree-cfg.c (gimple_duplicate_sese_region):
10304 Add update_dominance argument.
10305 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
10306 * tree-ssa-loop-ch.c (copy_loop_headers): Update
10307 gimple_duplicate_sese_region call.
10308 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
10309 Update copy_bbs call.
10310 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
10311 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
10313 2013-05-16 Jakub Jelinek <jakub@redhat.com>
10315 * tree-vectorizer.h (NUM_PATTERNS): Increment.
10316 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
10317 vect_recog_rotate_pattern.
10318 (vect_recog_rotate_pattern): New function.
10320 2013-05-16 Jason Merrill <jason@redhat.com>
10322 * Makefile.in (LLINKER): New variable.
10323 (mostlyclean): Remove link mutex.
10324 * configure.ac: Handle --enable-link-mutex.
10325 * lock-and-run.sh: New script.
10327 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10330 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
10333 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10335 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
10337 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
10339 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
10340 * config/arm/arm.c (next_consecutive_mem): New function.
10341 (gen_movmem_ldrd_strd): Likewise.
10342 * config/arm/arm.md (movmemqi): Update condition and code.
10343 (unaligned_loaddi, unaligned_storedi): New patterns.
10345 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10347 * config.gcc: Obsolete *-*-solaris2.9*.
10348 * doc/install.texi (Specific, *-*-solaris2*): Document it.
10350 2013-05-16 Richard Biener <rguenther@suse.de>
10352 * passes.c (init_optimization_passes): Move pass_parallelize_loops
10353 earlier, after GRAPHITE transforms and IV canonicalization.
10355 2013-05-16 Jakub Jelinek <jakub@redhat.com>
10357 * omp-low.c (extract_omp_for_data): For collapsed loops,
10358 if at least one of the loops is known at compile time to
10359 iterate zero times, set count to 0.
10360 (expand_omp_regimplify_p): New function.
10361 (expand_omp_for_generic): For collapsed loops, if at least
10362 one of the loops isn't known to iterate at least once,
10363 add runtime check with setting count to 0.
10364 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
10365 For unsigned types if it isn't known at compile time that
10366 the loop will iterate at least once, add runtime check to bypass
10367 the whole loop if initial condition isn't true.
10369 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
10371 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
10373 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
10375 PR middle-end/57286
10376 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
10377 transformations to avoid an infinite loop.
10379 2013-05-16 Marek Polacek <polacek@redhat.com>
10381 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
10383 2013-05-15 Leif Ekblad <leif@rdos.net>
10385 * config/i386/i386.c (ix86_decompose_address): Use
10386 DEFAULT_TLS_SEG_REG to access TLS segment register.
10387 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
10388 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
10389 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
10391 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
10394 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
10395 sibling calls to functions that would normally be lazily bound,
10396 unless $gp is call-clobbered.
10398 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
10400 * config/i386/i386.c (ix86_option_override_internal): Update
10401 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
10402 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
10403 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
10404 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
10406 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
10408 2013-05-15 Andreas Schwab <schwab@suse.de>
10410 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
10411 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
10414 2013-05-15 Teresa Johnson <tejohnson@google.com>
10416 * loop-unroll.c (report_unroll_peel): Check decision before
10417 emitting unroll/peel message.
10419 2013-05-15 Teresa Johnson <tejohnson@google.com>
10421 * function.h (has_bb_partition): New rtl_data flag.
10422 (bb_reorder_complete): Ditto.
10423 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
10424 instead of flag_reorder_blocks_and_partition.
10425 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
10426 with some enhancements.
10427 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
10428 * bb-reorder.c (connect_traces): Check for has_bb_partition
10429 instead of flag_reorder_blocks_and_partition.
10430 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
10431 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
10432 verify_hot_cold_block_grouping.
10433 (partition_hot_cold_basic_blocks): Set has_bb_partition.
10435 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10438 * config/arm/predicates.md (call_insn_operand): New predicate.
10439 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
10440 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
10441 if insn is not a tail call.
10442 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
10444 * config/arm/arm.h (enum reg_class): New caller save register class.
10445 (REG_CLASS_NAMES): Likewise.
10446 (REG_CLASS_CONTENTS): Likewise.
10447 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
10450 2013-05-15 Richard Biener <rguenther@suse.de>
10452 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
10453 of MSG_OPTIMIZED_LOCATIONS.
10454 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
10455 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
10457 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
10458 of MSG_OPTIMIZED_LOCATIONS.
10459 (execute_vect_slp): Likewise.
10460 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
10461 (vect_create_cond_for_alias_checks): Likewise.
10462 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
10463 (vect_recog_widen_mult_pattern): Likewise.
10464 (vect_recog_widen_sum_pattern): Likewise.
10465 (vect_recog_over_widening_pattern): Likewise.
10466 (vect_recog_widen_shift_pattern): Likewise.
10467 (vect_recog_vector_vector_shift_pattern): Likewise.
10468 (vect_recog_divmod_pattern): Likewise.
10469 (vect_recog_mixed_size_cond_pattern): Likewise.
10470 (vect_recog_bool_pattern): Likewise.
10471 (vect_pattern_recog_1): Likewise.
10473 2013-05-15 Martin Jambor <mjambor@suse.cz>
10475 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
10476 non-functions to builtin_unreachable.
10477 * ipa-inline-transform.c (inline_call): Do not assert estimates were
10478 correct when new direct edges were discovered.
10480 2013-05-15 Martin Jambor <mjambor@suse.cz>
10482 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
10483 header, print symbol order instead of node uid, print more information
10484 about indirect edge targets.
10485 (ipa_make_edge_direct_to_target): Print symbol order instead of node
10487 (ipa_make_edge_direct_to_target): Likewise.
10488 (remove_described_reference): Likewise.
10489 (propagate_controlled_uses): Likewise.
10490 (ipa_print_node_params): Also print symbol order.
10491 (ipcp_transform_function): Print symbol order instead of node uids.
10492 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
10493 (cgraph_get_create_real_symbol_node): Likewise.
10494 * ipa-cp.c (print_lattice): Likewise.
10495 (print_all_lattices): Likewise.
10496 (determine_versionability): Likewise.
10497 (initialize_node_lattices): Likewise.
10498 (estimate_local_effects): Likewise.
10499 (update_profiling_info): Likewise.
10500 (create_specialized_node): Likewise.
10501 (perhaps_add_new_callers): Likewise.
10502 (decide_about_value): Likewise.
10503 (decide_whether_version_node): Likewise.
10504 (identify_dead_nodes): Likewise.
10505 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
10506 (dump_inline_summary): Likewise.
10507 (estimate_node_size_and_time): Likewise.
10508 (inline_analyze_function): Likewise.
10509 * ipa-inline.c (report_inline_failed_reason): Likewise.
10510 (want_early_inline_function_p): Likewise.
10511 (edge_badness): Likewise.
10512 (update_edge_key): Likewise.
10513 (inline_small_functions): Likewise. Add dumping of order to two other
10515 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
10516 instead of node uids.
10517 (propagate_pure_const): Likewise.
10518 (propagate_pure_const): Likewise.
10519 * ipa-utils.c (dump_cgraph_node_set): Likewise.
10520 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
10521 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
10523 * tree-pretty-print.c (dump_function_header): Likewise.
10524 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
10525 Print symbol order instead of node uids.
10527 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10529 * config/s390/s390.c (s390_register_move_cost): Don't impose the
10530 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
10532 2013-05-15 Richard Biener <rguenther@suse.de>
10534 PR tree-optimization/57275
10535 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
10536 return value for fail to do runtime alias checks for gather loads.
10538 2013-05-15 Jan Hubicka <jh@suse.cz>
10542 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
10543 weakrefs are not external.
10544 (lto_symtab_merge_decls): Fix thinko when dealing with
10545 non-lto_symtab decls.
10546 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
10547 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
10548 * varpool.c (dump_varpool_node): Dump more flags.
10550 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10552 * config/i386/i386.c (processor_alias_table): Add instruction
10553 FSGSBASE for AMD bdver3 architecture.
10555 2013-05-14 Jakub Jelinek <jakub@redhat.com>
10557 * tree.c (warn_deprecated_use): Print file:line using locus color.
10558 * diagnostic.c (diagnostic_report_current_module): Print file:line
10559 and file:line:column using locus color.
10561 2013-05-14 Mike Stump <mikestump@comcast.net>
10563 * gdbinit.in: Add __null.
10565 2013-05-14 Mike Stump <mikestump@comcast.net>
10567 * recog.h: Rename struct recog_data to Recog_data.
10568 * recog.c: Likewise.
10569 * reload.c (can_reload_into): Likewise.
10570 * config/picochip/picochip.c: Likewise.
10572 2013-05-14 Mike Stump <mikestump@comcast.net>
10574 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
10576 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
10578 * resource.h (struct resources): Remove unch_memory member.
10579 (CLEAR_RESOURCE): Don't clear unch_memory.
10580 * resource.c (mark_referenced_resources): Don't set it.
10581 (mark_set_resources): Likewise.
10582 (mark_target_live_regs): Don't clear it.
10583 (init_resource_info): Likewise.
10584 * reorg.c (resource_conflicts_p): Don't compare it.
10585 (redundant_insn): Don't set it.
10587 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
10589 * emit-rtl.c (next_label): Remove unused function.
10590 (skip_consecutive_labels, link_cc0_insns): Move to ...
10591 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
10592 only place where these functions are used, and make them static.
10594 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
10596 * fold-const.c (fold_negate_expr): Handle vectors.
10597 (fold_truth_not_expr): Make it static.
10598 (fold_invert_truthvalue): New static function.
10599 (invert_truthvalue_loc): Handle vectors. Do not call
10600 fold_truth_not_expr directly.
10601 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
10602 <TRUTH_NOT_EXPR>: Do not cast to boolean.
10603 (fold_comparison): Handle vector constants.
10604 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
10605 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
10606 * tree.h (fold_truth_not_expr): Remove declaration.
10608 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
10610 * config/aarch64/aarch64-simd.md
10611 (aarch64_vcond_internal<mode>): Rename to...
10612 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
10613 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
10614 float modes. Clarify all iterator modes.
10615 (vcond<mode><mode>): Use new name for vcond expanders.
10616 (vcond<v_cmp_result><mode>): Likewise.
10617 (vcondu<mode><mode>: Likewise.
10618 * config/aarch64/iterators.md (VDQF_COND): New.
10620 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
10623 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
10624 variable for the shift amount. Check that we shift by non-negative
10627 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
10630 * config/arm/arm.h (EPILOGUE_USES): Only return true
10631 for LR_REGNUM after epilogue_completed.
10633 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
10635 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
10636 is error_mark_node.
10638 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10641 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
10642 and Solaris 11+/x86 with gld.
10643 * configure: Regenerate.
10645 2013-05-14 Jakub Jelinek <jakub@redhat.com>
10647 * expmed.c (expand_shift_1): Canonicalize rotates by
10648 constant bitsize / 2 to bitsize - 1.
10649 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
10650 case ROTATERT>: Likewise.
10653 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10655 * config/i386/i386.md (rotateinv): New code attr.
10656 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10657 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10658 roll $31, %eax, etc.
10660 2013-05-14 Richard Biener <rguenther@suse.de>
10662 PR middle-end/57235
10663 * tree-eh.c (sink_clobbers): Give up for successors with
10664 multiple predecessors and no virtual uses.
10666 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
10668 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
10669 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
10671 2013-05-14 Jakub Jelinek <jakub@redhat.com>
10673 PR middle-end/57251
10674 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
10675 the case when both op0 and op1 have VOIDmode.
10677 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10679 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
10680 in multiply-accumulate mode.
10682 2013-05-13 Guozhi Wei <carrot@google.com>
10684 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
10686 2013-05-13 Kai Tietz <ktietz@redhat.com>
10689 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
10690 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
10691 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
10692 * config/i386/i386.c (ix86_option_override_internal): Likewise.
10693 (ix86_expand_prologue): Likewise.
10694 (ix86_expand_split_stack_prologue): Likewise.
10695 (legitimate_pic_address_disp_p): Likewise.
10696 (legitimize_pic_address): Likewise.
10697 (legitimize_tls_address): Likewise.
10698 (legitimize_pe_coff_symbol): Likewise.
10699 (output_pic_addr_const): Likewise.
10700 (construct_plt_address): Likewise.
10701 (ix86_expand_call): Likewise.
10702 (x86_output_mi_thunk): Likewise.
10703 (x86_function_profiler): Likewise.
10705 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
10707 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
10708 similar switch cases.
10709 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
10710 (aarch64_simd_mov_to_<mode>low): Delete.
10711 (aarch64_simd_mov_to_<mode>high): Delete.
10712 (move_lo_quad_<mode>): Add w<-r alternative.
10713 (aarch64_simd_move_hi_quad_<mode>): Likewise.
10714 (aarch64_simd_mov_from_*): Update type attribute.
10715 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
10718 2013-05-13 Jan Hubicka <jh@suse.cz>
10720 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
10721 * config/i386/i386.c (ix86_compute_frame_layout,
10722 ix86_expand_epilogue, emit_i387_cw_initialization,
10723 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
10724 ix86_local_alignment): Fix use of size/speed predicates.
10726 2013-05-13 Jakub Jelinek <jakub@redhat.com>
10728 PR tree-optimization/45216
10729 PR tree-optimization/57157
10730 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
10731 the (-Y) & (B - 1) variant if OP is |.
10732 * expmed.c (expand_shift_1): For rotations by const0_rtx just
10733 return shifted. Use (-op1) & (prec - 1) as other_amount
10734 instead of prec - op1.
10736 2013-05-13 Martin Jambor <mjambor@suse.cz>
10738 PR middle-end/42371
10739 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
10740 (ipa_constant_data): New type.
10741 (ipa_jump_func): Use ipa_constant_data to hold information about
10742 constant jump functions.
10743 (ipa_get_jf_constant): Adjust to jump function type changes.
10744 (ipa_get_jf_constant_rdesc): New function.
10745 (ipa_param_descriptor): New field controlled_uses.
10746 (ipa_get_controlled_uses): New function.
10747 (ipa_set_controlled_uses): Likewise.
10748 * ipa-ref.h (ipa_find_reference): Declare.
10749 * ipa-prop.c (ipa_cst_ref_desc): New type.
10750 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
10752 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
10753 New parameter cs. Adjust all callers.
10754 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
10755 (remove_described_reference): New function.
10756 (jfunc_rdesc_usable): Likewise.
10757 (try_make_edge_direct_simple_call): Decrement controlled use count,
10758 attempt to remove reference if it hits zero.
10759 (combine_controlled_uses_counters): New function.
10760 (propagate_controlled_uses): Likewise.
10761 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
10762 (ipa_edge_duplication_hook): Duplicate reference descriptions.
10763 (ipa_print_node_params): Print described use counter.
10764 (ipa_write_jump_function): Adjust to jump function type changes.
10765 (ipa_read_jump_function): New parameter CS, pass it to
10766 ipa_set_jf_constant. Adjust caller.
10767 (ipa_write_node_info): Stream controlled use count
10768 (ipa_read_node_info): Likewise.
10769 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
10771 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
10772 count. Remove cloning-added reference if it reaches zero.
10773 * ipa-ref.c (ipa_find_reference): New function.
10775 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10777 * config/i386/i386.c (processor_target_table): Modified default
10778 alignment values for AMD BD and BT architectures.
10780 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
10782 * tree-vect-generic.c (uniform_vector_p): Move ...
10783 * tree.c (uniform_vector_p): ... here.
10784 * tree.h (uniform_vector_p): Declare it.
10785 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
10788 2013-05-13 Jakub Jelinek <jakub@redhat.com>
10790 PR tree-optimization/57230
10791 * tree-ssa-strlen.c (handle_char_store): Record length for
10792 array store from STRING_CST.
10794 PR tree-optimization/57230
10795 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
10798 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
10800 * config/epiphany/epiphany.c (epiphany_init): Check size of
10801 NUM_MODES_FOR_MODE_SWITCHING.
10802 (epiphany_expand_prologue):
10803 Remove CONFIG_REGNUM initial value handling code.
10804 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
10805 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
10806 (emit_set_fp_mode, epiphany_mode_after): Likewise.
10807 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
10808 Don't return 1 for FP_MODE_NONE.
10809 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
10810 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
10811 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
10812 * config/epiphany/epiphany.md (save_config): New pattern.
10814 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
10816 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
10818 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
10820 * config/i386/i386.md (memory): Handle sseishft1.
10821 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
10822 (*vec_extractv2di_1): Ditto.
10824 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
10826 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
10829 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
10831 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
10832 Add mthumb/march=armv7-a multilib.
10833 Add mthumb/march=armv7-r multilib.
10834 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
10836 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
10838 * config/v850/t-rtems: Add more multilibs.
10840 2013-05-10 Richard Biener <rguenther@suse.de>
10842 PR tree-optimization/57214
10843 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
10844 not propagate from SSA names that occur in abnormal PHI nodes.
10846 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
10848 * stor-layout.c (element_precision): New function.
10849 * machmode.h (element_precision): Declare it.
10850 * tree.c (build_minus_one_cst): New function.
10851 (element_precision): Likewise.
10852 * tree.h (build_minus_one_cst): Declare new function.
10853 (element_precision): Likewise.
10854 * fold-const.c (operand_equal_p): Use element_precision.
10855 (fold_binary_loc): Handle vector types.
10856 * convert.c (convert_to_integer): Use element_precision.
10857 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
10860 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
10862 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
10863 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
10864 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
10865 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10866 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
10867 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
10868 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
10869 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10870 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
10871 (Uuw8): New constraints.
10872 (Usb4): Move into alphabetical order.
10873 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
10874 (sd8_operand, ub8_operand, uw8_operand): New predicates.
10875 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
10876 previously unnamed patterns.
10877 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
10878 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
10879 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
10880 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
10881 of set_attr_alternative/if_then_else. Use extended_mips16 instead
10882 of specific lengths.
10884 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10886 * config/i386/i386.md (rotateinv): New code attr.
10887 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10888 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10889 roll $31, %eax, etc.
10891 PR tree-optimization/45216
10892 PR tree-optimization/57157
10893 * tree-ssa-forwprop.c (simplify_rotate): New function.
10894 (ssa_forward_propagate_and_combine): Call it.
10896 2013-05-10 Richard Biener <rguenther@suse.de>
10898 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
10899 disable peeling when we version for aliasing.
10900 (vector_alignment_reachable_p): Honor explicit user alignment.
10901 (vect_supportable_dr_alignment): Likewise.
10902 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
10903 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
10904 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
10905 then peeling to arrange for the cost-model check to come first.
10907 2013-05-10 Alan Modra <amodra@gmail.com>
10909 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
10910 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
10911 * configure: Regenerate.
10913 2013-05-10 Alan Modra <amodra@gmail.com>
10916 * varasm.c (default_elf_select_section): Move !DECL_P check..
10917 (get_named_section): ..to here before calling get_section_name.
10919 (default_section_type_flags): Add DECL_P check.
10920 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
10921 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
10923 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
10925 * config/epiphany/epiphany.c (epiphany_expand_prologue):
10926 When using gen_stack_adjust_str with a register offset, add a
10927 REG_FRAME_RELATED_EXPR note.
10929 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
10931 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
10932 (*vec_extractv4si_zext_mem): Ditto.
10933 (*vec_extractv2di): Add 0->x and x->x alternatives.
10934 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
10935 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
10937 2013-05-09 Jason Merrill <jason@redhat.com>
10939 N3639 C++1y VLA support
10940 * gimplify.c (gimplify_vla_decl): Don't touch an existing
10943 * tree.c (build_constructor_va): New.
10944 * tree.h: Declare it.
10946 2013-05-09 Martin Jambor <mjambor@suse.cz>
10949 * gimple-fold.c (canonicalize_constructor_val): Call
10950 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
10952 2013-05-09 Jan Hubicka <jh@suse.cz>
10953 Richard Biener <rguenther@suse.de>
10956 * symtab.c (symtab_make_decl_local): Do not add private names.
10958 2013-05-09 Jan Hubicka <jh@suse.cz>
10961 * symtab.c (insert_to_assembler_name_hash): Handle clones.
10962 (unlink_from_assembler_name_hash): Likewise.
10963 (symtab_prevail_in_asm_name_hash, symtab_register_node,
10964 symtab_unregister_node, symtab_initialize_asm_name_hash,
10965 change_decl_assembler_name): Update.
10967 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
10969 * config/aarch64/aarch64.md: New movtf split.
10970 (*movtf_aarch64): Update.
10971 (aarch64_movdi_tilow): Handle TF modes and rename to
10972 aarch64_movdi_<mode>low.
10973 (aarch64_movdi_tihigh): Handle TF modes and rename to
10974 aarch64_movdi_<mode>high
10975 (aarch64_movtihigh_di): Handle TF modes and rename to
10976 aarch64_mov<mode>high_di
10977 (aarch64_movtilow_di): Handle TF modes and rename to
10978 aarch64_mov<mode>low_di
10979 (aarch64_movtilow_tilow): Remove spurious whitespace.
10980 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
10982 (aarch64_print_operand): Update.
10984 2013-05-09 Alan Modra <amodra@gmail.com>
10986 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
10988 * configure: Regenerate.
10990 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10992 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
10993 splitter preparation statements.
10994 * config/i386/sse.md (*vec_extract* splitters): Ditto.
10995 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
10998 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11000 * gimple-ssa-strength-reduction.c (count_candidates): Change
11001 return value to int.
11002 (analyze_candidates_and_replace): Change type of length to int.
11004 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
11006 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
11007 (*vec_extract<mode>): Use VI12_128 mode iterator.
11008 (*vec_extract<mode>_mem): Ditto.
11009 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
11012 2013-05-08 Diego Novillo <dnovillo@google.com>
11017 2012-08-17 Diego Novillo <dnovillo@google.com>
11020 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
11021 * config.in: Regenerate.
11022 * configure: Regenerate.
11023 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
11025 2013-05-08 Jan Hubicka <jh@suse.cz>
11028 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
11029 * cgraph.h (symtab_node_base): Add unique_name.
11030 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11031 input_overwrite_node, input_varpool_node): Stream unique_name.
11032 * cgraphclones.c (cgraph_create_virtual_clone,
11033 cgraph_function_versioning): Set unique_name.
11034 * ipa.c (function_and_variable_visibility): Set unique_name.
11036 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11038 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
11039 (alloc_cand_and_find_basis): Restrict conditional candidate
11040 processing to CAND_MULTs.
11042 2013-05-08 Jan Hubicka <jh@suse.cz>
11045 lto-symtab.c (lto_symtab_symbol_p): New function.
11046 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
11047 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
11048 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
11049 Skip static symbols.
11051 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
11053 PR tree-optimization/57200
11054 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
11055 Only call inform if the preceding warning_at returns true.
11057 2013-05-07 Han Shen <shenhan@google.com>
11059 * cfgexpand.c (record_or_union_type_has_array_p): New function.
11060 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
11061 * common.opt (fstack-protector-strong): New option.
11062 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
11063 * doc/invoke.texi (Optimization Options): Document
11064 "-fstack-protector-strong".
11065 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
11067 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11069 * config/mips/mips.c (mips_machine_reorg2): Return 0.
11071 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
11073 * ira.c (update_equiv_regs): Add insn having equiv memory even if
11074 it is not lhs of the insn.
11075 (setup_reg_equiv): Remove insn having equiv memory which it is not
11077 * lra-constraints.c (process_address): Try to improve generation
11078 code for address base + disp.
11079 (lra_constraints): Make correct the code for checking insn setting
11080 up backward equivalence. Remove insn only if it is in the init
11082 * lra-eliminations.c (update_reg_eliminate): Change return value.
11083 (lra_eliminate): Use the result.
11085 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
11087 * config/i386/sse.md (ssescalarnummask): New mode attribute.
11088 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
11089 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
11090 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
11091 register target operands.
11092 (*vec_extractv8hi_sse2): New pattern.
11093 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
11094 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
11095 (*vec_extract<mode>_mem): New insn and split pattern.
11097 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
11099 * config/arm/arm.c (arm_asan_shadow_offset): New function.
11100 (TARGET_ASAN_SHADOW_OFFSET): Define.
11101 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
11102 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
11104 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11106 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
11107 (incr_vec_index): Return -1 if increment not found.
11108 (create_add_on_incoming_edge): Assert if increment not found.
11109 (record_increment): Limit number of increments recorded.
11110 (all_phi_incrs_profitable): Return false if an increment not found.
11111 (replace_profitable_candidates): Don't process increments that were
11113 (analyze_candidates_and_replace): Limit size of incr_vec.
11115 2013-05-07 Richard Biener <rguenther@suse.de>
11117 * calls.c (special_function_p): setjmp-like functions are leaf.
11118 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
11119 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
11121 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
11123 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
11124 (aarch64_simd_mov<mode>): New expander.
11125 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
11126 (aarch64_simd_mov_to_<mode>high): Likewise.
11127 (aarch64_simd_mov_from_<mode>low): Likewise.
11128 (aarch64_simd_mov_from_<mode>high): Likewise.
11129 (aarch64_dup_lane<mode>): Update.
11130 (aarch64_dup_lanedi): New instruction pattern.
11131 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
11132 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
11134 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11136 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
11137 (replace_mult_candidate): Remove unnecessary argument; remove
11138 unnecessary parameter from call to introduce_cast_before_cand.
11139 (replace_unconditional_candidate): Remove unnecessary parameter
11140 from call to replace_mult_candidate.
11141 (replace_conditional_candidate): Likewise.
11142 (insert_initializers): Use make_temp_ssa_name.
11143 (introduce_cast_before_cand): Remove unnecessary argument; use
11144 make_temp_ssa_name.
11145 (replace_one_candidate): Remove unnecessary argument; remove
11146 unnecessary parameter from calls to introduce_cast_before_cand.
11147 (replace_profitable_candidates): Remove unnecessary parameters
11148 from calls to replace_one_candidate.
11150 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11152 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
11153 phi def as possibly hiding a basis for a CAND_ADD whose operands
11154 have been commuted in the analysis.
11155 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
11157 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11159 * config/aarch64/aarch64.md
11160 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
11161 shift value between 0-4.
11163 2013-05-07 Richard Biener <rguenther@suse.de>
11165 * double-int.h (rshift): New overload.
11166 * double-int.c (rshift): New function.
11167 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
11168 (create_reference_ops_from_ref): Remove.
11169 (vn_reference_insert): Use shared ops for constructing the
11170 reference and copy it.
11172 2013-05-07 Richard Biener <rguenther@suse.de>
11174 PR middle-end/57190
11175 * tree-eh.c (sink_clobbers): Properly propagate
11176 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
11178 2013-05-07 Jakub Jelinek <jakub@redhat.com>
11180 PR tree-optimization/57149
11181 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
11182 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
11183 collect_phi_def_edges, execute_late_warn_uninitialized): Use
11184 uninit_undefined_value_p instead of ssa_undefined_value_p.
11187 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
11188 for modifier == EXPAND_INITIALIZER.
11190 2013-05-07 Anton Blanchard <anton@samba.org>
11192 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
11193 for powerpc64 little endian.
11194 * configure: Regenerate.
11196 2013-05-06 Graham Stott <grahams@btinternet.com>
11198 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
11199 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
11200 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
11201 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
11203 2013-05-06 Graham Stott <grahams@btinternet.com>
11205 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
11206 codes which allow non-lvalues.
11208 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
11210 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
11211 components are all 1s.
11212 (integer_minus_onep): New function.
11213 * tree.h (integer_minus_onep): Declare it.
11214 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
11215 integer_minus_onep instead of integer_all_onesp.
11217 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11220 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
11221 variations of these patterns.
11223 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
11225 * config/i386/i386.md (isa): Add x64_sse4 member.
11226 (enabled): Handle x64_sse4.
11227 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
11228 instruction for 64bit SSE4_1 targets. Update insn attributes.
11229 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
11230 instruction for SSE4_1 targets. Update insn attributes.
11231 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11232 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
11233 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
11235 (*vec_extractv4si): Rename from *sse4_1_pextrd.
11236 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
11237 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
11239 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11242 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
11244 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
11246 * final.c (do_assembler_dialects): Don't handle curly braces and
11247 vertical bar escaped by % as dialect delimiters.
11248 (output_asm_insn): Print curly braces and vertical bar if escaped
11249 by % and ASSEMBLER_DIALECT defined.
11250 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
11251 * doc/tm.texi: Regenerated.
11253 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11255 * config/mips/mips.c: Include tree-pass.h.
11256 (mips_reorg): Split in pre- and post-dbr_schedule parts.
11257 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
11258 (pass_mips_machine_reorg2): New machine specific pass.
11259 (insert_pass_mips_machine_reorg2): New pass plugin definition.
11260 (mips_option_override): Register the new pass.
11261 * rtl.h (cleanup_barriers): Remove prototype.
11262 (dbr_schedule): Likewise.
11263 * jump.c (cleanup_barriers): Make static.
11264 * reorg.c (dbr_schedule): Likewise.
11266 2013-05-06 Richard Biener <rguenther@suse.de>
11268 PR tree-optimization/57185
11269 * tree-parloops.c (add_field_for_reduction): Handle anonymous
11270 SSA names properly.
11272 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
11275 * config/i386/i386.c (add_parameter_dependencies): Add dependence
11276 between "first_arg" and "insn", not "last" and "insn".
11278 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11280 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
11281 (find_candidates_in_block): Re-enable slsr_process_phi.
11282 (create_phi_basis): Fix double counting of candidate adjustment.
11284 2013-05-06 Richard Biener <rguenther@suse.de>
11286 PR middle-end/57147
11287 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
11288 the edge is also fallthru, preserve it and just clear the
11290 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
11291 also complex, preserve that and just clear the fallthru flag.
11292 * tree-inline.c (update_ssa_across_abnormal_edges): Also
11293 update virtual operands.
11295 2013-05-06 Alan Modra <amodra@gmail.com>
11297 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
11298 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
11299 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
11300 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
11301 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11302 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
11304 2013-05-06 Alan Modra <amodra@gmail.com>
11306 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
11307 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
11308 (DEFAULT_ASM_ENDIAN): Define.
11309 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11310 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
11311 Update -K PIC clause from sysv4.h.
11312 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
11313 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
11315 2013-05-06 Alan Modra <amodra@gmail.com>
11317 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
11318 twice for little-endian.
11319 (ashrdi3_no_power, ashrdi3): Support little-endian.
11321 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11324 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
11325 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
11327 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
11328 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
11329 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
11332 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
11333 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11335 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
11336 * config/pa/pa.opt: Make mbig-switch a no-op.
11337 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
11338 (CASE_VECTOR_MODE): Always return SImode.
11339 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
11340 for the !TARGET_BIG_SWITCH case.
11341 * config/pa/pa-linux.h: Likewise.
11342 * config/pa/pa-openbsd.h: Likewise.
11343 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
11344 * config/pa/pa.md (short_jump): Remove define_insn.
11345 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
11346 (casesi0): Remove define_insn.
11347 (type): Remove btable_branch.
11348 (pa_combine_type): Likewise.
11349 (in_nullified_branch_delay): Likewise.
11350 (in_call_delay): Likewise.
11351 (define_delay): Likewise.
11352 (define_insn_reservation "Z3"): Likewise.
11353 (define_insn_reservation "Z4"): Likewise.
11354 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
11355 (pa_adjust_insn_length): Remove adjustment for btable branches.
11356 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
11359 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
11361 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11362 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
11363 Add m->r,x alternatives.
11364 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
11365 splitters using SWI48x mode iterator.
11366 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
11367 TARGET_64BIT. Add m->x alternative.
11368 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
11369 Add o->x alternative. Enable for TARGET_SSE.
11370 (sse_storeq): Remove expander.
11371 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
11372 with memory input operand.
11373 (*vec_extractv2di_1 splitter): New.
11374 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
11375 * config/i386/i386.md (ssevecmodelower): New mode attribute.
11377 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11379 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
11380 (INT_LOWPART): Delete.
11381 (extract_MB): Adjust.
11382 (extract_ME): Adjust.
11383 (print_operand): Adjust.
11385 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11387 * config/rs6000/predicates.md (reg_or_add_cint_operand,
11388 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11389 (reg_or_logical_cint_operand, easy_fp_constant,
11390 logical_const_operand): Delete "CONST_DOUBLE" case.
11391 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
11392 "HOST_BITS_PER_WIDE_INT == 64" test.
11393 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
11394 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
11395 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
11396 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
11398 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
11399 CONST_DOUBLE DImode/VOIDmode case.
11400 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
11401 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11402 CONST_DOUBLE VOIDmode case.
11403 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11404 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
11405 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
11406 Delete CONST_DOUBLE case.
11407 (splitters for mov FMOVE64 const_double): Delete
11408 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11409 "HOST_BITS_PER_WIDE_INT >= 64" test.
11410 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
11412 (mov DI const_double): Delete.
11414 2013-05-04 Jakub Jelinek <jakub@redhat.com>
11416 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
11417 on op shows all bits zero in mode of a lowpart subreg, return zero.
11419 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
11422 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
11423 to save TFmode registers and DImode to save TImode registers for
11424 caller save operations.
11425 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
11426 mark being partially clobbered since they only use the first
11429 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
11430 and TDmode only use the upper 64-bits of each VSX register.
11432 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11434 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
11435 (find_candidates_in_block): Disable slsr_process_phi.
11437 2013-05-03 Guozhi Wei <carrot@google.com>
11439 * coverage.c (coverage_obj_init): Move the construction of gcov
11441 (build_init_ctor): ... here.
11443 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11445 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
11446 (slsr_cand_d): Redefine def_phi.
11447 (stride_status, phi_adjust_status, count_phis_status): New enums.
11448 (find_phi_def): New.
11449 (find_basis_for_base_expr): New.
11450 (find_basis_for_candidate): Handle hidden bases.
11451 (alloc_cand_and_find_basis): Handle phi candidates.
11452 (slsr_process_phi): New.
11453 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
11454 (create_mul_imm_cand): Likewise.
11455 (create_add_ssa_cand): Exclude phi base candidates.
11456 (create_add_imm_cand): Likewise.
11457 (slsr_process_cast): Likewise.
11458 (slsr_process_copy): Likewise.
11459 (find_candidates_in_block): Handle phi candidates.
11460 (dump_candidate): Likewise.
11461 (unconditional_cands): Delete.
11462 (unconditional_cands_with_known_stride_p): Delete.
11463 (phi_dependent_cand_p): New.
11464 (cand_increment): Handle phi-dependent candidates.
11465 (replace_dependent): Delete.
11466 (replace_mult_candidate): New.
11467 (replace_unconditional_candidate): New.
11468 (incr_vec_index): Move to avoid forward reference.
11469 (create_add_on_incoming_edge): New.
11470 (create_phi_basis): New.
11471 (replace_dependents): Delete.
11472 (replace_conditional_candidate): New.
11473 (phi_add_costs): New.
11474 (replace_uncond_cands_and_profitable_phis): New.
11475 (record_increment): Handle phi adjustments.
11476 (record_phi_increments): New.
11477 (record_increments): Handle phi adjustments.
11478 (phi_incr_cost): New.
11479 (lowest_cost_path): Handle phis.
11480 (total_savings): Likewise.
11481 (analyze_increments): Likewise.
11482 (ncd_with_phi): New.
11483 (ncd_of_cand_and_phis): New.
11484 (nearest_common_dominator_for_cands): Handle phi increments.
11485 (all_phi_incrs_profitable): New.
11486 (replace_profitable_candidates): Handle phi-dependent candidates.
11487 (analyze_candidates_and_replace): Likewise.
11489 2013-05-03 Teresa Johnson <tejohnson@google.com>
11492 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
11493 do not exceed REG_BR_PROB_BASE.
11495 2013-05-03 Jeff Law <law@redhat.com>
11497 PR tree-optimization/57144
11498 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
11499 operand of the condition will bit into the new type when eliminating
11500 a cast feeding a condition.
11502 2013-05-03 Jakub Jelinek <jakub@redhat.com>
11504 PR rtl-optimization/57130
11505 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
11506 of COMPARE as in_code to the recursive call if needed.
11508 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
11510 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
11511 (enabled): Handle new members.
11512 * config/i386/sse.md (*vec_concatv2si): Merge from
11513 *vec_concatv2si_sse2 and vec_concatv2si_sse.
11514 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
11516 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
11518 PR tree-optimization/57027
11519 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
11520 for fnms opportunity, check we got the prerequisite kind
11521 of tree / gimple before using accessor functions.
11523 2013-05-03 Richard Biener <rguenther@suse.de>
11525 * double-int.h (lshift): New overload without precision
11526 and arith argument.
11527 (operator *=, operator +=, operator -=): Move ...
11528 * double-int.c (operator *=, operator +=, operator -=): ... here
11529 and implement more efficiently.
11530 (mul_double_with_sign): Remove.
11531 (lshift_double): Adjust to take unsinged shift argument, push
11532 dispatching code to callers.
11533 (mul_double_wide_with_sign): Add early out for callers that
11534 are not interested in high parts or overflow.
11535 (lshift): New function.
11536 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
11537 dispatch code here.
11538 (lrotate, rrotate): Use logical shifts.
11539 * expr.c (get_inner_reference): Use lshift.
11540 * fixed-value.c (do_fixed_divide): Likewise.
11541 * tree-dfa.c (get_ref_base_and_extent): Likewise.
11542 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
11543 (indirect_refs_may_alias_p): Likewise.
11544 (stmt_kills_ref_p_1): Likewise.
11546 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
11548 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
11550 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
11552 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
11553 scalar form of FABD instruction.
11555 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
11557 * lra-constraints.c (process_alt_operands): Add checking alt
11558 number to choose the best alternative.
11560 2013-05-02 Richard Biener <rguenther@suse.de>
11562 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
11563 bitmap and its handling.
11564 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
11566 2013-05-02 Richard Biener <rguenther@suse.de>
11568 PR middle-end/57140
11569 * tree-inline.c (copy_loops): Properly handle removed loops.
11570 (copy_cfg_body): Mark destination loops for fixup if source
11571 loops needed fixup.
11573 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
11576 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
11577 generating simple_return for naked functions.
11579 2013-05-02 Martin Jambor <mjambor@suse.cz>
11581 PR middle-end/56988
11582 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
11583 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
11585 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
11586 ipa_agg_replacement_value structures.
11587 (known_aggs_to_agg_replacement_list): Likewise.
11588 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
11589 (read_agg_replacement_chain): Likewise.
11590 (ipcp_transform_function): Also check that by_ref flags match.
11592 2013-05-02 Richard Biener <rguenther@suse.de>
11594 * graphds.h (struct graph): Add obstack member.
11595 * graphds.c (new_graph): Initialize obstack and allocate
11597 (add_edge): Allocate edge from the obstack.
11598 (free_graph): Free the obstack instead of all edges and vertices.
11600 2013-05-02 Teresa Johnson <tejohnson@google.com>
11602 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
11604 * cfg.c (update_bb_profile_for_threading): Ditto.
11605 * tree-inline.c (copy_bb): Ditto.
11606 (copy_edges_for_bb): Ditto.
11607 (initialize_cfun): Ditto.
11608 (copy_cfg_body): Ditto.
