]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
[Patch AArch64] Fix register constraints for lane intrinsics.
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
1c83b673
JG
12013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2
3 * config/aarch64/aarch64-simd.md
4 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
5 <vwx> iterator to ensure correct register choice.
6 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7 (aarch64_sqdmull_n<mode>): Likewise.
8 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9 * config/aarch64/arm_neon.h
10 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
11 (vml<as><q>_n_<su>16): Likewise.
12 (vml<as>l_high_lane<q>_<su>16): Likewise.
13 (vml<as>l_high_n_<su>16): Likewise.
14 (vml<as>l_lane<q>_<su>16): Likewise.
15 (vml<as>l_n_<su>16): Likewise.
16 (vmul<q>_lane<q>_<su>16): Likewise.
17 (vmul<q>_n_<su>16): Likewise.
18 (vmull_lane<q>_<su>16): Likewise.
19 (vmull_n_<su>16): Likewise.
20 (vmull_high_lane<q>_<su>16): Likewise.
21 (vmull_high_n_<su>16): Likewise.
22 (vqrdmulh<q>_n_s16): Likewise.
23
bb1ae543
TB
242013-09-06 Tejas Belagod <tejas.belagod@arm.com>
25
26 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
27 have the correct lane parameter.
28
f23c0742
RB
292013-09-06 Richard Biener <rguenther@suse.de>
30
31 * cfganal.c (control_dependences::~control_dependences):
32 Properly free all of the vector.
33
c1d49770
KY
342013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
35
36 PR target/58269
37 * config/i386/i386.c (ix86_conditional_register_usage):
38 Proper initialize extended SSE registers.
39
6da47f52
JH
402013-09-06 Jan Hubicka <jh@suse.cz>
41
42 PR tree-optimization/58311
43 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
44
452013-09-06 Jan Hubicka <jh@suse.cz>
46
47 * Makefile.in (tree-sra.o): Update dependencies.
48 * tree-sra.c: Include ipa-utils.h
49 (scan_function): Use recursive_call_p.
50 (has_caller_p): New function.
51 (cgraph_for_node_and_aliases): Count also callers of aliases.
52
fc11f321
JH
532013-09-06 Jan Hubicka <jh@suse.cz>
54
55 PR middle-end/58094
56 * cgraph.h (symtab_semantically_equivalent_p): Declare.
57 * tree-tailcall.c: Include ipa-utils.h.
58 (find_tail_calls): Use it.
59 * ipa-pure-const.c (check_call): Likewise.
60 * ipa-utils.c (recursive_call_p): New function.
61 * ipa-utils.h (recursive_call_p): Dclare.
62 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
63 (symtab_semantically_equivalent_p): New function.
64 * Makefile.in (tree-tailcall.o): Update dependencies.
65
9a6af450
EB
662013-09-06 Eric Botcazou <ebotcazou@adacore.com>
67
68 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
69 non-inlinable part.
70
35f32ad4
RB
712013-09-06 Richard Biener <rguenther@suse.de>
72
73 * lto-streamer.h (lto_global_var_decls): Remove.
74 * Makefile.in (OBJS): Remove lto-symtab.o.
75 (lto-symtab.o): Remove.
76 (GTFILES): Remove lto-symtab.c
77 * lto-symtab.c: Move to lto/
78
d12a76f3
AK
792013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
80
81 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
82 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
83 (UNSPEC_FPINT_RINT): New constant definitions.
84 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
85 definition with 2 attributes.
86 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
87 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
88 definitions.
89
902013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
91
92 * config/s390/s390.md: Add "bcr_flush" value to mnemonic
93 attribute.
94 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
95 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
96 attribute to "z196_alone".
97 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
98 "zEC12_simple".
99
c8e9d8c3
RB
1002013-09-06 Richard Biener <rguenther@suse.de>
101
102 * basic-block.h (class control_dependences): New.
103 * tree-ssa-dce.c (control_dependence_map): Remove.
104 (cd): New global.
105 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
106 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
107 find_pdom, find_control_dependence, find_all_control_dependences):
108 Move to cfganal.c.
109 (mark_control_dependent_edges_necessary, find_obviously_necessary_stmts,
110 propagate_necessity, tree_dce_init, tree_dce_done,
111 perform_tree_ssa_dce): Adjust.
112 * cfganal.c (set_control_dependence_map_bit,
113 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
114 find_all_control_dependences): Move from tree-ssa-dce.c and
115 implement as methods of control_dependences class.
116 (control_dependences::control_dependences): New.
117 (control_dependences::~control_dependences): Likewise.
118 (control_dependences::get_edges_dependent_on): Likewise.
119 (control_dependences::get_edge): Likewise.
120
7d2ba471
JH
1212013-09-04 Jan Hubicka <jh@suse.cz>
122
123 * tree.c (types_same_for_odr): Drop overactive check.
124 * ipa-devirt.c (hash_type_name): Likewise.
125
e18412fc
JH
1262013-09-04 Jan Hubicka <jh@suse.cz>
127
128 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
129 (analyze_functions): ... here.
130
d352b245
JH
1312013-09-04 Jan Hubicka <jh@suse.cz>
132
133 PR middle-end/58201
134 * cgraphunit.c (analyze_functions): Clear AUX fields
135 after processing; initialize assembler name has.
136
770da076
JL
1372013-09-05 Jeff Law <law@redhat.com>
138
139 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
140 from thread_around_empty_block. Record threading path into PATH.
141 Recurse if threading through the initial block is successful.
142 (thread_across_edge): Corresponding changes to slightly simplify.
143
fe6f68e2
JG
1442013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
145
146 * config/aarch64/aarch64.md
147 (type): Remove frecpe, frecps, frecpx.
148 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
149 fix to be a TARGET_SIMD instruction.
150 (aarch64_frecps): Remove.
151 * config/aarch64/aarch64-simd.md
152 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
153 (aarch64_frecps<mode>): Handle all float/vector of float modes.
154
6e4150e1
JG
1552013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
156 Sofiane Naci <sofiane.naci@arm.com>
157
158 * config/arm/types.md (define_attr "type"):
159 Expand "arlo_imm"
160 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
161 Expand "arlo_reg"
162 into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm", "alu_ext",
163 "alu_reg", "alus_ext", "alus_reg", "bfm", "csel", "logic_reg",
164 "logics_reg", "rev".
165 Expand "arlo_shift"
166 into "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
167 "logics_shift_imm".
168 Expand "arlo_shift_reg"
169 into "alu_shift_reg", "alus_shift_reg", "logic_shift_reg",
170 "logics_shift_reg".
171 Expand "clz" into "clz, "rbit".
172 Rename "shift" to "shift_imm".
173 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
174 changes.
175 Update for attribute changes all occurrences of arlo_* and
176 shift* types.
177 * config/arm/arm-fixed.md: Update for attribute changes
178 all occurrences of arlo_* types.
179 * config/arm/thumb2.md: Update for attribute changes all occurrences
180 of arlo_* types.
181 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
182 (cortexa7_older_only): Likewise.
183 (cortexa7_younger): Likewise.
184 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
185 (1020alu_shift_op): Likewise.
186 (1020alu_shift_reg_op): Likewise.
187 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
188 (alu_shift_op): Likewise.
189 (alu_shift_reg_op): Likewise.
190 * config/arm/arm1136jfs.md (11_alu_op): Update for
191 attribute changes.
192 (11_alu_shift_op): Likewise.
193 (11_alu_shift_reg_op): Likewise.
194 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
195 (9_alu_shift_reg_op): Likewise.
196 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
197 attribute changes.
198 (cortex_a15_alu_shift): Likewise.
199 (cortex_a15_alu_shift_reg): Likewise.
200 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
201 attribute changes.
202 (cortex_a5_alu_shift): Likewise.
203 * config/arm/cortex-a53.md
204 (cortex_a53_alu): Update for attribute changes.
205 (cortex_a53_alu_shift): Likewise.
206 * config/arm/cortex-a7.md
207 (cortex_a7_alu_imm): Update for attribute changes.
208 (cortex_a7_alu_reg): Likewise.
209 (cortex_a7_alu_shift): Likewise.
210 * config/arm/cortex-a8.md
211 (cortex_a8_alu): Update for attribute changes.
212 (cortex_a8_alu_shift): Likewise.
213 (cortex_a8_alu_shift_reg): Likewise.
214 * config/arm/cortex-a9.md
215 (cortex_a9_dp): Update for attribute changes.
216 (cortex_a9_dp_shift): Likewise.
217 * config/arm/cortex-m4.md
218 (cortex_m4_alu): Update for attribute changes.
219 * config/arm/cortex-r4.md
220 (cortex_r4_alu): Update for attribute changes.
221 (cortex_r4_mov): Likewise.
222 (cortex_r4_alu_shift_reg): Likewise.
223 * config/arm/fa526.md
224 (526_alu_op): Update for attribute changes.
225 (526_alu_shift_op): Likewise.
226 * config/arm/fa606te.md
227 (606te_alu_op): Update for attribute changes.
228 * config/arm/fa626te.md
229 (626te_alu_op): Update for attribute changes.
230 (626te_alu_shift_op): Likewise.
231 * config/arm/fa726te.md
232 (726te_alu_op): Update for attribute changes.
233 (726te_alu_shift_op): Likewise.
234 (726te_alu_shift_reg_op): Likewise.
235 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
236 (mp626_alu_shift_op): Likewise.
237 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
238 (pj4_alu_conds): Likewise.
239 (pj4_shift): Likewise.
240 (pj4_shift_conds): Likewise.
241 (pj4_alu_shift): Likewise.
242 (pj4_alu_shift_conds): Likewise.
243 * config/aarch64/aarch64.md: Update for attribute change
244 all occurrences of arlo_* and shift* types.
245
e6523306
MS
2462013-09-05 Mike Stump <mikestump@comcast.net>
247
248 * tree.h: Move documentation for tree_function_decl to tree-core.h
249 with the declaration.
250
80a18161
PB
2512013-09-05 Peter Bergner <bergner@vnet.ibm.com>
252
253 PR target/58139
254 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
255 looking for widest mode.
256
6994430a
EB
2572013-09-05 Eric Botcazou <ebotcazou@adacore.com>
258
259 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
260
e67f39f7
RB
2612013-09-05 Richard Biener <rguenther@suse.de>
262
263 PR tree-optimization/58137
264 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
265 Do not create vectors of pointers.
266 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
267 types for the components of the vector initializer.
268 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
269 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
270
568cda29
MJ
2712013-09-05 Martin Jambor <mjambor@suse.cz>
272
273 * ipa-prop.c (remove_described_reference): Accept missing references,
274 return false if that hppens, otherwise return true.
275 (cgraph_node_for_jfunc): New function.
276 (try_decrement_rdesc_refcount): Likewise.
277 (try_make_edge_direct_simple_call): Use them.
278 (ipa_edge_removal_hook): Remove references from rdescs.
279 (ipa_edge_duplication_hook): Clone rdescs and their references
280 when the new edge has the same caller as the old one.
281 * cgraph.c (cgraph_resolve_speculation): Remove speculative
282 reference before removing any edges.
283
5a200acb
RE
2842013-09-05 Richard Earnshaw <rearnsha@arm.com>
285
286 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
287 initial store.
288 * thumb2.md (thumb2_storewb_parisi): New pattern.
289
5922847b
YZ
2902013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
291
292 * config/aarch64/aarch64-option-extensions.def: Add
293 AARCH64_OPT_EXTENSION of 'crc'.
294 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
295 (AARCH64_ISA_CRC): Ditto.
296 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
297 description of the CRC extension.
298
996746aa
AI
2992013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
300
301 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
302 * config/rs6000/linux.h: Ditto.
303 * alpha/linux.h: Ditto.
304 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
305 no_c99_libc_has_function.
306 * config/c6x/uclinux-elf.h: Ditto.
307 * config/lm32/uclinux-elf.h: Ditto.
308 * config/m68k/uclinux.h: Ditto.
309 * config/moxie/uclinux.h: Ditto.
310 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
311 (crisv32-*-linux*, cris-*-linux*): Ditto.
312 * config/bfin/bfin.c: Include "tm_p.h".
313
c370d99f
RB
3142013-09-05 Richard Biener <rguenther@suse.de>
315
316 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
317 check for a definition without a basic-block.
318
003bb7f3
JG
3192013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
320 Sofiane Naci <sofiane.naci@arm.com>
321
322 * config/aarch64/aarch64.md
323 (*movti_aarch64): Rename r_2_f and f_2_r.
324 (*movsf_aarch64): Likewise.
325 (*movdf_aarch64): Likewise.
326 (*movtf_aarch64): Likewise.
327 (aarch64_movdi_<mode>low): Likewise.
328 (aarch64_movdi_<mode>high): Likewise.
329 (aarch64_mov<mode>high_di): Likewise.
330 (aarch64_mov<mode>low_di): Likewise.
331 (aarch64_movtilow_tilow): Likewise.
332 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
333 values to config/arm/types.md
334 (attribute "conds"): Update for attribute change.
335 (anddi3_insn): Likewise.
336 (iordi3_insn): Likewise.
337 (xordi3_insn): Likewise.
338 (one_cmpldi2): Likewise.
339 * config/arm/types.md (type): Add Neon types.
340 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
341 use "type" attribute.
342 (movmisalign<mode>_neon_store): Likewise.
343 (movmisalign<mode>_neon_load): Likewise.
344 (vec_set<mode>_internal): Likewise.
345 (vec_setv2di_internal): Likewise.
346 (vec_extract<mode>): Likewise.
347 (vec_extractv2di): Likewise.
348 (add<mode>3_neon): Likewise.
349 (adddi3_neon): Likewise.
350 (sub<mode>3_neon): Likewise.
351 (subdi3_neon): Likewise.
352 (mul<mode>3_neon): Likewise.
353 (mul<mode>3add<mode>_neon): Likewise.
354 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
355 (fma<VCVTF:mode>4)): Likewise.
356 (fma<VCVTF:mode>4_intrinsic): Likewise.
357 (fmsub<VCVTF:mode>4)): Likewise.
358 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
359 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
360 (ior<mode>3): Likewise.
361 (and<mode>3): Likewise.
362 (anddi3_neon): Likewise.
363 (orn<mode>3_neon): Likewise.
364 (orndi3_neon): Likewise.
365 (bic<mode>3_neon): Likewise.
366 (bicdi3_neon): Likewise.
367 (xor<mode>3): Likewise.
368 (one_cmpl<mode>2): Likewise.
369 (abs<mode>2): Likewise.
370 (neg<mode>2): Likewise.
371 (umin<mode>3_neon): Likewise.
372 (umax<mode>3_neon): Likewise.
373 (smin<mode>3_neon): Likewise.
374 (smax<mode>3_neon): Likewise.
375 (vashl<mode>3): Likewise.
376 (vashr<mode>3_imm): Likewise.
377 (vlshr<mode>3_imm): Likewise.
378 (ashl<mode>3_signed): Likewise.
379 (ashl<mode>3_unsigned): Likewise.
380 (neon_load_count): Likewise.
381 (ashldi3_neon_noclobber): Likewise.
382 (signed_shift_di3_neon): Likewise.
383 (unsigned_shift_di3_neon): Likewise.
384 (ashrdi3_neon_imm_noclobber): Likewise.
385 (lshrdi3_neon_imm_noclobber): Likewise.
386 (widen_ssum<mode>3): Likewise.
387 (widen_usum<mode>3): Likewise.
388 (quad_halves_<code>v4si): Likewise.
389 (quad_halves_<code>v4sf): Likewise.
390 (quad_halves_<code>v8hi): Likewise.
391 (quad_halves_<code>v16qi): Likewise.
392 (reduc_splus_v2di): Likewise.
393 (neon_vpadd_internal<mode>): Likewise.
394 (neon_vpsmin<mode>): Likewise.
395 (neon_vpsmax<mode>): Likewise.
396 (neon_vpumin<mode>): Likewise.
397 (neon_vpumax<mode>): Likewise.
398 (ss_add<mode>_neon): Likewise.
399 (us_add<mode>_neon): Likewise.
400 (ss_sub<mode>_neon): Likewise.
401 (us_sub<mode>_neon): Likewise.
402 (neon_vadd<mode>_unspec): Likewise.
403 (neon_vaddl<mode>): Likewise.
404 (neon_vaddw<mode>): Likewise.
405 (neon_vhadd<mode>): Likewise.
406 (neon_vqadd<mode>): Likewise.
407 (neon_vaddhn<mode>): Likewise.
408 (neon_vmul<mode>): Likewise.
409 (neon_vmla<mode>): Likewise.
410 (neon_vmlal<mode>): Likewise.
411 (neon_vmls<mode>): Likewise.
412 (neon_vmlsl<mode>): Likewise.
413 (neon_vqdmulh<mode>): Likewise.
414 (neon_vqdmlal<mode>): Likewise.
415 (neon_vqdmlsl<mode>): Likewise.
416 (neon_vmull<mode>): Likewise.
417 (neon_vqdmull<mode>): Likewise.
418 (neon_vsub<mode>_unspec): Likewise.
419 (neon_vsubl<mode>): Likewise.
420 (neon_vsubw<mode>): Likewise.
421 (neon_vqsub<mode>): Likewise.
422 (neon_vhsub<mode>): Likewise.
423 (neon_vsubhn<mode>): Likewise.
424 (neon_vceq<mode>): Likewise.
425 (neon_vcge<mode>): Likewise.
426 (neon_vcgeu<mode>): Likewise.
427 (neon_vcgt<mode>): Likewise.
428 (neon_vcgtu<mode>): Likewise.
429 (neon_vcle<mode>): Likewise.
430 (neon_vclt<mode>): Likewise.
431 (neon_vcage<mode>): Likewise.
432 (neon_vcagt<mode>): Likewise.
433 (neon_vtst<mode>): Likewise.
434 (neon_vabd<mode>): Likewise.
435 (neon_vabdl<mode>): Likewise.
436 (neon_vaba<mode>): Likewise.
437 (neon_vabal<mode>): Likewise.
438 (neon_vmax<mode>): Likewise.
439 (neon_vmin<mode>): Likewise.
440 (neon_vpaddl<mode>): Likewise.
441 (neon_vpadal<mode>): Likewise.
442 (neon_vpmax<mode>): Likewise.
443 (neon_vpmin<mode>): Likewise.
444 (neon_vrecps<mode>): Likewise.
445 (neon_vrsqrts<mode>): Likewise.
446 (neon_vqabs<mode>): Likewise.
447 (neon_vqneg<mode>): Likewise.
448 (neon_vcls<mode>): Likewise.
449 (clz<mode>2): Likewise.
450 (popcount<mode>2): Likewise.
451 (neon_vrecpe): Likewise.
452 (neon_vrsqrte): Likewise.
453 (neon_vget_lane<mode>_sext_internal): Likewise.
454 (neon_vget_lane<mode>_zext_internal): Likewise.
455 (neon_vdup_n<mode>): Likewise.
456 (neon_vdup_nv2di): Likewise.
457 (neon_vdpu_lane<mode>_internal): Likewise.
458 (neon_vswp<mode>): Likewise.
459 (float<mode><V_cvtto>2): Likewise.
460 (floatuns<mode><V_cvtto>2): Likewise.
461 (fix_trunc<mode><V_cvtto>)2): Likewise
462 (fixuns_trunc<mode><V_cvtto)2): Likewise.
463 (neon_vcvt<mode>): Likewise.
464 (neon_vcvtv4sfv4hf): Likewise.
465 (neon_vcvtv4hfv4sf): Likewise.
466 (neon_vcvt_n<mode>): Likewise.
467 (neon_vmovn<mode>): Likewise.
468 (neon_vqmovn<mode>): Likewise.
469 (neon_vqmovun<mode>): Likewise.
470 (neon_vmovl<mode>): Likewise.
471 (neon_vmul_lane<mode>): Likewise.
472 (neon_vmull_lane<mode>): Likewise.
473 (neon_vqdmull_lane<mode>): Likewise.
474 (neon_vqdmulh_lane<mode>): Likewise.
475 (neon_vmla_lane<mode>): Likewise.
476 (neon_vmlal_lane<mode>): Likewise.
477 (neon_vqdmlal_lane<mode>): Likewise.
478 (neon_vmls_lane<mode>): Likewise.
479 (neon_vmlsl_lane<mode>): Likewise.
480 (neon_vqdmlsl_lane<mode>): Likewise.
481 (neon_vext<mode>): Likewise.
482 (neon_vrev64<mode>): Likewise.
483 (neon_vrev32<mode>): Likewise.
484 (neon_vrev16<mode>): Likewise.
485 (neon_vbsl<mode>_internal): Likewise.
486 (neon_vshl<mode>): Likewise.
487 (neon_vqshl<mode>): Likewise.
488 (neon_vshr_n<mode>): Likewise.
489 (neon_vshrn_n<mode>): Likewise.
490 (neon_vqshrn_n<mode>): Likewise.
491 (neon_vqshrun_n<mode>): Likewise.
492 (neon_vshl_n<mode>): Likewise.
493 (neon_vqshl_n<mode>): Likewise.
494 (neon_vqshlu_n<mode>): Likewise.
495 (neon_vshll_n<mode>): Likewise.
496 (neon_vsra_n<mode>): Likewise.
497 (neon_vsri_n<mode>): Likewise.
498 (neon_vsli_n<mode>): Likewise.
499 (neon_vtbl1v8qi): Likewise.
500 (neon_vtbl2v8qi): Likewise.
501 (neon_vtbl3v8qi): Likewise.
502 (neon_vtbl4v8qi): Likewise.
503 (neon_vtbx1v8qi): Likewise.
504 (neon_vtbx2v8qi): Likewise.
505 (neon_vtbx3v8qi): Likewise.
506 (neon_vtbx4v8qi): Likewise.
507 (neon_vtrn<mode>_internal): Likewise.
508 (neon_vzip<mode>_internal): Likewise.
509 (neon_vuzp<mode>_internal): Likewise.
510 (neon_vld1<mode>): Likewise.
511 (neon_vld1_lane<mode>): Likewise.
512 (neon_vld1_dup<mode>): Likewise.
513 (neon_vld1_dupv2di): Likewise.
514 (neon_vst1<mode>): Likewise.
515 (neon_vst1_lane<mode>): Likewise.
516 (neon_vld2<mode>): Likewise.
517 (neon_vld2_lane<mode>): Likewise.
518 (neon_vld2_dup<mode>): Likewise.
519 (neon_vst2<mode>): Likewise.
520 (neon_vst2_lane<mode>): Likewise.
521 (neon_vld3<mode>): Likewise.
522 (neon_vld3qa<mode>): Likewise.
523 (neon_vld3qb<mode>): Likewise.
524 (neon_vld3_lane<mode>): Likewise.
525 (neon_vld3_dup<mode>): Likewise.
526 (neon_vst3<mode>): Likewise.
527 (neon_vst3qa<mode>): Likewise.
528 (neon_vst3qb<mode>): Likewise.
529 (neon_vst3_lane<mode>): Likewise.
530 (neon_vld4<mode>): Likewise.
531 (neon_vld4qa<mode>): Likewise.
532 (neon_vld4qb<mode>): Likewise.
533 (neon_vld4_lane<mode>): Likewise.
534 (neon_vld4_dup<mode>): Likewise.
535 (neon_vst4<mode>): Likewise.
536 (neon_vst4qa<mode>): Likewise.
537 (neon_vst4qb<mode>): Likewise.
538 (neon_vst4_lane<mode>): Likewise.
539 (neon_vec_unpack<US>_lo_<mode>): Likewise.
540 (neon_vec_unpack<US>_hi_<mode>): Likewise.
541 (neon_vec_<US>mult_lo_<mode>): Likewise.
542 (neon_vec_<US>mult_hi_<mode>): Likewise.
543 (neon_vec_<US>shiftl_<mode>): Likewise.
544 (neon_unpack<US>_<mode>): Likewise.
545 (neon_vec_<US>mult_<mode>): Likewise.
546 (vec_pack_trunc_<mode>): Likewise.
547 (neon_vec_pack_trunk_<mode>): Likewise.
548 (neon_vabd<mode>_2): Likewise.
549 (neon_vabd<mode>_3): Likewise.
550 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
551 (thumb2_movsi_vfp): Likewise.
552 (movdi_vfp): Likewise.
553 (movdi_vfp_cortexa8): Likewise.
554 (movhf_vfp_neon): Likewise.
555 (movhf_vfp): Likewiwse.
556 (movsf_vfp): Likewiwse.
557 (thumb2_movsf_vfp): Likewiwse.
558 (movdf_vfp): Likewise.
559 (thumb2_movdf_vfp): Likewise.
560 (movsfcc_vfp): Likewise.
561 (thumb2_movsfcc_vfp): Likewise.
562 (movdfcc_vfp): Likewise.
563 (thumb2_movdfcc_vfp): Likewise.
564 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
565 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
566 (v10_v2c): Likewise.
567 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
568 attribute change.
569 (cortex_a15_neon_int_2): Likewise.
570 (cortex_a15_neon_int_3): Likewise.
571 (cortex_a15_neon_int_4): Likewise.
572 (cortex_a15_neon_int_5): Likewise.
573 (cortex_a15_neon_vqneg_vqabs): Likewise.
574 (cortex_a15_neon_vmov): Likewise.
575 (cortex_a15_neon_vaba): Likewise.
576 (cortex_a15_neon_vaba_qqq): Likewise.
577 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
578 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
579 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
580 scalar_64_32_long_scalar): Likewise.
581 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
582 (cortex_a15_neon_mla_qqq_8_16): Likewise.
583 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
584 lotype_qdd_64_32_long): Likewise.
585 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
586 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
587 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
588 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
589 (cortex_a15_neon_shift_1): Likewise.
590 (cortex_a15_neon_shift_2): Likewise.
591 (cortex_a15_neon_shift_3): Likewise.
592 (cortex_a15_neon_vshl_ddd): Likewise.
593 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
594 (cortex_a15_neon_vsra_vrsra): Likewise.
595 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
596 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
597 (cortex_a15_neon_fp_vmul_ddd): Likewise.
598 (cortex_a15_neon_fp_vmul_qqd): Likewise.
599 (cortex_a15_neon_fp_vmla_ddd): Likewise.
600 (cortex_a15_neon_fp_vmla_qqq): Likewise.
601 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
602 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
603 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
604 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
605 (cortex_a15_neon_bp_simple): Likewise.
606 (cortex_a15_neon_bp_2cycle): Likewise.
607 (cortex_a15_neon_bp_3cycle): Likewise.
608 (cortex_a15_neon_vld1_1_2_regs): Likewise.
609 (cortex_a15_neon_vld1_3_4_regs): Likewise.
610 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
611 (cortex_a15_neon_vld2_4_regs): Likewise.
612 (cortex_a15_neon_vld3_vld4): Likewise.
613 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
614 (cortex_a15_neon_vst1_3_4_regs): Likewise.
615 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
616 (cortex_a15_neon_vst3_vst4): Likewise.
617 (cortex_a15_neon_vld1_vld2_lane): Likewise.
618 (cortex_a15_neon_vld3_vld4_lane" 10
619 (cortex_a15_neon_vst1_vst2_lane): Likewise.
620 (cortex_a15_neon_vst3_vst4_lane): Likewise.
621 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
622 (cortex_a15_neon_ldm_2): Likewise.0
623 (cortex_a15_neon_stm_2): Likewise.
624 (cortex_a15_neon_mcr): Likewise.
625 (cortex_a15_neon_mcr_2_mcrr): Likewise.
626 (cortex_a15_neon_mrc): Likewise.
627 (cortex_a15_neon_mrrc): Likewise.
628 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
629 change.
630 (cortex_a15_alu_shift): Likewise.
631 (cortex_a15_alu_shift_reg): Likewise.
632 (cortex_a15_mult32): Likewise.
633 (cortex_a15_mult64): Likewise.
634 (cortex_a15_block): Likewise.
635 (cortex_a15_branch): Likewise.
636 (cortex_a15_load1): Likewise.
637 (cortex_a15_load3): Likewise.
638 (cortex_a15_store1): Likewise.
639 (cortex_a15_store3): Likewise.
640 (cortex_a15_call): Likewise.
641 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute
642 change.
643 (cortex_a5_f2r): Likewise.
644 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
645 change.
646 (cortex_a53_f2r): Likewise.
647 * config/arm/cortex-a7.md
648 (cortex_a7_branch): Update for attribute change.
649 (cortex_a7_call): Likewise.
650 (cortex_a7_alu_imm): Likewise.
651 (cortex_a7_alu_reg): Likewise.
652 (cortex_a7_alu_shift): Likewise.
653 (cortex_a7_mul): Likewise.
654 (cortex_a7_load1): Likewise.
655 (cortex_a7_store1): Likewise.
656 (cortex_a7_load2): Likewise.
657 (cortex_a7_store2): Likewise.
658 (cortex_a7_load3): Likewise.
659 (cortex_a7_store3): Likewise.
660 (cortex_a7_load4): Likewise.
661 (cortex_a7_store4): Likewise.
662 (cortex_a7_fpalu): Likewise.
663 (cortex_a7_fconst): Likewise.
664 (cortex_a7_fpmuls): Likewise.
665 (cortex_a7_neon_mul): Likewise.
666 (cortex_a7_fpmacs): Likewise.
667 (cortex_a7_neon_mla: Likewise.
668 (cortex_a7_fpmuld: Likewise.
669 (cortex_a7_fpmacd: Likewise.
670 (cortex_a7_fpfmad: Likewise.
671 (cortex_a7_fdivs: Likewise.
672 (cortex_a7_fdivd: Likewise.
673 (cortex_a7_r2f: Likewise.
674 (cortex_a7_f2r: Likewise.
675 (cortex_a7_f_flags: Likewise.
676 (cortex_a7_f_loads: Likewise.
677 (cortex_a7_f_loadd: Likewise.
678 (cortex_a7_f_stores: Likewise.
679 (cortex_a7_f_stored: Likewise.
680 (cortex_a7_neon): Likewise.
681 * config/arm/cortex-a8-neon.md
682 (cortex_a8_neon_mrc): Update for attribute change.
683 (cortex_a8_neon_mrrc): Likewise.
684 (cortex_a8_neon_int_1): Likewise.
685 (cortex_a8_neon_int_2): Likewise.
686 (cortex_a8_neon_int_3): Likewise.
687 (cortex_a8_neon_int_4): Likewise.
688 (cortex_a8_neon_int_5): Likewise.
689 (cortex_a8_neon_vqneg_vqabs): Likewise.
690 (cortex_a8_neon_vmov): Likewise.
691 (cortex_a8_neon_vaba): Likewise.
692 (cortex_a8_neon_vaba_qqq): Likewise.
693 (cortex_a8_neon_vsma): Likewise.
694 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
695 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
696 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
697 Likewise.
698 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
699 (cortex_a8_neon_mla_qqq_8_16): Likewise.
700 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
701 long_scalar_qdd_64_32_long): Likewise.
702 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
703 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
704 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
705 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
706 (cortex_a8_neon_shift_1): Likewise.
707 (cortex_a8_neon_shift_2): Likewise.
708 (cortex_a8_neon_shift_3): Likewise.
709 (cortex_a8_neon_vshl_ddd): Likewise.
710 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
711 (cortex_a8_neon_vsra_vrsra): Likewise.
712 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
713 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
714 (cortex_a8_neon_fp_vsum): Likewise.
715 (cortex_a8_neon_fp_vmul_ddd): Likewise.
716 (cortex_a8_neon_fp_vmul_qqd): Likewise.
717 (cortex_a8_neon_fp_vmla_ddd): Likewise.
718 (cortex_a8_neon_fp_vmla_qqq): Likewise.
719 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
720 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
721 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
722 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
723 (cortex_a8_neon_bp_simple): Likewise.
724 (cortex_a8_neon_bp_2cycle): Likewise.
725 (cortex_a8_neon_bp_3cycle): Likewise.
726 (cortex_a8_neon_ldr): Likewise.
727 (cortex_a8_neon_str): Likewise.
728 (cortex_a8_neon_vld1_1_2_regs): Likewise.
729 (cortex_a8_neon_vld1_3_4_regs): Likewise.
730 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
731 (cortex_a8_neon_vld2_4_regs): Likewise.
732 (cortex_a8_neon_vld3_vld4): Likewise.
733 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
734 (cortex_a8_neon_vst1_3_4_regs): Likewise.
735 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
736 (cortex_a8_neon_vst3_vst4): Likewise.
737 (cortex_a8_neon_vld1_vld2_lane): Likewise.
738 (cortex_a8_neon_vld3_vld4_lane): Likewise.
739 (cortex_a8_neon_vst1_vst2_lane): Likewise.
740 (cortex_a8_neon_vst3_vst4_lane): Likewise.
741 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
742 (cortex_a8_neon_mcr): Likewise.
743 (cortex_a8_neon_mcr_2_mcrr): Likewise.
744 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
745 change.
746 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
747 change.
748 (ca9_neon_mrrc): Likewise.
749 (cortex_a9_neon_int_1): Likewise.
750 (cortex_a9_neon_int_2): Likewise.
751 (cortex_a9_neon_int_3): Likewise.
752 (cortex_a9_neon_int_4): Likewise.
753 (cortex_a9_neon_int_5): Likewise.
754 (cortex_a9_neon_vqneg_vqabs): Likewise.
755 (cortex_a9_neon_vmov): Likewise.
756 (cortex_a9_neon_vaba): Likewise.
757 (cortex_a9_neon_vaba_qqq): Likewise.
758 (cortex_a9_neon_vsma): Likewise.
759 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
760 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
761 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
762 Likewise.
763 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
764 (cortex_a9_neon_mla_qqq_8_16): Likewise.
765 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
766 long_scalar_qdd_64_32_long): Likewise.
767 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
768 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
769 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
770 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
771 (cortex_a9_neon_shift_1): Likewise.
772 (cortex_a9_neon_shift_2): Likewise.
773 (cortex_a9_neon_shift_3): Likewise.
774 (cortex_a9_neon_vshl_ddd): Likewise.
775 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
776 (cortex_a9_neon_vsra_vrsra): Likewise.
777 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
778 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
779 (cortex_a9_neon_fp_vsum): Likewise.
780 (cortex_a9_neon_fp_vmul_ddd): Likewise.
781 (cortex_a9_neon_fp_vmul_qqd): Likewise.
782 (cortex_a9_neon_fp_vmla_ddd): Likewise.
783 (cortex_a9_neon_fp_vmla_qqq): Likewise.
784 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
785 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
786 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
787 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
788 (cortex_a9_neon_bp_simple): Likewise.
789 (cortex_a9_neon_bp_2cycle): Likewise.
790 (cortex_a9_neon_bp_3cycle): Likewise.
791 (cortex_a9_neon_ldr): Likewise.
792 (cortex_a9_neon_str): Likewise.
793 (cortex_a9_neon_vld1_1_2_regs): Likewise.
794 (cortex_a9_neon_vld1_3_4_regs): Likewise.
795 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
796 (cortex_a9_neon_vld2_4_regs): Likewise.
797 (cortex_a9_neon_vld3_vld4): Likewise.
798 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
799 (cortex_a9_neon_vst1_3_4_regs): Likewise.
800 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
801 (cortex_a9_neon_vst3_vst4): Likewise.
802 (cortex_a9_neon_vld1_vld2_lane): Likewise.
803 (cortex_a9_neon_vld3_vld4_lane): Likewise.
804 (cortex_a9_neon_vst1_vst2_lane): Likewise.
805 (cortex_a9_neon_vst3_vst4_lane): Likewise.
806 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
807 (cortex_a9_neon_mcr): Likewise.
808 (cortex_a9_neon_mcr_2_mcrr): Likewise.
809 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
810 (cortex_a9_fps): Likewise.
811 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
812 change.
813 (cortex_m4_fmuls): Likewise.
814 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
815 change.
816 (cortex_r4_mrc): Likewise.
817 * config/arm/iterators.md: Update comment referring to neon_type.
818 * config/arm/iwmmxt.md
819 (iwmmxt_arm_movdi): Update for attribute change.
820 (iwmmxt_movsi_insn): Likewise.
821 * config/arm/marvell-pj4.md
822 (pj4_vfp_to_core): Update for attribute change.
823 (pj4_core_to_vfp): Likewise.
824 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
825 attribute change.
826 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
827 (vfp_fstore): Likewise.
828 * doc/md.texi: Change references to neon_type to refer to type.
829
fbe0dc78
DS
8302013-09-04 Dodji Seketeli <dodji@redhat.com>
831
832 * tree.h (DECL_BUILT_IN): Fix typo in comment.
833
c167bc5b
DE
8342013-09-04 David Edelsohn <dje.gcc@gmail.com>
835
836 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
837 lglobl if not weak.
838
035cb59f
ER
8392013-09-04 Easwaran Raman <eraman@google.com>
840
841 PR middle-end/57370
842 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
843 (build_and_add_sum): Use it.
844 (appears_later_in_bb): Simplify code.
845
71191083
TJ
8462013-09-04 Teresa Johnson <tejohnson@google.com>
847
848 * dumpfile.c (dump_finish): Don't close stderr/stdout.
849
0c488213
JG
8502013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
851
852 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
853
61a74079
JH
8542013-09-04 Jan Hubicka <jh@suse.cz>
855
856 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
857 * ipa-devirt.c: Include diganostic.h
858 (odr_type_d): Add types and types_set.
859 (hash_type_name): Work for types with vtables during LTO.
860 (odr_hasher::remove): Fix comment; destroy types_set.
861 (add_type_duplicate): New function,
862 (get_odr_type): Use it.
863 (dump_type_inheritance_graph): Dump type duplicates.
864 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
865 graph.
866 * tree.c (types_same_for_odr): Give exact answers on types with
867 virtual tables.
868
24ba8195
DS
8692013-09-04 Dodji Seketeli <dodji@redhat.com>
870
871 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
872 explaining their differences.
873
0a04869e
SKS
8742013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
875
876 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
877
3b18bc42
JL
8782013-09-03 Jeff Law <law@redhat.com>
879
880 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
881 when not threading through a joiner block. Pass joiner/no joiner
882 state to register_jump_thread.
883 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
884 state from argument rather than implying on path length.
885 Dump the entire jump thread path into debugging dump.
886 * tree-flow.h (register_jump_thread): Update prototype.
887
ae84e151
XDL
8882013-08-29 Xinliang David Li <davidxl@google.com>
889
890 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
891 Remove a trivial gcc_assert.
892
c716e67f
XDL
8932013-08-29 Xinliang David Li <davidxl@google.com>
894
895 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
896 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
897 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
898 Delay base decl alignment adjustment.
899 * tree-vectorizer.c (vect_destroy_datarefs): New function.
900 * tree-vectorizer.h: New data structure.
901 (set_dr_misalignment): New function.
902 (dr_misalignment): Ditto.
903 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
904 (vectorizable_load): Ditto.
905 (ensure_base_align): New function.
906 (vectorize_loops): Add dbg_cnt support.
907 (execute_vect_slp): Ditto.
908 * dbgcnt.def: New debug counter.
909 * Makefile: New dependency.
910
db8800bc
MI
9112013-09-03 Meador Inge <meadori@codesourcery.com>
912
913 Revert:
914
915 2013-08-30 Meador Inge <meadori@codesourcery.com>
916
917 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
918
0d9b4f5c
DE
9192013-09-03 David Edelsohn <dje.gcc@gmail.com>
920
921 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
922 function descriptor.
923
29e810b9
RB
9242013-09-03 Richard Biener <rguenther@suse.de>
925
926 * tree-affine.c (add_elt_to_tree): Fix association issue,
927 avoid useless converts and make sure to always return a
928 properly typed result.
929
dbfc2894
RB
9302013-09-03 Richard Biener <rguenther@suse.de>
931
932 PR middle-end/57656
933 * fold-const.c (negate_expr_p): Fix division case.
934 (negate_expr): Likewise.
935
fa3bf4ce
RB
9362013-09-03 Richard Biener <rguenther@suse.de>
937
938 PR lto/58285
939 * tree-streamer-out.c: Include tm.h.
940 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
941
a64bbb3f
JH
9422013-09-03 Jan Hubicka <jh@suse.cz>
943
944 * tree-profile.c (tree_profiling): Cleanup CFG when done.
945
93ed07e2
AM
9462013-09-03 Alan Modra <amodra@gmail.com>
947
948 * config.gcc (powerpc*-*-linux*): Add support for little-endian
949 multilibs to big-endian target and vice versa.
950 * config/rs6000/t-linux64: Use := assignment on all vars.
951 (MULTILIB_EXTRA_OPTS): Remove fPIC.
952 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
953 * config/rs6000/t-linux64le: New file.
954 * config/rs6000/t-linux64bele: New file.
955 * config/rs6000/t-linux64lebe: New file.
956
07d6cd64
JH
9572013-09-02 Jan Hubicka <jh@suse.cz>
958
959 * ipa-inline-transform.c (inline_transform): Do not
960 optimize_inline_calls when not optimizing.
961
8e4c9a10
JH
9622013-09-02 Jan Hubicka <jh@suse.cz>
963
964 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
965 duplicated nodes for assembler names.
966 * symtab.c (symtab_unregister_node): Do not attempt to unlink
967 hard registers from assembler name hash.
968
c91061e6
JH
9692013-09-02 Jan Hubicka <jh@suse.cz>
970
971 * ipa-split.c (execute_split_functions): Split externally visible
972 functions called once.
973
2fd0985c
MJ
9742013-09-02 Martin Jambor <mjambor@suse.cz>
975
976 PR ipa/58106
977 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
978 linked list. When finding the correct duplicate, also consider also
979 the caller in additon to its inlined_to node.
980
91bd4114
JG
9812013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
982
983 * config/aarch64/aarch64-simd-builtins.def
984 (dup_lane_scalar): Remove.
985 * config/aarch64/aarch64-simd.md
986 (aarch64_simd_dup): Add 'w->w' alternative.
987 (aarch64_dup_lane<mode>): Allow for VALL.
988 (aarch64_dup_lane_scalar<mode>): Remove.
989 (aarch64_dup_lane_<vswap_width_name><mode>): New.
990 (aarch64_get_lane_signed<mode>): Add w->w altenative.
991 (aarch64_get_lane_unsigned<mode>): Likewise.
992 (aarch64_get_lane<mode>): Likewise.
993 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
994 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
995 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
996 (VCON): Change container of V2SF.
997 (vswap_width_name): Likewise.
998 * config/aarch64/arm_neon.h
999 (__aarch64_vdup_lane_any): New.
1000 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1001 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
1002 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
1003
d617d2d8
EB
10042013-09-02 Eric Botcazou <ebotcazou@adacore.com>
1005
1006 PR middle-end/56382
1007 * expr.c (emit_move_complex): Do not move complex FP values as parts if
1008 the source or the destination is a single hard register.
1009
63fdb7be
RB
10102013-09-02 Richard Biener <rguenther@suse.de>
1011
1012 PR middle-end/57511
1013 * tree-scalar-evolution.c (instantiate_scev_name): Allow
1014 non-linear SCEVs.
1015
78de2333
RB
10162013-09-02 Richard Biener <rguenther@suse.de>
1017
1018 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
1019 arithmetic to sizetype.
1020
85ff4ec6
BC
10212013-09-02 Bin Cheng <bin.cheng@arm.com>
1022
1023 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
1024 Find auto-increment use both before and after candidate.
1025
576016fe
MP
10262013-09-02 Marek Polacek <polacek@redhat.com>
1027
1028 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
1029
08f835dc
JH
10302013-09-01 Jan Hubicka <jh@suse.cz>
1031
1032 * Makefile.in: Add ipa-profile.o
1033 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
1034 * cgraph.c (struct cgraph_propagate_frequency_data,
1035 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
1036 ipa-profile.c; replace cgraph_ by ipa_ prefix.
1037 * cgraph.h (cgraph_propagate_frequency): Remove.
1038 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
1039 (inline_update_callee_summaries): Update.
1040 * ipa-profile.c: New file.
1041 * ipa-utils.h (ipa_propagate_frequency): Declare.
1042 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
1043 data-streamer.h, value-prof.h
1044 (symtab_remove_unreachable_nodes): Update profile.
1045 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
1046 account_time_size, cmp_counts, dump_histogram,
1047 ipa_profile_generate_summary, ipa_profile_write_summary,
1048 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
1049 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
1050 Move to ipa-profile.c
1051
cec8583c
JDA
10522013-09-01 John David Anglin <danglin@gcc.gnu.org>
1053
1054 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
1055
bbc9396b
JH
10562013-09-01 Jan Hubicka <jh@suse.cz>
1057
1058 * common.opt (fdevirtualize-speculatively): New function.
1059 * invoke.texi (fdevirtualize-speculatively): Document.
1060 * ipa-devirt.c: Include ipa-inline.h
1061 (likely_target_p): New function.
1062 (ipa_devirt): New function.
1063 (gate_ipa_devirt): New function.
1064 (pass_data_ipa_devirt): New static var.
1065 (pass_ipa_devirt): Likewise.
1066 (make_pass_ipa_devirt): New function.
1067 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
1068 (common_handle_option): Disable devirtualization when
1069 value range profiling is available.
1070 * passes.def (pass_ipa_devirt): Add.
1071 * timever.def (TV_IPA_DEVIRT): New timevar.
1072 * tree-pass.h (make_pass_ipa_devirt):
1073
0cea211e
IS
10742013-09-01 Iain Sandoe <iain@codesourcery.com>
1075
1076 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
1077 include sanitize(undefined).
1078
5e351e96
DN
10792013-08-31 Diego Novillo <dnovillo@google.com>
1080
1081 * Makefile.in (TREE_CORE_H): Define.
1082 (TREE_H): Use.
1083 (GTFILES): Add tree-core.h.
1084 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
1085 size the array.
1086 * tree-core.h: New file.
1087 Move all data structures, enum, typedefs, global
1088 declarations and constants from ...
1089 * tree.h: ... here.
1090
63bf9a90
JH
10912013-08-31 Jan Hubicka <jh@suse.cz>
1092
1093 * bulitins.c (expand_builtin): Do not early exit for gcov
1094 instrumented functions.
1095
5c0abd6a
MP
10962013-08-31 Marek Polacek <polacek@redhat.com>
1097
1098 * ubsan.c: Include tm_p.h.
1099
4843f032
JH
11002013-08-31 Jan Hubicka <jh@suse.cz>
1101
8356c89c
JH
1102 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
1103 warning.
1104
4843f032
JH
1105 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
1106 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 1107 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH
1108 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
1109 (ipa_merge_profiles): New function.
1110 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
1111 (lto_input_function_body): Likewise.
5e581212
JH
1112 * ipa-utils.h (ipa_merge_profiles): Declare.
1113 * lto-streamer.h (lto_input_function_body): Update prototype.
1114 (emit_label_in_global_context_p): Remove.
1115 * lto-symtab.c: Include ipa-utils.h
1116 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 1117
57292ce9
JH
11182013-08-31 Jan Hubicka <jh@suse.cz>
1119
1120 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
1121
b9cbfeeb
JH
11222013-08-31 Jan Hubicka <jh@suse.cz>
1123
1124 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
1125
41dedebd
UB
11262013-08-31 Uros Bizjak <ubizjak@gmail.com>
1127
1128 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
1129 "cmp" RTX before signed_comparison_operator check to account
1130 for "code" changes.
1131
06d65050
JH
11322013-08-30 Jan Hubicka <jh@suse.cz>
1133
41dedebd 1134 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH
1135 (detect_type_change_1): Rename to ...
1136 (detect_type_change): ... this one; early return on non-polymorphic
1137 types.
41dedebd 1138 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH
1139 use of detect_type_change.
1140 (compute_complex_assign_jump_func): Add param_type parameter;
1141 update use of detect_type_change_ssa.
1142 (compute_complex_ancestor_jump_func): Likewise.
1143 (ipa_get_callee_param_type): New function.
1144 (ipa_compute_jump_functions_for_edge): Compute parameter type;
1145 update calls to the jump function computation functions.
1146
600b5b1d
TJ
11472013-08-30 Teresa Johnson <tejohnson@google.com>
1148 Steven Bosscher <steven@gcc.gnu.org>
1149
1150 * cfgrtl.c (fixup_new_cold_bb): New routine.
1151 (commit_edge_insertions): Invoke fixup_partitions.
1152 (find_partition_fixes): New routine.
1153 (fixup_partitions): Ditto.
1154 (verify_hot_cold_block_grouping): Update comments.
1155 (rtl_verify_edges): Invoke find_partition_fixes.
1156 (rtl_verify_bb_pointers): Update comments.
1157 (rtl_verify_bb_layout): Ditto.
1158 * basic-block.h (probably_never_executed_edge_p): Declare.
1159 (fixup_partitions): Ditto.
1160 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
1161 * bb-reorder.c (sanitize_hot_paths): New function.
1162 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
1163 sanitize_hot_paths.
1164 * predict.c (probably_never_executed_edge_p): New routine.
1165 * cfg.c (check_bb_profile): Add partition insanity warnings.
1166
0a6ea5c9
MI
11672013-08-30 Meador Inge <meadori@codesourcery.com>
1168
1169 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
1170
de5a5fa1
MP
11712013-08-30 Marek Polacek <polacek@redhat.com>
1172
1173 * Makefile.in (ubsan.o): Add.
1174 (c-family/c-ubsan.o): Add.
1175 (builtins.o): Add ubsan.h dependency.
1176 * ubsan.h: New file.
1177 * ubsan.c: New file.
1178 * common.opt: Add -fsanitize=undefined option.
1179 (flag_sanitize): Add variable.
1180 (fsanitize=): Add option. Add Driver.
1181 (fsanitize=thread): Remove option.
1182 (fsanitize=address): Likewise.
1183 (static-libubsan): New option.
1184 * doc/invoke.texi: Document the new flag and -static-libubsan.
1185 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
1186 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
1187 * builtin-attrs.def (ATTR_COLD): Define.
1188 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1189 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
1190 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
1191 * flag-types.h (sanitize_code): New enum.
1192 * opts.c (common_handle_option): Parse command line arguments
1193 of -fsanitize=. Add -fsanitize=unreachable option.
1194 * varasm.c (get_variable_section): Adjust.
1195 (assemble_noswitch_variable): Likewise.
1196 (assemble_variable): Likewise.
1197 (output_constant_def_contents): Likewise.
1198 (categorize_decl_for_section): Likewise.
1199 (place_block_symbol): Likewise.
1200 (output_object_block): Likewise.
1201 * builtins.def: Likewise.
1202 * toplev.c (compile_file): Likewise.
1203 (process_options): Likewise.
1204 * cppbuiltin.c: Likewise.
1205 * tsan.c (tsan_pass): Likewise.
1206 (tsan_gate): Likewise.
1207 (tsan_gate_O0): Likewise.
1208 * cfgexpand.c (partition_stack_vars): Likewise.
1209 (expand_stack_vars): Likewise.
1210 (defer_stack_allocation): Likewise.
1211 (expand_used_vars): Likewise.
1212 * cfgcleanup.c (old_insns_match_p): Likewise.
1213 * asan.c (asan_finish_file): Likewise.
1214 (asan_instrument): Likewise.
1215 (gate_asan): Likewise.
1216 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
1217 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
1218 (asan_global_struct): Use pointer_sized_int_node instead
1219 calling build_nonstandard_integer_type.
1220 (initialize_sanitizer_builtins): Likewise.
1221 (asan_finish_file): Likewise.
1222 * gcc.c: Document %{%:function(args):X}.
1223 (static_spec_functions): Add sanitize.
1224 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
1225 store funcval != NULL there.
1226 (do_spec_1): Adjust handle_spec_function caller.
1227 (handle_braces): Allow %:function(args) as condition.
1228 (sanitize_spec_function): New function.
1229 (ADD_STATIC_LIBUBSAN_LIBS): Define.
1230 (LIBUBSAN_SPEC): Likewise.
1231 (LIBUBSAN_EARLY_SPEC): Likewise.
1232 (SANITIZER_SPEC): Handle libubsan.
1233 (SANITIZER_EARLY_SPEC): Likewise.
1234 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
1235 instead of fsanitize=address.
1236 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
1237 instead of fsanitize=address*.
1238 * builtins.c: Include ubsan.h.
1239 (fold_builtin_0): Instrument __builtin_unreachable.
1240 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
1241 instead of flag_asan.
1242 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
1243 (pointer_sized_int_node): Define.
41dedebd 1244 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 1245
f07f30cf
MS
12462013-08-30 Mike Stump <mikestump@comcast.net>
1247
1248 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
1249 with RE patterns.
1250
8a41354f
JH
12512013-08-29 Jan Hubicka <jh@suse.cz>
1252
1253 * cgraph.c (cgraph_function_body_availability): Handle weakref
1254 correctly.
1255 * passes.def: Remove pass_fixup_cfg.
1256 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
1257 track when we need to remove functions.
1258 (gate_ipa_inline): Execute inlining always; add comment why.
1259 (pass_data_ipa_inline): Remove TODO_remove_functions.
1260 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
1261 do not produce summaries.
1262 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
1263 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
1264 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
1265 constant pool and vtable.
1266
dc8d7a0f
TB
12672013-08-30 Tejas Belagod <tejas.belagod@arm.com>
1268
41dedebd
UB
1269 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
1270 New arm_neon.h's internal macros to specify 64-bit constants.
1271 Avoid using stdint.h's macros.
dc8d7a0f 1272
ac7eacd2
JR
12732013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
1274
1275 * recog.c (verify_changes): Verify that changes[i].old is non-zero
1276 before applying REG_P.
1277
8b29fd4e
JJ
12782013-08-30 Jakub Jelinek <jakub@redhat.com>
1279
1280 PR tree-optimization/58277
1281 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
1282 after seeing too many stmts with vdef in between dombb and current
1283 bb, invalidate everything.
1284
26d75703
RB
12852013-08-30 Richard Biener <rguenther@suse.de>
1286
1287 * fold-const.c (fold_single_bit_test): Fix overflow test.
1288
f7e088e7
EB
12892013-08-30 Eric Botcazou <ebotcazou@adacore.com>
1290
1291 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
1292 and which can live in a register, always retrieve the value on entry.
1293 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
1294 passed by invisible reference specially.
1295 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
1296 (vt_add_function_parameter): Correctly deal with a parameter passed by
1297 invisible reference.
1298
12992013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH
1300
1301 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
1302
6e8dad05
RB
13032013-08-30 Richard Biener <rguenther@suse.de>
1304
1305 PR tree-optimization/58228
1306 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
1307 allow invariant loads in nested loop vectorization.
1308
062ef2c8
RB
13092013-08-30 Richard Biener <rguenther@suse.de>
1310
1311 PR tree-optimization/58223
1312 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
1313 (has_anti_or_output_dependence): ... this and adjust to also
1314 look for output dependences.
1315 (mark_nodes_having_upstream_mem_writes): Adjust.
1316 (rdg_flag_uses): Likewise.
1317
7a764c60
RB
13182013-08-30 Richard Biener <rguenther@suse.de>
1319
1320 PR tree-optimization/58010
1321 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
1322 assert that we have a loop-closed PHI.
1323
82e9d642
JH
13242013-08-29 Jan Hubicka <jh@suse.cz>
1325
1326 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
1327 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
1328 * lto-section-in.c (lto_free_function_in_decl_state): New function.
1329 (lto_free_function_in_decl_state_for_node): New function.
1330
9cc1fb4b
XDL
13312013-08-29 Xinliang David Li <davidxl@google.com>
1332
41dedebd 1333 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL
1334 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
1335 Emit alignment peeling message with default -fopt-info.
1336 (vect_loop_versioning): Emit loop version info message.
41dedebd 1337 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL
1338 (execute_vect_slp): Ditto.
1339
440a5082
EB
13402013-08-29 Eric Botcazou <ebotcazou@adacore.com>
1341
1342 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
1343 of the clone from the DECL_NAME of the original function.
1344
3fa3690d
OE
13452013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
1346
1347 * passes.c (register_pass): Add overload.
1348 * tree-pass.h (register_pass): Forward declare it. Add comment.
1349
0170f33c
JH
13502013-08-29 Jan Hubicka <jh@suse.cz>
1351
41dedebd
UB
1352 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
1353 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
1354 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
1355 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
1356 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
1357 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 1358 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
1359 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
1360
103ff0d6
TJ
13612013-08-29 Teresa Johnson <tejohnson@google.com>
1362
1363 * dumpfile.c (dump_loc): Output column number.
1364 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
1365 * doc/invoke.texi: Document optall -fopt-info flag.
1366 * profile.c (read_profile_edge_counts): Use new dump framework.
1367 (compute_branch_probabilities): Ditto.
1368 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
1369 when pass not in any opt group.
1370 * pass_manager.h (pass_manager::get_pass_profile): New method.
1371 * value-prof.c (check_counter): Use new dump framework.
1372 (check_ic_target): Ditto.
1373 * coverage.c (get_coverage_counts): Ditto.
1374 (coverage_init): Setup new dump framework.
1375
301bbc16
RB
13762013-08-29 Richard Biener <rguenther@suse.de>
1377
1378 PR tree-optimization/58246
1379 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
1380 handle the dominance check inside a basic-block.
1381
1b275000
RB
13822013-08-29 Richard Biener <rguenther@suse.de>
1383
1384 PR middle-end/57287
1385 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
1386 of default defs that appear in abnormal PHI nodes.
1387
6b1184ba
RB
13882013-08-29 Richard Biener <rguenther@suse.de>
1389
1390 PR tree-optimization/57685
1391 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
1392 single-use operands to avoid exponential complexity.
1393
1b08b734
DC
13942013-08-28 Dehao Chen <dehao@google.com>
1395
1396 * ipa-inline.c (edge_badness): Fix integer underflow.
1397
48a3fa69
UB
13982013-08-28 Uros Bizjak <ubizjak@gmail.com>
1399
1400 * gtm-builtins.def (_ITM_free): Declare leaf.
1401
215f73e6
JJ
14022013-08-28 Jakub Jelinek <jakub@redhat.com>
1403
1404 PR target/58067
1405 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
1406 (*tls_local_dynamic_base_64_largepic): Likewise.
1407 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
1408 Remove predicate from call operand.
1409 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
1410 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
1411
34554d1a
JL
14122013-08-28 Jeff Law <law@redhat.com>
1413
1414 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
1415 checks for the number of predecessors and successors allowed.
1416 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
1417 which require copying a joiner block if there is a request which
1418 is a subpath that requires no joiner block copying.
1419
c01c111b
JH
14202013-08-28 Jan Hubicka <jh@suse.cz>
1421
1422 * lto-streamer-out.c (DFS_write_tree_body): Drop
1423 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
1424 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
1425 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
1426 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
1427 Do not read DECL_ERROR_ISSUED.
1428 (unpack_ts_decl_with_vis_value_fields): Do not read
1429 DECL_DEFER_OUTPUT.
48a3fa69
UB
1430 (lto_input_ts_binfo_tree_pointers): Do not read
1431 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH
1432 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
1433 write DECL_ERROR_ISSUED..
1434 (pack_ts_decl_with_vis_value_fields): Do not write
1435 DECL_DEFER_OUTPUT.
1436 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 1437 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH
1438 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
1439 * tree.h (tree_decl_common): Update comment.
1440 (DECL_ERROR_ISSUED): Remove.
1441
4b128ece
JJ
14422013-08-28 Jakub Jelinek <jakub@redhat.com>
1443
1444 PR middle-end/58257
1445 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
1446
12211b99 14472013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH
1448
1449 * builtins.def (free): Declare leaf.
1450
8dce4dbc
DM
14512013-08-27 David Malcolm <dmalcolm@redhat.com>
1452
1453 * gdbhooks.py: New.
1454 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
1455 * configure: Regenerate.
1456
b8f6e610
MJ
14572013-08-27 Martin Jambor <mjambor@suse.cz>
1458
1459 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
1460 (ipa_ancestor_jf_data): Likewise.
1461 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
1462 (ipa_get_jf_pass_through_type_preserved): New function.
1463 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
1464 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 1465 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ
1466 (ipa_get_jf_ancestor_result): Likewise.
1467 (propagate_vals_accross_pass_through): Use
1468 ipa_get_jf_pass_through_result to do all the value mappings.
1469 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
1470 type_preserved flag.
1471 (ipa_set_jf_cst_copy): New function.
1472 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
1473 (ipa_set_jf_arith_pass_through): Likewise.
1474 (ipa_set_ancestor_jf): Likewise.
1475 (compute_complex_assign_jump_func): Set type_preserved instead of
1476 punting.
1477 (ipa_compute_jump_functions_for_edge): Likewise.
1478 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
1479 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 1480 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ
1481 (ipa_write_jump_function): Stream the type_preserved flags.
1482 (ipa_read_jump_function): Likewise.
1483
74bf76ed
JJ
14842013-08-27 Jakub Jelinek <jakub@redhat.com>
1485 Aldy Hernandez <aldyh@redhat.com>
1486
1487 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
1488 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
1489 * function.h (struct function): Add has_force_vect_loops and
1490 has_simduid_loops.
12211b99 1491 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ
1492 * gimple.c (gimple_build_omp_critical): Add KIND argument and
1493 handle it.
1494 * gimple.def: Update CLAUSES comments.
1495 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
1496 (gimple_build_omp_for): Add argument to prototype.
1497 (gimple_omp_for_kind): New.
1498 (gimple_omp_for_set_kind): New.
1499 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
1500 GOVD_DATA_SHARE_CLASS.
1501 (enum omp_region_type): Add ORT_SIMD.
1502 (gimple_add_tmp_var): Handle ORT_SIMD.
1503 (gimplify_var_or_parm_decl): Same.
1504 (is_gimple_stmt): Same.
1505 (omp_firstprivatize_variable): Same.
1506 (omp_add_variable): Only use splay_tree_insert if lookup failed.
1507 (omp_notice_variable): Handle ORT_SIMD.
12211b99 1508 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ
1509 (omp_check_private): Handle ORT_SIMD.
1510 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1511 OMP_CLAUSE_SAFELEN.
1512 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
1513 Handle OMP_CLAUSE_LASTPRIVATE.
1514 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
1515 OMP_CLAUSE_SAFELEN.
1516 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
1517 (gimplify_expr): Handle OMP_SIMD.
1518 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
1519 (expand_GOMP_SIMD_VF): New.
1520 (expand_GOMP_SIMD_LAST_LANE): New.
1521 * internal-fn.def (GOMP_SIMD_LANE): New.
1522 (GOMP_SIMD_VF): New.
1523 (GOMP_SIMD_LAST_LANE): New.
1524 * omp-low.c: Include target.h.
1525 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
1526 OMP_CLAUSE_SAFELEN.
1527 (check_omp_nesting_restrictions): Same.
1528 (omp_max_vf): New.
1529 (lower_rec_simd_input_clauses): New.
1530 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
1531 OMP_CLAUSE_LINEAR.
1532 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
1533 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
1534 (expand_omp_build_assign): New.
1535 (expand_omp_for_init_counts): New.
1536 (expand_omp_for_init_vars): New.
1537 (extract_omp_for_update_vars): New.
1538 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
1539 and rewrite accordingly.
1540 (expand_omp_simd): New.
1541 (expand_omp_for): Use expand_omp_simd.
1542 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
1543 (lower_omp_for): Do not lower the body.
1544 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
1545 in their own loops.
1546 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB
1547 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
1548 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ
1549 (gate_tree_if_conversion): Similarly.
1550 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
1551 gimple_build_omp_for.
1552 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
1553 * tree-parloops (create_parallel_loop): Pass kind argument to
1554 gimple_build_omp_for.
1555 * tree-pretty-print.c (dump_omp_clause): Add cases for
1556 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
1557 OMP_CLAUSE__SIMDUID_.
1558 (dump_generic_node): Handle OMP_SIMD.
1559 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
1560 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
1561 unroll OMP_SIMD loops here.
12211b99 1562 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 1563 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 1564 loop->safelen.
74bf76ed
JJ
1565 (vect_analyze_data_refs): Handle simd loops.
1566 * tree-vect-loop.c (vectorizable_live_operation): Handle
1567 IFN_GOMP_SIMD*.
12211b99 1568 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ
1569 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
1570 (vectorizable_load): Same.
12211b99 1571 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ
1572 (struct simduid_to_vf): New.
1573 (simduid_to_vf::hash): New.
1574 (simduid_to-vf::equal): New.
1575 (struct simd_array_to_simduid): New.
1576 (simd_array_to_simduid::hash): New.
1577 (simd_array_to_simduid::equal): New.
1578 (adjust_simduid_builtins): New.
1579 (struct note_simd_array_uses_struct): New.
1580 (note_simd_array_uses_cb): New.
1581 (note_simd_array_uses): New.
1582 (vectorize_loops): Handle simd hints and adjust simd builtins
1583 accordingly.
1584 * tree-vectorizer.h (struct _stmt_vec_info): Add
1585 simd_lane_access_p field.
1586 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
1587 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
1588 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
1589 (omp_clause_code_name): Same.
1590 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
1591 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
1592 * tree.def (OMP_SIMD): New entry.
12211b99
UB
1593 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
1594 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ
1595 (OMP_CLAUSE_DECL): Adjust range for new clauses.
1596 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
1597 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
1598 (OMP_CLAUSE_LINEAR_STEP): New.
1599 (OMP_CLAUSE_SAFELEN_EXPR): New.
1600 (OMP_CLAUSE__SIMDUID__DECL): New.
1601 (find_omp_clause): New prototype.
1602
d0cf4e84
L
16032013-08-27 H.J. Lu <hongjiu.lu@intel.com>
1604
1605 * config/i386/driver-i386.c (host_detect_local_cpu): Update
1606 Haswell processor detection.
1607
992592ec
CW
16082013-08-27 Christian Widmer <shadow@umbrox.de>
1609
1610 PR target/57927
1611 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
1612 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
1613 AVX2 capable processors.
1614
9116eb22
TB
16152013-08-27 Tejas Belagod <tejas.belagod@arm.com>
1616
1617 * config/aarch64/arm_neon.h: Replace all inline asm implementations
1618 of vget_low_* with implementations in terms of other intrinsics.
1619
1c5abb53
MG
16202013-08-27 Marc Glisse <marc.glisse@inria.fr>
1621
1622 PR middle-end/57219
1623 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
1624 values to -1, 0 and 1.
1625
2e100703
VP
16262013-08-27 Vidya Praveen <vidyapraveen@arm.com>
1627
1628 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
1629 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
1630 (<optab><mode>3_insn): Remove.
1631 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
1632 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
1633 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
1634 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
1635 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
1636 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
1637 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
1638 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
1639 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
1640 (ror<mode>3_insn): Likewise.
1641 * config/aarch64/predicates.md (aarch64_simd_register): New.
1642
4ded8276
RB
16432013-08-27 Richard Biener <rguenther@suse.de>
1644
1645 PR tree-optimization/57521
1646 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
1647 one edge is non-critical.
1648 (find_phi_replacement_condition): Make sure to use a non-critical
1649 edge. Cleanup and remove old bug workarounds.
1650 (bb_postdominates_preds): Remove.
1651 (if_convertible_loop_p_1): Do not compute post-dominators.
1652 (combine_blocks): Do not free post-dominators.
1653 (main_tree_if_conversion): Likewise.
1654 (pass_data_if_conversion): Add TODO_verify_ssa.
1655
5aa11061
DD
16562013-08-27 DJ Delorie <dj@redhat.com>
1657
1658 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
1659
fd91cfe3
YZ
16602013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
1661
1662 * function.c (assign_parm_find_data_types): Set passed_mode and
1663 nominal_mode to the TYPE_MODE of nominal_type for the built
1664 pointer type in case of the struct-pass-by-reference.
1665
907555ce
JR
16662013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1667
1668 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
1669 (UINT16_TYPE): Change default to "unsigned int".
1670
67518c93
JR
1671 * config/avr/avr.opt (mfract-convert-truncate): New option.
1672 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
1673 is set, round negative fractional integers according to n1169
1674 when converting to integer types.
1675
d6d989fc
JH
16762013-08-26 Jan Hubicka <jh@suse.cz>
1677
1678 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
1679 methods can not be called indirectly when their address is not taken.
1680
2aa3da06
JH
16812013-08-26 Jan Hubicka <jh@suse.cz>
1682
12211b99
UB
1683 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
1684 ctor_for_folding.
2aa3da06 1685
0987ffe7
JH
16862013-08-26 Jan Hubicka <jh@suse.cz>
1687
1688 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
1689 can be unshared.
1690
befe8647
JR
16912013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1692
1693 * reload.c (find_valid_class): Allow classes that do not include
1694 FIRST_PSEUDO_REGISTER - 1.
1695
a21e735e
JH
16962013-08-26 Jan Hubicka <jh@suse.cz>
1697
1698 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
1699 fix edge count/frequency when speculation failed; fix type check
1700 for the direct call.
1701
e067bd43
JH
17022013-08-26 Jan Hubicka <jh@suse.cz>
1703
1704 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
1705
e8aec975
JH
17062013-08-26 Jan Hubicka <jh@suse.cz>
1707
1708 * ipa-inline-transform.c (inline_transform): Be ready for basic block
1709 to be changed by edge redirection.
1710
d0b66480
JH
17112013-08-26 Jan Hubicka <jh@suse.cz>
1712
12211b99
UB
1713 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
1714 formating; add sanity check.
d0b66480
JH
1715 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
1716 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
1717 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
1718 (resolve_noninline_speculation): Update callee keys, too.
1719
0f9aaac7
JH
17202013-08-26 Jan Hubicka <jh@suse.cz>
1721
1722 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
1723 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
1724
4b37444e
JR
17252013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
1726
1727 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
1728 into proper place.
1729
5339fc0c
UB
17302013-08-26 Uros Bizjak <ubizjak@gmail.com>
1731
1732 * config/i386/i386.c (ix86_debug_options): Remove prototype.
1733 (x86_64_elf_select_section): Ditto.
1734 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
1735 arguments.
1736 (ix86_pass_by_reference): Ditto.
5339fc0c
UB
1737 (output_set_got): Ditto.
1738 (ix86_unary_operator_ok): Ditto.
1739 (ix86_expand_builtin): Ditto.
1740
d5c3d3ef
JH
17412013-08-23 Jan Hubicka <jh@suse.cz>
1742
5339fc0c 1743 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 1744
5b1e7435
JH
17452013-08-23 Jan Hubicka <jh@suse.cz>
1746
1747 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
1748 (tree_decl_with_vis): Add FINAL field.
1749
ead69dac
JL
17502013-08-23 Jeff Law <law@redhat.com>
1751
1752 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
1753 the debugging dump when the expression is fully redundant.
1754
025311c4
GDR
17552013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
1756
1757 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
1758 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
1759 * pretty-print.c (pp_formatted_text_data): Likewise.
1760 (pp_write_text_to_stream): Likewise.
1761 (pp_write_text_as_dot_label_to_stream): Likewise.
1762 (pp_append_r): Likewise.
1763 (pp_format): Likewise.
1764 (pp_flush): Likewise.
1765 (pp_clear_output_area): Likewise.
1766 (pp_append_text): Likewise.
1767 (pp_formatted_text): Likewise.
1768 (pp_remaining_character_count_for_line): Likewise.
1769 (pp_newline): Likewise.
1770 (pp_character): Likewise.
1771 (output_buffer::~output_buffer): Define.
1772 (pretty_printer::~pretty_printer): Destruct output buffer.
1773 * pretty-print.h (output_buffer::~output_buffer): Declare.
1774 (pretty_printer::~pretty_printer): Declare virtual.
1775
0cadbfaa
MG
17762013-08-24 Marc Glisse <marc.glisse@inria.fr>
1777
1778 PR other/57324
1779 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
1780 HOST_WIDE_INT_M1U): New macros.
1781 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
1782 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
1783 unsigned -1 for lshift.
1784 * cse.c (cse_insn): Likewise.
1785 * double-int.c (rshift_double, lshift_double): Likewise.
1786 * builtins.c (fold_builtin_bitop): Likewise.
1787 * combine.c (force_to_mode): Likewise.
1788 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
1789 * simplify-rtx.c (simplify_const_unary_operation,
1790 simplify_const_binary_operation): Likewise.
1791 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
1792 check_va_list_escapes): Likewise.
1793 * rtlanal.c (nonzero_bits1): Likewise.
1794 * expmed.c (expand_smod_pow2): Likewise.
1795 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
1796
009e5353
JH
17972013-08-23 Jan Hubicka <jh@suse.cz>
1798
1799 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
1800 as having address taken.
1801
c4be6568
JH
18022013-08-23 Jan Hubicka <jh@suse.cz>
1803
64cbf23d
JH
1804 * ipa-utils.h (method_class_type): Declare.
1805 * ipa-devirt.c (method_class_type): Export.
1806
c4be6568
JH
1807 * cgraphunit.c (analyze_functions): Do basic devirtualization;
1808 do not walk base classes of anonymous types.
1809
97aba8e9
KK
18102013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
1811
1812 PR rtl-optimization/58220
1813 PR regression/58221
1814 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
1815 handle SEQUENCE insns properly.
1816
0fc80001
GDR
18172013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
1818
1819 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
1820 definition.
1821 (pp_newline_and_indent): Likewise.
1822 (pp_separate_with): Likewise.
1823 * pretty-print.c (pp_newline_and_flush): Define.
1824 (pp_newline_and_indent): Likewise.
1825 (pp_separate_with): Likewise.
1826
520a5868
JJ
18272013-08-23 Jakub Jelinek <jakub@redhat.com>
1828
1829 PR target/58218
1830 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
1831 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
1832
e500c62a
KY
18332013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
1834
1835 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
1836 * gcc/config/i386/i386.md (*movti_internal): Use
1837 predicate to determine if EVEX is needed.
1838 (*movsi_internal): Ditto.
1839 (*movdf_internal): Ditto.
1840 (*movsf_internal): Ditto.
1841 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
1842
6b00d7dd
JJ
18432013-08-23 Jakub Jelinek <jakub@redhat.com>
1844
1845 PR tree-optimization/58209
1846 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
1847 (find_tail_calls): Give up for pointer result types if m is non-NULL.
1848 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
1849 emit POINTER_PLUS_EXPR.
1850 (create_tailcall_accumulator): For pointer result type accumulate in
1851 sizetype type.
1852
4f2a9f90
PC
18532013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
1854
1855 * configure.ac: Add backslashes missing from the last change.
1856 * configure: Regenerate.
1857
7bf4274e
JH
18582013-08-22 Jan Hubicka <jh@susue.cz>
1859
4f2a9f90
PC
1860 * ipa.c (function_and_variable_visibility): First remember function
1861 was global and then make it local.
7bf4274e 1862
1bea243d
JB
18632013-08-22 Julian Brown <julian@codesourcery.com>
1864
1865 * configure.ac: Add aarch64 to list of arches which use "nop" in
1866 debug_line test.
1867 * configure: Regenerate.
1868
bcba7ecc
AK
18692013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1870
1871 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
1872 gnu_libc_has_function.
1873 * config/s390/tpf.h: Likewise.
1874
4c495b0d
JH
18752013-08-22 Jan Hubicka <jh@susue.cz>
1876
1877 * timevar.c (validate_phases): Add cast.
1878
5e302bca
JH
18792013-08-22 Jan Hubicka <jh@susue.cz>
1880
1881 * timevar.c (validate_phases): Use size_t for memory.
1882 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
1883
da6ca2b5
GDR
18842013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
1885
1886 * pretty-print.h (output_buffer::output_buffer): Declare.
1887 (pretty_printer::pretty_printer): Likewise.
1888 (pp_construct): Remove.
1889 * pretty-print.c (output_buffer::output_buffer): Define.
1890 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
1891 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
1892 (print_gimple_expr): Likewise.
1893 (print_gimple_seq): Likewise.
1894 (gimple_dump_bb): Likewise.
1895 * sched-vis.c (dump_value_slim): Likewise.
1896 (dump_insn_slim): Likewise.
1897 (dump_rtl_slim): Likewise.
1898 (str_pattern_slim): Likewise.
1899 * tree-mudflap.c (mf_varname_tree): Likewise.
1900 * graph.c (print_graph_cfg): Likewise.
1901 (start_graph_dump): Likewise.
1902 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
1903 placement-new.
1904 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
1905 pretty printer initialization.
1906 * coretypes.h (diagnostic_context): Remove superflous type alias
1907 declaration.
1908 (pretty_printer): Likewise. Declare directly as a class.
1909 (pretty_print_info): Remove declaration as class.
1910 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
1911 and pp_clear_output_area.
1912 (asan_add_global): Likewise.
1913
0e1474e5
JH
19142013-08-22 Jan Hubicka <jh@suse.cz>
1915
1916 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
1917 * ipa-utils.h (update_type_inheritance_graph): Declare.
1918 (possible_polymorphic_call_target_p): Declare.
1919 (possible_polymorphic_call_target_p): New.
1920 * ipa-devirt.c: Update toplevel comments.
1921 (cached_polymorphic_call_targets): Move up.
1922 (odr_type_d): Move ID down.
1923 (polymorphic_type_binfo_p): Update comment.
1924 (odr_hasher::remove): Likewise;
1925 (get_odr_type): Set anonymous_namespace.
1926 (dump_odr_type): Dump it.
1927 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
1928 (maybe_record_node): Record node in cached_polymorphic_call_targets.
1929 (record_binfo): Add comment.
5339fc0c
UB
1930 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
1931 built.
0e1474e5
JH
1932 (devirt_node_removal_hook): Do not iCE when cache is freed.
1933 (possible_polymorphic_call_target_p): New predicate.
1934 (update_type_inheritance_graph): New function.
1935
3f97cb0b
AI
19362013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
1937 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1938 Sergey Lega <sergey.s.lega@intel.com>
1939 Anna Tikhonova <anna.tikhonova@intel.com>
1940 Ilya Tocar <ilya.tocar@intel.com>
1941 Andrey Turetskiy <andrey.turetskiy@intel.com>
1942 Ilya Verbin <ilya.verbin@intel.com>
1943 Kirill Yukhin <kirill.yukhin@intel.com>
1944 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1945
1946 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
1947 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
1948 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
1949 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
1950 (OPTION_MASK_ISA_AVX2_UNSET): Update.
1951 (OPTION_MASK_ISA_AVX512F_UNSET): New.
1952 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
1953 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
1954 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
1955 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
1956 OPT_mavx512pf, OPT_mavx512er cases.
1957 * config/i386/constraints.md (v): New constraint.
1958 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
1959 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
1960 (bit_AVX512CD): New.
1961 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
1962 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
1963 * config/i386/i386-c.c (ix86_target_macros_internal):
1964 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
1965 __AVX512PF__.
1966 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
1967 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
1968 * config/i386/i386.c (regclass_map, dbx_register_map)
1969 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
1970 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
1971 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
1972 -mavx512pf options.
1973 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
1974 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
1975 -mavx512pf options. Fix formatting.
1976 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
1977 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
1978 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
1979 -mavx512cd, -mavx512pf options.
1980 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
1981 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
1982 (ix86_preferred_output_reload_class): Replace SSE_REGS with
1983 ALL_SSE_REGS.
1984 (ix86_hard_regno_mode_ok): Support 512-bit registers.
1985 (ix86_set_reg_reg_cost): Ditto.
1986 (x86_order_regs_for_local_alloc): Ditto.
1987 (MAX_VECT_LEN): Extend to 64-byte.
1988 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
1989 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
1990 (TARGET_AVX512ER, TARGET_AVX512CD): New.
1991 (BIGGEST_ALIGNMENT): Extend to 512-bits.
1992 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
1993 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
1994 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
1995 (SSE_REG_MODE_P): Support new modes.
1996 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
1997 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
1998 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
1999 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
2000 (REG_CLASS_CONTENTS): Add new registers.
2001 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
2002 (EXT_REX_SSE_REGNO_P): New.
2003 (HI_REGISTER_NAMES): Add new registers.
2004 * config/i386/i386.md: Define constants for new registers.
2005 (mode): Add new 512-bit modes.
2006 (prefix): Support evex prefix.
2007 (isa): Support avx512f, noavx512f, fma_avx512f.
2008 (ssemodesuffix): Add new 512-bit modes.
2009 (movxi): New.
2010 (*movxi_internal_avx512f): Ditto.
2011 (*movdi_internal): Replace constraint "x" with the new constraint "v".
2012 Support MODE_XI.
2013 (*movsi_internal): Likewise.
2014 (*movdf_internal): Likewise.
2015 (*movsf_internal): Likewise.
2016 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
2017 (<code><mode>3): Likewise.
5339fc0c
UB
2018 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
2019 New.
3f97cb0b
AI
2020 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
2021 with the new constraint "v".
2022 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
2023 modes.
2024 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
2025 with the new constraint "v".
2026 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
2027 (<sse2>_storedqu<avxsizesuffix>): Likewise.
2028 (*<plusminus_insn><mode>3): Likewise.
2029 (<sse>_vm<plusminus_insn><mode>3): Likewise.
2030 (*mul<mode>3): Likewise.
2031 (<sse>_vmmul<mode>3): Likewise.
2032 (<sse>_div<mode>3): Likewise.
2033 (<sse>_vmdiv<mode>3): Likewise.
2034 (<sse>_sqrt<mode>2): Likewise.
2035 (<sse>_vmsqrt<mode>2): Likewise.
2036 (*<code><mode>3_finite): Likewise.
2037 (*<code><mode>3) <smaxmin>: Likewise.
2038 (<sse>_vm<code><mode>3): Likewise.
2039 (*<code><mode>3) <any_logic>: Likewise.
2040 (*fma_fmadd_<mode>): Likewise.
2041 (*fma_fmsub_<mode>): Likewise.
2042 (*fma_fnmadd_<mode>): Likewise.
2043 (*fma_fnmsub_<mode>): Likewise.
2044 (*fma_fmaddsub_<mode>): Likewise.
2045 (*fma_fmsubadd_<mode>): Likewise.
2046 (*fmai_fmadd_<mode>): Likewise.
2047 (*fmai_fmsub_<mode>): Likewise.
2048 (*fmai_fnmadd_<mode>): Likewise.
2049 (*fmai_fnmsub_<mode>): Likewise.
2050 (sse_cvtsi2ss): Likewise.
2051 (sse_cvtsi2ssq): Likewise.
2052 (sse_cvtss2si): Likewise.
2053 (sse_cvtss2si_2): Likewise.
2054 (sse_cvtss2siq): Likewise.
2055 (sse_cvtss2siq_2): Likewise.
2056 (sse_cvttss2si): Likewise.
2057 (sse_cvtss2siq_2): Likewise.
2058 (float<sseintvecmodelower><mode>2): Likewise.
2059 (sse2_cvtsd2si_2): Likewise.
2060 (sse2_cvtsd2siq_2): Likewise.
2061 (*<plusminus_insn><mode>3): Likewise.
2062 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
2063 (*<sse4_1_avx2>_mul<mode>3): Likewise.
2064 (ashr<mode>3): Likewise.
2065 (<shift_insn><mode>3): Likewise.
2066 (avx2_<code><mode>3): Likewise.
2067 (*avx2_<code><mode>3): Likewise.
2068 (*andnot<mode>3): Likewise.
2069 (*<code><mode>3) <any_logic>: Likewise.
2070 (abs<mode>2): Likewise.
2071 (avx2_permvar<mode>): Likewise.
2072 (avx2_perm<mode>_1): Likewise.
2073 (*avx_vpermilp<mode>): Likewise.
2074 (avx_vpermilvar<mode>3): Likewise.
2075 (avx2_ashrv<mode>): Likewise.
2076 (avx2_<shift_insn>v<mode>): Likewise.
2077 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
2078 -mavx512cd.
2079 * doc/rtl.texi: Document XImode.
2080
dd5e8423
JL
20812013-08-21 Jeff Law <law@redhat.com>
2082
b9ebee5d
JL
2083 * tree-flow.h (register_jump_thread): Pass vector of edges
2084 instead of each important edge.
2085 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
2086 thread path into a vector and pass that to register_jump_thread.
2087 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
2088 passed in edge vector to the current 3-edge form.
2089
dd5e8423
JL
2090 Revert:
2091 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2092
9bb6628e 2093 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL
2094 delete_unmarked_insns removed anything.
2095
445dc8df
JR
20962013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
2097
2098 * reload.h (struct reg_equivs): Rename to ..
2099 (struct reg_equivs_s): .. this.
2100
e0df53dd
ML
21012013-08-20 Martin Liska <marxin.liska@gmail.com>
2102
2103 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
2104
e7d1d3eb
RO
21052013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2106
5339fc0c 2107 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 2108
5562e26e
JL
21092013-08-21 Jeff Law <law@redhat.com>
2110
2111 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
2112 simplify assignments too. If the RHS collapses to a singleton
2113 range, then return the value for the range.
2114
c7ecdec6
KY
21152013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
2116
2117 * config/i386/sse.md (V16): Rename to...
2118 (VMOVE): this.
2119 (mov<mode>): Update iterator name.
2120 (*mov<mode>_internal): Ditto.
2121 (push<mode>1): Ditto.
2122 (movmisalign<mode>): Ditto.
2123
bfa3b50a
JH
21242013-08-20 Jan Hubicka <jh@suse.cz>
2125
2126 PR bootstrap/58186
2127 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
2128 entry for direct edges.
2129 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
2130
6cd4d135
DM
21312013-08-20 David Malcolm <dmalcolm@redhat.com>
2132
2133 Revert my last two changes, r201865 and r201864:
2134
2135 Revert r201865:
2136 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2137
2138 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2139 instances can own GC refs.
2140
2141 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2142 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2143 (gcc::context::gt_pch_nx): Likewise.
2144 (gcc::context::gt_pch_nx): Likewise.
2145 * ggc.h (gt_ggc_mx <T>): New.
2146 (gt_pch_nx_with_op <T>): New.
2147 (gt_pch_nx <T>): New.
2148 * passes.c (opt_pass::gt_ggc_mx): New.
2149 (opt_pass::gt_pch_nx): New.
2150 (opt_pass::gt_pch_nx_with_op): New.
2151 (pass_manager::gt_ggc_mx): New.
2152 (pass_manager::gt_pch_nx): New.
2153 (pass_manager::gt_pch_nx_with_op): New.
2154 (pass_manager::operator new): Use
2155 ggc_internal_cleared_alloc_stat rather than xcalloc.
2156 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2157 (pass_manager::gt_ggc_mx): New.
2158 (pass_manager::gt_pch_nx): New.
2159 (pass_manager::gt_pch_nx_with_op): New.
2160 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2161 (opt_pass::operator new): New.
2162 (opt_pass::gt_ggc_mx): New.
2163 (opt_pass::gt_pch_nx): New.
2164 (opt_pass::gt_pch_nx_with_op): New.
2165
2166 Revert r201864:
2167 2013-08-20 David Malcolm <dmalcolm@redhat.com>
2168
2169 * Makefile.in (GTFILES): Add context.h.
2170 * context.c (gcc::context::operator new): New.
2171 (gcc::context::gt_ggc_mx): New.
2172 (gcc::context::gt_pch_nx): New.
2173 (gcc::context::gt_pch_nx): New.
2174 * context.h (gcc::context): Add GTY((user)) marking.
2175 (gcc::context::operator new): New.
2176 (gcc::context::gt_ggc_mx): New.
2177 (gcc::context::gt_pch_nx): New.
2178 (gcc::context::gt_pch_nx): New.
2179 (g): Add GTY marking.
2180 (gt_ggc_mx (gcc::context *)): New.
2181 (gt_pch_nx (gcc::context *)): New.
2182 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2183 void *cookie)): New.
2184 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2185
a046136a
AM
21862013-08-20 Alexey Makhalov <makhaloff@gmail.com>
2187
2188 * dce.c (fini_dce): Call df_analyze again just in case
2189 delete_unmarked_insns removed anything.
2190
67598720
TJ
21912013-08-20 Teresa Johnson <tejohnson@google.com>
2192
2193 PR rtl-optimizations/57451
2194 * final.c (reemit_insn_block_notes): Prevent lexical blocks
2195 from crossing split section boundaries.
2196
f8693faf
MGD
21972013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
2198
2199 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
2200 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
2201 with MULTLIB_DEFAULTS.
2202
7d0b9a9c
NC
22032013-08-20 Nick Clifton <nickc@redhat.com>
2204
2205 * target.def (narrow_volatile_bitfield): Note that the default
2206 value is false, not !TARGET_STRICT_ALIGN.
2207 * doc/tm.texi: Regenerate.
2208
8edb8dc8
PC
22092013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
2210
2211 Fix LIB_SPEC for systems without libpthread.
2212
2213 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
2214 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
2215 for Android.
2216 * config/i386/linux-common.h: Likewise.
2217 * config/mips/linux-common.h: Likewise.
2218
12211b99 22192013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ
2220
2221 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
2222 checks.
2223
e42c64cb
DM
22242013-08-20 David Malcolm <dmalcolm@redhat.com>
2225
2226 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
2227 instances can own GC refs.
2228
2229 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
2230 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
2231 (gcc::context::gt_pch_nx): Likewise.
2232 (gcc::context::gt_pch_nx): Likewise.
2233 * ggc.h (gt_ggc_mx <T>): New.
2234 (gt_pch_nx_with_op <T>): New.
2235 (gt_pch_nx <T>): New.
2236 * passes.c (opt_pass::gt_ggc_mx): New.
2237 (opt_pass::gt_pch_nx): New.
2238 (opt_pass::gt_pch_nx_with_op): New.
2239 (pass_manager::gt_ggc_mx): New.
2240 (pass_manager::gt_pch_nx): New.
2241 (pass_manager::gt_pch_nx_with_op): New.
2242 (pass_manager::operator new): Use
2243 ggc_internal_cleared_alloc_stat rather than xcalloc.
2244 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
2245 (pass_manager::gt_ggc_mx): New.
2246 (pass_manager::gt_pch_nx): New.
2247 (pass_manager::gt_pch_nx_with_op): New.
2248 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
2249 (opt_pass::operator new): New.
2250 (opt_pass::gt_ggc_mx): New.
2251 (opt_pass::gt_pch_nx): New.
2252 (opt_pass::gt_pch_nx_with_op): New.
2253
910c02a0
DM
22542013-08-20 David Malcolm <dmalcolm@redhat.com>
2255
2256 * Makefile.in (GTFILES): Add context.h.
2257 * context.c (gcc::context::operator new): New.
2258 (gcc::context::gt_ggc_mx): New.
2259 (gcc::context::gt_pch_nx): New.
2260 (gcc::context::gt_pch_nx): New.
2261 * context.h (gcc::context): Add GTY((user)) marking.
2262 (gcc::context::operator new): New.
2263 (gcc::context::gt_ggc_mx): New.
2264 (gcc::context::gt_pch_nx): New.
2265 (gcc::context::gt_pch_nx): New.
2266 (g): Add GTY marking.
2267 (gt_ggc_mx (gcc::context *)): New.
2268 (gt_pch_nx (gcc::context *)): New.
2269 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
2270 void *cookie)): New.
2271 * gengtype.c (open_base_files) <ifiles>: Add context.h.
2272
af4b4236
AM
22732013-08-20 Alan Modra <amodra@gmail.com>
2274
2275 PR target/57865
2276 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
2277 (rs6000_emit_epilogue): Likewise.
2278
12211b99 22792013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC
2280
2281 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
2282
e2323f5b
PB
22832013-08-19 Peter Bergner <bergner@vnet.ibm.com>
2284 Jakub Jelinek <jakub@redhat.com>
2285
2286 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
2287 (BUILT_IN_FABSD64): Likewise.
2288 (BUILT_IN_FABSD128): Likewise.
2289 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
2290 (fold_builtin_1): Likewise.
5339fc0c
UB
2291 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
2292 destination and source operands.
e2323f5b
PB
2293 (*abstd2_fpr): Likewise.
2294 (*nabstd2_fpr): Likewise.
2295
cfbf3ee8
RS
22962013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
2297
2298 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
2299 JUMP_P and INSN_P.
2300
d3136aeb
AH
23012013-08-19 Aldy Hernandez <aldyh@redhat.com>
2302
2303 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
2304 incomplete.
2305
d33d9e47
AI
23062013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
2307
2308 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
2309 * builtins.c (default_libc_has_function): New.
2310 (gnu_libc_has_function): Ditto.
2311 (no_c99_libc_has_function): Ditto.
2312 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
2313 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
2314 (fold_builtin_sincos): Likewise.
2315 (fold_builtin_cexp): Likewise.
2316 * builtins.def (DEF_C94_BUILTIN): Likewise.
2317 (DEF_C99_BUILTIN): Likewise.
2318 (DEF_C99_C90RES_BUILTIN): Likewise.
2319 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
2320 definitions to using this define.
2321 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 2322 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI
2323 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
2324 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
2325 * config/darwin.h: Ditto.
2326 * config/elfos.h: Ditto.
2327 * config/freebsd.h: Ditto.
2328 * config/i386/cygming.h: Ditto.
2329 * config/i386/djgpp.h: Ditto.
2330 * config/i386/i386-interix.h: Ditto.
2331 * config/microblaze/microblaze.h: Ditto.
2332 * config/mmix/mmix.h: Ditto.
2333 * config/gnu-user.h: Ditto.
2334 * config/ia64/hpux.h: Ditto.
2335 * config/pa/pa-hpux.h: Ditto.
2336 * config/pdp11/pdp11.h: Ditto.
2337 * config/picochip/picochip.h: Ditto.
2338 * config/linux.h: Ditto.
2339 * config/netbsd.h: Ditto.
2340 * config/openbsd.h: Ditto.
2341 * config/rs6000/aix43.h: Ditto.
2342 * config/rs6000/aix51.h: Ditto.
2343 * config/rs6000/aix52.h: Ditto.
2344 * config/rs6000/aix53.h: Ditto.
2345 * config/rs6000/aix61.h: Ditto.
2346 * config/rs6000/darwin.h: Ditto.
2347 * config/rs6000/linux.h: Ditto.
2348 * config/rs6000/linux64.h: Ditto.
2349 * config/s390/tpf.h: Ditto.
2350 * config/sol2-10.h: Ditto.
2351 * config/sol2.h: Ditto.
2352 * config/vms/vms.h: Ditto.
2353 * config/vxworks.h: Ditto.
2354 * config/linux-android.c (linux_android_libc_has_function):
2355 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
2356 * config/linux-protos.h (linux_android_libc_has_function):
2357 New declaration.
2358 * config/i386/i386.c (ix86_libc_has_function): New.
2359 * config/i386/i386-protos.h
2360 (ix86_libc_has_function): New declaration.
2361 * config/i386/i386.md
2362 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
2363 ("isinf<mode>2): Likewise.
2364 * convert.c (convert_to_integer): Using new target hook
2365 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
2366 TARGET_C99_FUNCTIONS.
2367 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
2368 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
2369 * coretypes.h (function_class): New enum for different
2370 classes of functions.
2371 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
2372 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
2373 (TARGET_HAS_SINCOS): Likewise.
2374 (TARGET_LIBC_HAS_FUNCTION): New.
2375 * doc/tm.texi: Regenerated.
2376 * targhooks.h (default_libc_has_function): New declaration.
2377 (no_c99_libc_has_function): Ditto.
2378 (gnu_libc_has_function): Ditto.
2379 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
2380 and TARGET_HAS_SINCOS.
2381
eefe9a99
JH
23822013-08-18 Jan Hubicka <jh@suse.cz>
2383
2384 * Makeifle-in (ipa-devirt.o): New.
2385 (GTFILES): Add ipa-utils.h and ipa-devirt.c
2386 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
2387 (analyze_functions): Look into possible targets of polymorphic call.
2388 * dumpfile.c (dump_files): Add type-inheritance dump.
2389 * dumpfile.h (TDI_inheritance): New.
2390 * ipa-devirt.c: New file.
2391 * ipa-utils.h (odr_type_d): Forward declare.
2392 (odr_type): New type.
2393 (build_type_inheritance_graph): Declare.
2394 (possible_polymorphic_call_targets): Declare and introduce inline
2395 variant when only edge is pased.
2396 (dump_possible_polymorphic_call_targets): Likewise.
2397 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
2398 * tree.c (type_in_anonymous_namespace_p): Break out from ...
2399 (types_same_for_odr): ... here.
2400 * tree.h (type_in_anonymous_namespace_p): Declare.
2401
29b89442
JJ
24022013-08-18 Jakub Jelinek <jakub@redhat.com>
2403
2404 PR tree-optimization/58006
2405 * tree-parloops.c (take_address_of): Don't ICE if get_name
2406 returns NULL.
2407 (eliminate_local_variables_stmt): Remove clobber stmts.
2408
4f219961
EB
24092013-08-18 Eric Botcazou <ebotcazou@adacore.com>
2410
2411 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
2412 error message is issued for an alias to undefined symbol.
2413
1d5755ef
JH
24142013-08-18 Jan Hubicka <jh@suse.cz>
2415
2416 * cgraph.c (cgraph_create_indirect_edge): Discover
2417 polymorphic calls and record basic info into indirect_info.
2418 * gimple-fold.c (gimple_fold_call): When doing BINFO based
2419 devirtualization, ignore objc function calls.
2420 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
2421 call with no parm index info.
2422 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
2423 * tree.c (virtual_method_call_p): New function.
2424 * tree.h (virtual_method_call_p): Declare.
2425
5945bebf
JH
24262013-08-16 Jan Hubicka <jh@suse.cz>
2427
2428 PR middle-end/58179
2429 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
2430
47fa5d23
DE
24312013-08-16 David Edelsohn <dje.gcc@gmail.com>
2432
2433 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
2434 attribute.
2435
c74f54a0
DM
24362013-08-16 David Malcolm <dmalcolm@redhat.com>
2437
2438 * gengtype.c (type_for_name): Add special-case support for
2439 locating types within the "gcc::" namespace.
2440 (open_base_files): Emit a "using namespace gcc" directive.
2441
f758f299
MM
24422013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2443
2444 PR target/58160
2445 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
2446 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
2447
2448 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
2449 array instead of each individual operand as a separate argument.
2450 (emit_fusion_gpr_load): Likewise.
2451 (expand_fusion_gpr_load): Add new function declaration.
2452
2453 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
2454 signature to have the operands passed as an array, instead of as
2455 separate arguments. Allow ZERO_EXTEND to be in the memory
2456 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
2457 depend on the register live/dead flags when peepholes are run.
2458 (expand_fusion_gpr_load): New function to be called from the
2459 peephole2 pass, to change the register that addis sets to be the
2460 target register.
2461 (emit_fusion_gpr_load): Change the calling signature to have the
2462 operands passed as an array, instead of as separate arguments.
2463 Allow ZERO_EXTEND to be in the memory address, and also
2464 SIGN_EXTEND if -mpower8-fusion-sign.
2465
2466 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
2467 unspec enumeration.
2468 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
2469 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 2470 depend on the register live/dead state when the peephole pass is done.
f758f299 2471
158f4e4f
DM
24722013-08-16 David Malcolm <dmalcolm@redhat.com>
2473
2474 * gengtype.c (create_user_defined_type): Ensure that the kind
2475 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
2476 declaration is seen before the GTY((user)) marking.
2477
22f8cea5
BE
24782013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
2479
2480 PR target/58105
2481 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
2482
c49bdb2e
JH
24832013-08-16 Jan Hubicka <jh@suse.cz>
2484
2485 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
2486 arugment expected_type.
2487 (gimple_fold_call): Use it.
2488 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
2489 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 2490 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH
2491 (try_make_edge_direct_virtual_call): Likewise.
2492 * tree.c (obj_type_ref_class): New.
2493 * tree.h (obj_type_ref_class): Use it.
2494
4042dca9
GDR
24952013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
2496
2497 * sched-vis.c (rtl_slim_pp_initialized): Remove.
2498 (rtl_slim_pp): Likewise.
2499 (init_rtl_slim_pretty_print): Likewise.
2500 (dump_value_slim): Don't call it. Use local pretty printer.
2501 (dump_insn_slim): Likewise.
2502 (dump_rtl_slim): Likewise.
2503 (str_pattern_slim): Likewise.
2504 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
2505 Simplify.
2506
7a460594
JJ
25072013-08-16 Jakub Jelinek <jakub@redhat.com>
2508
639dc669
JJ
2509 PR tree-optimization/58164
2510 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
2511 walk gimple_goto_dest of GIMPLE_GOTO.
2512
7a460594
JJ
2513 PR tree-optimization/58165
2514 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
2515 bi_call must be the last stmt in a bb, don't split_block, instead
2516 use fallthru edge from it and give up if there is none.
2517 Release conds vector when returning early.
2518
e3c02cc7
XDL
25192013-08-14 Xinliang David Li <davidxl@google.com>
2520
2521 * config/i386/i386.c (ix86_option_override_internal):
2522 Remove unused variable and field.
2523
6469da2c
BS
25242013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2525
2526 PR target/57949
22f8cea5 2527 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS
2528 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
2529 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
2530 and Linux, correct BLKmode alignment when 128-bit alignment is
2531 required and compatibility flag is not set.
22f8cea5
BE
2532 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
2533 for zero-size arguments when compatibility flag is not set.
6469da2c 2534
4ca890e2
JJ
25352013-08-14 Jakub Jelinek <jakub@redhat.com>
2536
2537 PR tree-optimization/58145
2538 * tree-sra.c (build_ref_for_offset): If prev_base has
2539 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
2540
daec1759 25412013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 2542
daec1759 2543 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 2544 Fix uninitialized variable error.
daec1759 2545
3ad20bd4
XDL
25462013-08-14 Xinliang David Li <davidxl@google.com>
2547
2548 * config/i386/i386.opt: Define two new options.
2549 * config/i386/x86-tune.def: Add arch selector field in macros.
2550 * config/i386/i386.h: Adjust macro definition.
2551 * config/i386/i386.c (ix86_option_override_internal):
2552 Refactor the code.
2553 (parse_mtune_ctrl_str): New function.
2554 (set_ix86_tune_features): New function.
2555 (ix86_function_specific_restore): Call the new helper function.
2556
e839e2a9
AB
25572013-08-14 Andrey Belevantsev <abel@ispras.ru>
2558
2559 PR rtl-optimization/57662
2560 * sel-sched.c (code_motion_process_successors): When the current insn
2561 is removed after the recursive traversal, break from the loop.
2562 Add comments and debug printouts.
2563
27ed665f
JJ
25642013-08-14 Jakub Jelinek <jakub@redhat.com>
2565 Alexandre Oliva <aoliva@redhat.com>
2566
2567 PR target/58067
2568 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
2569 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
2570 there also UNSPEC_PLTOFF.