11609 (expand_call_inline): Ditto.
11610 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
11611 (estimate_node_size_and_time): Ditto.
11612 (inline_merge_summary): Ditto.
11613 * cgraphclones.c (cgraph_clone_edge): Ditto.
11614 (cgraph_clone_node): Ditto.
11615 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11616 (compute_trg_info): Ditto.
11618 2013-05-02 Ian Bolton <ian.bolton@arm.com>
11620 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
11621 S reg when fp attribute set.
11622 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
11624 2013-05-02 Ian Bolton <ian.bolton@arm.com>
11626 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
11628 (*and_one_cmplsi3_compare0_uxtw): Likewise.
11629 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
11630 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
11632 2013-05-02 Richard Biener <rguenther@suse.de>
11634 * tree-scalar-evolution.c (scev_info_hasher): Remove.
11635 (struct instantiate_cache_entry): New type.
11636 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
11637 (struct instantiate_cache_type): New type.
11638 (set_instantiated_value, get_instantiated_value): Remove.
11639 (get_instantiated_value_entry): New function.
11640 (instantiate_scev_name): Use the new cache and adjust.
11641 (instantiate_scev_poly): Adjust.
11642 (instantiate_scev_binary): Likewise.
11643 (instantiate_array_ref): Likewise.
11644 (instantiate_scev_convert): Likewise.
11645 (instantiate_scev_not): Likewise.
11646 (instantiate_scev_3): Likewise.
11647 (instantiate_scev_2): Likewise.
11648 (instantiate_scev_r): Likewise.
11649 (instantiate_scev): Likewise.
11650 (resolve_mixers): Likewise.
11652 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
11655 * lra-constraints.c (best_small_class_operands_num): Remove.
11656 (process_alt_operands): Remove small_class_operands_num. Take
11657 small classes operands into losers and only if the operand is not
11658 matched. Modify debugging output.
11659 (curr_insn_transform): Remove best_small_class_operands_num.
11662 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11664 * config/aarch64/aarch64-builtins.c
11665 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
11666 * config/aarch64/aarch64-simd-builtins.def
11667 (reduc_splus_): Add new modes.
11668 (reduc_uplus_): New.
11669 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
11670 (reduc_uplus_v4sf): Likewise.
11671 (reduc_splus_v4sf): Likewise.
11672 (aarch64_addv<mode>): Likewise.
11673 (reduc_uplus_<mode>): Likewise.
11674 (reduc_splus_<mode>): Likewise.
11675 (aarch64_addvv2di): Likewise.
11676 (reduc_uplus_v2di): Likewise.
11677 (reduc_splus_v2di): Likewise.
11678 (aarch64_addvv2si): Likewise.
11679 (reduc_uplus_v2si): Likewise.
11680 (reduc_splus_v2si): Likewise.
11681 (reduc_<sur>plus_<mode>): New.
11682 (reduc_<sur>plus_v2di): Likewise.
11683 (reduc_<sur>plus_v2si): Likewise.
11684 (reduc_<sur>plus_v4sf): Likewise.
11685 (aarch64_addpv4sf): Likewise.
11686 * config/aarch64/arm_neon.h
11687 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
11688 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
11689 add UNSPEC_SADDV, UNSPEC_UADDV.
11691 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
11693 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11695 * config/aarch64/arm_neon.h
11696 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
11698 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11700 * config/aarch64/aarch64-builtins
11701 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
11703 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11705 * config/aarch64/aarch64-simd-builtins.def
11706 (reduc_smax_): New.
11707 (reduc_smin_): Likewise.
11708 (reduc_umax_): Likewise.
11709 (reduc_umin_): Likewise.
11710 (reduc_smax_nan_): Likewise.
11711 (reduc_smin_nan_): Likewise.
11714 (smax): Update for V2SF, V4SF and V2DF modes.
11717 (smin_nan): Likewise.
11718 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
11719 (<su><maxmin><mode>3): ...This, refactor.
11720 (s<maxmin><mode>3): New.
11721 (<maxmin_uns><mode>3): Likewise.
11722 (reduc_<maxmin_uns>_<mode>): Refactor.
11723 (reduc_<maxmin_uns>_v4sf): Likewise.
11724 (reduc_<maxmin_uns>_v2si): Likewise.
11725 (aarch64_<fmaxmin><mode>: Remove.
11726 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
11728 (vmin<q>_f<32,64>): Likewise.
11729 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
11731 (su): Add mappings for smax, smin, umax, umin.
11733 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
11734 (FMAXMIN): Rename as...
11735 (FMAXMIN_UNS): ...This.
11737 (fmaxminv): Likewise.
11738 (fmaxmin): Likewise.
11740 (maxmin_uns_op): Likewise.
11742 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11744 * config/aarch64/arm_neon.h
11745 (vac<ge, gt><sd>_f<32, 64>): Rename to...
11746 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
11747 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
11749 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11751 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
11752 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
11754 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11756 * config/aarch64/aarch64-simd.md
11757 (vcond<mode>_internal): Handle special cases for constant masks.
11758 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
11759 (vcondu<mode><mode>): Likewise.
11760 (vcond<v_cmp_result><mode>): New.
11762 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11764 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
11765 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
11766 * config/aarch64/aarch64-simd-builtins.def
11767 (cmeq): Update to BUILTIN_VALLDI.
11772 * config/aarch64/arm_neon.h
11773 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
11774 to builtins or C as appropriate.
11776 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11778 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
11780 (cmhi): Rename to...
11782 * config/aarch64/aarch64-simd.md
11783 (simd_mode): Add SF.
11784 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
11785 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
11786 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
11787 (cstore<mode>_neg): ...This.
11788 * config/aarch64/iterators.md
11790 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
11791 (COMPARISONS): New.
11792 (UCOMPARISONS): Likewise.
11793 (optab): Add missing comparisons.
11799 (VCMP_S): Likewise.
11800 (VCMP_U): Likewise.
11801 (V_cmp_result): Add DF, SF modes.
11802 (v_cmp_result): Likewise.
11804 (vmtype): Likewise.
11805 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
11807 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
11809 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
11810 define_insn to define_insn_and_split.
11811 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
11812 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
11813 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
11814 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
11815 (thumb2_negscc): Likewise.
11817 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11819 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
11821 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11823 * config/arm/thumb2.md: Remove trailing whitespaces.
11825 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11827 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
11828 Use gen_int_mode rather than GEN_INT.
11830 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
11832 * value-prof.c (stream_in_histogram_value): Remove the strayed
11835 2013-04-30 Richard Biener <rguenther@suse.de>
11837 PR middle-end/57122
11838 * cfghooks.c (split_edge): Properly check for the loop latch edge.
11840 2013-04-30 Richard Biener <rguenther@suse.de>
11842 PR middle-end/57107
11843 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
11845 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
11847 PR rtl-optimization/56957
11848 PR rtl-optimization/57105
11849 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
11850 variable. Use just INSN_UID for determining whether an insn
11851 should be only disconnected from the insn stream.
11852 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
11854 2013-04-30 Jakub Jelinek <jakub@redhat.com>
11856 PR tree-optimization/57104
11857 * tsan.c (instrument_expr): Don't instrument accesses to
11858 DECL_HARD_REGISTER VAR_DECLs.
11860 2013-04-30 Richard Biener <rguenther@suse.de>
11862 * function.h (loops_for_fn): New inline function.
11863 (set_loops_for_fn): Likewise.
11864 * cfgloop.h (place_new_loop): Add struct function parameter.
11865 (get_loop): Likewise.
11866 (get_loops): Likewise.
11867 (number_of_loops): Likewise.
11868 (fel_next): Adjust.
11869 (fel_init): Likewise.
11870 * cfg.c (get_loop_copy): Adjust.
11871 * cfgloop.c (flow_loops_dump): Likewise.
11872 (record_loop_exits): Likewise.
11873 (verify_loop_structure): Likewise.
11874 * cfgloopanal.c (mark_irreducible_loops): Likewise.
11875 (estimate_reg_pressure_cost): Likewise.
11876 (mark_loop_exit_edges): Likewise.
11877 * cfgloopmanip.c (place_new_loop): Likewise.
11878 (add_loop): Likewise.
11879 (duplicate_loop): Likewise.
11880 * graph.c (draw_cfg_nodes): Likewise.
11881 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
11882 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
11883 (extract_affine_chrec): Likewise.
11884 (build_scop_iteration_domain): Likewise.
11885 * graphite.c (graphite_initialize): Likewise.
11886 * ira-build.c (create_loop_tree_nodes): Likewise.
11887 (more_one_region_p): Likewise.
11888 (rebuild_regno_allocno_maps): Likewise.
11889 (mark_loops_for_removal): Likewise.
11890 (mark_all_loops_for_removal): Likewise.
11891 (remove_unnecessary_regions): Likewise.
11892 (ira_build): Likewise.
11893 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
11894 * loop-init.c (fix_loop_structure): Likewise.
11895 (gate_rtl_move_loop_invariants): Likewise.
11896 (gate_rtl_unswitch): Likewise.
11897 (gate_rtl_unroll_and_peel_loops): Likewise.
11898 (rtl_doloop): Likewise.
11899 * lto-streamer-in.c (input_cfg): Likewise.
11900 * lto-streamer-out.c (output_cfg): Likewise.
11901 * modulo-sched.c (sms_schedule): Likewise.
11902 * predict.c (tree_estimate_probability): Likewise.
11903 (tree_estimate_probability_driver): Likewise.
11904 (estimate_loops): Likewise.
11905 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
11906 (move_sese_region_to_fn): Likewise.
11907 (debug_loop_num): Likewise.
11908 * tree-chrec.c (chrec_evaluate): Likewise.
11909 (hide_evolution_in_other_loops_than_loop): Likewise.
11910 (chrec_component_in_loop_num): Likewise.
11911 (reset_evolution_in_loop): Likewise.
11912 (evolution_function_is_invariant_rec_p): Likewise.
11913 * tree-if-conv.c (main_tree_if_conversion): Likewise.
11914 * tree-inline.c (copy_loops): Likewise.
11915 (copy_cfg_body): Likewise.
11916 (tree_function_versioning): Likewise.
11917 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
11918 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
11920 (add_to_evolution_1): Likewise.
11921 (scev_const_prop): Likewise.
11922 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
11923 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
11924 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
11925 (tree_ssa_lim_initialize): Likewise.
11926 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
11927 (verify_loop_closed_ssa): Likewise.
11928 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
11929 (tree_ssa_loop_im): Likewise.
11930 (tree_ssa_loop_unswitch): Likewise.
11931 (tree_vectorize): Likewise.
11932 (check_data_deps): Likewise.
11933 (tree_ssa_loop_ivcanon): Likewise.
11934 (tree_ssa_loop_bounds): Likewise.
11935 (tree_complete_unroll): Likewise.
11936 (tree_complete_unroll_inner): Likewise.
11937 (tree_parallelize_loops): Likewise.
11938 (tree_ssa_loop_prefetch): Likewise.
11939 (tree_ssa_loop_ivopts): Likewise.
11940 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
11941 * tree-vectorizer.c (vectorize_loops): Likewise.
11943 2013-04-29 Mike Frysinger <vapier@gentoo.org>
11945 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
11946 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
11947 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
11948 with EABI_LINK_SPEC.
11950 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11953 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
11956 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
11959 * lra-constraints.c (process_alt_operands): Discourage a bit more
11960 using memory for pseudos. Print cost dump for alternatives.
11961 Modify cost values for conflicts with early clobbers.
11962 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
11964 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11967 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
11969 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11971 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
11972 from/to S register.
11973 (movdi_aarch64): Support LDR/STR from/to D register.
11975 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11977 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
11978 or higher by default.
11980 2013-04-29 Richard Biener <rguenther@suse.de>
11982 PR middle-end/57075
11983 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
11984 even if not adding abnormal edges for calls that can make
11987 2013-04-29 Richard Biener <rguenther@suse.de>
11989 PR middle-end/57103
11990 * tree-cfg.c (move_stmt_op): Fix condition under which to update
11992 (move_stmt_r): Remove redundant checking.
11994 2013-04-29 Teresa Johnson <tejohnson@google.com>
11997 * basic-block.h (apply_scale): New function.
11998 (apply_probability): Use apply_scale.
11999 * gimple-streamer-in.c (input_bb): Ditto.
12000 * lto-streamer-in.c (input_cfg): Ditto.
12001 * lto-cgraph.c (merge_profile_summaries): Ditto.
12002 * tree-optimize.c (execute_fixup_cfg): Ditto.
12003 * tree-inline.c (copy_bb): Update comment to use apply_scale.
12004 (copy_edges_for_bb): Ditto.
12005 (copy_cfg_body): Ditto.
12007 2013-04-29 Tom de Vries <tom@codesourcery.com>
12009 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
12010 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
12011 (tail_merge_optimize): Handle current_loops == NULL.
12013 2013-04-26 Jeff Law <law@redhat.com>
12015 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
12016 (simplify_cond_using_ranges): Generalize code to simplify
12017 COND_EXPRs where one argument is a constant and the other
12018 is an SSA_NAME created by an integral type conversion.
12020 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12022 * config/arm/arm.md (store_minmaxsi): Use only when
12023 optimize_insn_for_size_p.
12025 2013-04-29 Christian Bruel <christian.bruel@st.com>
12028 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
12030 2013-04-29 Richard Biener <rguenther@suse.de>
12032 PR middle-end/57089
12033 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
12034 loop tree make sure to schedule a fixup for the child as well.
12035 (expand_omp_for_generic): Properly add loops.
12036 (expand_omp_for_static_nochunk): Likewise.
12037 (expand_omp_for_static_chunk): Likewise.
12038 (expand_omp_for): For the degenerate case fixup loops.
12039 (expand_omp_sections): Fix default bb placement in loops.
12040 (expand_omp_atomic_pipeline): Properly add loops.
12042 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12044 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
12046 2013-04-29 Tom de Vries <tom@codesourcery.com>
12048 * tree-ssa-tail-merge.c: Update header comment.
12050 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12052 * config/aarch64/arm_neon.h
12053 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
12054 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
12055 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
12056 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
12057 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
12058 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
12059 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
12060 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
12062 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12064 * config/aarch64/aarch64-simd.md
12065 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
12066 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
12067 fix_trunc, fixuns_trunc.
12068 (ftrunc<VDQF:mode>2): New.
12069 * config/aarch64/iterators.md (optab): Add fix, fixuns.
12070 (fix_trunc_optab): New.
12072 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12074 * config/aarch64/aarch64-builtins.c
12075 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
12076 iceilf, lround, iroundf.
12078 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
12081 * config/i386/i386.h (enum ix86_tune_indices)
12082 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
12083 New, split from X86_TUNE_INTER_UNIT_MOVES.
12084 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
12085 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
12086 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
12087 (TARGET_INTER_UNIT_MOVES): Remove.
12088 * config/i386/i386.c (initial_ix86_tune_features): Update.
12089 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
12090 (ix86_expand_convert_uns_didf_sse): Use
12091 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12092 (ix86_expand_vector_init_one_nonzero): Ditto.
12093 (ix86_expand_vector_init_interleave): Ditto.
12094 (inline_secondary_memory_needed): Return true for moves from SSE class
12095 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
12096 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
12097 * config/i386/constraints.md (Yi, Ym): Depend on
12098 TARGET_INTER_UNIT_MOVES_TO_VEC.
12099 (Yj, Yn): New constraints.
12100 * config/i386/i386.md (*movdi_internal): Change constraints of
12101 operand 1 from Yi to Yj and from Ym to Yn.
12102 (*movsi_internal): Ditto.
12103 (*movdf_internal): Ditto.
12104 (*movsf_internal): Ditto.
12105 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
12106 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12107 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
12108 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
12109 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
12110 * config/i386/sse.md (movdi_to_sse): Ditto.
12111 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
12112 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
12113 TARGET_INTER_UNIT_MOVES.
12114 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
12115 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
12116 instead of TARGET_INTER_UNIT_MOVES.
12117 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
12118 operand 1 from Yi to Yj and from Ym to Yn.
12120 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12122 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
12123 (float_truncate_hi_): Likewise.
12124 (float_extend_lo_): Likewise.
12125 (float_truncate_lo_): Likewise.
12126 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
12127 (aarch64_float_extend_lo_v2df): Likewise.
12128 (vec_unpacks_hi_v4sf): Likewise.
12129 (aarch64_float_truncate_lo_v2sf): Likewise.
12130 (aarch64_float_truncate_hi_v4sf): Likewise.
12131 (vec_pack_trunc_v2df): Likewise.
12132 (vec_pack_trunc_df): Likewise.
12134 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12136 * config/aarch64/aarch64-builtins.c
12137 (aarch64_fold_builtin): Fold float conversions.
12138 * config/aarch64/aarch64-simd-builtins.def
12139 (floatv2si, floatv4si, floatv2di): New.
12140 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
12141 * config/aarch64/aarch64-simd.md
12142 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
12143 * config/aarch64/iterators.md (FLOATUORS): New.
12144 (optab): Add float, floatuns.
12145 (su_optab): Likewise.
12147 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12149 * config/aarch64/aarch64-builtins.c
12150 (aarch64_builtin_vectorized_function): Use new names for
12152 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
12153 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
12154 (fcvtzu): Split as...
12155 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
12156 (fcvtas): Split as...
12157 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
12158 (fcvtau): Split as...
12159 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
12160 (fcvtps): Split as...
12161 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
12162 (fcvtpu): Split as...
12163 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
12164 (fcvtms): Split as...
12165 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
12166 (fcvtmu): Split as...
12167 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
12168 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
12169 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
12170 (lfrintnusf, lfrintnudf): Likewise.
12171 * config/aarch64/aarch64-simd.md
12172 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
12174 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
12175 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
12176 (fcvt_pattern): Likewise.
12178 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12180 * config/aarch64/aarch64-simd.md
12181 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
12182 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
12184 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12186 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
12187 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
12188 (vrnd<a,m,n,p>_f32): Implement using builtins.
12189 (vrnd<i,x><q>_f<32, 64>): New.
12191 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12193 * config/aarch64/aarch64-builtins.c
12194 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
12195 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
12196 (frintz): Rename to...
12198 (frintp): Rename to...
12200 (frintm): Rename to...
12202 (frinti): Rename to...
12203 (nearbyint): ...this.
12204 (frintx): Rename to...
12206 (frinta): Rename to...
12208 * config/aarch64/aarch64-simd.md
12209 (aarch64_frint<frint_suffix><mode>): Delete.
12210 (<frint_pattern><mode>2): Convert to insn.
12211 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
12212 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
12213 (frint_pattern): Likewise.
12214 (frint_suffix): Likewise.
12216 2013-04-29 Richard Biener <rguenther@suse.de>
12218 PR tree-optimization/57081
12219 * loop-init.c: Include tree-flow.h.
12220 (loop_optimizer_finalize): Free number of iteration estimates.
12221 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
12223 2013-04-29 Jakub Jelinek <jakub@redhat.com>
12225 PR tree-optimization/57083
12226 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
12227 non-singleton shift count range, zero extend low_bound for uns case.
12229 * config/i386/predicates.md (general_vector_operand): New predicate.
12230 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
12231 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
12232 if they aren't nonimmediate operands. If their original values
12233 satisfy const_vector_equal_evenodd_p, don't shift them.
12234 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
12235 predicates. For the SSE4.1 case force operands[{1,2}] into registers
12236 if not nonimmediate_operand.
12237 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
12238 instead of register_operand.
12239 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
12241 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
12243 * stor-layout.c (finalize_size_functions): Allocate a structure and
12244 reset cfun before dumping the functions.
12246 2013-04-27 Jakub Jelinek <jakub@redhat.com>
12248 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
12251 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
12252 use xop_pmacsdqh if uns_p.
12253 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
12254 the immediate rotate count.
12256 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
12258 * rtl.h (struct rtx_def): Add comment for field jump.
12259 (LRA_SUBREG_P): New macro.
12260 * recog.c (register_operand): Check LRA_SUBREG_P.
12261 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12263 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12264 (lra_final_code_change): Skip subreg change for operators.
12265 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12266 if there are no operand changes.
12267 * lra-constraints.c (curr_insn_set): New.
12268 (match_reload): Set LRA_SUBREG_P.
12269 (emit_spill_move): Ditto.
12270 (check_and_process_move): Use curr_insn_set. Process only single
12271 set insns. Don't initialize sec_mem_p and change_p.
12272 (simplify_operand_subreg): Use LRA_SUBREG_P.
12273 (reg_in_class_p): New function.
12274 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12275 of #ifdef. Add code to remove cycling.
12276 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12277 non-null disp. Reload inner instead of disp when base and index
12278 are null. Try to put lo_sum into register.
12279 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12280 (check_and_process_move): Move code for move cost check to
12281 simple_move_p. Remove equiv_substitution.
12282 (simple_move_p): New function.
12283 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12284 curr_insn_set. Call check_and_process_move only for single set
12285 insns. Use the new function. Move call of check_and_process_move
12286 after operand equiv substitution and address process.
12288 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12291 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
12292 with nonlocal goto receivers or returns twice calls, ignore
12293 unininitialized values from abnormal edges to nl goto receiver
12294 or returns twice call.
12296 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12298 PR tree-optimization/57051
12299 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
12300 and VEC_RSHIFT_EXPR if shift count is a multiple of element
12303 2013-04-26 Richard Biener <rguenther@suse.de>
12305 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
12306 (expand_omp_taskreg): Likewise. Mark loops for fixup.
12307 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
12308 (fixup_loop_arrays_after_move): New function.
12309 (move_sese_region_to_fn): Properly outline the loop tree parts
12310 of the SESE region.
12312 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
12314 * config/i386/i386.md (type, unit): Fix long lines.
12316 2013-04-26 Richard Biener <rguenther@suse.de>
12318 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
12319 (lto-streamer-out.o): Likewise.
12320 * cfgloop.c (init_loops_structure): Export, add struct function
12321 argument and adjust.
12322 (flow_loops_find): Adjust.
12323 * cfgloop.h (enum loop_estimation): Add EST_LAST.
12324 (init_loops_structure): Declare.
12325 * lto-streamer-in.c: Include cfgloop.h.
12326 (input_cfg): Input the loop tree.
12327 * lto-streamer-out.c: Include cfgloop.h.
12328 (output_cfg): Output the loop tree.
12329 (output_struct_function_base): Do not drop PROP_loops.
12331 2013-03-26 Richard Biener <rguenther@suse.de>
12333 * tree-cfg.c (execute_build_cfg): Build the loop tree.
12334 (pass_build_cfg): Provide PROP_loops.
12335 (move_sese_region_to_fn): Remove loops that are outlined into fn
12337 * tree-inline.c: Include cfgloop.h.
12338 (initialize_cfun): Do not drop PROP_loops.
12339 (copy_loops): New function.
12340 (copy_cfg_body): Copy loop structure.
12341 (tree_function_versioning): Initialize destination loop tree.
12342 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
12343 (pass_parallelize_loops): Do IL verification.
12344 * loop-init.c (loop_optimizer_init): Fixup loops if required.
12345 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
12346 the CFG make sure we fixup loops as well.
12347 * tree-ssa-tail-merge.c: Include cfgloop.h.
12348 (replace_block_by): When merging loop latches mark loops for fixup.
12349 * lto-streamer-out.c (output_struct_function_base): Drop
12350 PROP_loops for now.
12351 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
12352 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
12353 * ipa-split.c: Include cfgloop.h.
12354 (split_function): Add the new return block to the loop tree root.
12355 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
12356 whether we have removed the forwarder block.
12357 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
12358 * cfgloop.h (place_new_loop): Declare.
12359 * cfgloopmanip.c (place_new_loop): Export.
12360 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
12361 (tree-switch-conversion.o): Likewise.
12362 (tree-complex.o): Likewise.
12363 (tree-inline.o): Likewise.
12364 (tree-ssa-tailmerge.o): Likewise.
12365 (ipa-split.o): Likewise.
12366 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
12367 (tree-ssa-copy.o): Likewise.
12368 * tree-switch-conversion.c: Include cfgloop.h
12369 (process_switch): If we emit a bit-test cascade, schedule loops
12371 * tree-complex.c: Include cfgloop.h.
12372 (expand_complex_div_wide): Properly add new basic-blocks to loops.
12373 * asan.c: Include cfgloop.h.
12374 (create_cond_insert_point): Properly add new basic-blocks to
12375 loops, schedule loop fixup.
12376 * cfgloop.c (verify_loop_structure): Check that looks are not
12378 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
12380 (expand_omp_for_generic): Likewise.
12381 (expand_omp_sections): Likewise.
12382 (expand_omp_atomic_pipeline): Schedule loops for fixup.
12383 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
12384 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
12385 is initialized, not when loops are present.
12386 * tree-parloops.c (parallelize_loops): Remove checking here.
12387 * passes.c (init_optimization_passes): Schedule a copy-propagation
12388 pass before complete unrolling of inner loops.
12390 2013-04-26 Jakub Jelinek <jakub@redhat.com>
12392 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
12393 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
12394 (colorize_init): Add argument to _WIN32 version.
12395 * toplev.c: Include diagnostic-color.h.
12396 (process_options): Default to -fdiagnostics-color=auto if
12397 GCC_COLORS env var is in the environment.
12398 * common.opt (fdiagnostics-color=): Add Var and Init.
12399 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
12400 env var is in the environment, the default is auto rather than never.
12402 * diagnostic.h (file_name_as_prefix): Add context argument.
12403 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
12404 the string as locus.
12405 * langhooks.c (lhd_print_error_function): Adjust caller.
12407 2013-04-25 Lawrence Crowl <crowl@google.com>
12409 * var-tracking.c (shared_hash_def::htab):
12410 Change type to hash_table. Update dependent calls and types.
12412 2013-04-25 Lawrence Crowl <crowl@google.com>
12414 * Makefile.in: Update as needed below.
12416 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
12417 Move declaration to after the type's method definitons.
12419 * attribs.c (htab_t scoped_attributes::attribute_hash):
12420 Change type to hash_table. Update dependent calls and types.
12422 * bitmap.c (htab_t bitmap_desc_hash):
12423 Change type to hash_table. Update dependent calls and types.
12425 * cselib.c (htab_t cselib_hash_table):
12426 Change type to hash_table. Update dependent calls and types.
12428 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
12429 (hash_string_slot_node): Move implementation into lto-streamer.h
12430 struct string_slot_hasher.
12431 (eq_string_slot_node): Likewise.
12433 * data-streamer-out.c: Update output_block::string_hash_table
12434 dependent calls and types.
12436 * dwarf2cfi.c (htab_t trace_index):
12437 Change type to hash_table. Update dependent calls and types.
12439 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
12440 Change type to hash_table. Update dependent calls and types.
12441 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
12442 (htab_t optimize_external_refs::map): Likewise.
12443 (htab_t output_comp_unit::extern_map): Likewise.
12444 (htab_t output_comdat_type_unit::extern_map): Likewise.
12445 (htab_t output_macinfo::macinfo_htab): Likewise.
12446 (htab_t optimize_location_lists::htab): Likewise.
12447 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
12449 * except.c (htab_t ehspec_hash_type):
12450 Change type to hash_table. Update dependent calls and types.
12451 (assign_filter_values::ttypes): Likewise.
12452 (assign_filter_values::ehspec): Likewise.
12453 (sjlj_assign_call_site_values::ar_hash): Likewise.
12454 (convert_to_eh_region_ranges::ar_hash): Likewise.
12456 * gcse.c (htab_t pre_ldst_table):
12457 Change type to hash_table. Update dependent calls and types.
12459 * ggc-common.c (htab_t saving_htab):
12460 Change type to hash_table. Update dependent calls and types.
12461 (htab_t loc_hash): Likewise.
12462 (htab_t ptr_hash): Likewise.
12463 (call_count): Rename ggc_call_count.
12464 (call_alloc): Rename ggc_call_alloc.
12465 (loc_descriptor): Rename make_loc_descriptor.
12466 (add_statistics): Rename ggc_add_statistics.
12468 * ggc-common.c (saving_htab):
12469 Change type to hash_table. Update dependent calls and types.
12471 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
12472 (push_gimplify_context): Likewise.
12473 (pop_gimplify_context): Likewise.
12474 (struct gimple_temp_hash_elt): Added.
12475 (struct gimplify_hasher): Likewise.
12476 (struct gimplify_ctx.temp_htab):
12477 Change type to hash_table. Update dependent calls and types.
12479 * gimple-fold.c: Include gimplify-ctx.h.
12481 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
12482 Change type to hash_table. Update dependent calls and types.
12483 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
12484 avoid potential global name collision.
12486 * gimplify.c: Include gimplify-ctx.h.
12487 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
12488 (htab_t gimplify_ctx::temp_htab):
12489 Update dependent calls and types for new type hash_table.
12490 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
12491 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
12493 * gimplify-ctx.h: New.
12494 (struct gimple_temp_hash_elt): Move from gimplify.c.
12495 (class gimplify_hasher): New.
12496 (struct gimplify_ctx): Move from gimple.h.
12497 (htab_t gimplify_ctx::temp_htab):
12498 Change type to hash_table. Update dependent calls and types.
12500 * graphite-clast-to-gimple.c: Include graphite-htab.h.
12501 (htab_t ivs_params::newivs_index):
12502 Change type to hash_table. Update dependent calls and types.
12503 (htab_t ivs_params::params_index): Likewise.
12504 (htab_t print_generated_program::params_index): Likewise.
12505 (htab_t gloog::newivs_index): Likewise.
12506 (htab_t gloog::params_index): Likewise.
12508 * graphite.c: Include graphite-htab.h.
12509 4htab_t graphite_transform_loops::bb_pbb_mapping):
12510 Change type to hash_table. Update dependent calls and types.
12512 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
12513 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
12514 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
12516 * graphite-dependences.c: Include graphite-htab.h.
12517 (loop_is_parallel_p): Change hash table type of parameter.
12519 * graphite-htab.h: New.
12520 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
12521 (extern find_pbb_via_hash): Move from graphite-poly.h.
12522 (extern loop_is_parallel_p): Move from graphite-poly.h.
12523 (extern get_loop_body_pbbs): Move from graphite-poly.h.
12525 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
12526 (extern loop_is_parallel_p): Move to graphite-htab.h.
12527 (extern get_loop_body_pbbs): Move to graphite-htab.h.
12529 * haifa-sched.c (htab_t delay_htab):
12530 Change type to hash_table. Update dependent calls and types.
12531 (htab_t delay_htab_i2): Likewise.
12533 * ira-color.c (htab_t allocno_hard_regs_htab):
12534 Change type to hash_table. Update dependent calls and types.
12536 * ira-costs.c (htab_t cost_classes_htab):
12537 Change type to hash_table. Update dependent calls and types.
12539 * loop-invariant.c (htab_t merge_identical_invariants::eq):
12540 Change type to hash_table. Update dependent calls and types.
12542 * loop-iv.c (htab_t bivs):
12543 Change type to hash_table. Update dependent calls and types.
12545 * loop-unroll.c (htab_t opt_info::insns_to_split):
12546 Change type to hash_table. Update dependent calls and types.
12547 (htab_t opt_info::insns_with_var_to_expand): Likewise.
12549 * lto-streamer.h (struct string_slot): Move from data-streamer.h
12550 (struct string_slot_hasher): New.
12551 (htab_t output_block::string_hash_table):
12552 Change type to hash_table. Update dependent calls and types.
12554 * lto-streamer-in.c (freeing_string_slot_hasher): New.
12555 (htab_t file_name_hash_table):
12556 Change type to hash_table. Update dependent calls and types.
12558 * lto-streamer-out.c: Update output_block::string_hash_table dependent
12561 * lto-streamer.c (htab_t tree_htab):
12562 Change type to hash_table. Update dependent calls and types.
12564 * omp-low.c: Include gimplify-ctx.h.
12566 * passes.c (htab_t name_to_pass_map):
12567 Change type to hash_table. Update dependent calls and types.
12568 (pass_traverse): Rename to passes_pass_traverse.
12570 * plugin.c (htab_t event_tab):
12571 Change type to hash_table. Update dependent calls and types.
12573 * postreload-gcse.c (htab_t expr_table):
12574 Change type to hash_table. Update dependent calls and types.
12575 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
12577 * sese.c (debug_rename_map_1): Make extern.
12578 (htab_t copy_bb_and_scalar_dependences::rename_map):
12579 Change type to hash_table. Update dependent calls and types.
12581 * sese.h (extern debug_rename_map): Move to .c file.
12583 * store-motion.c (htab_t store_motion_mems_table):
12584 Change type to hash_table. Update dependent calls and types.
12586 * trans-mem.c (htab_t tm_new_mem_hash):
12587 Change type to hash_table. Update dependent calls and types.
12589 * tree-browser.c (htab_t TB_up_ht):
12590 Change type to hash_table. Update dependent calls and types.
12592 * tree-cfg.c (htab_t discriminator_per_locus):
12593 Change type to hash_table. Update dependent calls and types.
12595 * tree-complex.c: Include tree-hasher.h
12596 (htab_t complex_variable_components):
12597 Change type to hash_table. Update dependent calls and types.
12599 * tree-eh.c (htab_t finally_tree):
12600 Change type to hash_table. Update dependent calls and types.
12602 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
12603 struct int_tree_hasher.
12604 (extern int_tree_map_eq): Likewise.
12605 (uid_decl_map_hash): Removed.
12606 (extern decl_tree_map_eq): Likewise.
12608 * tree-hasher.h: New.
12609 (struct int_tree_hasher): New.
12610 (typedef int_tree_htab_type): New.
12612 * tree-inline.c: Include gimplify-ctx.h.
12614 * tree-mudflap.c: Include gimplify-ctx.h.
12616 * tree-parloops.c: Include tree-hasher.h.
12617 (htab_t eliminate_local_variables_stmt::decl_address):
12618 Change type to hash_table. Update dependent calls and types.
12619 (htab_t separate_decls_in_region::decl_copies): Likewise.
12621 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
12622 Change type to hash_table. Update dependent calls and types.
12624 * tree-sra.c (candidates):
12625 Change type to hash_table. Update dependent calls and types.
12627 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
12629 (int_tree_map_hash): Likewise.
12631 * tree-ssa-dom.c (htab_t avail_exprs):
12632 Change type to hash_table. Update dependent calls and types.
12634 * tree-ssa-live.c (var_map_base_init::tree_to_index):
12635 Change type to hash_table. Update dependent calls and types.
12637 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
12638 Change type to hash_table. Update dependent calls and types.
12640 * tree-ssa-phiopt.c (seen_ssa_names):
12641 Change type to hash_table. Update dependent calls and types.
12643 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
12644 Change type to hash_table. Update dependent calls and types.
12646 * tree-ssa-uncprop.c (equiv):
12647 Change type to hash_table. Update dependent calls and types.
12649 2013-04-25 Jakub Jelinek <jakub@redhat.com>
12651 PR rtl-optimization/57003
12652 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
12653 call note_stores with kill_clobbered_value callback again after
12654 killing regs_invalidated_by_call.
12656 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12658 * config/aarch64/aarch64-simd.md
12659 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
12660 (aarch64_simd_bsl<mode>): Likewise.
12661 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
12663 2013-04-25 Marek Polacek <polacek@redhat.com>
12665 PR tree-optimization/57066
12666 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
12668 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
12670 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
12672 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12674 * config/aarch64/aarch64-builtins.c
12675 (aarch64_fold_builtin): New.
12676 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
12677 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
12678 * config/aarch64/aarch64-simd-builtins.def (abs): New.
12679 * config/aarch64/arm_neon.h
12680 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
12682 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12683 Tejas Belagod <tejas.belagod@arm.com>
12685 * config/aarch64/aarch64-builtins.c
12686 (aarch64_gimple_fold_builtin): New.
12687 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
12688 * config/aarch64/aarch64-simd-builtins.def (addv): New.
12689 * config/aarch64/aarch64-simd.md (addpv4sf): New.
12690 (addvv4sf): Update.
12691 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
12693 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12695 * config/aarch64/aarch64.md
12696 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
12698 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12700 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
12701 (*ngcsi_uxtw): New pattern.
12703 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12704 Julian Brown <julian@codesourcery.com>
12706 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
12707 (TB_DREG): Add T_V4HF.
12708 (v4hf_UP): New macro.
12709 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
12710 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
12711 Handle initialisation of V4HF. Adjust initialisation of reinterpret
12713 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
12714 (arm_vector_mode_supported_p): Handle V4HF.
12715 (arm_mangle_map): Handle V4HFmode.
12716 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
12717 * config/arm/arm_neon_builtins.def: Add entries for
12718 vcvtv4hfv4sf, vcvtv4sfv4hf.
12719 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
12720 (neon_vcvtv4hfv4sf): Likewise.
12721 * config/arm/neon-gen.ml: Handle half-precision floating point
12723 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
12724 * config/arm/arm_neon.h: Regenerate.
12725 * config/arm/neon.ml (type elts): Add F16.
12726 (type vectype): Add T_float16x4, T_floatHF.
12727 (type vecmode): Add V4HF.
12728 (type features): Add Requires_FP_bit feature.
12729 (elt_width): Handle F16.
12730 (elt_class): Likewise.
12731 (elt_of_class_width): Likewise.
12732 (mode_of_elt): Refactor.
12733 (type_for_elt): Handle F16, fix error messages.
12734 (vectype_size): Handle T_float16x4.
12735 (vcvt_sh): New function.
12736 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
12737 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
12738 (string_of_mode): Handle V4HF.
12739 * doc/arm-neon-intrinsics.texi: Regenerate.
12741 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12743 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
12744 format specifier in 'X' case.
12746 2013-04-25 Alan Modra <amodra@gmail.com>
12749 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
12750 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
12751 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
12752 Repeat for many other rotate/shift and mask patterns using subregs.
12753 Name lshiftrt insns.
12754 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
12755 on WORDS_BIG_ENDIAN.
12757 2013-04-25 Alan Modra <amodra@gmail.com>
12759 * config.gcc: Support little-endian powerpc-linux targets.
12760 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
12761 (LINK_OS_LINUX_SPEC): Define.
12762 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
12763 Preserve MASK_LITTLE_ENDIAN.
12764 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
12765 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
12766 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
12767 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
12768 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
12769 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
12770 Correct fp word order for little-endian. Don't shift toc entries
12771 smaller than a word for little-endian.
12772 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
12773 (bswapdi2 splits): Correct low-part subreg for little-endian.
12774 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
12775 low/high where such is correct only for be.
12776 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
12777 little-endian for -mcall-aixdesc.
12779 2013-04-25 Alan Modra <amodra@gmail.com>
12781 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
12782 replace_equiv_address_nv.
12784 2013-04-25 Alan Modra <amodra@gmail.com>
12786 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
12788 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12791 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12792 * rtl.h (struct rtx_def): ...
12794 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12796 PR rtl-optimizations/57046
12797 * lra-constraints (split_reg): Set up lra_risky_transformations_p
12798 for multi-reg splits.
12800 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
12802 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
12804 2013-04-24 Sterling Augustine <saugustine@google.com>
12806 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
12807 (comp_dir_string, debug_str_dwo_section): New.
12808 (DEBUG_STR_DWO_SECTION): Rename to ...