2571
61c1a609
MP
25722013-08-14 Marek Polacek <polacek@redhat.com>
2573
2574 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
2575 NUM_CONDITIONS bit positions.
2576
0e901c67
CC
25772013-08-13 Cary Coutant <ccoutant@google.com>
2578
2579 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
2580 (attr_checksum): Hash vector contents instead of pointer.
2581 (attr_checksum_ordered): Likewise.
2582
061eff6d
UB
25832013-08-13 Uros Bizjak <ubizjak@gmail.com>
2584
2585 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
2586 when Pmode != word_mode. Add length_address attribute.
2587 (sse3_monitor_<mode>): Merge from sse3_monitor and
2588 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
2589 Pmode != word_mode. Update insn length attribute.
2590 * config/i386/i386.c (ix86_option_override_internal): Update
2591 ix86_gen_monitor selection for merged sse3_monitor insn.
2592
8ec9249b
JB
25932013-08-13 Julian Brown <julian@codesourcery.com>
2594
2595 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
2596 perform invalid legitimization on greater-than-word-size modes for
2597 TARGET_E500_DOUBLE.
2598
761a8eb7
VM
25992013-08-13 Vladimir Makarov <vmakarov@redhat.com>
2600
2601 * ira.c (setup_class_translate_array): Use aclass instead of cl
2602 for classes not fully covered by allocno classes.
2603
99c2bd54
JJ
26042013-08-13 Jakub Jelinek <jakub@redhat.com>
2605
4029a5e0
JJ
2606 PR tree-optimization/57661
2607 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
2608 * tree-inline.c (tree_function_versioning): Initialize it.
2609 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
2610 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
2611 that is not being copied.
2612
99c2bd54
JJ
2613 PR sanitizer/56417
2614 * asan.c (instrument_strlen_call): Fix typo in comment.
2615 Use char * type even for the lhs of POINTER_PLUS_EXPR.
2616
b5f8f063
SE
26172013-08-13 Steve Ellcey <sellcey@mips.com>
2618
2619 * config/mips/mips.md (prefetch): Use lw instead of ld on
2620 loongson in 32bit mode.
2621
4de80584
NC
26222013-08-13 Nick Clifton <nickc@redhat.com>
2623
2624 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
2625
09ce3660
JH
26262013-08-13 Jan Hubicka <jh@suse.cz>
2627
2628 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
2629 introduced edge; fix typo in sanity check.
2630 (cgraph_resolve_speculation): Export; improve diagnostic.
2631 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
2632 speculation at type mismatch.
2633 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
2634 (cgraph_resolve_speculation): Declare.
2635 (symtab_can_be_discarded): New function.
2636 * value-prof.c (gimple_ic_transform): Remove actual transform code.
2637 * ipa-inline-transform.c (speculation_removed): New global var.
2638 (clone_inlined_nodes): See if speculation can be removed.
2639 (inline_call): If speculations was removed, we growths may not match.
2640 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
2641 (speculation_useful_p): New function.
2642 (resolve_noninline_speculation): New function.
2643 (inline_small_functions): Resolve useless speculations.
2644 * ipa-inline.h (speculation_useful_p): Declare
2645 * ipa.c (can_replace_by_local_alias): Simplify.
2646 (ipa_profile): Produce speculative calls in non-lto, too;
2647 add simple cost model; produce local aliases.
2648
f971dc24
DM
26492013-08-13 David Malcolm <dmalcolm@redhat.com>
2650
2651 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 2652 PASS_MANAGER_H.
f971dc24 2653
2e507b9a
PC
26542013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
2655
2656 * config/i386/i386.c (ix86_function_versions): Use error + inform.
2657
c97fd4b8
UB
26582013-08-12 Uros Bizjak <ubizjak@gmail.com>
2659
2660 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
2661 iterator instead of X87MODEF.
2662
26632013-08-12 Perez Read <netfirewall@gmail.com>
2664
2665 PR target/58132
2666 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
2667 operand 0 for intel asm alternative.
2668 (*movabs<mode>_2): Ditto for operand 1.
2669
f4f4204c
JG
26702013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
2671
2672 * config/aarch64/arm_none.h
2673 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
2674
133b1a8e
NC
26752013-08-12 Nick Clifton <nickc@redhat.com>
2676
2677 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
2678
15dd8b3a
YR
26792013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
2680
2681 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
2682 expand for QI/HImode operand to produce more effictive code for
2683 unsigned char(short) --> float(double) conversion.
2684
94109a6a
AM
26852013-08-12 Alexander Monakov <amonakov@ispras.ru>
2686
2687 * doc/invoke.texi: Mention that -ftls-model does not force the final
2688 model.
2689
4e115102
MP
26902013-08-12 Marek Polacek <polacek@redhat.com>
2691 Marc Glisse <marc.glisse@inria.fr>
2692
2693 PR tree-optimization/57980
2694 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
2695 when creating -1 constant.
2696
748d71f3
JH
26972013-08-10 Jan Hubicka <jh@suse.cz>
2698
1f97e193 2699 Workaround binutils PR14342.
748d71f3
JH
2700 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
2701 (gimple_init_edge_profiler): Likewise.
2702 (gimple_gen_ic_func_profiler): Likewise.
2703
5979aa54
JH
27042013-08-09 Jan Hubicka <jh@suse.cz>
2705
2706 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
2707
ad83025e
XDL
27082013-08-09 Xinliang David Li <davidxl@google.com>
2709
2710 * config/i386/stringop.def: New file.
2711 * config/i386/stringop.opt: New file.
2712 * config/i386/i386-opts.h: Include stringopt.def.
2713 * config/i386/i386.opt: Include stringopt.opt.
2714 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 2715 Override default size based stringop inline strategies with options.
ad83025e
XDL
2716 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
2717 New function.
2718
e59df5fd
JH
27192013-08-09 Jan Hubicka <jh@suse.cz>
2720
c97fd4b8 2721 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 2722
634ab819
JH
27232013-08-09 Jan Hubicka <jh@suse.cz>
2724
2725 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
2726 CGRAPH_FREQ_MAX.
2727 (dump_cgraph_node): Dump profile-id.
2728 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
2729 and common_target_probability.
2730 * lto-cgraph.c (lto_output_edge): Stream common targets.
2731 (lto_output_node): Stream profile ids.
2732 (input_node): Stream profile ids.
2733 (input_edge): Stream common targets.
2734 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
2735 * ipa.c: Include value-prof.h
2736 (ipa_profile_generate_summary): Turn indirect call statement histograms
2737 into common targets.
2738 (ipa_profile): Turn common targets into speculative edges.
2739
2fa3d31b
JH
27402013-08-09 Jan Hubicka <jh@suse.cz>
2741
2742 * cgraph.h (cgraph_node): Add profile_id.
2743 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 2744 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH
2745 (del_node_map): Update.
2746 (find_func_by_funcdef_no): Replace by ...
2747 (find_func_by_profile_id): ... this one.
2748 (gimple_ic_transform): Do not remove useful histograms when
2749 speculation is not done; dump info when indirect call removal
2750 can happen at LTO.
2751 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
2752 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
2753 (__gcov_indirect_call_profiler_v2): .. this one.
2754 * profile.h (init_node_map): Update.
2755 * coverage.c (coverage_compute_profile_id): New function.
2756 * coverage.h (coverage_compute_profile_id): Declare.
2757 * tree-profile.c (init_ic_make_global_vars): Make
2758 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
2759 (gimple_init_edge_profiler): Update prototype of
2760 __gcov_indirect_call_profiler.
2761 (gimple_gen_ic_func_profiler): Simplify.
2762 (tree_profiling): Use init_node_map
2763
042ae7d2
JH
27642013-08-09 Jan Hubicka <jh@suse.cz>
2765
c97fd4b8
UB
2766 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
2767 non-speculative refs.
042ae7d2
JH
2768 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
2769 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
2770 (cgraph_set_call_stmt): Likewise.
2771 (cgraph_create_edge_1): Fix release checking compilatoin;
2772 clear lto_stmt_uid.
2773 (cgraph_free_edge): Free indirect info.
2774 (cgraph_turn_edge_to_speculative): New function.
2775 (cgraph_speculative_call_info): New function.
2776 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 2777 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH
2778 (dump_cgraph_node): Dump speculation.
2779 (verify_edge_count_and_frequency): Accept speculative edges.
2780 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
2781 (verify_cgraph_node): Handle speculation.
2782 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
2783 (cgraph_set_call_stmt): Update prototype.
2784 (cgraph_make_edge_direct): Update prototype.
2785 (cgraph_speculative_call_info): Declare.
2786 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
2787 to change; update call of ipa_find_references.
2788 * ipa-ref.c (ipa_record_reference): Fix return value; clear
2789 lto_stmt_uid and speculative flags.
2790 (ipa_dump_references): Dump speculation.
2791 (ipa_clone_references): Clone speculative flag.
2792 (ipa_clone_referring): Likewise.
2793 (ipa_clone_ref): New function.
2794 (ipa_find_reference): Look into lto_stmt_uids
2795 (ipa_clear_stmts_in_references): Do not clear speculative calls.
2796 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
2797 (ipa_find_reference): Update declaration.
2798 (ipa_clone_ref): Declare.
2799 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
2800 stream speculative flag.
2801 (lto_output_ref): Stream statements uids and speculation.
2802 (input_ref): Likewise.
2803 (input_edge): Stream speuclation.
2804 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
2805 (cgraph_set_call_stmt_including_clones): Handle speculation.
2806 * ipa-inline.c (heap_edge_removal_hook): New function.
2807 (inline_small_functions): Register it.
2808 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
2809 also initialize refs.
2810 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
2811 edge to change.
2812 (try_make_edge_direct_simple_call): Likewise.
2813 (try_make_edge_direct_simple_call): Likewise.
2814 (update_indirect_edges_after_inlining): Likewise.
2815 (remove_described_reference): Look proper lto_stmt_uid.
2816 (propagate_controlled_uses): Likewise.
2817 (propagate_controlled_uses): Liekwise.
2818 * tree-inline.c (copy_bb): Copy speculative edges.
2819 (redirect_all_calls): New function.
c97fd4b8
UB
2820 (copy_cfg_body): Do redirection after loop info is updated.
2821 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 2822
aa1e10cc
JH
28232013-08-09 Jan Hubicka <jh@suse.cz>
2824
2825 * lto-streamer-out.c (output_function): Renumber PHIs.
2826 * lto-streamer-in.c (input_function): Likewise.
2827
66adb8eb
JG
28282013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
2829
2830 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
2831 (get_lane_unsigned): Likewise.
2832 (dup_lane_scalar): Likewise.
2833 (get_lane): enable for VALL.
2834 * config/aarch64/aarch64-simd.md
2835 (aarch64_dup_lane_scalar<mode>): Remove.
2836 (aarch64_get_lane_signed<mode>): Likewise.
2837 (aarch64_get_lane_unsigned<mode>): Likewise.
2838 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
2839 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
2840 (aarch64_get_lane<mode>): Enable for all vector modes.
2841 (aarch64_get_lanedi): Remove misleading constraints.
2842 * config/aarch64/arm_neon.h
2843 (__aarch64_vget_lane_any): Define.
2844 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
2845 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
2846 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
2847 * config/aarch64/iterators.md (VDQQH): New.
2848 (VDQQHS): Likewise.
2849 (vwcore): Likewise.
2850
1f65ae7a
EB
28512013-08-09 Eric Botcazou <ebotcazou@adacore.com>
2852
2853 * configure.ac: Add GAS check for LEON instructions on SPARC.
2854 * configure: Regenerate.
2855 * config.in: Likewise.
2856 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
2857 sparc*-*-* block.
2858 * config/sparc/sparc.opt (LEON, LEON3): New masks.
2859 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
2860 for LEON or LEON3.
2861 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
2862 (AS_LEON_FLAG): New macro.
2863 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
2864 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
2865 Deal with LEON and LEON3 for the memory model.
23a6cb78 2866 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB
2867 (atomic_compare_and_swap<mode>_1): Likewise.
2868 (*atomic_compare_and_swap<mode>_1): Likewise.
2869
ff522f7f
ZC
28702013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2871
2872 * config/arm/neon.md (vcond): Fix floating-point vector
2873 comparisons against 0.
2874
cb1cca12
VM
28752013-08-08 Vladimir Makarov <vmakarov@redhat.com>
2876
2877 * lra-constraints.c (emit_spill_move): Remove assert.
2878 (process_alt_operands): Add more debugging
2879 output. Increase reject for spilling into memory. Decrease
2880 reject for reloading scratch.
2881 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
2882
a3719e31
SE
28832013-08-08 Steve Ellcey <sellcey@mips.com>
2884
2885 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
2886 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
2887 micromips incompatible. Add nan2008.
2888 (MULTILIB_DIRNAMES): Add nan2008.
2889 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
2890 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
2891 and micromips incompatible. Add nan2008.
2892 (MULTILIB_DIRNAMES): Add nan2008.
2893 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
2894
f40423e2
RS
28952013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
2896
2897 PR rtl-optimization/58079
2898 * combine.c (combine_simplify_rtx): Avoid using SUBST if
2899 simplify_comparison has widened a comparison with an integer.
2900
869b9125
KT
29012013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2902
2903 * config/arm/neon.md (movmisalign<mode>): Disable when we
2904 don't allow unaligned accesses.
2905 (*movmisalign<mode>_neon_store): Likewise.
2906 (*movmisalign<mode>_neon_load): Likewise.
2907 (*movmisalign<mode>_neon_store): Likewise.
2908 (*movmisalign<mode>_neon_load): Likewise.
2909
71cafea9
JH
29102013-08-08 Jan Hubicka <jh@suse.cz>
2911
2912 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
2913 (make_pass_rebuild_cgraph_edges): Also clear references.
2914 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
2915 * ipa-inline-transform.c (inline_transform): Remove all references
2916 after inlining.
c97fd4b8
UB
2917 * cgraphunit.c (expand_function): Remove all references after
2918 expansion.
71cafea9
JH
2919 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
2920 (ipa_find_reference): Rewrite to iterator.
2921 (remove_stmt_references): Likewise.
2922 (ipa_clear_stmts_in_references): New function.
2923 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB
2924 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
2925 clear references.
71cafea9
JH
2926 * ipa-split.c (split_function): Remove references in split function.
2927
27d2e612
RE
29282013-08-08 Richard Earnshaw <rearnsha@arm.com>
2929
2930 PR target/57431
1f65ae7a 2931 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE
2932 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
2933
5efc661c
RE
29342013-08-08 Richard Earnshaw <rearnsha@arm.com>
2935
2936 PR target/56979
1f65ae7a 2937 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE
2938 suggested mode for the assignment isn't compatible with the
2939 registers required.
2940
f276d31d
BE
29412013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2942
2943 PR target/58065
2944 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
2945
4b8bc035
XDL
29462013-08-07 Xinliang David Li <davidxl@google.com>
2947
2948 * config/i386/i386.opt: New option -mtune-ctrl=.
2949 * config/i386/x86-tune.def: New file.
2950 * config/i386/i386.h: include x86-tune.def.
2951 * config/i386/i386.c (ix86_option_override_internal):
2952 Parsing -mtune-ctrl= option and set tune features.
2953
12211b99 29542013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE
2955
2956 PR other/12081
03b0ee0a 2957 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE
2958 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
2959 to gen_2arg_fn_t.
2960
5cf6635b
EB
29612013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2962
2963 * rtl.h (update_alignments): Declare.
2964 * final.c (grow_label_align): New function extracted from...
2965 (shorten_branches): ...here. Call it.
2966 (update_alignments): New function.
2967 * reorg.c (sibling_labels): New variable.
2968 (get_label_before): Add SIBLING parameter. If it is non-zero, push
2969 the new label along with it onto the sibling_labels vector.
2970 (fill_simple_delay_slots): Adjust call to get_label_before.
2971 (fill_slots_from_thread): Likewise.
2972 (relax_delay_slots): Likewise.
2973 (make_return_insns): Likewise.
2974 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
2975
3c8ca1ab
EB
29762013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2977
2978 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
2979 document its semantics.
2980 (diagnostic_report_diagnostic): Adjust accordingly.
2981
5ee5b32c
DM
29822013-08-07 David Malcolm <dmalcolm@redhat.com>
2983
03b0ee0a 2984 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM
2985 (sparc_option_override): ...and port to new C++ pass API.
2986 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
2987
abd566fa
PB
29882013-08-07 Peter Bergner <bergner@vnet.ibm.com>
2989
2990 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
2991
2077db1b
CT
29922013-08-06 Caroline Tice <cmtice@google.com>
2993
2994 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
2995 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
2996 * tree-pass.h: Add pass_vtable_verify.
2997 * varasm.c (assemble_variable): Add code to properly set the comdat
2998 section and name for the .vtable_map_vars section.
2999 (assemble_vtyv_preinit_initializer): New function.
3000 (default_sectin_type_flags): Make sure .vtable_map_vars section has
3001 LINK_ONCE flag.
3002 * output.h: Add function decl for assemble_vtv_preinit_initializer.
3003 * vtable-verify.c: New file.
3004 * vtable-verify.h: New file.
3005 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
3006 initialiation levels.
3007 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
3008 * passes.def: Insert pass_vtable_verify.
3009 * aclocal.m4: Reorder includes.
3010 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
3011 -fvtv-counts options.
3012 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
3013 as appropriate, if -fvtable-verify=... is used.
3014 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
3015 -fvtable-verify=... is used.
3016 * Makefile.in (OBJS): Add vtable-verify.o to list.
3017 (vtable-verify.o): Add new build rule.
3018 (GTFILES): Add vtable-verify.c to list.
3019 * common.opt (fvtable-verify=): New flag.
3020 (vtv_priority): Values for fvtable-verify= flag.
3021 (fvtv-counts): New flag.
3022 (fvtv-debug): New flag.
3023 * tree.h (save_vtable_map_decl): New extern function decl.
3024
03085d1c
DM
30252013-08-07 David Malcolm <dmalcolm@redhat.com>
3026
3027 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
3028 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
3029 (pass_data_rl78_devirt): ...new pass_data instance and...
3030 (make_pass_rl78_devirt): ...new function.
3031 (rl78_asm_file_start): Port pass registration to new C++ API.
3032
05555c4a
DM
30332013-08-07 David Malcolm <dmalcolm@redhat.com>
3034
3035 * coretypes.h (rtl_opt_pass): Add.
3036 (gcc::context): Add.
3037 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
3038 (epiphany_init): Port to new C++ pass API.
3039 (epiphany_optimize_mode_switching): Likewise.
3040 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
3041 (pass_manager::get_pass_mode_switching): New.
3042 (pass_manager::get_pass_peephole2): New.
3043 * mode-switching.c (pass_mode_switching): Add clone method.
3044 * recog.c (pass_peephole2): Add clone method.
3045 (pass_split_all_insns): Add clone method.
3046
c5a12e26
DM
30472013-08-06 David Malcolm <dmalcolm@redhat.com>
3048
03b0ee0a
UB
3049 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
3050 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 3051
a2e2a668
JH
30522013-08-06 Jan Hubicka <jh@suse.cz>
3053
3054 * cgraph.c (cgraph_get_body): New function based on lto.c
3055 implementation.
3056 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB
3057 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
3058 LTO paths.
a2e2a668
JH
3059 * cgraphunit.c (expand_function): Get body prior expanding.
3060 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB
3061 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
3062 really need.
a2e2a668
JH
3063 * passes.c (do_per_function_toporder): Get body.
3064 * tree-inline.c (expand_call_inline): Get body prior inlining it.
3065 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
3066
15682f24
MJ
30672013-08-06 Martin Jambor <mjambor@suse.cz>
3068
3069 PR fortran/57987
3070 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
3071 is not re-finalized. Rename second parameter to no_collect.
3072
78f6dd68
MJ
30732013-08-06 Martin Jambor <mjambor@suse.cz>
3074
3075 PR middle-end/58041
3076 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
3077 MEM_REF has proper alignment information.
3078
12211b99 30792013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE
3080
3081 PR other/12081
3082 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
3083 class insn_gen_fn.
3084 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
3085 rtx (*) (rtx, ...) with insn_gen_fn.
3086 * genoutput.c (output_insn_data): Cast gen_? function pointers to
3087 insn_gen_fn::stored_funcptr. Add initializer braces.
3088
8ac69a6c
DM
30892013-08-05 David Malcolm <dmalcolm@redhat.com>
3090
3091 Rewrite how instances of passes are cloned to remove assumptions
3092 about their sizes (thus allowing pass subclasses to have
3093 additional data fields, albeit non-GC-managed ones at this point).
3094
3095 * passes.c (make_pass_instance): Now that passes have clone
3096 methods, rewrite this function to eliminate XNEW and memcpy
3097 calls that used hardcoded sizes. Since this function no longer
3098 creates pass instances, rename it to...
03b0ee0a
UB
3099 (add_pass_instance): ...this. Document the old way that passes were
3100 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM
3101 (next_pass_1): Add an initial_pass argument for use by
3102 add_pass_instance.
3103 (position_pass): When adding multiple instances of a pass, use
3104 the pass's clone method, rather than relying on the XNEW/memcpy
3105 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB
3106 (pass_manager::pass_manager): When invoking next_pass_1, also supply
3107 the initial instance of the current pass within the pass manager.
8ac69a6c 3108
27a4cd48
DM
31092013-08-05 David Malcolm <dmalcolm@redhat.com>
3110
3111 This is the automated part of the conversion of passes from C
3112 structs to C++ classes.
3113
3114 Patch autogenerated by refactor_passes.py from
3115 https://github.com/davidmalcolm/gcc-refactoring-scripts
3116 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
3117
3118 * asan.c (pass_asan): Convert from a global struct to a subclass of
3119 gimple_opt_pass along with...
3120 (pass_data_asan): ...new pass_data instance and...
3121 (make_pass_asan): ...new function.
3122 (pass_asan_O0): Convert from a global struct to a subclass of
3123 gimple_opt_pass along with...
3124 (pass_data_asan_O0): ...new pass_data instance and...
3125 (make_pass_asan_O0): ...new function.
3126 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
3127 subclass of rtl_opt_pass along with...
3128 (pass_data_inc_dec): ...new pass_data instance and...
3129 (make_pass_inc_dec): ...new function.
3130 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
3131 a subclass of rtl_opt_pass along with...
3132 (pass_data_reorder_blocks): ...new pass_data instance and...
3133 (make_pass_reorder_blocks): ...new function.
3134 (pass_duplicate_computed_gotos): Convert from a global struct to a
3135 subclass of rtl_opt_pass along with...
3136 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
3137 (make_pass_duplicate_computed_gotos): ...new function.
3138 (pass_partition_blocks): Convert from a global struct to a subclass of
3139 rtl_opt_pass along with...
3140 (pass_data_partition_blocks): ...new pass_data instance and...
3141 (make_pass_partition_blocks): ...new function.
3142 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
3143 struct to a subclass of rtl_opt_pass along with...
3144 (pass_data_branch_target_load_optimize1): ...new pass_data instance
3145 and...
3146 (make_pass_branch_target_load_optimize1): ...new function.
3147 (pass_branch_target_load_optimize2): Convert from a global struct to a
3148 subclass of rtl_opt_pass along with...
3149 (pass_data_branch_target_load_optimize2): ...new pass_data instance
3150 and...
3151 (make_pass_branch_target_load_optimize2): ...new function.
3152 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
3153 of rtl_opt_pass along with...
3154 (pass_data_jump): ...new pass_data instance and...
3155 (make_pass_jump): ...new function.
3156 (pass_jump2): Convert from a global struct to a subclass of
3157 rtl_opt_pass along with...
3158 (pass_data_jump2): ...new pass_data instance and...
3159 (make_pass_jump2): ...new function.
3160 * cfgexpand.c (pass_expand): Convert from a global struct to a
3161 subclass of rtl_opt_pass along with...
3162 (pass_data_expand): ...new pass_data instance and...
3163 (make_pass_expand): ...new function.
3164 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
3165 of rtl_opt_pass along with...
3166 (pass_data_free_cfg): ...new pass_data instance and...
3167 (make_pass_free_cfg): ...new function.
3168 (pass_into_cfg_layout_mode): Convert from a global struct to a
3169 subclass of rtl_opt_pass along with...
3170 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
3171 (make_pass_into_cfg_layout_mode): ...new function.
3172 (pass_outof_cfg_layout_mode): Convert from a global struct to a
3173 subclass of rtl_opt_pass along with...
3174 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
3175 (make_pass_outof_cfg_layout_mode): ...new function.
3176 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
3177 struct to a subclass of gimple_opt_pass along with...
3178 (pass_data_build_cgraph_edges): ...new pass_data instance and...
3179 (make_pass_build_cgraph_edges): ...new function.
3180 (pass_rebuild_cgraph_edges): Convert from a global struct to a
3181 subclass of gimple_opt_pass along with...
3182 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
3183 (make_pass_rebuild_cgraph_edges): ...new function.
3184 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
3185 subclass of gimple_opt_pass along with...
3186 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
3187 and...
3188 (make_pass_remove_cgraph_callee_edges): ...new function.
3189 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
3190 struct to a subclass of rtl_opt_pass along with...
3191 (pass_data_stack_adjustments): ...new pass_data instance and...
3192 (make_pass_stack_adjustments): ...new function.
3193 * combine.c (pass_combine): Convert from a global struct to a subclass
3194 of rtl_opt_pass along with...
3195 (pass_data_combine): ...new pass_data instance and...
3196 (make_pass_combine): ...new function.
3197 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
3198 global struct to a subclass of rtl_opt_pass along with...
3199 (pass_data_compare_elim_after_reload): ...new pass_data instance
3200 and...
3201 (make_pass_compare_elim_after_reload): ...new function.
3202 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
3203 of rtl_opt_pass along with...
3204 (pass_data_rtl_cprop): ...new pass_data instance and...
3205 (make_pass_rtl_cprop): ...new function.
3206 * cse.c (pass_cse): Convert from a global struct to a subclass of
3207 rtl_opt_pass along with...
3208 (pass_data_cse): ...new pass_data instance and...
3209 (make_pass_cse): ...new function.
3210 (pass_cse2): Convert from a global struct to a subclass of
3211 rtl_opt_pass along with...
3212 (pass_data_cse2): ...new pass_data instance and...
3213 (make_pass_cse2): ...new function.
3214 (pass_cse_after_global_opts): Convert from a global struct to a
3215 subclass of rtl_opt_pass along with...
3216 (pass_data_cse_after_global_opts): ...new pass_data instance and...
3217 (make_pass_cse_after_global_opts): ...new function.
3218 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
3219 of rtl_opt_pass along with...
3220 (pass_data_ud_rtl_dce): ...new pass_data instance and...
3221 (make_pass_ud_rtl_dce): ...new function.
3222 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
3223 rtl_opt_pass along with...
3224 (pass_data_fast_rtl_dce): ...new pass_data instance and...
3225 (make_pass_fast_rtl_dce): ...new function.
3226 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
3227 a subclass of rtl_opt_pass along with...
3228 (pass_data_df_initialize_opt): ...new pass_data instance and...
3229 (make_pass_df_initialize_opt): ...new function.
3230 (pass_df_initialize_no_opt): Convert from a global struct to a
3231 subclass of rtl_opt_pass along with...
3232 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
3233 (make_pass_df_initialize_no_opt): ...new function.
3234 (pass_df_finish): Convert from a global struct to a subclass of
3235 rtl_opt_pass along with...
3236 (pass_data_df_finish): ...new pass_data instance and...
3237 (make_pass_df_finish): ...new function.
3238 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
3239 rtl_opt_pass along with...
3240 (pass_data_rtl_dse1): ...new pass_data instance and...
3241 (make_pass_rtl_dse1): ...new function.
3242 (pass_rtl_dse2): Convert from a global struct to a subclass of
3243 rtl_opt_pass along with...
3244 (pass_data_rtl_dse2): ...new pass_data instance and...
3245 (make_pass_rtl_dse2): ...new function.
3246 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
3247 subclass of rtl_opt_pass along with...
3248 (pass_data_dwarf2_frame): ...new pass_data instance and...
3249 (make_pass_dwarf2_frame): ...new function.
3250 * except.c (pass_set_nothrow_function_flags): Convert from a global
3251 struct to a subclass of rtl_opt_pass along with...
3252 (pass_data_set_nothrow_function_flags): ...new pass_data instance
3253 and...
3254 (make_pass_set_nothrow_function_flags): ...new function.
3255 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
3256 subclass of rtl_opt_pass along with...
3257 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
3258 and...
3259 (make_pass_convert_to_eh_region_ranges): ...new function.
3260 * final.c (pass_compute_alignments): Convert from a global struct to a
3261 subclass of rtl_opt_pass along with...
3262 (pass_data_compute_alignments): ...new pass_data instance and...
3263 (make_pass_compute_alignments): ...new function.
3264 (pass_final): Convert from a global struct to a subclass of
3265 rtl_opt_pass along with...
3266 (pass_data_final): ...new pass_data instance and...
3267 (make_pass_final): ...new function.
3268 (pass_shorten_branches): Convert from a global struct to a subclass of
3269 rtl_opt_pass along with...
3270 (pass_data_shorten_branches): ...new pass_data instance and...
3271 (make_pass_shorten_branches): ...new function.
3272 (pass_clean_state): Convert from a global struct to a subclass of
3273 rtl_opt_pass along with...
3274 (pass_data_clean_state): ...new pass_data instance and...
3275 (make_pass_clean_state): ...new function.
3276 * function.c (pass_instantiate_virtual_regs): Convert from a global
3277 struct to a subclass of rtl_opt_pass along with...
3278 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
3279 (make_pass_instantiate_virtual_regs): ...new function.
3280 (pass_leaf_regs): Convert from a global struct to a subclass of
3281 rtl_opt_pass along with...
3282 (pass_data_leaf_regs): ...new pass_data instance and...
3283 (make_pass_leaf_regs): ...new function.
3284 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
3285 subclass of rtl_opt_pass along with...
3286 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
3287 and...
3288 (make_pass_thread_prologue_and_epilogue): ...new function.
3289 (pass_match_asm_constraints): Convert from a global struct to a
3290 subclass of rtl_opt_pass along with...
3291 (pass_data_match_asm_constraints): ...new pass_data instance and...
3292 (make_pass_match_asm_constraints): ...new function.
3293 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
3294 subclass of rtl_opt_pass along with...
3295 (pass_data_rtl_fwprop): ...new pass_data instance and...
3296 (make_pass_rtl_fwprop): ...new function.
3297 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
3298 rtl_opt_pass along with...
3299 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
3300 (make_pass_rtl_fwprop_addr): ...new function.
3301 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
3302 rtl_opt_pass along with...
3303 (pass_data_rtl_pre): ...new pass_data instance and...
3304 (make_pass_rtl_pre): ...new function.
3305 (pass_rtl_hoist): Convert from a global struct to a subclass of
3306 rtl_opt_pass along with...
3307 (pass_data_rtl_hoist): ...new pass_data instance and...
3308 (make_pass_rtl_hoist): ...new function.
3309 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
3310 subclass of gimple_opt_pass along with...
3311 (pass_data_lower_cf): ...new pass_data instance and...
3312 (make_pass_lower_cf): ...new function.
3313 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
3314 from a global struct to a subclass of gimple_opt_pass along with...
3315 (pass_data_strength_reduction): ...new pass_data instance and...
3316 (make_pass_strength_reduction): ...new function.
3317 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
3318 of rtl_opt_pass along with...
3319 (pass_data_rtl_ifcvt): ...new pass_data instance and...
3320 (make_pass_rtl_ifcvt): ...new function.
3321 (pass_if_after_combine): Convert from a global struct to a subclass of
3322 rtl_opt_pass along with...
3323 (pass_data_if_after_combine): ...new pass_data instance and...
3324 (make_pass_if_after_combine): ...new function.
3325 (pass_if_after_reload): Convert from a global struct to a subclass of
3326 rtl_opt_pass along with...
3327 (pass_data_if_after_reload): ...new pass_data instance and...
3328 (make_pass_if_after_reload): ...new function.
3329 * init-regs.c (pass_initialize_regs): Convert from a global struct to
3330 a subclass of rtl_opt_pass along with...
3331 (pass_data_initialize_regs): ...new pass_data instance and...
3332 (make_pass_initialize_regs): ...new function.
3333 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
3334 of ipa_opt_pass_d along with...
3335 (pass_data_ipa_cp): ...new pass_data instance and...
3336 (make_pass_ipa_cp): ...new function.
3337 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
3338 global struct to a subclass of gimple_opt_pass along with...
3339 (pass_data_inline_parameters): ...new pass_data instance and...
3340 (make_pass_inline_parameters): ...new function.
3341 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
3342 subclass of gimple_opt_pass along with...
3343 (pass_data_early_inline): ...new pass_data instance and...
3344 (make_pass_early_inline): ...new function.
3345 (pass_ipa_inline): Convert from a global struct to a subclass of
3346 ipa_opt_pass_d along with...
3347 (pass_data_ipa_inline): ...new pass_data instance and...
3348 (make_pass_ipa_inline): ...new function.
3349 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
3350 struct to a subclass of gimple_opt_pass along with...
3351 (pass_data_local_pure_const): ...new pass_data instance and...
3352 (make_pass_local_pure_const): ...new function.
3353 (pass_ipa_pure_const): Convert from a global struct to a subclass of
3354 ipa_opt_pass_d along with...
3355 (pass_data_ipa_pure_const): ...new pass_data instance and...
3356 (make_pass_ipa_pure_const): ...new function.
3357 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
3358 to a subclass of ipa_opt_pass_d along with...
3359 (pass_data_ipa_reference): ...new pass_data instance and...
3360 (make_pass_ipa_reference): ...new function.
3361 * ipa-split.c (pass_split_functions): Convert from a global struct to
3362 a subclass of gimple_opt_pass along with...
3363 (pass_data_split_functions): ...new pass_data instance and...
3364 (make_pass_split_functions): ...new function.
3365 (pass_feedback_split_functions): Convert from a global struct to a
3366 subclass of gimple_opt_pass along with...
3367 (pass_data_feedback_split_functions): ...new pass_data instance and...
3368 (make_pass_feedback_split_functions): ...new function.
3369 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
3370 global struct to a subclass of simple_ipa_opt_pass along with...
3371 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
3372 instance and...
3373 (make_pass_ipa_function_and_variable_visibility): ...new function.
3374 (pass_ipa_free_inline_summary): Convert from a global struct to a
3375 subclass of simple_ipa_opt_pass along with...
3376 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
3377 (make_pass_ipa_free_inline_summary): ...new function.
3378 (pass_ipa_whole_program_visibility): Convert from a global struct to a
3379 subclass of ipa_opt_pass_d along with...
3380 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
3381 and...
3382 (make_pass_ipa_whole_program_visibility): ...new function.
3383 (pass_ipa_profile): Convert from a global struct to a subclass of
3384 ipa_opt_pass_d along with...
3385 (pass_data_ipa_profile): ...new pass_data instance and...
3386 (make_pass_ipa_profile): ...new function.
3387 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
3388 ipa_opt_pass_d along with...
3389 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
3390 (make_pass_ipa_cdtor_merge): ...new function.
3391 * ira.c (pass_ira): Convert from a global struct to a subclass of
3392 rtl_opt_pass along with...
3393 (pass_data_ira): ...new pass_data instance and...
3394 (make_pass_ira): ...new function.
3395 (pass_reload): Convert from a global struct to a subclass of
3396 rtl_opt_pass along with...
3397 (pass_data_reload): ...new pass_data instance and...
3398 (make_pass_reload): ...new function.
3399 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
3400 subclass of rtl_opt_pass along with...
3401 (pass_data_cleanup_barriers): ...new pass_data instance and...
3402 (make_pass_cleanup_barriers): ...new function.
3403 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
3404 of rtl_opt_pass along with...
3405 (pass_data_loop2): ...new pass_data instance and...
3406 (make_pass_loop2): ...new function.
3407 (pass_rtl_loop_init): Convert from a global struct to a subclass of
3408 rtl_opt_pass along with...
3409 (pass_data_rtl_loop_init): ...new pass_data instance and...
3410 (make_pass_rtl_loop_init): ...new function.
3411 (pass_rtl_loop_done): Convert from a global struct to a subclass of
3412 rtl_opt_pass along with...
3413 (pass_data_rtl_loop_done): ...new pass_data instance and...
3414 (make_pass_rtl_loop_done): ...new function.
3415 (pass_rtl_move_loop_invariants): Convert from a global struct to a
3416 subclass of rtl_opt_pass along with...
3417 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
3418 (make_pass_rtl_move_loop_invariants): ...new function.
3419 (pass_rtl_unswitch): Convert from a global struct to a subclass of
3420 rtl_opt_pass along with...
3421 (pass_data_rtl_unswitch): ...new pass_data instance and...
3422 (make_pass_rtl_unswitch): ...new function.
3423 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
3424 subclass of rtl_opt_pass along with...
3425 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
3426 and...
3427 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3428 (pass_rtl_doloop): Convert from a global struct to a subclass of
3429 rtl_opt_pass along with...
3430 (pass_data_rtl_doloop): ...new pass_data instance and...
3431 (make_pass_rtl_doloop): ...new function.
3432 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
3433 a subclass of rtl_opt_pass along with...
3434 (pass_data_lower_subreg): ...new pass_data instance and...
3435 (make_pass_lower_subreg): ...new function.
3436 (pass_lower_subreg2): Convert from a global struct to a subclass of
3437 rtl_opt_pass along with...
3438 (pass_data_lower_subreg2): ...new pass_data instance and...
3439 (make_pass_lower_subreg2): ...new function.
3440 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
3441 struct to a subclass of ipa_opt_pass_d along with...
3442 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
3443 (make_pass_ipa_lto_gimple_out): ...new function.
3444 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
3445 of ipa_opt_pass_d along with...
3446 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
3447 (make_pass_ipa_lto_finish_out): ...new function.
3448 * mode-switching.c (pass_mode_switching): Convert from a global struct
3449 to a subclass of rtl_opt_pass along with...
3450 (pass_data_mode_switching): ...new pass_data instance and...
3451 (make_pass_mode_switching): ...new function.
3452 * modulo-sched.c (pass_sms): Convert from a global struct to a
3453 subclass of rtl_opt_pass along with...
3454 (pass_data_sms): ...new pass_data instance and...
3455 (make_pass_sms): ...new function.
3456 * omp-low.c (pass_expand_omp): Convert from a global struct to a
3457 subclass of gimple_opt_pass along with...
3458 (pass_data_expand_omp): ...new pass_data instance and...
3459 (make_pass_expand_omp): ...new function.
3460 (pass_lower_omp): Convert from a global struct to a subclass of
3461 gimple_opt_pass along with...
3462 (pass_data_lower_omp): ...new pass_data instance and...
3463 (make_pass_lower_omp): ...new function.
3464 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
3465 of gimple_opt_pass along with...
3466 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
3467 (make_pass_diagnose_omp_blocks): ...new function.
3468 * passes.c (pass_early_local_passes): Convert from a global struct to
3469 a subclass of simple_ipa_opt_pass along with...
3470 (pass_data_early_local_passes): ...new pass_data instance and...
3471 (make_pass_early_local_passes): ...new function.
3472 (pass_all_early_optimizations): Convert from a global struct to a
3473 subclass of gimple_opt_pass along with...
3474 (pass_data_all_early_optimizations): ...new pass_data instance and...
3475 (make_pass_all_early_optimizations): ...new function.
3476 (pass_all_optimizations): Convert from a global struct to a subclass
3477 of gimple_opt_pass along with...
3478 (pass_data_all_optimizations): ...new pass_data instance and...
3479 (make_pass_all_optimizations): ...new function.
3480 (pass_all_optimizations_g): Convert from a global struct to a subclass
3481 of gimple_opt_pass along with...
3482 (pass_data_all_optimizations_g): ...new pass_data instance and...
3483 (make_pass_all_optimizations_g): ...new function.
3484 (pass_rest_of_compilation): Convert from a global struct to a subclass
3485 of rtl_opt_pass along with...
3486 (pass_data_rest_of_compilation): ...new pass_data instance and...
3487 (make_pass_rest_of_compilation): ...new function.
3488 (pass_postreload): Convert from a global struct to a subclass of
3489 rtl_opt_pass along with...
3490 (pass_data_postreload): ...new pass_data instance and...
3491 (make_pass_postreload): ...new function.
3492 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
3493 subclass of rtl_opt_pass along with...
3494 (pass_data_gcse2): ...new pass_data instance and...
3495 (make_pass_gcse2): ...new function.
3496 * postreload.c (pass_postreload_cse): Convert from a global struct to
3497 a subclass of rtl_opt_pass along with...
3498 (pass_data_postreload_cse): ...new pass_data instance and...
3499 (make_pass_postreload_cse): ...new function.
3500 * predict.c (pass_profile): Convert from a global struct to a subclass
3501 of gimple_opt_pass along with...
3502 (pass_data_profile): ...new pass_data instance and...
3503 (make_pass_profile): ...new function.
3504 (pass_strip_predict_hints): Convert from a global struct to a subclass
3505 of gimple_opt_pass along with...
3506 (pass_data_strip_predict_hints): ...new pass_data instance and...
3507 (make_pass_strip_predict_hints): ...new function.
3508 * recog.c (pass_peephole2): Convert from a global struct to a subclass
3509 of rtl_opt_pass along with...
3510 (pass_data_peephole2): ...new pass_data instance and...
3511 (make_pass_peephole2): ...new function.
3512 (pass_split_all_insns): Convert from a global struct to a subclass of
3513 rtl_opt_pass along with...
3514 (pass_data_split_all_insns): ...new pass_data instance and...
3515 (make_pass_split_all_insns): ...new function.
3516 (pass_split_after_reload): Convert from a global struct to a subclass
3517 of rtl_opt_pass along with...
3518 (pass_data_split_after_reload): ...new pass_data instance and...
3519 (make_pass_split_after_reload): ...new function.
3520 (pass_split_before_regstack): Convert from a global struct to a
3521 subclass of rtl_opt_pass along with...
3522 (pass_data_split_before_regstack): ...new pass_data instance and...
3523 (make_pass_split_before_regstack): ...new function.
3524 (pass_split_before_sched2): Convert from a global struct to a subclass
3525 of rtl_opt_pass along with...
3526 (pass_data_split_before_sched2): ...new pass_data instance and...
3527 (make_pass_split_before_sched2): ...new function.
3528 (pass_split_for_shorten_branches): Convert from a global struct to a
3529 subclass of rtl_opt_pass along with...
3530 (pass_data_split_for_shorten_branches): ...new pass_data instance
3531 and...
3532 (make_pass_split_for_shorten_branches): ...new function.
3533 * ree.c (pass_ree): Convert from a global struct to a subclass of
3534 rtl_opt_pass along with...
3535 (pass_data_ree): ...new pass_data instance and...
3536 (make_pass_ree): ...new function.
3537 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
3538 subclass of rtl_opt_pass along with...
3539 (pass_data_stack_regs): ...new pass_data instance and...
3540 (make_pass_stack_regs): ...new function.
3541 (pass_stack_regs_run): Convert from a global struct to a subclass of
3542 rtl_opt_pass along with...
3543 (pass_data_stack_regs_run): ...new pass_data instance and...
3544 (make_pass_stack_regs_run): ...new function.
3545 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
3546 subclass of rtl_opt_pass along with...
3547 (pass_data_cprop_hardreg): ...new pass_data instance and...
3548 (make_pass_cprop_hardreg): ...new function.
3549 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
3550 subclass of rtl_opt_pass along with...
3551 (pass_data_reginfo_init): ...new pass_data instance and...
3552 (make_pass_reginfo_init): ...new function.
3553 * regmove.c (pass_regmove): Convert from a global struct to a subclass
3554 of rtl_opt_pass along with...
3555 (pass_data_regmove): ...new pass_data instance and...
3556 (make_pass_regmove): ...new function.
3557 * regrename.c (pass_regrename): Convert from a global struct to a
3558 subclass of rtl_opt_pass along with...
3559 (pass_data_regrename): ...new pass_data instance and...
3560 (make_pass_regrename): ...new function.
3561 * reorg.c (pass_delay_slots): Convert from a global struct to a
3562 subclass of rtl_opt_pass along with...
3563 (pass_data_delay_slots): ...new pass_data instance and...
3564 (make_pass_delay_slots): ...new function.
3565 (pass_machine_reorg): Convert from a global struct to a subclass of
3566 rtl_opt_pass along with...
3567 (pass_data_machine_reorg): ...new pass_data instance and...
3568 (make_pass_machine_reorg): ...new function.
3569 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
3570 of rtl_opt_pass along with...
3571 (pass_data_sched): ...new pass_data instance and...
3572 (make_pass_sched): ...new function.
3573 (pass_sched2): Convert from a global struct to a subclass of
3574 rtl_opt_pass along with...
3575 (pass_data_sched2): ...new pass_data instance and...
3576 (make_pass_sched2): ...new function.
3577 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
3578 to a subclass of rtl_opt_pass along with...
3579 (pass_data_stack_ptr_mod): ...new pass_data instance and...
3580 (make_pass_stack_ptr_mod): ...new function.
3581 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
3582 to a subclass of rtl_opt_pass along with...
3583 (pass_data_rtl_store_motion): ...new pass_data instance and...
3584 (make_pass_rtl_store_motion): ...new function.
3585 * tracer.c (pass_tracer): Convert from a global struct to a subclass
3586 of gimple_opt_pass along with...
3587 (pass_data_tracer): ...new pass_data instance and...
3588 (make_pass_tracer): ...new function.
3589 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
3590 to a subclass of gimple_opt_pass along with...
3591 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
3592 (make_pass_diagnose_tm_blocks): ...new function.
3593 (pass_lower_tm): Convert from a global struct to a subclass of
3594 gimple_opt_pass along with...
3595 (pass_data_lower_tm): ...new pass_data instance and...
3596 (make_pass_lower_tm): ...new function.
3597 (pass_tm_init): Convert from a global struct to a subclass of
3598 gimple_opt_pass along with...
3599 (pass_data_tm_init): ...new pass_data instance and...
3600 (make_pass_tm_init): ...new function.
3601 (pass_tm_mark): Convert from a global struct to a subclass of
3602 gimple_opt_pass along with...
3603 (pass_data_tm_mark): ...new pass_data instance and...
3604 (make_pass_tm_mark): ...new function.
3605 (pass_tm_edges): Convert from a global struct to a subclass of
3606 gimple_opt_pass along with...
3607 (pass_data_tm_edges): ...new pass_data instance and...
3608 (make_pass_tm_edges): ...new function.
3609 (pass_tm_memopt): Convert from a global struct to a subclass of
3610 gimple_opt_pass along with...
3611 (pass_data_tm_memopt): ...new pass_data instance and...
3612 (make_pass_tm_memopt): ...new function.
3613 (pass_ipa_tm): Convert from a global struct to a subclass of
3614 simple_ipa_opt_pass along with...
3615 (pass_data_ipa_tm): ...new pass_data instance and...
3616 (make_pass_ipa_tm): ...new function.
3617 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
3618 subclass of gimple_opt_pass along with...
3619 (pass_data_call_cdce): ...new pass_data instance and...
3620 (make_pass_call_cdce): ...new function.
3621 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
3622 subclass of gimple_opt_pass along with...
3623 (pass_data_build_cfg): ...new pass_data instance and...
3624 (make_pass_build_cfg): ...new function.
3625 (pass_split_crit_edges): Convert from a global struct to a subclass of
3626 gimple_opt_pass along with...
3627 (pass_data_split_crit_edges): ...new pass_data instance and...
3628 (make_pass_split_crit_edges): ...new function.
3629 (pass_warn_function_return): Convert from a global struct to a
3630 subclass of gimple_opt_pass along with...
3631 (pass_data_warn_function_return): ...new pass_data instance and...
3632 (make_pass_warn_function_return): ...new function.
3633 (pass_warn_function_noreturn): Convert from a global struct to a
3634 subclass of gimple_opt_pass along with...
3635 (pass_data_warn_function_noreturn): ...new pass_data instance and...
3636 (make_pass_warn_function_noreturn): ...new function.
3637 (pass_warn_unused_result): Convert from a global struct to a subclass
3638 of gimple_opt_pass along with...
3639 (pass_data_warn_unused_result): ...new pass_data instance and...
3640 (make_pass_warn_unused_result): ...new function.
3641 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
3642 a subclass of gimple_opt_pass along with...
3643 (pass_data_merge_phi): ...new pass_data instance and...
3644 (make_pass_merge_phi): ...new function.
3645 * tree-complex.c (pass_lower_complex): Convert from a global struct to
3646 a subclass of gimple_opt_pass along with...
3647 (pass_data_lower_complex): ...new pass_data instance and...
3648 (make_pass_lower_complex): ...new function.
3649 (pass_lower_complex_O0): Convert from a global struct to a subclass of
3650 gimple_opt_pass along with...
3651 (pass_data_lower_complex_O0): ...new pass_data instance and...
3652 (make_pass_lower_complex_O0): ...new function.
3653 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
3654 subclass of gimple_opt_pass along with...
3655 (pass_data_lower_eh): ...new pass_data instance and...
3656 (make_pass_lower_eh): ...new function.
3657 (pass_refactor_eh): Convert from a global struct to a subclass of
3658 gimple_opt_pass along with...
3659 (pass_data_refactor_eh): ...new pass_data instance and...
3660 (make_pass_refactor_eh): ...new function.
3661 (pass_lower_resx): Convert from a global struct to a subclass of
3662 gimple_opt_pass along with...
3663 (pass_data_lower_resx): ...new pass_data instance and...
3664 (make_pass_lower_resx): ...new function.
3665 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
3666 of gimple_opt_pass along with...
3667 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
3668 (make_pass_lower_eh_dispatch): ...new function.
3669 (pass_cleanup_eh): Convert from a global struct to a subclass of
3670 gimple_opt_pass along with...
3671 (pass_data_cleanup_eh): ...new pass_data instance and...
3672 (make_pass_cleanup_eh): ...new function.
3673 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
3674 to a subclass of simple_ipa_opt_pass along with...
3675 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
3676 (make_pass_ipa_lower_emutls): ...new function.
3677 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
3678 a subclass of gimple_opt_pass along with...
3679 (pass_data_if_conversion): ...new pass_data instance and...
3680 (make_pass_if_conversion): ...new function.
3681 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
3682 subclass of gimple_opt_pass along with...
3683 (pass_data_build_ssa): ...new pass_data instance and...
3684 (make_pass_build_ssa): ...new function.
3685 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
3686 global struct to a subclass of gimple_opt_pass along with...
3687 (pass_data_loop_distribution): ...new pass_data instance and...
3688 (make_pass_loop_distribution): ...new function.
3689 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
3690 subclass of gimple_opt_pass along with...
3691 (pass_data_mudflap_1): ...new pass_data instance and...
3692 (make_pass_mudflap_1): ...new function.
3693 (pass_mudflap_2): Convert from a global struct to a subclass of
3694 gimple_opt_pass along with...
3695 (pass_data_mudflap_2): ...new pass_data instance and...
3696 (make_pass_mudflap_2): ...new function.
3697 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
3698 subclass of gimple_opt_pass along with...
3699 (pass_data_mudflap_1): ...new pass_data instance and...
3700 (make_pass_mudflap_1): ...new function.
3701 (pass_mudflap_2): Convert from a global struct to a subclass of
3702 gimple_opt_pass along with...
3703 (pass_data_mudflap_2): ...new pass_data instance and...
3704 (make_pass_mudflap_2): ...new function.
3705 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
3706 gimple_opt_pass along with...
3707 (pass_data_nrv): ...new pass_data instance and...
3708 (make_pass_nrv): ...new function.
3709 (pass_return_slot): Convert from a global struct to a subclass of
3710 gimple_opt_pass along with...
3711 (pass_data_return_slot): ...new pass_data instance and...
3712 (make_pass_return_slot): ...new function.
3713 * tree-object-size.c (pass_object_sizes): Convert from a global struct
3714 to a subclass of gimple_opt_pass along with...
3715 (pass_data_object_sizes): ...new pass_data instance and...
3716 (make_pass_object_sizes): ...new function.
3717 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
3718 global struct to a subclass of gimple_opt_pass along with...
3719 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
3720 and...
3721 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3722 (pass_fixup_cfg): Convert from a global struct to a subclass of
3723 gimple_opt_pass along with...
3724 (pass_data_fixup_cfg): ...new pass_data instance and...
3725 (make_pass_fixup_cfg): ...new function.
3726 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
3727 (make_pass_mudflap_1): ...new function.
3728 (pass_mudflap_2): Replace declaration with that of...
3729 (make_pass_mudflap_2): ...new function.
3730 (pass_asan): Replace declaration with that of...
3731 (make_pass_asan): ...new function.
3732 (pass_asan_O0): Replace declaration with that of...
3733 (make_pass_asan_O0): ...new function.
3734 (pass_tsan): Replace declaration with that of...
3735 (make_pass_tsan): ...new function.
3736 (pass_tsan_O0): Replace declaration with that of...
3737 (make_pass_tsan_O0): ...new function.
3738 (pass_lower_cf): Replace declaration with that of...
3739 (make_pass_lower_cf): ...new function.
3740 (pass_refactor_eh): Replace declaration with that of...
3741 (make_pass_refactor_eh): ...new function.
3742 (pass_lower_eh): Replace declaration with that of...
3743 (make_pass_lower_eh): ...new function.
3744 (pass_lower_eh_dispatch): Replace declaration with that of...
3745 (make_pass_lower_eh_dispatch): ...new function.
3746 (pass_lower_resx): Replace declaration with that of...
3747 (make_pass_lower_resx): ...new function.
3748 (pass_build_cfg): Replace declaration with that of...
3749 (make_pass_build_cfg): ...new function.
3750 (pass_early_tree_profile): Replace declaration with that of...
3751 (make_pass_early_tree_profile): ...new function.
3752 (pass_cleanup_eh): Replace declaration with that of...
3753 (make_pass_cleanup_eh): ...new function.
3754 (pass_sra): Replace declaration with that of...
3755 (make_pass_sra): ...new function.
3756 (pass_sra_early): Replace declaration with that of...
3757 (make_pass_sra_early): ...new function.
3758 (pass_early_ipa_sra): Replace declaration with that of...
3759 (make_pass_early_ipa_sra): ...new function.
3760 (pass_tail_recursion): Replace declaration with that of...
3761 (make_pass_tail_recursion): ...new function.
3762 (pass_tail_calls): Replace declaration with that of...
3763 (make_pass_tail_calls): ...new function.
3764 (pass_tree_loop): Replace declaration with that of...
3765 (make_pass_tree_loop): ...new function.
3766 (pass_tree_loop_init): Replace declaration with that of...
3767 (make_pass_tree_loop_init): ...new function.
3768 (pass_lim): Replace declaration with that of...
3769 (make_pass_lim): ...new function.
3770 (pass_tree_unswitch): Replace declaration with that of...
3771 (make_pass_tree_unswitch): ...new function.
3772 (pass_predcom): Replace declaration with that of...
3773 (make_pass_predcom): ...new function.
3774 (pass_iv_canon): Replace declaration with that of...
3775 (make_pass_iv_canon): ...new function.
3776 (pass_scev_cprop): Replace declaration with that of...
3777 (make_pass_scev_cprop): ...new function.
3778 (pass_empty_loop): Replace declaration with that of...
3779 (make_pass_empty_loop): ...new function.
3780 (pass_record_bounds): Replace declaration with that of...
3781 (make_pass_record_bounds): ...new function.
3782 (pass_graphite): Replace declaration with that of...
3783 (make_pass_graphite): ...new function.
3784 (pass_graphite_transforms): Replace declaration with that of...
3785 (make_pass_graphite_transforms): ...new function.
3786 (pass_if_conversion): Replace declaration with that of...
3787 (make_pass_if_conversion): ...new function.
3788 (pass_loop_distribution): Replace declaration with that of...
3789 (make_pass_loop_distribution): ...new function.
3790 (pass_vectorize): Replace declaration with that of...
3791 (make_pass_vectorize): ...new function.
3792 (pass_slp_vectorize): Replace declaration with that of...
3793 (make_pass_slp_vectorize): ...new function.
3794 (pass_complete_unroll): Replace declaration with that of...
3795 (make_pass_complete_unroll): ...new function.
3796 (pass_complete_unrolli): Replace declaration with that of...
3797 (make_pass_complete_unrolli): ...new function.
3798 (pass_parallelize_loops): Replace declaration with that of...
3799 (make_pass_parallelize_loops): ...new function.
3800 (pass_loop_prefetch): Replace declaration with that of...
3801 (make_pass_loop_prefetch): ...new function.
3802 (pass_iv_optimize): Replace declaration with that of...
3803 (make_pass_iv_optimize): ...new function.
3804 (pass_tree_loop_done): Replace declaration with that of...
3805 (make_pass_tree_loop_done): ...new function.
3806 (pass_ch): Replace declaration with that of...
3807 (make_pass_ch): ...new function.
3808 (pass_ccp): Replace declaration with that of...
3809 (make_pass_ccp): ...new function.
3810 (pass_phi_only_cprop): Replace declaration with that of...
3811 (make_pass_phi_only_cprop): ...new function.
3812 (pass_build_ssa): Replace declaration with that of...
3813 (make_pass_build_ssa): ...new function.
3814 (pass_build_alias): Replace declaration with that of...
3815 (make_pass_build_alias): ...new function.
3816 (pass_build_ealias): Replace declaration with that of...
3817 (make_pass_build_ealias): ...new function.
3818 (pass_dominator): Replace declaration with that of...
3819 (make_pass_dominator): ...new function.
3820 (pass_dce): Replace declaration with that of...
3821 (make_pass_dce): ...new function.
3822 (pass_dce_loop): Replace declaration with that of...
3823 (make_pass_dce_loop): ...new function.
3824 (pass_cd_dce): Replace declaration with that of...
3825 (make_pass_cd_dce): ...new function.
3826 (pass_call_cdce): Replace declaration with that of...
3827 (make_pass_call_cdce): ...new function.
3828 (pass_merge_phi): Replace declaration with that of...
3829 (make_pass_merge_phi): ...new function.
3830 (pass_split_crit_edges): Replace declaration with that of...
3831 (make_pass_split_crit_edges): ...new function.
3832 (pass_pre): Replace declaration with that of...
3833 (make_pass_pre): ...new function.
3834 (pass_profile): Replace declaration with that of...
3835 (make_pass_profile): ...new function.
3836 (pass_strip_predict_hints): Replace declaration with that of...
3837 (make_pass_strip_predict_hints): ...new function.
3838 (pass_lower_complex_O0): Replace declaration with that of...
3839 (make_pass_lower_complex_O0): ...new function.
3840 (pass_lower_complex): Replace declaration with that of...
3841 (make_pass_lower_complex): ...new function.
3842 (pass_lower_vector): Replace declaration with that of...
3843 (make_pass_lower_vector): ...new function.
3844 (pass_lower_vector_ssa): Replace declaration with that of...
3845 (make_pass_lower_vector_ssa): ...new function.
3846 (pass_lower_omp): Replace declaration with that of...
3847 (make_pass_lower_omp): ...new function.
3848 (pass_diagnose_omp_blocks): Replace declaration with that of...
3849 (make_pass_diagnose_omp_blocks): ...new function.
3850 (pass_expand_omp): Replace declaration with that of...
3851 (make_pass_expand_omp): ...new function.
3852 (pass_expand_omp_ssa): Replace declaration with that of...
3853 (make_pass_expand_omp_ssa): ...new function.
3854 (pass_object_sizes): Replace declaration with that of...
3855 (make_pass_object_sizes): ...new function.
3856 (pass_strlen): Replace declaration with that of...
3857 (make_pass_strlen): ...new function.
3858 (pass_fold_builtins): Replace declaration with that of...
3859 (make_pass_fold_builtins): ...new function.
3860 (pass_stdarg): Replace declaration with that of...
3861 (make_pass_stdarg): ...new function.
3862 (pass_early_warn_uninitialized): Replace declaration with that of...
3863 (make_pass_early_warn_uninitialized): ...new function.
3864 (pass_late_warn_uninitialized): Replace declaration with that of...
3865 (make_pass_late_warn_uninitialized): ...new function.
3866 (pass_cse_reciprocals): Replace declaration with that of...
3867 (make_pass_cse_reciprocals): ...new function.
3868 (pass_cse_sincos): Replace declaration with that of...
3869 (make_pass_cse_sincos): ...new function.
3870 (pass_optimize_bswap): Replace declaration with that of...
3871 (make_pass_optimize_bswap): ...new function.
3872 (pass_optimize_widening_mul): Replace declaration with that of...
3873 (make_pass_optimize_widening_mul): ...new function.
3874 (pass_warn_function_return): Replace declaration with that of...
3875 (make_pass_warn_function_return): ...new function.
3876 (pass_warn_function_noreturn): Replace declaration with that of...
3877 (make_pass_warn_function_noreturn): ...new function.
3878 (pass_cselim): Replace declaration with that of...
3879 (make_pass_cselim): ...new function.
3880 (pass_phiopt): Replace declaration with that of...
3881 (make_pass_phiopt): ...new function.
3882 (pass_forwprop): Replace declaration with that of...
3883 (make_pass_forwprop): ...new function.
3884 (pass_phiprop): Replace declaration with that of...
3885 (make_pass_phiprop): ...new function.
3886 (pass_tree_ifcombine): Replace declaration with that of...
3887 (make_pass_tree_ifcombine): ...new function.
3888 (pass_dse): Replace declaration with that of...
3889 (make_pass_dse): ...new function.
3890 (pass_nrv): Replace declaration with that of...
3891 (make_pass_nrv): ...new function.
3892 (pass_rename_ssa_copies): Replace declaration with that of...
3893 (make_pass_rename_ssa_copies): ...new function.
3894 (pass_sink_code): Replace declaration with that of...
3895 (make_pass_sink_code): ...new function.
3896 (pass_fre): Replace declaration with that of...
3897 (make_pass_fre): ...new function.
3898 (pass_check_data_deps): Replace declaration with that of...
3899 (make_pass_check_data_deps): ...new function.
3900 (pass_copy_prop): Replace declaration with that of...
3901 (make_pass_copy_prop): ...new function.
3902 (pass_vrp): Replace declaration with that of...
3903 (make_pass_vrp): ...new function.
3904 (pass_uncprop): Replace declaration with that of...
3905 (make_pass_uncprop): ...new function.
3906 (pass_return_slot): Replace declaration with that of...
3907 (make_pass_return_slot): ...new function.
3908 (pass_reassoc): Replace declaration with that of...
3909 (make_pass_reassoc): ...new function.
3910 (pass_rebuild_cgraph_edges): Replace declaration with that of...
3911 (make_pass_rebuild_cgraph_edges): ...new function.
3912 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
3913 (make_pass_remove_cgraph_callee_edges): ...new function.
3914 (pass_build_cgraph_edges): Replace declaration with that of...
3915 (make_pass_build_cgraph_edges): ...new function.
3916 (pass_local_pure_const): Replace declaration with that of...
3917 (make_pass_local_pure_const): ...new function.
3918 (pass_tracer): Replace declaration with that of...
3919 (make_pass_tracer): ...new function.
3920 (pass_warn_unused_result): Replace declaration with that of...
3921 (make_pass_warn_unused_result): ...new function.
3922 (pass_diagnose_tm_blocks): Replace declaration with that of...
3923 (make_pass_diagnose_tm_blocks): ...new function.
3924 (pass_lower_tm): Replace declaration with that of...
3925 (make_pass_lower_tm): ...new function.
3926 (pass_tm_init): Replace declaration with that of...
3927 (make_pass_tm_init): ...new function.
3928 (pass_tm_mark): Replace declaration with that of...
3929 (make_pass_tm_mark): ...new function.
3930 (pass_tm_memopt): Replace declaration with that of...
3931 (make_pass_tm_memopt): ...new function.
3932 (pass_tm_edges): Replace declaration with that of...
3933 (make_pass_tm_edges): ...new function.
3934 (pass_split_functions): Replace declaration with that of...
3935 (make_pass_split_functions): ...new function.
3936 (pass_feedback_split_functions): Replace declaration with that of...
3937 (make_pass_feedback_split_functions): ...new function.
3938 (pass_strength_reduction): Replace declaration with that of...
3939 (make_pass_strength_reduction): ...new function.
3940 (pass_ipa_lower_emutls): Replace declaration with that of...
3941 (make_pass_ipa_lower_emutls): ...new function.
3942 (pass_ipa_function_and_variable_visibility): Replace declaration with
3943 that of...
3944 (make_pass_ipa_function_and_variable_visibility): ...new function.
3945 (pass_ipa_tree_profile): Replace declaration with that of...
3946 (make_pass_ipa_tree_profile): ...new function.
3947 (pass_early_local_passes): Replace declaration with that of...
3948 (make_pass_early_local_passes): ...new function.
3949 (pass_ipa_whole_program_visibility): Replace declaration with that
3950 of...
3951 (make_pass_ipa_whole_program_visibility): ...new function.
3952 (pass_ipa_lto_gimple_out): Replace declaration with that of...
3953 (make_pass_ipa_lto_gimple_out): ...new function.
3954 (pass_ipa_increase_alignment): Replace declaration with that of...
3955 (make_pass_ipa_increase_alignment): ...new function.
3956 (pass_ipa_inline): Replace declaration with that of...
3957 (make_pass_ipa_inline): ...new function.
3958 (pass_ipa_free_lang_data): Replace declaration with that of...
3959 (make_pass_ipa_free_lang_data): ...new function.
3960 (pass_ipa_free_inline_summary): Replace declaration with that of...
3961 (make_pass_ipa_free_inline_summary): ...new function.
3962 (pass_ipa_cp): Replace declaration with that of...
3963 (make_pass_ipa_cp): ...new function.
3964 (pass_ipa_reference): Replace declaration with that of...
3965 (make_pass_ipa_reference): ...new function.
3966 (pass_ipa_pure_const): Replace declaration with that of...
3967 (make_pass_ipa_pure_const): ...new function.
3968 (pass_ipa_pta): Replace declaration with that of...
3969 (make_pass_ipa_pta): ...new function.
3970 (pass_ipa_lto_finish_out): Replace declaration with that of...
3971 (make_pass_ipa_lto_finish_out): ...new function.
3972 (pass_ipa_tm): Replace declaration with that of...
3973 (make_pass_ipa_tm): ...new function.
3974 (pass_ipa_profile): Replace declaration with that of...
3975 (make_pass_ipa_profile): ...new function.
3976 (pass_ipa_cdtor_merge): Replace declaration with that of...
3977 (make_pass_ipa_cdtor_merge): ...new function.
3978 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
3979 of...
3980 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3981 (pass_init_datastructures): Replace declaration with that of...
3982 (make_pass_init_datastructures): ...new function.
3983 (pass_fixup_cfg): Replace declaration with that of...
3984 (make_pass_fixup_cfg): ...new function.
3985 (pass_expand): Replace declaration with that of...
3986 (make_pass_expand): ...new function.
3987 (pass_instantiate_virtual_regs): Replace declaration with that of...
3988 (make_pass_instantiate_virtual_regs): ...new function.
3989 (pass_rtl_fwprop): Replace declaration with that of...
3990 (make_pass_rtl_fwprop): ...new function.
3991 (pass_rtl_fwprop_addr): Replace declaration with that of...
3992 (make_pass_rtl_fwprop_addr): ...new function.
3993 (pass_jump): Replace declaration with that of...
3994 (make_pass_jump): ...new function.
3995 (pass_jump2): Replace declaration with that of...
3996 (make_pass_jump2): ...new function.
3997 (pass_lower_subreg): Replace declaration with that of...
3998 (make_pass_lower_subreg): ...new function.
3999 (pass_cse): Replace declaration with that of...
4000 (make_pass_cse): ...new function.
4001 (pass_fast_rtl_dce): Replace declaration with that of...
4002 (make_pass_fast_rtl_dce): ...new function.
4003 (pass_ud_rtl_dce): Replace declaration with that of...
4004 (make_pass_ud_rtl_dce): ...new function.
4005 (pass_rtl_dce): Replace declaration with that of...
4006 (make_pass_rtl_dce): ...new function.
4007 (pass_rtl_dse1): Replace declaration with that of...
4008 (make_pass_rtl_dse1): ...new function.
4009 (pass_rtl_dse2): Replace declaration with that of...
4010 (make_pass_rtl_dse2): ...new function.
4011 (pass_rtl_dse3): Replace declaration with that of...
4012 (make_pass_rtl_dse3): ...new function.
4013 (pass_rtl_cprop): Replace declaration with that of...
4014 (make_pass_rtl_cprop): ...new function.
4015 (pass_rtl_pre): Replace declaration with that of...
4016 (make_pass_rtl_pre): ...new function.
4017 (pass_rtl_hoist): Replace declaration with that of...
4018 (make_pass_rtl_hoist): ...new function.
4019 (pass_rtl_store_motion): Replace declaration with that of...
4020 (make_pass_rtl_store_motion): ...new function.
4021 (pass_cse_after_global_opts): Replace declaration with that of...
4022 (make_pass_cse_after_global_opts): ...new function.
4023 (pass_rtl_ifcvt): Replace declaration with that of...
4024 (make_pass_rtl_ifcvt): ...new function.
4025 (pass_into_cfg_layout_mode): Replace declaration with that of...
4026 (make_pass_into_cfg_layout_mode): ...new function.
4027 (pass_outof_cfg_layout_mode): Replace declaration with that of...
4028 (make_pass_outof_cfg_layout_mode): ...new function.
4029 (pass_loop2): Replace declaration with that of...
4030 (make_pass_loop2): ...new function.
4031 (pass_rtl_loop_init): Replace declaration with that of...
4032 (make_pass_rtl_loop_init): ...new function.
4033 (pass_rtl_move_loop_invariants): Replace declaration with that of...
4034 (make_pass_rtl_move_loop_invariants): ...new function.
4035 (pass_rtl_unswitch): Replace declaration with that of...
4036 (make_pass_rtl_unswitch): ...new function.
4037 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
4038 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4039 (pass_rtl_doloop): Replace declaration with that of...
4040 (make_pass_rtl_doloop): ...new function.
4041 (pass_rtl_loop_done): Replace declaration with that of...
4042 (make_pass_rtl_loop_done): ...new function.
4043 (pass_web): Replace declaration with that of...
4044 (make_pass_web): ...new function.
4045 (pass_cse2): Replace declaration with that of...
4046 (make_pass_cse2): ...new function.
4047 (pass_df_initialize_opt): Replace declaration with that of...
4048 (make_pass_df_initialize_opt): ...new function.
4049 (pass_df_initialize_no_opt): Replace declaration with that of...
4050 (make_pass_df_initialize_no_opt): ...new function.
4051 (pass_reginfo_init): Replace declaration with that of...
4052 (make_pass_reginfo_init): ...new function.
4053 (pass_inc_dec): Replace declaration with that of...
4054 (make_pass_inc_dec): ...new function.
4055 (pass_stack_ptr_mod): Replace declaration with that of...
4056 (make_pass_stack_ptr_mod): ...new function.
4057 (pass_initialize_regs): Replace declaration with that of...
4058 (make_pass_initialize_regs): ...new function.
4059 (pass_combine): Replace declaration with that of...
4060 (make_pass_combine): ...new function.
4061 (pass_if_after_combine): Replace declaration with that of...
4062 (make_pass_if_after_combine): ...new function.
4063 (pass_ree): Replace declaration with that of...
4064 (make_pass_ree): ...new function.
4065 (pass_partition_blocks): Replace declaration with that of...
4066 (make_pass_partition_blocks): ...new function.
4067 (pass_match_asm_constraints): Replace declaration with that of...
4068 (make_pass_match_asm_constraints): ...new function.
4069 (pass_regmove): Replace declaration with that of...
4070 (make_pass_regmove): ...new function.
4071 (pass_split_all_insns): Replace declaration with that of...
4072 (make_pass_split_all_insns): ...new function.
4073 (pass_fast_rtl_byte_dce): Replace declaration with that of...
4074 (make_pass_fast_rtl_byte_dce): ...new function.
4075 (pass_lower_subreg2): Replace declaration with that of...
4076 (make_pass_lower_subreg2): ...new function.
4077 (pass_mode_switching): Replace declaration with that of...
4078 (make_pass_mode_switching): ...new function.
4079 (pass_sms): Replace declaration with that of...
4080 (make_pass_sms): ...new function.
4081 (pass_sched): Replace declaration with that of...
4082 (make_pass_sched): ...new function.
4083 (pass_ira): Replace declaration with that of...
4084 (make_pass_ira): ...new function.
4085 (pass_reload): Replace declaration with that of...
4086 (make_pass_reload): ...new function.
4087 (pass_clean_state): Replace declaration with that of...
4088 (make_pass_clean_state): ...new function.
4089 (pass_branch_prob): Replace declaration with that of...
4090 (make_pass_branch_prob): ...new function.
4091 (pass_value_profile_transformations): Replace declaration with that
4092 of...
4093 (make_pass_value_profile_transformations): ...new function.
4094 (pass_postreload_cse): Replace declaration with that of...
4095 (make_pass_postreload_cse): ...new function.
4096 (pass_gcse2): Replace declaration with that of...
4097 (make_pass_gcse2): ...new function.
4098 (pass_split_after_reload): Replace declaration with that of...
4099 (make_pass_split_after_reload): ...new function.
4100 (pass_branch_target_load_optimize1): Replace declaration with that
4101 of...
4102 (make_pass_branch_target_load_optimize1): ...new function.
4103 (pass_thread_prologue_and_epilogue): Replace declaration with that
4104 of...
4105 (make_pass_thread_prologue_and_epilogue): ...new function.
4106 (pass_stack_adjustments): Replace declaration with that of...
4107 (make_pass_stack_adjustments): ...new function.
4108 (pass_peephole2): Replace declaration with that of...
4109 (make_pass_peephole2): ...new function.
4110 (pass_if_after_reload): Replace declaration with that of...
4111 (make_pass_if_after_reload): ...new function.
4112 (pass_regrename): Replace declaration with that of...
4113 (make_pass_regrename): ...new function.
4114 (pass_cprop_hardreg): Replace declaration with that of...
4115 (make_pass_cprop_hardreg): ...new function.
4116 (pass_reorder_blocks): Replace declaration with that of...
4117 (make_pass_reorder_blocks): ...new function.
4118 (pass_branch_target_load_optimize2): Replace declaration with that
4119 of...
4120 (make_pass_branch_target_load_optimize2): ...new function.
4121 (pass_leaf_regs): Replace declaration with that of...
4122 (make_pass_leaf_regs): ...new function.
4123 (pass_split_before_sched2): Replace declaration with that of...
4124 (make_pass_split_before_sched2): ...new function.
4125 (pass_compare_elim_after_reload): Replace declaration with that of...
4126 (make_pass_compare_elim_after_reload): ...new function.
4127 (pass_sched2): Replace declaration with that of...
4128 (make_pass_sched2): ...new function.
4129 (pass_stack_regs): Replace declaration with that of...
4130 (make_pass_stack_regs): ...new function.
4131 (pass_stack_regs_run): Replace declaration with that of...
4132 (make_pass_stack_regs_run): ...new function.
4133 (pass_df_finish): Replace declaration with that of...
4134 (make_pass_df_finish): ...new function.
4135 (pass_compute_alignments): Replace declaration with that of...
4136 (make_pass_compute_alignments): ...new function.
4137 (pass_duplicate_computed_gotos): Replace declaration with that of...
4138 (make_pass_duplicate_computed_gotos): ...new function.
4139 (pass_variable_tracking): Replace declaration with that of...
4140 (make_pass_variable_tracking): ...new function.
4141 (pass_free_cfg): Replace declaration with that of...
4142 (make_pass_free_cfg): ...new function.
4143 (pass_machine_reorg): Replace declaration with that of...
4144 (make_pass_machine_reorg): ...new function.
4145 (pass_cleanup_barriers): Replace declaration with that of...
4146 (make_pass_cleanup_barriers): ...new function.
4147 (pass_delay_slots): Replace declaration with that of...
4148 (make_pass_delay_slots): ...new function.
4149 (pass_split_for_shorten_branches): Replace declaration with that of...
4150 (make_pass_split_for_shorten_branches): ...new function.
4151 (pass_split_before_regstack): Replace declaration with that of...
4152 (make_pass_split_before_regstack): ...new function.
4153 (pass_convert_to_eh_region_ranges): Replace declaration with that
4154 of...
4155 (make_pass_convert_to_eh_region_ranges): ...new function.
4156 (pass_shorten_branches): Replace declaration with that of...
4157 (make_pass_shorten_branches): ...new function.
4158 (pass_set_nothrow_function_flags): Replace declaration with that of...
4159 (make_pass_set_nothrow_function_flags): ...new function.
4160 (pass_dwarf2_frame): Replace declaration with that of...
4161 (make_pass_dwarf2_frame): ...new function.
4162 (pass_final): Replace declaration with that of...
4163 (make_pass_final): ...new function.
4164 (pass_rtl_seqabstr): Replace declaration with that of...
4165 (make_pass_rtl_seqabstr): ...new function.
4166 (pass_release_ssa_names): Replace declaration with that of...
4167 (make_pass_release_ssa_names): ...new function.
4168 (pass_early_inline): Replace declaration with that of...
4169 (make_pass_early_inline): ...new function.
4170 (pass_inline_parameters): Replace declaration with that of...
4171 (make_pass_inline_parameters): ...new function.
4172 (pass_update_address_taken): Replace declaration with that of...
4173 (make_pass_update_address_taken): ...new function.
4174 (pass_convert_switch): Replace declaration with that of...
4175 (make_pass_convert_switch): ...new function.
4176 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
4177 to a subclass of simple_ipa_opt_pass along with...
4178 (pass_data_ipa_tree_profile): ...new pass_data instance and...
4179 (make_pass_ipa_tree_profile): ...new function.
4180 * tree-sra.c (pass_sra_early): Convert from a global struct to a
4181 subclass of gimple_opt_pass along with...
4182 (pass_data_sra_early): ...new pass_data instance and...
4183 (make_pass_sra_early): ...new function.
4184 (pass_sra): Convert from a global struct to a subclass of
4185 gimple_opt_pass along with...
4186 (pass_data_sra): ...new pass_data instance and...
4187 (make_pass_sra): ...new function.
4188 (pass_early_ipa_sra): Convert from a global struct to a subclass of
4189 gimple_opt_pass along with...
4190 (pass_data_early_ipa_sra): ...new pass_data instance and...
4191 (make_pass_early_ipa_sra): ...new function.
4192 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
4193 subclass of gimple_opt_pass along with...
4194 (pass_data_ccp): ...new pass_data instance and...
4195 (make_pass_ccp): ...new function.
4196 (pass_fold_builtins): Convert from a global struct to a subclass of
4197 gimple_opt_pass along with...
4198 (pass_data_fold_builtins): ...new pass_data instance and...
4199 (make_pass_fold_builtins): ...new function.
4200 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
4201 subclass of gimple_opt_pass along with...
4202 (pass_data_copy_prop): ...new pass_data instance and...
4203 (make_pass_copy_prop): ...new function.
4204 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
4205 global struct to a subclass of gimple_opt_pass along with...
4206 (pass_data_rename_ssa_copies): ...new pass_data instance and...
4207 (make_pass_rename_ssa_copies): ...new function.
4208 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
4209 subclass of gimple_opt_pass along with...
4210 (pass_data_dce): ...new pass_data instance and...
4211 (make_pass_dce): ...new function.
4212 (pass_dce_loop): Convert from a global struct to a subclass of
4213 gimple_opt_pass along with...
4214 (pass_data_dce_loop): ...new pass_data instance and...
4215 (make_pass_dce_loop): ...new function.
4216 (pass_cd_dce): Convert from a global struct to a subclass of
4217 gimple_opt_pass along with...
4218 (pass_data_cd_dce): ...new pass_data instance and...
4219 (make_pass_cd_dce): ...new function.
4220 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
4221 subclass of gimple_opt_pass along with...
4222 (pass_data_dominator): ...new pass_data instance and...
4223 (make_pass_dominator): ...new function.
4224 (pass_phi_only_cprop): Convert from a global struct to a subclass of
4225 gimple_opt_pass along with...
4226 (pass_data_phi_only_cprop): ...new pass_data instance and...
4227 (make_pass_phi_only_cprop): ...new function.
4228 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
4229 subclass of gimple_opt_pass along with...
4230 (pass_data_dse): ...new pass_data instance and...
4231 (make_pass_dse): ...new function.
4232 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
4233 a subclass of gimple_opt_pass along with...
4234 (pass_data_forwprop): ...new pass_data instance and...
4235 (make_pass_forwprop): ...new function.
4236 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
4237 struct to a subclass of gimple_opt_pass along with...
4238 (pass_data_tree_ifcombine): ...new pass_data instance and...
4239 (make_pass_tree_ifcombine): ...new function.
4240 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
4241 subclass of gimple_opt_pass along with...
4242 (pass_data_ch): ...new pass_data instance and...
4243 (make_pass_ch): ...new function.
4244 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
4245 subclass of gimple_opt_pass along with...
4246 (pass_data_tree_loop): ...new pass_data instance and...
4247 (make_pass_tree_loop): ...new function.
4248 (pass_tree_loop_init): Convert from a global struct to a subclass of
4249 gimple_opt_pass along with...
4250 (pass_data_tree_loop_init): ...new pass_data instance and...
4251 (make_pass_tree_loop_init): ...new function.
4252 (pass_lim): Convert from a global struct to a subclass of
4253 gimple_opt_pass along with...
4254 (pass_data_lim): ...new pass_data instance and...
4255 (make_pass_lim): ...new function.
4256 (pass_tree_unswitch): Convert from a global struct to a subclass of
4257 gimple_opt_pass along with...
4258 (pass_data_tree_unswitch): ...new pass_data instance and...
4259 (make_pass_tree_unswitch): ...new function.
4260 (pass_predcom): Convert from a global struct to a subclass of
4261 gimple_opt_pass along with...
4262 (pass_data_predcom): ...new pass_data instance and...
4263 (make_pass_predcom): ...new function.
4264 (pass_vectorize): Convert from a global struct to a subclass of
4265 gimple_opt_pass along with...
4266 (pass_data_vectorize): ...new pass_data instance and...
4267 (make_pass_vectorize): ...new function.
4268 (pass_graphite): Convert from a global struct to a subclass of
4269 gimple_opt_pass along with...
4270 (pass_data_graphite): ...new pass_data instance and...
4271 (make_pass_graphite): ...new function.
4272 (pass_graphite_transforms): Convert from a global struct to a subclass
4273 of gimple_opt_pass along with...
4274 (pass_data_graphite_transforms): ...new pass_data instance and...
4275 (make_pass_graphite_transforms): ...new function.
4276 (pass_check_data_deps): Convert from a global struct to a subclass of
4277 gimple_opt_pass along with...
4278 (pass_data_check_data_deps): ...new pass_data instance and...
4279 (make_pass_check_data_deps): ...new function.
4280 (pass_iv_canon): Convert from a global struct to a subclass of
4281 gimple_opt_pass along with...
4282 (pass_data_iv_canon): ...new pass_data instance and...
4283 (make_pass_iv_canon): ...new function.
4284 (pass_scev_cprop): Convert from a global struct to a subclass of
4285 gimple_opt_pass along with...
4286 (pass_data_scev_cprop): ...new pass_data instance and...
4287 (make_pass_scev_cprop): ...new function.
4288 (pass_record_bounds): Convert from a global struct to a subclass of
4289 gimple_opt_pass along with...
4290 (pass_data_record_bounds): ...new pass_data instance and...
4291 (make_pass_record_bounds): ...new function.
4292 (pass_complete_unroll): Convert from a global struct to a subclass of
4293 gimple_opt_pass along with...
4294 (pass_data_complete_unroll): ...new pass_data instance and...
4295 (make_pass_complete_unroll): ...new function.
4296 (pass_complete_unrolli): Convert from a global struct to a subclass of
4297 gimple_opt_pass along with...
4298 (pass_data_complete_unrolli): ...new pass_data instance and...
4299 (make_pass_complete_unrolli): ...new function.
4300 (pass_parallelize_loops): Convert from a global struct to a subclass
4301 of gimple_opt_pass along with...
4302 (pass_data_parallelize_loops): ...new pass_data instance and...
4303 (make_pass_parallelize_loops): ...new function.
4304 (pass_loop_prefetch): Convert from a global struct to a subclass of
4305 gimple_opt_pass along with...
4306 (pass_data_loop_prefetch): ...new pass_data instance and...
4307 (make_pass_loop_prefetch): ...new function.
4308 (pass_iv_optimize): Convert from a global struct to a subclass of
4309 gimple_opt_pass along with...
4310 (pass_data_iv_optimize): ...new pass_data instance and...
4311 (make_pass_iv_optimize): ...new function.
4312 (pass_tree_loop_done): Convert from a global struct to a subclass of
4313 gimple_opt_pass along with...
4314 (pass_data_tree_loop_done): ...new pass_data instance and...
4315 (make_pass_tree_loop_done): ...new function.
4316 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
4317 struct to a subclass of gimple_opt_pass along with...
4318 (pass_data_cse_reciprocals): ...new pass_data instance and...
4319 (make_pass_cse_reciprocals): ...new function.
4320 (pass_cse_sincos): Convert from a global struct to a subclass of
4321 gimple_opt_pass along with...
4322 (pass_data_cse_sincos): ...new pass_data instance and...
4323 (make_pass_cse_sincos): ...new function.
4324 (pass_optimize_bswap): Convert from a global struct to a subclass of
4325 gimple_opt_pass along with...
4326 (pass_data_optimize_bswap): ...new pass_data instance and...
4327 (make_pass_optimize_bswap): ...new function.
4328 (pass_optimize_widening_mul): Convert from a global struct to a
4329 subclass of gimple_opt_pass along with...
4330 (pass_data_optimize_widening_mul): ...new pass_data instance and...
4331 (make_pass_optimize_widening_mul): ...new function.
4332 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
4333 subclass of gimple_opt_pass along with...
4334 (pass_data_phiopt): ...new pass_data instance and...
4335 (make_pass_phiopt): ...new function.
4336 (pass_cselim): Convert from a global struct to a subclass of
4337 gimple_opt_pass along with...
4338 (pass_data_cselim): ...new pass_data instance and...
4339 (make_pass_cselim): ...new function.
4340 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
4341 subclass of gimple_opt_pass along with...
4342 (pass_data_phiprop): ...new pass_data instance and...
4343 (make_pass_phiprop): ...new function.
4344 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
4345 subclass of gimple_opt_pass along with...
4346 (pass_data_pre): ...new pass_data instance and...
4347 (make_pass_pre): ...new function.
4348 (pass_fre): Convert from a global struct to a subclass of
4349 gimple_opt_pass along with...
4350 (pass_data_fre): ...new pass_data instance and...
4351 (make_pass_fre): ...new function.
4352 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
4353 subclass of gimple_opt_pass along with...
4354 (pass_data_reassoc): ...new pass_data instance and...
4355 (make_pass_reassoc): ...new function.
4356 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
4357 subclass of gimple_opt_pass along with...
4358 (pass_data_sink_code): ...new pass_data instance and...
4359 (make_pass_sink_code): ...new function.
4360 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
4361 subclass of gimple_opt_pass along with...
4362 (pass_data_strlen): ...new pass_data instance and...
4363 (make_pass_strlen): ...new function.
4364 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
4365 struct to a subclass of gimple_opt_pass along with...
4366 (pass_data_build_alias): ...new pass_data instance and...
4367 (make_pass_build_alias): ...new function.
4368 (pass_build_ealias): Convert from a global struct to a subclass of
4369 gimple_opt_pass along with...
4370 (pass_data_build_ealias): ...new pass_data instance and...
4371 (make_pass_build_ealias): ...new function.
4372 (pass_ipa_pta): Convert from a global struct to a subclass of
4373 simple_ipa_opt_pass along with...
4374 (pass_data_ipa_pta): ...new pass_data instance and...
4375 (make_pass_ipa_pta): ...new function.
4376 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
4377 subclass of gimple_opt_pass along with...
4378 (pass_data_uncprop): ...new pass_data instance and...
4379 (make_pass_uncprop): ...new function.
4380 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
4381 global struct to a subclass of gimple_opt_pass along with...
4382 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
4383 (make_pass_late_warn_uninitialized): ...new function.
4384 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
4385 to a subclass of gimple_opt_pass along with...
4386 (pass_data_init_datastructures): ...new pass_data instance and...
4387 (make_pass_init_datastructures): ...new function.
4388 (pass_early_warn_uninitialized): Convert from a global struct to a
4389 subclass of gimple_opt_pass along with...
4390 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
4391 (make_pass_early_warn_uninitialized): ...new function.
4392 (pass_update_address_taken): Convert from a global struct to a
4393 subclass of gimple_opt_pass along with...
4394 (pass_data_update_address_taken): ...new pass_data instance and...
4395 (make_pass_update_address_taken): ...new function.
4396 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
4397 struct to a subclass of gimple_opt_pass along with...
4398 (pass_data_release_ssa_names): ...new pass_data instance and...
4399 (make_pass_release_ssa_names): ...new function.
4400 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
4401 subclass of gimple_opt_pass along with...
4402 (pass_data_stdarg): ...new pass_data instance and...
4403 (make_pass_stdarg): ...new function.
4404 * tree-switch-conversion.c (pass_convert_switch): Convert from a
4405 global struct to a subclass of gimple_opt_pass along with...
4406 (pass_data_convert_switch): ...new pass_data instance and...
4407 (make_pass_convert_switch): ...new function.
4408 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
4409 to a subclass of gimple_opt_pass along with...
4410 (pass_data_tail_recursion): ...new pass_data instance and...
4411 (make_pass_tail_recursion): ...new function.
4412 (pass_tail_calls): Convert from a global struct to a subclass of
4413 gimple_opt_pass along with...
4414 (pass_data_tail_calls): ...new pass_data instance and...
4415 (make_pass_tail_calls): ...new function.
4416 * tree-vect-generic.c (pass_lower_vector): Convert from a global
4417 struct to a subclass of gimple_opt_pass along with...
4418 (pass_data_lower_vector): ...new pass_data instance and...
4419 (make_pass_lower_vector): ...new function.
4420 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
4421 gimple_opt_pass along with...
4422 (pass_data_lower_vector_ssa): ...new pass_data instance and...
4423 (make_pass_lower_vector_ssa): ...new function.
4424 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
4425 to a subclass of gimple_opt_pass along with...
4426 (pass_data_slp_vectorize): ...new pass_data instance and...
4427 (make_pass_slp_vectorize): ...new function.
4428 (pass_ipa_increase_alignment): Convert from a global struct to a
4429 subclass of simple_ipa_opt_pass along with...
4430 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
4431 (make_pass_ipa_increase_alignment): ...new function.
4432 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
4433 gimple_opt_pass along with...
4434 (pass_data_vrp): ...new pass_data instance and...
4435 (make_pass_vrp): ...new function.
4436 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
4437 subclass of simple_ipa_opt_pass along with...
4438 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
4439 (make_pass_ipa_free_lang_data): ...new function.
4440 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
4441 gimple_opt_pass along with...
4442 (pass_data_tsan): ...new pass_data instance and...
4443 (make_pass_tsan): ...new function.
4444 (pass_tsan_O0): Convert from a global struct to a subclass of
4445 gimple_opt_pass along with...
4446 (pass_data_tsan_O0): ...new pass_data instance and...
4447 (make_pass_tsan_O0): ...new function.
4448 * var-tracking.c (pass_variable_tracking): Convert from a global
4449 struct to a subclass of rtl_opt_pass along with...
4450 (pass_data_variable_tracking): ...new pass_data instance and...
4451 (make_pass_variable_tracking): ...new function.
4452 * web.c (pass_web): Convert from a global struct to a subclass of
4453 rtl_opt_pass along with...
4454 (pass_data_web): ...new pass_data instance and...
4455 (make_pass_web): ...new function.
4456 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
4457 declaration with that of...
4458 (make_pass_mode_switch_use): ...new function.
4459 (pass_resolve_sw_modes): Replace declaration with that of...
4460 (make_pass_resolve_sw_modes): ...new function.
4461 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
4462 from a global struct to a subclass of rtl_opt_pass along with...
4463 (pass_data_mode_switch_use): ...new pass_data instance and...
4464 (make_pass_mode_switch_use): ...new function.
4465 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
4466 from a global struct to a subclass of rtl_opt_pass along with...
4467 (pass_data_resolve_sw_modes): ...new pass_data instance and...
4468 (make_pass_resolve_sw_modes): ...new function.
4469 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
4470 struct to a subclass of rtl_opt_pass along with...
4471 (pass_data_insert_vzeroupper): ...new pass_data instance and...
4472 (make_pass_insert_vzeroupper): ...new function.
4473 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
4474 global struct to a subclass of rtl_opt_pass along with...
4475 (pass_data_work_around_errata): ...new pass_data instance and...
4476 (make_pass_work_around_errata): ...new function.
4477 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
4478 struct to a subclass of rtl_opt_pass along with...
4479 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
4480 (make_pass_mips_machine_reorg2): ...new function.
4481
6a389ed5
DM
44822013-08-05 David Malcolm <dmalcolm@redhat.com>
4483
4484 * passes.c (pass_manager::operator new): New.
4485
f7695dbf
DM
44862013-08-05 David Malcolm <dmalcolm@redhat.com>
4487
4488 Handwritten part of conversion of passes to C++ classes.
4489
4490 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
4491 (toplev.o): Add dep on PASS_MANAGER_H.
4492 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
4493 of early local pases to reflect this moving from a global to a
4494 member of gcc::pass_manager.
4495 (cgraph_add_new_function): Likewise.
4496 * lto-cgraph.c (lto_output_node): Update for conversion of
4497 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
4498 * passes.c (opt_pass::clone): New.
4499 (opt_pass::gate): New.
4500 (opt_pass::execute): New.
4501 (opt_pass::opt_pass): New.
4502 (pass_manager::execute_early_local_passes): New.
4503 (pass_manager::execute_pass_mode_switching): new.
4504 (finish_optimization_passes): Convert to...
4505 (pass_manager::finish_optimization_passes): ...this.
4506 (finish_optimization_passes): Update for conversion of passes to
4507 C++ classes.
4508 (register_dump_files_1): Use has_gate since we cannot portably
4509 check a vtable entry against NULL.
4510 (dump_one_pass): Likewise.
4511 (ipa_write_summaries_2): Likewise.
4512 (ipa_write_optimization_summaries_1): Likewise.
4513 (ipa_read_summaries_1): Likewise.
4514 (ipa_read_optimization_summaries_1): Likewise.
4515 (execute_ipa_stmt_fixups): Likewise.
4516 (pass_manager::pass_manager): Rewrite pass-creation, invoking
4517 pass-creation functions rather than wiring up globals, and
4518 storing the results in fields of pass_manager generated using
4519 pass-instances.def.
4520 (pass_manager::dump_profile_report): Update for conversion of
4521 passes to C++ classes.
4522 (pass_manager::execute_ipa_summary_passes): Likewise.
4523 (execute_one_ipa_transform_pass): Likewise.
4524 (execute_one_pass): Use has_gate and has_execute since we cannot
4525 portably check a vtable entry against NULL.
4526 * pass_manager.h (pass_manager::finish_optimization_passes): New.
4527 (pass_manager): Use pass-instances.def to add fields for the
4528 various pass instances.
4529 * toplev.c (finalize): Update for move of
4530 finish_optimization_passes to a method of gcc::pass_manager.
4531 * toplev.h (finish_optimization_passes): Move to method of class
4532 pass_manager.
4533 * tree-pass.h (struct pass_data): New.
03b0ee0a 4534 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM
4535 (opt_pass::gate): Convert to virtual function.
4536 (opt_pass::~opt_pass): New.
4537 (opt_pass::clone): New.
4538 (opt_pass::execute): Convert to virtual function.
4539 (opt_pass::opt_pass): New.
4540 (opt_pass::ctxt_): new.
4541 (gimple_opt_pass): Convert to subclass of opt_pass.
4542 (gimple_opt_pass::gimple_opt_pass): New.
4543 (rtl_opt_pass): Convert to subclass of opt_pass.
4544 (rtl_opt_pass::rtl_opt_pass): New.
4545 (ipa_opt_pass_d): Convert to subclass of opt_pass.
4546 (ipa_opt_pass_d::ipa_opt_pass_d): New.
4547 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
4548 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
4549 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
4550 invocation of pass_mode_switching to reflect this moving from a
4551 global to a member of gcc::pass_manager.
4552 (ix86_option_override): Rework how pass_insert_vzeroupper is
4553 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 4554 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 4555
d0b2f831
RE
45562013-08-05 Richard Earnshaw <rearnsha@arm.com>
4557
4558 PR rtl-optimization/57708
4559 * recog.c (peep2_find_free_register): Validate all regs in a
4560 multi-reg mode.
4561
51a5c0c2
JH
45622013-08-05 Jan Hubicka <jh@suse.cz>
4563
4564 PR lto/57602
03b0ee0a
UB
4565 * cgraph.c (verify_cgraph_node): Accept local flags from other
4566 partitions.
51a5c0c2
JH
4567 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
4568 (function_and_variable_visibility): Likewise.
4569 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
4570
65f0a120
GDR
45712013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4572
4573 * graph.c (init_graph_slim_pretty_print): Remove.
4574 (print_graph_cfg): Do not call it. Use local pretty printer.
4575 (start_graph_dump): Likewise.
4576
b3f80694
GDR
45772013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4578
4579 * gimple-pretty-print.c (buffer): Remove.
4580 (initialized): Likewise.
4581 (maybe_init_pretty_print): Likewise.
4582 (print_gimple_stmt): Do not call it. Use non-static local
4583 pretty_printer variable.
4584 (print_gimple_expr): Likewise.
4585 (print_gimple_seq): Likewise.
4586 (gimple_dump_bb): Likewise.
4587
11a877b3
GDR
45882013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
4589
4590 * asan.c (asan_pp): Remove.
4591 (asan_pp_initialized): Likewise.
4592 (asan_pp_initialize): Likewise.
4593 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
4594 (asan_emit_stack_protection): Tidy. Use local pretty printer.
4595 (asan_add_global): Likewise.
4596
b066401f
GDR
45972013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
4598
4599 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
4600 * diagnostic.h (diagnostic_flush_buffer): Adjust.
4601 * pretty-print.c (pp_formatted_text_data): Likewise.
4602 (pp_indent): Rename from pp_base_indent.
4603 (pp_format): Rename from pp_base_format.
4604 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
4605 (pp_format_verbatim): Rename from pp_base_format_verbatim.
4606 (pp_flush): Rename from pp_base_flush.
4607 (pp_set_line_maximum_length): Rename from
4608 pp_base_set_line_maximum_length.
4609 (pp_clear_output_area): Rename from pp_base_clear_output_area.
4610 (pp_set_prefix): Rename from pp_base_set_prefix.
4611 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
4612 (pp_emit_prefix): Rename from pp_base_emit_prefix.
4613 (pp_append_text): Rename from pp_base_append_text.
4614 (pp_formatted_text): Rename from pp_base_formatted_text.
4615 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
4616 (pp_remaining_character_count_for_line): Rename from
4617 pp_base_remaining_character_count_for_line.
4618 (pp_newline): Rename from pp_base_newline.
4619 (pp_character): Rename from pp_base_character.
4620 (pp_string): Rename from pp_base_string.
4621 (pp_maybe_space): Rename from pp_base_maybe_space.
4622 * asan.c (asan_pp_string): Adjust.
4623 (asan_emit_stack_protection): Likewise.
4624 (asan_add_global): Likewise.
4625 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
4626 * tree-mudflap.c (mf_varname_tree): Likewise.
4627 * tree-pretty-print.c (pp_tree_identifier): Rename from
4628 pp_base_tree_identifier.
4629 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
4630 Declare as function.
4631
137a1a27
GDR
46322013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
4633
4634 * pretty-print.h (pp_bar_bar): New.
4635 (pp_ampersand_ampersand): Likewise.
4636 (pp_less_equal): Likewise.
4637 (pp_greater_equal): Likewise.
4638 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
4639 printer functions instead of pp_string or operators and punctuators.
4640 (dump_gimple_call): Likewise.
4641 (dump_gimple_omp_for): Likewise.
4642 (dump_gimple_transaction): Likewise.
4643 (dump_gimple_phi): Likewise.
4644 (pp_gimple_stmt_1): Likewise.
4645 * sched-vis.c (print_insn): Likewise.
4646 * tree-mudflap.c (mf_varname_tree): Likewise.
4647 * tree-pretty-print.c (dump_block_node): Likewise.
4648 (dump_generic_node): Likewise.
4649
815effe1
JH
46502013-08-02 Jan Hubicka <jh@suse.cz>
4651
4652 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
4653 boundaries.
4654 * lto-streamer-out.c (tree_is_indexable): Results decls and
4655 parm decls are not indexable.
4656 (DFS_write_tree_body): Do not follow args and results.
4657 (hash_tree): Likewise.
4658 (output_functions): Rearrange so struct function is needed
4659 only when real body is output; be able to also ouptut abstract
4660 functions; output DECL_ARGUMENTS and DECL_RESULT.
4661 (lto_output): When not in WPA, ale store abstract functions.
4662 (write_symbol): Do not care about RESULT_DECL.
4663 (output_symbol_p): Handle correctly sbtract decls.
4664 * lto-streamer-in.c (input_function): Rearrange so struct
4665 function can be NULL at entry; allow streaming of
4666 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
4667 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
4668 sanity check during LTO.
4669 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
4670 RESULT_DECl and DECL_ARGUMENTS.
4671 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
4672 Likewise.
4673
07838b13
GDR
46742013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
4675
4676 * pretty-print.h (pp_underscore): New.
4677 (pp_comma): Tidy.
4678 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
4679 printer functions instead of pp_character.
4680 (dump_binary_rhs): Likewise.
4681 (dump_ternary_rhs): Likewise.
4682 (dump_gimple_call_args): Likewise.
4683 (pp_points_to_solution): Likewise.
4684 (dump_gimple_call): Likewise.
4685 (dump_gimple_switch): Likewise.