12809 (DEBUG_DWO_STR_SECTION): ... this.
12810 (DEBUG_NORM_STR_SECTION): Delete.
12811 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
12812 (DEBUG_STR_DWO_SECTION_FLAGS): New.
12813 (find_AT_string): Move most logic to ...
12814 (find_AT_string_in_table): ... here. New.
12815 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
12816 add_skeleton_AT_string. Delete logic.
12817 (output_skeleton_debug_sections): Remove call to
12818 add_top_level_skeleton_die_attrs.
12819 (add_comp_dir_attribute): Move logic to comp_dir_string.
12820 (dwarf2out_init): Initialize debug_str_dwo_section.
12821 (output_indirect_string): Call find_string_form.
12822 (output_indirect_strings): Rewrite.
12823 (prune_unused_types): Empty skeleton_debug_str_hash.
12824 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
12825 (dwarf2out_finish): Call output_indirect_strings.
12827 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12829 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
12831 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12833 * rtl.h (struct rtx_def): Add comment for field jump.
12834 (LRA_SUBREG_P): New macro.
12835 * recog.c (register_operand): Check LRA_SUBREG_P.
12836 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12838 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12839 (lra_final_code_change): Skip subreg change for operators.
12840 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12841 if there are no operand changes.
12842 * lra-constraints.c (curr_insn_set): New.
12843 (match_reload): Set LRA_SUBREG_P.
12844 (emit_spill_move): Ditto.
12845 (check_and_process_move): Use curr_insn_set. Process only single
12846 set insns. Don't initialize sec_mem_p and change_p.
12847 (simplify_operand_subreg): Use LRA_SUBREG_P.
12848 (reg_in_class_p): New function.
12849 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12850 of #ifdef. Add code to remove cycling.
12851 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12852 non-null disp. Reload inner instead of disp when base and index
12853 are null. Try to put lo_sum into register.
12854 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12855 (check_and_process_move): Move code for move cost check to
12856 simple_move_p. Remove equiv_substitution.
12857 (simple_move_p): New function.
12858 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12859 curr_insn_set. Call check_and_process_move only for single set
12860 insns. Use the new function. Move call of check_and_process_move
12861 after operand equiv substitution and address process.
12863 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
12865 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
12866 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
12867 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
12869 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12871 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
12873 2013-04-24 Marek Polacek <polacek@redhat.com>
12875 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
12876 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
12877 (select_loops_exit_conditions): Likewise.
12878 (number_of_iterations_for_all_loops): Likewise.
12879 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
12880 (scev_analysis): Likewise.
12882 2013-04-02 Catherine Moore <clm@codesourcery.com>
12883 Chao-ying Fu <fu@mips.com>
12885 * config/mips/micromips.md (jraddiusp): New pattern.
12886 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
12887 instruction if possible.
12889 2013-04-24 Alan Modra <amodra@gmail.com>
12891 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
12893 2013-04-24 Julian Brown <julian@codesourcery.com>
12894 Chung-Lin Tang <cltang@codesourcery.com>
12896 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
12897 dependency behavior in enumeration type DIE generation. Add TODO note
12898 to comments about future DW_FORM_sdata/udata re-work of related code.
12900 2013-04-23 Lawrence Crowl <crowl@google.com>
12902 * Makefile.in: Update as needed below.
12904 * hash-table.h (class hash_table):
12905 Correct many methods with parameter types compare_type to the correct
12906 value_type. (Correct code was unlikely to notice the change.)
12907 (hash_table::elements_with_deleted) New.
12908 (class hashtable::iterator): New.
12909 (hashtable::begin()): New.
12910 (hashtable::end()): New.
12911 (FOR_EACH_HASH_TABLE_ELEMENT): New.
12913 * statistics.c (statistics_hashes):
12914 Change type to hash_table. Update dependent calls and types.
12916 * tree-into-ssa.c (var_infos):
12917 Change type to hash_table. Update dependent calls and types.
12919 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
12920 Change type to hash_table. Update dependent calls and types.
12922 * tree-ssa-loop-im.c (struct mem_ref.refs):
12923 Change type to hash_table. Update dependent calls and types.
12925 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
12926 Change type to hash_table. Update dependent calls and types.
12928 * tree-ssa-sccvn.c (vn_tables_s::nary):
12929 Change type to hash_table. Update dependent calls and types.
12930 (vn_tables_s::phis): Likewise.
12931 (vn_tables_s::references): Likewise.
12933 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
12934 (vn_reference_eq): Update parameter and return types.
12936 * tree-ssa-structalias.c (pointer_equiv_class_table):
12937 Change type to hash_table. Update dependent calls and types.
12938 (location_equiv_class_table): Likewise.
12940 * tree-vect-data-refs.c: Consequential changes for making
12941 peeling a hash_table.
12943 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
12944 (destroy_loop_vec_info): Dependent hash_table update.
12946 * tree-vectorizer.h (peeling_htab):
12947 Change type to hash_table. Update dependent calls and types.
12949 2013-04-23 Shiva Chen <shiva0217@gmail.com>
12951 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
12952 to check the register content is equal or not.
12953 * lra-constraints.c (match_reload): Use lra_assign_reg_val
12954 to assign register content record.
12955 * lra-eliminations.c (update_reg_eliminate): Use
12956 lra_update_reg_val_offset to update register content offset.
12957 * lra-int.h (struct lra_reg): Add offset member.
12958 (lra_reg_val_equal_p): New static inline function.
12959 (lra_update_reg_val_offset): New static inline function.
12960 (lra_assign_reg_val): New static inline function.
12961 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
12962 to assign register content record.
12963 (initialize_lra_reg_info_element): Initial offset to zero.
12965 2013-04-23 Catherine Moore <clm@codesourcery.com>
12967 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
12968 operands. Record compression.
12970 2013-04-23 Xinliang David Li <davidxl@google.com>
12972 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
12974 2013-04-23 Richard Biener <rguenther@suse.de>
12976 PR middle-end/57036
12977 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
12978 parameter, only add abnormal goto edges from the copied body
12979 if the call could perform abnormal gotos.
12980 (copy_cfg_body): Adjust.
12982 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
12984 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
12986 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
12988 * coretypes.h (gimple_stmt_iterator): Add struct to make
12991 2013-04-23 Richard Biener <rguenther@suse.de>
12993 PR tree-optimization/57026
12994 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
12995 from SSA names occuring in abnormal PHI nodes.
12997 2013-04-22 Andi Kleen <ak@linux.intel.com>
12999 * lto/lto.c (print_lto_report_1): Fix LTO report names.
13001 2013-04-22 Andi Kleen <ak@linux.intel.com>
13003 * lto/lto.c (print_lto_report_1): Declare early.
13004 (read_cgraph_and_symbols): Call print_lto_report_1 early.
13006 2013-04-22 Andi Kleen <ak@linux.intel.com>
13008 * common.opt (-flto-report-wpa): Add.
13009 * doc/invoke.texi (-flto-report-wpa): Add.
13010 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
13013 2013-04-22 Xinliang David Li <davidxl@google.com>
13015 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
13016 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
13017 * Makefile.in: New dependency
13019 David Daney <ddaney.cavm@gmail.com>
13021 * configure.ac (gcc_cv_as_micromips_support): Use the
13022 --fatal-warnings option.
13023 * configure: Regenerate.
13025 2013-04-22 Marek Polacek <polacek@redhat.com>
13028 * tsan.c (instrument_expr): Don't instrument expression
13029 in case its size is zero.
13031 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
13035 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13037 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13039 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13041 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
13042 (gimple_stmt_iterator): New typedef.
13043 * gimple.h (gimple_stmt_iterator): Rename to...
13044 (gimple_stmt_iterator_d): ... This.
13045 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
13046 trees be valid for GIMPLE and GENERIC.
13047 (TARGET_GIMPLE_FOLD_BUILTIN): New.
13048 * gimple-fold.c (gimple_fold_call): Call target hook
13049 gimple_fold_builtin.
13050 * hooks.c (hook_bool_gsiptr_false): New.
13051 * hooks.h (hook_bool_gsiptr_false): New.
13052 * target.def (fold_stmt): New.
13053 * doc/tm.texi: Regenerate.
13055 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
13058 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
13059 a set sp if no stack realignment.
13061 2013-04-22 Nick Clifton <nickc@redhat.com>
13063 * config.gcc (tilegx-linux): Extend extra_objs rather than
13065 (tilepro-linux): Likewise.
13067 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13069 * config/aarch64/aarch64-builtins.c
13071 (CF0, CF1, CF2, CF3, CF4, CF10): New.
13072 (VAR<1-12>): Add MAP parameter.
13073 (BUILTIN_*): Likewise.
13074 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
13075 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
13076 (aarch64_ushl_n<mode>): Likewise.
13077 (aarch64_sshr_n<mode>): Likewise.
13078 (aarch64_ushr_n<mode>): Likewise.
13079 (aarch64_<maxmin><mode>): Likewise.
13080 (aarch64_sqrt<mode>): Likewise.
13081 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
13082 (vshr<q>_n_*): Likewise.
13084 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13086 * config/aarch64/aarch64-builtins.c
13087 (aarch64_simd_builtin_type_mode): Handle SF types.
13089 (BUILTIN_GPF): Define.
13090 (aarch64_init_simd_builtins): Handle SF types.
13091 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
13092 (frecps): Likewise.
13093 (frecpx): Likewise.
13094 * config/aarch64/aarch64-simd.md
13095 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
13096 (aarch64_frecpe<mode>): New.
13097 (aarch64_frecps<mode>): Likewise.
13098 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
13099 (v8type): Add frecp<esx>.
13100 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
13101 (aarch64_frecps<mode>): Likewise.
13102 * config/aarch64/iterators.md (FRECP): New.
13103 (frecp_suffix): Likewise.
13104 * config/aarch64/arm_neon.h
13105 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
13107 2013-04-22 Christian Bruel <christian.bruel@st.com>
13110 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
13111 (REG_CLASS_NAMES): Idem.
13112 (REG_CLASS_CONTENTS): Idem.
13113 (REGCLASS_HAS_FP_REG): Idem.
13114 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
13115 (sh_conditional_register_usage): Idem.
13117 2013-04-21 Jeff Law <law@redhat.com>
13119 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
13120 (ssa_forward_propagate_and_combine): Use it.
13122 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13124 * lra.c: Update the flow chart diagram.
13126 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13128 PR rtl-optimization/56847
13129 * lra-constraints.c (process_alt_operands): Discourage alternative
13130 with non-matche doffsettable memory constraint fro memory with
13133 2013-04-19 Richard Biener <rguenther@suse.de>
13135 PR tree-optimization/56982
13136 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
13138 * gimplify.c (gimplify_call_expr): Notice special calls.
13139 (gimplify_modify_expr): Likewise.
13140 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
13141 abnormal control flow receivers.
13142 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
13143 in the same way as cfun->has_nonlocal_labels.
13144 (gimple_purge_dead_abnormal_call_edges): Likewise.
13145 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
13146 receivers start a basic-block.
13148 2013-04-19 Richard Biener <rguenther@suse.de>
13150 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
13152 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
13153 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
13154 (SLP_TREE_LOAD_PERMUTATION): Add.
13155 (vect_transform_slp_perm_load): Adjust prototype.
13156 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
13157 (vect_free_slp_instance): Likewise.
13158 (vect_create_new_slp_node): Likewise.
13159 (vect_supported_slp_permutation_p): Remove.
13160 (vect_slp_rearrange_stmts): Adjust.
13161 (vect_supported_load_permutation_p): Likewise. Inline
13162 vect_supported_slp_permutation_p here.
13163 (vect_analyze_slp_instance): Compute load permutations per
13164 slp node instead of per instance.
13165 (vect_get_slp_defs): Adjust.
13166 (vect_transform_slp_perm_load): Likewise.
13167 (vect_schedule_slp_instance): Remove redundant code.
13168 (vect_schedule_slp): Remove hack for PR56270, add it ...
13169 * tree-vect-stmts.c (vectorizable_load): ... here, do not
13170 CSE loads for SLP. Adjust.
13172 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13174 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
13175 spelling in two comments.
13177 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13180 * config/arm/arm.c (load_multiple_sequence): Require SP
13181 as base register for loads if SP is in the register list.
13183 2013-04-19 Martin Jambor <mjambor@suse.cz>
13185 PR tree-optimization/56718
13186 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
13187 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
13188 and made public. Adjusted all callers.
13189 (ipa_intraprocedural_devirtualization): New function.
13190 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
13191 (ipa_intraprocedural_devirtualization): Likewise.
13192 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
13194 2013-04-19 Richard Biener <rguenther@suse.de>
13196 PR tree-optimization/57000
13197 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
13199 2013-04-19 Terry Guo <terry.guo@arm.com>
13201 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
13203 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
13204 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
13205 (cortex_m4_fmacs): Use new reservations.
13206 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
13208 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
13210 PR rtl-optimization/56999
13211 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
13213 (lra_coalesce): Remove split_origin_bitmap and related code.
13214 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
13215 ranges if necessary.
13217 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
13219 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
13221 (ix86_expand_call): Remove clobbered_registers array and use
13222 x86_64_ms_sysv_extra_clobbered_registers instead.
13223 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
13225 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
13227 * config/i386/i386.md (*call_rex64_ms_sysv): Use
13228 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
13229 (*call_value_rex64_ms_sysv): Ditto.
13231 2013-04-18 Cary Coutant <ccoutant@google.com>
13233 * dwarf2out.c (output_pubnames): Check die_perennial_p of
13234 parent instead of die_mark.
13236 2013-04-18 Diego Novillo <dnovillo@google.com>
13238 * gimple.c (create_gimple_tmp): New.
13239 (get_expr_type): New.
13240 (build_assign): New.
13241 (build_type_cast): New.
13242 * gimple.h (enum ssa_mode): Define.
13243 (gimple_seq_set_location): New.
13244 * asan.c (build_check_stmt): Change some gimple_build_* calls
13245 to use build_assign and build_type_cast.
13247 2013-04-18 Richard Biener <rguenther@suse.de>
13249 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
13250 handle negative step. Remove redundant checks.
13251 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
13252 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
13253 for negative step and grouped loads fail to vectorize.
13255 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13257 * emit-rtl.c (reset_insn_used_flags): New function.
13258 (reset_all_used_flags): Use it.
13259 (verify_insn_sharing): New function.
13260 (verify_rtl_sharing): Fix verification for SEQUENCEs.
13262 2013-04-18 Jakub Jelinek <jakub@redhat.com>
13264 PR tree-optimization/56984
13265 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
13266 and (x >> M) >= N don't register any assertion if N << M is the
13269 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13271 * lower-subreg.c (resolve_simple_move): If called self-recursive,
13272 do not delete_insn insns that have not yet been emitted, only
13273 unlink them with remove_insn.
13274 * df-scan.c (df_insn_delete): Revert r197492.
13276 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
13278 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
13279 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
13281 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
13283 * config/arm/arm.md (movsicc_insn): Convert define_insn into
13284 define_insn_and_split.
13285 (and_scc,ior_scc,negscc): Likewise.
13286 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
13288 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
13290 * config/arm/arm.c (use_return_insn): Return 0 for targets that
13291 can benefit from using a sequence of LDRD instructions in epilogue
13292 instead of a single LDM instruction.
13294 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
13297 * doc/extend.texi: Fix typo.
13299 2013-04-17 Richard Biener <rguenther@suse.de>
13301 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
13302 (vect_build_slp_tree): ... here.
13303 (vect_build_slp_tree_1): Compute which stmts of the SLP group
13304 match. Remove special-casing of mismatched complex loads.
13305 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
13306 re-try the match with swapped commutative operands.
13307 (vect_supported_load_permutation_p): Remove special-casing of
13308 mismatched complex loads.
13309 (vect_analyze_slp_instance): Adjust.
13311 2013-04-17 Richard Biener <rguenther@suse.de>
13313 PR rtl-optimization/56921
13314 * cfgloop.h (struct loop): Add simple_loop_desc member.
13315 (struct niter_desc): Mark with GTY(()).
13316 (simple_loop_desc): Do not use aux field but simple_loop_desc.
13317 * loop-iv.c (get_simple_loop_desc): Likewise.
13318 (free_simple_loop_desc): Likewise.
13321 2013-04-16 Richard Biener <rguenther@suse.de>
13323 PR rtl-optimization/56921
13324 * loop-init.c (pass_rtl_move_loop_invariants): Add
13325 TODO_do_not_ggc_collect to todo_flags_finish.
13326 (pass_rtl_unswitch): Same.
13327 (pass_rtl_unroll_and_peel_loops): Same.
13328 (pass_rtl_doloop): Same.
13330 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
13332 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
13333 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
13334 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
13336 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
13337 * tree-streamer.c (record_common_node): Adjust reference in comment.
13339 2013-04-17 Terry Guo <terry.guo@arm.com>
13341 * config/arm/cortex-m4.md: Add a new bypass.
13343 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13345 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
13347 (*subs_<optab><mode>_multp2): New pattern.
13348 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13349 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13351 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13353 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
13354 (*subs_mul_imm_<mode>): New pattern.
13356 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
13359 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
13360 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
13361 (vsx_movti_32bit): Change j->wa to O->wa.
13363 2013-04-16 Richard Biener <rguenther@suse.de>
13365 PR rtl-optimization/56921
13366 * loop-init.c (pass_rtl_move_loop_invariants): Add
13367 TODO_do_not_ggc_collect to todo_flags_finish.
13368 (pass_rtl_unswitch): Same.
13369 (pass_rtl_unroll_and_peel_loops): Same.
13370 (pass_rtl_doloop): Same.
13372 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
13374 * config/arm/arm.c (emit_multi_reg_push): New declaration
13375 for an existing function.
13376 (arm_emit_strd_push): New function.
13377 (arm_expand_prologue): Used here.
13378 (arm_emit_ldrd_pop): New function.
13379 (arm_expand_epilogue): Used here.
13380 (arm_get_frame_offsets): Update condition.
13381 (arm_emit_multi_reg_pop): Add a special case for load of a single
13382 register with writeback.
13384 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
13386 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
13389 2013-04-16 Richard Biener <rguenther@suse.de>
13391 PR tree-optimization/56756
13392 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
13393 (first_mem_ref_loc): New.
13394 (execute_sm): Place the load temporarily before a previous
13395 access instead of in the latch edge to ensure its SSA dependencies
13396 are defined at points dominating the load.
13398 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
13400 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
13401 correct fix by moving header and footer insn to the footer of
13402 the merged basic block. Clear BB_END of the merged-away block.
13404 PR middle-end/43631
13405 * emit-rtl.c (make_note_raw): New function.
13406 (link_insn_into_chain): New static inline function.
13407 (add_insn): Use it.
13408 (add_insn_before, add_insn_after): Factor insn chain linking code...
13409 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
13410 using link_insn_into_chain.
13411 (note_outside_basic_block_p): New helper function for emit_note_after
13412 and emit_note_before.
13413 (emit_note_after): Use nobb variant of add_insn_after if the note
13414 should not be contained in a basic block.
13415 (emit_note_before): Use nobb variant of add_insn_before if the note
13416 should not be contained in a basic block.
13417 (emit_note_copy): Use make_note_raw.
13418 (emit_note): Likewise.
13419 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
13420 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
13421 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
13422 the moved barrier the tail of the basic block it follows.
13423 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
13425 2013-04-15 Jakub Jelinek <jakub@redhat.com>
13427 PR tree-optimization/56962
13428 * gimple-ssa-strength-reduction.c (record_increment): Only set
13429 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
13430 either rhs1 or rhs2 is equal to c->base_expr.
13432 2013-04-15 Richard Biener <rguenther@suse.de>
13434 PR tree-optimization/56933
13435 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
13437 (GROUP_READ_WRITE_DEPENDENCE): Remove.
13438 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
13439 * tree-vect-data-refs.c (vect_analyze_group_access): Move
13440 dependence check ...
13441 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
13443 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
13444 GROUP_READ_WRITE_DEPENDENCE.
13446 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13448 * emit-rtl.c (reset_all_used_flags): New function.
13449 (verify_rtl_sharing): Call reset_all_used_flags before and after
13450 performing the checks.
13452 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13454 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
13455 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
13456 * config/arm/constraints.md (De): New constraint.
13457 * config/arm/neon.md (anddi3_neon): Delete.
13458 (neon_vand<mode>): Expand to standard anddi3 pattern.
13459 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
13460 Move earlier in the file.
13461 (neon_inv_logic_op2): Likewise.
13462 (arm_anddi_operand_neon): New predicate.
13464 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13466 * configure.ac (gcc_cv_ld_as_needed): Set
13467 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
13468 Use -z ignore, -z record on *-*-solaris2*.
13469 (HAVE_LD_AS_NEEDED): Update comment.
13470 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
13471 * configure: Regenerate.
13472 * config.in: Regenerate.
13473 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
13474 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
13475 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
13476 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
13477 equivalents. Fix markup.
13478 * doc/tm.texi: Regenerate.
13480 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
13482 * config/i386/i386.opt: New option mstack-protector-guard=.
13483 * config/i386/i386-opts.h: Add enum stack_protector_guard.
13484 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
13485 TARGET_SSP_TLS_GUARD.
13486 * config/i386/i386.c (ix86_option_override_internal): Set
13487 ix86_stack_protector_guard.
13488 * config/i386/i386.md (stack_protect_set): Enable for
13489 TARGET_SSP_TLS_GUARD only.
13490 (stack_protect_set_<mode>): Ditto.
13491 (stack_protect_test): Ditto.
13492 (stack_protect_test_<mode>): Ditto.
13493 * doc/invoke.texi (i386 Option): Document.
13495 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
13498 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
13499 (S_MODES): Set H_MODE bit.
13500 (SF_MODES): Set only S_MODE and SF_MODE bits.
13501 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
13502 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
13503 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
13504 <MODE_FLOAT>: Likewise.
13506 2013-04-15 Joey Ye <joey.ye@arm.com>
13508 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
13510 2013-04-15 Joey Ye <joey.ye@arm.com>
13512 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
13514 (thumb_far_jump_used_p): Count instruction size and set
13517 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
13519 * reorg.c (fill_simple_delay_slots): Reindent block of code.
13520 * resource.c (mark_target_live_regs): Reformat conditional block.
13522 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
13524 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
13525 notes, they are emitted only just before final.
13526 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
13528 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
13530 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
13531 * cfgrtl.c (delete_insn): Call it here instead.
13532 * lra-spills.c (lra_final_code_change): Use delete_insn.
13533 * haifa-sched.c (sched_remove_insn): Likewise.
13534 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
13535 returning to the nop pool.
13536 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
13537 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
13539 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
13541 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
13542 * doc/tm.texi: Regenerated.
13544 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
13546 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
13549 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
13551 * df-core.c (df_find_def): Compare register numbers.
13552 (df_find_use): Likewise.
13554 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
13557 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
13558 lra_in_progress for return.
13560 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
13562 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
13563 define_insn into define_insn_and_split and emit movsicc patterns.
13565 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
13567 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
13569 2013-04-12 Richard Biener <rguenther@suse.de>
13571 * tree-pass.h (TODO_do_not_ggc_collect): New.
13572 * passes.c (execute_one_ipa_transform_pass): Honor
13573 TODO_do_not_ggc_collect.
13574 (execute_one_pass): Likewise.
13577 2013-04-10 Richard Biener <rguenther@suse.de>
13579 * passes.c (init_optimization_passes): Remove reload pass.
13580 * ira.c (do_reload): Merge into ...
13582 (rest_of_handle_reload): Remove.
13583 (pass_reload): Likewise.
13584 * config/i386/i386.c (ix86_option_override): Refer to ira instead
13585 of reload for vzeroupper pass placement.
13587 2013-04-12 Jakub Jelinek <jakub@redhat.com>
13589 PR tree-optimization/56918
13590 PR tree-optimization/56920
13591 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
13592 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
13593 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
13594 use wide_mul_with_sign method.
13596 2013-04-12 Richard Biener <rguenther@suse.de>
13598 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
13599 not be considered a gimple constant.
13601 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
13603 * fold-const.c (const_binop): Handle vector shifts by a scalar.
13604 (fold_binary_loc): Call const_binop also for mixed vector-scalar
13607 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
13608 Jakub Jelinek <jakub@redhat.com>
13610 * opts.c: Include diagnostic-color.h.
13611 (common_handle_option): Handle OPT_fdiagnostics_color_.
13612 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
13613 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
13614 (diagnostic-color.o): New.
13615 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
13616 (diagnostic_color_rule): New enum.
13617 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
13618 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
13619 the location string.
13620 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
13621 either NULL, or color kind.
13622 * diagnostic-color.c: New file.
13623 * diagnostic-color.h: New file.
13624 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
13626 * doc/invoke.texi (-fdiagnostics-color): Document.
13627 * pretty-print.h (pp_show_color): Define.
13628 (struct pretty_print_info): Add show_color field.
13629 * diagnostic.c: Include diagnostic-color.h.
13630 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
13631 macros. Colorize error:, warning: etc. strings and also the location
13633 (diagnostic_show_locus): Colorize the caret line.
13634 * pretty-print.c: Include diagnostic-color.h.
13635 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
13636 inside of %< %> quotes or quoted through q format modifier.
13638 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13640 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
13642 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13644 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
13645 code in CC_NZ mode.
13646 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
13649 2013-04-11 Marek Polacek <polacek@redhat.com>
13651 PR tree-optimization/48184
13652 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
13654 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
13656 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
13657 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
13658 (skip_simple_arithmetic): Tidy up.
13659 * tree.h (skip_simple_constant_arithmetic): Declare.
13661 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13663 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
13665 2013-04-11 Richard Biener <rguenther@suse.de>
13667 * tree-vect-loop.c (get_initial_def_for_induction): Properly
13668 generate vector constants.
13670 2013-04-11 Richard Biener <rguenther@suse.de>
13672 PR tree-optimization/56878
13673 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
13674 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
13676 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
13677 Prefer to align the DR with the most invariant base address.
13679 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
13681 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
13684 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
13686 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
13687 floating-point vector comparisons against 0.
13689 2013-04-11 Jakub Jelinek <jakub@redhat.com>
13691 PR tree-optimization/56899
13692 * fold-const.c (extract_muldiv_1): Apply distributive law
13693 only if TYPE_OVERFLOW_WRAPS (ctype).
13695 2013-04-11 Bin Cheng <bin.cheng@arm.com>
13698 * ira-costs.c (scan_one_insn): Check whether the source rtx of
13699 loading has side effect.
13701 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
13703 * config/sparc/sparc.c: Include tree-pass.h.
13704 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
13705 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
13706 head of file. Change return type. Split off gate function.
13707 (sparc_gate_work_around_errata): New function.
13708 (pass_work_around_errata): New pass definition.
13709 (insert_pass_work_around_errata) New pass insert definition to
13710 insert pass_work_around_errata just after delayed-branch scheduling.
13711 (sparc_option_override): Insert the pass.
13712 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
13714 2013-04-10 David S. Miller <davem@davemloft.net>
13716 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
13717 or -mcpu=hypersparc.
13719 * target.def (cstore_mode): New hook.
13720 * target.h: Include insn-codes.h
13721 * targhooks.c: Likewise.
13722 (default_cstore_mode): New function.
13723 * targhooks.h: Declare it.
13724 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
13725 * doc/tm.texi: Rebuild.
13726 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
13727 target hook, rather than inspecting the insn_data.
13728 * config/sparc/sparc.c (sparc_cstore_mode): New function.
13729 (TARGET_CSTORE_MODE): Redefine.
13730 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
13732 * config/sparc/predicates.md (cstore_result_operand): New special
13734 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
13735 Use it for operand 0.
13736 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
13737 (*snesi_special): Likewise.
13738 (*snesi_zero): Likewise.
13739 (*seqsi_zero): Likewise.
13740 (*sltu_insn): Likewise.
13741 (*sgeu_insn): Likewise.
13742 (*seqdi_special): Make operand 0 and comparison operation be of
13744 (*snedi_special): Likewise.
13745 (*snedi_special_vis3): Likewise.
13746 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
13747 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
13748 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
13749 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
13750 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
13751 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
13752 (*sltu_extend_sp64): Likewise.
13753 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
13754 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
13755 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
13756 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
13757 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
13759 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
13761 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
13762 (aarch64_start_file): Use the new function.
13764 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
13765 Jason Merrill <jason@redhat.com>
13767 * common.opt: Add -gdwarf.
13768 * opts.c (common_handle_option): Handle it.
13769 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
13771 2013-04-10 Richard Biener <rguenther@suse.de>
13773 * passes.c (execute_todo): Do not call ggc_collect conditional here.
13774 (execute_one_ipa_transform_pass): But unconditionally here.
13775 (execute_one_pass): And here.
13776 (init_optimization_passes): Remove reload pass.
13777 * tree-pass.h (TODO_ggc_collect): Remove.
13778 (pass_reload): Likewise.
13779 * ira.c (do_reload): Merge into ...
13781 (rest_of_handle_reload): Remove.
13782 (pass_reload): Likewise.
13783 * config/i386/i386.c (ix86_option_override): Refer to ira instead
13784 of reload for vzeroupper pass placement.
13785 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
13786 and todo_flags_finish of all passes.
13788 2013-04-10 Richard Biener <rguenther@suse.de>
13790 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
13791 first_const_oprnd field, rename first_def_type to first_op_type.
13792 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
13793 (vect_get_and_check_slp_defs): Always use the type of the
13794 operand. Allow mixed vect_external_def, vect_constant_def types.
13795 (vect_get_constant_vectors): Handle mixed vect_external_def,
13796 vect_constant_def types.
13798 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
13800 PR tree-optimization/55524
13801 * tree-ssa-math-opts.c
13802 (convert_mult_to_fma): Don't use an fms construct
13803 when we don't have an fms operation, but fnma, and it looks
13804 likely that we'll be able to use the latter.
13806 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
13808 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
13810 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
13811 inline fail caused by overwritable functions.
13813 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
13815 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
13816 unnecessary bits in the constant power of two case.
13818 2013-04-10 Richard Biener <rguenther@suse.de>
13820 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
13821 broken code swapping operands.
13822 (vect_build_slp_tree): Do not compute load permutations here.
13823 (vect_analyze_slp_instance): Compute load permutations here,
13824 after building the SLP tree.
13826 2013-04-09 Christian Bruel <christian.bruel@st.com>
13828 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
13829 of next/prev_real_insn.
13831 2013-04-09 Jan Hubicka <jh@suse.cz>
13833 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13834 Drop aliased parameter.
13835 (function_and_variable_visibility): Do not handle alias pairs.
13836 * cgraph.c (varpool_externally_visible_p): Update prototype.
13837 * varpool.c (varpool_add_new_variable): Update.
13839 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13841 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
13843 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
13845 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
13847 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
13849 2013-04-09 Marek Polacek <polacek@redhat.com>
13851 PR tree-optimization/48762
13852 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
13854 2013-04-09 Richard Biener <rguenther@suse.de>
13856 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
13858 (vect_build_slp_tree): Likewise.
13859 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
13860 calculating the cost of a SLP instance.
13861 (vect_analyze_slp_instance): Use it from here, after building
13864 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13866 PR middle-end/56883
13867 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
13868 expand_omp_for_static_chunk): Use simple_p = true in
13869 force_gimple_operand_gsi calls when assigning to addressable decls.
13871 2013-04-09 Jeff Law <law@redhat.com>
13873 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
13874 when the boolean was created by converting a wider object which
13875 had a boolean range.
13877 2013-04-09 Richard Biener <rguenther@suse.de>
13879 * tree-vectorizer.h (slp_void_p): Remove.
13880 (slp_tree): Typedef before _slp_tree declaration.
13881 (struct _slp_tree): Use a vector of slp_tree as children.
13882 (vect_get_place_in_interleaving_chain): Remove.
13883 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
13885 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
13887 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
13888 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
13889 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
13890 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
13891 Use slp_node instead of slp_void_p and adjust.
13893 2013-04-09 Richard Biener <rguenther@suse.de>
13895 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
13896 work that is not necessary.
13898 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13900 PR tree-optimization/56854
13901 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
13902 forward into clobber stmts if it would change MEM_REF lhs into
13905 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
13907 * tree.c (type_hash_lookup, type_hash_add): Make static.
13908 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
13910 2013-04-09 Richard Biener <rguenther@suse.de>
13912 * tree.h (unsave_expr_now): Remove.
13913 * tree-inline.c (mark_local_for_remap_r): Remove.
13914 (unsave_expr_1): Likewise.
13915 (unsave_r): Likewise.
13916 (unsave_expr_now): Likewise.
13917 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
13918 (propagate_tree_value): Likewise.
13920 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
13922 * doc/rtl.texi (sequence): Rewrite documentation to match the
13923 current use of SEQUENCE rtl objects.
13924 * rtl.def (SEQUENCE): Likewise.
13926 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
13927 Update documentation.
13928 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
13929 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
13931 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
13933 2013-04-08 Teresa Johnson <tejohnson@google.com>
13935 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
13936 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
13938 (estimate_edge_size_and_time): Add comment to suggest using rounding
13940 (estimate_node_size_and_time): Ditto.
13941 (remap_edge_change_prob): Use helper rounding divide methods.
13942 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
13943 (gimple_mod_pow2_value_transform): Ditto.
13944 (gimple_mod_subtract_transform): Ditto.
13945 (gimple_ic_transform): Ditto.
13946 (gimple_stringops_transform): Ditto.
13947 * stmt.c (conditional_probability): Ditto.
13948 (emit_case_dispatch_table): Ditto.
13949 * lto-cgraph.c (merge_profile_summaries): Ditto.
13950 * tree-optimize.c (execute_fixup_cfg): Ditto.
13951 * cfgcleanup.c (try_forward_edges): Ditto.
13952 * cfgloopmanip.c (scale_loop_profile): Ditto.
13954 (duplicate_loop_to_header_edge): Ditto.
13955 (lv_adjust_loop_entry_edge): Ditto.
13956 * tree-vect-loop.c (vect_transform_loop): Ditto.
13957 * profile.c (compute_branch_probabilities): Ditto.
13958 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
13959 * lto-streamer-in.c (input_cfg): Ditto.
13960 * gimple-streamer-in.c (input_bb): Ditto.
13961 * ipa-cp.c (update_profiling_info): Ditto.
13962 (update_specialized_profile): Ditto.
13963 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
13964 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
13966 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13967 (compute_trg_info): Ditto.
13968 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
13969 (purge_dead_edges): Ditto.
13970 * loop-unswitch.c (unswitch_loop): Ditto.
13971 * cgraphclones.c (cgraph_clone_edge): Ditto.
13972 (cgraph_clone_node): Ditto.
13973 * tree-inline.c (copy_bb): Ditto.
13974 (copy_edges_for_bb): Ditto.
13975 (initialize_cfun): Ditto.
13976 (copy_cfg_body): Ditto.
13977 (expand_call_inline): Ditto.
13979 2013-04-08 Kai Tietz <ktietz@redhat.com>
13981 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
13982 TARGET_CYGWIN64 by TARGET_64BIT.
13984 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13986 * config/epiphany/epiphany.md (GPR_1): New constant.
13987 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
13988 * config/epiphany/epiphany.c (gen_compare_reg):
13989 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
13990 is already in place.
13991 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
13992 Don't require being called during rtl expansion; If y operlaps r0,
13994 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
13995 (epiphany_expand_epilogue): Likewise.
13997 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
13998 Don't use CC_FPmode for ORDERED / UNORDERED.
13999 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
14001 * config/epiphany/constraints.md (CnL): New constraint.
14002 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
14003 * config/epiphany/predicates.md (add_operand): Allow 1024.
14005 * config/epiphany/epiphany.md (logical_op): New code iterator.
14006 (op_mnc): New code attribute.
14007 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
14008 (mov_f+1, mov_f+2): New peephole2 patterns.
14010 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
14011 (cstoresi4): Also allow re-use of zero result when doing a NE
14012 comparison to a non-zero operand.
14013 Use (clobber (scratch)) for first insn if the gpr output is not needed.
14015 * config/epiphany/epiphany.md (<insn_opname>v2si3):
14016 Use gen_addsi3_i / gen_subsi3_i.
14018 2013-04-08 Jakub Jelinek <jakub@redhat.com>
14022 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
14023 contain anything but clobbers, at most one __builtin_stack_restore,
14024 optionally debug stmts and final resx, and if it has at least one
14025 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
14026 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
14027 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
14028 which isn't defaut definition, remove them.
14029 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
14030 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
14031 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
14032 with MEM_REF LHS with SSA_NAME address.
14034 2013-04-08 Jeff Law <law@redhat.com>
14036 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
14038 2013-04-08 Richard Biener <rguenther@suse.de>
14040 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
14042 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
14043 determined vector type.
14044 (vect_analyze_data_refs): Likewise.
14045 (vect_get_new_vect_var): Adjust.
14046 (vect_create_destination_var): Preserve SSA name versions.
14047 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
14048 not dump anything here.
14050 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14052 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
14053 Add member lr_slot_known.
14054 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
14056 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14057 Remove code that sets lr_slot_offset according to what a previous
14058 version of epiphany_emit_save_restore used to do.
14059 (epiphany_emit_save_restore): When doing an lr save or restore,
14060 set/verify lr_slot_known and lr_slot_offset.
14062 2013-04-08 Xinyu Qi <xyqi@marvell.com>
14065 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
14068 2013-04-08 Richard Biener <rguenther@suse.de>
14070 * alias.c (find_base_term): Fix thinko in previous change.
14072 2013-04-08 Jakub Jelinek <jakub@redhat.com>
14074 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
14075 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
14076 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
14077 if possible to compute val.
14078 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
14079 For QImode integers don't require anything about precision. Use
14080 const_with_all_bytes_same to find out if the constant doesn't have
14081 repeated bytes in it.
14083 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14085 * config/s390/s390.c (s390_expand_insv): Only accept insertions
14088 2013-04-08 Marek Polacek <polacek@redhat.com>
14090 PR rtl-optimization/48182
14091 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
14094 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14097 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
14098 nuses, make sure we have a label.
14100 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14103 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
14104 (rs6000_emit_swdiv_low_precision): Remove.
14105 (rs6000_emit_swdiv): Rewrite to handle between one and four
14106 iterations of Newton-Raphson generally; modify required number of
14107 iterations for some cases.
14108 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
14110 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
14112 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
14113 set-but-unused variable.
14115 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
14116 basic blocks of released function bodies garbage-collectable.
14118 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
14119 (struct rtl_opt_pass): Add TODO_df_finish.
14121 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
14123 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14125 * config/arm/constraints.md (q): New constraint.
14126 * config/arm/ldrdstrd.md: New file.
14127 * config/arm/arm.md (ldrdstrd.md) New include.
14128 (arm_movdi): Use "q" instead of "r" constraint
14129 for double-word memory access.
14130 (movdf_soft_insn): Likewise.
14131 * config/arm/vfp.md (movdi_vfp): Likewise.
14132 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
14133 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
14134 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
14135 (mem_ok_for_ldrd_strd): Likewise.
14136 (output_move_double): Update assertion.
14138 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14140 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
14142 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14144 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
14145 define_insn_and_split.
14146 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
14148 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14150 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
14151 define_insn_and_split.