4686 (dump_gimple_cond): Likewise.
4687 (dump_gimple_bind): Likewise.
4688 (dump_gimple_try): Likewise.
4689 (dump_gimple_omp_for): Likewise.
4690 (dump_gimple_omp_continue): Likewise.
4691 (dump_gimple_omp_single): Likewise.
4692 (dump_gimple_omp_sections): Likewise.
4693 (dump_gimple_omp_block): Likewise.
4694 (dump_gimple_omp_critical): Likewise.
4695 (dump_gimple_transaction): Likewise.
4696 (dump_gimple_asm): Likewise.
4697 (dump_gimple_phi): Likewise.
4698 (dump_gimple_omp_parallel): Likewise.
4699 (dump_gimple_omp_task): Likewise.
4700 (dump_gimple_omp_atomic_load): Likewise.
4701 (dump_gimple_omp_atomic_store): Likewise.
4702 (dump_gimple_mem_ops): Likewise.
4703 (pp_gimple_stmt_1): Likewise.
4704 (pp_cfg_jump): Likewise.
4705 (dump_implicit_edges): Likewise.
4706 (gimple_dump_bb_for_graph): Likewise.
4707 * graph.c (draw_cfg_node): Likewise.
4708 * langhooks.c (lhd_print_error_function): Likewise.
4709 * sched-vis.c (print_exp): Likewise.
4710 (print_value): Likewise.
4711 (print_pattern): Likewise.
4712 (print_insn): Likewise.
4713 (rtl_dump_bb_for_graph): Likewise.
4714 * tree-pretty-print.c (dump_function_declaration): Likewise.
4715 (dump_array_domain): Likewise.
4716 (dump_omp_clause): Likewise.
4717 (dump_location): Likewise.
4718 (dump_generic_node): Likewise.
4719 (print_struct_decl): Likewise.
4720 * diagnostic.c (diagnostic_show_locus): Use pp_space.
4721
bb0d2039
BS
47222013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4723
4724 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
4725 candidate table when replacing a candidate statement.
4726 (replace_rhs_if_not_dup): Likewise.
4727 (replace_one_candidate): Likewise.
4728
bc0ec027 47292013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 4730 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH
4731
4732 * cgraphunit.c (add_new_function): Fix logic when adding from
4733 late IPA pass.
4734 (assemble_thunk): Rename to ...
03b0ee0a
UB
4735 (expand_thunk); .. this one; export; get it working with
4736 general functions; make produced gimple valid.
bc0ec027
JH
4737 * cgraph.h (expand_thunk): Declare.
4738
0e8853ee
JH
47392013-08-02 Jan Hubicka <jh@suse.cz>
4740
03b0ee0a
UB
4741 * ipa-cp.c (gather_context_independent_values): Use
4742 ipa_get_param_move_cost.
4743 (get_replacement_map): Remove PARAM; move parameter folding
4744 into tree-inline.c
0e8853ee
JH
4745 (create_specialized_node): Update.
4746 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
4747 assert that we have gimple body; update move_cost.
4748 (count_formal_params): Assert that we have gimple body.
4749 (ipa_dump_param): New function.
4750 (ipa_alloc_node_params): Break out from ...
4751 (ipa_initialize_node_params): ... here.
4752 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
4753 (ipa_write_node_info): Stream move costs.
4754 (ipa_read_node_info): Read move costs.
4755 (ipa_update_after_lto_read): Do not recompute node params.
4756 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
4757 (ipa_get_param): Check we are not in WPA.
4758 (ipa_get_param_move_cost): New.
4759 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
4760 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
4761 parm numbers to be present.
4762
ecee672b
VM
47632013-08-02 Vladimir Makarov <vmakarov@redhat.com>
4764
4765 PR rtl-optimization/58048
4766 * lra-constraints.c (process_alt_operands): Don't check asm
4767 operand on register.
4768
13ec0527
EB
47692013-08-02 Eric Botcazou <ebotcazou@adacore.com>
4770
4771 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
4772 the implied StoreLoad barrier for atomic operations if before.
4773
af15184a
JH
47742013-08-02 Jan Hubicka <jh@suse.cz>
4775 Martin Liska <marxin.liska@gmail.com>
4776
03b0ee0a
UB
4777 * cgraph.c (cgraph_function_body_availability): Do not check
4778 cgraph flags.
af15184a
JH
4779 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
4780 symtab_node_availability): Declare.
4781 * ipa.c (can_replace_by_local_alias): New.
4782 (function_and_variable_visibility): Use it.
03b0ee0a
UB
4783 * symtab.c (symtab_for_node_and_aliases,
4784 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 4785
01e54ef8
VM
47862013-08-02 Vladimir Makarov <vmakarov@redhat.com>
4787
4788 PR rtl-optimization/57963
03b0ee0a 4789 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM
4790 (lra_constraints): Use them.
4791
e2fd7ca7
SN
47922013-08-02 Sofiane Naci <sofiane.naci@arm.com>
4793
03b0ee0a
UB
4794 * config/arm/types.md (define_attr "type"): Add "load_acq"
4795 and "store_rel".
e2fd7ca7
SN
4796 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
4797 changes.
4798 (cortex_a53_store1): Likewise.
4799
0dcc26c3
JH
48002013-08-01 Jan Hubicka <jh@suse.cz>
4801
03b0ee0a
UB
4802 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
4803 partitions are not needed.
0dcc26c3 4804
0bd72901
UB
48052013-08-01 Uros Bizjak <ubizjak@gmail.com>
4806
4807 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
4808 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
4809 MAYBE_NON_Q_CLASS_P where appropriate.
4810
12123452
JH
48112013-08-01 Jan Hubicka <jh@suse.cz>
4812
4813 * cgraph.h (release_function_body): Declare.
4814 * tree.c (free_lang_data_in_decl): Free, parameters and return values
4815 of unused delcarations.
4816
ce852f9c
KT
48172013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4818
03b0ee0a
UB
4819 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
4820 RTL form when subtracting a constant.
ce852f9c 4821
c743b246
KT
48222013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4823
4824 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
4825 Generate canonical plus rtx with negated immediate instead of minus
4826 where appropriate.
4827 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
4828
c0c123ef
JH
48292013-08-01 Jan Hubicka <jh@suse.cz>
4830
4831 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
4832 (cgraph_release_function_body): Likewise.
4833 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
4834 * cgraph.h (cgrpah_node): Rename abstract_and_needed
4835 to used_as_abstract_origin.
4836 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
4837 symbols used as abstract origins.
4838 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB
4839 * ipa.c (symtab_remove_unreachable_nodes): Recompute
4840 used_as_abstract_origin.
c0c123ef 4841 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB
4842 used_as_abstract_origin; be ready for DECL_RESULT and
4843 DECL_ARGUMENTS to be NULL.
c0c123ef 4844
0bd72901
UB
4845 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
4846 for abstract functions.
4847 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
4848 real symbols.
c0c123ef 4849
1f029433
JH
48502013-08-01 Jan Hubicka <jh@suse.cz>
4851
4852 * profile.c (compute_value_histograms): Fix thinko.
4853
c451f4d6
SN
48542013-08-01 Sofiane Naci <sofiane.naci@arm.com>
4855
4856 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
4857 aarch-common-protos.h to extra_headers.
4858 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
4859 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
4860 * config/aarch64/t-aarch64 (aarch-common.o): Define.
4861
34b512f5
SN
48622013-08-01 Sofiane Naci <sofiane.naci@arm.com>
4863
4864 * config/aarch64/aarch64.md (define_attr "type"): Delete.
4865 Include "../arm/types.md". Define "type" attribute for all patterns.
4866 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
4867 attribute changes.
4868
d86e633a
MM
48692013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
4870
4871 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
4872 to support power8 load fusion.
4873 (fusion_gpr_mem_load): Likewise.
4874
4875 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
4876
4877 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
4878 declarations for power8 load fusion.
4879 (emit_fusion_gpr_load): Likewise.
4880
4881 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
4882 tuning for power8, turn on fusion mode by default. Turn on sign
4883 extending fusion mode if normal fusion mode is on, and we are at
4884 -O2 or -O3.
4885 (fusion_gpr_load_p): New function, return true if we can fuse an
4886 addis instruction with a dependent load to a GPR.
4887 (emit_fusion_gpr_load): Emit the instructions for power8 load
4888 fusion to GPRs.
4889
0bd72901 4890 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM
4891 (VSX load fusion peepholes): New peepholes to fuse together an
4892 addi instruction with a VSX load instruction.
4893
4894 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
4895 peepholes to fuse an addis instruction with a load to a GPR base
4896 register. If we are supporting sign extending fusions, convert
4897 sign extending loads to zero extending loads and add an explicit
4898 sign extension.
4899
c3f35647
SN
49002013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4901
4902 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
4903 aarch-common-protos.h to extra_headers.
4904 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
4905 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
4906 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
4907 (arm_no_early_alu_shift_dep): Likewise.
4908 (arm_no_early_alu_shift_value_dep): Likewise.
4909 (arm_no_early_mul_dep): Likewise.
4910 (arm_no_early_store_addr_dep): Likewise.
4911 (arm_mac_accumulator_is_mul_result): Likewise.
4912 (arm_mac_accumulator_is_result): Likewise.
c3f35647 4913 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB
4914 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
4915 here to ...
c3f35647 4916 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
4917 (arm_no_early_alu_shift_dep): Likewise.
4918 (arm_no_early_alu_shift_value_dep): Likewise.
4919 (arm_no_early_mul_dep): Likewise.
4920 (arm_no_early_store_addr_dep): Likewise.
4921 (arm_mac_accumulator_is_mul_result): Likewise.
4922 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN
4923 * config/arm/aarch-common-protos.h: ... here. New file.
4924 * config/arm/t-arm (aarch-common.o): Define.
4925
49262013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4927
4928 * config/arm/arm.md: Include new file "types.md".
4929 (define_attr "type"): Move from here to ...
4930 (define_attr "mul32"): Likewise.
4931 (define_attr "mul64"): Likewise.
4932 * config/arm/types.md: ... here. New file.
4933
228c1313
SH
49342013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
4935
4936 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
4937 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
4938
d6e9046d 49392013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG
4940
4941 * gen-pass-instances.awk: Fix offset of substr().
4942
a167b052
DM
49432013-07-31 David Malcolm <dmalcolm@redhat.com>
4944
4945 * Makefile.in (pass-instances.def): New.
4946 (passes.o): Replace dependency on passes.def with one on
4947 pass-instances.def
4948
4949 * gen-pass-instances.awk: New.
4950
4951 * passes.c (pass_manager::pass_manager): Use pass-instances.def
4952 rather than passes.def, updating local definition of NEXT_PASS
4953 macro to add an extra NUM parameter (currently unused).
4954
315f8c0e
DM
49552013-07-30 David Malcolm <dmalcolm@redhat.com>
4956
4957 * Makefile.in (PASS_MANAGER_H): New.
4958 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
4959 (passes.o): Likewise.
4960 (statistics.o): Likewise.
4961 (cgraphunit.o): Likewise.
4962 (context.o): Depend on PASS_MANAGER_H.
4963
4964 * pass_manager.h: New.
4965
4966 * cgraphunit.c (cgraph_add_new_function): Update for moves
4967 of globals to fields of pass_manager.
4968 (analyze_function): Likewise.
4969 (expand_function): Likewise.
4970 (ipa_passes): Likewise.
4971 (compile): Likewise.
4972
4973 * context.c (context::context): New.
4974 * context.h (context::context): New.
4975 (context::get_passes): New.
4976 (context::passes_): New.
4977
4978 * lto-cgraph.c (input_node): Update for moves of globals to
4979 fields of pass_manager.
4980
4981 * passes.c (all_passes): Remove, in favor of a field of the
4982 same name within the new class pass_manager.
4983 (all_small_ipa_passes): Likewise.
4984 (all_lowering_passes): Likewise.
4985 (all_regular_ipa_passes): Likewise.
4986 (all_late_ipa_passes): Likewise.
4987 (all_lto_gen_passes): Likewise.
4988 (passes_by_id): Likewise.
4989 (passes_by_id_size): Likewise.
4990 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4991 the new class pass_manager.
4992 (set_pass_for_id): Convert to...
4993 (pass_manager::set_pass_for_id): ...method.
4994 (get_pass_for_id): Convert to...
4995 (pass_manager::get_pass_for_id): ...method.
4996 (register_one_dump_file): Move body of implementation into...
4997 (pass_manager::register_one_dump_file): ...here.
4998 (register_dump_files_1): Convert to...
4999 (pass_manager::register_dump_files_1): ...method.
5000 (register_dump_files): Convert to...
5001 (pass_manager::register_dump_files): ...method.
5002 (create_pass_tab): Update for moves of globals to fields of
5003 pass_manager.
5004 (dump_passes): Move body of implementation into...
5005 (pass_manager::dump_passes): ...here.
5006 (register_pass): Move body of implementation into...
5007 (pass_manager::register_pass): ...here.
5008 (init_optimization_passes): Convert into...
5009 (pass_manager::pass_manager): ...constructor for new
5010 pass_manager class, and initialize the pass_lists array.
5011 (check_profile_consistency): Update for moves of globals to
5012 fields of pass_manager.
5013 (dump_profile_report): Move body of implementation into...
5014 (pass_manager::dump_profile_report): ...here.
5015 (ipa_write_summaries_1): Update for moves of pass lists from
5016 being globals to fields of pass_manager.
5017 (ipa_write_optimization_summaries): Likewise.
5018 (ipa_read_summaries): Likewise.
5019 (ipa_read_optimization_summaries): Likewise.
5020 (execute_all_ipa_stmt_fixups): Likewise.
5021
5022 * statistics.c (statistics_fini): Update for moves of globals to
5023 fields of pass_manager.
5024
5025 * toplev.c (general_init): Replace call to
5026 init_optimization_passes with construction of the pass_manager
5027 instance.
5028
5029 * tree-pass.h (all_passes): Remove, in favor of a field of the
5030 same name within the new class pass_manager.
5031 (all_small_ipa_passes): Likewise.
5032 (all_lowering_passes): Likewise.
5033 (all_regular_ipa_passes): Likewise.
5034 (all_lto_gen_passes): Likewise.
5035 (all_late_ipa_passes): Likewise.
5036 (passes_by_id): Likewise.
5037 (passes_by_id_size): Likewise.
5038 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
5039 the new class pass_manager.
5040 (get_pass_for_id): Remove.
5041
02cab1c1
RE
50422013-07-30 Richard Earnshaw <rearnsha@arm.com>
5043
5044 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
5045 configs.
5046
b2d3f886
RE
50472013-07-30 Richard Earnshaw <rearnsha@arm.com>
5048
5049 * arm.md (mulhi3): New expand pattern.
5050
605e86fa
JH
50512013-07-30 Jan Hubicka <jh@suse.cz>
5052 Martin Liska <marxin.liska@gmail.com>
5053
5054 * profile.c (compute_value_histograms): Do not ICE when
5055 there is mismatch only on some counters.
5056
14a87636
ZC
50572013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5058
5059 PR rtl-optimization/57637
5060 * function.c (move_insn_for_shrink_wrap): Also check the
5061 GEN set of the LIVE problem for the liveness analysis
5062 if it exists, otherwise give up.
5063
0100cd3f
BS
50642013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5065
5066 PR tree-optimization/57993
5067 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
5068 replaced statement in the candidate table.
5069 (phi_add_costs): Return infinite cost when the hidden basis does
5070 not dominate all phis on which the candidate is dependent.
5071 (replace_one_candidate): Record replaced statement in the
5072 candidate table.
5073
f0d811f7
JR
50742013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
5075
5076 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
5077 (ashlv2si3): New expander.
5078 (*ashlv2si3_i): New define_insn_and_split.
5079 * predicates.md (float_operation): Allow patterns with three
5080 basic sub-patterns.
5081
bba33211
JR
5082 PR rtl-optimization/58021
5083 * mode-switching.c (create_pre_exit): Always split off preceding
5084 insns if we are not at the basic block head.
5085
ff3f3951
MR
50862013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
5087
5088 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
5089 (UCLIBC_DYNAMIC_LINKER): New macro.
5090 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
5091 `-mnan=2008'.
5092 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
5093 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
5094 `-mnan=2008'.
5095 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
5096 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
5097 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
5098 for SF and DF modes. Use ieee_quad_format for TF mode.
5099 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
5100 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
5101 (mips_option_override): Handle `-mnan=legacy'.
5102 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
5103 `-mabs=2008' and `-mnan=2008'.
5104 (OPTION_DEFAULT_SPECS): Add "nan" default.
5105 (ASM_SPEC): Handle `-mnan='.
5106 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
5107 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
5108 comment accordingly.
5109 (neg<mode>2): Likewise.
5110 * config/mips/mips.opt (mabs, mnan): New options.
5111 * doc/install.texi (Configuration): Document `--with-nan=' option.
5112 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
5113 `-mnan=' options.
5114 (MIPS Options): Document them.
5115 * config.gcc <mips*-*-*>: Handle `--with-nan='.
5116 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
5117 * configure: Regenerate.
5118 * config.in: Regenerate.
5119
576e2f37
UB
51202013-07-29 Uros Bizjak <ubizjak@gmail.com>
5121
5122 * config/i386/i386.md (float post-reload splitters): Do not check
5123 for subregs of SSE registers.
5124
51252013-07-29 Uros Bizjak <ubizjak@gmail.com>
5126 H.J. Lu <hongjiu.lu@intel.com>
5127
5128 PR target/57954
5129 PR target/57988
5130 * config/i386/i386.md (post-reload splitter
5131 to avoid partial SSE reg dependency stalls): New pattern.
5132
a71f0749
DV
51332013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
5134
5135 * config/s390/s390.md ("movcc"): Swap load and store instructions.
5136
4c97f1cc
JR
51372013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
5138
5139 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
5140 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
5141
b0c733d8
CC
51422013-07-26 Cary Coutant <ccoutant@google.com>
5143
5144 * dwarf2out.c (die_checksum_ordered): Don't include template
5145 instantiations in signature.
5146 (is_template_parameter): New function.
5147 (is_template_instantiation): New function.
5148 (generate_skeleton_bottom_up): Don't include template instantiations
5149 in type unit DIE.
5150 (generate_skeleton): Likewise.
5151 (break_out_comdat_types): Move recursive call to break out nested
5152 types earlier.
5153 (prune_unused_types_mark_generic_parms_dies): Call
5154 is_template_parameter.
5155
040d8a1c
IB
51562013-07-26 Ian Bolton <ian.bolton@arm.com>
5157
5158 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
5159 uses vector registers.
89fdc743 5160 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 5161
85bd4ac6 51622013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 5163 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT
5164
5165 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 5166 where appropriate.
85bd4ac6
KT
5167 * config/arm/arm.md (movcond_addsi): New splitter.
5168
b6051207
SA
51692013-07-25 Sterling Augustine <saugustine@google.com>
5170
5171 * dwarf2out.c (size_of_pubnames): Move code to...
5172 (include_pubname_in_output): ...here. New.
5173 (want_pubnames): Rearrange.
5174 (output_pubnames): Call include_pubname_in_output. Move assertion.
5175
41a7c215
CM
51762013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5177
5178 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
5179
51802013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
5181
5182 PR target/38836
5183 * doc/extend.texi: Remove obsolete builtins. Fix
5184 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
5185
49bde175
JH
51862013-07-25 Jan Hubicka <jh@suse.cz>
5187
5188 * cgraph.c (release_function_body): Break out from ...
5189 (cgraph_release_function_body): ... this one; also release DECL_RESULT
5190 and DECL_ARGUMENTS.
5191 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
5192 old_tree in the map.
5193 (create_specialized_node): Update.
5194 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
5195 into index.
41a7c215
CM
5196 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
5197 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH
5198 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
5199 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
5200 DECL_RESULT.
5201
4870352d
KT
52022013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5203
5204 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
5205 addsi3_carryin_alt2_<optab>): Correct output template.
5206
1d6aee1c
KT
52072013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5208
5209 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
5210 Adjust for arm_restrict_it.
5211 Remove trailing whitespace.
5212
c7cafd75
MK
52132013-07-25  Mark Kettenis  <kettenis@openbsd.org>
5214
017d38f5
MK
5215 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
5216 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
5217
c7cafd75
MK
5218 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
5219
00b7527b
VM
52202013-07-25 Vladimir Makarov <vmakarov@redhat.com>
5221
5222 PR rtl-optimization/57960
5223 * lra-constraints.c (process_alt_operands): Use the right mode
5224 when checking strict_low.
5225
4f63dfc6
JH
52262013-07-25 Jan Hubicka <jh@suse.cz>
5227
5228 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
5229 * cgraph.c (cgraph_remove_node): Do not release function body
5230 when in cgraph streaming.
5231 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM
5232 in other partitions are not considered reachable; fix handling of
5233 clones.
4f63dfc6 5234
e6dcfa18
RR
52352013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5236
5237 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
5238
0b93d3b6
RR
52392013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5240
5241 PR target/19599
5242 PR target/57731
bb80c2eb 5243 PR target/57837
e6dcfa18 5244 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 5245 Ss with US. Adjust output for v5 and v4t.
41a7c215 5246 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 5247
0b93d3b6
RR
5248 * config/arm/constraints.md ("Ss"): Rename to US.
5249
690688b3
TG
52502013-07-25 Terry Guo <terry.guo@arm.com>
5251
5252 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
5253 shift_add/shift_sub0/shift_sub1 RTXs.
5254
a4ad093b 52552013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 5256 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS
5257
5258 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
5259 (altivec_vpks<VI_char>ss): Likewise.
5260 (altivec_vpks<VI_char>us): Likewise.
5261 (altivec_vpku<VI_char>us): Likewise.
5262 (altivec_vpku<VI_char>um): Likewise.
5263
8c5005ce
DM
52642013-07-24 David Malcolm <dmalcolm@redhat.com>
5265
5266 Introduce context class.
5267
5268 * Makefile.in (CONTEXT_H): New.
5269 (OBJS): Add context.o.
5270 (toplev.o): Add CONTEXT_H to dependencies.
5271 (context.o): New.
5272
41a7c215 5273 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM
5274
5275 * context.c: New.
5276
5277 * context.h: New.
5278
ce4a9422
JR
52792013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
5280
5281 PR rtl-optimization/57968
5282 * mode-switching.c (create_pre_exit): Allow instructions that
5283 don't set a return register to need a non-exit mode.
5284
493f4c9e 52852013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 5286 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 5287
b5860fd3
WS
5288 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
5289 operands to vperm for little endian.
5290 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
5291 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 5292
73792b92 52932013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 5294 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS
5295
5296 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
5297 two operands for little-endian.
5298
14c21302
SE
52992013-07-23 Steve Ellcey <sellcey@mips.com>
5300
5301 * config/mips/mips.c (mips_case_values_threshold): New.
5302 (TARGET_CASE_VALUES_THRESHOLD): Define.
5303
a6056198 53042013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 5305 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS
5306
5307 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
5308 selection of field for vector splat in little endian mode.
5309
dd7a40e1
MM
53102013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
5311
5312 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
5313 expanders to rs6000.md.
5314 (ior<mode>3): Likewise.
5315 (and<mode>3): Likewise.
5316 (one_cmpl<mode>2): Likewise.
5317 (nor<mode>3): Likewise.
5318 (andc<mode>3): Likewise.
5319 (eqv<mode>3): Likewise.
5320 (nand<mode>3): Likewise.
5321 (orc<mode>3): Likewise.
5322
5323 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
5324 declaration.
5325
5326 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
5327 to split multi-word logical operations.
5328 (rs6000_split_logical_di): Likewise.
5329 (rs6000_split_logical): Likewise.
5330
5331 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
5332 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
5333 and allow TImode operations in 32-bit.
5334 (vsx_and<mode>3_64bit): Likewise.
5335 (vsx_ior<mode>3_32bit): Likewise.
5336 (vsx_ior<mode>3_64bit): Likewise.
5337 (vsx_xor<mode>3_32bit): Likewise.
5338 (vsx_xor<mode>3_64bit): Likewise.
5339 (vsx_one_cmpl<mode>2_32bit): Likewise.
5340 (vsx_one_cmpl<mode>2_64bit): Likewise.
5341 (vsx_nor<mode>3_32bit): Likewise.
5342 (vsx_nor<mode>3_64bit): Likewise.
5343 (vsx_andc<mode>3_32bit): Likewise.
5344 (vsx_andc<mode>3_64bit): Likewise.
5345 (vsx_eqv<mode>3_32bit): Likewise.
5346 (vsx_eqv<mode>3_64bit): Likewise.
5347 (vsx_nand<mode>3_32bit): Likewise.
5348 (vsx_nand<mode>3_64bit): Likewise.
5349 (vsx_orc<mode>3_32bit): Likewise.
5350 (vsx_orc<mode>3_64bit): Likewise.
5351
5352 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
5353 logical types in GPRs.
5354
5355 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
5356 logical insns to rs6000.md, and allow TImode operations in
5357 32-bit.
5358 (altivec_ior<mode>3): Likewise.
5359 (altivec_xor<mode>3): Likewise.
5360 (altivec_one_cmpl<mode>2): Likewise.
5361 (altivec_nor<mode>3): Likewise.
5362 (altivec_andc<mode>3): Likewise.
5363
5364 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
5365 attributes for moving the 128-bit logical operations into
5366 rs6000.md.
5367 (BOOL_REGS_OUTPUT): Likewise.
5368 (BOOL_REGS_OP1): Likewise.
5369 (BOOL_REGS_OP2): Likewise.
5370 (BOOL_REGS_UNARY): Likewise.
5371 (BOOL_REGS_AND_CR0): Likewise.
5372 (one_cmpl<mode>2): Add support for DI logical operations on
5373 32-bit, splitting the operations to 32-bit.
5374 (anddi3): Likewise.
5375 (iordi3): Likewise.
5376 (xordi3): Likewise.
5377 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
5378 changes to combine the 32/64-bit code, allow logical operations on
5379 TI mode in 32-bit, and to use similar match_operator patterns like
5380 scalar mode uses. Combine the Altivec and VSX code for logical
5381 operations, and move it here.
5382 (ior<mode>3, 128-bit types): Likewise.
5383 (xor<mode>3, 128-bit types): Likewise.
5384 (one_cmpl<mode>3, 128-bit types): Likewise.
5385 (nor<mode>3, 128-bit types): Likewise.
5386 (andc<mode>3, 128-bit types): Likewise.
5387 (eqv<mode>3, 128-bit types): Likewise.
5388 (nand<mode>3, 128-bit types): Likewise.
5389 (orc<mode>3, 128-bit types): Likewise.
5390 (and<mode>3_internal): Likewise.
5391 (bool<mode>3_internal): Likewise.
5392 (boolc<mode>3_internal1): Likewise.
5393 (boolc<mode>3_internal2): Likewise.
5394 (boolcc<mode>3_internal1): Likewise.
5395 (boolcc<mode>3_internal2): Likewise.
5396 (eqv<mode>3_internal1): Likewise.
5397 (eqv<mode>3_internal2): Likewise.
5398 (one_cmpl1<mode>3_internal): Likewise.
5399
12211b99 54002013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH
5401
5402 * config/microblaze/microblaze.c (microblaze_expand_prologue):
5403 Rename flag_stack_usage to flag_stack_usage_info.
5404
12211b99 54052013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 5406
a5f257fd
ME
5407 * config/microblaze/sync.md: New file.
5408 * config/microblaze/microblaze.md: Include sync.md
5409 * config/microblaze/microblaze.c: Add print_operand 'y'.
5410 * config/microblaze/constraints.md: Add memory_contraint
5411 'Q' which is a single register.
450b0ebb 5412
49fba14f
EB
54132013-07-23 Eric Botcazou <ebotcazou@adacore.com>
5414
5415 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
5416
c7b3b99f
PCC
54172013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
5418
5419 * reload.c (find_reloads): Exit loop once we find this operand
5420 cannot be reloaded somehow for this alternative.
5421
5422 * reload.c (find_reloads): Exit loop once we find a hard register.
5423
5424 * rtlanal.c (computed_jump_p): Exit loop once we find label
5425 reference is used.
5426
5427 * i386.c (ix86_pad_returns): Exit loop after setting replace.
5428
5429 * cfgloopmanip.c (remove_path): Exit loop after setting
5430 irred_invalidated.
5431
5432 * gensupport.c (subst_dup): Avoid loop if code is not
5433 MATCH_DUP nor MATCH_OP_DUP.
5434
c8fbf1fa
NBJ
54352013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
5436
5437 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
5438
9259db42
YZ
54392013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5440
5441 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
5442 true for SP_REGNUM if mode == ptr_mode.
5443 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
5444 with value R0_REGNUM + 31.
5445
00edcfbe
YZ
54462013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5447
5448 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
5449 pad pointer-typed argument downward.
5450
43be9a95
YZ
54512013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5452
5453 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
5454 and __ILP32__ when the ILP32 model is in use.
5455
28514dda
YZ
54562013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5457
5458 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
5459 (aarch64_load_symref_appropriately): In the case of
5460 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
5461 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
5462 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
5463 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
5464 if 'mode' doesn't equal to 'ptr_mode'.
5465 (aarch64_output_mi_thunk): Add an assertion on the alignment of
5466 'vcall_offset'; change to call aarch64_emit_move differently depending
5467 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
5468 to calculate the upper bound of 'vcall_offset'.
5469 (aarch64_cannot_force_const_mem): Change to also return true if
5470 mode != ptr_mode.
5471 (aarch64_legitimize_reload_address): In the case of large
5472 displacements, add new local variable 'xmode' and an assertion
5473 based on it; change to use 'xmode' to generate the new rtx and
5474 reload.
5475 (aarch64_asm_trampoline_template): Change to generate the template
5476 differently depending on TARGET_ILP32 or not; change to use
5477 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
5478 (aarch64_trampoline_size): Removed.
5479 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
5480 and replace immediate literals with it. Change to use 'ptr_mode'
5481 instead of 'DImode' and call convert_memory_address if the mode
5482 of 'fnaddr' doesn't equal to 'ptr_mode'.
5483 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
5484 to output symbol.
5485 (aarch64_elf_asm_destructor): Likewise.
5486 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
5487 on TARGET_ILP32 instead of aarch64_trampoline_size.
5488 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
5489 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
5490 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
5491 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
5492 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
5493 (storewb_pair<GPI:mode>_<P:mode>): ... this.
5494 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
5495 depending on the value of 'mode'.
5496 (add_losym_<mode>): New.
5497 (ldr_got_small_<mode>): New, based on ldr_got_small.
5498 (ldr_got_small): Remove.
5499 (ldr_got_small_sidi): New.
5500 * config/aarch64/iterators.md (P): New.
5501 (PTR): Change to 'ptr_mode' in the condition.
5502
17a819cb
YZ
55032013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
5504
5505 * config.gcc (aarch64*-*-*): Support --with-abi.
5506 (aarch64*-*-elf): Support --with-multilib-list.
5507 (aarch64*-*-linux*): Likewise.
5508 (supported_defaults): Add abi to aarch64*-*-*.
5509 * configure.ac: Mention AArch64 for --with-multilib-list.
5510 * configure: Re-generated.
5511 * config/aarch64/biarchilp32.h: New file.
5512 * config/aarch64/biarchlp64.h: New file.
5513 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
5514 (ABI_SPEC): Ditto.
5515 (MULTILIB_DEFAULTS): Ditto.
5516 (DRIVER_SELF_SPECS): Ditto.
5517 (ASM_SPEC): Update to also substitute -mabi.
5518 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
5519 file whose name depends on -mabi= and -mbig-endian.
5520 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
5521 TARGET_ILP32.
5522 (POINTER_SIZE): New define.
5523 (POINTERS_EXTEND_UNSIGNED): Ditto.
5524 (enum aarch64_abi_type): New enumeration tag.
5525 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
5526 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
5527 (TARGET_ILP32): New define.
5528 * config/aarch64/aarch64.opt (mabi): New.
5529 (aarch64_abi): New.
5530 (ilp32, lp64): New values for -mabi.
5531 * config/aarch64/t-aarch64 (comma): New define.
5532 (MULTILIB_OPTIONS): Ditto.
5533 (MULTILIB_DIRNAMES): Ditto.
5534 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
5535 * doc/invoke.texi: Document -mabi for AArch64.
5536
9f9cbdce
GJL
55372013-07-23 Georg-Johann Lay <avr@gjlay.de>
5538
5539 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
5540
7061977b 55412013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 5542 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS
5543
5544 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
5545 endianness when selecting field to splat.
5546
75a41878
EC
55472013-07-22 Eric Christopher <echristo@gmail.com>
5548
5549 * dwarf2out.c (die_odr_checksum): New function to use
5550 CHECKSUM_ macros and ULEB128 for DIE tag.
5551 (generate_type_signature): Use.
41a7c215 5552
38ae58ca
EB
55532013-07-22 Eric Botcazou <ebotcazou@adacore.com>
5554
5555 * config.gcc (sparc*-*-*): Accept leon3 processor.
5556 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
5557 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
5558 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
5559 * config/sparc/sparc.opt (enum processor_type): Add leon3.
5560 (mfix-ut699): Adjust comment.
5561 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
5562 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
5563 (CPP_CPU_SPEC): Likewise.
5564 (ASM_CPU_SPEC): Likewise.
5565 * config/sparc/sparc.c (leon3_cost): New constant.
5566 (sparc_option_override): Add leon3 support.
5567 (mem_ref): New function.
5568 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
5569 (sparc_do_work_around_errata): Look into the instruction in the delay
5570 slot and adjust accordingly. Add fix for the data cache nullify issues
5571 of the UT699. Change insertion position for the NOP.
5572 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
5573 (leon3_load): New reservation.
5574 (leon_store): Bump latency to 2.
5575 (grfpu): New automaton.
5576 (grfpu_alu): New unit.
5577 (grfpu_ds): Likewise.
5578 (leon_fp_alu): Adjust.
5579 (leon_fp_mult): Delete.
5580 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
5581 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
5582 * config/sparc/sparc.md (cpu): Add leon3.
5583 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
5584 (swapsi): Likewise.
5585 (atomic_test_and_set): Likewise.
5586 (ldstub): Likewise.
5587
04dfc6df
JU
55882013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
5589
5590 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
5591 default for R5900 targets.
5592 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
5593 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
5594 * config/mips/mips.c (mips_option_override): Report an error for
5595 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
5596 for -march=r5900 -mhard-float.
5597
c7b3b99f 55982013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC
5599
5600 * df-problems.c (can_move_insns_across): Exit loop once we
5601 find a non-fixed, non-global register.
5602
5603 * ipa-pure-const.c (propagate_nothrow): Exit loop after
5604 setting can_throw.
5605
5606 * omega.c (omega_eliminate_red): Break after setting red_found.
5607 (omega_problem_has_red_equations): Similarly after setting found.
5608 (omega_query_variable): Similarly after setting coupled.
5609
9b6e6981
MP
56102013-07-22 Marek Polacek <polacek@redhat.com>
5611
5612 * gimplify.c: Don't include gimple.h twice.
5613
5e5f7673
KT
56142013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5615
5616 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
5617 instead of TARGET_THUMB1.
5618 (Pz): New constraint.
5619 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
5620 encodings.
5621 (compare_negsi_si): Likewise.
5622 (compare_addsi2_op0): Likewise.
5623 (compare_addsi2_op1): Likewise.
5624 (addsi3_carryin_<optab>): Likewise.
5625 (addsi3_carryin_alt2_<optab>): Likewise.
5626 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
5627 for arm_restrict_it.
5628 (subsi3_carryin): Likewise.
5629 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
5630 (minmax_arithsi): Disable for arm_restrict_it.
5631 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
5632 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
5633 (satsi_<SAT:code>_shift): Likewise.
5634 (arm_shiftsi3): Add alternative for 16-bit encoding.
5635 (arm32_movhf): Disable for arm_restrict_it.
5636 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
5637 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
5638
859abddd
SN
56392013-07-22 Sofiane Naci <sofiane.naci@arm.com>
5640
5641 * config/arm/arm.md (attribute "insn"): Delete.
5642 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
5643 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
5644 (not_shiftsi): Update for attribute change.
5645 (not_shiftsi_compare0): Likewise.
5646 (not_shiftsi_compare0_scratch): Likewise.
5647 (arm_one_cmplsi2): Likewise.
5648 (thumb1_one_cmplsi2): Likewise.
5649 (notsi_compare0): Likewise.
5650 (notsi_compare0_scratch): Likewise.
5651 (thumb1_movdi_insn): Likewise.
5652 (arm_movsi_insn): Likewise.
5653 (movhi_insn_arch4): Likewise.
5654 (movhi_bytes): Likewise.
5655 (arm_movqi_insn): Likewise.
5656 (thumb1_movqi_insn): Likewise.
5657 (arm32_movhf): Likewise.
5658 (thumb1_movhf): Likewise.
5659 (arm_movsf_soft_insn): Likewise.
5660 (thumb1_movsf_insn): Likewise.
5661 (thumb_movdf_insn): Likewise.
5662 (movsicc_insn): Likewise.
5663 (movsfcc_soft_insn): Likewise.
5664 (and_scc): Likewise.
5665 (cond_move): Likewise.
5666 (if_move_not): Likewise.
5667 (if_not_move): Likewise.
5668 (if_shift_move): Likewise.
5669 (if_move_shift): Likewise.
5670 (if_shift_shift): Likewise.
5671 (if_not_arith): Likewise.
5672 (if_arith_not): Likewise.
5673 (cond_move_not): Likewise.
5674 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
5675 (neon_mov<mode>): Likewise.
5676 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
5677 (thumb2_movsi_vfp): Likewise.
5678 (movsf_vfp): Likewise.
5679 (thumb2_movsf_vfp): Likewise.
371e77e3
UB
5680 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5681 change.
859abddd
SN
5682 (cortexa7_older_only): Likewise.
5683 (cortexa7_younger): Likewise.
5684 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
5685 (1020alu_shift_op): Likewise.
5686 (1020alu_shift_reg_op): Likewise.
5687 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
5688 (alu_shift_op): Likewise.
5689 (alu_shift_reg_op): Likewise.
5690 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
5691 (11_alu_shift_op): Likewise.
5692 (11_alu_shift_reg_op): Likewise.
5693 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
5694 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
5695 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5696 change.
859abddd
SN
5697 (cortex_a15_alu_shift): Likewise.
5698 (cortex_a15_alu_shift_reg): Likewise.
5699 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
5700 (cortex_a5_alu_shift): Likewise.
371e77e3
UB
5701 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
5702 change.
859abddd 5703 (cortex_a53_alu_shift): Likewise.
371e77e3
UB
5704 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5705 change.
859abddd
SN
5706 (cortex_a7_alu_reg): Likewise.
5707 (cortex_a7_alu_shift): Likewise.
5708 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5709 (cortex_a8_alu_shift): Likewise.
5710 (cortex_a8_alu_shift_reg): Likewise.
5711 (cortex_a8_mov): Likewise.
5712 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
5713 (cortex_a9_dp_shift): Likewise.
5714 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
5715 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
5716 (cortex_r4_mov): Likewise.
5717 (cortex_r4_alu_shift): Likewise.
5718 (cortex_r4_alu_shift_reg): Likewise.
5719 * config/arm/fa526.md (526_alu_op): Update for attribute change.
5720 (526_alu_shift_op): Likewise.
5721 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
5722 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
5723 (626te_alu_shift_op): Likewise.
5724 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
5725 (726te_alu_op): Likewise.
5726 (726te_alu_shift_op): Likewise.
5727 (726te_alu_shift_reg_op): Likewise.
5728 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
5729 (mp626_alu_shift_op): Likewise.
5730 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
5731 (pj4_alu_e1_conds): Likewise.
5732 (pj4_alu): Likewise.
5733 (pj4_alu_conds): Likewise.
5734 (pj4_shift): Likewise.
5735 (pj4_shift_conds): Likewise.
5736 (pj4_alu_shift): Likewise.
5737 (pj4_alu_shift_conds): Likewise.
5738
95b97fac
KT
57392013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5740
5741 * config/arm/predicates.md (shiftable_operator_strict_it):
5742 New predicate.
5743 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
5744 Disable cond_exec version for arm_restrict_it.
5745 (thumb2_smaxsi3): Convert to generate cond_exec.
5746 (thumb2_sminsi3): Likewise.
5747 (thumb32_umaxsi3): Likewise.
5748 (thumb2_uminsi3): Likewise.
5749 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
5750 (thumb2_neg_abssi2): Likewise.
5751 (thumb2_mov_scc): Add alternative for 16-bit encoding.
5752 (thumb2_movsicc_insn): Adjust alternatives.
5753 (thumb2_mov_negscc): Disable for arm_restrict_it.
5754 (thumb2_mov_negscc_strict_it): New pattern.
5755 (thumb2_mov_notscc_strict_it): New pattern.
5756 (thumb2_mov_notscc): Disable for arm_restrict_it.
5757 (thumb2_ior_scc): Likewise.
5758 (thumb2_ior_scc_strict_it): New pattern.
5759 (thumb2_cond_move): Adjust for arm_restrict_it.
5760 (thumb2_cond_arith): Disable for arm_restrict_it.
5761 (thumb2_cond_arith_strict_it): New pattern.
5762 (thumb2_cond_sub): Adjust for arm_restrict_it.
5763 (thumb2_movcond): Likewise.
5764 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
5765 (thumb2_zero_extendhisi2_v6): Likewise.
5766 (thumb2_zero_extendqisi2_v6): Likewise.
5767 (orsi_notsi_si): Likewise.
5768 (orsi_not_shiftsi_si): Likewise.
5769
d5a6ef82
GJL
57702013-07-22 Georg-Johann Lay <avr@gjlay.de>
5771
5772 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
5773 instruction sequence is 1 byte shorter.
5774
d9e71154
UB
57752013-07-22 Uros Bizjak <ubizjak@gmail.com>
5776
5777 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
5778 it is not needed after split.
5779
8c6381ec
IS
57802013-07-20 Iain Sandoe <iain@codesourcery.com>
5781
5782 PR target/51784
5783 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
5784 second label for nonlocal goto receivers. Don't output pic base labels
5785 unless we're producing PIC; mark that action unreachable().
5786 (ix86_save_reg): If the function contains a nonlocal label, save the
5787 PIC base reg.
5788 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
371e77e3 5789 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS
5790 (update_pic_label_number_if_needed): New.
5791 (machopic_output_function_base_name): Adjust for nonlocal receiver
5792 case.
5793 (machopic_should_output_picbase_label): New.
5794 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
5795 (nonlocal_goto_receiver): New insn and split.
5796
d05d0709
JG
57972013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
5798
5799 * config/aarch64/aarch64-builtins.c
5800 (aarch64_fold_builtin): Fold abs in all modes.
5801 * config/aarch64/aarch64-simd-builtins.def
5802 (abs): Enable for all modes.
5803 * config/aarch64/arm_neon.h
5804 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
5805 (vabs_f64): Add missing intrinsic.
5806
ba590e63
IB
58072013-07-19 Ian Bolton <ian.bolton@arm.com>
5808
5809 * config/aarch64/arm_neon.h (vabs_s64): New function
5810
02371798
GJL
58112013-07-19 Georg-Johann Lay <avr@gjlay.de>
5812
5813 PR target/57516
5814 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
5815 * config/avr/avr.md (adjust_len): Add `round'.
5816 * config/avr/avr-protos.h (avr_out_round): New prototype.
5817 (avr_out_plus): Add `out_label' argument.
5818 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
5819 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
5820 Handle the case where `insn' is just a pattern.
5821 (avr_out_bitop): Handle the case where `insn' is just a pattern.
5822 (avr_out_round): New function.
5823 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
5824
12211b99 58252013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH
5826
5827 * config/microblaze/microblaze.c (microblaze_expand_prologue):
5828 Add check for flag_stack_usage to handle -fstack-usage support
5829
12211b99 58302013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH
5831
5832 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
5833 interaction for new Power8 flags and VSX.
5834
df9dda2d
ST
58352013-07-18 Sriraman Tallam <tmsriram@google.com>
5836
5837 PR middle-end/57698
5838 * tree-inline.c (expand_call_inline): Emit errors during
5839 early_inlining only if optimization is not turned on.
5840
0f9cdef4
DM
58412013-07-18 David Malcolm <dmalcolm@redhat.com>
5842
5843 * passes.def: New.
5844
5845 * passes.c (init_optimization_passes): Move the construction of
5846 the pass hierarchy into a new passes.def file.
5847
5848 * Makefile.in (passes.o): Add dependency on passes.def.
5849
2efa4087
DM
58502013-07-18 David Malcolm <dmalcolm@redhat.com>
5851
5852 * passes.c (init_optimization_passes): Introduce macros for
5853 constructing the tree of passes (INSERT_PASSES_AFTER,
5854 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
5855 TERMINATE_PASS_LIST).
5856
bc404e1b
VM
58572013-07-18 Vladimir Makarov <vmakarov@redhat.com>
5858 Wei Mi <wmi@google.com>
5859
5860 PR rtl-optimization/57878
5861 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
5862 top.
5863 (reload_pseudo_compare_func): Check nregs first for reload
5864 pseudos.
5865
d675843f
DM
58662013-07-18 David Malcolm <dmalcolm@redhat.com>
5867
5868 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
5869
c7b3b99f 58702013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC
5871
5872 * read-rtl.c (validate_const_int): Once an invalid character is
5873 seen, quit the loop.
5874
5875 * gengtype.c (write_roots): Similarly once we find the "deletable"
5876 or "if_marked" option.
5877
52fceb44
SN
58782013-07-18 Sofiane Naci <sofiane.naci@arm.com>
5879
5880 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
5881 "xtab" and "sat". Move value "clz" from here to ...
5882 (attriubte "type"): ... here.
5883 (satsi_<SAT:code>): Delete "insn" attribute.
5884 (satsi_<SAT:code>_shift): Likewise.
5885 (arm_zero_extendqisi2addsi): Likewise.
5886 (arm_extendqisi2addsi): Likewise.
5887 (clzsi2): Update for attribute changes.
5888 (rbitsi2): Likewise.
371e77e3
UB
5889 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
5890 attribute.
52fceb44
SN
5891 (arm_usatsihi): Likewise.
5892 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
5893
006bd006
SN
58942013-07-18 Sofiane Naci <sofiane.naci@arm.com>
5895
5896 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB
5897 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
5898 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN
5899 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
5900 in alphabetical order.
5901 (attribute "core_cycles"): Update for attribute changes.
5902 (arm_addsi3): Likewise.
5903 (addsi3_compare0): Likewise.
5904 (addsi3_compare0_scratch): Likewise.
5905 (addsi3_compare_op1): Likewise.
5906 (addsi3_compare_op2): Likewise.
5907 (compare_addsi2_op0): Likewise.
5908 (compare_addsi2_op1): Likewise.
5909 (addsi3_carryin_shift_<optab>): Likewise.
5910 (subsi3_carryin_shift): Likewise.
5911 (rsbsi3_carryin_shift): Likewise.
5912 (arm_subsi3_insn): Likewise.
5913 (subsi3_compare0): Likewise.
5914 (subsi3_compare): Likewise.
5915 (arm_andsi3_insn): Likewise.
5916 (thumb1_andsi3_insn): Likewise.
5917 (andsi3_compare0): Likewise.
5918 (andsi3_compare0_scratch): Likewise.
5919 (zeroextractsi_compare0_scratch
5920 (andsi_not_shiftsi_si): Likewise.
5921 (iorsi3_insn): Likewise.
5922 (iorsi3_compare0): Likewise.
5923 (iorsi3_compare0_scratch): Likewise.
5924 (arm_xorsi3): Likewise.
5925 (thumb1_xorsi3_insn): Likewise.
5926 (xorsi3_compare0): Likewise.
5927 (xorsi3_compare0_scratch): Likewise.
5928 (satsi_<SAT:code>_shift): Likewise.
5929 (rrx): Likewise.
5930 (arm_shiftsi3): Likewise.
5931 (shiftsi3_compare0): Likewise.
5932 (not_shiftsi): Likewise.
5933 (not_shiftsi_compare0): Likewise.
5934 (not_shiftsi_compare0_scratch): Likewise.
5935 (arm_one_cmplsi2): Likewise.
5936 (thumb_one_complsi2): Likewise.
5937 (notsi_compare0): Likewise.
5938 (notsi_compare0_scratch): Likewise.
5939 (thumb1_zero_extendhisi2): Likewise.
5940 (arm_zero_extendhisi2): Likewise.
5941 (arm_zero_extendhisi2_v6): Likewise.
5942 (arm_zero_extendhisi2addsi): Likewise.
5943 (thumb1_zero_extendqisi2): Likewise.
5944 (thumb1_zero_extendqisi2_v6): Likewise.
5945 (arm_zero_extendqisi2): Likewise.
5946 (arm_zero_extendqisi2_v6): Likewise.
5947 (arm_zero_extendqisi2addsi): Likewise.
5948 (thumb1_extendhisi2): Likewise.
5949 (arm_extendhisi2): Likewise.
5950 (arm_extendhisi2_v6): Likewise.
5951 (arm_extendqisi): Likewise.
5952 (arm_extendqisi_v6): Likewise.
5953 (arm_extendqisi2addsi): Likewise.
5954 (thumb1_extendqisi2): Likewise.
5955 (thumb1_movdi_insn): Likewise.
5956 (arm_movsi_insn): Likewise.
5957 (movsi_compare0): Likewise.
5958 (movhi_insn_arch4): Likewise.
5959 (movhi_bytes): Likewise.
5960 (arm_movqi_insn): Likewise.
5961 (thumb1_movqi_insn): Likewise.
5962 (arm32_movhf): Likewise.
5963 (thumb1_movhf): Likewise.
5964 (arm_movsf_soft_insn): Likewise.
5965 (thumb1_movsf_insn): Likewise.
5966 (movdf_soft_insn): Likewise.
5967 (thumb_movdf_insn): Likewise.
5968 (arm_cmpsi_insn): Likewise.
5969 (cmpsi_shiftsi): Likewise.
5970 (cmpsi_shiftsi_swp): Likewise.
5971 (arm_cmpsi_negshiftsi_si): Likewise.
5972 (movsicc_insn): Likewise.
5973 (movsfcc_soft_insn): Likewise.
5974 (arith_shiftsi): Likewise.
5975 (arith_shiftsi_compare0
5976 (arith_shiftsi_compare0_scratch
5977 (sub_shiftsi): Likewise.
5978 (sub_shiftsi_compare0
5979 (sub_shiftsi_compare0_scratch
5980 (and_scc): Likewise.
5981 (cond_move): Likewise.
5982 (if_plus_move): Likewise.
5983 (if_move_plus): Likewise.
5984 (if_move_not): Likewise.
5985 (if_not_move): Likewise.
5986 (if_shift_move): Likewise.
5987 (if_move_shift): Likewise.
5988 (if_shift_shift): Likewise.
5989 (if_not_arith): Likewise.
5990 (if_arith_not): Likewise.
5991 (cond_move_not): Likewise.
5992 (thumb1_ashlsi3): Set type attribute.
5993 (thumb1_ashrsi3): Likewise.
5994 (thumb1_lshrsi3): Likewise.
5995 (thumb1_rotrsi3): Likewise.
5996 (shiftsi3_compare0_scratch): Likewise.
5997 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
5998 (neon_mov<mode>): Likewise.
371e77e3
UB
5999 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
6000 attribute changes.
006bd006
SN
6001 (thumb2_movsi_insn): Likewise.
6002 (thumb2_cmpsi_neg_shiftsi): Likewise.
6003 (thumb2_extendqisi_v6): Likewise.
6004 (thumb2_zero_extendhisi2_v6): Likewise.
6005 (thumb2_zero_extendqisi2_v6): Likewise.
6006 (thumb2_shiftsi3_short): Likewise.
6007 (thumb2_addsi3_compare0_scratch): Likewise.
6008 (orsi_not_shiftsi_si): Likewise.
6009 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6010 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
6011 changes.
6012 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6013 (1020alu_shift_op): Likewise.
6014 (1020alu_shift_reg_op): Likewise.
6015 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6016 (alu_shift_op): Likewise.
6017 (alu_shift_reg_op): Likewise.
6018 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6019 (11_alu_shift_op): Likewise.
6020 (11_alu_shift_reg_op): Likewise.
6021 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6022 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
6023 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6024 changes.
006bd006
SN
6025 (cortex_a15_alu_shift): Likewise.
6026 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB
6027 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
6028 changes.
006bd006
SN
6029 (cortex_a5_alu_shift): Likewise.
6030 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
6031 changes.
6032 (cortex_a53_alu_shift): Likewise.
6033 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6034 changes.
6035 (cortex_a7_alu_reg): Likewise.
6036 (cortex_a7_alu_shift): Likewise.
371e77e3
UB
6037 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
6038 changes.
006bd006
SN
6039 (cortex_a8_alu_shift): Likewise.
6040 (cortex_a8_alu_shift_reg): Likewise.
6041 (cortex_a8_mov): Likewise.
6042 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6043 (cortex_a9_dp_shift): Likewise.
371e77e3
UB
6044 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
6045 changes.
6046 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
6047 changes.
006bd006
SN
6048 (cortex_r4_mov): Likewise.
6049 (cortex_r4_alu_shift): Likewise.
6050 (cortex_r4_alu_shift_reg): Likewise.
6051 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6052 (526_alu_shift_op): Likewise.
6053 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6054 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6055 (626te_alu_shift_op): Likewise.
6056 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
6057 (726te_alu_op): Likewise.
6058 (726te_alu_shift_op): Likewise.
6059 (726te_alu_shift_reg_op): Likewise.
6060 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6061 (mp626_alu_shift_op): Likewise.
6062 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
6063 (pj4_alu_e1_conds): Likewise.
6064 (pj4_alu): Likewise.
6065 (pj4_alu_conds): Likewise.
6066 (pj4_shift): Likewise.
6067 (pj4_shift_conds): Likewise.
6068 (pj4_alu_shift): Likewise.
6069 (pj4_alu_shift_conds): Likewise.
371e77e3
UB
6070 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6071 changes.
006bd006
SN
6072 (cortexa7_older_only): Likewise.
6073 (cortexa7_younger): Likewise.
6074
651df1b2
DM
60752013-07-18 David Malcolm <dmalcolm@redhat.com>
6076
6077 * ipa-pure-const.c (generate_summary): Rename to...
6078 (pure_const_generate_summary): ... this.
6079
ea717bd7
IS
60802013-07-17 Iain Sandoe <iain@codesourcery.com>
6081
6082 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
6083
a1259a13
YR
60842013-07-17 Yvan Roux <yvan.roux@linaro.org>
6085
6086 PR target/57909
6087 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
6088 usage in HI mode.
6089
167f68ed
AK
60902013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6091
6092 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
6093 enabled without -march=zEC12.
6094 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
6095 flags to be set.
6096
f900a982
MR
60972013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6098
6099 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
6100 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
6101 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
6102 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
6103 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
6104 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
6105 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
6106 ISA_HAS_FP4.
6107 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
6108 and ISA_HAS_NMADD3_NMSUB3.
6109 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
6110 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
6111 (nmsub4<mode>, nmsub3<mode>): Likewise.
6112 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
6113
d4ed27eb
MR
61142013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6115
6116 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
6117 TARGET_MIPS5400 checking.
6118
7a5add18
PB
61192013-07-16 Jakub Jelinek <jakub@redhat.com>
6120 Peter Bergner <bergner@vnet.ibm.com>
6121
6122 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
6123 registers in the comment.
6124 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
6125 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
6126 rather than FIRST_PSEUDO_REGISTERS.
6127
61282013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB
6129
6130 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
6131 enable extra ISA flags with TARGET_HTM.
6132
d2ab0929
MR
61332013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
6134
6135 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
6136 Fix comment typos.
6137
839c74bc
CH
61382013-07-15 Cong Hou <congh@google.com>
6139
6140 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
6141 in compare function for sorting.
6142
7a5add18 61432013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB
6144
6145 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
6146 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
6147 * config/rs6000/rs6000.opt: Add -mhtm option.
6148 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
6149 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
6150 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
6151 __HTM__ if the HTM instructions are available.
6152 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
6153 htm_spr_reg_operand): New define_predicates.
6154 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
6155 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
6156 Include htm.md.
6157 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
6158 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
6159 HTM builtin functions.
6160 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
6161 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
6162 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
6163 (rs6000_builtin_mask_calculate): Likewise.
6164 (rs6000_option_override_internal): Likewise.
6165 (bdesc_htm): Add new HTM builtin support.
6166 (htm_spr_num): New function.
6167 (htm_spr_regno): Likewise.
6168 (rs6000_htm_spr_icode): Likewise.
6169 (htm_expand_builtin): Likewise.
6170 (htm_init_builtins): Likewise.
6171 (rs6000_expand_builtin): Add support for HTM builtin functions.
6172 (rs6000_init_builtins): Likewise.
371e77e3
UB
6173 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
6174 option.
0258b6e4
PB
6175 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
6176 (TARGET_HTM, MASK_HTM): Define macros.
6177 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
6178 (FIXED_REGISTERS): Likewise.
6179 (CALL_USED_REGISTERS): Likewise.
6180 (CALL_REALLY_USED_REGISTERS): Likewise.
6181 (REG_ALLOC_ORDER): Likewise.
6182 (enum reg_class): Likewise.
6183 (REG_CLASS_NAMES): Likewise.
6184 (REG_CLASS_CONTENTS): Likewise.
6185 (REGISTER_NAMES): Likewise.
6186 (ADDITIONAL_REGISTER_NAMES): Likewise.
6187 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
6188 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
6189 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
6190 * config/rs6000/htm.md: New file.
6191 * config/rs6000/htmintrin.h: New file.
6192 * config/rs6000/htmxlintrin.h: New file.
6193
87dd8ab0
MS
61942013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
6195
6196 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
6197 Define SYMBOL_TINY_GOT, update comment.
6198 * config/aarch64/aarch64.c
6199 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
6200 (aarch64_expand_mov_immediate): Likewise.
6201 (aarch64_print_operand): Likewise.
6202 (aarch64_classify_symbol): Likewise.
6203 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
6204 (ldr_got_tiny): Define.
6205
52d676b6
TG
62062013-07-13 Tobias Grosser <tobias@grosser.es>
6207
6208 PR tree-optimization/54094
6209 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
6210 scheduling dimension for the parallelism check from the polyhedral
6211 information in the AST.
6212 * graphite-dependences.c (carries_deps): Do not assume the schedule is
6213 in 2D + 1 form.
6214
286e8fc1
JM
62152013-07-13 Jason Merrill <jason@redhat.com>
6216
6217 * print-tree.c (debug_vec_tree): Use debug_raw.
6218 (debug_raw (vec<tree, va_gc> &)): New.
6219 (debug_raw (vec<tree, va_gc> *)): New.
6220 * tree.h: Declare them.
6221
e43257e8
BC
62222013-07-13 Bin Cheng <bin.cheng@arm.com>
6223
6224 * ifcvt.c (ifcvt_after_combine): New static variable.
6225 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
6226 for size.
6227 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
6228 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
6229 rest_of_handle_if_after_reload): Pass new argument for if_convert.
6230
da582d46
MR
62312013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
6232
6233 * config/mips/mips.c (mips_expand_call): Remove empty statement.
6234
18b0ea8f
MM
62352013-07-12 Michael Matz <matz@suse.de>
6236
6237 PR middle-end/55771
6238 * convert.c (convert_to_real): Reject non-float inner types.
6239
e4f0f84d
TB
62402013-07-12 Tejas Belagod <tejas.belagod@arm.com>
6241
6242 * config/aarch64/aarch64-protos.h
6243 (aarch64_simd_immediate_valid_for_move): Remove.
6244 * config/aarch64/aarch64.c (simd_immediate_info): New member.
6245 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
6246 cases.
6247 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
6248
76500022
SE
62492013-07-11 Steve Ellcey <sellcey@mips.com>
6250
6251 * config/mips/mips.c (mips_conditional_register_usage): Do not
6252 use t[0-7] registers in MIPS16 mode when optimizing for size.
6253
78039734
ST
62542013-07-11 Sriraman Tallam <tmsriram@google.com>
6255
6256 * config/i386/i386.c (dispatch_function_versions): Fix array
6257 indexing of function_version_info to match actual_versions.
6258
7b24b675
TJ
62592013-07-11 Teresa Johnson <tejohnson@google.com>
6260
6261 * vec.h (struct va_gc): Move release out-of-line.
6262 (va_gc::release): Call ggc_free on released vec.
6263
85d4cbb8
UW
62642013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6265
6266 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
6267 Require GOT register as additional operand in UNSPEC.
6268 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
6269 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6270 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
6271 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
6272 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
6273 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
6274 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
6275
52b3f9e2
GJL
62762013-07-11 Georg-Johann Lay <avr@gjlay.de>
6277
6278 PR target/57631
6279 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
6280 name seen by assembler/linker rather if available.
6281
3f3e266e
AS
62822013-07-11 Andreas Schwab <schwab@suse.de>
6283
6284 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
6285
21231ca6
VM
62862013-07-10 Vladimir Makarov <vmakarov@redhat.com>
6287
371e77e3 6288 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 6289
d331d374
JM
62902013-07-10 Joseph Myers <joseph@codesourcery.com>
6291
673c2f63
JM
6292 * doc/tm.texi.in: Move hook documentation to ....
6293 * target.def: ... here.
6294
d331d374
JM
6295 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
6296 text on @hook line.
6297 * doc/tm.texi: Regenerate.
6298
dfeadaa0
PC
62992013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
6300
6301 PR c++/57869
6302 * doc/invoke.texi: Document -Wconditionally-supported.
6303
bb6f2bac
GJL
63042013-07-10 Georg-Johann Lay <avr@gjlay.de>
6305
6306 PR target/57844
6307 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
6308 of my_fp.
6309
5219b232
GJL
63102013-07-10 Georg-Johann Lay <avr@gjlay.de>
6311
6312 PR target/57506
6313 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
6314 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
6315 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
6316 Remove duplicate devices.
6317 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
6318 * config/avr/t-multilib: Regenerate.
6319 * config/avr/avr-tables.opt: Regenerate.
6320 * doc/avr-mmcu.texi: Regenerate.
6321
283045c0
GJL
63222013-07-10 Georg-Johann Lay <avr@gjlay.de>
6323
6324 PR target/56987
6325 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
6326
ceb2d59f 63272013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 6328
ceb2d59f
GS
6329 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
6330 the cost of MULT when optimizing for size.
6331
055e0a99
JBG
63322013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6333
6334 * config/cr16/cr16-protos.h: Don't include target.h.
6335
ef43e868
JM
63362013-07-09 Joseph Myers <joseph@codesourcery.com>
6337
6338 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
6339 adjust register size for TDmode and TFmode for VSX registers.
6340
4a283fdf
KT
63412013-07-08 Kai Tietz <ktietz@redhat.com>
6342
6343 PR target/56892
6344 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
6345 hook_bool_const_tree_true.
6346
2cf4c39e
AK
63472013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6348
6349 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
6350 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
6351 * config/s390/s390.md: Define FPR*_REGNUM constants.
6352 Fix FPR2_REGNUM constant (18 -> 17).
6353 ("*trunc<BFP:mode><DFP_ALL:mode>2")
6354 ("*trunc<DFP_ALL:mode><BFP:mode>2")
6355 ("trunc<BFP:mode><DFP_ALL:mode>2")
6356 ("trunc<DFP_ALL:mode><BFP:mode>2")
6357 ("*extend<BFP:mode><DFP_ALL:mode>2")
6358 ("*extend<DFP_ALL:mode><BFP:mode>2")
6359 ("extend<BFP:mode><DFP_ALL:mode>2")
6360 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
6361 FPR4_REGNUM.
6362
e6ac0270 63632013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 6364
6a2fa4b2 6365 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 6366
b89b22fc
AK
63672013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6368
6369 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
6370 and cfun_fpr_bit_p to cfun_fpr_save_p.
6371 (s390_frame_area, s390_register_info, s390_frame_info)
6372 (s390_emit_prologue, s390_emit_epilogue)
6373 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
6374 register numbers.
6375 * config/s390/s390.h: Define *_REGNUM macros for floating point
6376 register numbers.
6377
07f398aa
EB
63782013-07-08 Eric Botcazou <ebotcazou@adacore.com>
6379
6380 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
6381
75fefa91
PCC
63822013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
6383
6384 PR rtl-optimization/57786
6385 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
6386 and break out of the loop when it is set to false.
6387
046f1eee
JJ
63882013-07-08 Jakub Jelinek <jakub@redhat.com>
6389
8140c065
JJ
6390 PR target/57819
6391 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
6392 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
6393 (const_int 63)) 0)).
6394 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
6395 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
6396 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
6397
046f1eee
JJ
6398 PR rtl-optimization/57829
6399 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
6400 mask bits outside of mode are just sign-extension from mode to HWI.
6401
930b700b
MZ
64022013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6403
6404 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
6405 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
6406 adjust_address instead of change_address to keep info about alignment.
6407 (emit_strmov): Remove.
6408 (emit_memmov): New function.
6409 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
6410 (expand_movmem_epilogue): Likewise and return updated rtx for
6411 destination.
6412 (expand_constant_movmem_prologue): Likewise and return updated rtx for
6413 destination and source.
6414 (decide_alignment): Refactor, handle vector_loop.
6415 (ix86_expand_movmem): Likewise.
6416 (ix86_expand_setmem): Likewise.
6417 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
6418
7b9d1bd8
UB
64192013-07-07 Uros Bizjak <ubizjak@gmail.com>
6420
6421 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
6422 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
6423
eabb5f48
UB
64242013-07-06 Uros Bizjak <ubizjak@gmail.com>
6425
6426 * config/i386/sse.md (sse_movlhps): Change alternative 3
6427 of operand 2 to "m".
6428
64292013-07-06 Uros Bizjak <ubizjak@gmail.com>
6430
6431 PR target/57807
6432 * config/i386/sse.md (iptr): New mode attribute.
6433 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
6434 (<sse>_vm<plusminus_insn><mode>3): Ditto.
6435 (<sse>_vmmul<mode>3): Ditto.
6436 (<sse>_vmdiv<mode>3): Ditto.
6437 (sse_vmrcpv4sf2): Ditto.
6438 (<sse>_vmsqrt<mode>2): Ditto.
6439 (sse_vmrsqrtv4sf2): Ditto.
6440 (<sse>_vm<code><mode>3): Ditto.
6441 (avx_vmcmp<mode>3): Ditto.
6442 (<sse>_vmmaskcmp<mode>3): Ditto.
6443 (<sse>_comi): Ditto.
6444 (<sse>_ucomi): Ditto.
6445 (*xop_vmfrcz_<mode>): Ditto.
6446 (*fmai_fmadd_<mode>): Ditto.
6447 (*fmai_fmsub_<mode>): Ditto.
6448 (*fmai_fnmadd_<mode>): Ditto.
6449 (*fmai_fnmsub_<mode>): Ditto.
6450 (*fma4i_vmfmadd_<mode>): Ditto.
6451 (*fma4i_vmfmsub_<mode>): Ditto.
6452 (*fma4i_vmfnmadd_<mode>): Ditto.
6453 (*fma4i_vmfnmsub_<mode>): Ditto.
6454 (*xop_vmfrcz_<mode>): Ditto.
6455 (sse_cvtps2pi): Ditto.
6456 (sse_cvttps2pi): Ditto.
6457 (sse_cvtss2si): Ditto.
6458 (sse_cvtss2si_2): Ditto.
6459 (sse_cvtss2siq_2): Ditto.
6460 (sse_cvttss2si): Ditto.
6461 (sse_cvttss2siq): Ditto.
6462 (sse_cvtsd2si): Ditto.
6463 (sse_cvtsd2si_2): Ditto.
6464 (sse_cvtsd2siq_2): Ditto.
6465 (sse_cvttsd2si): Ditto.
6466 (sse_cvttsd2siq): Ditto.
6467 (sse_cvtsd2ss): Ditto.
6468 (sse_cvtss2sd): Ditto.
6469 (avx2_pbroadcast<mode>): Ditto.
6470 (avx2_pbroadcast<mode>_1): Ditto.
6471 (*avx_vperm_broadcast_v4sf): Ditto.
6472
6473 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
6474 (sse_movlhps): Ditto.
6475 (sse_storehps): Ditto.
6476 (sse_loadhps): Ditto.
6477 (sse_storelps): Ditto.
6478 (sse_loadlps): Ditto.
6479 (*vec_concatv4sf): Ditto.
6480 (*vec_interleave_highv2df): Ditto.
6481 (*vec_interleave_lowv2df): Ditto.
6482 (*vec_extractv2df_1_sse): Ditto.
6483 (*vec_extractv2df_0_sse): Ditto.
6484 (sse2_storelpd): Ditto.
6485 (sse2_loadlpd): Ditto.
6486 (sse2_movsd): Ditto.
6487 (*vec_concatv4si): Ditto.
6488 (vec_concatv2di): Ditto.
6489
6490 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
6491 for Intel asm dialect.
6492 (mmx_punpcklwd): Ditto.
6493 (mmx_punpckldq): Ditto.
6494
6495 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
6496 for intel assembler dialect.
6497
1f6eac90
JJ
64982013-07-06 Jakub Jelinek <jakub@redhat.com>
6499
6500 PR target/29776
6501 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
6502 for BUILT_IN_C{LZ,LRSB}*.
6503 * tree.h (CASE_INT_FN): Add FN##IMAX case.
6504 * tree-vrp.c (extract_range_basic): Handle
6505 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
6506 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
6507 fall thru to code calling set_value*.
6508 * builtins.c (expand_builtin): Remove *IMAX cases.
6509 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
6510 if width is bigger than 2*HWI.
6511
2b778c9d
VM
65122013-07-05 Vladimir Makarov <vmakarov@redhat.com>
6513
6514 PR rtl-optimization/55342
6515 * lra-int.h (lra_subreg_reload_pseudos): New.
6516 * lra.c: Add undoing optional reloads to the block diagram.
6517 (lra_subreg_reload_pseudos): New.
6518 (lra_optional_reload_pseudos): Change comments.
6519 (lra): Init and clear lra_subreg_reload_pseudos. Clear
6520 lra_optional_reload_pseudos after undo transformations.
6521 * lra-assigns.c (pseudo_prefix_title): New.
6522 (lra_setup_reg_renumber): Use it.
6523 (spill_for): Ditto. Check subreg reload pseudos too.
6524 (assign_by_spills): Consider subreg reload pseudos too.
6525 * lra-constraints.c (simplify_operand_subreg): Use
6526 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
6527 (curr_insn_transform): Recognize and do optional reloads.
6528 (undo_optional_reloads): New.
6529 (lra_undo_inheritance): Call undo_optional_reloads.
6530
207156e4
TQ
65312013-07-05 Thomas Quinot <quinot@adacore.com>
6532
6533 * tree-complex.c (expand_complex_operations_1): Fix typo.
6534
8990e73a
TB
65352013-07-04 Tejas Belagod <tejas.belagod@arm.com>
6536
6537 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
6538 (tune_params): New member 'const vec_costs'.
6539 * config/aarch64/aarch64.c (generic_vector_cost): New.
6540 (generic_tunings): New member 'generic_vector_cost'.
6541 (aarch64_builtin_vectorization_cost): New.
6542 (aarch64_add_stmt_cost): New.
6543 (TARGET_VECTORIZE_ADD_STMT_COST): New.
6544 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
6545
f7bc421d
JJ
65462013-07-03 Jakub Jelinek <jakub@redhat.com>
6547
6548 PR target/57777
6549 * config/i386/predicates.md (vsib_address_operand): Disallow
6550 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
6551
e90d1568
HPN
65522013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
6553
6554 PR middle-end/55030
6555 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
6556 expand_builtin_setjmp_receiver.
6557 (expand_label): Adjust, call expand_builtin_setjmp_receiver
6558 with NULL for the label parameter.
6559 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
6560 the frame-pointer. Adjust comments.
6561 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
6562 only if LABEL is non-NULL.
6563
cbd7413b
YZ
65642013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
6565
6566 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
6567 (ARM_ABI_AAPCS64): Ditto.
6568 (arm_abi): Ditto.
6569 (ARM_DEFAULT_ABI): Ditto.
6570
dec11868
JG
65712013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
6572
6573 * config/aarch64/aarch64-builtins.c
6574 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
6575 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
6576 (st1): Likewise.
6577 * config/aarch64/aarch64-simd.md
6578 (aarch64_ld1<VALL:mode>): New.
6579 (aarch64_st1<VALL:mode>): Likewise.
6580 * config/aarch64/arm_neon.h
6581 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
6582
786fb9b6
ST
65832013-07-02 Sriraman Tallam <tmsriram@google.com>
6584
0ba6aff3
UB
6585 * config/i386/i386.c (gate_insert_vzeroupper): Check if
6586 target ISA is AVX.
786fb9b6
ST
6587 (ix86_option_override_internal):Turn on all -mavx target flags by
6588 default as they are dependent on AVX anyway.
6589
4b043553
CC
65902013-07-02 Cary Coutant <ccoutant@google.com>
6591
6592 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
6593 deterministic hash.
6594 (loc_checksum_ordered): Likewise.
6595 (hash_loc_operands): Remove inline keyword.
6596
c4e87a13
JJ
65972013-07-02 Jakub Jelinek <jakub@redhat.com>
6598
6599 PR tree-optimization/57741
6600 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
6601 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
6602 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
6603 Allow REAL_CST step_exprs if flag_associative_math.
6604 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
6605
26366d28
IB
66062013-07-02 Ian Bolton <ian.bolton@arm.com>
6607
0ba6aff3 6608 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 6609
2879bb2b
IB
66102013-07-02 Ian Bolton <ian.bolton@arm.com>
6611
6612 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
6613
afc5e8a6
KT
66142013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6615
6616 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
6617 encoding.
6618 (iorsi3_insn): Likewise.
6619 (arm_xorsi3): Likewise.
6620
31f8442b
SN
66212013-07-01 Sofiane Naci <sofiane.naci@arm.com>
6622
6623 * arm.md (attribute "wtype"): Delete. Move attribute values from here
6624 to ...
6625 (attribute "type"): ... here, and prefix with "wmmx_".
6626 (attribute "core_cycles"): Update for attribute changes.
6627 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
6628 (tbcstv4hi): Likewise.
6629 (tbcstv2si): Likewise.
6630 (iwmmxt_iordi3): Likewise.
6631 (iwmmxt_xordi3): Likewise.
6632 (iwmmxt_anddi3): Likewise.
6633 (iwmmxt_nanddi3): Likewise.
6634 (iwmmxt_arm_movdi): Likewise.
6635 (iwmmxt_movsi_insn): Likewise.
6636 (mov<mode>_internal): Likewise.
6637 (and<mode>3_iwmmxt): Likewise.
6638 (ior<mode>3_iwmmxt): Likewise.
6639 (xor<mode>3_iwmmxt): Likewise.
6640 (add<mode>3_iwmmxt): Likewise.
6641 (ssaddv8qi3): Likewise.
6642 (ssaddv4hi3): Likewise.
6643 (ssaddv2si3): Likewise.
6644 (usaddv8qi3): Likewise.
6645 (usaddv4hi3): Likewise.
6646 (usaddv2si3): Likewise.
6647 (sub<mode>3_iwmmxt): Likewise.
6648 (sssubv8qi3): Likewise.
6649 (sssubv4hi3): Likewise.
6650 (sssubv2si3): Likewise.
6651 (ussubv8qi3): Likewise.
6652 (ussubv4hi3): Likewise.
6653 (ussubv2si3): Likewise.
6654 (mulv4hi3_iwmmxt): Likewise.
6655 (smulv4hi3_highpart): Likewise.
6656 (umulv4hi3_highpart): Likewise.
6657 (iwmmxt_wmacs): Likewise.
6658 (iwmmxt_wmacsz): Likewise.
6659 (iwmmxt_wmacu): Likewise.
6660 (iwmmxt_wmacuz): Likewise.
6661 (iwmmxt_clrdi): Likewise.
6662 (iwmmxt_clrv8qi): Likewise.
6663 (iwmmxt_clr4hi): Likewise.
6664 (iwmmxt_clr2si): Likewise.
6665 (iwmmxt_uavgrndv8qi3): Likewise.
6666 (iwmmxt_uavgrndv4hi3): Likewise.
6667 (iwmmxt_uavgv8qi3): Likewise.
6668 (iwmmxt_uavgv4hi3): Likewise.
6669 (iwmmxt_tinsrb): Likewise.
6670 (iwmmxt_tinsrh): Likewise.
6671 (iwmmxt_tinsrw): Likewise.
6672 (iwmmxt_textrmub): Likewise.
6673 (iwmmxt_textrmsb): Likewise.
6674 (iwmmxt_textrmuh): Likewise.
6675 (iwmmxt_textrmsh): Likewise.
6676 (iwmmxt_textrmw): Likewise.
6677 (iwmxxt_wshufh): Likewise.
6678 (eqv8qi3): Likewise.
6679 (eqv4hi3): Likewise.
6680 (eqv2si3): Likewise.
6681 (gtuv8qi3): Likewise.
6682 (gtuv4hi3): Likewise.
6683 (gtuv2si3): Likewise.
6684 (gtv8qi3): Likewise.
6685 (gtv4hi3): Likewise.
6686 (gtv2si3): Likewise.
6687 (smax<mode>3_iwmmxt): Likewise.
6688 (umax<mode>3_iwmmxt): Likewise.
6689 (smin<mode>3_iwmmxt): Likewise.
6690 (umin<mode>3_iwmmxt): Likewise.
6691 (iwmmxt_wpackhss): Likewise.
6692 (iwmmxt_wpackwss): Likewise.
6693 (iwmmxt_wpackdss): Likewise.
6694 (iwmmxt_wpackhus): Likewise.
6695 (iwmmxt_wpackwus): Likewise.
6696 (iwmmxt_wpackdus): Likewise.
6697 (iwmmxt_wunpckihb): Likewise.
6698 (iwmmxt_wunpckihh): Likewise.
6699 (iwmmxt_wunpckihw): Likewise.
6700 (iwmmxt_wunpckilb): Likewise.
6701 (iwmmxt_wunpckilh): Likewise.
6702 (iwmmxt_wunpckilw): Likewise.
6703 (iwmmxt_wunpckehub): Likewise.
6704 (iwmmxt_wunpckehuh): Likewise.
6705 (iwmmxt_wunpckehuw): Likewise.
6706 (iwmmxt_wunpckehsb): Likewise.
6707 (iwmmxt_wunpckehsh): Likewise.
6708 (iwmmxt_wunpckehsw): Likewise.
6709 (iwmmxt_wunpckelub): Likewise.
6710 (iwmmxt_wunpckeluh): Likewise.
6711 (iwmmxt_wunpckeluw): Likewise.
6712 (iwmmxt_wunpckelsb): Likewise.
6713 (iwmmxt_wunpckelsh): Likewise.
6714 (iwmmxt_wunpckelsw): Likewise.
6715 (ror<mode>3): Likewise.
6716 (ashr<mode>3_iwmmxt): Likewise.
6717 (lshr<mode>3_iwmmxt): Likewise.
6718 (ashl<mode>3_iwmmxt): Likewise.
6719 (ror<mode>3_di): Likewise.
6720 (ashr<mode>3_di): Likewise.
6721 (lshr<mode>3_di): Likewise.
6722 (ashl<mode>3_di): Likewise.
6723 (iwmmxt_wmadds): Likewise.
6724 (iwmmxt_wmaddu): Likewise.
6725 (iwmmxt_tmia): Likewise.
6726 (iwmmxt_tmiaph): Likewise.
6727 (iwmmxt_tmiabb): Likewise.
6728 (iwmmxt_tmiatb): Likewise.
6729 (iwmmxt_tmiabt): Likewise.
6730 (iwmmxt_tmiatt): Likewise.
6731 (iwmmxt_tmovmskb): Likewise.
6732 (iwmmxt_tmovmskh): Likewise.
6733 (iwmmxt_tmovmskw): Likewise.
6734 (iwmmxt_waccb): Likewise.
6735 (iwmmxt_wacch): Likewise.
6736 (iwmmxt_waccw): Likewise.
6737 (iwmmxt_waligni): Likewise.
6738 (iwmmxt_walignr): Likewise.
6739 (iwmmxt_walignr0): Likewise.
6740 (iwmmxt_walignr1): Likewise.
6741 (iwmmxt_walignr2): Likewise.
6742 (iwmmxt_walignr3): Likewise.
6743 (iwmmxt_wsadb): Likewise.
6744 (iwmmxt_wsadh): Likewise.
6745 (iwmmxt_wsadbz): Likewise.
6746 (iwmmxt_wsadhz): Likewise.
6747 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
6748 (iwmmxt_wabsdiffb): Likewise.
6749 (iwmmxt_wabsdiffh): Likewise.
6750 (iwmmxt_wabsdiffw): Likewise.
6751 (iwmmxt_waddsubhx): Likewise
6752 (iwmmxt_wsubaddhx): Likewise.
6753 (addc<mode>3): Likewise.
6754 (iwmmxt_avg4): Likewise.
6755 (iwmmxt_avg4r): Likewise.
6756 (iwmmxt_wmaddsx): Likewise.
6757 (iwmmxt_wmaddux): Likewise.
6758 (iwmmxt_wmaddsn): Likewise.
6759 (iwmmxt_wmaddun): Likewise.
6760 (iwmmxt_wmulwsm): Likewise.
6761 (iwmmxt_wmulwum): Likewise.
6762 (iwmmxt_wmulsmr): Likewise.
6763 (iwmmxt_wmulumr): Likewise.
6764 (iwmmxt_wmulwsmr): Likewise.
6765 (iwmmxt_wmulwumr): Likewise.
6766 (iwmmxt_wmulwl): Likewise.
6767 (iwmmxt_wqmulm): Likewise.
6768 (iwmmxt_wqmulwm): Likewise.
6769 (iwmmxt_wqmulmr): Likewise.
6770 (iwmmxt_wqmulwmr): Likewise.
6771 (iwmmxt_waddbhusm): Likewise.
6772 (iwmmxt_waddbhusl): Likewise.
6773 (iwmmxt_wqmiabb): Likewise.
6774 (iwmmxt_wqmiabt): Likewise.
6775 (iwmmxt_wqmiatb): Likewise.
6776 (iwmmxt_wqmiatt): Likewise.
6777 (iwmmxt_wqmiabbn): Likewise.
6778 (iwmmxt_wqmiabtn): Likewise.
6779 (iwmmxt_wqmiatbn): Likewise.
6780 (iwmmxt_wqmiattn): Likewise.
6781 (iwmmxt_wmiabb): Likewise.
6782 (iwmmxt_wmiabt): Likewise.
6783 (iwmmxt_wmiatb): Likewise.
6784 (iwmmxt_wmiatt): Likewise.
6785 (iwmmxt_wmiabbn): Likewise.
6786 (iwmmxt_wmiabtn): Likewise.
6787 (iwmmxt_wmiatbn): Likewise.
6788 (iwmmxt_wmiattn): Likewise.
6789 (iwmmxt_wmiawbb): Likewise.
6790 (iwmmxt_wmiawbt): Likewise.
6791 (iwmmxt_wmiawtb): Likewise.
6792 (iwmmxt_wmiawtt): Likewise.
6793 (iwmmxt_wmiawbbn): Likewise.
6794 (iwmmxt_wmiawbtn): Likewise.
6795 (iwmmxt_wmiawtbn): Likewise.
6796 (iwmmxt_wmiawttn): Likewise.
6797 (iwmmxt_wmerge): Likewise.
6798 (iwmmxt_tandc<mode>3): Likewise.
6799 (iwmmxt_torc<mode>3): Likewise.
6800 (iwmmxt_torvsc<mode>3): Likewise.
6801 (iwmmxt_textrc<mode>3): Likewise.
6802 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
6803 (wmmxt_pack): Likewise.
6804 (wmmxt_mult_c1): Likewise.
6805 (wmmxt_mult_c2): Likewise.
6806 (wmmxt_alu_c1): Likewise.
6807 (wmmxt_alu_c2): Likewise.
6808 (wmmxt_alu_c3): Likewise.
6809 (wmmxt_transfer_c1): Likewise.
6810 (wmmxt_transfer_c2): Likewise.
6811 (wmmxt_transfer_c3): Likewise.
6812 (marvell_f_iwmmxt_wstr): Likewise.
6813 (marvell_f_iwmmxt_wldr): Likewise.
6814
bdb7bf8a
YZ
68152013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
6816
0ba6aff3 6817 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 6818
e6631d38
VM
68192013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6820
6821 Revert:
6822 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6823 * lra-constraints.c (need_for_split_p): Check call used hard regs
6824 living through calls.
6825
6826 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
6827 call used regs for call insn.
6828
53c1275b
JJ
68292013-06-28 Jakub Jelinek <jakub@redhat.com>
6830
6831 PR target/57736
0ba6aff3
UB
6832 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
6833 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 6834
1141ed3f
BI
68352013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
6836
6837 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 6838
e32e4c4a
VM
68392013-06-28 Vladimir Makarov <vmakarov@redhat.com>
6840
6841 * lra-constraints.c (need_for_split_p): Check call used hard regs
6842 living through calls.
6843
f161bfd3
MM
68442013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
6845
6846 PR target/57744
6847 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
6848 to tie with any other modes. Eliminate Altivec vector mode tests,
6849 since these are a subset of ALTIVEC or VSX vector modes. Simplify
6850 code, to return 0 if testing MODE2 for a condition, if we've
6851 already tested MODE1 for the same condition.
6852
7eda14e1
MS
68532013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6854
6855 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
6856 layout.
6857
c822f852
MS
68582013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6859
6860 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
6861 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
6862
da4f13a4
MS
68632013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
6864
0ba6aff3
UB
6865 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
6866 Define.
da4f13a4
MS
6867 (aarch64_symbolic_constant_p): Remove.
6868 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
6869 static. Fix line length and white space.
6870 (aarch64_symbolic_constant_p): Remove.
6871 * config/aarch64/predicates.md (aarch64_valid_symref):
6872 Use aarch64_classify_symbol_expression.
6873
956a95a5
KT
68742013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6875
6876 * config/arm/constraints.md (Ts): New constraint.
6877 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
6878 16-bit encodings.
6879 (compare_scc): Use "Ts" constraint for operand 0.
6880 (ior_scc_scc): Likewise.
6881 (and_scc_scc): Likewise.
6882 (and_scc_scc_nodom): Likewise.
6883 (ior_scc_scc_cmp): Likewise for operand 7.
6884 (and_scc_scc_cmp): Likewise.
6885 * config/arm/thumb2.md (thumb2_movsi_insn):
6886 Add alternatives for 16-bit encodings.
6887 (thumb2_movhi_insn): Likewise.
6888 (thumb2_movsicc_insn): Likewise.
6889 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
6890 (thumb2_negscc): Use "Ts" constraint.
6891 Move mvn instruction outside cond_exec block.
6892 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
6893 for 16-bit encodings.
6894
31dfce10
KT
68952013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6896
6897 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
6898 encoding.
6899 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
6900 (mulsi3subsi): Likewise.
6901 (mulsidi3adddi): Likewise.
6902 (mulsidi3_v6): Likewise.
6903 (umulsidi3_v6): Likewise.
6904 (umulsidi3adddi_v6): Likewise.
6905 (smulsi3_highpart_v6): Likewise.
6906 (umulsi3_highpart_v6): Likewise.
6907 (mulhisi3tb): Likewise.
6908 (mulhisi3bt): Likewise.
6909 (mulhisi3tt): Likewise.
6910 (maddhisi4): Likewise.
6911 (maddhisi4tb): Likewise.
6912 (maddhisi4tt): Likewise.
6913 (maddhidi4): Likewise.
6914 (maddhidi4tb): Likewise.
6915 (maddhidi4tt): Likewise.
6916 (zeroextractsi_compare0_scratch): Likewise.
6917 (insv_zero): Likewise.
6918 (insv_t2): Likewise.
6919 (anddi_notzesidi_di): Likewise.
6920 (anddi_notsesidi_di): Likewise.
6921 (andsi_notsi_si): Likewise.
6922 (iordi_zesidi_di): Likewise.
6923 (xordi_zesidi_di): Likewise.
6924 (andsi_iorsi3_notsi): Likewise.
6925 (smax_0): Likewise.
6926 (smax_m1): Likewise.
6927 (smin_0): Likewise.
6928 (not_shiftsi): Likewise.
6929 (unaligned_loadsi): Likewise.
6930 (unaligned_loadhis): Likewise.
6931 (unaligned_loadhiu): Likewise.
6932 (unaligned_storesi): Likewise.
6933 (unaligned_storehi): Likewise.
6934 (extv_reg): Likewise.
6935 (extzv_t2): Likewise.
6936 (divsi3): Likewise.
6937 (udivsi3): Likewise.
6938 (arm_zero_extendhisi2addsi): Likewise.
6939 (arm_zero_extendqisi2addsi): Likewise.
6940 (compareqi_eq0): Likewise.
6941 (arm_extendhisi2_v6): Likewise.
6942 (arm_extendqisi2addsi): Likewise.
6943 (arm_movt): Likewise.
6944 (thumb2_ldrd): Likewise.
6945 (thumb2_ldrd_base): Likewise.
6946 (thumb2_ldrd_base_neg): Likewise.
6947 (thumb2_strd): Likewise.
6948 (thumb2_strd_base): Likewise.
6949 (thumb2_strd_base_neg): Likewise.
6950 (arm_negsi2): Add alternative for 16-bit encoding.
6951 (arm_one_cmplsi2): Likewise.
6952
1572e697
KT
69532013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6954
6955 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
6956 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
6957 (movdfcc): Likewise.
6958 * config/arm/vfp.md (*thumb2_movsf_vfp):
6959 Disable predication for arm_restrict_it.
6960 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
6961 (*thumb2_movdfcc_vfp): Likewise.
6962 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
6963 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
6964 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
6965 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
6966 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
6967 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
6968 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
6969 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
6970 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
6971 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
6972 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
6973 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
6974 Disable predication for arm_restrict_it.
6975
ebf8f0ea
KY
69762013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
6977
6978 * config/i386/bmiintrin.h (_bextr_u32): New.
6979 (_bextr_u64): Ditto.
6980
a290fcda
RS
69812013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
6982
6983 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
6984 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
6985 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
6986 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
6987 * config/mips/n32-elf.h: ...this new file.
6988
ea169996
MG
69892013-06-27 Marc Glisse <marc.glisse@inria.fr>
6990
6991 PR target/57224
6992 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
6993 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
6994
12211b99 69952013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM
6996
6997 * config/mips/mips-tables.opt: Regenerate.
6998 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 6999 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 7000 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 7001
2444244e
JJ
70022013-06-27 Jakub Jelinek <jakub@redhat.com>
7003
cd06d2a2
JJ
7004 PR target/57623
7005 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
7006 constraints of operand 1 and 2.
7007
2444244e
JJ
7008 PR target/57623
7009 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
7010 to match RTL canonicalization. Swap predicates and
7011 constraints of operand 1 and 2.
7012
2f259720
VM
70132013-06-27 Vladimir Makarov <vmakarov@redhat.com>
7014
0ba6aff3
UB
7015 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
7016 Process OP_INOUT regs for splitting too.
2f259720 7017
fef4d2b3
JJ
70182013-06-27 Jakub Jelinek <jakub@redhat.com>
7019
7020 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
7021 decl before the loop, initialize to NULL.
7022 (vectorizable_load): Initialize ptr_incr to NULL.
7023
79ee9826
MJ
70242013-06-27 Martin Jambor <mjambor@suse.cz>
7025
7026 PR lto/57208
7027 * ipa-ref.h (ipa_maybe_record_reference): Declare.
7028 * ipa-ref.c (ipa_maybe_record_reference): New function.
7029 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
7030 * ipa-cp.c (create_specialized_node): Record potential references from
7031 aggvals.
7032 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
7033
e18b4a81
YZ
70342013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7035
7036 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
7037 parameter 'mode' of type 'enum machine_mode mode'; change to pass
7038 'mode' to force_reg.
7039 (aarch64_add_offset): Update calls to aarch64_force_temporary.
7040 (aarch64_expand_mov_immediate): Likewise.
7041
9c023bf0
YZ
70422013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
7043
7044 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
7045 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
7046
5a3fe9b6
AK
70472013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7048
7049 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
7050 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7051 (struct machine_function): Add tbegin_p.
7052 (s390_canonicalize_comparison): Fold CC mode compares to
7053 conditional jump if possible.
7054 (s390_emit_jump): Return the emitted jump.
7055 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
7056 Handle CCRAWmode compares.
7057 (s390_option_override): Default to -mhtm if available.
7058 (s390_reg_clobbered_rtx): Handle floating point regs as well.
7059 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
7060 FPRs instead of df_regs_ever_live_p.
7061 (s390_optimize_nonescaping_tx): New function.
7062 (s390_init_frame_layout): Extend clobbered_regs array to cover
7063 FPRs as well.
7064 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
7065 (s390_expand_tbegin): New function.
7066 (enum s390_builtin): New enum definition.
7067 (code_for_builtin): New array definition.
7068 (s390_init_builtins): New function.
7069 (s390_expand_builtin): New function.
7070 (TARGET_INIT_BUILTINS): Define.
7071 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 7072 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK
7073 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
7074 (s390_alc_comparison): Likewise.
7075 * config/s390/s390-modes.def: Add CCRAWmode.
7076 * config/s390/s390.h (processor_flags): Add PF_TX.
7077 (TARGET_CPU_HTM): Define macro.
7078 (TARGET_HTM): Define macro.
7079 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
7080 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
7081 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
7082 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 7083 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK
7084 (TBEGIN_MASK, TBEGINC_MASK): New constants.
7085 ("*cc_to_int"): Move up.
7086 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
7087 constants other than 0.
7088 ("*ccraw_to_int"): New insn and splitter definition.
7089 ("tbegin", "tbegin_nofloat", "tbegin_retry")
7090 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
7091 ("tx_assist"): New expander.
7092 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
7093 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
7094 * config/s390/s390.opt: Add -mhtm option.
7095 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
7096 * config/s390/htmxlintrin.h: New file.
7097 * config/s390/htmintrin.h: New file.
7098 * config/s390/s390intrin.h: New file.
7099 * doc/extend.texi: Document htm builtins.
7100 * config.gcc: Add the new header files to extra_headers.
7101
9e216629
TS
71022013-06-26 Thomas Schwinge <thomas@codesourcery.com>
7103
7104 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 7105 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 7106
fd8c65e7 71072013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
7108 Pat Haugen <pthaugen@us.ibm.com>
7109 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM
7110
7111 * config/rs6000/power8.md: New.
7112 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
7113 setting for power8 entry.
7114 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
7115 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
7116 test for Power4/Power5 only.
7117 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
7118 support.
7119 (force_new_group): Adjust comment.
7120 * config/rs6000/rs6000.md: Include power8.md.
7121
b24a2ce5
GY
71222013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
7123
7124 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
7125 * config/arm/arm-protos.h (arm_max_conditional_execute): New
7126 declaration.
7127 (tune_params): Update comment.
7128 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
7129 (arm_max_conditional_execute): New function.
7130 (thumb2_final_prescan_insn): Use max_insn_skipped and
7131 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
7132
f6fdeb62
JJ
71332013-06-25 Jakub Jelinek <jakub@redhat.com>
7134
7135 PR tree-optimization/57705
7136 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
7137 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 7138 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ
7139 (get_initial_def_for_induction): Handle SSA_NAME IV step.
7140
c13bc3d9
MJ
71412013-06-25 Martin Jambor <mjambor@suse.cz>
7142
7143 PR middle-end/57670
7144 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
7145 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
7146 calls in the dump.
7147 (ipa_note_param_call): Initialize member_ptr flag.
7148 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
7149 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
7150 (ipa_write_indirect_edge_info): Stream member_ptr flag.
7151 (ipa_read_indirect_edge_info): Likewise.
7152
8a845901
RB
71532013-06-25 Richard Biener <rguenther@suse.de>
7154
7155 PR middle-end/56977
7156 * passes.c (init_optimization_passes): Move pass_fold_builtins
7157 and pass_dce earlier with -Og.
7158
33be0bec
EB
71592013-06-25 Eric Botcazou <ebotcazou@adacore.com>
7160
7161 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
7162 <BIT_FIELD_REF>: Remove trailing TAB.
7163 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
7164 remove blank line.
7165
5fe8e757
MJ
71662013-06-24 Martin Jambor <mjambor@suse.cz>
7167
7168 PR tree-optimization/57358
7169 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
7170 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
7171 (ipa_analyze_params_uses): Generate pessimistic info when true.
7172
44a60244
MJ
71732013-06-24 Martin Jambor <mjambor@suse.cz>
7174
7175 PR tree-optimization/57539
7176 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
7177 global.inlined_to of the new node to it. All callers changed.
7178 * ipa-inline-transform.c (clone_inlined_nodes): New variable
7179 inlining_into, pass it to cgraph_clone_node.
7180 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
7181 ipa_free_edge_args_substructures.
7182 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
7183 rdesc linked list. Do not assert rdesc edges have inlined caller.
7184 Assert we have found an rdesc in the rdesc list.
7185
7c5848b8
RB
71862013-06-24 Richard Biener <rguenther@suse.de>
7187
0ba6aff3 7188 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB
7189 (pointer_set_lookup): Declare.
7190 (class pointer_map): New template class implementing a
7191 generic pointer to T map.
7192 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
7193 pointer_map<T>::contains, pointer_map<T>::insert,
7194 pointer_map<T>::traverse): New functions.
7195 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
7196 (pointer_set_lookup): New function.
7197 (pointer_set_contains): Use pointer_set_lookup.
7198 (pointer_set_insert): Likewise.
7199 (insert_aux): Remove.
7200 (struct pointer_map_t): Embed a pointer_set_t.
7201 (pointer_map_create): Adjust.
7202 (pointer_map_destroy): Likewise.
7203 (pointer_map_contains): Likewise.
7204 (pointer_map_insert): Likewise.
7205 (pointer_map_traverse): Likewise.
7206 * tree-streamer.h (struct streamer_tree_cache_d): Use a
7207 pointer_map<unsigned> instead of a pointer_map_t.
7208 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
7209 (streamer_tree_cache_lookup): Likewise.
7210 (streamer_tree_cache_create): Likewise.
7211 (streamer_tree_cache_delete): Likewise.
7212 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
7213 pointer_map<unsigned> instead of a pointer_map_t.
7214 (lto_init_tree_ref_encoder): Adjust.
7215 (lto_destroy_tree_ref_encoder): Likewise.
7216 * lto-section-out.c (lto_output_decl_index): Likewise.
7217 (lto_record_function_out_decl_state): Likewise.
7218 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
7219
3dbc97a9
RB
72202013-06-24 Richard Biener <rguenther@suse.de>
7221
7222 PR tree-optimization/57488
7223 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
7224
8345965a
AM
72252013-06-24 Alan Modra <amodra@gmail.com>
7226
7227 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
7228 (gen_easy_altivec_constant): Likewise.
7229 * config/rs6000/predicates.md (easy_vector_constant_add_self,
7230 easy_vector_constant_msb): Likewise.
7231
b7df379f
JJ
72322013-06-23 Jakub Jelinek <jakub@redhat.com>
7233
7234 PR target/57688
7235 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
7236 add missing return true.
7237
12211b99 72382013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE
7239
7240 PR target/52483
7241 * config/sh/predicates.md (general_extend_operand): Invoke
7242 general_movsrc_operand for memory operands.
7243 (general_movsrc_operand): Allow reg+reg addressing, do not use
7244 general_operand for memory operands.
7245
97db2bf7
ST
72462013-06-23 Sriraman Tallam <tmsriram@google.com>
7247
7248 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
7249 when current target options does not apply.
7250 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
7251 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB
7252 * config/i386/bmiintrin.h: Pass appropriate target
7253 attributes to header.
97db2bf7
ST
7254 * config/i386/mmintrin.h: Ditto.
7255 * config/i386/nmmintrin.h: Ditto.
7256 * config/i386/avx2intrin.h: Ditto.
7257 * config/i386/fxsrintrin.h: Ditto.
7258 * config/i386/tbmintrin.h: Ditto.
7259 * config/i386/xsaveintrin.h: Ditto.
7260 * config/i386/f16cintrin.h: Ditto.
7261 * config/i386/xtestintrin.h: Ditto.
7262 * config/i386/xsaveoptintrin.h: Ditto.
7263 * config/i386/bmi2intrin.h: Ditto.
7264 * config/i386/lzcntintrin.h: Ditto.
7265 * config/i386/smmintrin.h: Ditto.
7266 * config/i386/wmmintrin.h: Ditto.
7267 * config/i386/x86intrin.h: Remove all header include guards.
7268 * config/i386/prfchwintrin.h: Ditto.
7269 * config/i386/pmmintrin.h: Ditto.
7270 * config/i386/tmmintrin.h: Ditto.
7271 * config/i386/xmmintrin.h: Ditto.
7272 * config/i386/popcntintrin.h: Ditto.
7273 * config/i386/rdseedintrin.h: Ditto.
7274 * config/i386/ammintrin.h: Ditto.
7275 * config/i386/emmintrin.h: Ditto.
7276 * config/i386/immintrin.h: Remove all header include guards.
7277 * config/i386/fma4intrin.h: Ditto.
7278 * config/i386/lwpintrin.h: Ditto.
7279 * config/i386/xopintrin.h: Ditto.
7280 * config/i386/ia32intrin.h: Ditto.
7281 * config/i386/avxintrin.h: Ditto.
7282 * config/i386/rtmintrin.h: Ditto.
7283 * config/i386/fmaintrin.h: Ditto.
7284 * config/i386/mm3dnow.h: Ditto.
7285
495e6879
ST
72862013-06-22 Sriraman Tallam <tmsriram@google.com>
7287
7288 * common/config/i386/i386-common.c: Handle LZCNT.
7289
11e8f687
AK
72902013-06-22 Andi Kleen <ak@linux.intel.com>
7291
7292 * doc/extend.texi: Use __atomic_store_n instead of
7293 __atomic_store in HLE example.
7294
12211b99 72952013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE
7296
7297 * config/sh/sh.c: Remove <cstdlib> workaround.