14152 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
14153 (shiftsi3_compare): New pattern.
14154 (rrx): New pattern.
14155 * config/arm/unspecs.md (UNSPEC_RRX): New.
14157 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14159 * config/arm/arm.md (negdi_extendsidi): New pattern.
14160 (negdi_zero_extendsidi): Likewise.
14162 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14164 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
14165 define_insn_and_split.
14166 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
14167 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
14169 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14171 * config/arm/arm.md (arm_subdi3): Convert define_insn into
14172 define_insn_and_split.
14173 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
14174 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
14176 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14178 * config/arm/arm.md (subsi3_carryin): New pattern.
14179 (subsi3_carryin_const): Likewise.
14180 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
14181 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
14183 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14185 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
14187 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14189 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
14190 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
14192 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14194 * config/arm/arm.c (arm_expand_builtin): Change fcode
14195 type to unsigned int.
14197 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14199 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
14201 2013-04-04 Ian Lance Taylor <iant@google.com>
14203 * doc/standards.texi (Standards): The Go frontend supports the Go 1
14206 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
14208 PR middle-end/56729
14209 * df-scan.c (df_insn_delete): Disable failing assert.
14211 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14213 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
14214 New function prototype.
14215 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
14216 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
14217 (arm_builtin_vectorized_function): New function.
14219 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14221 * config/arm/arm_neon_builtins.def: New file.
14222 * config/arm/arm.c (neon_builtin_data): Move contents to
14223 arm_neon_builtins.def.
14224 (enum arm_builtins): Include neon builtin definitions.
14225 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
14226 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
14228 2013-04-04 Marek Polacek <polacek@redhat.com>
14230 PR tree-optimization/48186
14231 * predict.c (maybe_hot_frequency_p): Return false if
14232 HOT_BB_FREQUENCY_FRACTION is 0.
14233 (cgraph_maybe_hot_edge_p): Likewise.
14235 2013-04-04 Richard Biener <rguenther@suse.de>
14237 PR tree-optimization/56826
14238 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
14241 2013-04-04 Richard Biener <rguenther@suse.de>
14243 PR tree-optimization/56213
14244 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
14245 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
14247 2013-04-04 Richard Biener <rguenther@suse.de>
14249 PR tree-optimization/56837
14250 * tree-loop-distribution.c (classify_partition): For non-zero
14251 values require that the value has the same precision as its
14252 mode to be useful as memset value.
14254 2013-04-03 Nick Clifton <nickc@redhat.com>
14256 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
14257 (fmssf4): Use fmsf.s on E3V5 architectures.
14258 (fnmasf4): Use fnmaf.s on E3V5 architectures.
14259 (fnmssf4): Use fnmsf.s on E3V5 architectures.
14261 2013-04-03 Jeff Law <law@redhat.com>
14263 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
14264 (lra-eliminations.o): Likewise.
14266 2013-04-03 Teresa Johnson <tejohnson@google.com>
14268 * gcov-io.c (compute_working_sets): Moved most of body of old
14269 compute_working_sets here from profile.c.
14270 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
14271 (gcov_working_set_t): Moved typedef here from basic-block.h
14272 (compute_working_set): Declare.
14273 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
14274 (get_working_sets): Renamed from compute_working_set,
14275 replace most of body with call to new compute_working_sets.
14276 (get_exec_counts): Replace call to compute_working_sets
14277 to get_working_sets.
14278 * profile.h (get_working_sets): Renamed from compute_working_set.
14279 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
14280 to get_working_sets.
14281 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
14282 * gcov-dump.c (dump_working_sets): New function.
14284 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
14286 * hwint.c (sext_hwi, zext_hwi): New functions.
14287 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
14288 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
14289 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
14290 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
14291 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
14292 (sext_hwi, zext_hwi): New functions.
14294 2013-04-03 Jeff Law <law@redhat.com>
14296 PR tree-optimization/56799
14297 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
14298 back test for widening conversion erroneously dropped in prior change.
14300 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14303 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
14304 instead of next_real_insn.
14306 2013-04-03 Marek Polacek <polacek@redhat.com>
14309 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
14311 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14314 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
14316 (thumb1_output_casesi): Likewise.
14317 (thumb2_output_casesi): Likewise.
14319 2013-04-03 Richard Biener <rguenther@suse.de>
14321 PR tree-optimization/56817
14322 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
14324 (tree_unroll_loops_completely_1): ... new function to manually
14325 walk the loop tree, properly defering outer loops of unrolled
14326 loops to later iterations.
14328 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14330 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
14331 (vectorizable_load): Likewise.
14332 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
14333 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
14335 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14337 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
14340 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
14342 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
14344 2013-04-03 Bin Cheng <bin.cheng@arm.com>
14346 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
14348 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14350 PR tree-optimization/56790
14351 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
14354 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
14356 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
14358 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
14359 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
14362 2013-04-03 Jakub Jelinek <jakub@redhat.com>
14365 * tree.h (force_folding_builtin_constant_p): New decl.
14366 * builtins.c (force_folding_builtin_constant_p): New variable.
14367 (fold_builtin_constant_p): Fold immediately also if
14368 force_folding_builtin_constant_p.
14370 2013-04-03 Richard Biener <rguenther@suse.de>
14372 PR tree-optimization/56812
14373 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
14374 DRs of the same interleaving chain are independent.
14376 2013-04-02 Jason Merrill <jason@redhat.com>
14378 * gdbinit.in (pbb): Use debug fn.
14380 2013-04-02 Lawrence Crowl <crowl@google.com>
14382 * sese.h (struct ivtype_map_elt_s): Remove unused.
14383 (extern debug_ivtype_map): Remove unused.
14384 (extern eq_ivtype_map_elts): Remove unused.
14385 * sese.c (debug_ivtype_map): Removed unused.
14386 (debug_ivtype_map_1): Removed unused.
14387 (debug_ivtype_elt): Remove unused.
14388 (eq_ivtype_map_elts): Remove unused.
14391 2013-04-02 Kai Tietz <ktietz@redhat.com>
14394 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
14395 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
14396 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
14398 (legitimize_pe_coff_symbol): Likewise.
14399 (is_imported_p): New helper-function.
14400 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
14401 for Windows x64 targets.
14402 (ix86_expand_prologue): Optimize for pe-coff targets.
14403 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
14404 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
14405 medium/large code-model.
14406 (legitimize_pic_address): Likewise.
14407 (legitimize_tls_address): Likewise.
14408 (ix86_expand_call): Likewise.
14409 (x86_output_mi_thunk): Likewise.
14410 (get_dllimport_decl): Add new beimport argument.
14411 (construct_plt_address): Don't assert for x64 pe-coff targets.
14412 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
14414 (SYMBOL_FLAG_STUBVAR): New macro.
14415 (SYMBOL_REF_STUBVAR_P): Likewise.
14416 * config/i386/winnt.c (stub_list): New structure.
14417 (stub_head): New local variable.
14418 (i386_pe_record_stub): New function.
14419 (i386_pe_file_end): Emit refptr-stubs.
14421 2013-04-02 Jakub Jelinek <jakub@redhat.com>
14423 PR rtl-optimization/56745
14424 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
14425 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
14428 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
14429 and both of them are MEM_REFs, just compare first argument for
14430 equality and attempt to deal even with differing offsets.
14433 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
14434 of gimple_clobber_p to be MEM_REF.
14435 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
14436 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
14437 after gimplification.
14438 * asan.c (get_mem_ref_of_assignment): Don't instrument
14439 gimple_clobber_p stmts.
14440 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
14441 gimple_clobber_p stmt if they have MEM_REF lhs and
14442 are dead because of another gimple_clobber_p stmt.
14443 * tree-ssa-live.c (clear_unused_block_pointer): Treat
14444 gimple_clobber_p stmts like debug stmts.
14445 (remove_unused_locals): Remove clobbers with MEM_REF lhs
14446 that refer to unused VAR_DECLs or uninitialized values.
14447 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
14448 gimple_clobber_p stmts if they refer to removed parameters.
14449 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
14452 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
14454 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
14455 using SWI48 mode attribute.
14457 2013-04-02 Wei Mi <wmi@google.com>
14459 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
14460 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
14461 *<rotate_insn><mode>3_mask in i386.md.
14463 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
14465 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
14467 2013-04-02 Richard Biener <rguenther@suse.de>
14469 PR tree-optimization/56778
14470 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
14471 Runtime alias tests are not supported for gather loads.
14472 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
14473 stmts referenced from SSA operands before updating SSA form.
14475 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
14476 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14478 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
14479 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
14480 * config/arm/cortex-a53.md: New file.
14481 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
14482 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
14483 * config/arm/arm.c (arm_issue_rate): Likewise.
14484 * config/arm/arm-tune.md: Regenerate
14485 * config/arm/arm-tables.opt: Regenerate.
14486 * config/arm/arm-cores.def: Add cortex-a53.
14488 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
14490 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
14493 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
14495 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
14496 scalar load/store operations using B/H registers.
14497 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
14499 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
14501 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
14503 * config/aarch64/aarch64.c
14504 (aarch64_simd_scalar_immediate_valid_for_move): New.
14505 * config/aarch64/aarch64-protos.h
14506 (aarch64_simd_scalar_immediate_valid_for_move): New.
14507 * config/aarch64/constraints.md (Dh, Dq): New.
14508 * config/aarch64/iterators.md (hq): New.
14510 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
14512 * reorg.c (get_branch_condition): Deal with conditional returns.
14513 (fill_simple_delay_slots): Remove dead code dealing with jumps.
14515 2013-04-01 Wei Mi <wmi@google.com>
14517 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
14518 Truncate operand 2 using %b asm operand modifier.
14519 (*<shift_insn><mode>3_mask): Ditto.
14520 (*<rotate_insn><mode>3_mask): Ditto.
14522 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
14524 PR middle-end/56798
14525 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
14527 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
14529 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
14531 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
14533 2013-03-30 Lawrence Crowl <crowl@google.com>
14535 * dse.c (clear_alias_sets): Remove never set.
14536 (disqualified_clear_alias_sets): Remove never set.
14537 (clear_alias_mode_pool): Remove never set.
14538 (dse_step0): Remove condition that is never true.
14539 (canon_address): Remove condition that is never true.
14540 (dse_step7): Remove condition that is never true.
14541 (rest_of_handle_dse): Remove condition that is never true.
14542 (rest_of_handle_dse::did_global): Remove never read from above.
14543 (dse_step2_spill): Remove never called from above.
14544 (dse_step5_spill): Remove never called from above.
14546 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
14548 * doc/md.texi (Standard Names) <casesi>: Update documentation for
14549 JUMP_TABLE_DATA changes.
14550 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
14551 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
14552 (Insns) <jump_table_data>: New entry.
14553 * doc/tm.texi: Regenerate.
14555 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
14557 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
14558 for table jump at the end of a basic block using tablejump_p.
14559 * targhooks.c (default_invalid_within_doloop): Likewise.
14560 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
14561 target hook implementation that is identical to the default hook.
14562 (rs6000_invalid_within_doloop): Remove.
14564 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
14565 unused variable from tablejump_p call.
14567 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
14568 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
14569 (INSN_DELETED_P): Likewise.
14570 (emit_jump_table_data): New prototype.
14571 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
14572 after 4th as unused.
14573 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
14574 * sched-vis.c (print_insn): Likewise.
14575 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
14576 insn for compatibility with back ends that use next_active_insn to
14577 identify jump table data.
14578 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
14579 (remove_insn): Likewise.
14580 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
14582 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
14583 (emit_jump_table_data): New function.
14585 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
14586 basic block, a JUMP_TABLE_DATA never is.
14587 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
14588 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
14589 off from code handling real insns.
14590 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
14591 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
14592 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
14593 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
14594 is not a NONDEBUG_INSN_P.
14595 * ira-costs.c (scan_one_insn): Likewise.
14596 * jump.c (mark_all_labels): Likewise.
14597 (mark_jump_label_1): Likewise.
14598 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
14599 * lra.c (get_insn_freq): Expect all insns reaching here to be in
14601 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
14602 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
14603 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
14604 JUMP_TABLE_DATA_P insns.
14605 (calculate_elim_costs_all_insns): Likewise.
14606 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
14607 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
14608 (delete_output_reload): Code style fixups.
14609 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
14610 insn flags on this non-insn.
14611 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
14612 as scheduling barriers, for pre-change compatibility.
14613 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
14614 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
14616 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
14617 redundant JUMP_TABLE_DATA_P test.
14618 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
14619 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
14620 (frv_for_each_packet): Likewise.
14621 * config/i386/i386.c (min_insn_size): Likewise.
14622 (ix86_avoid_jump_mispredicts): Likewise.
14623 * config/m32r/m32r.c (m32r_is_insn): Likewise.
14624 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
14625 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
14626 (mips16_insn_length): Robustify.
14627 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
14628 (mips16_split_long_branches): Likewise.
14629 * config/pa/pa.c (pa_combine_instructions): Likewise.
14630 * config/rs6000/rs6000.c (get_next_active_insn): Treat
14631 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
14632 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
14633 as contributing to pool range lengths.
14634 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
14635 Remove redundant JUMP_TABLE_DATA_P test.
14636 (sh_loop_align): Likewise.
14637 (split_branches): Likewise.
14638 (sh_insn_length_adjustment): Likewise.
14639 * config/spu/spu.c (get_branch_target): Likewise.
14641 2013-03-29 Jan Hubicka <jh@suse.cz>
14643 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
14644 gcov streaming; stream hot bb threshold to ltrans.
14645 * predict.c (get_hot_bb_threshold): Break out from ....
14646 (maybe_hot_count_p): ... here.
14647 (set_hot_bb_threshold): New function.
14648 * lto-section-in.c (lto_section_name): Add profile.
14649 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
14650 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
14651 and data-streamer.h
14652 (histogram_entry): New structure.
14653 (histogram, histogram_pool): New global vars.
14654 (histogram_hash): New structure.
14655 (histogram_hash::hash): New method.
14656 (histogram_hash::equal): Likewise.
14657 (account_time_size): New function.
14658 (cmp_counts): New function.
14659 (dump_histogram): New function.
14660 (ipa_profile_generate_summary): New function.
14661 (ipa_profile_write_summary): New function.
14662 (ipa_profile_read_summary): New function.
14663 (ipa_profile): Decide on threshold.
14664 (pass_ipa_profile): Add ipa_profile_write_summary and
14665 ipa_profile_read_summary.
14666 * Makefile.in (ipa.o): Update dependencies.
14667 * lto-streamer.h (LTO_section_ipa_profile): New section.
14669 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
14671 * tree.h (VAR_P): New.
14673 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
14676 * doc/invoke.texi ([-fwhole-program]): Fix typo.
14678 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
14680 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
14681 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
14682 (control_flow_insn_p): Likewise.
14683 * cfgrtl.c (duplicate_insn_chain): Likewise.
14684 * final.c (get_attr_length_1): Likewise.
14685 (shorten_branches): Likewise.
14686 (final_scan_insn): Likewise.
14687 * function.c (instantiate_virtual_regs): Likewise.
14688 * gcse.c (insert_insn_end_basic_block): Likewise.
14689 * ira-costs.c (scan_one_insn): Likewise.
14690 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
14691 * lra.c (check_rtl): Likewise.
14692 * reload1.c (elimination_costs_in_insn): Likewise.
14693 * reorg.c (follow_jumps): Likewise.
14695 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
14696 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
14697 (thumb_far_jump_used_p): Likewise.
14698 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
14699 (workaround_speculation): Likewise.
14700 (add_sched_insns_for_speculation): Likewise.
14701 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
14702 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
14703 (frv_for_each_packet): Likewise.
14704 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
14705 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
14706 (final_emit_insn_group_barriers): Likewise.
14707 * config/m32r/m32r.c (m32r_is_insn): Likewise.
14708 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
14709 (mips16_insn_length): Likewise.
14710 * config/pa/pa.c (pa_reorg): Likewise.
14711 (pa_combine_instructions): Likewise.
14712 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
14713 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
14714 (sh_reorg): Likewise.
14715 (split_branches): Likewise.
14716 * config/spu/spu.c (get_branch_target): Likewise.
14718 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
14721 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
14723 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
14724 Fix declaration name.
14726 2013-03-28 Lawrence Crowl <crowl@google.com>
14728 * graphds.h (struct graph.indicies): Remove unused.
14729 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
14730 (SCOP_ORIGINAL_PDDRS): Remove unused.
14731 * sese.h (extern insert_loop_close_phis): Removed unused.
14732 (extern insert_guard_phis): Removed unused.
14733 (extern ivtype_map_elt_info): Removed unused.
14734 (new_ivtype_map_elt): Removed unused.
14735 * sese.c (ivtype_map_elt_info): Removed unused.
14737 2013-03-28 Lawrence Crowl <crowl@google.com>
14739 * Makefile.in: Add several missing include dependences.
14741 (test-dump.o): New. This object is not added to any executable,
14742 but is present for ad-hoc testing.
14744 (debug (const bitmap_head_def &)): New.
14745 (debug (const bitmap_head_def *)): New.
14747 (extern debug (const bitmap_head_def &)): New.
14748 (extern debug (const bitmap_head_def *)): New.
14750 (debug (edge_def &)): New.
14751 (debug (edge_def *)): New.
14753 (debug (basic_block_def &)): New.
14754 (debug (basic_block_def *)): New.
14756 (dump_node (const_tree, int, FILE *)): Correct source file.
14758 (debug (die_struct &)): New.
14759 (debug (die_struct *)): New.
14761 (extern debug (die_struct &)): New.
14762 (extern debug (die_struct *)): New.
14763 * gimple-pretty-print.c
14764 (debug (gimple_statement_d &)): New.
14765 (debug (gimple_statement_d *)): New.
14766 * gimple-pretty-print.h
14767 (extern debug (gimple_statement_d &)): New.
14768 (extern debug (gimple_statement_d *)): New.
14770 (debug (ira_allocno_copy &)): New.
14771 (debug (ira_allocno_copy *)): New.
14772 (debug (ira_allocno &)): New.
14773 (debug (ira_allocno *)): New.
14775 (extern debug (ira_allocno_copy &)): New.
14776 (extern debug (ira_allocno_copy *)): New.
14777 (extern debug (ira_allocno &)): New.
14778 (extern debug (ira_allocno *)): New.
14780 (debug (live_range &)): New.
14781 (debug (live_range *)): New.
14783 (debug (lra_live_range &)): New.
14784 (debug (lra_live_range *)): New.
14786 (debug (lra_live_range &)): New.
14787 (debug (lra_live_range *)): New.
14789 (debug (omega_pb_d &)): New.
14790 (debug (omega_pb_d *)): New.
14792 (extern debug (omega_pb_d &)): New.
14793 (extern debug (omega_pb_d *)): New.
14795 (debug (const rtx_def &)): New.
14796 (debug (const rtx_def *)): New.
14798 (debug_tree (tree): Move within file.
14799 (debug_raw (const tree_node &)): New.
14800 (debug_raw (const tree_node *)): New.
14801 (dump_tree_via_hooks (const tree_node *, int)): New.
14802 (debug (const tree_node &)): New.
14803 (debug (const tree_node *)): New.
14804 (debug_verbose (const tree_node &)): New.
14805 (debug_verbose (const tree_node *)): New.
14806 (debug_head (const tree_node &)): New.
14807 (debug_head (const tree_node *)): New.
14808 (debug_body (const tree_node &)): New.
14809 (debug_body (const tree_node *)): New.
14810 (debug_vec_tree (tree): Move and reimplement in terms of dump.
14811 (debug (vec<tree, va_gc> &)): New.
14812 (debug (vec<tree, va_gc> *)): New.
14814 (extern debug (const rtx_def &)): New.
14815 (extern debug (const rtx_def *)): New.
14817 (debug_raw (simple_bitmap_def &)): New.
14818 (debug_raw (simple_bitmap_def *)): New.
14819 (debug (simple_bitmap_def &)): New.
14820 (debug (simple_bitmap_def *)): New.
14822 (extern debug (simple_bitmap_def &)): New.
14823 (extern debug (simple_bitmap_def *)): New.
14824 (extern debug_raw (simple_bitmap_def &)): New.
14825 (extern debug_raw (simple_bitmap_def *)): New.
14827 (debug (vinsn_def &)): New.
14828 (debug (vinsn_def *)): New.
14829 (debug_verbose (vinsn_def &)): New.
14830 (debug_verbose (vinsn_def *)): New.
14831 (debug (expr_def &)): New.
14832 (debug (expr_def *)): New.
14833 (debug_verbose (expr_def &)): New.
14834 (debug_verbose (expr_def *)): New.
14835 (debug (vec<rtx> &)): New.
14836 (debug (vec<rtx> *)): New.
14838 (extern debug (vinsn_def &)): New.
14839 (extern debug (vinsn_def *)): New.
14840 (extern debug_verbose (vinsn_def &)): New.
14841 (extern debug_verbose (vinsn_def *)): New.
14842 (extern debug (expr_def &)): New.
14843 (extern debug (expr_def *)): New.
14844 (extern debug_verbose (expr_def &)): New.
14845 (extern debug_verbose (expr_def *)): New.
14846 (extern debug (vec<rtx> &)): New.
14847 (extern debug (vec<rtx> *)): New.
14849 (_list_iter_cond_expr): Make inline instead of static.
14851 (debug (sreal &)): New.
14852 (debug (sreal *)): New.
14854 (extern debug (sreal &)): New.
14855 (extern debug (sreal *)): New.
14857 (extern debug_raw (const tree_node &)): New.
14858 (extern debug_raw (const tree_node *)): New.
14859 (extern debug (const tree_node &)): New.
14860 (extern debug (const tree_node *)): New.
14861 (extern debug_verbose (const tree_node &)): New.
14862 (extern debug_verbose (const tree_node *)): New.
14863 (extern debug_head (const tree_node &)): New.
14864 (extern debug_head (const tree_node *)): New.
14865 (extern debug_body (const tree_node &)): New.
14866 (extern debug_body (const tree_node *)): New.
14867 (extern debug (vec<tree, va_gc> &)): New.
14868 (extern debug (vec<tree, va_gc> *)): New.
14870 (debug (struct loop &)): New.
14871 (debug (struct loop *)): New.
14872 (debug_verbose (struct loop &)): New.
14873 (debug_verbose (struct loop *)): New.
14874 * tree-dump.c: Add header dependence.
14876 (extern debug (struct loop &)): New.
14877 (extern debug (struct loop *)): New.
14878 (extern debug_verbose (struct loop &)): New.
14879 (extern debug_verbose (struct loop *)): New.
14881 (debug (data_reference &)): New.
14882 (debug (data_reference *)): New.
14883 (debug (vec<data_reference_p> &)): New.
14884 (debug (vec<data_reference_p> *)): New.
14885 (debug (vec<ddr_p> &)): New.
14886 (debug (vec<ddr_p> *)): New.
14888 (extern debug (data_reference &)): New.
14889 (extern debug (data_reference *)): New.
14890 (extern debug (vec<data_reference_p> &)): New.
14891 (extern debug (vec<data_reference_p> *)): New.
14892 (extern debug (vec<ddr_p> &)): New.
14893 (extern debug (vec<ddr_p> *)): New.
14895 (debug (pt_solution &)): New.
14896 (debug (pt_solution *)): New.
14898 (extern debug (pt_solution &)): New.
14899 (extern debug (pt_solution *)): New.
14901 (debug (_var_map &)): New.
14902 (debug (_var_map *)): New.
14903 (debug (tree_live_info_d &)): New.
14904 (debug (tree_live_info_d *)): New.
14906 (extern debug (_var_map &)): New.
14907 (extern debug (_var_map *)): New.
14908 (extern debug (tree_live_info_d &)): New.
14909 (extern debug (tree_live_info_d *)): New.
14911 2013-03-28 Jan Hubicka <jh@suse.cz>
14913 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
14915 2013-03-28 Ian Bolton <ian.bolton@arm.com>
14917 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
14918 record only when desired or required.
14920 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
14922 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
14923 *vec_extractv2di_1_rex64. Use x64 isa attribute.
14925 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14927 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
14928 (*andsi3_compare0_uxtw): New pattern.
14929 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
14930 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
14932 2013-03-28 Jan Hubicka <jh@suse.cz>
14934 * data-streamer-in.c (streamer_read_gcov_count): New function.
14935 * gimple-streamer-out.c: Include value-prof.h.
14936 (output_gimple_stmt): Output histogram.
14937 (output_bb): Use streamer_write_gcov_count.
14938 * value-prof.c: Include data-streamer.h
14939 (dump_histogram_value): Add HIST_TYPE_MAX.
14940 (stream_out_histogram_value): New function.
14941 (stream_in_histogram_value): New function.
14942 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
14943 (stream_out_histogram_value, stream_in_histogram_value): Declare.
14944 * data-streamer-out.c (streamer_write_gcov_count): New function.
14945 (streamer_write_gcov_count_stream): New function.
14946 * lto-cgraph.c (lto_output_edge): Update counter streaming.
14947 (lto_output_node): Likewise.
14948 (input_node, input_edge): Likewise.
14949 * lto-streamer-out.c (output_cfg): Update streaming.
14950 * lto-streamer-in.c (input_cfg): Likewise.
14951 * data-streamer.h (streamer_write_gcov_count,
14952 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
14953 * gimple-streamer-in.c: Include value-prof.h
14954 (input_gimple_stmt): Input histograms.
14955 (input_bb): Update profile streaming.
14957 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14959 * genmodes.c (emit_max_int): New function.
14960 (emit_insn_modes_h): Added call to emit_max_function.
14961 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
14963 * machmode.def: Fixed comment.
14965 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14967 * combine.c (try_combine): Removed useless assert.
14968 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
14970 2013-03-28 Marek Polacek <polacek@redhat.com>
14971 Richard Biener <rguenther@suse.de>
14973 PR tree-optimization/56695
14974 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
14975 build signed result of a vector comparison.
14976 * tree-cfg.c (verify_gimple_comparison): Check that a result
14977 of a vector comparison has signed type.
14979 2013-03-28 Richard Biener <rguenther@suse.de>
14981 PR tree-optimization/37021
14982 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
14983 do not restrict gaps between groups.
14984 * tree-vect-stmts.c (vectorizable_load): Properly account for
14985 a gap between groups.
14987 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
14989 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
14990 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
14993 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
14995 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
14996 * doc/extend.texi (Named Address Spaces): Ditto.
14997 (Variable Attributes): Ditto.
14999 2013-03-27 Kai Tietz <ktietz@redhat.com>
15001 * config.build: Add support for cygwin x64 target.
15002 * config.gcc: Likewise.
15003 * config.host: Likewise.
15004 * configure.ac: Likewise
15005 * configure: Regenerated.
15007 2013-03-27 Kai Tietz <ktietz@redhat.com>
15009 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
15010 * config/i386/t-cygwin-w64: New file.
15011 * config/i386/cygwin-w64.h: New file.
15012 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
15013 and add support for x64-cygwin target.
15014 (CPP_SPEC): Likewise.
15015 (CXX_WRAP_SPEC_LIST): Undefine before define.
15016 (LIBGCJ_SONAME): Use 15 as version.
15018 2013-03-27 Richard Biener <rguenther@suse.de>
15020 PR tree-optimization/56716
15021 * tree-ssa-structalias.c (perform_var_substitution): Adjust
15022 dumping for ref nodes.
15024 2013-03-27 Martin Jambor <mjambor@suse.cz>
15026 PR tree-optimization/55334
15027 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
15028 restricted pointers to arrays.
15030 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
15032 * Makefile.in (.SUFFIXES): Add .cc.
15033 (.c.o): Apply same recipe for implicit rule .cc.o.
15035 2013-03-27 Richard Biener <rguenther@suse.de>
15037 PR tree-optimization/37021
15038 * tree-vect-data-refs.c (vect_check_strided_load): Allow
15039 REALPART/IMAGPART_EXPRs around the supported refs.
15040 * tree-ssa-structalias.c (find_func_aliases): Assume that
15041 floating-point values are not used to transfer pointers.
15043 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
15045 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
15046 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
15047 * gcc/doc/tm.texi: Regenerate.
15048 * gcc/targhooks.h (default_has_ifunc_p): New.
15049 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
15050 * gcc/config/linux-protos.h: New file.
15051 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
15052 this hook for linux which disables support of indirect functions in
15054 * gcc/config/linux-android.c: New file.
15055 * gcc/config/t-linux-android.c: Ditto.
15056 * gcc/config.gcc: Added new object file linux-android.o.
15057 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
15058 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
15059 * gcc/varasm.c (do_assemble_alias): Likewise.
15060 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
15061 doesn't support indirect functions.
15062 * configure: Regenerate.
15064 2013-03-27 Bin Cheng <bin.cheng@arm.com>
15067 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
15068 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
15071 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15073 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
15075 2013-03-27 Terry Guo <terry.guo@arm.com>
15077 * config/arm/arm-cores.def: Added core cortex-r7.
15078 * config/arm/arm-tune.md: Regenerated.
15079 * config/arm/arm-tables.opt: Regenerated.
15080 * doc/invoke.texi: Added entry for core cortex-r7.
15082 2013-03-27 Walter Lee <walt@tilera.com>
15084 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
15085 double-decrement of next_scratch_regno.
15087 2013-03-27 Walter Lee <walt@tilera.com>
15089 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
15091 (insn_v1mulus): Ditto.
15092 (insn_v2muls): Ditto.
15094 2013-03-27 Walter Lee <walt@tilera.com>
15096 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
15097 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
15099 2013-03-27 Walter Lee <walt@tilera.com>
15101 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
15102 (*sibcall_value): Ditto.
15104 2013-03-27 Walter Lee <walt@tilera.com>
15106 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
15107 (insn_mnz_v8qi): ... this ...
15108 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
15110 (insn_v<n>mnz): Replaced by ...
15111 (insn_v1mnz): ... this ...
15112 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
15114 (insn_mz_<mode>): Replaced by ...
15115 (insn_mz_v8qi): ... this ...
15116 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
15118 (insn_v<n>mz): Replaced by ...
15119 (insn_v1mz): ... this ...
15120 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
15123 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
15125 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
15127 2013-03-26 Roland McGrath <mcgrathr@google.com>
15129 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
15130 than fprintf with a non-constant, non-format string.
15132 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
15134 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
15135 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
15136 operand 0 predicate.
15137 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
15138 attribute. Use general_x64nomem_operand as operand 1 predicate.
15139 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
15140 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15141 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
15142 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15143 (mov<mode>_insv_1): Remove expander. Merge insn with
15144 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
15145 Use general_x64nomem_operand as operand 1 predicate.
15146 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
15147 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
15148 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15149 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
15150 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
15151 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
15152 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15153 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
15154 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
15155 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
15156 (general_x64nomem_operand): Ditto.
15158 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
15160 * config/rtems.opt: Add -pthread option.
15162 2013-03-26 Richard Biener <rguenther@suse.de>
15164 * alias.c (find_base_term): Avoid redundant and not used recursion.
15165 (base_alias_check): Get the initial base term from the caller.
15166 (true_dependence_1): Compute and pass base terms to base_alias_check.
15167 (write_dependence_p): Likewise.
15168 (may_alias_p): Likewise.
15170 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
15172 * config/aarch64/aarch64.c (aarch64_classify_address): Support
15173 PC-relative load in SI modes and above only.
15175 2013-03-26 Xinyu Qi <xyqi@marvell.com>
15177 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
15178 * config/arm/iwmmxt.md (WCGR0): Update.
15179 (WCGR1, WCGR2, WCGR3): Likewise.
15181 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
15183 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
15184 Use x64 and nox64 isa attributes.
15186 2013-03-26 Richard Biener <rguenther@suse.de>
15188 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
15189 alignment computations and rely on get_object_alignment_1
15190 for the !TYPE_P case.
15191 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
15193 2013-03-26 Walter Lee <walt@tilera.com>
15195 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
15196 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
15198 2013-03-25 Jeff Law <law@redhat.com>
15200 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
15201 check for INTEGRAL_TYPE_P that was missing due to checking in
15202 wrong version of prior patch.
15204 2013-03-25 Walter Lee <walt@tilera.com>
15206 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
15207 TILEGX_INSN_SHUFFLEBYTES1.
15208 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
15210 (tilegx_builtins): Ditto.
15211 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
15213 2013-03-25 Walter Lee <walt@tilera.com>
15215 * config/tilegx/tilegx.md (floatsisf2): New pattern.
15216 (floatunssisf2): New pattern.
15217 (floatsidf2): New pattern.
15218 (floatunssidf2): New pattern.
15220 2013-03-25 Walter Lee <walt@tilera.com>
15222 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
15223 tests for constraint J, K, N, P.
15225 2013-03-25 Walter Lee <walt@tilera.com>
15227 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
15228 Use indirect/pcrel encoding.
15229 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
15232 2013-03-25 Steve Ellcey <sellcey@mips.com>
15234 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
15235 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
15236 * config/mips/mips.c (mips_option_override): Set IMADD default.
15237 * config/mips/mips.h (PTF_AVOID_IMADD): New.
15238 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
15239 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
15240 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
15241 * doc/invoke.texi (-mimadd/-mno-imadd): New.
15243 2013-03-25 Jeff Law <law@redhat.com>
15245 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
15246 slightly to avoid creating and folding useless trees. Simplify
15247 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
15249 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
15251 * config/i386/i386.md (*zero_extendsidi2): Merge with
15252 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
15253 * config/i386/predicates.md (x86_64_zext_operand): Rename from
15254 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
15255 targets. Clarify comment.
15257 2013-03-25 Martin Jambor <mjambor@suse.cz>
15259 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
15260 pass-through jump functions differently.
15261 (ipa_read_jump_function): Likewise. Also use setter functions to set
15264 2013-03-25 Martin Jambor <mjambor@suse.cz>
15266 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
15267 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
15269 (ipa_get_indirect_edge_target): New function.
15270 (devirtualization_time_bonus): New parameter known_aggs, pass it to
15271 ipa_get_indirect_edge_target. Update all callers.
15272 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
15273 ipa_get_indirect_edge_target_1 instead of calling
15274 ipa_get_indirect_edge_target.
15275 (create_specialized_node): Pass aggvlas to
15276 ipcp_discover_new_direct_edges.
15278 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15280 * config/arm/arm.md (f_sels, f_seld): New types.
15281 (*cmov<mode>): New pattern.
15282 * config/arm/predicates.md (arm_vsel_comparison_operator): New
15285 2013-03-25 Kai Tietz <ktietz@redhat.com>
15287 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
15288 POSIX-printf for mingw-hosted builds.
15290 2013-03-25 Richard Biener <rguenther@suse.de>
15292 PR middle-end/56694
15293 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
15294 must-not-throw stmt location.
15296 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15298 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
15299 Emit load-acquire versions when acq is true.
15300 (arm_emit_store_exclusive): Add rel parameter.
15301 Emit store-release versions when rel is true.
15302 (arm_split_compare_and_swap): Use acquire-release instructions
15304 of barriers when appropriate.
15305 (arm_split_atomic_op): Likewise.
15306 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
15307 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
15308 (VUNSPEC_SLX): Likewise.
15309 (VUNSPEC_LDA): Likewise.
15310 (VUNSPEC_STL): Likewise.
15311 * config/arm/sync.md (atomic_load<mode>): New pattern.
15312 (atomic_store<mode>): Likewise.
15313 (arm_load_acquire_exclusive<mode>): Likewise.
15314 (arm_load_acquire_exclusivesi): Likewise.
15315 (arm_load_acquire_exclusivedi): Likewise.
15316 (arm_store_release_exclusive<mode>): Likewise.
15318 2013-03-25 Catherine Moore <clm@codesourcery.com>
15320 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
15321 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
15322 * config/mip/predicates.md (lwsp_swsp_operand,
15323 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
15324 sb16_operand, db4_operand, db7_operand, ib3_operand,
15325 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
15326 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
15327 andi16_operand): New predicates.
15328 * config/mips/mips.md (compression): New attribute.
15329 (enabled): New attribute.
15330 (length): Consider compression in computing length.
15331 (shift_compression): New code attribute.
15332 (*add<mode>3): New operands. Record compression.
15333 (sub<mode>3): Likewise.
15334 (one_cmpl<mode>2): Likewise.
15335 (*and<mode>3): Likewise.
15336 (*ior<mode>3): Likewise.
15337 (unnamed pattern for xor): Likewise.
15338 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
15339 (*<optab><mode>3): Likewise.
15340 (*mov<mode>_internal: Likewise.
15341 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
15342 (mips_unsigned_immediate_p): New.
15343 (umips_lwsp_swsp_address_p): New.
15344 (m16_based_address_p): New.
15345 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
15346 (mips_unsigned_immediate_p): New prototype.
15347 (lwsp_swsp_address_p): New prototype.
15348 (m16_based_address_p): New prototype.
15349 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
15350 (mips_signed_immediate_p): New function.
15351 (m16_based_address_p): New function.
15352 (lwsp_swsp_address_p): New function.
15353 (mips_print_operand_punctuation): Recognize short delay slot insns
15354 for microMIPS.add<mode>3"
15356 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15359 * config/arm/iterators.md (v_cmp_result): New mode attribute.
15360 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
15362 2013-03-25 Richard Biener <rguenther@suse.de>
15364 PR tree-optimization/56689
15365 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
15368 2013-03-25 Richard Biener <rguenther@suse.de>
15370 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
15372 (memory_references): Likewise.
15373 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
15374 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
15375 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
15376 (gather_mem_refs_in_loops): Fold into ...
15377 (analyze_memory_references): ... this. Move initialization
15378 to tree_ssa_lim_initialize.
15379 (fill_always_executed_in): Rename to ...
15380 (fill_always_executed_in_1): ... this.
15381 (fill_always_executed_in): Move contains_call computation to
15382 this new function from ...
15383 (tree_ssa_lim_initialize): ... here.
15384 (tree_ssa_lim): Call fill_always_executed_in.
15386 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15388 * postreload.c (reload_combine): Fix code detecting returns.
15390 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15392 * function.c (emit_use_return_register_into_block): On cc0 targets,
15393 do not emit the sequence between cc0 setter and user.
15395 2013-03-25 Kai Tietz <ktietz@redhat.com>
15397 * config/i386/predicates.md (local_symbolic_operand): Interpret
15398 dll-imported symbols as none-local.
15400 2013-03-25 Richard Biener <rguenther@suse.de>
15402 * tree-ssa-loop-im.c (struct depend): Remove.
15403 (struct lim_aux_data): Make depends a vec of gimples.
15404 (free_lim_aux_data): Adjust.
15405 (add_dependency): Likewise.
15406 (set_level): Likewise.
15408 2013-03-25 Richard Biener <rguenther@suse.de>
15410 PR middle-end/56434
15411 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
15412 the pointer returned by calls with ECF_MALLOC set.
15414 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
15416 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
15418 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
15420 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
15421 using MMXMODE mode iterator.
15422 (*move<mode>_internal): Merge with *movv2sf_internal and
15423 *movv2sf_internal_rex64 using MMXMODE mode iterator.
15425 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15427 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
15428 (record_last_mem_set_info): Likewise.
15430 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
15431 of XNEWVEC followed by memset.
15432 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
15434 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15436 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
15437 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
15438 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
15439 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
15440 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
15441 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
15442 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
15443 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
15444 BARRIER_P instead of GET_CODE.
15446 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15448 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
15449 inaccuracy in the probing code.
15451 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
15452 (ctrapdi4): Likewise.
15454 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15456 * calls.c (expand_call): Add missing guard to code handling return
15457 of non-BLKmode structures in MSB.
15458 * function.c (expand_function_end): Likewise.
15460 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15462 * combine.c (try_combine): Adjust comment. Do not add the set of
15463 insn #0 if the destination indirectly is set or dies in insn #2.
15464 Tidy up code to distribute a new note.
15466 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
15468 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
15469 also for alternatives 16 and 17.
15471 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
15473 * config/i386/sse.md (*mov<mode>_internal): Merge with
15474 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
15475 Emit insn template depending on type attribute. Use
15476 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
15477 movd instead of movq mnemonic for interunit moves. Rewrite mode
15478 attribute calculation. Remove unit attribute calculation.
15479 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
15480 Set prefix_data16 attribute for DImode ssemov types.
15481 Use Ym instead of y for SSE-MMX conversion alternatives.
15482 Reorder operand constraints.
15484 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
15486 * df.h (df_insn_delete): Adjust prototype.
15487 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
15488 and let it decide whether mark the basic block dirty.
15489 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
15490 * df-scan.c (df_insn_info_delete): New helper function, split
15491 off from df_insn_delete.
15492 (df_scan_free_bb_info): Use it.
15493 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
15495 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
15496 that the insn is actually an insn and it has a non-NULL basic block.
15497 Do not mark basic block dirty if only deleting a DEBUG_INSN.
15499 2013-03-22 Richard Biener <rguenther@suse.de>
15501 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
15503 (mem_ref_alloc): Do not allocate them.
15504 (refs_independent_p): Do not query or maintain a cache.
15506 2013-03-22 Richard Biener <rguenther@suse.de>
15508 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
15509 (gather_mem_refs_in_loops): Do not compute it.
15510 (analyze_memory_references): Do not allocate it.
15511 (tree_ssa_lim_finalize): Do not free it.
15512 (for_all_locs_in_loop): Do not query all_refs_in_loop.
15514 2013-03-22 Richard Biener <rguenther@suse.de>
15516 * is-a.h (as_a): Use gcc_checking_assert.
15518 2013-03-22 Ian Bolton <ian.bolton@arm.com>
15520 * config/aarch64/aarch64.c (aarch64_print_operand): New
15521 format specifier for printing a constant in hex.
15522 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
15523 format specifier for printing second operand.
15525 2013-03-22 Richard Biener <rguenther@suse.de>
15527 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
15529 (gather_mem_refs_in_loops): Perform store accumulation here.
15530 (create_vop_ref_mapping_loop): Remove.
15531 (create_vop_ref_mapping): Likewise.
15532 (analyze_memory_references): Initialize refs_stored_in_loop.
15533 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
15534 (record_indep_loop): Remove.
15535 (record_dep_loop): New function.
15536 (ref_indep_loop_p_1): Adjust to only walk over references
15537 in the loop, not its subloops.
15538 (ref_indep_loop_p): Rename to ...
15539 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
15540 maintaining a more fine-grained cache.
15541 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
15542 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
15544 2013-03-22 Richard Biener <rguenther@suse.de>
15546 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
15547 (struct mem_ref): Make accesses_in_loop a vec of a vec of
15548 aggregate mem_ref_loc.
15549 (free_mem_ref_locs): Inline into ...
15550 (memref_free): ... this and adjust.
15551 (mem_ref_alloc): Adjust.
15552 (mem_ref_locs_alloc): Remove.
15553 (record_mem_ref_loc): Adjust.
15554 (get_all_locs_in_loop): Rewrite into ...
15555 (for_all_locs_in_loop): ... this iterator.
15556 (rewrite_mem_ref_loc): New functor.
15557 (rewrite_mem_refs): Use for_all_locs_in_loop.
15558 (sm_set_flag_if_changed): New functor.
15559 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
15560 (ref_always_accessed): New functor.
15561 (ref_always_accessed_p): Use for_all_locs_in_loop.
15563 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
15565 * tree-pass.h (PROP_gimple_lvec): New.
15566 * passes.c (dump_properties): Handle PROP_gimple_lvec.
15567 (init_optimization_passes): Move pass_lower_vector.
15568 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
15570 (pass_lower_vector): Provide PROP_gimple_lvec.
15571 (pass_lower_vector_ssa): Likewise.
15572 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
15574 2013-03-21 Mark Wielaard <mjw@redhat.com>
15576 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
15578 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
15580 * config/i386/i386.md (*movdi_internal): Disparage slightly
15581 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
15582 conversion alternatives.
15584 2013-03-21 Jakub Jelinek <jakub@redhat.com>
15586 PR middle-end/48087
15587 * diagnostic.def (DK_WERROR): New kind.
15588 * diagnostic.h (werrorcount): Define.
15589 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
15590 promoted to DK_ERROR, increment DK_WERROR counter instead of
15592 * toplev.c (toplev_main): Call print_ignored_options even if
15593 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
15594 even if just werrorcount is non-zero.
15597 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
15599 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
15600 (string_cst_pool_decl): New function.
15601 (optimize_one_addr_into_implicit_ptr): New function.
15602 (resolve_addr_in_expr): Optimize DWARF location expression
15603 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
15604 which doesn't live in memory, but has DW_AT_location or
15605 DW_AT_const_value, or refers to a string literal, into
15606 DW_OP_GNU_implicit_pointer.
15607 (optimize_location_into_implicit_ptr): New function.
15608 (resolve_addr): If removing DW_AT_location of a variable because
15609 it was DW_OP_addr of address of the variable, but the variable doesn't
15610 live in memory, try to emit const value attribute for the initializer.
15612 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
15614 * tree.h (VECTOR_TYPE_P): New macro.
15615 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
15616 TYPE_MODE): Use it.
15617 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
15618 VEC_COND_EXPR cannot be lvalues.
15619 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
15621 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
15623 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
15624 Restrict the transformation to equal modes.
15626 2013-03-21 Richard Biener <rguenther@suse.de>
15628 PR tree-optimization/39326
15629 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
15630 (MEM_ANALYZABLE): Adjust.
15631 (record_mem_ref_loc): Move bitmap ops ...
15632 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
15633 unanalyzable refs, do not record locations for it.
15634 (analyze_memory_references): Allocate ref zero as shared
15636 (refs_independent_p): Do not test for unanalyzed mems here.
15637 (ref_indep_loop_p_1): Special-case disambiguation against
15638 the unanalyzed ref.
15639 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
15641 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
15643 * config/arm/arm-protos.h (tune_params): Add
15644 prefer_neon_for_64bits field.
15645 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
15646 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
15647 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
15648 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
15649 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
15650 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
15651 (arm_option_override): Handle -mneon-for-64bits new option.
15652 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
15653 (prefer_neon_for_64bits): Declare new variable.
15654 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
15655 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
15656 (arch_enabled): Handle new arch types. Remove support for onlya8
15658 (one_cmpldi2): Use new arch names.
15659 (zero_extend<mode>di2, extend<mode>di2): Ditto.
15660 * config/arm/arm.opt (mneon-for-64bits): Add option.
15661 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
15662 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
15663 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
15665 * doc/invoke.texi (-mneon-for-64bits): Document.
15667 2013-03-21 Richard Biener <rguenther@suse.de>
15669 PR tree-optimization/39326
15670 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
15671 (sort_bbs_in_loop_postorder_cmp): New function.
15672 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
15674 2013-03-21 Richard Biener <rguenther@suse.de>
15676 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
15677 (vect_insert_into_interleaving_chain): Likewise.
15678 (vect_drs_dependent_in_basic_block): Inline ...
15679 (vect_slp_analyze_data_ref_dependence): ... here. New function,
15681 (vect_analyze_data_ref_dependence): ... here. Simplify.
15682 (vect_check_interleaving): Simplify.
15683 (vect_analyze_data_ref_dependences): Likewise. Split out ...
15684 (vect_slp_analyze_data_ref_dependences): ... this new function.
15685 (dr_group_sort_cmp): New function.
15686 (vect_analyze_data_ref_accesses): Compute data-reference groups
15687 here instead of in vect_analyze_data_ref_dependence. Use
15688 a more efficient algorithm.
15689 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
15690 vect_slp_analyze_data_ref_dependences. Call
15691 vect_analyze_data_ref_accesses earlier.
15692 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
15693 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
15694 (vect_slp_analyze_data_ref_dependences): New prototype.
15696 2013-03-21 Richard Biener <rguenther@suse.de>
15698 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
15699 ref is stored in the loop.
15700 (find_refs_for_sm): Walk only over all stores.
15701 (store_motion_loop): Allocate from lim_bitmap_obstack.
15702 (store_motion): Likewise.
15704 2013-03-21 Richard Biener <rguenther@suse.de>
15706 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
15707 Update virtual SSA form.
15709 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15711 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
15712 * configure: Regenerate.
15713 * config.in: Regenerate.
15714 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
15715 if !HAVE_LD_EH_FRAME_CIEV3.
15717 2013-03-21 Richard Biener <rguenther@suse.de>
15719 * tree-cfg.c (verify_expr_no_block): New function.
15720 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
15721 nor DECL_VALUE_EXPR have locations with associated blocks.
15722 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
15723 (clear_unused_block_pointer): Remove code dealing with
15724 blocks in DECL_DEBUG_EXPR locations.
15726 2013-03-21 Richard Biener <rguenther@suse.de>
15728 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
15729 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
15730 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
15731 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
15732 instead of DECL_DEBUG_EXPR_IS_FROM.
15733 * gimplify.c (gimplify_modify_expr): Likewise.
15734 * tree-cfg.c (verify_expr_location_1): Likewise.
15735 * tree-complex.c (create_one_component_var): Likewise.
15736 * tree-sra.c (create_access_replacement): Likewise.
15737 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
15738 (clear_unused_block_pointer): Likewise.
15739 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
15740 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
15741 * var-tracking.c (var_debug_decl): Likewise.
15742 (track_expr_p): Likewise.
15743 * tree-inline.c (add_local_variables): Likewise. Set
15744 DECL_HAS_DEBUG_EXPR_P after copying it.
15745 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
15746 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
15748 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
15751 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
15752 * configure: Regenerate.
15753 * config.in: Regenerate.
15754 * config/i386/i386.md (*movdf_internal): Use
15755 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
15756 movd instead of movq mnemonic for interunit moves.
15757 (*movdi_internal): Ditto.
15759 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15761 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
15762 (abd<mode>_3): New pattern.
15763 (aba<mode>_3): New pattern.
15764 (fabd<mode>_3): New pattern.
15766 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15768 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
15769 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
15770 occurrence of REGISTER_PREFIX as its empty string.
15772 2013-03-20 Jeff Law <law@redhat.com>
15774 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
15775 addititional equivalences for equality comparisons between an SSA_NAME
15776 and a constant where the SSA_NAME was set from a widening conversion.
15778 2013-03-20 Walter Lee <walt@tilera.com>
15780 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
15782 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15784 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
15785 depending on type attribute.
15786 (*movti_internal): Ditto.
15787 (*movtf_internal): Ditto.
15788 (*movxf_internal): Ditto.
15789 (*movdf_internal): Ditto.
15790 (*movsf_internal): Ditto.
15792 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15794 * config/i386/i386.md (*movti_internal): Set prefix attribute to
15795 maybe_vex for sselog1 and ssemov types.
15796 (*movdi_internal): Reorder operand constraints.
15797 (*movsi_internal): Ditto. Set prefix attribute to
15798 maybe_vex for sselog1 and ssemov types.
15799 (*movtf_internal): Set prefix attribute to maybe_vex
15800 for sselog1 and ssemov types.
15801 (*movdf_internal): Ditto. Set prefix_data16 attribute for
15802 DImode ssemov types. Reorder operand constraints.
15803 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
15804 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
15805 attribute for SImode ssemov types. Reorder operand constraints.
15807 2013-03-20 Martin Jambor <mjambor@suse.cz>
15809 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
15810 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
15812 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
15814 * config/rs6000/predicates.md (indexed_address, update_address_mem
15815 update_indexed_address_mem): New predicates.
15816 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
15817 attribute for load/store instructions.
15818 * config/rs6000/dfp.md (movsd_store): Likewise.
15819 (movsd_load): Likewise.
15820 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
15821 (unnamed HI->DI extend define_insn): Likewise.
15822 (unnamed SI->DI extend define_insn): Likewise.
15823 (unnamed QI->SI extend define_insn): Likewise.
15824 (unnamed QI->HI extend define_insn): Likewise.
15825 (unnamed HI->SI extend define_insn): Likewise.
15826 (unnamed HI->SI extend define_insn): Likewise.
15827 (extendsfdf2_fpr): Likewise.
15828 (movsi_internal1): Likewise.
15829 (movsi_internal1_single): Likewise.
15830 (movhi_internal): Likewise.
15831 (movqi_internal): Likewise.
15832 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
15833 attribute for load/store instructions.
15834 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
15836 (mov<mode>_softfloat): Likewise.
15837 (mov<mode>_hardfloat32): Likewise.
15838 (mov<mode>_hardfloat64): Likewise.
15839 (mov<mode>_softfloat64): Likewise.
15840 (movdi_internal32): Likewise.
15841 (movdi_internal64): Likewise.
15842 (probe_stack_<mode>): Likewise.
15844 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
15846 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
15847 floating point, and decimal floating point to reload iterator.
15849 * config/rs6000/constraints.md (wl constraint): New constraints to
15850 return FLOAT_REGS if certain options are used to reduce the number
15851 of separate patterns that exist in the file.
15852 (wx constraint): Likewise.
15853 (wz constraint): Likewise.
15855 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15856 -mdebug=reg, print wg, wl, wx, and wz constraints.
15857 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15858 Initialize the reload functions for 64-bit binary/decimal floating
15860 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15861 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15862 create the buffer on the stack to overcome not having a 32-bit
15864 (rs6000_emit_move): Likewise.
15865 (rs6000_secondary_memory_needed_rtx): Likewise.
15866 (rs6000_alloc_sdmode_stack_slot): Likewise.
15867 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15868 via xxlxor, just like DFmode 0.0.
15870 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
15871 define as 1 if we are running on a power7 or newer.
15872 (enum r6000_reg_class_enum): Add new constraints.
15874 * config/rs6000/dfp.md (movsd): Delete, combine with binary
15875 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15876 with other moves by using conditional constraits (wg). Use LFIWZX
15877 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
15878 (movsd splitter): Likewise.
15879 (movsd_hardfloat): Likewise.
15880 (movsd_softfloat): Likewise.
15882 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
15883 binary and decimal floating point moves.
15884 (fmove_ok): New attributes to combine binary and decimal floating
15885 point moves, and to combine power6x (mfpgpr) moves along normal
15887 (real_value_to_target): Likewise.
15888 (f32_lr): Likewise.
15889 (f32_lm): Likewise.
15890 (f32_li): Likewise.
15891 (f32_sr): Likewise.
15892 (f32_sm): Likewise.
15893 (f32_si): Likewise.
15894 (movsf): Combine binary and decimal floating point moves. Combine
15895 power6x (mfpgpr) moves with other moves by using conditional
15896 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
15897 (mov<mode> for SFmode/SDmode); Likewise.
15898 (SFmode/SDmode splitters): Likewise.
15899 (movsf_hardfloat): Likewise.
15900 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
15901 (movsf_softfloat): Likewise.
15902 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
15904 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
15905 wx and wz constraints.
15907 * config/rs6000/constraints.md (wg constraint): New constraint to
15908 return FLOAT_REGS if -mmfpgpr (power6x) was used.
15910 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
15913 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15914 -mdebug=reg, print wg, wl, wx, and wz constraints.
15915 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15916 Initialize the reload functions for 64-bit binary/decimal floating
15918 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15919 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15920 create the buffer on the stack to overcome not having a 32-bit
15922 (rs6000_emit_move): Likewise.
15923 (rs6000_secondary_memory_needed_rtx): Likewise.
15924 (rs6000_alloc_sdmode_stack_slot): Likewise.
15925 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15926 via xxlxor, just like DFmode 0.0.
15928 * config/rs6000/dfp.md (movdd): Delete, combine with binary
15929 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15930 with other moves by using conditional constraits (wg). Use LFIWZX
15931 and STFIWX for loading SDmode on power7.
15932 (movdd splitters): Likewise.
15933 (movdd_hardfloat32): Likewise.
15934 (movdd_softfloat32): Likewise.
15935 (movdd_hardfloat64_mfpgpr): Likewise.
15936 (movdd_hardfloat64): Likewise.
15937 (movdd_softfloat64): Likewise.
15939 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
15940 64-bit binary and decimal floating point moves.
15941 (FMOVE64X): Likewise.
15942 (movdf): Combine 64-bit binary and decimal floating point moves.
15943 Combine power6x (mfpgpr) moves with other moves by using
15944 conditional constraits (wg).
15945 (mov<mode> for DFmode/DDmode): Likewise.
15946 (DFmode/DDmode splitters): Likewise.
15947 (movdf_hardfloat32): Likewise.
15948 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
15949 (movdf_softfloat32): Likewise.
15950 (movdf_hardfloat64_mfpgpr): Likewise.
15951 (movdf_hardfloat64): Likewise.
15952 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
15953 (movdf_softfloat64): Likewise.
15954 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
15955 (reload_<mode>_load): Move to later in the file so they aren't in
15956 the middle of the floating point move insns.
15957 (reload_<mode>_store): Likewise.
15959 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
15962 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
15963 constraint if -mdebug=reg.
15964 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
15965 Enable using dd reload support if needed.
15967 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
15968 binary and decimal floating point moves in rs6000.md.
15969 (movtd_internal): Likewise.
15971 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
15972 decimal floating point moves.
15974 (movtf_internal): Likewise.
15975 (mov<mode>_internal, TDmode/TFmode): Likewise.
15976 (movtf_softfloat): Likewise.
15977 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
15979 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
15980 movdi_internal64, using wg constraint for move direct operations.
15981 (movdi_internal64): Likewise.
15983 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
15984 MODES_TIEABLE_P for selected modes. Print the numerical value of
15985 the various virtual registers. Use GPR/FPR first/last values,
15986 instead of hard coding the register numbers. Print which modes
15987 have reload functions registered.
15988 (rs6000_option_override_internal): If -mdebug=reg, trace the options
15989 settings before/after setting cpu, target and subtarget settings.
15990 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
15991 and for secondary reload failures in rs6000_secondary_reload_inner.
15992 (rs6000_secondary_reload_fail): Likewise.
15993 (rs6000_secondary_reload_inner): Likewise.
15995 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
15996 macros for first/last GPR and FPR registers.
15997 (LAST_GPR_REGNO): Likewise.
15998 (FIRST_FPR_REGNO): Likewise.
15999 (LAST_FPR_REGNO): Likewise.
16001 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
16002 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
16003 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
16004 (vcond<mode><mode>): Likewise.
16005 (vcondu<mode><mode>): Likewise.
16006 (vector_gtu<mode>): Likewise.
16007 (vector_gte<mode>): Likewise.
16008 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
16009 to prevent the compiler from converting DImode operations to TImode.
16010 (ior<mode>3): Likewise.
16011 (and<mode>3): Likewise.
16012 (one_cmpl<mode>2): Likewise.
16013 (nor<mode>3): Likewise.
16014 (andc<mode>3): Likewise.
16016 * config/rs6000/constraints.md (wt constraint): New constraint
16017 that returns VSX_REGS if TImode is allowed in VSX registers.
16019 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
16020 constant under VSX.
16022 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
16023 similar to TImode, but it is restricted to being in the GPRs.
16025 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
16026 TImode to occupy a single VSX register.
16028 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
16029 -mvsx-timode for power7/power8.
16030 (power7 cpu): Likewise.
16031 (power8 cpu): Likewise.
16033 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
16034 sure that TFmode/TDmode take up two registers if they are ever
16035 allowed in the upper VSX registers.
16036 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
16038 (rs6000_init_hard_regno_mode_ok): Likewise.
16039 (rs6000_debug_reg_global): Add debugging for PTImode and wt
16040 constraint. Print if LRA is turned on.
16041 (rs6000_option_override_internal): Give an error if -mvsx-timode
16042 and VSX is not enabled.
16043 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
16044 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
16045 to reg+offset addressing. Use PTImode when checking offset
16046 addresses for validity.
16047 (reg_offset_addressing_ok_p): Likewise.
16048 (rs6000_legitimate_offset_address_p): Likewise.
16049 (rs6000_legitimize_address): Likewise.
16050 (rs6000_legitimize_reload_address): Likewise.
16051 (rs6000_legitimate_address_p): Likewise.
16052 (rs6000_eliminate_indexed_memrefs): Likewise.
16053 (rs6000_emit_move): Likewise.
16054 (rs6000_secondary_reload): Likewise.
16055 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
16056 reloads to fpr registers to continue to use reg+offset addressing,
16057 but 64-bit reloads to altivec registers need reg+reg addressing.
16058 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
16059 it. Treat LO_SUM like a PLUS operation.
16060 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
16061 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
16062 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
16063 registers to share a register with a smaller sized type, since VSX
16064 puts scalars in the upper 64-bits.
16065 (print_operand): Add support for PTImode.
16066 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
16067 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
16068 registers, but don't have arithmetic support.
16069 (rs6000_memory_move_cost): Add test for VSX.
16070 (rs6000_opt_masks): Add -mvsx-timode.
16072 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
16075 (VSr): Use wt constraint for TImode.
16076 (VSv): Drop TImode support.
16077 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
16078 (vsx_movti_64bit): Likewise.
16079 (vsx_movti_32bit): Likewise.
16080 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
16081 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
16082 one '?' on the appropriate output constraint. Do not allow TImode
16083 logical operations on 32-bit systems.
16084 (vsx_ior<mode>3): Likewise.
16085 (vsx_xor<mode>3): Likewise.
16086 (vsx_one_cmpl<mode>2): Likewise.
16087 (vsx_nor<mode>3): Likewise.
16088 (vsx_andc<mode>3): Likewise.
16089 (vsx_concat_<mode>): Likewise.
16090 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
16092 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
16093 OPTION_MASK_VSX_TIMODE.
16094 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
16095 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
16097 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
16098 (TI2 iterator): New iterator for TImode, PTImode.
16099 (wd mode attribute): Add values for vector types.
16100 (movti_string): Replace TI move operations with operations for TImode
16101 and PTImode. Add support for TImode being allowed in VSX registers.
16102 (mov<mode>_string, TImode/PTImode): Likewise.
16103 (movti_ppc64): Likewise.
16104 (mov<mode>_ppc64, TImode/PTImode): Likewise.
16105 (TI mode splitters): Likewise.
16107 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
16110 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
16112 PR tree-optimization/56355
16113 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
16114 Also handle integers with undefined overflow.
16116 2013-03-20 Catherine Moore <clm@codesourcery.com>
16117 Maciej W. Rozycki <macro@codesourcery.com>
16118 Tom de Vries <tom@codesourcery.com>
16119 Nathan Sidwell <nathan@codesourcery.com>
16120 Iain Sandoe <iain@codesourcery.com>
16121 Nathan Froyd <froydnj@codesourcery.com>
16122 Chao-ying Fu <fu@mips.com>
16124 * doc/extend.texi: (micromips, nomicromips, nocompression):
16125 Document new function attributes.
16126 * doc/invoke.texi (minterlink-compressed, mmicromips,
16127 m14k, m14ke, m14kec): Document new options.
16128 (minterlink-mips16): Update documentation.
16129 * doc/md.texi (ZC, ZD): Document new constraints.
16130 * configure.ac (gcc_cv_as_micromips): Check if linker
16131 supports the .set micromips directive.
16132 * configure: Regenerate.
16133 * config.in: Regenerate.
16134 * config/mips/mips-tables.opt: Regenerate.
16135 * config/mips/micromips.md: New file.
16136 * constraints.md (ZC, ZD): New constraints.
16137 * config/mips/predicates.md (movep_src_register): New predicate.
16138 (movep_src_operand): New predicate.
16139 (non_volatile_mem_operand): New predicate.
16140 * config/mips/mips.md (multimem): New type.
16141 (length): Differentiate between 17-bit and 18-bit branch offsets.
16142 (MOVEP1, MOVEP2): New mode iterator.
16143 (mov_<load>l): Use ZC constraint.
16144 (mov_<load>r): Likewise.
16145 (mov_<store>l): Likewise.
16146 (mov_<store>r): Likewise.
16147 (*branch_equality<mode>_inverted): Add microMIPS support.
16148 (*branch_equality<mode>): Likewise.
16149 (*jump_absolute): Likewise.
16150 (indirect_jump_<mode>): Likewise.
16151 (tablejump_<mode>): Likewise.
16152 (<optab>_internal): Likewise.
16153 (sibcall_internal): Likewise.
16154 (sibcall_value_internal): Likewise.
16155 (prefetch): Use constraint ZD.
16156 * config/mips/mips.opt (minterlink-compressed): New option.
16157 (minterlink-mips16): Now an alias for minterlink-compressed.
16158 (mmicromips): New option.
16159 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16160 (compare_and_swap_12): Likewise.
16161 (sync_add<mode>): Likewise.
16162 (sync_<optab>_12): Likewise.
16163 (sync_old_<optab>_12): Likewise.
16164 (sync_new_<optab>_12): Likewise.
16165 (sync_nand_12): Likewise.
16166 (sync_old_nand_12): Likewise.
16167 (sync_new_nand_12): Likewise.
16168 (sync_sub<mode>): Likewise.
16169 (sync_old_add<mode>): Likewise.
16170 (sync_old_sub<mode>): Likewise.
16171 (sync_new_add<mode>): Likewise.
16172 (sync_new_sub<mode>): Likewise.
16173 (sync_<optab><mode>): Likewise.
16174 (sync_old_<optab><mode>): Likewise.
16175 (sync_new_<optab><mode>): Likewise.
16176 (sync_nand<mode>): Likewise.
16177 (sync_old_nand<mode>): Likewise.
16178 (sync_new_nand<mode>): Likewise.
16179 (sync_lock_test_and_set<mode>): Likewise.
16180 (test_and_set_12): Likewise.
16181 (atomic_compare_and_swap<mode>): Likewise.
16182 (atomic_exchange<mode>_llsc): Likewise.
16183 (atomic_fetch_add<mode>_llsc): Likewise.
16184 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16185 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16186 (umips_save_restore_pattern_p): Likewise.
16187 (umips_load_store_pair_p): Likewise.
16188 (umips_output_load_store_pair): Likewise.
16189 (umips_movep_target_p): Likewise.
16190 (umips_12bit_offset_address_p): Likewise.
16191 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16192 (mips_base_mips16): Rename this...
16193 (mips_base_compression_flags): ...to this. Update all uses.
16194 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16195 (mips_mips16_decl_p): Delete.
16196 (mips_nomips16_decl_p): Delete.
16197 (mips_get_compress_on_flags): New function.
16198 (mips_get_compress_off_flags): New function.
16199 (mips_get_compress_mode): New function.
16200 (mips_get_compress_on_name): New function.
16201 (mips_get_compress_off_name): New function.
16202 (mips_insert_attributes): Support multiple compression types.
16203 (mips_merge_decl_attributes): Likewise.
16204 (umips_12bit_offset_address_p): New function.
16205 (mips_start_function_definition): Emit .set micromips directive.
16206 (mips_call_may_need_jalx_p): New function.
16207 (mips_function_ok_for_sibcall): Add microMIPS support.
16208 (mips_print_operand_punctuation): Support short delay slots and
16210 (umips_swm_mask, umips_swm_encoding): New.
16211 (umips_build_save_restore): New function.
16212 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16213 (was_mips16_p): Remove.
16214 (old_compression_mode): New.
16215 (mips_set_compression_mode): New function.
16216 (mips_set_current_function): Add microMIPS support.
16217 (mips_option_override): Likewise.
16218 (umips_save_restore_pattern_p): New function.
16219 (umips_output_save_restore): New function.
16220 (umips_load_store_pair_p_1): New function.
16221 (umips_load_store_pair_p): New function.
16222 (umips_output_load_store_pair_1): New function.
16223 (umips_output_load_store_pair): New function.
16224 (umips_movep_target_p) New function.
16225 (mips_prepare_pch_save): Add microMIPS support.
16226 * config/mips/mips.h (TARGET_COMPRESSION): New.
16227 (TARGET_CPU_CPP_BUILTINS): Update macro
16228 to use new compression flags and to support microMIPS.
16229 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16230 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16231 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16232 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16233 (ASM_SPEC): Support mmicromips and mno-micromips.
16234 (M16STORE_REG_P): New macro.
16235 (MIPS_CALL): Support TARGET_MICROMIPS.
16236 (MICROMIPS_J): New macro.
16237 (mips_base_mips16): Rename this...
16238 (mips_base_compression_flags): ...to this.
16239 (UMIPS_12BIT_OFFSET_P): New macro.
16240 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16241 (MULTILIB_DIRNAMES): Likewise.
16242 2013-03-20 Richard Biener <rguenther@suse.de>
16244 PR tree-optimization/56661
16245 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
16246 the result does not have to be distinct.
16248 2013-03-20 Richard Biener <rguenther@suse.de>
16250 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
16253 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16254 Steven Bosscher <steven@gcc.gnu.org>
16256 PR rtl-optimization/56605
16257 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
16259 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
16262 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
16263 that require movd instead of movq.
16265 2013-03-20 Richard Biener <rguenther@suse.de>
16267 * tree-ssa-structalias.c (struct variable_info): Add pointer
16268 to the first field of an aggregate with sub-vars. Make
16269 this and the pointer to the next subfield its ID.
16270 (vi_next): New function.
16271 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
16272 storedanything_id, integer_id): Increment by one.
16273 (new_var_info, get_call_vi, lookup_call_clobber_vi,
16274 get_call_clobber_vi): Adjust.
16275 (solution_set_expand): Simplify and speedup.
16276 (solution_set_add): Inline into ...
16277 (set_union_with_increment): ... this. Adjust accordingly.
16278 (do_sd_constraint): Likewise.
16279 (do_ds_constraint): Likewise.
16280 (do_complex_constraint): Simplify.
16281 (build_pred_graph): Adjust.
16282 (solve_graph): Likewise. Simplify and speedup.
16283 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
16284 get_constraint_for_component_ref, get_constraint_for_1,
16285 first_vi_for_offset, first_or_preceding_vi_for_offset,
16286 create_function_info_for, create_variable_info_for_1,
16287 create_variable_info_for, intra_create_variable_infos): Adjust.
16288 (init_base_vars): Push NULL for ID zero.
16289 (compute_points_to_sets): Adjust.
16291 2013-03-20 Richard Biener <rguenther@suse.de>
16293 * cfgloop.c (verify_loop_structure): Streamline and avoid
16294 ICEing on corrupt loop tree.
16295 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
16298 2013-03-20 Richard Biener <rguenther@suse.de>
16300 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
16301 check whether an SSA update is needed.
16303 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
16305 * config/mips/constraints.md (T): Rename to...
16309 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
16310 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
16312 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16314 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
16315 (*subsi3_carryin_uxtw): Likewise.
16317 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16319 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
16320 (*rorsi3_insn_uxtw): Likewise.
16322 2013-03-19 Ian Bolton <ian.bolton@arm.com>
16324 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
16325 (*extrsi5_insn_uxtw): Likewise.
16327 2013-03-19 Richard Biener <rguenther@suse.de>
16329 PR tree-optimization/56273
16330 * passes.c (init_optimization_passes): Move second VRP after DOM.
16332 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
16334 * config/i386/i386.md (*movti_internal): Merge from
16335 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
16336 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
16337 nox64 isa attributes.
16339 2013-03-18 Richard Biener <rguenther@suse.de>
16341 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
16343 (merge_node_constraints): Likewise.
16344 (build_succ_graph): Likewise.
16345 (valid_graph_edge): Inline into single caller.
16346 (unify_nodes): Likewise. Use bitmap_set_bit return value
16348 (scc_visit): Fix formatting and variable use.
16349 (do_sd_constraint): Use gcc_checking_assert.
16350 (do_ds_constraint): Likewise.
16351 (do_complex_constraint): Likewise.
16352 (condense_visit): Likewise. Cleanup.
16353 (dump_pred_graph): New function.
16354 (perform_var_substitution): Dump the pred-graph before
16355 variable substitution.
16356 (find_equivalent_node): Use gcc_checking_assert.
16357 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
16359 2013-03-18 Richard Biener <rguenther@suse.de>
16361 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
16362 Remove cond_expr_stmt_list argument and do not gimplify the
16364 (vect_loop_versioning): Adjust.
16365 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
16366 Cleanup to use less temporaries.
16367 (vect_create_data_ref_ptr): Cleanup.
16369 2013-03-18 Jakub Jelinek <jakub@redhat.com>
16371 PR tree-optimization/56635
16372 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
16373 require types_compatible_p types.
16375 2013-03-18 Nick Clifton <nickc@redhat.com>
16377 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
16378 spurious backslash.
16380 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
16381 Add missing line to comment describing function.
16383 2013-03-18 Richard Biener <rguenther@suse.de>
16385 PR tree-optimization/56210
16386 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
16387 Handle string / character search functions.
16388 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
16390 2013-03-18 Richard Biener <rguenther@suse.de>
16392 PR middle-end/56483
16393 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
16394 and implement properly.
16395 * gimple.h (gimple_cond_single_var_p): Remove.
16397 2013-03-18 Richard Biener <rguenther@suse.de>
16399 * tree-data-ref.h (find_data_references_in_loop): Declare.
16400 * tree-data-ref.c (get_references_in_stmt): Use a stack
16401 vector pre-allocated in the callers.
16402 (find_data_references_in_stmt): Adjust.
16403 (graphite_find_data_references_in_stmt): Likewise.
16404 (create_rdg_vertices): Likewise.
16405 (find_data_references_in_loop): Export.
16406 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
16407 Compute dependences here...
16408 (vect_analyze_data_refs): ...not here. When we encounter
16409 a non-vectorizable data reference in basic-block vectorization
16410 truncate the data reference vector. Do not bother to
16411 fixup data-dependence information for gather loads.
16412 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
16413 of data references, as reported.
16415 2013-03-18 Richard Biener <rguenther@suse.de>
16417 PR tree-optimization/3713
16418 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
16419 has_constants and expr.
16420 (stmt_has_constants): Properly valueize SSA names when deciding
16421 whether the stmt has constants.
16423 2013-03-18 Richard Biener <rguenther@suse.de>
16425 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
16426 whole function when there is nothing to do.
16427 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
16428 * tree-vectorizer.c (vectorize_loops): Update virtual and
16429 loop-closed SSA once.
16430 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
16432 2013-03-18 Richard Biener <rguenther@suse.de>
16434 PR middle-end/56113
16435 * domwalk.c (bb_postorder): New global static.
16436 (cmp_bb_postorder): New function.
16437 (walk_dominator_tree): Replace scheme imposing an order for
16438 visiting dominator sons by one sorting them at the time they
16439 are pushed on the stack.
16441 2013-03-18 Richard Biener <rguenther@suse.de>
16443 PR tree-optimization/39326
16444 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
16445 (struct mem_ref): Replace mem member with ao_ref typed member.
16446 (MEM_ANALYZABLE): Adjust.
16447 (memref_eq): Likewise.
16448 (mem_ref_alloc): Likewise.
16449 (gather_mem_refs_stmt): Likewise.
16450 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
16451 (execute_sm_if_changed_flag_set): Adjust.
16452 (execute_sm): Likewise.
16453 (ref_always_accessed_p): Likewise.
16454 (refs_independent_p): Likewise.
16455 (can_sm_ref_p): Likewise.
16457 2013-03-18 Jakub Jelinek <jakub@redhat.com>
16460 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
16461 return 1 even for !unsignedp.
16463 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16465 * config/i386/i386.md (isa): Add x64 and nox64.
16466 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
16467 (*pushtf): Enable *roF alternative for x64 isa only.
16468 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
16469 mode attribute of integer alternatives to DImode for TARGET_64BIT.
16470 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
16471 (*movtf_internal): Merge from *movtf_internal_rex64 and
16472 *movtf_internal_sse. Use x64 and nox64 isa attributes.
16473 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
16474 nox64 isa attributes.
16475 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
16476 nox64 isa attributes.
16477 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
16479 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
16481 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16483 2013-03-17 Jakub Jelinek <jakub@redhat.com>
16486 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
16487 class names. Remove trailing comma after "ALL_REGS".
16489 2013-03-16 Jan Hubicka <jh@suse.cz>
16491 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
16492 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
16493 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
16494 of cgraph_get_create_node.
16495 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
16497 2013-03-16 Jason Merrill <jason@redhat.com>
16500 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
16501 with DW_AT_default_value.
16503 2013-03-16 Jakub Jelinek <jakub@redhat.com>
16505 * BASE-VER: Set to 4.9.0.
16507 2013-03-14 Andi Kleen <ak@linux.intel.com>
16510 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
16511 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
16512 Document _x* TSX intrinsics.
16514 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
16515 David Holsgrove <david.holsgrove@xilinx.com>
16517 * configure.ac: Add MicroBlaze TLS support detection.
16518 * configure: Regenerate.
16519 * config/microblaze/microblaze-protos.h
16520 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
16521 symbol_mentioned_p, label_mentioned_p): Add prototypes.
16522 * config/microblaze/microblaze.c (microblaze_address_type): Add
16523 ADDRESS_TLS and tls_reloc address types.
16524 (microblaze_address_info): Add tls_reloc.
16525 (TARGET_HAVE_TLS): Define.
16526 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
16527 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
16528 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
16529 load_tls_operand, microblaze_call_tls_get_addr,
16530 microblaze_legitimize_tls_address): New functions.
16531 (microblaze_classify_unspec): Handle UNSPEC_TLS.
16532 (get_base_reg): Use microblaze_tls_symbol_p.
16533 (microblaze_classify_address): Handle TLS.
16534 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
16535 label_mentioned_p and microblaze_tls_referenced_p.
16536 (microblaze_legitimize_address): Handle TLS.
16537 (microblaze_address_insns): Handle ADDRESS_TLS.
16538 (pic_address_needs_scratch): Handle TLS.
16539 (print_operand_address): Handle TLS.
16540 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
16541 (microblaze_expand_move): Handle TLS.
16542 (microblaze_legitimate_constant_p): Check
16543 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
16544 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
16545 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
16546 (PIC_OFFSET_TABLE_REGNUM): Set.
16547 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
16548 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
16549 (addsi3, movsi_internal2, movdf_internal): Update constraints
16550 * config/microblaze/predicates.md (arith_plus_operand): Define
16551 (move_operand): Redefine as move_src_operand,
16552 check microblaze_tls_referenced_p.
16554 2013-03-14 Ian Bolton <ian.bolton@arm.com>
16556 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
16557 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
16559 2013-03-14 Ian Bolton <ian.bolton@arm.com>
16561 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
16564 2013-03-14 Jakub Jelinek <jakub@redhat.com>
16566 PR tree-optimization/53265
16567 * common.opt (Waggressive-loop-optimizations): New option.