7298
927f908b
AK
72992013-06-21 Andi Kleen <ak@linux.intel.com>
7300
0ba6aff3 7301 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 7302
cca410b6
AK
73032013-06-21 Andi Kleen <ak@linux.intel.com>
7304
7305 * doc/extend.texi: Document that __atomic_clear and
7306 __atomic_test_and_set should only be used with bool.
7307
d84db1ed
JH
73082013-06-20 Jan Hubicka <jh@suse.cz>
7309
7310 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
7311 types_same_for_odr.
7312 * tree.c (decls_same_for_odr): New function.
7313 (same_for_edr): New function.
7314 (types_same_for_odr): New function.
7315 (get_binfo_at_offset): Use it.
7316 * tree.h (types_same_for_odr): Declare.
7317
12211b99 73182013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE
7319 Jason Merrill <jason@redhat.com>
7320
7321 * system.h: Include <cstdlib> as well as <stdlib.h>.
7322
49b4ff63
UB
73232013-06-20 Uros Bizjak <ubizjak@gmail.com>
7324
7325 PR target/57655
7326 * config/i386/i386.c (construct_container): Report error if
7327 long double is used with disabled x87 float returns.
7328
7a89ad00
JH
73292013-06-20 Jan Hubicka <jh@suse.cz>
7330
7331 * lto-cgraph.c (input_symtab): Do not set cgraph state.
7332
12211b99 73332013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR
7334
7335 PR rtl-optimization/57425
7336 PR rtl-optimization/57569
7337 * alias.c (write_dependence_p): Remove parameters mem_mode and
7338 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
7339 Changed all callers.
7340 (canon_anti_dependence): Get comments and semantics in sync.
7341 Add parameter mem_canonicalized. Changed all callers.
7342 * rtl.h (canon_anti_dependence): Update prototype.
7343
d16e9a99
RB
73442013-06-20 Richard Biener <rguenther@suse.de>
7345
7346 * data-streamer-in.c (streamer_read_uhwi): Optimize single
7347 byte case, inline streamer_read_uchar and defer section
7348 overrun check.
7349
c3a9b91b
RB
73502013-06-20 Richard Biener <rguenther@suse.de>
7351
7352 PR tree-optimization/57584
7353 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
7354 SSA names into the expanded expression that take part in
7355 abnormal coalescing.
7356
371e77e3 73572013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS
7358
7359 * gcov.c (print_usage): Handle new option.
7360 (process_args): Ditto.
7361 (get_gcov_intermediate_filename): New function.
7362 (output_intermediate_file): New function.
7363 (output_gcov_file): New function
7364 (generate_results): Handle new option.
7365 (release_function): Relase demangled name.
7366 (read_graph_file): Handle demangled name.
7367 (output_lines): Ditto.
7368 * doc/gcov.texi: Document gcov intermediate format.
7369
11f5b71a
VM
73702013-06-19 Vladimir Makarov <vmakarov@redhat.com>
7371
7372 PR bootstrap/57604
7373 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
7374 (lra_emit_add): Use the functions. Add comment about Y as an
7375 address segment.
7376
e1f9340b
DE
73772013-06-19 David Edelsohn <dje.gcc@gmail.com>
7378
7379 PR driver/57652
7380 * collect2.c (collect_atexit): New.
7381 (collect_exit): Delete.
7382 (main): Register collect_atexit with atexit.
7383 (collect_wait): Change collect_exit to exit.
7384 (do_wait): Same.
7385 * collect2.h (collect_exit): Delete.
7386 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
7387
e3f9e0ac
WM
73882013-06-19 Wei Mi <wmi@google.com>
7389
7390 PR rtl-optimization/57518
7391 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
7392 if regno is used in paradoxical subreg.
7393 (update_equiv_regs): Check pdx_subregs[regno] before
7394 set a reg to be equivalent with a mem.
7395
a09f1a76
MK
73962013-06-19 Matthias Klose <doko@ubuntu.com>
7397
7398 PR driver/57651
7399 * file-find.h (find_a_file): Add a mode parameter.
7400 * file-find.c (find_a_file): Likewise.
7401 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
7402 with X_OK for the executables.
7403 * collect2.c (main): Call find_a_file with X_OK.
7404
7d18b0ad
SE
74052013-06-19 Steve Ellcey <sellcey@mips.com>
7406
7407 PR target/56942
7408 * config/mips/mips.md (casesi_internal_mips16_<mode>):
7409 Use NEXT_INSN instead of next_real_insn.
7410
6a6dac52
JH
74112013-06-19 Jan Hubicka <jh@suse.cz>
7412
7413 * cgraph.h (const_value_known_p): Replace by ...
7414 (ctor_for_folding): .. this one.
7415 * cgraphunit.c (process_function_and_variable_attributes): Use it.
7416 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
7417 * expr.c (expand_expr_real_1): Likewise.
7418 (string_constant): Likewise.
7419 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
7420 * ipa.c (process_references): Likewise.
7421 (symtab_remove_unreachable_nodes): Likewise.
7422 * ipa-inline-analysis.c (param_change_prob): Likewise.
7423 * gimple-fold.c (canonicalize_constructor_val): Likewise.
7424 (get_base_constructor): Likwise.
7425 * varpool.c (varpool_remove_node): Likewise.
7426 (varpool_remove_initializer): LIkewise.
7427 (dump_varpool_node): LIkwise.
7428 (const_value_known_p): Rewrite to ...
7429 (ctor_for_folding): ... this one.
7430
216c12ab
JJ
74312013-06-19 Jakub Jelinek <jakub@redhat.com>
7432
7433 PR driver/57651
7434 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
7435 PERSONALITY in $PATH derived prefixes.
7436
d53e2f99
JL
74372013-06-19 Jeff Law <law@redhat.com>
7438
ecdbf306
JL
7439 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
7440 in comment.
7441
d53e2f99
JL
7442 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
7443 (simplify_bitwise_binary): Use it to simpify certain binary ops on
7444 booleans.
7445
89e7fc04
SN
74462013-06-19 Sofiane Naci <sofiane.naci@arm.com>
7447
7448 * config/arm/vfp.md: Move VFP instruction classification documentation
7449 to ...
7450 * config/arm/arm.md: ... here. Update instruction classification
7451 documentation.
7452
5e7f6aaa
RE
74532013-06-19 Richard Earnshaw <rearnsha@arm.com>
7454
7455 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
7456 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
7457 pattern. Use more efficient sequences on ARMv5 and Thumb2.
7458
5fdc1e5d
SB
74592013-06-19 Steven Bosscher <steven@gcc.gnu.org>
7460
7461 PR target/57609
7462 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7463 with NEXT_INSN. Use tablejump_p to check for jump table data
7464 insns.
7465
e952afba
PC
74662013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
7467
7468 PR c++/56544
7469 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
7470 that now in C++ the value is correct per the C++ standards.
7471
3fd0303a
RB
74722013-06-19 Richard Biener <rguenther@suse.de>
7473
7474 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
7475 for global context.
7476
b04e4b08
AK
74772013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7478
7479 Revert:
7480 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7481
7482 PR target/57609
7483 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7484 with next_active_insn.
7485
bef8491a
ST
74862013-06-18 Sriraman Tallam <tmsriram@google.com>
7487
7488 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
7489 functions are inlined during failures to flag an error.
7490 * tree-inline.c (expand_call_inline): Allow the error to be flagged
7491 in early inline pass.
22fc7d3d 7492
c21b257d
L
74932013-06-18 H.J. Lu <hongjiu.lu@intel.com>
7494
7495 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
7496 in comments.
7497
33255ae3
JB
74982013-06-18 Julian Brown <julian@codesourcery.com>
7499
7500 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
7501 Permit virtual register pre-reload if !strict.
7502 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
7503 change.
7504 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
7505 prototype.
7506 * config/arm/neon.md (movmisalign<mode>): Use
7507 neon_perm_struct_or_reg_operand instead of
7508 neon_struct_or_register_operand.
7509 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
7510 neon_permissive_struct_operand instead of neon_struct_operand.
7511 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
7512 neon_vector_mem_operand.
7513 * config/arm/predicates.md (neon_struct_operand): Adjust call to
7514 neon_vector_mem_operand.
7515 (neon_permissive_struct_operand): New.
7516 (neon_struct_or_register_operand): Rename to...
7517 (neon_perm_struct_or_reg_operand): This. Adjust call to
7518 neon_vector_mem_operand.
7519
d579fcda
RB
75202013-06-18 Richard Biener <rguenther@suse.de>
7521
7522 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
7523 * lto-streamer.h: Include pointer-set.h.
7524 (struct lto_decl_slot): Remove.
7525 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
7526 Remove next_index entry.
7527 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7528 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7529 (lto_init_tree_ref_encoder): Adjust.
7530 (lto_destroy_tree_ref_encoder): Likewise.
7531 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
7532 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
7533 (lto_output_decl_index): Adjust.
7534 (lto_new_out_decl_state): Likewise.
7535 (lto_record_function_out_decl_state): Likewise.
7536 * lto-streamer-out.c (copy_function): Likewise.
7537
31ee20ba
RB
75382013-06-18 Richard Biener <rguenther@suse.de>
7539
7540 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
7541 * cgraphunit.c: Include cfgloop.h.
7542 (init_lowered_empty_function): Initialize the loop tree.
7543 (assemble_thunk): Insert new BBs into loops.
7544
ac0511f2
RB
75452013-06-18 Richard Biener <rguenther@suse.de>
7546
7547 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
7548 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
7549 the map from cache entry to cache index optional.
7550 (streamer_tree_cache_replace_tree): Adjust accordingly.
7551 (streamer_tree_cache_append): Likewise.
7552 (streamer_tree_cache_delete): Likewise.
7553 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
7554 streamer cache map from cache entry to cache index.
7555 * lto-streamer-out.c (create_output_block): Adjust.
7556
09485a08
SN
75572013-06-18 Sofiane Naci <sofiane.naci@arm.com>
7558
22fc7d3d
UB
7559 * config/arm/arm.md (attribute "insn"): Move multiplication and
7560 division attributes to...
09485a08
SN
7561 (attribute "type"): ... here. Remove mult.
7562 (attribute "mul32"): New attribute.
7563 (attribute "mul64"): Add umaal.
7564 (*arm_mulsi3): Update attributes.
7565 (*arm_mulsi3_v6): Likewise.
7566 (*thumb_mulsi3): Likewise.
7567 (*thumb_mulsi3_v6): Likewise.
7568 (*mulsi3_compare0): Likewise.
7569 (*mulsi3_compare0_v6): Likewise.
7570 (*mulsi_compare0_scratch): Likewise.
7571 (*mulsi_compare0_scratch_v6): Likewise.
7572 (*mulsi3addsi): Likewise.
7573 (*mulsi3addsi_v6): Likewise.
7574 (*mulsi3addsi_compare0): Likewise.
7575 (*mulsi3addsi_compare0_v6): Likewise.
7576 (*mulsi3addsi_compare0_scratch): Likewise.
7577 (*mulsi3addsi_compare0_scratch_v6): Likewise.
7578 (*mulsi3subsi): Likewise.
7579 (*mulsidi3adddi): Likewise.
7580 (*mulsi3addsi_v6): Likewise.
7581 (*mulsidi3adddi_v6): Likewise.
7582 (*mulsidi3_nov6): Likewise.
7583 (*mulsidi3_v6): Likewise.
7584 (*umulsidi3_nov6): Likewise.
7585 (*umulsidi3_v6): Likewise.
7586 (*umulsidi3adddi): Likewise.
7587 (*umulsidi3adddi_v6): Likewise.
7588 (*smulsi3_highpart_nov6): Likewise.
7589 (*smulsi3_highpart_v6): Likewise.
7590 (*umulsi3_highpart_nov6): Likewise.
7591 (*umulsi3_highpart_v6): Likewise.
7592 (mulhisi3): Likewise.
7593 (*mulhisi3tb): Likewise.
7594 (*mulhisi3bt): Likewise.
7595 (*mulhisi3tt): Likewise.
7596 (maddhisi4): Likewise.
7597 (*maddhisi4tb): Likewise.
7598 (*maddhisi4tt): Likewise.
7599 (maddhidi4): Likewise.
7600 (*maddhidi4tb): Likewise.
7601 (*maddhidi4tt): Likewise.
7602 (divsi3): Likewise.
7603 (udivsi3): Likewise.
7604 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
7605 (thumb2_mulsi_short_compare0): Likewise.
7606 (thumb2_mulsi_short_compare0_scratch): Likewise.
7607 * config/arm/arm1020e.md (1020mult1): Update attribute change.
7608 (1020mult2): Likewise.
7609 (1020mult3): Likewise.
7610 (1020mult4): Likewise.
7611 (1020mult5): Likewise.
7612 (1020mult6): Likewise.
22fc7d3d
UB
7613 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
7614 change.
09485a08
SN
7615 (cortex_a15_mult64): Likewise.
7616 (cortex_a15_sdiv): Likewise.
7617 (cortex_a15_udiv): Likewise.
7618 * config/arm/arm1026ejs.md (mult1): Update attribute change.
7619 (mult2): Likewise.
7620 (mult3): Likewise.
7621 (mult4): Likewise.
7622 (mult5): Likewise.
7623 (mult6): Likewise.
7624 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
7625 (pj4_ir_div): Likewise.
7626 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
7627 (11_mult2): Likewise.
7628 (11_mult3): Likewise.
7629 (11_mult4): Likewise.
7630 (11_mult5): Likewise.
7631 (11_mult6): Likewise.
7632 (11_mult7): Likewise.
7633 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
7634 (cortex_a8_mla): Likewise.
7635 (cortex_a8_mull): Likewise.
7636 (cortex_a8_smulwy): Likewise.
7637 (cortex_a8_smlald): Likewise.
7638 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
7639 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
7640 (cortex_r4_mul_3): Likewise.
7641 (cortex_r4_mla_4): Likewise.
7642 (cortex_r4_mla_3): Likewise.
7643 (cortex_r4_smlald): Likewise.
7644 (cortex_r4_mull): Likewise.
7645 (cortex_r4_sdiv): Likewise.
7646 (cortex_r4_udiv): Likewise.
7647 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
7648 (cortex_a7_idiv): Likewise.
7649 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
7650 (9_mult2): Likewise.
7651 (9_mult3): Likewise.
7652 (9_mult4): Likewise.
7653 (9_mult5): Likewise.
7654 (9_mult6): Likewise.
7655 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
7656 (cortex_a53_sdiv): Likewise.
7657 (cortex_a53_udiv): Likewise.
7658 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
7659 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
7660 (mp626_mult2): Likewise.
7661 (mp626_mult3): Likewise.
7662 (mp626_mult4): Likewise.
7663 * config/arm/fa526.md (526_mult1): Update attribute change.
7664 (526_mult2): Likewise.
7665 * config/arm/arm-generic.md (mult): Update attribute change.
7666 (mult_ldsched_strongarm): Likewise.
7667 (mult_ldsched): Likewise.
7668 (multi_cycle): Likewise.
7669 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
7670 * config/arm/fa606te.md (606te_mult1): Update attribute change.
7671 (606te_mult2): Likewise.
7672 (606te_mult3): Likewise.
7673 (606te_mult4): Likewise.
7674 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
7675 (cortex_a9_mac16): Likewise.
7676 (cortex_a9_multiply): Likewise.
7677 (cortex_a9_mac): Likewise.
7678 (cortex_a9_multiply_long): Likewise.
7679 * config/arm/fa626te.md (626te_mult1): Update attribute change.
7680 (626te_mult2): Likewise.
7681 (626te_mult3): Likewise.
7682 (626te_mult4): Likewise.
7683
8951f345
RB
76842013-06-18 Richard Biener <rguenther@suse.de>
7685
7686 PR lto/57334
7687 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
7688
d8101d05
AK
76892013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7690
7691 PR target/57609
7692 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
7693 with next_active_insn.
7694
69eff9da
AM
76952013-06-18 Alan Modra <amodra@gmail.com>
7696
7697 * config/rs6000/rs6000.h (enum data_align): New.
7698 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
7699 (DATA_ABI_ALIGNMENT): Define.
7700 (CONSTANT_ALIGNMENT): Correct comment.
7701 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
7702 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
7703
8ed00d76
DM
77042013-06-17 David Malcolm <dmalcolm@redhat.com>
7705
7706 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
7707 ATTRIBUTE_UNUSED marking.
7708
3d33d06b
SN
77092013-06-17 Sofiane Naci <sofiane.naci@arm.com>
7710
7711 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
7712 alternative and update.
7713 (aarch64_dup_lanedi): Delete.
7714 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
7715 * config/aarch64/aarch64-simd-builtins.def: Update.
7716
ee03e71d
RB
77172013-06-17 Richard Biener <rguenther@suse.de>
7718
7719 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
7720 (lto_input_scc): Declare.
7721 (lto_input_tree_1): Likewise.
7722 (struct lto_stats_d): Add num_tree_bodies_output and
7723 num_pickle_refs_output.
7724 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
7725 (lto_read_tree_1): Split out from ...
7726 (lto_read_tree): ... this.
7727 (lto_input_scc): New function.
7728 (lto_input_tree_1): Split out from ...
7729 (lto_input_tree): ... this. Handle LTO_tree_scc.
7730 (lto_data_in_create): Create the streamer cache without hashes.
7731 * lto-streamer-out.c (create_output_block): Create the streamer
7732 cache with hashes when not doing WPA.
7733 (lto_write_tree_1): Split out from ...
7734 (lto_write_tree): ... this.
7735 (get_symbol_initial_value): New function.
7736 (lto_output_tree_1): Split out from ...
7737 (lto_output_tree): ... this. Write trees as series of SCCs
7738 using a DFS walk via DFS_write_tree.
7739 (struct sccs, struct scc_entry): New types.
7740 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
7741 (DFS_write_tree_body): New function.
7742 (DFS_write_tree): Likewise.
7743 (hash_tree): Likewise.
7744 (scc_entry_compare): Likewise.
7745 (hash_scc): Likewise.
7746 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
7747 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
7748 TREE_CHAIN as regular reference.
7749 (streamer_read_integer_cst): Remove.
7750 (streamer_get_pickled_tree): Adjust.
7751 * tree-streamer-out.c (streamer_write_chain): Disable streaming
7752 of DECL_EXTERNALs in BLOCK_VARS for now.
7753 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
7754 reference.
7755 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
7756 Add hash value argument and record that if hashes are recorded
7757 in the cache.
7758 (streamer_tree_cache_insert_1): Adjust.
7759 (streamer_tree_cache_insert): Likewise.
7760 (streamer_tree_cache_insert_at): Rename to ...
7761 (streamer_tree_cache_replace_tree): ... this and adjust.
7762 (streamer_tree_cache_append): Adjust.
7763 (record_common_node): Likewise.
7764 (streamer_tree_cache_create): Add argument whether to
7765 record hash values together with trees.
7766 (streamer_tree_cache_delete): Adjust.
7767 * tree-streamer.h (struct streamer_tree_cache_d): Add
7768 vector of hashes.
7769 (streamer_read_integer_cst): Remove.
7770 (streamer_tree_cache_insert): Adjust.
7771 (streamer_tree_cache_append): Likewise.
7772 (streamer_tree_cache_insert_at): Rename to ...
7773 (streamer_tree_cache_replace_tree): ... this and adjust.
7774 (streamer_tree_cache_create): Add argument whether to record hashes.
7775 (streamer_tree_cache_get): Rename to ...
7776 (streamer_tree_cache_get_tree): ... this.
7777 (streamer_tree_cache_get_hash): New function.
7778 * tree.c (cache_integer_cst): New function.
7779 * tree.h (cache_integer_cst): Declare.
7780 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
7781 * lto-symtab.c (lto_varpool_replace_node): Only release
7782 DECL_INITIAL of non-prevailing decls.
7783 * varpool.c (varpool_remove_initializer): Do not release
7784 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
7785
0de86a92
JU
77862013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
7787
7788 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
7789 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
7790 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
7791 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
7792 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
7793 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
7794 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
7795 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
7796 instead of TARGET_64BIT.
7797 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
7798 Require ISA_HAS_<D>DIV.
7799
74240413
RS
78002013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
7801
7802 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
7803 (mips*-*-linux*): Move default with_llsc setting to where other
7804 defaults are set.
7805 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
7806 with_arch block.
7807 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
7808 Likewise. Remove default with_tune setting. Move default float
7809 setting to its own block. Handle with_llsc in the same block as above.
7810
12211b99 78112013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR
7812
7813 PR rtl-optimization/57425
7814 PR rtl-optimization/57569
299a5f6a 7815 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR
7816 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
7817 Changed all callers.
7818 (canon_anti_dependence): New function.
7819 * cse.c (check_dependence): Use canon_anti_dependence.
7820 * cselib.c (cselib_invalidate_mem): Likewise.
7821 * rtl.h (canon_anti_dependence): Declare.
7822
b259d352
JU
78232013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
7824
7825 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
7826 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
7827 ".set mips3" for 64-bit targets.
7828
4de09b85
DC
78292013-06-15 Dehao Chen <dehao@google.com>
7830
7831 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
7832 * gimple-low.c (gimple_check_call_matching_types): Likewise.
7833 (gimple_check_call_args): Likewise.
7834 * value-prof.c (check_ic_target): Likewise.
7835 * ipa-inline.c (early_inliner): Likewise.
7836 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
7837 * cgraph.c (cgraph_create_edge_1): Likewise.
7838 (cgraph_make_edge_direct): Likewise.
7839
d1aa4a76
MM
78402013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
7841
7842 PR target/57615
7843 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
7844 rs6000_output_move_128bit to handle emitting quad memory
7845 operations. Set attribute length to 8 bytes.
7846
12211b99 78472013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP
7848
7849 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
7850 New pattern.
7851 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
7852 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
7853 (aarch64_<su>mlsl<mode>): Likewise.
7854
f5273574
MS
78552013-06-14 Mike Stump <mikestump@comcast.net>
7856
7857 * Makefile.in (TARGET_H): Add insn-codes.h.
7858
ef7e7ab2
AM
78592013-06-14 Alan Modra <amodra@gmail.com>
7860
7861 PR middle-end/57134
7862 PR middle-end/57586
7863 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
7864 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
7865 bitfield expansion when EXPAND_MEMORY.
7866 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
7867
135faab6
MM
78682013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
7869
7870 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
7871 test for clearing quad memory on 32-bit later.
7872
948a1fd9
MG
78732013-06-13 Marc Glisse <marc.glisse@inria.fr>
7874
7875 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
7876 (fold_negate_expr): Likewise.
7877 (fold_real_zero_addition_p): Handle vectors.
7878 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
7879
0597fb9c
AM
78802013-06-14 Alan Modra <amodra@gmail.com>
7881
7882 * varasm.c (force_const_mem): Revert 2013-06-07 change.
7883
e5b962d0
JH
78842013-06-13 Jan Hubicka <jh@suse.cz>
7885
7886 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
7887 Local comdats are not externally visible.
7888 * symtab.c (dump_symtab_base): Dump externally visible.
7889 (verify_symtab_base): Verify back links in the symtab hash.
7890
a8355e51
BC
78912013-06-13 Bin Cheng <bin.cheng@arm.com>
7892
7893 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
7894 CONVERT_EXPR as equal nodes.
7895
8f7e6e33
BC
78962013-06-13 Bin Cheng <bin.cheng@arm.com>
7897
22fc7d3d 7898 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 7899
5d418483
MG
79002013-06-13 Marc Glisse <marc.glisse@inria.fr>
7901
7902 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
7903 Generalize to complex and vector.
7904 * tree.c (build_all_ones_cst): New function.
7905 * tree.h (build_all_ones_cst): Declare it.
7906
50751417
AM
79072013-06-13 Alan Modra <amodra@gmail.com>
7908
7909 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
7910 * config/rs6000/rs6000.md (signbittf2): New insn.
7911 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
7912 (abstf2_internal, cmptf_internal2): Likewise.
7913 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
7914
4b02c962 79152013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
7916 Pat Haugen <pthaugen@us.ibm.com>
7917 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM
7918
7919 * config/rs6000/rs6000.c (emit_load_locked): Add support for
7920 power8 byte, half-word, and quad-word atomic instructions.
7921 (emit_store_conditional): Likewise.
7922 (rs6000_expand_atomic_compare_and_swap): Likewise.
7923 (rs6000_expand_atomic_op): Likewise.
7924
7925 * config/rs6000/sync.md (larx): Add new modes for power8.
7926 (stcx): Likewise.
7927 (AINT): New mode iterator to include TImode as well as normal
7928 integer modes on power8.
7929 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
7930 that VSX registers are not considered. Use AINT mode iterator
7931 instead of INT1 to allow inclusion of quad word atomic operations
7932 on power8.
7933 (load_locked<mode>): Likewise.
7934 (store_conditional<mode>): Likewise.
7935 (atomic_compare_and_swap<mode>): Likewise.
7936 (atomic_exchange<mode>): Likewise.
7937 (atomic_nand<mode>): Likewise.
7938 (atomic_fetch_<fetchop_name><mode>): Likewise.
7939 (atomic_nand_fetch<mode>): Likewise.
7940 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
7941 each type.
7942 (ATOMIC): On power8, add QImode, HImode modes.
7943 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
7944 modes that promote to SImode.
7945 (load_lockedti): Convert TImode arguments to PTImode, so that we
7946 get a guaranteed even/odd register pair.
7947 (load_lockedpti): Likewise.
7948 (store_conditionalti): Likewise.
7949 (store_conditionalpti): Likewise.
7950
7951 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
7952 atomic load/store instructions.
7953 (HSI): Likewise.
7954
710b6cc1
RS
79552013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
7956
7957 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
7958 loads.
7959 (insn_count): New attribute, with most cases extracted from...
7960 (length): ...here. Redefine most cases in terms of insn_count.
7961 (single_insn): Delete.
7962 (can_delay): Use insn_count to check for single instructions.
7963 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
7964 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
7965 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
7966 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
7967 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
7968 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
7969 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
7970 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
7971 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
7972 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
7973 rather than "length".
7974 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
7975 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
7976 Use "insn_count" rather than "length".
7977 * config/mips/mips-dsp.md
7978 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
7979 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
7980 length attributes.
7981
1951f101
MG
79822013-06-12 Marc Glisse <marc.glisse@inria.fr>
7983
7984 PR tree-optimization/57361
7985 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
7986
8b033a8a
SN
79872013-06-12 Sofiane Naci <sofiane.naci@arm.com>
7988
22fc7d3d
UB
7989 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
7990 to split.
8b033a8a
SN
7991 (aarch64_simd_combine<mode>): New instruction expansion.
7992 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
7993 function prototype.
7994 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
7995 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
7996
ca0f62a8
JH
79972013-06-12 Jan Hubicka <jh@suse.cz>
7998
7999 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
8000 decl has when in streaming stage.
8001 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
8002 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
8003
12211b99 80042013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS
8005
8006 PR target/57578
8007 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
8008
bfdeda2c
JJ
80092013-06-12 Jakub Jelinek <jakub@redhat.com>
8010
8011 PR tree-optimization/57537
8012 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
8013 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
8014
a4fa02d1
RB
80152013-06-12 Richard Biener <rguenther@suse.de>
8016
8017 * data-streamer.h (streamer_write_char_stream): CSE
8018 obs->current_pointer.
8019 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
8020 streamer_write_char_stream manually and optimize the resulting loop.
8021 (streamer_write_hwi_stream): Likewise.
8022
bbf9ad07
JH
80232013-06-12 Jan Hubicka <jh@suse.cz>
8024
8025 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
8026 * cgraph.h (varpool_create_empty_node): Declare.
8027 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
8028 duplicated nodes.
8029 * symtab.c (symtab_unregister_node): Be lax about missin entries
8030 in node hash.
8031 (symtab_get_node): Update comment.
8032 * varpool.c (varpool_create_empty_node): Break out from ...
8033 (varpool_node_for_decl): ... here.
8034 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
8035
a53f90ad
EB
80362013-06-12 Eric Botcazou <ebotcazou@adacore.com>
8037
8038 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
8039 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
8040 part. Use straight-line flow at the end.
8041 <COMPONENT_REF>: Remove superfluous else.
8042 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
8043
946f9306
JJ
80442013-06-12 Jakub Jelinek <jakub@redhat.com>
8045
8046 PR target/56564
8047 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
8048 target hook even for !TREE_PUBLIC decls. If no resolution info
8049 is available, return false for common and external decls.
8050
a7fca6f0
KP
80512013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8052
8053 * config/rl78/constraints.md (U): New constraint.
8054 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
8055 valloc attribute.
8056
c7d8f446
MM
80572013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
8058
8059 PR target/57589
8060 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
8061 to allow returning address to AT_PLATFORM name.
8062
08346abd
JH
80632013-06-11 Jan Hubicka <jh@suse.cz>
8064
8065 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
8066 * cgraph.h (symtab_node_base): Add weakref flag.
8067 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
8068 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
8069 (output_weakrefs): Use weakref flag.
8070 * fold-const.c (simple_operand_p): Handle WEAK.
8071 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
8072 * ipa.c (varpool_externally_visible_p): Drop weakref.
8073 (function_and_variable_visibility): Update comment; fix weakref
8074 sanity checks; do not clear DECL_WEAK on them.
8075 * lto-cgraph.c (lto_output_node): update.
8076 (lto_output_varpool_node): Update.
8077 (input_overwrite_node): Update.
8078 (input_node): Update.
8079 (input_varpool_node): Update.
8080 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
8081 (lto_symtab_merge_symbols): Add sanity check.
8082 (lto_symtab_prevailing_decl): Do not special case weakrefs.
8083 * passes.c (rest_of_decl_compilation): Set static flag, too.
8084 * symtab.c (dump_symtab_base): Dump weakref.
8085 (verify_symtab_base): Sanity check weakrefs.
8086 (symtab_make_decl_local): Remove duplicated code.
8087 (symtab_alias_ultimate_target): Simplify.
8088 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 8089
d9af3652
DD
80902013-06-11 DJ Delorie <dj@redhat.com>
8091
8092 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
8093 (rl78_unwind_word_mode): New.
8094
ffa66012
DM
80952013-06-11 David Malcolm <dmalcolm@redhat.com>
8096
8097 * final.c (debug_prefix_maps): Make static.
8098
783f2d35
DM
80992013-06-11 David Malcolm <dmalcolm@redhat.com>
8100
8101 * function.c (initial_trampoline): Remove stray copy.
8102
7af26def
SN
81032013-06-11 Sofiane Naci <sofiane.naci@arm.com>
8104
8105 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
8106
97756c0e
MJ
81072013-06-11 Martin Jambor <mjambor@suse.cz>
8108
8109 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
8110 within bounds at the beginning of the function.
8111
2c7b8bf6
AM
81122013-06-11 Alan Modra <amodra@gmail.com>
8113
8114 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
8115 reporting.
8116 (get_named_section): Don't NULL !DECL_P decl.
8117
77cea46e
IZ
81182013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
8119
8120 * doc/invoke.texi (core-avx2): Document.
8121 (slm): Likewise.
8122 (atom): Updated with MOVBE.
8123
888f0920
RB
81242013-06-11 Richard Biener <rguenther@suse.de>
8125
22fc7d3d 8126 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 8127
6b438d58
AB
81282013-06-11 Anton Blanchard <anton@samba.org>
8129
8130 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
8131 correct shift value in little-endian mode.
8132
3bc66938
JJ
81332013-06-11 Jakub Jelinek <jakub@redhat.com>
8134
8135 PR target/56564
8136 * varasm.c (get_variable_align): Move #endif to the right place.
8137
a5dc7f88
CC
81382013-06-10 Cary Coutant <ccoutant@google.com>
8139
8140 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
8141 for hash so that hash table traversal order is deterministic.
8142
19be72ab 81432013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
8144 Pat Haugen <pthaugen@us.ibm.com>
8145 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM
8146
8147 * config/rs6000/vector.md (GPR move splitter): Do not split moves
8148 of vectors in GPRS if they are direct moves or quad word load or
8149 store moves.
8150
8151 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
8152 declaration.
8153 (direct_move_p): Likewise.
8154 (quad_load_store_p): Likewise.
8155
8156 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
8157 classes into bins based on the physical register type.
8158 (reg_class_to_reg_type): Likewise.
8159 (IS_STD_REG_TYPE): Likewise.
8160 (IS_FP_VECT_REG_TYPE): Likewise.
8161 (reload_fpr_gpr): Arrays to determine what insn to use if we can
8162 use direct move instructions.
8163 (reload_gpr_vsx): Likewise.
8164 (reload_vsx_gpr): Likewise.
8165 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
8166 information that is a simplification of register classes. Also
8167 precalculate direct move reload helpers.
8168 (direct_move_p): New function to return true if the operation can
8169 be done as a direct move instruciton.
8170 (quad_load_store_p): New function to return true if the operation
8171 is a quad memory operation.
8172 (rs6000_legitimize_address): If quad memory, only allow register
8173 indirect for TImode addresses.
8174 (rs6000_legitimate_address_p): Likewise.
8175 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
8176 (rs6000_reload_register_type): Likewise.
8177 (register_to_reg_type): Return register type.
8178 (rs6000_secondary_reload_simple_move): New helper function for
8179 secondary reload and secondary memory needed to identify anything
8180 that is a simple move, and does not need reloading.
8181 (rs6000_secondary_reload_direct_move): New helper function for
8182 secondary reload to identify cases that can be done with several
8183 instructions via the direct move instructions.
8184 (rs6000_secondary_reload_move): New helper function for secondary
8185 reload to identify moves between register types that can be done.
8186 (rs6000_secondary_reload): Add support for quad memory operations
8187 and for direct move.
8188 (rs6000_secondary_memory_needed): Likewise.
8189 (rs6000_debug_secondary_memory_needed): Change argument names.
8190 (rs6000_output_move_128bit): New function to return the move to
8191 use for 128-bit moves, including knowing about the various
8192 limitations of quad memory operations.
8193
8194 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
8195 memory operations. call rs6000_output_move_128bit for the actual
8196 instruciton(s) to generate.
8197 (vsx_movti_64bit): Likewise.
8198
8199 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
8200 (UNSPEC_P8V_MTVSRWZ): Likewise.
8201 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
8202 (UNSPEC_P8V_MTVSRD): Likewise.
8203 (UNSPEC_P8V_XXPERMDI): Likewise.
8204 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
8205 (UNSPEC_FUSION_GPR): Likewise.
8206 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 8207 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM
8208 (f32_sv): Likewise.
8209 (f32_dm): Likewise.
8210 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
8211 loads and direct move instructions.
8212 (zero_extendsidi2_lfiwzx): Likewise.
8213 (extendsidi2_lfiwax): Likewise.
8214 (extendsidi2_nocell): Likewise.
8215 (floatsi<mode>2_lfiwax): Likewise.
8216 (lfiwax): Likewise.
8217 (floatunssi<mode>2_lfiwzx): Likewise.
8218 (lfiwzx): Likewise.
8219 (fix_trunc<mode>_stfiwx): Likewise.
8220 (fixuns_trunc<mode>_stfiwx): Likewise.
8221 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
8222 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
8223 (parity<mode>2_cmpb): Set length/type attr.
8224 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
8225 for 'mr.' to fast_compare.
8226 (bpermd_<mode>): Change type attr to popcnt.
8227 (p8_fmrgow_<mode>): New insns for power8 direct move support.
8228 (p8_mtvsrwz_1): Likewise.
8229 (p8_mtvsrwz_2): Likewise.
8230 (reload_fpr_from_gpr<mode>): Likewise.
8231 (p8_mtvsrd_1): Likewise.
8232 (p8_mtvsrd_2): Likewise.
8233 (p8_xxpermdi_<mode>): Likewise.
8234 (reload_vsx_from_gpr<mode>): Likewise.
8235 (reload_vsx_from_gprsf): Likewise.
8236 (p8_mfvsrd_3_<mode>): LIkewise.
8237 (reload_gpr_from_vsx<mode>): Likewise.
8238 (reload_gpr_from_vsxsf): Likewise.
8239 (p8_mfvsrd_4_disf): Likewise.
8240 (multi-word GPR splits): Do not split direct moves or quad memory
8241 operations.
8242
16876bdc
DM
82432013-06-10 David Malcolm <dmalcolm@redhat.com>
8244
8245 * tree-into-ssa.c (interesting_blocks): Make static.
8246
df8a1d28
JJ
82472013-06-10 Jakub Jelinek <jakub@redhat.com>
8248
8249 PR target/56564
8250 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
8251 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
8252 Use DATA_ABI_ALIGNMENT for that case instead if defined.
8253 (get_variable_align): New function.
8254 (get_variable_section, emit_bss, emit_common,
8255 assemble_variable_contents, place_block_symbol): Use
8256 get_variable_align instead of DECL_ALIGN.
8257 (assemble_noswitch_variable): Add align argument, use it
8258 instead of DECL_ALIGN.
8259 (assemble_variable): Adjust caller. Use get_variable_align
8260 instead of DECL_ALIGN.
8261 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
8262 caller.
8263 (DATA_ABI_ALIGNMENT): Define.
8264 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
8265 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
8266 opt is false, only return the psABI mandated alignment increase.
8267 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
8268 (DATA_ABI_ALIGNMENT): ... this.
8269 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
8270 (DATA_ABI_ALIGNMENT): ... this.
8271 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
8272 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
8273 (DATA_ABI_ALIGNMENT): ... this.
8274 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
8275 * doc/tm.texi: Regenerated.
8276
57b29ca6
UB
82772013-06-10 Uros Bizjak <ubizjak@gmail.com>
8278
8279 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
8280 cmp_code to construct REG_EQUAL note.
8281
3e56ed50
JJ
82822013-06-09 Jakub Jelinek <jakub@redhat.com>
8283
8284 PR target/57568
8285 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
8286 that operands[2] doesn't overlap with operands[0].
8287
74fe2a1d
DE
82882013-06-09 David Edelsohn <dje.gcc@gmail.com>
8289 Jan Hubicka <jh@suse.cz>
8290
8291 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
8292 hack to mark symbols as used.
8293
ed52a84e
VM
82942013-06-08 Vladimir Makarov <vmakarov@redhat.com>
8295
8296 PR rtl-optimization/57559
8297 * lra-constraints.c (process_alt_operands): Don't discourage
8298 memory with known offset for offsetable memory constraint.
8299 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
8300
299316ed
EB
83012013-06-08 Eric Botcazou <ebotcazou@adacore.com>
8302
8303 * varasm.c (struct oc_local_state): Reorder fields.
8304 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
8305 and adjust accordingly.
8306 (output_constructor): Reorder initialization code and adjust call to
8307 output_constructor_bitfield.
8308
88ac60d3
JH
83092013-06-07 Jan Hubicka <jh@suse.cz>
8310
8311 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
8312
4cdce1a8
DM
83132013-06-07 David Malcolm <dmalcolm@redhat.com>
8314
8315 * tree-object-size.c (unknown): Make const.
8316
b5e0425c
AK
83172013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8318
8319 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
8320 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
8321 for last alternative in the cpu_facility attribute.
8322
79678d04
KT
83232013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8324
8325 PR target/56315
8326 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
8327 (xordi3): Change operand 2 constraint to arm_xordi_operand.
8328 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
8329 * config/arm/constraints.md (Dg): New constraint.
8330 * config/arm/neon.md (xordi3_neon): Remove.
8331 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
8332 * config/arm/predicates.md (arm_xordi_operand): New predicate.
8333
b31ddbdb
KT
83342013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8335
8336 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
8337 Clean up alternatives.
8338
8f90b7d4
AM
83392013-06-07 Alan Modra <amodra@gmail.com>
8340
8341 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
8342 va_list_gpr_size.
8343
961ce119
AM
83442013-06-07 Alan Modra <amodra@gmail.com>
8345
8346 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
8347
33e49835
KT
83482013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8349
8350 * config/arm/constraints.md (Df): New constraint.
8351 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
8352 Correct length attribute for last two alternatives.
8353
acd17ae6
AM
83542013-06-07 Alan Modra <amodra@gmail.com>
8355
8356 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
8357 override user -mfp-in-toc.
8358 (offsettable_ok_by_alignment): Consider just the current access
8359 rather than the whole object, unless BLKmode. Handle
8360 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
8361 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
8362 for -mcmodel=medium.
8363 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
8364 override user -mfp-in-toc or -msum-in-toc. Default to
8365 -mno-fp-in-toc for -mcmodel=medium.
8366
73310b0e
DD
83672013-06-06 DJ Delorie <dj@redhat.com>
8368
8369 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
8370 TARGET_VALID_POINTER_MODE.
8371
0bd62dca 83722013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
8373 Pat Haugen <pthaugen@us.ibm.com>
8374 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM
8375
8376 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8377 Document new power8 builtins.
8378
8379 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
8380 condition code register, to allow 128-bit logical operations to be
8381 done in the VSX or GPR registers.
8382 (nor<mode>3): Use the canonical form for nor.
8383 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
8384 vclz*, and vpopcnt* vector instructions.
8385 (nand<mode>3): Likewise.
8386 (orc<mode>3): Likewise.
8387 (clz<mode>2): LIkewise.
8388 (popcount<mode>2): Likewise.
8389
8390 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
8391 that only the GPRs are recognized.
8392
8393 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8394 support for new power8 builtins.
8395
8396 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
8397 builtin functions.
8398 (xscvdpspn): Likewise.
8399 (vclz): Likewise.
8400 (vclzb): Likewise.
8401 (vclzh): Likewise.
8402 (vclzw): Likewise.
8403 (vclzd): Likewise.
8404 (vpopcnt): Likewise.
8405 (vpopcntb): Likewise.
8406 (vpopcnth): Likewise.
8407 (vpopcntw): Likewise.
8408 (vpopcntd): Likewise.
8409 (vgbbd): Likewise.
8410 (vmrgew): Likewise.
8411 (vmrgow): Likewise.
8412 (eqv): Likewise.
8413 (eqv_v16qi3): Likewise.
8414 (eqv_v8hi3): Likewise.
8415 (eqv_v4si3): Likewise.
8416 (eqv_v2di3): Likewise.
8417 (eqv_v4sf3): Likewise.
8418 (eqv_v2df3): Likewise.
8419 (nand): Likewise.
8420 (nand_v16qi3): Likewise.
8421 (nand_v8hi3): Likewise.
8422 (nand_v4si3): Likewise.
8423 (nand_v2di3): Likewise.
8424 (nand_v4sf3): Likewise.
8425 (nand_v2df3): Likewise.
8426 (orc): Likewise.
8427 (orc_v16qi3): Likewise.
8428 (orc_v8hi3): Likewise.
8429 (orc_v4si3): Likewise.
8430 (orc_v2di3): Likewise.
8431 (orc_v4sf3): Likewise.
8432 (orc_v2df3): Likewise.
8433
8434 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
8435 allow power8 quad mode in 64-bit.
8436 (rs6000_builtin_vectorized_function): Add support to vectorize
8437 ISA 2.07 count leading zeros, population count builtins.
8438 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
8439 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
8440 (builtin_function_type): Add vgbbd builtin function which takes an
8441 unsigned argument.
8442 (altivec_expand_vec_perm_const): Add support for new power8 merge
8443 instructions.
8444
8445 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
8446 that does not include TImdoe for use with 32-bit.
8447 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
8448 instructions.
8449 (UNSPEC_VSX_CVDPSPN): Likewise.
8450 (vsx_xscvdpspn): Likewise.
8451 (vsx_xscvspdpn): Likewise.
8452 (vsx_xscvdpspn_scalar): Likewise.
8453 (vsx_xscvspdpn_directmove): Likewise.
8454 (vsx_and<mode>3): Split logical operations into 32-bit and
8455 64-bit. Add support to do logical operations on TImode as well as
8456 VSX vector types. Allow logical operations to be done in either
8457 VSX registers or in general purpose registers in 64-bit mode. Add
8458 splitters if GPRs were used. For AND, add clobber of CCmode to
8459 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
8460 encoding.
8461 (vsx_and<mode>3_32bit): Likewise.
8462 (vsx_and<mode>3_64bit): Likewise.
8463 (vsx_ior<mode>3): Likewise.
8464 (vsx_ior<mode>3_32bit): Likewise.
8465 (vsx_ior<mode>3_64bit): Likewise.
8466 (vsx_xor<mode>3): Likewise.
8467 (vsx_xor<mode>3_32bit): Likewise.
8468 (vsx_xor<mode>3_64bit): Likewise.
8469 (vsx_one_cmpl<mode>2): Likewise.
8470 (vsx_one_cmpl<mode>2_32bit): Likewise.
8471 (vsx_one_cmpl<mode>2_64bit): Likewise.
8472 (vsx_nor<mode>3): Likewise.
8473 (vsx_nor<mode>3_32bit): Likewise.
8474 (vsx_nor<mode>3_64bit): Likewise.
8475 (vsx_andc<mode>3): Likewise.
8476 (vsx_andc<mode>3_32bit): Likewise.
8477 (vsx_andc<mode>3_64bit): Likewise.
8478 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
8479 and xxlorc instructions.
8480 (vsx_eqv<mode>3_64bit): Likewise.
8481 (vsx_nand<mode>3_32bit): Likewise.
8482 (vsx_nand<mode>3_64bit): Likewise.
8483 (vsx_orc<mode>3_32bit): Likewise.
8484 (vsx_orc<mode>3_64bit): Likewise.
8485
8486 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
8487
8488 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
8489 instruction.
8490 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
8491 (p8_vmrgow): Likewise.
8492 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
8493 GPRs to be split under VSX.
8494 (p8v_clz<mode>2): Add power8 count leading zero support.
8495 (p8v_popcount<mode>2): Add power8 population count support.
8496 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
8497 support.
8498
8499 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
8500 instruction.
8501
8502 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
8503 builtin functions.
8504 (vec_nand): Likewise.
8505 (vec_vclz): Likewise.
8506 (vec_vclzb): Likewise.
8507 (vec_vclzd): Likewise.
8508 (vec_vclzh): Likewise.
8509 (vec_vclzw): Likewise.
8510 (vec_vgbbd): Likewise.
8511 (vec_vmrgew): Likewise.
8512 (vec_vmrgow): Likewise.
8513 (vec_vpopcnt): Likewise.
8514 (vec_vpopcntb): Likewise.
8515 (vec_vpopcntd): Likewise.
8516 (vec_vpopcnth): Likewise.
8517 (vec_vpopcntw): Likewise.
8518
37684c46
VM
85192013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8520
8521 PR rtl-optimization/57468
8522 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
8523 spilled pseudos.
8524
d89ae27c
VM
85252013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8526
8527 PR rtl-optimization/57459
8528 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
8529 type when setting live regs.
8530
3597e113
VM
85312013-06-06 Vladimir Makarov <vmakarov@redhat.com>
8532
8533 * config/s390/s390.opt (mlra): New option.
8534 * config/s390/s390.c (s390_decompose_address): Check displacement
8535 for all registers for LRA.
8536 (s390_secondary_reload): Don't used secondary reloads for LRA.
8537 (s390_lra_p): New function.
8538 (TARGET_LRA_P): Define.
8539 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
8540 of attribute cpu_facility to zarch for the last alternative.
8541 (*cmpmem_short): Ditto.
8542
01037aeb
EB
85432013-06-06 Eric Botcazou <ebotcazou@adacore.com>
8544
8545 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
8546 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
8547 (arm_expand_prologue): Likewise.
8548
3371a64f
TJ
85492013-06-06 Teresa Johnson <tejohnson@google.com>
8550
8551 PR c++/53743
8552 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
8553 as this is now done by redirect_edge_and_branch_force.
8554 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
8555 barriers, and fix interaction with splitting.
8556 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
8557 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
8558 reflect changes made in the routine.
8559 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
8560 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
8561 since this is called in cfglayout mode, and replace partition fixup
8562 with assert as that is now done by force_nonfallthru_and_redirect.
8563 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
8564 already be marked with region crossing note.
8565 (insert_section_boundary_note): Make non-static, gate on flag
8566 has_bb_partition, rewrite to also check for multiple partitions.
8567 (rest_of_handle_reorder_blocks): Remove call to
8568 insert_section_boundary_note, now done later during free_cfg.
8569 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
8570 * bb-reorder.h (insert_section_boundary_note): Declare.
8571 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
8572 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
8573 invoke insert_section_boundary_note.
8574 (try_redirect_by_replacing_jump): Remove unnecessary
8575 check for region crossing note.
8576 (fixup_partition_crossing): New function.
8577 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
8578 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
8579 in non-cfglayout mode.
8580 (force_nonfallthru_and_redirect): Fixup partition boundaries,
8581 remove old code that tried to do this. Emit barrier correctly
8582 when we are in cfglayout mode.
8583 (last_bb_in_partition): New function.
8584 (rtl_split_edge): Correctly fixup partition boundaries.
8585 (commit_one_edge_insertion): Remove old code that tried to
8586 fixup region crossing edge since this is now handled in
8587 split_block, and set up insertion point correctly since
8588 block may now end in a jump.
8589 (verify_hot_cold_block_grouping): Guard against checking when not in
8590 linearized RTL mode.
8591 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
8592 notes.
8593 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
8594 rtl_verify_flow_info, so not called in cfglayout mode.
8595 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
8596 (fixup_reorder_chain): Remove old code that attempted to fixup region
8597 crossing note as this is now handled in force_nonfallthru_and_redirect.
8598 (duplicate_insn_chain): Don't duplicate switch section notes.
8599 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
8600 note.
8601 * basic-block.h (emit_barrier_after_bb): Declare.
8602
66071e10
KT
86032013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8604
8605 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
8606 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
8607 arm_usatsihi): Adjust alternatives for arm_restrict_it.
8608
0e26bf3d
KT
86092013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8610
8611 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
8612 where appropriate.
8613 * config/arm/ldmstm.md: Regenerate.
8614
12b4e7ef
KT
86152013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8616
8617 * config/arm/sync.md (atomic_loaddi_1):
8618 Disable predication for arm_restrict_it.
8619 (arm_load_exclusive<mode>): Likewise.
8620 (arm_load_exclusivesi): Likewise.
8621 (arm_load_exclusivedi): Likewise.
8622 (arm_load_acquire_exclusive<mode>): Likewise.
8623 (arm_load_acquire_exclusivesi): Likewise.
8624 (arm_load_acquire_exclusivedi): Likewise.
8625 (arm_store_exclusive<mode>): Likewise.
8626 (arm_store_exclusive<mode>): Likewise.
8627 (arm_store_release_exclusivedi): Likewise.
8628 (arm_store_release_exclusive<mode>): Likewise.
8629
15d16c8a
RB
86302013-06-06 Richard Biener <rguenther@suse.de>
8631
8632 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
8633 after LTO_null.
8634 (lto_tag_is_tree_code_p): Adjust.
8635 (lto_tag_is_gimple_code_p): Likewise.
8636 (lto_gimple_code_to_tag): Likewise.
8637 (lto_tag_to_gimple_code): Likewise.
8638 (lto_tree_code_to_tag): Likewise.
8639 (lto_tag_to_tree_code): Likewise.
8640 * data-streamer.h (streamer_write_hwi_in_range): Use
8641 uhwi streaming to stream the normalized range.
8642 (streamer_read_hwi_in_range): Likewise.
8643
17a7fc37
KT
86442013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8645
8646 * config/arm/arm.md (enabled_for_depr_it): New attribute.