16568 * tree-ssa-loop-niter.c: Include tree-pass.h.
16569 (do_warn_aggressive_loop_optimizations): New function.
16570 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
16571 if number_of_latch_executions returned constant.
16572 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
16573 early. If number_of_latch_executions returned constant, set
16574 nb_iterations_upper_bound back to it.
16575 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
16577 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
16578 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
16580 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
16581 (MULTILIB_OSDIRNAMES): Set.
16582 * genmultilib: If defaultosdirname doesn't start with :: , set
16583 defaultosdirname2 instead, clear it and emit two . multilib_raw
16584 entries instead of just one.
16586 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
16588 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
16589 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
16590 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
16591 (SUBTARGET_OVERRIDE_OPTIONS): New.
16593 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
16596 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
16597 (musermode): Convert to Var(TARGET_USERMODE).
16598 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
16599 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
16600 * config/sh/sh.c (sh_option_override): Use
16601 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
16602 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
16604 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
16606 (udivsi3_i4_single, divsi3_i4_single): Use
16607 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
16609 2013-03-13 Dave Korn <dave.korn.cygwin@....>
16611 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
16614 2013-03-13 Richard Biener <rguenther@suse.de>
16616 PR tree-optimization/56608
16617 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
16618 calls when vectorizing basic-blocks.
16620 2013-03-13 Jakub Jelinek <jakub@redhat.com>
16623 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
16626 2013-03-12 Jakub Jelinek <jakub@redhat.com>
16628 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
16630 2013-03-11 Jan Hubicka <jh@suse.cz>
16633 * lto-streamer-out.c (output_symbol_p): Skip references from
16634 constructors of external variables.
16636 2013-03-11 Jan Hubicka <jh@suse.cz>
16638 PR middle-end/56571
16639 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
16641 * emit-rtl.c (verify_rtx_sharing): Likewise.
16642 (copy_insn_1): Likewise.
16643 * rtl.c (copy_rtx): Likewise.
16645 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
16648 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
16649 output_operand_lossage message.
16651 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
16654 * arm.c (shift_op): Validate RTL pattern on the fly.
16655 (arm_print_operand, case 'S'): Don't use shift_operator to validate
16658 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16661 * config/pa/pa.md (call_value): Check for calls to powf and direct to
16662 new call patterns that clobber %fr12.
16663 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
16664 split and postreload patterns.
16665 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
16666 registers %fr12 and %fr12R as call used.
16668 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
16670 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
16671 (canon_address, record_store, replace_read, check_mem_read_rtx,
16672 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
16673 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
16674 rest_of_handle_dse): Likewise.
16676 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
16678 PR middle-end/56524
16679 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
16681 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
16682 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
16683 (save_optabs_if_changed): Replace with...
16684 (init_tree_optimization_optabs): ...this.
16685 * optabs.c (save_optabs_if_changed): Rename to...
16686 (init_tree_optimization_optabs): ...this. Take the optimization node
16687 as argument. Do nothing if the base optabs are already correct.
16688 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
16689 to recompute optabs.
16690 * function.h (function): Remove optabs field.
16691 * function.c (invoke_set_current_function_hook): Call
16692 init_tree_optimization_optabs. Use the result to initialize
16695 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
16697 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
16698 if GTMA_HAS_NO_INSTRUMENTATION.
16699 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
16700 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
16701 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
16702 * gimple-pretty-print.c (dump_gimple_transaction): Handle
16703 GTMA_HAS_NO_INSTRUMENTATION.
16705 2013-03-08 Jakub Jelinek <jakub@redhat.com>
16707 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
16710 2013-03-08 Marek Polacek <polacek@redhat.com>
16711 Jakub Jelinek <jakub@redhat.com>
16713 PR tree-optimization/56478
16714 * predict.c (is_comparison_with_loop_invariant_p): Change the
16715 type of loop_step to tree.
16716 (predict_loops): Adjust.
16717 (predict_iv_comparison): Perform the computations on double_ints.
16719 2013-03-08 Richard Biener <rguenther@suse.de>
16721 PR tree-optimization/56570
16722 * tree-cfg.c (verify_expr_location_1): Verify locations for
16724 * tree-sra.c (create_access_replacement): Strip locations
16725 from DECL_DEBUG_EXPRs.
16727 2013-03-08 Richard Biener <rguenther@suse.de>
16729 * tree-inline.c (expand_call_inline): Do not associate
16730 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
16731 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
16733 2013-03-08 Richard Biener <rguenther@suse.de>
16735 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
16736 or block changes with -Og. Fix for location / block encoding
16737 changes and PHI arguments with locations.
16739 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
16741 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
16743 (struct output_info): Likewise.
16744 (register_overhead): Remove bad gcc_assert.
16745 (bitmap_find_bit): If there is only a single bitmap element, do not
16746 count a miss as a search.
16747 (print_statistics): Update for counter type changes.
16748 (dump_bitmap_statistics): Likewise. Print headers such that they
16749 are properly lined up with the printed counters.
16751 2013-03-07 Jakub Jelinek <jakub@redhat.com>
16753 PR tree-optimization/56559
16754 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
16755 check that it has only a single use.
16757 2013-03-07 Richard Biener <rguenther@suse.de>
16759 * doc/invoke.texi (fwhole-program): Discourage use in combination
16762 2013-03-06 Jakub Jelinek <jakub@redhat.com>
16764 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
16766 PR tree-optimization/56539
16767 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
16768 instead of GSI_CONTINUE_LINKING as last argument to
16769 force_gimple_operand_gsi. Adjust function comment.
16771 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
16774 PR middle-end/56548
16775 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
16776 promoted mode, convert the result back to the original mode.
16778 2013-03-06 Richard Biener <rguenther@suse.de>
16780 PR middle-end/56294
16781 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
16782 (insert_updated_phi_nodes_compare_uids): New function.
16783 (update_ssa): Sort symbols_to_rename after UID before
16784 traversing it to insert PHI nodes.
16786 2013-03-06 Richard Biener <rguenther@suse.de>
16788 PR middle-end/50494
16789 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
16790 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
16793 2013-02-13 Richard Biener <rguenther@suse.de>
16796 * varasm.c (output_constant_def_1): Get the decl representing
16797 the constant as argument.
16798 (output_constant_def): Wrap output_constant_def_1.
16799 (make_decl_rtl): Use output_constant_def_1 with the decl
16800 representing the constant.
16801 (build_constant_desc): Optionally re-use a decl already
16802 representing the constant.
16803 (tree_output_constant_def): Adjust.
16805 2013-03-06 Joey Ye <joey.ye@arm.com>
16808 * gcc.c (convert_white_space): New function.
16809 (main): Handles white space in function name.
16811 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
16814 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
16815 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
16816 to SH_DIV_CALL_TABLE for TARGET_SH2.
16817 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
16819 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
16820 call-table options.
16822 2013-03-05 Sterling Augustine <saugustine@google.com>
16823 Cary Coutant <ccoutant@google.com>
16826 * dwarf2out.c (resolve_addr): Don't call
16827 remove_loc_list_addr_table_entries a second time for the same
16830 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16833 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
16834 (avoid_complex_debug_insns): New function.
16835 (expand_debug_locations): Call it.
16837 PR rtl-optimization/56484
16838 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
16839 lifetimes of hard registers on small register class machines.
16841 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16843 * config/microblaze/microblaze-protos.h: Rename
16844 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
16845 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
16847 (microblaze_fast_interrupt_function_p): New function.
16848 (microblaze_is_interrupt_handler): Rename to
16849 microblaze_is_interrupt_variant and add fast_interrupt check.
16850 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
16851 (save_restore_insns): Likewise.
16852 (compute_frame_size): Likewise.
16853 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
16854 (microblaze_globalize_label): Likewise.
16855 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
16856 * config/microblaze/microblaze.md: Use wrapper
16857 microblaze_is_interrupt_variant.
16859 2013-03-05 Kai Tietz <ktietz@redhat.com>
16861 * sdbout.c (sdbout_one_type): Switch to current function's section
16862 supporting cold/hot.
16864 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16866 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
16869 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16871 PR middle-end/56461
16872 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
16873 if VALGRIND_GET_VBITS is defined, temporarily make object
16874 memory all defined, and restore previous valgrind addressability
16875 and definability afterwards. Free this_object at the end.
16877 PR middle-end/56461
16878 * lra.c (lra): Call lra_clear_live_ranges if live_p,
16879 right before calling lra_create_live_ranges, also call it
16880 when clearing live_p. Only call lra_clear_live_ranges
16881 at the end if live_p.
16883 PR middle-end/56461
16884 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
16886 2013-03-05 Richard Biener <rguenther@suse.de>
16888 PR tree-optimization/56521
16889 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
16892 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
16895 * except.h (remove_unreachable_eh_regions): New prototype.
16896 * except.c (remove_eh_handler_splicer): New function, split out
16897 of remove_eh_handler.
16898 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
16899 warning about running it on many EH regions one at a time.
16900 (remove_unreachable_eh_regions_worker): New function, walk the
16901 EH tree in depth-first order and remove non-marked regions.
16902 (remove_unreachable_eh_regions): New function.
16903 * tree-eh.c (mark_reachable_handlers): New function, split out
16904 from remove_unreachable_handlers.
16905 (remove_unreachable_handlers): Use mark_reachable_handlers and
16906 remove_unreachable_eh_regions.
16907 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
16908 and remove_unreachable_eh_regions.
16910 2013-03-05 Richard Biener <rguenther@suse.de>
16912 PR middle-end/56525
16913 * loop-init.c (fix_loop_structure): Remove loops in two stages,
16914 not freeing them until the end.
16916 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16918 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
16920 2013-03-05 Richard Biener <rguenther@suse.de>
16922 PR tree-optimization/56270
16923 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
16924 of loads after scheduling an SLP instance.
16926 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16928 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
16930 (check_gcc_parallelize): Run guality.exp as a separate job from
16931 vect.exp with unsorted.exp and $(dg_target_exps) separately from
16932 struct-layout-1.exp with stackalign.exp.
16934 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
16936 PR middle-end/56461
16937 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
16938 load_index sbitmap even if some bit in it isn't set.
16940 PR middle-end/56461
16941 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
16942 (discover_iteration_bound_by_body_walk): Change queues to
16943 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
16944 spelling in comment. Call safe_push on queues[bound_index] directly.
16945 Release queues[queue_index] in every iteration unconditionally.
16946 Release bounds vector.
16948 PR middle-end/56461
16949 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
16950 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
16951 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
16954 2013-03-05 Richard Biener <rguenther@suse.de>
16957 * tree-inline.c (remap_blocks_to_null): New function.
16958 (expand_call_inline): When expanding a call stmt without
16959 an associated block inline remap all callee blocks to NULL.
16961 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16963 PR rtl-optimization/56494
16964 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
16965 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
16966 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
16968 PR middle-end/56461
16969 * sel-sched-ir.c (free_sched_pools): Release
16970 succs_info_pool.stack[succs_info_pool.max_top] vectors too
16971 if succs_info_pool.max_top isn't -1.
16974 * opts.c (opts_obstack, opts_concat): Moved to...
16975 * opts-common.c (opts_obstack, opts_concat): ... here.
16977 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16979 PR middle-end/56461
16980 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
16982 2013-03-04 Martin Jambor <mjambor@suse.cz>
16984 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
16985 all appropriate places.
16987 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
16989 PR tree-optimization/56424
16990 * ipa-split.c (split_function): Do not set the RSO flag if result is
16991 not by reference and its type is a register type.
16993 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
16995 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
16996 (microblaze_legitimate_pic_operand): Likewise
16997 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
16998 new function microblaze_legitimate_pic_operand
16999 * config/microblaze/microblaze-protos.h
17000 (microblaze_legitimate_pic_operand): Declare.
17002 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17004 * config/microblaze/predicates.md (call_insn_simple_operand):
17005 New predicate for supported rtx code types.
17006 * config/microblaze/microblaze.md (call_internal1): Use
17007 call_insn_simple_operand predicate.
17009 2013-03-04 Jakub Jelinek <jakub@redhat.com>
17011 PR middle-end/56461
17012 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
17013 partitions.ordered_remove.
17015 PR middle-end/56461
17016 * tree-vect-stmts.c (vectorizable_conversion): Don't call
17017 vec_oprnds0.create (1) for modifier == NONE.
17019 PR middle-end/56461
17020 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
17021 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
17022 vec_oprnds1 right before pushing anything to it for
17025 PR middle-end/56461
17026 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
17027 set nbbs to 0 instead of having separate code path.
17028 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
17029 instead of false as last argument if returning NULL.
17031 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
17033 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
17034 the attribute is now called "target" instead of "option".
17035 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
17036 * doc/tm.texi.in (Target Attributes): Likewise document the correct
17037 attribute/pragma name for TARGET_OPTION_VALID_P and
17038 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
17039 * doc/tm.texi: Regenerated.
17041 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
17043 * config/microblaze/microblaze.c:
17044 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
17045 * config/microblaze/microblaze.h: Add -mxl-reorder to
17047 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
17048 instructions emitted if TARGET_REORDER.
17049 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
17050 or 0 for -m/-mno case, but initialises as 2 to detect default use case
17053 2013-03-01 Xinliang David Li <davidxl@google.com>
17055 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
17058 2013-03-01 Jakub Jelinek <jakub@redhat.com>
17060 PR middle-end/56461
17061 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
17062 vector even when returning true. Fix up function comment formatting.
17064 PR middle-end/56461
17065 * ira-build.c (ira_loop_nodes_count): New variable.
17066 (create_loop_tree_nodes): Initialize it.
17067 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
17069 PR middle-end/56461
17070 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
17071 method on dr_chain and result_chain.
17072 * tree-vect-stmts.c (vectorizable_store): Only call
17073 result_chain.create if j == 0.
17075 PR middle-end/56461
17076 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
17077 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
17078 before overwriting it.
17080 2013-03-01 Tobias Burnus <burnus@net-b.de>
17082 * doc/extended.texi (C Extensions): Change order in @menu
17084 (Other MIPS Built-in Functions): Move last MIPS entry before
17085 "picoChip Built-in Functions".
17086 (SH Built-in Functions): Move after RX Built-in Functions.
17087 * doc/gcc.texi (Introduction): Change order in @menu
17089 * doc/md.texi (Constraints): Ditto.
17090 * gty.texi (Type Information): Ditto.
17091 (User-provided marking routines for template types): Make
17093 * doc/invoke.texi (AArch64 Options): Move before
17094 "Adapteva Epiphany Options".
17096 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
17097 Jakub Jelinek <jakub@redhat.com>
17100 * asan.c (gate_asan): Lookup no_sanitize_address instead of
17101 no_address_safety_analysis attribute.
17102 * doc/extend.texi (no_address_safety_attribute): Rename to
17103 no_sanitize_address attribute, mention no_address_safety_analysis
17104 attribute as deprecated alias.
17106 2013-02-28 Jakub Jelinek <jakub@redhat.com>
17108 PR middle-end/56461
17109 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
17110 type to vec<vec<tree> > *.
17111 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
17112 to be vec<tree> instead of vec<tree> *, set vec_defs
17113 to vNULL and call vec_defs.create (number_of_vects), adjust other
17115 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
17116 vectorizable_condition): Adjust vect_get_slp_defs callers.
17118 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17120 * config/aarch64/aarch64.c
17121 (aarch64_float_const_representable): Remove unused variable.
17123 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17125 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
17127 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17129 * config/aarch64/aarch64-builtins.c
17130 (aarch64_init_simd_builtins): Make static.
17132 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17134 * config/aarch64/aarch64.c
17135 (aarch64_simd_make_constant): Make static.
17137 2013-02-28 Martin Jambor <mjambor@suse.cz>
17139 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
17140 with no initialization to the RHS of debug statements.
17142 2013-02-28 Martin Jambor <mjambor@suse.cz>
17144 PR tree-optimization/56294
17145 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
17147 (get_access_replacement): Do not call create_access_replacement.
17148 Assert a replacement exists.
17149 (get_repl_default_def_ssa_name): Create the replacement declaration
17152 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17154 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
17155 final_end_function.
17157 2013-02-28 Marek Polacek <polacek@redhat.com>
17159 PR rtl-optimization/56466
17160 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
17161 if we're changing a loop.
17162 (peel_loops_completely): Likewise.
17164 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
17167 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
17169 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17172 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
17173 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
17174 INTX_FTYPE_FX, FX_FTYPE_INTX.
17175 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
17177 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17179 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
17180 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
17181 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
17182 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
17183 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
17184 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
17185 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
17186 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
17187 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
17188 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
17189 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
17190 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
17191 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
17192 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
17193 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
17194 (avrxmega6): Increase max flash segments from 5 to 6.
17195 * config/avr/t-multilib: Regenerate.
17196 * config/avr/avr-tables.opt: Regenerate.
17197 * doc/avr-mmcu.texi: Regenerate.
17199 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
17201 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
17202 (avr_device_to_arch): Rename to avr_device_to_ld.
17203 (avr_device_to_as): New prototype.
17204 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
17205 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
17206 * config/avr/driver-avr.c (avr_device_to_as): New.
17207 (avr_device_to_arch): Rename to avr_device_to_ld.
17209 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17211 PR middle-end/56461
17212 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
17213 method on dr_chain and result_chain.
17215 PR middle-end/56461
17216 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
17217 pointer_set_destroy on not_executed_last_iteration.
17219 PR middle-end/56461
17220 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
17222 PR middle-end/56461
17223 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
17224 FOR_EACH_DEFINED_FUNCTION when freeing state.
17226 PR middle-end/56461
17227 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
17229 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
17232 PR middle-end/56461
17233 * ipa-cp.c (decide_whether_version_node): Call vec_free on
17234 known_aggs[i].items and release known_aggs vector.
17236 PR middle-end/56461
17237 * ipa-reference.c (propagate): Free node_info even for alias nodes.
17239 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17241 * config/microblaze/microblaze.c (microblaze_emit_compare):
17242 Use xor for EQ/NE comparisions.
17243 * config/microblaze/microblaze.md (cstoresf4): Add constraints
17244 (cbranchsf4): Adjust operator to comparison_operator.
17246 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17248 PR middle-end/56461
17249 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
17251 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
17252 vec_safe_push, always update *slot.
17253 (redirect_edge_var_map_clear): Use vec_free.
17254 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
17255 (free_var_map_entry): Use vec_free.
17256 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
17257 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
17259 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
17261 PR middle-end/45472
17262 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
17263 when the may_trap_p bit of the exprs being merged differs.
17264 Reorder tests for speculativeness in the logical and operator.
17266 2013-02-27 Jakub Jelinek <jakub@redhat.com>
17268 * incpath.c (add_standard_paths): Use reconcat instead of concat
17269 where appropriate and avoid leaking memory.
17271 * opts.h: Include obstack.h.
17272 (opts_concat): New prototype.
17273 (opts_obstack): New declaration.
17274 * opts.c (opts_concat): New function.
17275 (opts_obstack): New variable.
17276 (init_options_struct): Call gcc_init_obstack on opts_obstack.
17277 (finish_options): Use opts_concat instead of concat
17278 and XOBNEWVEC instead of XNEWVEC.
17279 * opts-common.c (generate_canonical_option, decode_cmdline_option,
17280 generate_option): Likewise.
17281 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
17282 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
17285 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
17286 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
17288 2013-02-26 Jakub Jelinek <jakub@redhat.com>
17290 PR middle-end/56461
17291 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
17293 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17295 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
17296 (arm_block_move_unaligned_straight): Likewise.
17297 (arm_adjust_block_mem): Likewise.
17299 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17302 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
17303 temp, cond and label.
17304 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
17307 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
17308 * config/c6x/c6x.h (dbx_register_map): Update declaration.
17311 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
17312 of prologue/epilogue functions.
17315 * config/tilegx/tilegx.c (tilegx_expand_prologue):
17316 Remove unused variable cfa_offset.
17317 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
17320 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
17321 type promotion to unsigned.
17324 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
17325 for HOST_WIDE_INT of 32 bit / same size as int.
17326 (arm_block_move_unaligned_straight): Likewise.
17327 (arm_adjust_block_mem): Likewise.
17330 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
17333 2013-02-26 Marek Polacek <polacek@redhat.com>
17335 PR tree-optimization/56426
17336 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
17338 2013-02-26 Richard Biener <rguenther@suse.de>
17341 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
17342 unused variable loops.
17344 2013-02-26 Jakub Jelinek <jakub@redhat.com>
17346 PR tree-optimization/56448
17347 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
17348 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
17349 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
17350 later operands of the references, or even first operand for
17351 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
17353 PR tree-optimization/56443
17354 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
17355 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
17356 to type_for_mode langhook.
17358 2013-02-25 Matt Turner <mattst88@gmail.com>
17360 * doc/invoke.texi: Document r4700.
17362 2013-02-25 Richard Biener <rguenther@suse.de>
17364 PR tree-optimization/56175
17365 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
17367 (simplify_bitwise_binary): ... here. Also guard the conversion
17368 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
17370 2013-02-25 Catherine Moore <clm@codesourcery.com>
17373 2013-02-24 Catherine Moore <clm@codesourcery.com>
17374 Maciej W. Rozycki <macro@codesourcery.com>
17375 Tom de Vries <tom@codesourcery.com>
17376 Nathan Sidwell <nathan@codesourcery.com>
17377 Iain Sandoe <iain@codesourcery.com>
17378 Nathan Froyd <froydnj@codesourcery.com>
17379 Chao-ying Fu <fu@mips.com>
17381 * doc/extend.texi: (micromips, nomicromips, nocompression):
17382 Document new function attributes.
17383 * doc/invoke.texi (minterlink-compressed, mmicromips,
17384 m14k, m14ke, m14kec): Document new options.
17385 (minterlink-mips16): Update documentation.
17386 * doc/md.texi (ZC, ZD): Document new constraints.
17387 * configure.ac (gcc_cv_as_micromips): Check if linker
17388 supports the .set micromips directive.
17389 * configure: Regenerate.
17390 * config.in: Regenerate.
17391 * config/mips/mips-tables.opt: Regenerate.
17392 * config/mips/micromips.md: New file.
17393 * constraints.md (ZC, AD): New constraints.
17394 * config/mips/predicates.md (movep_src_register): New predicate.
17395 (movep_src_operand): New predicate.
17396 (non_volatile_mem_operand): New predicate.
17397 * config/mips/mips.md (multimem): New type.
17398 (length): Differentiate between 17-bit and 18-bit branch offsets.
17399 (MOVEP1, MOVEP2): New mode iterator.
17400 (mov_<load>l): Use ZC constraint.
17401 (mov_<load>r): Likewise.
17402 (mov_<store>l): Likewise.
17403 (mov_<store>r): Likewise.
17404 (*branch_equality<mode>_inverted): Add microMIPS support.
17405 (*branch_equality<mode>): Likewise.
17406 (*jump_absolute): Likewise.
17407 (indirect_jump_<mode>): Likewise.
17408 (tablejump_<mode>): Likewise.
17409 (<optab>_internal): Likewise.
17410 (sibcall_internal): Likewise.
17411 (sibcall_value_internal): Likewise.
17412 (prefetch): Use constraint ZD.
17413 * config/mips/mips.opt (minterlink-compressed): New option.
17414 (minterlink-mips16): Now an alias for minterlink-compressed.
17415 (mmicromips): New option.
17416 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17417 (compare_and_swap_12): Likewise.
17418 (sync_add<mode>): Likewise.
17419 (sync_<optab>_12): Likewise.
17420 (sync_old_<optab>_12): Likewise.
17421 (sync_new_<optab>_12): Likewise.
17422 (sync_nand_12): Likewise.
17423 (sync_old_nand_12): Likewise.
17424 (sync_new_nand_12): Likewise.
17425 (sync_sub<mode>): Likewise.
17426 (sync_old_add<mode>): Likewise.
17427 (sync_old_sub<mode>): Likewise.
17428 (sync_new_add<mode>): Likewise.
17429 (sync_new_sub<mode>): Likewise.
17430 (sync_<optab><mode>): Likewise.
17431 (sync_old_<optab><mode>): Likewise.
17432 (sync_new_<optab><mode>): Likewise.
17433 (sync_nand<mode>): Likewise.
17434 (sync_old_nand<mode>): Likewise.
17435 (sync_new_nand<mode>): Likewise.
17436 (sync_lock_test_and_set<mode>): Likewise.
17437 (test_and_set_12): Likewise.
17438 (atomic_compare_and_swap<mode>): Likewise.
17439 (atomic_exchange<mode>_llsc): Likewise.
17440 (atomic_fetch_add<mode>_llsc): Likewise.
17441 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17442 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17443 (umips_save_restore_pattern_p): Likewise.
17444 (umips_load_store_pair_p): Likewise.
17445 (umips_output_load_store_pair): Likewise.
17446 (umips_movep_target_p): Likewise.
17447 (umips_12bit_offset_address_p): Likewise.
17448 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17449 (mips_base_mips16): Rename this...
17450 (mips_base_compression_flags): ...to this. Update all uses.
17451 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17452 (mips_mips16_decl_p): Delete.
17453 (mips_nomips16_decl_p): Delete.
17454 (mips_get_compress_on_flags): New function.
17455 (mips_get_compress_off_flags): New function.
17456 (mips_get_compress_mode): New function.
17457 (mips_get_compress_on_name): New function.
17458 (mips_get_compress_off_name): New function.
17459 (mips_insert_attributes): Support multiple compression types.
17460 (mips_merge_decl_attributes): Likewise.
17461 (umips_12bit_offset_address_p): New function.
17462 (mips_start_function_definition): Emit .set micromips directive.
17463 (mips_call_may_need_jalx_p): New function.
17464 (mips_function_ok_for_sibcall): Add microMIPS support.
17465 (mips_print_operand_punctuation): Support short delay slots and
17467 (umips_swm_mask, umips_swm_encoding): New.
17468 (umips_build_save_restore): New function.
17469 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17470 (was_mips16_p): Remove.
17471 (old_compression_mode): New.
17472 (mips_set_compression_mode): New function.
17473 (mips_set_current_function): Add microMIPS support.
17474 (mips_option_override): Likewise.
17475 (umips_save_restore_pattern_p): New function.
17476 (umips_output_save_restore): New function.
17477 (umips_load_store_pair_p_1): New function.
17478 (umips_load_store_pair_p): New function.
17479 (umips_output_load_store_pair_1): New function.
17480 (umips_output_load_store_pair): New function.
17481 (umips_movep_target_p) New function.
17482 (mips_prepare_pch_save): Add microMIPS support.
17483 * config/mips/mips.h (TARGET_COMPRESSION): New.
17484 (TARGET_CPU_CPP_BUILTINS): Update macro
17485 to use new compression flags and to support microMIPS.
17486 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17487 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17488 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17489 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17490 (ASM_SPEC): Support mmicromips and mno-micromips.
17491 (M16STORE_REG_P): New macro.
17492 (MIPS_CALL): Support TARGET_MICROMIPS.
17493 (MICROMIPS_J): New macro.
17494 (mips_base_mips16): Rename this...
17495 (mips_base_compression_flags): ...to this.
17496 (UMIPS_12BIT_OFFSET_P): New macro.
17497 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17498 (MULTILIB_DIRNAMES): Likewise.
17500 2013-02-25 Tom de Vries <tom@codesourcery.com>
17502 PR rtl-optimization/56131
17503 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
17504 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
17505 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
17507 2013-02-25 Tobias Burnus <burnus@net-b.de>
17509 * doc/invoke.texi (-fsanitize=): Move from optimization
17510 to debugging options.
17512 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
17514 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
17516 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
17517 Alexander Monakov <amonakov@ispras.ru>
17519 PR middle-end/56077
17520 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
17521 flush pending lists also on non-jumps. Adjust comment.
17523 2013-02-24 Catherine Moore <clm@codesourcery.com>
17524 Maciej W. Rozycki <macro@codesourcery.com>
17525 Tom de Vries <tom@codesourcery.com>
17526 Nathan Sidwell <nathan@codesourcery.com>
17527 Iain Sandoe <iain@codesourcery.com>
17528 Nathan Froyd <froydnj@codesourcery.com>
17529 Chao-ying Fu <fu@mips.com>
17531 * doc/extend.texi: (micromips, nomicromips, nocompression):
17532 Document new function attributes.
17533 * doc/invoke.texi (minterlink-compressed, mmicromips,
17534 m14k, m14ke, m14kec): Document new options.
17535 (minterlink-mips16): Update documentation.
17536 * doc/md.texi (ZC, ZD): Document new constraints.
17537 * configure.ac (gcc_cv_as_micromips): Check if linker
17538 supports the .set micromips directive.
17539 * configure: Regenerate.
17540 * config.in: Regenerate.
17541 * config/mips/mips-tables.opt: Regenerate.
17542 * config/mips/micromips.md: New file.
17543 * constraints.md (ZC, AD): New constraints.
17544 * config/mips/predicates.md (movep_src_register): New predicate.
17545 (movep_src_operand): New predicate.
17546 (non_volatile_mem_operand): New predicate.
17547 * config/mips/mips.md (multimem): New type.
17548 (length): Differentiate between 17-bit and 18-bit branch offsets.
17549 (MOVEP1, MOVEP2): New mode iterator.
17550 (mov_<load>l): Use ZC constraint.
17551 (mov_<load>r): Likewise.
17552 (mov_<store>l): Likewise.
17553 (mov_<store>r): Likewise.
17554 (*branch_equality<mode>_inverted): Add microMIPS support.
17555 (*branch_equality<mode>): Likewise.
17556 (*jump_absolute): Likewise.
17557 (indirect_jump_<mode>): Likewise.
17558 (tablejump_<mode>): Likewise.
17559 (<optab>_internal): Likewise.
17560 (sibcall_internal): Likewise.
17561 (sibcall_value_internal): Likewise.
17562 (prefetch): Use constraint ZD.
17563 * config/mips/mips.opt (minterlink-compressed): New option.
17564 (minterlink-mips16): Now an alias for minterlink-compressed.
17565 (mmicromips): New option.
17566 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17567 (compare_and_swap_12): Likewise.
17568 (sync_add<mode>): Likewise.
17569 (sync_<optab>_12): Likewise.
17570 (sync_old_<optab>_12): Likewise.
17571 (sync_new_<optab>_12): Likewise.
17572 (sync_nand_12): Likewise.
17573 (sync_old_nand_12): Likewise.
17574 (sync_new_nand_12): Likewise.
17575 (sync_sub<mode>): Likewise.
17576 (sync_old_add<mode>): Likewise.
17577 (sync_old_sub<mode>): Likewise.
17578 (sync_new_add<mode>): Likewise.
17579 (sync_new_sub<mode>): Likewise.
17580 (sync_<optab><mode>): Likewise.
17581 (sync_old_<optab><mode>): Likewise.
17582 (sync_new_<optab><mode>): Likewise.
17583 (sync_nand<mode>): Likewise.
17584 (sync_old_nand<mode>): Likewise.
17585 (sync_new_nand<mode>): Likewise.
17586 (sync_lock_test_and_set<mode>): Likewise.
17587 (test_and_set_12): Likewise.
17588 (atomic_compare_and_swap<mode>): Likewise.
17589 (atomic_exchange<mode>_llsc): Likewise.
17590 (atomic_fetch_add<mode>_llsc): Likewise.
17591 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17592 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17593 (umips_save_restore_pattern_p): Likewise.
17594 (umips_load_store_pair_p): Likewise.
17595 (umips_output_load_store_pair): Likewise.
17596 (umips_movep_target_p): Likewise.
17597 (umips_12bit_offset_address_p): Likewise.
17598 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17599 (mips_base_mips16): Rename this...
17600 (mips_base_compression_flags): ...to this. Update all uses.
17601 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17602 (mips_mips16_decl_p): Delete.
17603 (mips_nomips16_decl_p): Delete.
17604 (mips_get_compress_on_flags): New function.
17605 (mips_get_compress_off_flags): New function.
17606 (mips_get_compress_mode): New function.
17607 (mips_get_compress_on_name): New function.
17608 (mips_get_compress_off_name): New function.
17609 (mips_insert_attributes): Support multiple compression types.
17610 (mips_merge_decl_attributes): Likewise.
17611 (umips_12bit_offset_address_p): New function.
17612 (mips_start_function_definition): Emit .set micromips directive.
17613 (mips_call_may_need_jalx_p): New function.
17614 (mips_function_ok_for_sibcall): Add microMIPS support.
17615 (mips_print_operand_punctuation): Support short delay slots and
17617 (umips_swm_mask, umips_swm_encoding): New.
17618 (umips_build_save_restore): New function.
17619 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17620 (was_mips16_p): Remove.
17621 (old_compression_mode): New.
17622 (mips_set_compression_mode): New function.
17623 (mips_set_current_function): Add microMIPS support.
17624 (mips_option_override): Likewise.
17625 (umips_save_restore_pattern_p): New function.
17626 (umips_output_save_restore): New function.
17627 (umips_load_store_pair_p_1): New function.
17628 (umips_load_store_pair_p): New function.
17629 (umips_output_load_store_pair_1): New function.
17630 (umips_output_load_store_pair): New function.
17631 (umips_movep_target_p) New function.
17632 (mips_prepare_pch_save): Add microMIPS support.
17633 * config/mips/mips.h (TARGET_COMPRESSION): New.
17634 (TARGET_CPU_CPP_BUILTINS): Update macro
17635 to use new compression flags and to support microMIPS.
17636 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17637 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17638 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17639 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17640 (ASM_SPEC): Support mmicromips and mno-micromips.
17641 (M16STORE_REG_P): New macro.
17642 (MIPS_CALL): Support TARGET_MICROMIPS.
17643 (MICROMIPS_J): New macro.
17644 (mips_base_mips16): Rename this...
17645 (mips_base_compression_flags): ...to this.
17646 (UMIPS_12BIT_OFFSET_P): New macro.
17647 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17648 (MULTILIB_DIRNAMES): Likewise.
17650 2013-02-24 Jakub Jelinek <jakub@redhat.com>
17653 * target-globals.c (save_target_globals): For init_reg_sets and
17654 target_reinit remporarily set this_fn_optabs to this_target_optabs.
17656 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
17658 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
17659 * config/aarch64/t-aarch64
17660 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
17662 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
17664 PR inline-asm/56148
17665 * lra-constraints.c (process_alt_operands): Reload operand
17666 conflicting with earlier clobber only if no more other conflicting
17669 2013-02-22 Jakub Jelinek <jakub@redhat.com>
17672 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
17673 if not linking a shared library.
17675 2013-02-22 Seth LaForge <sethml@google.com>
17677 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
17679 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
17681 * config/arm/arm.md (split for extendsidi): Update condition.
17682 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
17683 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
17684 (qhs_zextenddi_cstr): Likewise.
17686 2013-02-21 Jakub Jelinek <jakub@redhat.com>
17688 PR middle-end/56420
17689 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
17690 avoid signed wrapping.
17691 (expand_mult): Handle properly multiplication by
17692 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
17693 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
17694 in the compiler if coeff is HOST_WIDE_INT_MIN.
17695 (expand_divmod): Don't make ext_op1 static, change it's type to
17696 uhwi. Avoid undefined behavior in -INTVAL (op1).
17698 PR rtl-optimization/50339
17699 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
17701 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
17702 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
17703 into splitting_ashiftrt field.
17704 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
17706 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
17709 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
17711 PR middle-end/56108
17712 * trans-mem.c (execute_tm_mark): Do not expand transactions that
17713 are sure to go irrevocable.
17715 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
17717 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
17718 scalars are valid operands.
17720 2013-02-21 Martin Jambor <mjambor@suse.cz>
17722 PR tree-optimization/56310
17723 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
17724 only matching indices and non-negative final offsets.
17725 (intersect_aggregates_with_edge): Pass src_idx to
17726 agg_replacements_to_vector. Pass src_idx insstead of index to
17727 intersect_with_agg_replacements.
17729 2013-02-21 Martin Jambor <mjambor@suse.cz>
17731 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
17732 instead of hard-wired defaults.
17734 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
17736 * doc/invoke.texi (MIPS Options): Update documentation of the
17737 floating-point multiply-accumulate instruction restrictions.
17739 2013-02-21 Kostya Serebryany <kcc@google.com>
17741 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
17742 asan_shadow_offset on x86_64 linux.
17744 2013-02-21 Richard Biener <rguenther@suse.de>
17746 PR tree-optimization/56415
17748 2013-02-11 Richard Biener <rguenther@suse.de>
17750 PR tree-optimization/56273
17751 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
17754 2013-02-21 Jakub Jelinek <jakub@redhat.com>
17757 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
17760 PR inline-asm/56405
17761 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
17762 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
17764 2013-02-20 Jan Hubicka <jh@suse.cz>
17766 PR tree-optimization/56265
17767 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
17768 when target is referenced for first time.
17770 2013-02-20 Richard Biener <rguenther@suse.de>
17772 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
17773 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
17774 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
17775 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
17776 not return anything.
17777 (rename_ssa_copies): Do not remove unused locals.
17778 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
17779 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
17780 * passes.c (execute_function_todo): Do not schedule unused locals
17781 removal if cleanup_tree_cfg did something.
17782 * tree-ssa-live.c (remove_unused_locals): Dump statistics
17783 about the number of removed locals.
17785 2013-02-20 Richard Biener <rguenther@suse.de>
17787 PR tree-optimization/56398
17788 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
17790 2013-02-20 Martin Jambor <mjambor@suse.cz>
17792 PR tree-optimization/55334
17793 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
17794 restricted pointers to arrays.
17796 2013-02-20 Richard Biener <rguenther@suse.de>
17797 Jakub Jelinek <jakub@redhat.com>
17799 PR tree-optimization/56396
17800 * tree-ssa-ccp.c (n_const_val): New static variable.
17801 (get_value): Return NULL for SSA names we don't have a lattice
17803 (ccp_initialize): Initialize n_const_val.
17804 * tree-ssa-copy.c (n_copy_of): New static variable.
17805 (init_copy_prop): Initialize n_copy_of.
17806 (get_value): Return NULL_TREE for SSA names we don't have a
17809 2013-02-20 Martin Jambor <mjambor@suse.cz>
17811 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
17813 2013-02-20 Richard Biener <rguenther@suse.de>
17815 * genpreds.c (write_lookup_constraint): Do not compare first
17816 letter of the constraint again.
17818 2013-02-20 Richard Biener <rguenther@suse.de>
17820 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
17822 (get_use_iv_cost): Terminate hashtable walk when coming across
17825 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
17827 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
17828 reassociation for avx2 targets.
17830 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17832 * config/microblaze/microblaze.c: microblaze_has_clz = 0
17833 Add version check for v8.10.a to enable microblaze_has_clz
17834 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
17835 version and TARGET_PATTERN_COMPARE check
17836 * config/microblaze/microblaze.md: New clzsi2 instruction
17838 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17840 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
17841 function before branching.
17843 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17845 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
17847 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
17849 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17851 PR middle-end/55889
17852 * sel-sched.c: Include ira.h.
17853 (implicit_clobber_conflict_p): New function.
17854 (moveup_expr): Use it.
17855 * Makefile.in (sel-sched.o): Depend on ira.h.
17857 2013-02-19 Richard Biener <rguenther@suse.de>
17859 PR tree-optimization/56384
17860 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
17861 (vn_hash_type): Split out from ...
17862 (vn_hash_constant_with_type): ... here.
17863 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
17864 (vn_phi_eq): Compare types from vn_phi_s structure.
17865 (vn_phi_lookup): Populate vn_phi_s type.
17866 (vn_phi_insert): Likewise.
17868 2013-02-19 Jakub Jelinek <jakub@redhat.com>
17870 PR tree-optimization/56350
17871 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
17872 if haven't found reduction or nested cycle operand, rather than
17873 asserting we must find it.
17875 PR tree-optimization/56381
17876 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
17879 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
17880 Jakub Jelinek <jakub@redhat.com>
17883 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
17884 (swap_optab_enable): Same.
17885 (init_all_optabs): Use argument instead of global.
17886 * tree.h (struct tree_optimization_option): New field target_optabs.
17887 * expr.h (init_all_optabs): Add argument to prototype.
17888 (TREE_OPTIMIZATION_OPTABS): New.
17889 (save_optabs_if_changed): Protoize.
17890 * optabs.h: Declare this_fn_optabs.
17891 * optabs.c (save_optabs_if_changed): New.
17892 Declare this_fn_optabs.
17893 (init_optabs): Add argument to init_all_optabs() call.
17894 * function.c (invoke_set_current_function_hook): Handle per
17896 * function.h (struct function): New field optabs.
17897 * config/mips/mips.c (mips_set_mips16_mode): Handle when
17898 optimization_current_node has changed.
17899 * target-globals.h (save_target_globals_default_opts): Protoize.
17900 * target-globals.c (save_target_globals_default_opts): New.
17902 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17905 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
17906 registers %fr12 and %fr12R as call used.
17909 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
17910 and HImode, require all displacements to be an integer multiple of
17912 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
17913 only allow QImode and HImode when reload is in progress and strict is
17914 true. Likewise for symbolic addresses. Use base14_operand to check
17915 displacements in REG+BASE addresses.
17917 2013-02-18 Richard Biener <rguenther@suse.de>
17919 PR tree-optimization/56366
17920 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17921 handle sign-conversion of outer-loop initial induction value.
17923 2013-02-18 Richard Biener <rguenther@suse.de>
17925 PR middle-end/56349
17926 * cfghooks.c (merge_blocks): If we merge a latch into another
17927 block adjust references to it.
17928 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
17929 (verify_loop_structure): Verify that a recorded latch is in fact
17932 2013-02-18 Richard Biener <rguenther@suse.de>
17934 PR tree-optimization/56321
17935 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
17936 order SSA name release and virtual operand unlinking.
17938 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17940 * config/microblaze/microblaze.md (save_stack_block): Define.
17941 (restore_stack_block): Likewise.
17943 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17945 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
17946 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
17947 * config/microblaze/microblaze.c (microblaze_option_override):
17948 Bail out early for PIC modes when target does not support PIC.
17950 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17952 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
17953 Replace with a microblaze version.
17954 (microblaze_trampoline_init): Adapt for microblaze.
17955 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
17958 2013-02-16 Jakub Jelinek <jakub@redhat.com>
17959 Dodji Seketeli <dodji@redhat.com>
17962 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
17963 (instrument_mem_region_access): Do not forget to always put
17964 instrumentation of the of 'base' and 'base + len' in a "if (len !=
17965 0) statement, even for cases where either 'base' or 'base + len'
17966 are not instrumented -- because they have been previously
17967 instrumented. Simplify the logic by putting all the statements
17968 instrument 'base + len' inside a sequence, and then insert that
17969 sequence right before the current insertion point. Then, to
17970 instrument 'base + len', just get an iterator on that statement.
17971 And do not forget to update the pointer to iterator the function
17972 received as argument.
17974 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
17976 PR rtl-optimization/56348
17977 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
17979 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
17981 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
17982 (clean_graph_dump_file): Pass base to start_graph_dump.
17984 2013-02-14 Richard Henderson <rth@redhat.com>
17987 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
17989 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
17991 * collect2-aix.h: Define F_LOADONLY.
17993 2013-02-14 Richard Biener <rguenther@suse.de>
17996 * varasm.c (output_constant_def_1): Get the decl representing
17997 the constant as argument.
17998 (output_constant_def): Wrap output_constant_def_1.
17999 (make_decl_rtl): Use output_constant_def_1 with the decl
18000 representing the constant.
18001 (build_constant_desc): Optionally re-use a decl already
18002 representing the constant.
18003 (tree_output_constant_def): Adjust.
18005 2013-02-14 Dodji Seketeli <dodji@redhat.com>
18008 * asan.c (instrument_builtin_call): Really put the length of the
18009 second source argument into src1_len.
18011 2013-02-13 Jakub Jelinek <jakub@redhat.com>
18013 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
18014 argument. If it is false, don't create edge from then_bb to
18016 (insert_if_then_before_iter): Pass true to it.
18017 (build_check_stmt): Pass false to it.
18018 (transform_statements): Flush hash table only on extended basic
18019 block boundaries, rather than at the beginning of every bb.
18020 Don't flush hash table on nonfreeing_call_p calls.
18021 * tree-flow.h (nonfreeing_call_p): New prototype.
18022 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
18024 2013-02-13 David S. Miller <davem@davemloft.net>
18026 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
18028 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
18031 * ira.c (max_regno_before_ira): Move from ...
18033 (fix_reg_equiv_init): Use max_regno_before_ira instead of
18036 2013-02-13 Jakub Jelinek <jakub@redhat.com>
18038 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
18040 2013-02-13 Richard Biener <rguenther@suse.de>
18043 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
18044 globals in MEM_REFs.
18046 2013-02-13 Richard Biener <rguenther@suse.de>
18048 * loop-init.c (loop_optimizer_init): Clear loop state when
18049 re-initializing preserved loops.
18050 * loop-unswitch.c (unswitch_single_loop): Return whether
18051 we unswitched the loop. Do not verify loop state here.
18052 (unswitch_loops): When we unswitched a loop discover new loops.
18054 2013-02-13 Kostya Serebryany <kcc@google.com>
18056 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
18058 * sanitizer.def: Rename __asan_init to __asan_init_v1.
18060 2013-02-12 Dodji Seketeli <dodji@redhat.com>
18062 Avoid instrumenting duplicated memory access in the same basic block
18063 * Makefile.in (asan.o): Add new dependency on hash-table.h
18064 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
18065 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
18066 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
18067 (free_mem_ref_resources, has_mem_ref_been_instrumented)
18068 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
18069 (get_mem_ref_of_assignment): New functions.
18070 (get_mem_refs_of_builtin_call): Extract from
18071 instrument_builtin_call and tweak a little bit to make it fit with
18073 (instrument_builtin_call): Use the new
18074 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
18075 of is_gimple_builtin_call.
18076 (instrument_derefs, instrument_mem_region_access): Insert the
18077 instrumented memory reference into the hash table.
18078 (maybe_instrument_assignment): Renamed instrument_assignment into
18079 this, and change it to advance the iterator when instrumentation
18080 actually happened and return true in that case. This makes it
18081 homogeneous with maybe_instrument_assignment, and thus give a
18082 chance to callers to be more 'regular'.
18083 (transform_statements): Clear the memory reference hash table
18084 whenever we enter a new BB, when we cross a function call, or when
18085 we are done transforming statements. Use
18086 maybe_instrument_assignment instead of instrumentation. No more
18087 need to special case maybe_instrument_assignment and advance the
18088 iterator after calling it; it's now handled just like
18089 maybe_instrument_call. Update comment.
18091 2013-02-13 Richard Biener <rguenther@suse.de>
18093 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
18094 Fix loop discovery code.
18096 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
18098 PR inline-asm/56148
18099 * lra-constraints.c (process_alt_operands): Match early clobber
18100 operand with itself. Check conflicts with earlyclobber only if
18101 the operand is not reloaded. Prefer to reload conflicting operand
18102 if earlyclobber and matching operands are the same.
18104 2013-02-12 Richard Biener <rguenther@suse.de>
18107 * lto-streamer-out.c (write_symbol): Do not output symbols
18108 for hard register variables.
18110 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
18113 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
18114 (umulsidi3_insn, mulsidi3_insn): New insns.
18116 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
18118 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
18119 (struct tune_params): Add vec_costs field.
18120 * config/arm/arm.c (arm_builtin_vectorization_cost)
18121 (arm_add_stmt_cost): New functions.
18122 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
18123 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
18124 (arm_default_vec_cost): New struct of type cpu_vec_costs.
18125 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
18126 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
18127 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
18128 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
18130 2013-02-12 Richard Biener <rguenther@suse.de>
18133 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
18134 decls again if possible.
18136 2013-02-12 Richard Biener <rguenther@suse.de>
18138 PR middle-end/56288
18139 * tree-ssa.c (verify_ssa_name): Fix check, move
18140 SSA_NAME_IN_FREE_LIST check up.
18142 2013-02-12 Jakub Jelinek <jakub@redhat.com>
18143 Steven Bosscher <steven@gcc.gnu.org>
18145 PR rtl-optimization/56151
18146 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
18147 equal to op0 or op1, and last_insn pattern is CODE operation
18148 with MEM dest and one of the operands matches that MEM.
18150 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
18152 * doc/extend.texi: Document Function Multiversioning and "default"
18153 parameter string to target attribute.
18154 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
18155 target attribute parameter is "default".
18156 (ix86_compare_version_priority): Remove checks for target attribute.
18157 (ix86_mangle_function_version_assembler_name): Change error to sorry.
18158 Remove check for target attribute equal to NULL. Add assert.
18159 (ix86_generate_version_dispatcher_body): Change error to sorry.
18161 2013-02-11 Iain Sandoe <iain@codesourcery.com>
18162 Jack Howarth <howarth@bromo.med.uc.edu>
18163 Patrick Marlier <patrick.marlier@gmail.com>
18166 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
18167 define ENDFILE_SPEC as TM_DESTRUCTOR.
18168 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
18170 2013-02-11 Alexander Potapenko <glider@google.com>
18171 Jack Howarth <howarth@bromo.med.uc.edu>
18172 Jakub Jelinek <jakub@redhat.com>
18175 * config/darwin.c (cdtor_record): Rename ctor_record.
18176 (sort_cdtor_records): Rename sort_ctor_records.
18177 (finalize_dtors): New routine to sort destructors by
18178 priority before use in assemble_integer.
18179 (machopic_asm_out_destructor): Use finalize_dtors if needed.
18181 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
18183 PR rtl-optimization/56275
18184 * simplify-rtx.c (avoid_constant_pool_reference): Check that
18185 offset is non-negative and less than cmode size before
18186 calling simplify_subreg.
18188 2013-02-11 Richard Biener <rguenther@suse.de>
18190 PR tree-optimization/56264
18191 * cfgloop.h (fix_loop_structure): Adjust prototype.
18192 * loop-init.c (fix_loop_structure): Return the number of
18193 newly discovered loops.
18194 * tree-cfgcleanup.c (repair_loop_structures): When new loops
18195 are discovered, do a full loop-closed SSA rewrite.
18197 2013-02-11 Richard Biener <rguenther@suse.de>
18199 PR tree-optimization/56273
18200 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18202 (check_array_ref): Fix missing newline in dumps.
18203 (search_for_addr_array): Likewise.
18205 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
18207 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
18209 2013-02-09 Jakub Jelinek <jakub@redhat.com>
18212 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
18214 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
18216 PR rtl-optimization/56246
18217 * lra-constraints.c (simplify_operand_subreg): Try to reuse
18219 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
18220 constraints are satisfied.
18222 2013-02-08 Jeff Law <law@redhat.com>
18225 * emit-rtl.c (reg_is_parm_p): New function.
18226 * regs.h (reg_is_parm_p): New prototype.
18227 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
18228 callee-clobbered registers.
18230 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
18233 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
18234 If there is no implicit builtin declaration, just return NULL.
18236 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18238 * config/i386/sse.md (FMAMODEM): New mode iterator.
18239 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
18240 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
18242 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18244 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
18245 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
18246 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
18248 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18250 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
18251 (microblaze*-*-elf): Likewise.
18252 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
18254 * config/microblaze/microblaze-c.c: Add builtin defines for
18255 _LITTLE_ENDIAN and _BIG_ENDIAN.
18256 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
18257 add to TARGET_DEFAULT flags.
18258 Expand ASM_SPEC and LINK_SPEC.
18259 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
18260 * config/microblaze/microblaze.md: Update extendsidi2 and
18261 movdi_internal instructions to use low-order / high-order reg
18263 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
18264 options and inversemask / mask of LITTLE_ENDIAN.
18265 * config/microblaze/t-microblaze: Expand multilib options to
18266 include mlittle-endian (le) and update exceptions patterns.
18268 2013-02-08 Jakub Jelinek <jakub@redhat.com>
18270 PR rtl-optimization/56195
18271 * lra-constraints.c (get_reload_reg): Don't reuse regs
18272 if they have smaller mode than requested, if they have
18273 wider mode than requested, try to return a SUBREG.
18275 PR tree-optimization/56250
18276 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
18277 if type is unsigned and code isn't MULT_EXPR.
18279 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
18281 PR tree-optimization/56064
18282 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
18283 bits according to mode.
18284 * fixed-value.h (fixed_from_double_int)
18285 (const_fixed_from_double_int): Adjust comments.
18287 2013-02-08 Richard Biener <rguenther@suse.de>
18290 * lto-streamer.h (struct data_in): Remove current_file, current_line
18291 and current_col members.
18292 * lto-streamer-out.c (lto_output_location): Stream changed bits
18293 en-block for efficiency.
18294 * lto-streamer-in.c (clear_line_info): Remove.
18295 (lto_input_location): Cache current file, line and column
18296 globally via local statics. Read changed bits en-block.
18297 (input_function): Do not call clear_line_info.
18298 (lto_read_body): Likewise.
18299 (lto_input_toplevel_asms): Likewise.
18301 2013-02-08 Michael Matz <matz@suse.de>
18303 PR tree-optimization/52448
18304 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
18305 (nt_call_phase): New static.
18306 (add_or_mark_expr): Only mark accesses with newer phase than any
18308 (nonfreeing_call_p): New.
18309 (nt_init_block): Update nt_call_phase, mark blocks as visited.
18310 (nt_fini_block): Keep blocks marked as visited.
18311 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
18313 2013-02-08 Richard Biener <rguenther@suse.de>
18315 * ira.c (ira): Free broken dominator information.
18317 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
18319 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
18321 2013-02-08 Marek Polacek <polacek@redhat.com>
18323 * cfgloop.c (verify_loop_structure): Add more checking of headers.
18325 2013-02-08 Richard Biener <rguenther@suse.de>
18327 PR middle-end/56181
18328 * cfgloop.h (flow_loops_find): Adjust.
18329 (bb_loop_header_p): Declare.
18330 * cfgloop.c (bb_loop_header_p): New function split out from ...
18331 (flow_loops_find): ... here. Adjust function signature,
18332 support incremental loop structure update.
18333 (verify_loop_structure): Cleanup. Verify a loop is a loop.
18334 * cfgloopmanip.c (fix_loop_structure): Move ...
18335 * loop-init.c (fix_loop_structure): ... here.
18336 (apply_loop_flags): Split out from ...
18337 (loop_optimizer_init): ... here.
18338 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
18339 in incremental mode, only remove dead loops here.
18341 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
18344 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
18345 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
18346 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
18347 (*round<mode>3.libgcc): New insns for fixed-modes.
18348 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
18349 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
18350 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
18351 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
18352 implementations. Define to __builtin_avr_absFX,
18353 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
18354 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
18355 __builtin_avr_countlsFX, respectively.
18356 * config/avr/avr-c.c (target.h): Include it.
18357 (enum avr_builtin_id): New enum.
18358 (avr_resolve_overloaded_builtin): New static function.
18359 (avr_register_target_pragmas): Use it to set
18360 targetm.resolve_overloaded_builtin.
18361 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
18362 tree nodes used by DEF_BUILTIN.
18363 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
18364 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
18365 <AVR_BUILTIN_xxBITS>: Same.
18367 2013-02-08 Richard Biener <rguenther@suse.de>
18369 * cfgloop.c (verify_loop_structure): Properly handle
18370 a loop exiting to another loop header.
18371 * ira-int.h (ira_loops): Remove.
18372 * ira.c (ira_loops): Remove.
18373 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
18374 (do_reload): Use loop_optimizer_finalize.
18375 * ira-build.c (create_loop_tree_nodes): Use get_loops and
18376 number_of_loops to access the loop tree.
18377 (more_one_region_p): Likewise.
18378 (finish_loop_tree_nodes): Likewise.
18379 (rebuild_regno_allocno_maps): Likewise.
18380 (mark_loops_for_removal): Likewise.
18381 (mark_all_loops_for_removal): Likewise.
18382 (remove_unnecessary_regions): Likewise.
18383 (ira_build): Likewise.
18384 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18386 2013-02-08 Richard Biener <rguenther@suse.de>
18388 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
18389 * ipa-pure-const.c (analyze_function): Avoid calling
18390 mark_irreducible_loops twice.
18391 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
18393 2013-02-07 David S. Miller <davem@davemloft.net>
18395 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
18397 * var-tracking.c (vt_add_function_parameter): Test the presence of
18398 HAVE_window_save properly and do not remap argument registers when
18399 we have a leaf function.
18401 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
18404 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
18406 * config/i386/i386.c (ix86_print_operand): Ditto.
18408 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18410 * lra-constraints.c (process_alt_operands): Fix recently added comment.
18412 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18414 PR rtl-optimization/56225
18415 * lra-constraints.c (process_alt_operands): Check that reload hard
18416 reg can hold value for strict_low_part.
18418 2013-02-07 Jakub Jelinek <jakub@redhat.com>
18421 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
18422 dwarf2out_end_function.
18423 (in_first_function_p, maybe_at_text_label_p,
18424 first_loclabel_num_not_at_text_label): New variables.
18425 (dwarf2out_var_location): In the first function find out
18426 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
18427 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
18430 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
18432 PR rtl-optimization/56178
18433 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
18434 SUBREG of a register. Tidy up related block of code.
18435 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
18436 note if the source is a register or a SUBREG of a register.
18438 2013-02-07 Jakub Jelinek <jakub@redhat.com>
18441 * config/rs6000/rs6000.md (ptrm): New mode attr.
18442 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
18443 call_value_indirect_aix<pttrsize>,
18444 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
18447 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18449 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
18450 if -bnortl. Convert to strcmp and strncmp.
18452 2013-02-07 Alan Modra <amodra@gmail.com>
18455 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
18456 addresses won't wrap when offsetting.
18457 (rs6000_secondary_reload): Provide secondary reloads needed for
18458 wrapping LO_SUM addresses.
18460 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
18462 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
18463 MACH, just __MACH__.
18465 2013-02-06 Richard Biener <rguenther@suse.de>
18467 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
18468 instead of calling fix_loop_structure.
18470 2013-02-06 Jakub Jelinek <jakub@redhat.com>
18472 PR middle-end/56217
18473 * omp-low.c (use_pointer_for_field): Return false if
18474 lower_send_shared_vars doesn't generate any copy-out code.
18476 2013-02-06 Tom de Vries <tom@codesourcery.com>
18478 PR rtl-optimization/56131
18479 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
18480 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
18481 of the label is NULL. Add comment.
18483 2013-02-05 Jakub Jelinek <jakub@redhat.com>
18485 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
18488 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
18489 (STATIC_LIBTSAN_LIBS): Likewise.
18490 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
18491 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
18492 is defined, don't add anything else beyond that.
18493 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
18494 (LINK_COMMAND_SPEC): Use them.
18496 PR tree-optimization/56205
18497 * tree-stdarg.c (check_all_va_list_escapes): Return true if
18498 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
18499 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
18501 2013-02-05 Richard Biener <rguenther@suse.de>
18503 PR tree-optimization/53342
18504 PR tree-optimization/53185
18505 * tree-vectorizer.h (vect_check_strided_load): Remove.
18506 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
18507 not disallow peeling for vectorized strided loads.
18508 (vect_check_strided_load): Make static and simplify.
18509 (vect_analyze_data_refs): Adjust.
18510 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
18511 correctly when vectorizing strided loads.
18513 2013-02-05 Richard Biener <rguenther@suse.de>
18515 * doc/install.texi: Refer to ISL, not PPL.
18517 2013-02-05 Jan Hubicka <jh@suse.cz>
18519 PR tree-optimization/55789
18520 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
18522 2013-02-05 Jan Hubicka <jh@suse.cz>
18524 PR tree-optimization/55789
18525 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
18526 the dead call anyway.
18528 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
18531 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
18533 2013-02-04 Alexander Potapenko <glider@google.com>
18534 Jack Howarth <howarth@bromo.med.uc.edu>
18535 Jakub Jelinek <jakub@redhat.com>
18538 * config/darwin.c (sort_ctor_records): Stabilized qsort
18539 on constructor priority by using original position.
18540 (finalize_ctors): New routine to sort constructors by
18541 priority before use in assemble_integer.
18542 (machopic_asm_out_constructor): Use finalize_ctors if needed.
18544 2013-02-04 Jakub Jelinek <jakub@redhat.com>
18547 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
18548 about visibility on artificial decls.
18549 * config/sol2.c (solaris_assemble_visibility): Likewise.
18551 2013-02-04 Kai Tietz <ktietz@redhat.com>
18554 * config/i386/i386.c (function_value_ms_64): Add additional valtype
18555 argument and improve checking of return-argument types for 16-byte
18557 (ix86_function_value_1): Add additional valtype argument on call
18558 of function_value_64.
18559 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
18560 handling infunction_value_64 function.
18562 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
18564 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
18566 2013-02-04 Richard Biener <rguenther@suse.de>
18568 PR tree-optimization/56188
18569 * tree-ssa-structalias.c (label_visit): Consider case with
18570 initially non-empty points-to set.
18571 (perform_var_substitution): Dump node mapping and clean up.
18573 2013-02-04 Richard Guenther <rguenther@suse.de>
18576 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
18577 node prevail as last resort.
18578 (lto_symtab_merge_decls): Remove guard on LTRANS here.
18579 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
18581 2013-02-04 Richard Biener <rguenther@suse.de>
18583 PR tree-optimization/56113
18584 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
18586 (equiv_class_lookup_or_add): ... this.
18587 (label_visit): Adjust and fix error in previous patch.
18588 (perform_var_substitution): Adjust.
18590 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
18592 * config/sh/divtab.c: Fix formatting and comments throughout the file.
18593 * config/sh/sh4-300.md: Likewise.
18594 * config/sh/sh4a.md: Likewise.
18595 * config/sh/constraints.md: Likewise.
18596 * config/sh/sh.md: Likewise.
18597 * config/sh/netbsd-elf.h: Likewise.
18598 * config/sh/predicates.md: Likewise.
18599 * config/sh/sh-protos.h: Likewise.
18600 * config/sh/ushmedia.h: Likewise.
18601 * config/sh/linux.h: Likewise.
18602 * config/sh/sh.c: Likewise.
18603 * config/sh/superh.h: Likewise.
18604 * config/sh/elf.h: Likewise.
18605 * config/sh/sh4.md: Likewise.
18606 * config/sh/sh.h: Likewise.
18608 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18610 * config/pa/constraints.md: Adjust unused letters. Change "T"
18611 constraint to match_test floating_point_store_memory_operand().
18612 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
18613 (base14_operand): New.
18614 (floating_point_store_memory_operand): New.
18615 (integer_store_memory_operand): Revise to use base14_operand and
18616 reg_plus_base_memory_operand.
18617 (move_dest_operand): Allow symbolic_memory_operands.
18618 (symbolic_memory_operand): Check for LO_SOM.
18619 (symbolic_operand): Change default case to break.
18620 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
18621 CONST_DOUBLE values to be reloaded by putting them into memory when
18622 the destination is a floating point register.
18623 (movdf): Remove code to handle CONST_DOUBLE.
18625 (reload_indf_r1): New.
18626 (reload_insf_r1): New.
18627 Consistently use "Q" and "T" constraints with integer and floating
18628 point move instructions, respectively.
18629 (movdi): Remove FAIL.
18630 Change predicate for source operand unamed DImode move from
18631 general_operand to move_src_operand.
18632 (umulsidi3): Change predicate for destination operand to
18634 Likewise for similar unamed patterns.
18635 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
18636 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
18637 (hppa_legitimize_address): Simplify mask calculation.
18638 (pa_emit_move_sequence): Revised handling of secondary reloads from
18639 REG+D addresses for floating point loads and stores. Directly handle
18640 loading CONST0_RTX (mode) to a floating point register.
18641 (pa_secondary_reload): Handle reloading DF and SFmode constant values
18642 to floating point registers. Don't restrict secondary reloads to
18643 floating point registers to integer modes. Revise some comments and
18645 (TARGET_LEGITIMATE_ADDRESS_P): Define.
18646 (pa_legitimate_address_p): New.
18647 (pa_legitimize_reload_address): New.
18648 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
18649 (STRICT_REG_OK_FOR_BASE_P): New.
18650 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
18651 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
18653 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
18654 Andrew Dixie <andrewd@gentrack.com>
18656 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
18659 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
18661 * expmed.c (extract_bit_field_1): Pass the full width of the
18662 structure to get_best_reg_extraction_insn.
18664 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
18667 * configure.ac (use_cxa_atexit): Add AIX.
18668 * configure: Regenerate.
18670 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
18672 2013-02-01 Jakub Jelinek <jakub@redhat.com>
18675 * final.c (need_profile_function): New variable.
18676 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
18677 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
18678 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
18679 notes, targetm.asm_out.function_prologue doesn't emit anything,
18680 HAVE_prologue and profiler should be emitted before prologue,
18681 set need_profile_function instead of emitting it.
18682 (final_scan_insn): If need_profile_function, emit
18683 profile_function on the first NOTE_INSN_BASIC_BLOCK or
18684 NOTE_INSN_FUNCTION_BEG note.
18686 2013-02-01 Richard Henderson <rth@redhat.com>
18688 * config/rs6000/rs6000.md (smulditi3): New.
18691 * config/alpha/alpha.md (umulditi3): New.
18693 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
18695 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
18696 (ASM_OUTPUT_ALIGNED_LOCAL): New.
18698 2013-02-01 Richard Biener <rguenther@suse.de>
18700 PR tree-optimization/56113
18701 * tree-ssa-structalias.c (label_visit): Reduce work for
18702 single-predecessor nodes.
18704 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
18706 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
18707 range isn't testing for zero.
18709 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
18711 PR middle-end/56113
18712 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
18714 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
18715 Nick Clifton <nickc@redhat.com>
18717 * config/v850/constraints.md (Q): Define as a memory constraint.
18718 * config/v850/predicates.md (label_ref_operand): New predicate.
18719 (e3v5_shift_operand): New predicate.
18720 (ior_operator): New predicate.
18721 * config/v850/t-v850: Add e3v5 multilib.
18722 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
18723 (v850_gen_movdi): Prototype.
18724 * config/v850/v850.c: Add support for e3v5 architecture.
18725 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
18727 (construct_save_jarl): Add e3v5 long JARL support.
18728 (v850_adjust_insn_length): New function. Adjust length of call
18729 insns when using e3v5 instructions.
18730 (v850_gen_movdi): New function: Generate instructions to move a
18732 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
18733 (CPP_SPEC): Define __v850e3v5__ as appropriate.
18734 (TARGET_USE_FPU): Enable for e3v5.
18735 (CONST_OK_FOR_W): New macro.
18736 (ADJUST_INSN_LENGTH): Define.
18737 * config/v850/v850.md (UNSPEC_LOOP): Define.
18738 (attr cpu): Add v850e3v5.
18739 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
18740 (movdi): New pattern.
18741 (movdi_internal): New pattern.
18742 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
18743 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
18744 (cstoresf4): Likewise.
18745 (cstoredf4): Likewise.
18746 (insv): New pattern.
18747 (rotlso3_a): New pattern.
18748 (rotlsi3_b): New pattern
18749 (rotlsi3_v850e3v5): New pattern.
18750 (doloop_begin): New pattern.
18751 (fix_loop_counter): New pattern.
18752 (doloop_end): New pattern.
18753 (branch_normal): Add e3v5 long branch support.
18754 (branch_invert): Likewise.
18755 (branch_z_normal): Likewise.
18756 (branch_z_invert): Likewise.
18757 (branch_nz_normal): Likewise.
18758 (branch_nz_invert): Likewise.
18759 (call_internal_short): Add e3v5 register-indirect JARL support.
18760 (call_internal_long): Likewise.
18761 (call_value_internal_short): Likewise.
18762 (call_value_internal_long): Likewise.
18763 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
18764 (mloop): New option.
18765 * config.gcc: Add support for configuring v840e3v5 target.
18766 * doc/invoke.texi: Document new v850 specific command line options.
18768 2013-01-31 Paul Koning <ni1d@arrl.net>
18772 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
18773 children if parent is a class.
18774 (prune_unused_types_prune): Don't add DW_AT_declaration.
18776 2013-01-31 Richard Biener <rguenther@suse.de>
18778 PR tree-optimization/56157
18779 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
18780 match up operand with SLP child.
18782 2013-01-31 Jason Merrill <jason@redhat.com>
18785 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
18786 parameters the first time.
18787 (gen_scheduled_generic_parms_dies): Check completeness here.
18789 2013-01-31 Richard Biener <rguenther@suse.de>
18791 PR middle-end/53073
18792 * common.opt (faggressive-loop-optimizations): New flag,
18793 enabled by default.
18794 * doc/invoke.texi (faggressive-loop-optimizations): Document.
18795 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
18796 infer_loop_bounds_from_undefined by it.
18798 2013-01-31 Richard Biener <rguenther@suse.de>
18800 PR tree-optimization/56150
18801 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
18802 visit virtual operands.
18803 (find_uses_to_rename_bb): Likewise.
18805 2013-01-31 Richard Biener <rguenther@suse.de>
18807 PR tree-optimization/56150
18808 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
18809 mixed store non-store stmts.
18811 2013-01-30 Jakub Jelinek <jakub@redhat.com>
18814 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
18815 LIBASAN_EARLY_SPEC is defined.
18816 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
18817 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
18819 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
18822 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
18823 invalid args instead of ICEing on it.
18824 (ix86_valid_target_attribute_tree): Return error_mark_node if
18825 ix86_valid_target_attribute_inner_p failed.
18826 (ix86_valid_target_attribute_p): Return false only if
18827 ix86_valid_target_attribute_tree returned error_mark_node. Allow
18828 target("default") attribute.
18829 (sorted_attr_string): Change argument from const char * to tree,
18830 merge in all target attribute arguments rather than just one.
18831 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
18832 instead of free. Avoid using strcat.
18833 (ix86_mangle_function_version_assembler_name): Mangle
18834 target("default") as if no target attribute is present. Adjust
18835 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
18836 instead of xmalloc and XDELETEVEC instead of free.
18837 (ix86_function_versions): Don't return true if one of the decls
18838 doesn't have target attribute. If they don't and one of the decls
18839 is DECL_FUNCTION_VERSIONED, report an error. Adjust
18840 sorted_attr_string caller. Use XDELETEVEC instead of free.
18841 (ix86_supports_function_versions): Remove.
18842 (make_name): Fix up formatting.
18843 (make_dispatcher_decl): Remove resolver_name and its initialization.
18844 Avoid leaking memory.
18845 (is_function_default_version): Return true if there is
18846 target("default") attribute rather than no target attribute at all.
18847 (make_resolver_func): Avoid leaking memory.
18848 (ix86_generate_version_dispatcher_body): Likewise.
18849 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
18850 * target.def (supports_function_versions): Remove.
18851 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
18852 * doc/tm.texi: Regenerated.
18854 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
18856 PR rtl-optimization/56144
18857 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
18858 for values with side effects.
18860 2013-01-30 Richard Biener <rguenther@suse.de>
18862 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
18863 (sparseset_pop): Likewise.
18864 * cfganal.c (compute_idf): Likewise. Increase work-stack size
18865 to be able to use quick_push in the worker loop.
18867 2013-01-30 Marek Polacek <polacek@redhat.com>
18869 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
18871 2013-01-30 Richard Biener <rguenther@suse.de>
18874 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
18876 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
18878 PR tree-optimization/56064
18879 * fixed-value.c (fixed_from_double_int): New function.
18880 * fixed-value.h (fixed_from_double_int): New prototype.
18881 (const_fixed_from_double_int): New static inline function.
18882 * fold-const.c (native_interpret_fixed): New static function.
18883 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
18884 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
18885 (native_encode_fixed): New static function.
18886 (native_encode_expr) <FIXED_CST>: Use it.
18887 (native_interpret_int): Move double_int worker code to...
18888 * double-int.c (double_int::from_buffer): ...this new static method.
18889 * double-int.h (double_int::from_buffer): Prototype it.
18891 2013-01-30 Richard Biener <rguenther@suse.de>
18893 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
18894 New pointer-map and obstack.
18895 (init_alias_vars): Allocate pointer-map and obstack.
18896 (delete_points_to_sets): Free them.
18897 (find_what_var_points_to): Cache result.
18898 (find_what_p_points_to): Adjust for changed interface of
18899 find_what_var_points_to.
18900 (compute_points_to_sets): Likewise.
18901 (ipa_pta_execute): Likewise.
18903 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18905 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
18906 * configure: Regenerate.
18907 * config.in: Regenerate.
18908 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
18909 #nobits/#progbits if supported.
18911 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
18914 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
18915 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
18916 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
18918 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18920 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
18921 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
18923 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18925 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
18926 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
18928 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18930 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
18932 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
18933 * config/arm/cortex-a7.md: New bypasses using
18934 arm_mac_accumulator_is_result.
18936 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18938 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
18939 (cortex_a7_neon_mla): Likewise.
18940 (cortex_a7_fpfmad): New reservation.
18941 (cortex_a7_fpmacs): Use ffmas and update required units.
18942 (cortex_a7_fpmuld): Update required units and latency.
18943 (cortex_a7_fpmacd): Likewise.
18944 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
18945 (cortex_a7_neon). Likewise.
18946 (bypass) Update participating units.
18948 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18950 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
18951 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
18953 * config/arm/vfp11.md (vfp_farith): Use ffmas.
18954 (vfp_fmul): Use ffmad.
18955 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
18956 (cortex_r4_fmacd): Use ffmad.
18957 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
18958 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
18959 (cortex_a9_fmacd): Use ffmad.
18960 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
18961 (cortex_a8_vfp_macd): Use ffmad.
18962 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
18963 (cortex_a5_fpmacd): Use ffmad.
18964 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
18965 (cortex_a15_vfp_macd): Use ffmad.
18966 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
18968 2013-01-29 Jason Merrill <jason@redhat.com>
18971 * varasm.c (default_assemble_visibility): Don't warn about
18972 visibility on artificial decls.
18974 2013-01-29 Richard Biener <rguenther@suse.de>
18976 PR tree-optimization/56113
18977 * tree-ssa-structalias.c (equiv_class_lookup): Also return
18979 (label_visit): Free duplicate bitmaps and record the leader instead.
18980 (perform_var_substitution): Adjust.
18982 2013-01-29 Richard Biener <rguenther@suse.de>
18984 PR tree-optimization/55270
18985 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
18986 the CFG, schedule loops for fixup.
18988 2013-01-29 Nick Clifton <nickc@redhat.com>
18990 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
18993 2013-01-28 Leif Ekblad <leif@rdos.net>
18995 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
18996 * config/i386/i386.h (TARGET_RDOS): New macro.
18997 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
18998 * config/i386/i386.c (ix86_option_override_internal): For 64bit
18999 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
19000 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
19001 DEFAULT_LARGE_SECTION_THRESHOLD.
19002 * config/i386/i386.md (R14_REG, R15_REG): New constants.
19003 * config/i386/rdos.h: New file.
19004 * config/i386/rdos64.h: New file.
19006 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
19009 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
19012 2013-01-28 Jakub Jelinek <jakub@redhat.com>
19014 PR rtl-optimization/56117
19015 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
19016 call cselib_lookup_from_insn on the MEM before calling
19017 add_insn_mem_dependence.
19019 2013-01-28 Richard Biener <rguenther@suse.de>
19021 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
19022 to a stmt that didn't have one.
19023 (copy_phis_for_bb): Likewise for PHI arguments.
19024 (copy_debug_stmt): Likewise for debug stmts.
19026 2013-01-28 Richard Biener <rguenther@suse.de>
19028 PR tree-optimization/56034
19029 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
19030 (partition_builtin_p): Adjust.
19031 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
19032 it is the last partition.
19033 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
19034 up the vertex for the definition.
19035 (classify_partition): Classify whether a partition is a
19036 PKIND_REDUCTION, thus has uses outside of the loop.
19037 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
19038 Merge all PKIND_REDUCTION partitions into the last partition.
19039 (tree_loop_distribution): Seed partitions from reductions as well.
19041 2013-01-28 Jakub Jelinek <jakub@redhat.com>
19043 PR tree-optimization/56125
19044 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
19045 pow(x,c) into sqrt(x) * powi(x, n/2) or
19046 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
19047 optimizing for size.
19048 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
19049 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
19052 PR tree-optimization/56094
19053 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
19054 to UNKNOWN_LOCATION while gimplifying expr.
19056 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
19059 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
19060 operand 0 in movabs insn template for -masm=intel asm alternative.
19061 (*movabs<mode>_2): Ditto for operand 1.
19063 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
19066 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
19069 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19071 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
19074 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19076 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
19077 Undef to avoid warning.
19079 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19081 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
19082 * configure: Regenerate.
19084 2013-01-25 Jakub Jelinek <jakub@redhat.com>
19086 PR tree-optimization/56098
19087 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
19088 for stmts with volatile ops.
19089 (cond_store_replacement): Don't optimize if assign has volatile ops.
19090 (cond_if_else_store_replacement_1): Don't optimize if either
19091 then_assign or else_assign have volatile ops.
19092 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
19095 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
19097 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
19099 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
19101 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
19102 missing ':' in asm example.
19104 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
19106 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
19107 entries into lane and laneq entries.
19108 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
19109 Remove AdvSIMD scalar modes.
19110 (aarch64_sq<r>dmulh_laneq<mode>): New.
19111 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
19113 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
19114 builtin implementations to relfect changes in RTL in aarch64-simd.md.
19115 * config/aarch64/iterators.md (VCOND): New.
19118 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
19121 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
19122 Add NULL LIBNAME argument to existing definitions.
19123 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
19124 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
19125 * config/avr/avr.c (DEF_BUILTIN): Same.
19126 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
19127 (avr_expand_builtin): Expand to a vanilla call if a libgcc
19128 implementation is available (DECL_ASSEMBLER_NAME is set).
19129 (avr_fold_absfx): New static function.
19130 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
19131 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
19132 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
19133 AVR_BUILTIN_ABSLLK.
19134 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
19135 (abshk, absk, abslk, absllk): Provide as static inline functions.
19137 2013-01-25 Marek Polacek <polacek@redhat.com>
19139 PR tree-optimization/56035
19140 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
19142 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
19144 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
19145 (*movtf_internal_rex64): Add (!o,C) alternative
19146 (*movxf_internal_rex64): Ditto.
19147 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
19149 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
19151 * doc/invoke.texi: fix typo.
19152 * doc/objc.texi: fix typo.
19154 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
19156 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
19157 for the first two alternatives.
19159 2013-01-24 Diego Novillo <dnovillo@google.com>
19161 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
19162 (ggc-zone.o): Remove.
19163 * configure.ac: Remove option --with-gc.
19164 * configure: Re-generate.
19165 * doc/install.texi: Remove documentation for --with-gc.
19166 * gengtype.c (write_enum_defn): Remove. Update all users.
19167 (write_Types_process_field): Remove generation of gt_e_* argument.
19168 (output_type_enum): Remove. Update all users.
19169 (write_enum_defn): Remove. Update all users.
19170 (enum alloc_zone): Remove. Update all users.
19171 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
19172 * ggc-common.c (ggc_splay_alloc): Remove first argument.
19173 Update all callers.
19174 (struct ptr_data): Remove field TYPE. Update all users.
19175 (gt_pch_note_object): Remove argument TYPE. Update all users.
19176 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
19178 * ggc-none.c (ggc_alloc_typed_stat): Remove.
19179 (struct alloc_zone): Remove.
19180 (ggc_internal_alloc_zone_stat): Remove.
19181 (ggc_internal_cleared_alloc_zone_stat): Remove.
19182 * ggc-page.c (ggc_alloc_typed_stat): Remove.
19183 (ggc_pch_count_object): Remove last argument. Update all users.
19184 (ggc_pch_alloc_object): Remove last argument. Update all users.
19185 (struct alloc_zone): Remove.
19186 * ggc-zone.c: Remove.
19187 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
19188 (struct alloc_zone): Remove.
19189 (ggc_alloc_typed_stat): Remove.
19190 (ggc_alloc_typed): Remove.
19191 (ggc_splay_alloc): Remove first argument.
19192 (rtl_zone): Remove. Update all users.
19193 (tree_zone): Remove. Update all users.
19194 (tree_id_zone): Remove. Update all users.
19195 (ggc_internal_zone_alloc_stat): Remove. Update all users.
19196 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
19197 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
19198 * tree-ssanames.c: Remove references to zone allocator in comments.
19200 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
19202 * config/avr/avr.c (avr_out_fract): Make register numbers that
19203 might be outside of source operand signed.
19205 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
19207 * config/i386/constraints.md (Yf): New constraint.
19208 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
19209 of f constraint to conditionaly disable x87 register preferences.
19210 (*movdf_internal): Ditto.
19211 (*movsf_internal): Ditto.
19213 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
19215 PR inline-asm/55934
19216 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
19217 that have operands with impossible constraints.
19218 Add a FIXME for a speed-up opportunity.
19219 * lra-constraints.c (process_alt_operands): Verify that a class
19220 selected from constraints on asms is valid for the operand mode.
19221 (curr_insn_transform): Remove incorrect comment.
19223 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
19225 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
19226 TOC operand is a valid symbol ref in the constant pool.
19228 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19230 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
19232 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
19235 * config/avr/stdfix.h: New file.
19236 * t-avr (stdfix-gcc.h): New rule to build it.
19237 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
19239 2013-01-23 Kostya Serebryany <kcc@google.com>
19241 * config/darwin.h: remove dependency on
19242 CoreFoundation (asan on Mac OS).
19244 2013-01-23 Jakub Jelinek <jakub@redhat.com>
19247 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
19248 instead of cmpdi_operand for first comparison operand.
19249 Don't assert that comparison operands aren't both constants.
19251 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19253 * doc/install.texi (Downloading the Source): Update references to
19254 downloading separate components.
19256 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19258 * doc/extend.texi (__int128): Improve grammar.
19260 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
19263 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
19264 alternative to (o,r).
19265 (*movdi_internal_rex64): Remove (!o,n) alternative.
19266 (DImode immediate->memory splitter): Remove.
19267 (DImode immediate->memory peephole2): Remove.
19268 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
19269 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
19270 alternative to (!o,*r).
19271 (*movtf_internal_sse): New pattern.
19272 (*movxf_internal_rex64): New pattern.
19273 (*movxf_internal): Disable for TARGET_64BIT.
19274 (*movdf_internal_rex64): Remove (!o,F) alternative.
19276 2013-01-22 Jakub Jelinek <jakub@redhat.com>
19278 PR middle-end/56074
19279 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
19280 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
19281 * tree-vect-loop-manip.c (find_loop_location): Also ignore
19282 stmt locations where LOCATION_LOCUS of the stmt location is
19283 UNKNOWN_LOCATION or BUILTINS_LOCATION.
19286 * config/i386/i386.md (UNSPEC_STOS): New.
19287 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
19288 *strsetqi_1): Add UNSPEC_STOS.
19290 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
19293 * doc/invoke.texi: Remove left over -Wsynth example.
19295 2013-01-21 Jakub Jelinek <jakub@redhat.com>
19297 PR tree-optimization/56051
19298 * fold-const.c (fold_binary_loc): Don't fold
19299 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
19300 a narrowing conversion, or widening conversion from signed
19303 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
19305 PR rtl-optimization/56023
19306 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
19307 dependent on debug instruction.
19309 2013-01-21 Martin Jambor <mjambor@suse.cz>
19311 PR middle-end/56022
19312 * function.c (allocate_struct_function): Call
19313 invoke_set_current_function_hook earlier.
19315 2013-01-21 Jakub Jelinek <jakub@redhat.com>
19317 * reload1.c (init_reload): Only initialize reload_obstack
19318 during the first call.
19320 2013-01-21 Marek Polacek <polacek@redhat.com>
19322 * cfgloop.c (verify_loop_structure): Fix up grammar.
19324 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
19326 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
19327 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
19329 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19332 * config/arm/marvell-pj4.md: Update copyright year.
19333 Fix up use of alu to alu_reg and simple_alu_imm.
19335 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
19337 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
19339 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
19342 * lra-constraints.c (curr_insn_transform): Don't reuse original
19343 insn for secondary memory move when memory mode should be different.
19345 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19347 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
19348 atomic_storedi_1): New patterns.
19350 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19352 btver2 pipeline descriptions.
19353 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
19355 * config/i386/i386.md (btver2_decode): New type attributes.
19356 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
19358 * config/i386/btver2.md: New file describing btver2 pipelines.
19360 2013-01-19 Andrew Pinski <apinski@cavium.com>
19362 PR tree-optimization/52631
19363 * tree-ssa-sccvn (visit_use): Before looking up the original
19364 statement, try looking up the simplified expression.
19366 2013-01-19 Anthony Green <green@moxielogic.com>
19368 * config/moxie/moxie.c (moxie_expand_prologue): Set
19369 current_function_static_stack_size.
19371 2013-01-18 Jakub Jelinek <jakub@redhat.com>
19373 PR tree-optimization/56029
19374 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
19375 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
19377 2013-01-18 Sharad Singhai <singhai@google.com>
19379 PR tree-optimization/55995
19380 * dumpfile.c (dump_loc): Print location only if available.
19381 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
19383 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
19386 * lra-constraints.c (curr_insn_transform): Reuse original insn for
19387 secondary memory move.
19388 (inherit_reload_reg): Use rclass instead of cl for
19389 check_secondary_memory_needed_p.
19391 2013-01-18 Jakub Jelinek <jakub@redhat.com>
19393 PR middle-end/56015
19394 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
19395 the case where writing real complex part of target modifies op1.
19397 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
19399 * config/aarch64/aarch64-simd.md
19400 (aarch64_vcond_internal<mode>): Handle unordered cases.
19401 * config/aarch64/iterators.md (v_cmp_result): New.
19403 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
19404 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19406 * config/arm/marvell-pj4.md: New file.
19407 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
19408 * config/arm/arm.md (generic_sched): Add marvell_pj4.
19409 (generic_vfp): Likewise.
19410 * config/arm/arm-cores.def: Add marvell-pj4.
19411 * config/arm/arm-tune.md: Regenerate.
19412 * config/arm/arm-tables.opt: Regenerate.
19413 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
19414 * doc/invoke.texi: Document marvell-pj4.
19416 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
19418 * config/aarch64/arm_neon.h: Map scalar types to standard types.
19420 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
19425 * var-tracking.c (negative_power_of_two_p): New.
19426 (global_get_addr_cache, local_get_addr_cache): New.
19427 (get_addr_from_global_cache, get_addr_from_local_cache): New.
19428 (vt_canonicalize_addr): Rewrite using the above. Adjust the
19430 (vt_stack_offset_p): Remove.
19431 (vt_canon_true_dep): Always canonicalize loc's address.
19432 (clobber_overlapping_mems): Make sure we have a MEM.
19433 (local_get_addr_clear_given_value): New.
19434 (val_reset): Clear local cached entries.
19435 (compute_bb_dataflow): Create and release the local cache.
19436 Disable duplicate MEMs clobbering.
19437 (emit_notes_in_bb): Clobber MEMs likewise.
19438 (vt_emit_notes): Create and release the local cache.
19439 (vt_initialize, vt_finalize): Create and release the global
19440 cache, respectively.
19441 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
19443 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
19445 PR libmudflap/53359
19446 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
19447 not found in the symtab.
19449 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
19452 PR rtl-optimization/55547
19453 PR rtl-optimization/53827
19456 * alias.c (offset_overlap_p): New, factored out of...
19457 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
19458 the conservative special case for symbolic constants. Don't
19459 adjust zero sizes on alignment.
19461 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
19463 PR rtl-optimization/52573
19464 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
19465 REG_UNUSED for the same register.
19467 2013-01-17 Richard Biener <rguenther@suse.de>
19468 Marek Polacek <polacek@redhat.com>
19470 PR rtl-optimization/55833
19471 * loop-unswitch.c (unswitch_loops): Move loop verification...
19472 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
19473 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
19474 Set it to true when we're removing a loop from hierarchy tree in
19475 an irreducible region.
19476 (fix_bb_placements): Adjust caller.
19477 (fix_loop_placements): Likewise.
19479 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
19481 * config/avr/builtins.def (DEF_BUILTIN): Factor out
19482 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
19483 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
19484 Remove ID. Adjust comments.
19485 * config/avr/avr-c.c (avr_builtin_name): Remove.
19486 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
19487 * config/avr/avr.c (avr_tolower): New static function.
19488 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
19489 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
19490 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
19493 2013-01-17 Jan Hubicka <jh@suse.cz>
19495 PR tree-optimization/55273
19496 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
19498 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
19501 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
19502 store through atomic_store<mode>_1.
19503 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
19505 2013-01-17 Martin Jambor <mjambor@suse.cz>
19507 PR tree-optimizations/55264
19508 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
19509 for virtual methods.
19510 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
19511 virtual methods before inlining is over.
19512 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
19514 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
19517 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
19519 PR rtl-optimization/56005
19520 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
19521 pending reads for prefetch.
19523 2013-01-16 Ian Bolton <ian.bolton@arm.com>
19525 * config/aarch64/aarch64.md
19526 (*cstoresi_neg_uxtw): New pattern.
19527 (*cmovsi_insn_uxtw): New pattern.
19528 (*<optab>si3_uxtw): New pattern.
19529 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
19530 (*<optab>si3_insn_uxtw): New pattern.
19531 (*bswapsi2_uxtw): New pattern.
19533 2013-01-16 Richard Biener <rguenther@suse.de>
19535 * tree-inline.c (tree_function_versioning): Remove set but
19536 never used variable.
19538 2013-01-16 Richard Biener <rguenther@suse.de>
19540 PR tree-optimization/55964
19541 * tree-flow.h (rename_variables_in_loop): Remove.
19542 (rename_variables_in_bb): Likewise.
19543 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
19544 (copy_loop_before): Adjust and delete update-ssa status.
19545 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
19546 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
19547 (rename_variables_in_loop): Remove.
19548 (slpeel_update_phis_for_duplicate_loop): Likewise.
19549 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
19550 use available cfg machinery instead of duplicating it.
19551 Update PHI nodes and perform poor-mans SSA update here.
19552 (slpeel_tree_peel_loop_to_edge): Adjust.
19554 2013-01-16 Richard Biener <rguenther@suse.de>
19556 PR tree-optimization/54767
19557 PR tree-optimization/53465
19558 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
19559 (vrp_visit_phi_node): For PHI arguments coming via backedges
19560 drop all symbolical range information.
19561 (execute_vrp): Compute backedges.
19563 2013-01-16 Richard Biener <rguenther@suse.de>
19565 * doc/install.texi: Update CLooG and ISL requirements to
19568 2013-01-16 Christian Bruel <christian.bruel@st.com>
19571 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
19572 (broken_move): Handle UNSPECV_SP_SWITCH_B.
19573 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
19575 2013-01-16 DJ Delorie <dj@redhat.com>
19577 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
19578 (UNSPECV_SP_SWITCH_E): New.
19579 (sp_switch_1): Change to an unspec.
19580 (sp_switch_2): Change to an unspec. Don't use post-inc when we
19583 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
19585 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
19586 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
19587 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
19588 (expand_mem_thread_fence): Ditto.
19589 (expand_mem_signal_fence): Ditto.
19590 (expand_atomic_load): Ditto.
19591 (expand_atomic_store): Ditto.
19593 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
19595 PR rtl-optimization/55547
19596 PR rtl-optimization/53827
19599 * alias.c (memrefs_conflict_p): Set sizes to negative after
19602 2013-01-15 Jakub Jelinek <jakub@redhat.com>
19605 * function.c (thread_prologue_and_epilogue_insns): Always
19606 add crtl->drap_reg to set_up_by_prologue.set, even if
19607 stack_realign_drap is false.
19609 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
19611 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
19612 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
19613 *call): Fix indention.
19615 2013-01-15 Tom de Vries <tom@codesourcery.com>
19618 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
19621 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
19623 PR rtl-optimization/55153
19624 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
19626 2013-01-15 Martin Jambor <mjambor@suse.cz>
19628 PR tree-optimization/55920
19629 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
19630 accesses as grp_to_be_debug_replaced.
19632 2013-01-15 Jakub Jelinek <jakub@redhat.com>
19634 PR tree-optimization/55920
19635 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
19636 there is non-useless type conversion needed from debug rhs to lhs,
19637 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
19639 2013-01-15 Joseph Myers <joseph@codesourcery.com>
19640 Mikael Pettersson <mikpe@it.uu.se>
19643 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
19645 (ASM_OUTPUT_CASE_LABEL): Remove.
19646 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
19647 * final.c (shorten_branches): Update alignment of labels before
19648 jump tables if CASE_VECTOR_SHORTEN_MODE.
19650 2013-01-15 Richard Biener <rguenther@suse.de>
19653 * system.h: Do not include gmp.h for building host tools.
19655 2013-01-15 Richard Biener <rguenther@suse.de>
19657 PR middle-end/55882
19658 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
19659 account for bitpos when computing alignment.
19661 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
19663 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
19664 (ix86_target_macros_internal): Likewise.
19666 * config/i386/i386.c (m_CORE2I7): Removed.
19667 (m_CORE_HASWELL): New macro.
19668 (m_CORE_ALL): Likewise.
19669 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
19670 (initial_ix86_arch_features): Likewise.
19671 (processor_target_table): Initializations for Core avx2.
19672 (cpu_names): New names "core-avx2".
19673 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
19674 PROCESSOR_CORE_HASWELL.
19675 (ix86_issue_rate): New case.
19676 (ia32_multipass_dfa_lookahead): Likewise.
19677 (ix86_sched_init_global): Likewise.
19679 * config/i386/i386.h (TARGET_HASWELL): New macro.
19680 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
19681 (processor_type): New PROCESSOR_HASWELL.
19683 2013-01-15 Jakub Jelinek <jakub@redhat.com>
19685 PR tree-optimization/55955
19686 * tree-vect-loop.c (vectorizable_reduction): Give up early on
19687 *SHIFT_EXPR and *ROTATE_EXPR codes.
19689 PR tree-optimization/48766
19690 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
19691 -ftrapv disable -fwrapv.
19693 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19696 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
19697 etc. to 1 and not to __flash.
19698 Use LL suffix for __INT24_MAX__ with -mint8.
19699 Use ULL suffix for __UINT24_MAX__ with -mint8.
19701 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19703 * config/avr/avr-arch.h
19704 (struct base_arch_s): Use typedef avr_arch_t instead.
19705 (struct arch_info_s): Use typedef avr_arch_info_t instead.
19706 (struct mcu_type_s): Use typedef avr_mcu_t instead.
19707 * config/avr/avr.c: Same.
19708 * config/avr/avr-devices.c: Same.
19709 * config/avr/driver-avr.c: Same.
19710 * config/avr/gen-avr-mmcu-texi.c: Same.
19711 * config/avr/avr-mcus.def: Adjust comment.
19713 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
19715 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
19716 * config/aarch64/iterators.md (VALLDI): New.
19718 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
19719 Andi Kleen <ak@linux.intel.com>
19722 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
19723 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
19726 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19728 * config/avr/avr-stdint.h: Remove trailing blanks.
19729 * config/avr/avr-log.h: Same.
19730 * config/avr/avr-arch.h: Same.
19731 * config/avr/avr-devices.c: Same.
19732 * config/avr/avr-dimode.md: Same.
19733 * config/avr/predicates.md: Same.
19734 * config/avr/avr-c.c: Same. And fix typo.
19736 * config/avr/avr-protos.h: Same. And:
19737 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
19738 (init_cumulative_args): Rename to avr_init_cumulative_args.
19739 (expand_prologue): Rename to avr_expand_prologue.
19740 (expand_epilogue): Rename to avr_expand_epilogue.
19741 (adjust_insn_length): Rename to avr_adjust_insn_length.
19742 (notice_update_cc): Rename to avr_notice_update_cc.
19743 (final_prescan_insn): Rename to avr_final_prescan_insn.
19744 * config/avr/avr.c: Same.
19745 * config/avr/avr.h: Same.
19746 * config/avr/avr.md: Remove trailing blanks.
19747 (prologue): Use avr_expand_prologue.
19748 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
19750 2013-01-14 Richard Biener <rguenther@suse.de>
19752 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
19753 verify_location, collect_subblocks): New functions.
19754 (verify_gimple_in_cfg): Verify that locations only reference
19755 BLOCKs in the functions BLOCK tree.
19757 2013-01-14 Richard Biener <rguenther@suse.de>
19759 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
19761 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
19763 (insert_out_of_ssa_copy_on_edge): Likewise.
19764 (rewrite_close_phi_out_of_ssa): Likewise.
19765 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
19767 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
19768 propagated constants.
19769 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
19772 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
19774 * config/avr/avr-modes.def: Add GPL copyright notice.
19776 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
19778 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
19779 MEMMODEL_MASK to determine memory model.
19780 (atomic_store<mode>): Ditto from operands[2].
19781 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
19783 2013-01-13 Jakub Jelinek <jakub@redhat.com>
19786 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
19787 (fold_gimple_assign): Don't call unshare_expr here.
19788 (fold_ctor_reference): Call unshare_expr.
19790 2013-01-13 Terry Guo <terry.guo@arm.com>
19792 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
19793 * doc/fragments.texi: Document MULTILIB_REUSE.
19794 * gcc.c (multilib_reuse): New internal spec.
19795 (set_multilib_dir): Also search multilib from multilib_reuse.
19796 * genmultilib (tmpmultilib3): Refactor code.
19797 (tmpmultilib4): Ditto.
19798 (multilib_reuse): New multilib argument.
19800 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
19802 * Makefile.in: Update copyright.
19804 2013-01-12 Tom de Vries <tom@codesourcery.com>
19806 PR middle-end/55890
19807 * calls.c (expand_call): Check if arg_nr is valid.
19809 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19811 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
19812 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
19813 documentation. Add missing '__' in front of
19814 __builtin_ia32_packssdw256.
19816 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19819 * config/s390/s390.c (s390_preferred_reload_class): Do not return
19820 NO_REGS for larl operands.
19821 (s390_reload_larl_operand): Use s390_load_address instead of
19824 2013-01-11 Richard Biener <rguenther@suse.de>
19826 * tree-cfg.c (verify_node_sharing_1): Split out from ...
19827 (verify_node_sharing): ... here.
19828 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
19830 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
19832 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
19833 Substitute TREECHECKING.
19834 * configure: Regenerate.
19835 * Makefile.in (TREECHECKING): New.
19837 2013-01-11 Richard Guenther <rguenther@suse.de>
19839 PR tree-optimization/44061
19840 * tree-vrp.c (extract_range_basic): Compute zero as
19841 value-range for __builtin_constant_p of function parameters.
19843 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
19845 Update copyright years.
19847 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
19849 PR rtl-optimization/55672
19850 * lra-eliminations.c (mark_not_eliminable): Permit addition with
19851 const to be eliminable.
19853 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
19855 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
19856 * configure: Regenerate.
19858 2013-01-10 Richard Biener <rguenther@suse.de>
19860 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
19862 2013-01-10 Richard Biener <rguenther@suse.de>
19865 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
19866 locations for virtual PHI arguments.
19867 (rewrite_update_phi_arguments): Likewise.
19869 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
19871 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
19874 2013-01-10 Jakub Jelinek <jakub@redhat.com>
19876 PR tree-optimization/55921
19877 * tree-complex.c (expand_complex_asm): New function.
19878 (expand_complex_operations_1): Call it for GIMPLE_ASM.
19880 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19883 * config/s390/s390.c (s390_symref_operand_p)
19884 (s390_loadrelative_operand_p): Merge the two functions.
19885 (s390_check_qrst_address, print_operand_address): Add parameters
19886 to s390_loadrelative_operand_p invokation.
19887 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
19888 (s390_reload_larl_operand, s390_secondary_reload): Use
19889 s390_loadrelative_operand_p instead of s390_symref_operand_p.
19890 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
19892 2013-01-09 Mike Stump <mikestump@comcast.net>
19894 * dse.c (record_store): Remove unnecessary assert.
19896 2013-01-09 Jan Hubicka <jh@suse.cz>
19898 PR tree-optimization/55569
19899 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
19900 * cfgloop.h (scale_loop_profile): Likewise.
19902 2013-01-09 Jan Hubicka <jh@suse.cz>
19905 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
19907 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
19909 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
19911 PR middle-end/55114
19912 * expr.h (maybe_emit_group_store): Declare.
19913 * expr.c (maybe_emit_group_store): New function.
19914 * builtins.c (expand_builtin_int_roundingfn): Call it.
19915 (expand_builtin_int_roundingfn_2): Likewise.
19917 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
19919 PR rtl-optimization/55829
19920 * lra-constraints.c (match_reload): Add code for absent output.
19921 (curr_insn_transform): Add code for reloads of matched inputs
19924 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
19926 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
19927 attribute of movddup insn to DF.
19928 (*vec_interleave_lowv2df): Ditto.
19929 (vec_dupv2df): Ditto.
19931 2013-01-09 Jan Hubicka <jh@suse.cz>
19933 PR tree-optimiation/55875
19934 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
19935 EVERY_ITERATION parameter.
19936 (number_of_iterations_exit): Check if exit is executed every iteration.
19937 (idx_infer_loop_bounds): Similarly here.
19938 (n_of_executions_at_most): Simplify
19939 to only test for cases where statement is dominated by the
19940 particular bound; handle correctly the "postdominance" test.
19941 (scev_probably_wraps_p): Use max loop iterations info
19942 as a global bound first.
19944 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
19945 Nick Clifton <nickc@redhat.com>
19947 * config/v850/v850.md (cbranchsf4): New pattern.
19948 (cstoresf4): New pattern.
19949 (cbranchdf4): New pattern.
19950 (cstoredf4): New pattern.
19951 (movsicc): Disallow floating point comparisons.
19952 (cmpsf_le_insn): Fix order of operators.
19953 (cmpsf_lt_insn): Likewise.
19954 (cmpsf_eq_insn): Likewise.
19955 (cmpdf_le_insn): Likewise.
19956 (cmpdf_lt_insn): Likewise.
19957 (cmpdf_eq_insn): Likewise.
19958 (cmpsf_ge_insn): Use LE comparison.
19959 (cmpdf_ge_insn): Likewise.
19960 (cmpsf_gt_insn): Use LT comparison.
19961 (cmpdf_gt_insn): Likewise.
19962 (cmpsf_ne_insn): Delete pattern.
19963 (cmpdf_ne_insn): Delete pattern.
19964 * config/v850/v850.c (v850_gen_float_compare): Use
19965 gen_cmpdf_eq_insn for NE comparison.
19966 (v850_float_z_comparison_operator)
19967 (v850_float_nz_comparison_operator): Move from here ...
19968 * config/v850/predicates.md: ... to here. Move GT and GE
19969 comparisons into v850_float_z_comparison_operator.
19970 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
19972 (v850_float_nz_comparison_operator): Likewise.
19974 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19976 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
19977 with calls to gen_insvsi/gen_insvdi.
19979 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19981 * config/i386/i386.c (initial_ix86_tune_features): Set up
19982 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
19984 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
19985 Jakub Jelinek <jakub@redhat.com>
19987 PR tree-optimization/48189
19988 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
19989 If nitercst is 0, don't predict the exit edge.
19991 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19993 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
19994 in asm_fprintf with reg_names.
19995 (aarch64_print_operand_address): Likewise.
19996 (aarch64_return_addr): Likewise.
19997 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
19999 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20001 * config/pa/pa.h (VAL_U6_BITS_P): Define.
20002 (INT_U6_BITS): Likewise.
20003 * config/pa/predicates.md (uint6_operand): New predicate.
20004 (shift5_operand, shift6_operand): Likewise.
20005 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
20007 (lshrdi3): Use shift6_operand.
20008 (shrpsi4, shrpdi4): New insn patterns.
20009 (extzv): Delete expander.
20010 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
20011 predicates in unamed zero extract patterns. Tighten common constraint.
20012 (extv): Delete expander.
20013 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
20014 predicates in unamed sign extract patterns. Tighten common constraint.
20015 (insv): Delete expander.
20016 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
20017 predicates in unamed insert patterns. Tighten common constraint.
20018 Change uint32_operand predicate to uint6_operand predicate in unamed
20019 DImode pattern to insert constant values of type 1...1xxxx.
20021 2013-01-04 Jan Hubicka <jh@suse.cz>
20023 PR tree-optimization/55823
20024 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
20027 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20028 Uros Bizjak <ubizjak@gmail.com>
20030 PR rtl-optimization/55845
20031 * df-problems.c (can_move_insns_across): Stop scanning at
20032 volatile_insn_p source instruction or give up if
20033 across_from .. across_to range contains any volatile_insn_p
20036 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20038 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
20039 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
20041 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
20042 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
20044 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20047 * asan.c (asan_clear_shadow): New function.
20048 (asan_emit_stack_protection): Use it.
20050 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20052 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
20053 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
20054 with tab instead of space.
20056 2013-01-08 Nick Clifton <nickc@redhat.com>
20058 * config/rl78/rl78.c (rl78_expand_prologue): Always select
20059 register bank 0 at the start of an interrupt handler.
20060 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
20063 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20065 * config/aarch64/aarch64-simd.md
20066 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
20067 (aarch64_simd_bsl): Likewise.
20068 (aarch64_vcond_internal<mode>): Likewise.
20069 (vcond<mode><mode>): Likewise.
20070 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
20071 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
20073 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20075 * config/aarch64/aarch64-builtins.c
20076 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
20078 2013-01-08 Martin Jambor <mjambor@suse.cz>
20081 * tree-sra.c (analyze_access_subtree): Return true also after
20082 potentially creating a debug-only replacement.
20084 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20086 PR middle-end/55890
20087 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
20089 PR tree-optimization/54120
20090 * tree-vrp.c (range_fits_type_p): Don't allow
20091 src_precision < precision from signed vr to unsigned_p
20092 if vr->min or vr->max is negative.
20093 (simplify_float_conversion_using_ranges): Test can_float_p
20094 against CODE_FOR_nothing.
20096 2013-01-08 Jakub Jelinek <jakub@redhat.com>
20097 Richard Biener <rguenther@suse.de>
20099 PR middle-end/55851
20100 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
20101 types instead of just INTEGER_TYPE types.
20103 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
20105 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
20108 2013-01-07 Steve Ellcey <sellcey@mips.com>
20111 * config/mips/mips.opt: Change mad to mmad to match documentation.
20113 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20116 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
20117 .progmemx.data now.
20119 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20122 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
20123 (avr_addrspace_t): Add .section_name field.
20124 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
20126 (avr_addrspace): Same. Initialize .section_name. Remove last
20127 NULL entry. Put __memx into .progmemx.data.
20128 (progmem_section_prefix): Remove.
20129 (avr_asm_init_sections): No need to initialize progmem_section.
20130 (avr_asm_named_section): Use avr_addrspace[].section_name to get
20131 section name prefix.
20132 (avr_asm_select_section): Ditto. And use get_unnamed_section to
20133 retrieve the progmem section.
20134 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
20135 boundary to run over avr_addrspace[].
20136 (avr_register_target_pragmas): Ditto.
20138 2013-01-06 Jakub Jelinek <jakub@redhat.com>
20140 * varasm.c (output_constant_def_contents): For asan_protect_global
20141 protected strings, adjust DECL_ALIGN if needed, before testing for
20143 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
20144 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
20146 (output_object_block): For asan protected decls, emit asan padding
20147 after their contents.
20148 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
20149 (asan_finish_file): Test it here instead.
20151 2013-01-07 Nick Clifton <nickc@redhat.com>
20152 Matthias Klose <doko@debian.org>
20153 Doug Kwan <dougkwan@google.com>
20154 H.J. Lu <hongjiu.lu@intel.com>
20157 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
20159 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
20161 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
20163 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
20165 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
20167 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20170 * doc/install.texi (Cross-Compiler-Specific Options): Document
20173 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
20175 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
20176 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
20177 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
20178 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
20179 vqmovun_high_s64): Fix source operand number and update copyright.
20181 2013-01-07 Richard Biener <rguenther@suse.de>
20183 PR middle-end/55890
20184 * gimple.h (gimple_call_builtin_p): New overload.
20185 * gimple.c (validate_call): New function.
20186 (gimple_call_builtin_p): Likewise.
20187 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20188 Use gimple_call_builtin_p.
20189 (find_func_clobbers): Likewise.
20190 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
20191 (strlen_optimize_stmt): Likewise.
20193 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20195 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
20196 (vld1q_dup_*): Likewise.
20197 (vld1_*): Likewise.
20198 (vld1q_*): Likewise.
20199 (vld1_lane_*): Likewise.
20200 (vld1q_lane_*): Likewise.
20202 2013-01-07 Richard Biener <rguenther@suse.de>
20204 * lto-streamer.h (LTO_minor_version): Bump to 2.
20206 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20208 * config/aarch64/aarch64-protos.h
20209 (aarch64_const_double_zero_rtx_p): Rename to...
20210 (aarch64_float_const_zero_rtx_p): ...this.
20211 (aarch64_float_const_representable_p): New.
20212 (aarch64_output_simd_mov_immediate): Likewise.
20213 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
20214 move immediate case.
20215 * config/aarch64/aarch64.c
20216 (aarch64_const_double_zero_rtx_p): Rename to...
20217 (aarch64_float_const_zero_rtx_p): ...this.
20218 (aarch64_print_operand): Allow printing of new constants.
20219 (aarch64_valid_floating_const): New.
20220 (aarch64_legitimate_constant_p): Check for valid floating-point
20222 (aarch64_simd_valid_immediate): Likewise.
20223 (aarch64_vect_float_const_representable_p): New.
20224 (aarch64_float_const_representable_p): Likewise.
20225 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
20226 (aarch64_output_simd_mov_immediate): New.
20227 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
20228 (*movdf_aarch64): Likewise.
20229 * config/aarch64/constraints.md (Ufc): New.
20230 (Y): call aarch64_float_const_zero_rtx.
20231 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
20233 2013-01-07 Richard Biener <rguenther@suse.de>
20235 PR tree-optimization/55888
20236 PR tree-optimization/55862
20237 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
20238 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
20239 not if it is contained therein.
20241 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20243 * config/avr/t-avr: Typo.
20245 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
20248 * config/avr/t-avr: Don't automatically rebuild
20249 $(srcdir)/config/avr/t-multilib
20250 $(srcdir)/config/avr/avr-tables.opt
20251 $(srcdir)/doc/avr-mmcu.texi
20252 (avr-mcus): New phony target to build them on request.
20253 (s-avr-mlib, s-avr-mmcu-texi): Remove.
20254 * avr/avr-mcus.def: Adjust comments.
20256 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
20258 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
20260 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20262 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
20264 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20266 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
20268 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
20270 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
20271 to generate profiling.
20272 * config/rs6000/aix64.h (LIB_SPEC): Same.
20274 2013-01-04 Andrew Pinski <apinski@cavium.com>
20276 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
20278 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
20280 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
20282 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
20284 (ix86_expand_move): Ditto.
20285 (ix86_zero_extend_to_Pmode): Ditto.
20286 (ix86_expand_call): Ditto.
20287 (ix86_expand_special_args_builtin): Ditto.
20288 (ix86_expand_builtin): Ditto.
20290 2013-01-04 Richard Biener <rguenther@suse.de>
20292 PR tree-optimization/55862
20293 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
20294 translating them through PHI nodes.
20296 2013-01-04 Martin Jambor <mjambor@suse.cz>
20298 PR tree-optimization/55755
20299 * tree-sra.c (sra_modify_assign): Do not check that an access has no
20300 children when trying to avoid producing a VIEW_CONVERT_EXPR.
20302 2013-01-04 Marek Polacek <polacek@redhat.com>
20304 PR middle-end/55859
20305 * opts.c (default_options_optimization): Clarify error message.
20307 2013-01-04 Richard Biener <rguenther@suse.de>
20309 PR middle-end/55863
20310 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
20313 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20316 * config/pa/pa.md (movsi): Revert previous change.
20317 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
20320 2013-01-03 Richard Henderson <rth@redhat.com>
20322 * config/i386/i386.c (ix86_expand_move): Always assign to op1
20323 after eliminating TLS symbols.
20325 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
20328 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
20329 * graphite-poly.c (debug_gmp_value): Likewise.
20331 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
20334 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
20335 selected code model, define __code_mode_small__, __code_model_medium__,
20336 __code_model_large__, __code_model_32__ or __code_model_kernel__.
20337 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
20338 xchg temporary register with %k. Declare temporary register as
20340 [__x86_64__]: For medium and large code models, preserve %rbx register.
20342 2013-01-03 Richard Biener <rguenther@suse.de>
20344 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
20345 (dump_subscript): Adjust.
20346 (finalize_ddr_dependent): Do not dump redundant info.
20347 (analyze_siv_subscript): Adjust.
20348 (subscript_dependence_tester): Likewise.
20349 (compute_affine_dependence): Likewise.
20351 2013-01-03 Richard Biener <rguenther@suse.de>
20354 2013-01-03 Richard Biener <rguenther@suse.de>
20356 PR tree-optimization/55857
20357 * tree-vect-stmts.c (vectorizable_load): Do not setup
20358 re-alignment for invariant loads.
20360 2013-01-02 Richard Biener <rguenther@suse.de>
20362 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
20363 invariant load do not generate a vector load from the scalar location.
20365 2013-01-03 Richard Biener <rguenther@suse.de>
20367 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
20368 for not vectorizing.
20369 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
20370 not build INDIRECT_REFs, call get_name once only.
20371 (vect_create_data_ref_ptr): Likewise. Dump base object kind
20372 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
20374 2013-01-03 Richard Biener <rguenther@suse.de>
20376 PR tree-optimization/55857
20377 * tree-vect-stmts.c (vectorizable_load): Do not setup
20378 re-alignment for invariant loads.
20380 2013-01-03 Richard Biener <rguenther@suse.de>
20383 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
20384 prefer a built-in decl.
20386 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20388 * gcc.c (process_command): Update copyright notice dates.
20389 * gcov.c (print_version): Likewise.
20390 * gcov-dump.c (print_version): Likewise.
20392 PR rtl-optimization/55838
20393 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
20394 iv0.step, iv1.step and step.
20396 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20397 Marc Glisse <marc.glisse@inria.fr>
20399 PR tree-optimization/55832
20400 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
20401 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
20402 integer_{one,zero}_node.
20404 2013-01-03 Jakub Jelinek <jakub@redhat.com>
20407 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
20408 * var-tracking.c (reverse_op): Don't add reverse ops to
20409 VALUEs that have already
20410 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
20412 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
20414 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
20416 2013-01-02 Teresa Johnson <tejohnson@google.com>
20418 * dumpfile.c (dump_loc): Print filename with location.
20419 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
20420 new location_t parameter to emit complete unroll message with
20421 new dump framework.
20422 (canonicalize_loop_induction_variables): Compute loops location
20423 and pass to try_unroll_loop_completely.
20424 * loop-unroll.c (report_unroll_peel): New function.
20425 (peel_loops_completely): Use new dump format with location
20426 for main dumpfile message, and invoke report_unroll_peel on success.
20427 (decide_unrolling_and_peeling): Ditto.
20428 (decide_peel_once_rolling): Remove old dumpfile message subsumed
20429 by report_unroll_peel.
20430 (decide_peel_completely): Ditto.
20431 (decide_unroll_constant_iterations): Ditto.
20432 (decide_unroll_runtime_iterations): Ditto.
20433 (decide_peel_simple): Ditto.
20434 (decide_unroll_stupid): Ditto.
20435 * cfgloop.c (get_loop_location): New function.
20436 * cfgloop.h (get_loop_location): Declare.
20438 2013-01-02 Sriraman Tallam <tmsriram@google.com>
20440 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
20443 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20445 PR middle-end/55198
20446 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
20447 BLKmode objects when EXPAND_MEMORY is specified.
20449 2013-01-02 Sriraman Tallam <tmsriram@google.com>
20451 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
20453 (fold_builtin_cpu): Do not share cpu model decls across statements.
20455 2013-01-02 Jason Merrill <jason@redhat.com>
20458 * tree.c (build_array_type_1): Revert earlier change.
20460 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
20462 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
20464 * config/aarch64/aarch64-tune.md: Re-generate.
20466 2013-01-02 Richard Biener <rguenther@suse.de>
20468 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
20469 invariant load do not generate a vector load from the scalar location.
20471 2013-01-02 Richard Biener <rguenther@suse.de>
20474 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
20475 * configure: Regenerate.
20477 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
20479 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
20480 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
20481 (expand_builtin_int_roundingfn_2): Keep the original target around
20482 for the fallback case.
20484 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
20486 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
20487 to be clear for sign changes.
20489 2013-01-01 Jan Hubicka <jh@suse.cz>
20491 * ipa-inline-analysis.c: Fix formatting.
20493 2013-01-01 Jakub Jelinek <jakub@redhat.com>
20495 PR tree-optimization/55831
20496 * tree-vect-loop.c (get_initial_def_for_induction): Use
20497 gsi_after_labels instead of gsi_start_bb.
20499 Copyright (C) 2013 Free Software Foundation, Inc.
20501 Copying and distribution of this file, with or without modification,
20502 are permitted in any medium without royalty provided the copyright
20503 notice and this notice are preserved.