8647 (predicable_short_it): Likewise.
8648 (predicated): Likewise.
8649 (enabled): Handle above.
8650 (define_cond_exec): Set predicated attribute to yes.
8651
b57ca59b
MS
86522013-06-05 Mike Stump <mikestump@comcast.net>
8653
8654 * gdbinit.in (__FUNCTION__): Add.
8655
c1e183a9
UB
86562013-06-05 Uros Bizjak <ubizjak@gmail.com>
8657
8658 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
8659 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
8660
0378bddb
JH
86612013-06-05 Jan Hubicka <jh@suse.cz>
8662
8663 * varasm.c (mark_decl_referenced): Revert the removal until targets
8664 are fixed.
8665
8f8a2057
DE
86662013-06-05 David Edelsohn <dje.gcc@gmail.com>
8667
8668 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
8669 instead of mark_decl_referenced.
8670
edb983b2
JH
86712013-06-05 Jan Hubicka <jh@suse.cz>
8672
8673 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
8674 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
8675 and symtab_used_from_object_file_p.
8676 (cgraph_make_node_local_1): Clear forced_by_abi.
8677 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
8678 * cgraph.h (symtab_node_base): Add forced_by_abi.
8679 (decide_is_variable_needed): Remove.
8680 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
8681 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
8682 (decide_is_symbol_needed): ... this one; handle symbols in general;
8683 always analyze virtuals; honnor forced_by_abi.
8684 (cgraph_finalize_function): Update.
8685 (varpool_finalize_decl): Update.
8686 (symbol_defined_and_needed): Remove.
8687 (analyze_functions): Update.
8688 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
8689 output_refs, input_overwrite_node): Handle forced_by_abi.
8690 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
8691 (address_taken_from_non_vtable_p): ... this one.
8692 (comdat_can_be_unshared_p_1): New function.
8693 (cgraph_comdat_can_be_unshared_p): Rename to ...
8694 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
8695 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
8696 (function_and_variable_visibility): Clear forced_by_abi as needed.
8697 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
8698 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH
8699 * symtab.c (dump_symtab_base): Dump forced_by_abi.
8700 * varpool.c (decide_is_variable_needed): Remove.
8701
9912dbe5
KT
87022013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8703
8704 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
8705 (arm_option_override): Override arm_restrict_it where appropriate.
8706 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
8707 * config/arm/arm.opt (mrestrict-it): New command-line option.
8708 * doc/invoke.texi: Document -mrestrict-it.
8709
34c136b6
DM
87102013-06-05 David Malcolm <dmalcolm@redhat.com>
8711
8712 * tsan.c (tsan_atomic_table): Make const.
8713
a367df53
RB
87142013-06-05 Richard Biener <rguenther@suse.de>
8715
8716 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
8717 index associated with the tree we are supposed to replace.
8718 * tree-streamer-out.c (pack_ts_base_value_fields): Output
8719 TREE_ASM_WRITTEN as zero for everything but SSA names.
8720
70d51a19
DM
87212013-06-05 David Malcolm <dmalcolm@redhat.com>
8722
8723 * tree-ssa-structalias.c (call_stmt_vars): Make static.
8724
b9bd2075
JH
87252013-06-04 Jan Hubicka <jh@suse.cz>
8726
8727 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
8728 (input_node, input_varpool_node): Handle correctly external same
8729 body aliases.
8730 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
8731 nodes at ltrans stage.
8732
aaae719d
JH
87332013-06-04 Jan Hubicka <jh@suse.cz>
8734
8735 * ipa-inline.c (update_caller_keys): Fix availability test.
8736 (update_callee_keys): Likewise.
8737 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
8738 to follow ELF standard.
8739
107eea2c
JU
87402013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
8741
8742 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
8743 (mips64r5900el-*-elf*): New configurations.
8744 * config/mips/mips-cpus.def (r5900): New processor.
8745 * config/mips/mips-tables.opt: Regenerate.
8746 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
8747 (mips_issue_rate): Handle PROCESSOR_R5900.
8748 (mips_reorg_process_insns): Force reorder mode for the R5900.
8749 * config/mips/mips.h (TARGET_MIPS5900): Define.
8750 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
8751 TARGET_MIPS5900.
8752 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
8753 TARGET_MIPS5900.
8754 * config/mips/mips.md (processor): Add r5900.
8755 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
8756
b7342d25
IB
87572013-06-04 Ian Bolton <ian.bolton@arm.com>
8758
8759 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
8760 into function to generate MOVI instruction.
c1e183a9 8761 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB
8762 (aarch64_preferred_simd_mode): Turn into wrapper.
8763 (aarch64_output_scalar_simd_mov_immediate): New function.
8764 * config/aarch64/aarch64-protos.h: Add prototype for above.
8765
81c2dfb9
IB
87662013-06-04 Ian Bolton <ian.bolton@arm.com>
8767
8768 * config/aarch64/aarch64.c (simd_immediate_info): Remove
8769 element_char member.
8770 (sizetochar): Return signed char.
8771 (aarch64_simd_valid_immediate): Remove elchar and other
8772 unnecessary variables.
8773 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
8774 Calculate element_char as required.
8775 * config/aarch64/aarch64-protos.h: Update and move prototype
8776 for aarch64_output_simd_mov_immediate.
8777 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
8778 Update arguments.
8779
48063b9d
IB
87802013-06-04 Ian Bolton <ian.bolton@arm.com>
8781
8782 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
8783 information completed by aarch64_simd_valid_immediate.
8784 (aarch64_legitimate_constant_p): Update arguments.
8785 (aarch64_simd_valid_immediate): Work with struct rather than many
8786 pointers.
8787 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
8788 (aarch64_simd_make_constant): Update arguments.
8789 (aarch64_output_simd_mov_immediate): Work with struct rather than
8790 many pointers. Output immediate directly rather than as operand.
8791 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
8792 Update prototype.
8793 * config/aarch64/constraints.md (Dn): Update arguments.
8794
3ea63f60
IB
87952013-06-04 Ian Bolton <ian.bolton@arm.com>
8796
8797 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
8798 longer static.
8799 (aarch64_simd_immediate_valid_for_move): Remove.
8800 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
8801 (aarch64_simd_make_constant): Update call.
8802 (aarch64_output_simd_mov_immediate): Update call.
8803 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
8804 Add prototype.
8805 * config/aarch64/constraints.md (Dn): Update call.
8806
d8edd899
IB
88072013-06-04 Ian Bolton <ian.bolton@arm.com>
8808
8809 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
8810 return type to bool for prototype.
8811 (aarch64_legitimate_constant_p): Check for true instead of not -1.
8812 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
8813 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
8814
44b20bb8
CM
88152013-06-04 Catherine Moore <clm@codesourcery.com>
8816
8817 * config/mips/mips.opt (meva): New.
c1e183a9 8818 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM
8819 (ASM_SPEC): Handle -meva.
8820 * doc/invoke.texi (meva): Document.
8821
52befbd8
AM
88222013-06-04 Alan Modra <amodra@gmail.com>
8823
8824 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
8825 constant output.
8826
aadaf24e
KT
88272013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8828
8829 * rtl.def: Add extra fourth optional field to define_cond_exec.
8830 * gensupport.c (process_one_cond_exec): Process attributes from
8831 define_cond_exec.
8832 * doc/md.texi: Document fourth field in define_cond_exec.
8833
0cd9e9ee
EB
88342013-06-04 Eric Botcazou <ebotcazou@adacore.com>
8835
8836 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
8837 out the processing order as in store_bit_field_1.
8838
cb0f665d
JH
88392013-06-04 Jan Hubicka <jh@suse.cz>
8840
8841 PR middle-end/57500
8842 * cgraphunit.c (cgraph_process_same_body_aliases): Create
8843 non-VAR_DECL node if it does not exist yet.
8844
53984b9b
RS
88452013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
8846
8847 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
8848 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
8849 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
8850 target_cpu_default setting.
8851
5d318fd4
TJ
88522013-06-03 Teresa Johnson <tejohnson@google.com>
8853
8854 * dumpfile.c (opt_info_switch_p): Change -fopt-info
8855 default to -fopt-info=optimized instead of all.
8856 * doc/invoke.texi: Ditto.
8857 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
8858 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
8859 (execute_vect_slp): Emit BB vectorization success under
8860 MSG_OPTIMIZED_LOCATIONS.
8861 * tree-vect-slp.c (vect_slp_transform_bb): Change
8862 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
8863 * tree-vect-loop.c (vect_transform_loop): Ditto.
8864
ec3af349
JM
88652013-06-03 Jason Merrill <jason@redhat.com>
8866
8867 PR c++/57415
8868 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8869 Use TARGET_EXPR for C++.
8870
1baf9159
JJ
88712013-06-03 Jakub Jelinek <jakub@redhat.com>
8872
8873 PR rtl-optimization/57268
0cd9e9ee 8874 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ
8875 if DEBUG_INSN_P (insn).
8876
8877 Reapply
8878 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8879
8880 PR rtl-optimization/57268
8881 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8882 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8883
0f1d3965
YR
88842013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8885
8886 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
8887 (ix86_avoid_lea_for_addr): Likewise.
8888 (exact_dependency_1): Likewise.
8889 (ix86_adjust_cost): Likewise.
8890 (swap_top_of_ready_list): Fix formatting and !reload_completed check
8891 removed.
8892 (do_reorder_for_imul): Fix typo, formatting and
8893 !reload_completed check removed.
8894 (ix86_sched_reorder): Fix typo and formatting.
8895 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
8896 list.
8897
3bc30db4
SN
88982013-06-03 Sofiane Naci <sofiane.naci@arm.com>
8899
8900 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
8901
19cc6fac
EB
89022013-06-03 Eric Botcazou <ebotcazou@adacore.com>
8903
8904 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
8905 <STRING_CST>: Likewise.
8906 <VECTOR_CST>: Likewise.
8907
c68a6e08
JW
89082013-06-01 Janus Weil <janus@gcc.gnu.org>
8909 Mikael Morin <mikael@gcc.gnu.org>
8910
8911 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
8912 * config.in: Regenerated.
8913 * configure: Regenerated.
8914
38e55e5c
JH
89152013-06-01 Jan Hubicka <jh@suse.cz>
8916
8917 PR middle-end/57366
8918 * cgraphunit.c (compile): When weakref is not supported,
8919 set up transparent aliases before final output pass.
8920 * varasm.c (assemble_alias): Do not try to do it here.
8921
eb51d2ff
JH
89222013-06-01 Jan Hubicka <jh@suse.cz>
8923
8924 PR middle-end/57467
8925 * passes.c (for_per_function): Skip unanalyzed functions.
8926
40a7fe1e
JH
89272013-06-01 Jan Hubicka <jh@suse.cz>
8928
8929 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
8930 (lto_symtab_merge_symbols_1): ... this one.
8931 (lto_symtab_merge_cgraph_nodes): Rename to ...
8932 (lto_symtab_merge_symbols): ... this one; simplify.
8933 * cgraph.c (same_body_aliases_done): Rename to ...
8934 (cpp_implicit_aliases_done): ... this one.
8935 (cgraph_create_function_alias): Update.
8936 (cgraph_same_body_alias): Update.
c1e183a9 8937 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 8938 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 8939 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH
8940 (cgraph_node): Remove same_body_alias.
8941 (varpool_node): Remove alias_of and extra_name_alias.
8942 (same_body_aliases_done): Rename to ..
8943 (cpp_implicit_aliases_done): ... this one.
8944 (symtab_alias_ultimate_target): Add default parameter.
8945 (symtab_resolve_alias): New function.
8946 (fixup_same_cpp_alias_visibility): Declare.
8947 (cgraph_function_node): Add default parameter.
8948 (cgraph_node_asm_name): Likewise.
8949 (cgraph_function_or_thunk_node): Add default parameter; do
8950 not ICE when it is NULL.
8951 (varpool_variable_node): Likewise.
8952 * tree-emutls.c (create_emultls_var): Update.
8953 (ipa_lower_emutls): Update.
8954 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
8955 (cgraph_reset_node): Reset alias info.
8956 (cgraph_finalize_function): Update.
8957 (fixup_same_cpp_alias_visibility): Move to symtab.c.
8958 (analyze_function): Simplify.
8959 (cgraph_process_same_body_aliases): Simplify.
8960 (analyze_functions): Fixup same body aliases.
8961 (handle_alias_pairs): Simplify.
8962 (assemble_thunk): Update.
8963 (assemble_thunks_and_aliases): Update.
8964 (output_weakrefs): Rewrite.
8965 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
8966 (lto_output_varpool_node): Likewise.
8967 (compute_ltrans_boundary): Remve assert.
8968 (get_alias_symbol): New functoin.
8969 (input_node): Rewrite alias handling.
8970 (input_varpool_node): Likewise.
8971 * ipa-pure-const.c (propagate_pure_const): Fix formating.
8972 * ipa.c (process_references): Handle weakrefs correctly.
8973 (symtab_remove_unreachable_nodes): Likewise.
8974 * trans-mem.c (get_cg_data): Update.
8975 (ipa_tm_create_version_alias): Update.
8976 (ipa_tm_execute): Update.
8977 * symtab.c (dump_symtab_base): Dump aliases.
8978 (verify_symtab_base): Verify aliases.
8979 (symtab_node_availability): New function.
8980 (symtab_alias_ultimate_target): Simplify.
8981 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
8982 handle all the fixup cases.
8983 (symtab_resolve_alias): New function.
8984 * passes.c (ipa_write_summaries): Handle weakrefs.
8985 * varpool.c (varpool_analyze_node): Simplify.
8986 (assemble_aliases): Update.
8987 (varpool_create_variable_alias): Simplify.
8988 (varpool_extra_name_alias): Simplify.
8989 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
8990 (lto_symtab_merge_symbols): ... this one.
8991
21759881
DT
89922013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
8993
8994 Revert
8995 PR rtl-optimization/57268
8996 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8997 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8998
daf4e940
TB
89992013-06-01 Tobias Burnus <burnus@net-b.de>
9000
9001 Partially reverted:
9002 2013-05-31 Tobias Burnus <burnus@net-b.de>
9003
9004 PR middle-end/57073
9005 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
9006 further up.
9007
c359d8d0
DT
90082013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
9009
9010 PR rtl-optimization/57268
9011 * sched-deps.c (sched_analyze_2): Flush dependence lists if
9012 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
9013
4172215d
EB
90142013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9015
9016 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
9017 unordered comparison operators when -fno-trapping-math is in effect
9018 on the e500.
9019 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
9020 and implement unordered comparison operators properly on the e500.
9021
a8c50132
EB
90222013-05-31 Eric Botcazou <ebotcazou@adacore.com>
9023
9024 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
9025 for constant scalar integers.
9026 (simplify_relational_operation_1): Likewise.
9027
44626634
SB
90282013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
9029
9030 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
9031 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
9032 Fix comment.
9033
55805e54
YR
90342013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9035 Igor Zamyatin <igor.zamyatin@intel.com>
9036
9037 Silvermont (SLM) architecture performance tuning.
9038 * config/i386/i386.h (enum ix86_tune_indices): Add
9039 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
9040 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
9041
9042 * config/i386/i386.c (initial_ix86_tune_features)
9043 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
9044 (ix86_lea_outperforms): Handle Silvermont tuning.
9045 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
9046 call.
9047 (ix86_use_lea_for_mov): Likewise.
9048 (ix86_avoid_lea_for_addr): Likewise.
9049 (ix86_lea_for_add_ok): Likewise.
9050 (exact_dependency_1): New function.
9051 (exact_store_load_dependency): Likewise.
9052 (ix86_adjust_cost): Handle Silvermont tuning.
9053 (do_reoder_for_imul): Likewise.
9054 (swap_top_of_ready_list): New function.
9055 (ix86_sched_reorder): Changed to handle Silvermont tuning.
9056
9057 * config/i386/i386.md (peepholes that split memory operand in fp
9058 converts): New.
9059
e19c9de2
MS
90602013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9061
9062 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
9063 Remove un-necessary braces.
9064
38e6c9a6
MS
90652013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
9066
9067 * config/aarch64/aarch64.c (aarch64_classify_symbol):
9068 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
9069
e3530904
TB
90702013-05-31 Tobias Burnus <burnus@net-b.de>
9071
9072 PR middle-end/57073
c1e183a9 9073 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 9074
b6af05a9
KT
90752013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9076
9077 PR target/56315
9078 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
9079 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
9080 * config/arm/neon.md (iordi3_neon): Remove.
9081 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
9082 * config/arm/predicates.md (imm_for_neon_logic_operand):
9083 Move to earlier in the file.
9084 (neon_logic_op2): Likewise.
9085 (arm_iordi_operand_neon): New predicate.
9086
f800c166
RB
90872013-05-31 Richard Biener <rguenther@suse.de>
9088
9089 PR tree-optimization/57478
9090 PR tree-optimization/57453
9091 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
9092 are life as well.
9093
e571fa59
KP
90942013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9095
9096 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
9097 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
9098
0fa6e0ef
TB
90992013-05-30 Tobias Burnus <burnus@net-b.de>
9100 Thomas Koenig <tkoenig@gcc.gnu.org>
9101
9102 PR middle-end/57073
9103 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
9104 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
9105
4688ddf5
SB
91062013-05-30 Steven Bosscher <steven@gcc.gnu.org>
9107
0fa6e0ef 9108 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 9109
3b9ceb4b
VM
91102013-05-30 Vladimir Makarov <vmakarov@redhat.com>
9111
9112 * target.def (register_usage_leveling_p): New hook.
9113 * targhooks.c (default_register_usage_leveling_p): New.
9114 * targhooks.h (default_register_usage_leveling_p): New prototype.
9115 * lra-assigns.c (register_usage_leveling_p): Use the hook.
9116 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
9117 * doc/tm.texi: Update.
9118 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
9119
65f9e789
IB
91202013-05-30 Ian Bolton <ian.bolton@arm.com>
9121
9122 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
9123 (*insv_reg<mode>): New define_insn.
9124
12211b99 91252013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR
9126
9127 PR rtl-optimization/57439
9128 * postreload.c (move2add_valid_value_p): Check that we have
9129 a zero subreg_regno_offset when accessing the register in
9130 the requested mode.
9131
0b871ccf
YR
91322013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
9133 Igor Zamyatin <igor.zamyatin@intel.com>
9134
9135 Silvermont (SLM) architecture pipeline model, tuning and
9136 insn selection.
9137 * config.gcc: Add slm config options and target.
9138
9139 * config/i386/slm.md: New.
9140
9141 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
9142
9143 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
9144 PROCESSOR_SLM.
9145 (ix86_target_macros_internal): Likewise.
9146
9147 * gcc/config/i386/i386.c (slm_cost): New cost.
9148 (m_SLM): New macro flag.
9149 (initial_ix86_tune_features): Set m_SLM.
9150 (x86_accumulate_outgoing_args): Likewise.
9151 (x86_arch_always_fancy_math_387): Likewise.
9152 (processor_target_table): Add slm cost.
9153 (cpu_names): Add slm cpu name.
9154 (x86_option_override_internal): Set SLM ISA.
9155 (ix86_issue_rate): New case PROCESSOR_SLM.
9156 (ia32_multipass_dfa_lookahead): Likewise.
9157 (fold_builtin_cpu): Add slm.
9158
9159 * config/i386/i386.h (TARGET_SLM): New target macro.
9160 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
9161 (processor_type): Add PROCESSOR_SLM.
9162
9163 * config/i386/i386.md (cpu): Add new value "slm".
9164 (slm.md): Include slm.md.
9165
24d5b097
XG
91662013-05-30 Bernd Schmidt <bernds@codesourcery.com>
9167 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9168
9169 * config/arm/arm-protos.h: Add and update function protos.
9170 * config/arm/arm.c (use_simple_return_p): New added.
9171 (thumb2_expand_return): Check simple_return flag.
9172 * config/arm/arm.md: Add simple_return and conditional simple_return.
9173 * config/arm/iterators.md: Add iterator for return and simple_return.
9174
c1cccc15
ZC
91752013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9176
9177 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
9178 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
9179 (arm_emit_vfp_multi_reg_pop): Likewise.
9180 (thumb2_emit_ldrd_pop): Likewise.
9181 (arm_expand_epilogue): Add misc REG_CFA notes.
9182 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
9183
3a4f280b
LC
91842013-05-29 Lawrence Crowl <crowl@google.com>
9185
9186 * config/arm/t-arm: Update for below.
9187
9188 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
9189 Change type to hash_table. Update dependent calls and types.
9190
9191 * config/i386/t-cygming: Update for below.
9192
9193 * config/i386/t-interix: Update for below.
9194
9195 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
9196 Change type to hash_table. Update dependent calls and types.
9197 (i386_find_on_wrapper_list::wrappers): Likewise.
9198
9199 * config/ia64/t-ia64: Update for below.
9200
9201 * config/ia64/ia64.c (bundle_state_table):
9202 Change type to hash_table. Update dependent calls and types.
9203
9204 * config/mips/mips.c (mips_reorg_process_insns::htab):
9205 Change type to hash_table. Update dependent calls and types.
9206
9207 * config/sol2.c (solaris_comdat_htab):
9208 Change type to hash_table. Update dependent calls and types.
9209
9210 * config/t-sol2: Update for above.
9211
1388a0e3
TJ
92122013-05-29 Teresa Johnson <tejohnson@google.com>
9213
9214 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
9215 functions are not yet marked as defined.
9216
a5965b52 92172013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
9218 Pat Haugen <pthaugen@us.ibm.com>
9219 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM
9220
9221 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
9222 instructions.
9223 (VEC_A): Likewise.
9224 (VEC_C): Likewise.
9225 (vrotl<mode>3): Likewise.
9226 (vashl<mode>3): Likewise.
9227 (vlshr<mode>3): Likewise.
9228 (vashr<mode>3): Likewise.
9229
9230 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9231 support for power8 V2DI builtins.
9232
9233 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
9234 power8 V2DI builtins.
9235 (vupkhsw): Likewise.
9236 (vupklsw): Likewise.
9237 (vaddudm): Likewise.
9238 (vminsd): Likewise.
9239 (vmaxsd): Likewise.
9240 (vminud): Likewise.
9241 (vmaxud): Likewise.
9242 (vpkudum): Likewise.
9243 (vpksdss): Likewise.
9244 (vpkudus): Likewise.
9245 (vpksdus): Likewise.
9246 (vrld): Likewise.
9247 (vsld): Likewise.
9248 (vsrd): Likewise.
9249 (vsrad): Likewise.
9250 (vsubudm): Likewise.
9251 (vcmpequd): Likewise.
9252 (vcmpgtsd): Likewise.
9253 (vcmpgtud): Likewise.
9254 (vcmpequd_p): Likewise.
9255 (vcmpgtsd_p): Likewise.
9256 (vcmpgtud_p): Likewise.
9257 (vupkhsw): Likewise.
9258 (vupklsw): Likewise.
9259 (vaddudm): Likewise.
9260 (vmaxsd): Likewise.
9261 (vmaxud): Likewise.
9262 (vminsd): Likewise.
9263 (vminud): Likewise.
9264 (vpksdss): Likewise.
9265 (vpksdus): Likewise.
9266 (vpkudum): Likewise.
9267 (vpkudus): Likewise.
9268 (vrld): Likewise.
9269 (vsld): Likewise.
9270 (vsrad): Likewise.
9271 (vsrd): Likewise.
9272 (vsubudm): Likewise.
9273
9274 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
9275 support for power8 V2DI instructions.
9276
9277 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
9278 power8 V2DI instructions. Combine pack and unpack insns to use an
9279 iterator for each mode. Check whether a particular mode supports
9280 Altivec instructions instead of just checking TARGET_ALTIVEC.
9281 (UNSPEC_VPKUWUM): Likewise.
9282 (UNSPEC_VPKSHSS): Likewise.
9283 (UNSPEC_VPKSWSS): Likewise.
9284 (UNSPEC_VPKUHUS): Likewise.
9285 (UNSPEC_VPKSHUS): Likewise.
9286 (UNSPEC_VPKUWUS): Likewise.
9287 (UNSPEC_VPKSWUS): Likewise.
9288 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
9289 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
9290 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
9291 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
9292 (UNSPEC_VUPKHSB): Likewise.
9293 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
9294 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
9295 (UNSPEC_VUPKHSH): Likewise.
9296 (UNSPEC_VUPKLSB): Likewise.
9297 (UNSPEC_VUPKLSH): Likewise.
9298 (VI2): Likewise.
9299 (VI_char): Likewise.
9300 (VI_scalar): Likewise.
9301 (VI_unit): Likewise.
9302 (VP): Likewise.
9303 (VP_small): Likewise.
9304 (VP_small_lc): Likewise.
9305 (VU_char): Likewise.
9306 (add<mode>3): Likewise.
9307 (altivec_vaddcuw): Likewise.
9308 (altivec_vaddu<VI_char>s): Likewise.
9309 (altivec_vadds<VI_char>s): Likewise.
9310 (sub<mode>3): Likewise.
9311 (altivec_vsubcuw): Likewise.
9312 (altivec_vsubu<VI_char>s): Likewise.
9313 (altivec_vsubs<VI_char>s): Likewise.
9314 (altivec_vavgs<VI_char>): Likewise.
9315 (altivec_vcmpbfp): Likewise.
9316 (altivec_eq<mode>): Likewise.
9317 (altivec_gt<mode>): Likewise.
9318 (altivec_gtu<mode>): Likewise.
9319 (umax<mode>3): Likewise.
9320 (smax<mode>3): Likewise.
9321 (umin<mode>3): Likewise.
9322 (smin<mode>3): Likewise.
9323 (altivec_vpkuhum): Likewise.
9324 (altivec_vpkuwum): Likewise.
9325 (altivec_vpkshss): Likewise.
9326 (altivec_vpkswss): Likewise.
9327 (altivec_vpkuhus): Likewise.
9328 (altivec_vpkshus): Likewise.
9329 (altivec_vpkuwus): Likewise.
9330 (altivec_vpkswus): Likewise.
9331 (altivec_vpks<VI_char>ss): Likewise.
9332 (altivec_vpks<VI_char>us): Likewise.
9333 (altivec_vpku<VI_char>us): Likewise.
9334 (altivec_vpku<VI_char>um): Likewise.
9335 (altivec_vrl<VI_char>): Likewise.
9336 (altivec_vsl<VI_char>): Likewise.
9337 (altivec_vsr<VI_char>): Likewise.
9338 (altivec_vsra<VI_char>): Likewise.
9339 (altivec_vsldoi_<mode>): Likewise.
9340 (altivec_vupkhsb): Likewise.
9341 (altivec_vupkhs<VU_char>): Likewise.
9342 (altivec_vupkls<VU_char>): Likewise.
9343 (altivec_vupkhsh): Likewise.
9344 (altivec_vupklsb): Likewise.
9345 (altivec_vupklsh): Likewise.
9346 (altivec_vcmpequ<VI_char>_p): Likewise.
9347 (altivec_vcmpgts<VI_char>_p): Likewise.
9348 (altivec_vcmpgtu<VI_char>_p): Likewise.
9349 (abs<mode>2): Likewise.
9350 (vec_unpacks_hi_v16qi): Likewise.
9351 (vec_unpacks_hi_v8hi): Likewise.
9352 (vec_unpacks_lo_v16qi): Likewise.
9353 (vec_unpacks_hi_<VP_small_lc>): Likewise.
9354 (vec_unpacks_lo_v8hi): Likewise.
9355 (vec_unpacks_lo_<VP_small_lc>): Likewise.
9356 (vec_pack_trunc_v8h): Likewise.
9357 (vec_pack_trunc_v4si): Likewise.
9358 (vec_pack_trunc_<mode>): Likewise.
9359
9360 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
9361 V2DI builtins.
9362 (vec_vmaxsd): Likewise.
9363 (vec_vmaxud): Likewise.
9364 (vec_vminsd): Likewise.
9365 (vec_vminud): Likewise.
9366 (vec_vpksdss): Likewise.
9367 (vec_vpksdus): Likewise.
9368 (vec_vpkudum): Likewise.
9369 (vec_vpkudus): Likewise.
9370 (vec_vrld): Likewise.
9371 (vec_vsld): Likewise.
9372 (vec_vsrad): Likewise.
9373 (vec_vsrd): Likewise.
9374 (vec_vsubudm): Likewise.
9375 (vec_vupkhsw): Likewise.
9376 (vec_vupklsw): Likewise.
9377
e70670cf
JH
93782013-05-29 Jan Hubicka <jh@suse.cz>
9379
9380 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
9381 flags; reorder rest of fields in more consistent way.
9382 (varpool_node): Remove analyzed, finalized and alias.
9383 (cgraph_ndoe): Likewise.
9384 (symtab_alias_ultimate_target): New function.
9385 (cgraph_function_node): Move offline.
9386 (cgraph_reset_node): Declare.
9387 (cgraph_comdat_can_be_unshared_p): Remove.
9388 (varpool_remove_initializer): Declare.
9389 (varpool_first_defined_variable, varpool_next_defined_variable
9390 cgraph_first_defined_function, cgraph_next_defined_function): Update.
9391 (cgraph_function_with_gimple_body_p): Update.
9392 (varpool_all_refs_explicit_p): Update.
9393 (symtab_alias_target): New function.
9394 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
9395 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB
9396 (cgraph_function_or_thunk_node): Simplify using
9397 symtab_alias_ultimate_target.
e70670cf
JH
9398 (varpool_variable_node): Likewise.
9399 * cgraph.c (cgraph_create_function_alias): Update.
9400 (cgraph_add_thunk): Update.
9401 (cgraph_remove_node): Update.
9402 (dump_cgraph_node): Do not dump removed flags.
9403 (cgraph_function_body_availability): Update.
9404 (cgraph_propagate_frequency): Update.
9405 (verify_cgraph_node): Check sanity of local flag.
9406 (cgraph_function_node): Move here from cgraph.h; revamp for
9407 cgraph_function_or_thunk_node.
9408 * lto-symtab.c (lto_varpool_replace_node): Update.
9409 (lto_symtab_resolve_can_prevail_p): Update.
9410 (lto_symtab_merge_cgraph_nodes): Update.
9411 * ipa-cp.c (determine_versionability, initialize_node_lattices,
9412 propagate_constants_accross_call, devirtualization_time_bonus,
9413 ipcp_propagate_stage): Update.
9414 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB
9415 * ipa-inline-transform.c (clone_inlined_nodes,
9416 preserve_function_body_p): Update.
e70670cf
JH
9417 * ipa-reference.c (propagate): Update.
9418 (write_node_summary_p): Update.
9419 * toplev.c (wrapup_global_declaration_2): Update.
9420 * cgraphunit.c (cgraph_analyze_function): Rename to ...
9421 (analyze_function) ... this one.
9422 (cgraph_process_new_functions): Update.
9423 (cgraph_reset_node): Export.
9424 (cgraph_finalize_function): Update.
9425 (cgraph_add_new_function): Update.
9426 (process_function_and_variable_attributes): Update.
9427 (varpool_finalize_decl): Update.
9428 (symbol_finalized): Remove.
9429 (symbol_finalized_and_needed): Rename to ...
9430 (symbol_defined_and_needed): ... update.
9431 (cgraph_analyze_functions): Update.
9432 (handle_alias_pairs): Update.
9433 (mark_functions_to_output): Update.
9434 (assemble_thunk): Update.
9435 (output_in_order): Update.
9436 (output_weakrefs): Update.
9437 (finalize_compilation_unit): Update.
9438 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
9439 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
9440 input_node, input_varpool_node): Update.
9441 * dbxout.c (dbxout_expand_expr): Update.
9442 * cgraphclones.c (cgraph_clone_node): Update.
9443 (cgraph_copy_node_for_versioning): Update.
9444 (cgraph_materialize_clone): Update.
9445 (cgraph_materialize_all_clones): Update.
9446 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
9447 propagate_pure_const, propagate_nothrow): Update.
9448 * lto-streamer-out.c (lto_output, write_symbol): Update.
9449 * ipa-utils.c (ipa_reverse_postorder): Update.
9450 * ipa-inline.c (can_inline_edge_p): Update.
9451 (update_caller_keys, ipa_inline): Update.
9452 * dwarf2out.c (reference_to_unused,
9453 premark_types_used_by_global_vars_helper): Update.
9454 * tree-eh.c (tree_could_trap_p): Update.
9455 * ipa-split.c (consider_split, execute_split_functions): Update.
9456 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB
9457 has_addr_references_p): Update; move ahead in file for better
9458 readability.
e70670cf
JH
9459 (process_references): Simplify.
9460 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
9461 bodies are removed.
9462 (cgraph_comdat_can_be_unshared_p): Make static.
9463 (cgraph_externally_visible_p): Update.
9464 (varpool_externally_visible_p): Update.
9465 (function_and_variable_visibility): Update.
9466 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
9467 ipa_tm_mark_force_output_node): Update.
9468 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
9469 estimate_edge_devirt_benefit, inline_generate_summary,
9470 inline_write_summary): Update.
9471 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
9472 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB
9473 (ipa_print_node_params, ipa_prop_read_section,
9474 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH
9475 * varasm.c (mark_decl_referenced): Update.
9476 (assemble_alias, dump_tm_clone_pairs): Update.
9477 * tree-inline.c (copy_bb): Update.
9478 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
9479 Update.
9480 * symtab.c (dump_symtab_base): Print new flags.
9481 (verify_symtab_base): Verify new flags.
9482 (symtab_alias_ultimate_target): New function.
9483 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
9484 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
9485 Update.
c1e183a9
UB
9486 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
9487 Update.
e70670cf
JH
9488 * i386.c (ix86_get_function_versions_dispatcher,
9489 ix86_generate_version_dispatcher_body): Update.
9490 (fold_builtin_cpu): Use varpool_add_new_variable.
9491 * varpool.c (varpool_remove_initializer): Break out from ...
9492 (varpool_remove_node): ... this one.
9493 (dump_varpool_node, varpool_node_for_asm,
9494 cgraph_variable_initializer_availability, varpool_analyze_node,
9495 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB
9496 varpool_finalize_named_section_flags, varpool_create_variable_alias):
9497 Update.
e70670cf 9498
182802ad
JH
94992013-05-29 Jan Hubicka <jh@suse.cz>
9500
9501 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
9502
8190b609
ER
95032013-05-29 Easwaran Raman <eraman@google.com>
9504
9505 PR tree-optimization/57442
9506 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
9507 when control exits the main loop.
9508
69f5aa9b
SKS
95092013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
9510
9511 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
9512 and RX600.
c1e183a9 9513 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS
9514 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
9515 * rx/t-rx: Add rx100 under multi library matches option for nofpu
9516 option.
9517
4b847da9
BS
95182013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9519
9520 PR tree-optimization/57441
9521 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
9522 Don't limit size of incr_vec to number of candidates.
9523
4fc43c39
SE
95242013-05-29 Steve Ellcey <sellcey@imgtec.com>
9525
9526 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
9527 and mips16 directories.
c1e183a9 9528 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE
9529 (MULTILIB_DIRNAMES): Ditto.
9530 (MULTILIB_EXCEPTIONS): Add new exceptions.
9531 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
9532 (MULTILIB_DIRNAMES): Ditto.
9533 (MULTILIB_EXCEPTIONS): Add new exceptions.
9534
12211b99 95352012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS
9536 Marcus Shawcroft <marcus.shawcroft@arm.com>
9537
9538 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
9539 SYMBOL_TINY_ABSOLUTE.
9540 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
9541 SYMBOL_TINY_ABSOLUTE.
9542 (aarch64_expand_mov_immediate): Likewise.
9543 (aarch64_classify_symbol): Likewise.
9544 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
9545 Permit SYMBOL_TINY_ABSOLUTE.
9546 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
9547
12211b99 95482013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS
9549 Marcus Shawcroft <marcus.shawcroft@arm.com>
9550
9551 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
9552 Refactor if/switch. Replace gcc_assert with if.
9553
c0186656
GG
95542013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9555
9556 * config/i386/i386.c (initial_ix86_tune_features): Enable
9557 FP Reassociation for AMD bdver1 and bdver2.
9558
d20188f3
MJ
95592013-05-29 Martin Jambor <mjambor@suse.cz>
9560
9561 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
9562 and IMAGPART_EXPR do not occur within other handled_components.
9563
292cba13
RB
95642013-05-29 Richard Biener <rguenther@suse.de>
9565
9566 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
9567 access on whether the use is in the BB we currently try to
9568 vectorize.
9569 (vect_bb_vectorization_profitable_p): Pass the BB we currently
9570 vectorize to vect_bb_slp_scalar_cost.
9571
6eddf228
RB
95722013-05-29 Richard Biener <rguenther@suse.de>
9573
9574 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
9575 computing scalar cost offsetted by stmts that are kept live
9576 by scalar uses.
9577 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
9578 for computation of scalar cost.
9579
7df36117
SE
95802013-05-28 Steve Ellcey <sellcey@mips.com>
9581
9582 * config/mips/mips-cpus.def (mips32r2): Change processor type.
9583
36536d79
BI
95842013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9585
9586 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
9587 array notation built-in reduction functions.
9588 * doc/passes.texi (Passes): Added documentation about changes done
9589 for Cilk Plus.
9590 * doc/invoke.texi (C Dialect Options): Added documentation about
9591 the -fcilkplus flag.
9592 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
9593 (BUILTINS_DEF): Depend on cilkplus.def.
9594 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
9595 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
9596 * cilkplus.def: New file.
9597
12211b99 95982013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR
9599
9600 PR rtl-optimization/57439
9601 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
9602
9daf6dbb
ER
96032013-05-28 Easwaran Raman <eraman@google.com>
9604
9605 PR tree-optimization/57337
9606 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
9607 (find_insert_point): Correctly identify the insertion point
9608 when two statements with the same UID is compared.
9609
fbd7e877
RB
96102013-05-28 Richard Biener <rguenther@suse.de>
9611
9612 PR tree-optimization/56787
9613 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
9614 from the list of data references.
9615 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
9616 clobbers.
9617 (vect_analyze_loop_operations): Likewise.
9618 (vect_transform_loop): Remove clobbers.
9619
bbba1117
MJ
96202013-05-28 Martin Jambor <mjambor@suse.cz>
9621
9622 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
9623 and REALPART_EXPRs have scalar type.
9624
bd388c2a
RB
96252013-05-28 Richard Biener <rguenther@suse.de>
9626
9627 PR tree-optimization/57411
9628 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
9629 virtual operands.
9630 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
9631 virtual operand propagation.
9632
2f56a311
EB
96332013-05-28 Eric Botcazou <ebotcazou@adacore.com>
9634
9635 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
9636 destination register for bmasksi_vis.
9637 (vector_init_bshuffle): Likewise.
9638 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
9639
5c3eacbb
EB
96402013-05-28 Eric Botcazou <ebotcazou@adacore.com>
9641
9642 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
9643 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
9644 mode if the instruction isn't available in the original mode.
9645 * config/sparc/sparc.opt (mfix-ut699): New option.
9646 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
9647 (divdf3): Turn into expander.
9648 (divdf3_nofix): New insn.
9649 (divdf3_fix): Likewise.
9650 (divsf3): Disable if -mfix-ut699.
9651 (sqrtdf2): Turn into expander.
9652 (sqrtdf2_nofix): New insn.
9653 (sqrtdf2_fix): Likewise.
9654 (sqrtsf2): Disable if -mfix-ut699.
9655
a1756c0a
RB
96562013-05-27 Richard Biener <rguenther@suse.de>
9657
9658 PR middle-end/57412
9659 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
9660 block for the new loop.
9661
5a892248
RB
96622013-05-27 Richard Biener <rguenther@suse.de>
9663
9664 PR tree-optimization/57343
9665 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
9666 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
9667 (number_of_iterations_cond): Do not build the folded tree.
9668
d1de852b
RB
96692013-05-27 Richard Biener <rguenther@suse.de>
9670
9671 Revert
9672 PR middle-end/57381
9673 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
9674 OEP_CONSTANT_ADDRESS_OF retained.
9675
9676 PR tree-optimization/57417
9677 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
9678 for unchanged base.
9679 (set_ssa_val_to): Compare addresses using
9680 get_addr_base_and_unit_offset.
9681
12211b99 96822013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR
9683
9684 PR rtl-optimization/56833
9685 * postreload.c (move2add_record_mode): New function.
9686 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
9687 (move2add_use_add2_insn): Use move2add_record_sym_value.
9688 (move2add_use_add3_insn): Likewise.
9689 (reload_cse_move2add): Use move2add_valid_value_p and
9690 move2add_record_mode. Invalidate call-clobbered and REG_INC
9691 affected regs by setting reg_mode to VOIDmode.
9692 (move2add_note_store): Don't pretend the inside of a SUBREG is
9693 the actual destination. Invalidate single/leading registers by
9694 setting reg_mode to VOIDmode.
9695 Use move2add_record_sym_value, move2add_valid_value_p and
9696 move2add_record_mode.
9697
b03be25f
RB
96982013-05-27 Richard Biener <rguenther@suse.de>
9699
9700 PR tree-optimization/57396
9701 * tree-affine.c (double_int_constant_multiple_p): Properly
9702 return false for val == 0 and div != 0.
9703
44e88db2
RS
97042013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9705
9706 * config/mips/mips.h: Use #elif in preprocessor conditions.
9707
3b859704
RS
97082013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9709
9710 PR target/53916
9711 * config/mips/constraints.md (kl): New constraint.
9712 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
9713 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
9714 constraint for operand 0. Split after CSE for MIPS16. Emit a move
9715 from LO for MIPS16.
9716 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
9717
c979d5f5
RS
97182013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
9719
9720 PR target/55777
9721 * config/mips/mips.c (mips_can_inline_p): New function.
9722 (TARGET_CAN_INLINE_P): Define.
9723
8e90de43
SB
97242013-05-25 Steven Bosscher <steven@gcc.gnu.org>
9725
9726 * sched-int.h (ds_t, dw_t): Make unsigned int.
9727 Fix documentation that describes how all the ds_t bits are used.
9728 Reserve the last bit for delayed-branch scheduling.
9729 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
9730 (BITS_PER_DEP_WEAK): Fix definition and documentation.
9731 (gen_dep_weak_1): Remove prototype.
9732 * sched-deps.c (get_dep_weak_1): Make static.
9733 * target.def (speculate_insn, needs_block_p, gen_spec_check,
9734 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
9735 * doc/tm.texi: Regenerate.
9736 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
9737
cb5cb194
SB
97382013-05-24 Steven Bosscher <steven@gcc.gnu.org>
9739
9740 PR debug/56950
9741 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
9742
5bd93ff6
NS
97432013-05-24 Nathan Sidwell <nathan@codesourcery.com>
9744 Sandra Loosemore <sandra@codesourcery.com>
9745
9746 * config.gcc (powerpc-*): Allow native for with-cpu.
9747
2343af65
JL
97482013-05-24 Jeff Law <law@redhat.com>
9749
9750 PR tree-optimization/57124
9751 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
9752 conversion feeding a condition if the range has an overflow
9753 if -fstrict-overflow. Add warnings for when we do make the
9754 transformation.
9755
3ad695b9
DC
97562013-05-24 Dehao Chen <dehao@google.com>
9757
04960246 9758 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC
9759 (locus_discrim_hasher): Likewise.
9760 (locus_discrim_hasher::hash): Likewise.
9761 (locus_discrim_hasher::equal): Likewise.
9762
82338059
MJ
97632013-05-24 Martin Jambor <mjambor@suse.cz>
9764
9765 PR tree-optimization/57294
9766 * cgraph.h (ipa_record_stmt_references): Declare.
9767 * cgraphbuild.c (ipa_record_stmt_references): New function.
9768 (build_cgraph_edges): Use ipa_record_stmt_references.
9769 (rebuild_cgraph_edges): Likewise.
9770 (cgraph_rebuild_references): Likewise.
9771 * ipa-prop.c (ipa_modify_call_arguments): Discard references
9772 associated with the old statement and build references from the
9773 newly built statements.
9774 * ipa-ref.c (ipa_remove_stmt_references): New function.
9775 * ipa-ref.h (ipa_remove_stmt_references): Declare.
9776
1ccd4874
VM
97772013-05-24 Vladimir Makarov <vmakarov@redhat.com>
9778
55805e54 9779 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 9780 mem-mem moves.
55805e54 9781 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 9782 too.
55805e54 9783 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM
9784 secondary memory moves.
9785 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
9786 reg set up in the current insn.
9787
25e25c73
DC
97882013-05-24 Dehao Chen <dehao@google.com>
9789
04960246 9790 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC
9791 hash function.
9792 (locus_descrim_hasher::equal): Likewise.
04960246 9793 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC
9794 (make_edges): Likewise.
9795 (next_discriminator_for_locus): Likewise.
9796 (same_line_p): Likewise.
9797 (assign_discriminators): Likewise.
9798 (make_cond_expr_edges): Likewise.
9799 (make_gimple_switch_edges): Likewise.
9800 (make_goto_expr_edges): Likewise.
9801 (make_gimple_asm_edges): Likewise.
9802
50d38551
IB
98032013-05-24 Ian Bolton <ian.bolton@arm.com>
9804
9805 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
9806 X format specifier to only display bottom 16 bits.
9807 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
9808 immediate to match for operand 2, since it will be masked.
9809
aea0101d
RB
98102013-05-24 Richard Biener <rguenther@suse.de>
9811
9812 PR tree-optimization/57287
9813 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
9814 all SSA names that occur in abnormal PHIs.
9815
634e03d3
AI
98162013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
9817
9818 PR tree-ssa/57385
9819 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
9820 that index is not negative.
9821
b17c024f
EB
98222013-05-24 Eric Botcazou <ebotcazou@adacore.com>
9823
9824 PR rtl-optimization/55177
9825 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
9826 (simplify_byte_swapping_operation): New.
9827 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
9828 (simplify_relational_operation_1): Deal with BSWAP.
9829
46aeac1b
RH
98302013-05-23 Richard Henderson <rth@redhat.com>
9831
9832 PR target/56742
9833 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
9834 (ix86_reorg): Call it.
9835
70cc1536
UB
98362013-05-23 Uros Bizjak <ubizjak@gmail.com>
9837
9838 PR target/57379
9839 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
9840 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
9841 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
9842
b3851501
CB
98432013-05-23 Christian Bruel <christian.bruel@st.com>
9844
9845 PR debug/57351
9846 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
9847
12211b99 98482013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS
9849 Marcus Shawcroft <marcus.shawcroft@arm.com>
9850
9851 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
9852 * config/aarch64/constraints.md (Usa): Remove.
9853 * doc/md.texi (AArch64 Usa): Remove.
9854
12211b99 98552013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS
9856 Marcus Shawcroft <marcus.shawcroft@arm.com>
9857
9858 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
9859 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
9860 * config/aarch64/predicates.md (aarch64_const_address): Remove.
9861 (aarch64_mov_operand): Use aarch64_mov_operand_p.
9862
12211b99 98632013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP
9864
9865 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
9866 instruction (AdvSIMD).
9867 * config/aarch64/aarch64-builtins.c
9868 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
9869 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
9870
76c36cb1 98712013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ
9872
9873 PR middle-end/57347
9874 * tree.h (contains_bitfld_component_ref_p): Declare.
9875 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB
9876 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
9877 caller.
7d2fb524
MJ
9878 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
9879 not access a bit-field. Assert all final offsets are byte-aligned.
9880
ce521ae6
RB
98812013-05-23 Richard Biener <rguenther@suse.de>
9882
9883 PR tree-optimization/57380
9884 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
9885 least one invariant or re-used load.
9886 * passes.c (init_optimization_passes): Move pass_phiprop before
9887 pass_forwprop.
9888
75c7257f
JG
98892013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
9890
9891 * config/aarch64/aarch64-simd.md
9892 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
9893
af6d44b5
RB
98942013-05-23 Richard Biener <rguenther@suse.de>
9895
9896 PR middle-end/57381
9897 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
9898 OEP_CONSTANT_ADDRESS_OF retained.
9899
bd3647bf
JJ
99002013-05-23 Jakub Jelinek <jakub@redhat.com>
9901
9902 PR middle-end/57344
70cc1536
UB
9903 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
9904 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 9905
a55757ea
RB
99062013-05-23 Richard Biener <rguenther@suse.de>
9907
9908 PR rtl-optimization/57341
9909 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
9910 instead of true_dependence.
9911
cfb00b41
DM
99122013-05-22 David Malcolm <dmalcolm@redhat.com>
9913
dd1c676f
DM
9914 * bb-reorder.c (branch_threshold): Make const.
9915 (exec_threshold): Ditto.
cfb00b41 9916
f62511da 99172013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
9918 Pat Haugen <pthaugen@us.ibm.com>
9919 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
9920
9921 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
9922 documentation for the power8 crypto builtins.
9923
9924 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
9925
9926 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
9927 macros for defining power8 builtin functions.
9928 (BU_P8V_AV_2): Likewise.
9929 (BU_P8V_AV_P): Likewise.
9930 (BU_P8V_VSX_1): Likewise.
9931 (BU_P8V_OVERLOAD_1): Likewise.
9932 (BU_P8V_OVERLOAD_2): Likewise.
9933 (BU_CRYPTO_1): Likewise.
9934 (BU_CRYPTO_2): Likewise.
9935 (BU_CRYPTO_3): Likewise.
9936 (BU_CRYPTO_OVERLOAD_1): Likewise.
9937 (BU_CRYPTO_OVERLOAD_2): Likewise.
9938 (XSCVSPDP): Fix typo, point to the correct instruction.
9939 (VCIPHER): Add power8 crypto builtins.
9940 (VCIPHERLAST): Likewise.
9941 (VNCIPHER): Likewise.
9942 (VNCIPHERLAST): Likewise.
9943 (VPMSUMB): Likewise.
9944 (VPMSUMH): Likewise.
9945 (VPMSUMW): Likewise.
9946 (VPERMXOR_V2DI): Likewise.
9947 (VPERMXOR_V4SI: Likewise.
9948 (VPERMXOR_V8HI: Likewise.
9949 (VPERMXOR_V16QI: Likewise.
9950 (VSHASIGMAW): Likewise.
9951 (VSHASIGMAD): Likewise.
9952 (VPMSUM): Likewise.
9953 (VPERMXOR): Likewise.
9954 (VSHASIGMA): Likewise.
9955
9956 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9957 __CRYPTO__ if the crypto instructions are available.
9958 (altivec_overloaded_builtins): Add support for overloaded power8
9959 builtins.
9960
9961 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
9962 support for power8 crypto builtins.
9963 (builtin_function_type): Likewise.
9964 (altivec_init_builtins): Add support for builtins that take vector
9965 long long (V2DI) arguments.
9966
9967 * config/rs6000/crypto.md: New file, define power8 crypto
9968 instructions.
9969
99702013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
9971 Pat Haugen <pthaugen@us.ibm.com>
9972 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
9973
9974 * doc/invoke.texi (Option Summary): Add power8 options.
9975 (RS/6000 and PowerPC Options): Likewise.
9976
9977 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
9978 constraints.md instead of rs6000.h. Reorder w* constraints. Add
9979 wm, wn, wr documentation.
9980
9981 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
9982 registers if direct move instructions are enabled.
9983 (wn): New constraint for no registers.
9984 (wq): New constraint for quad word even GPR registers.
9985 (wr): New constraint if 64-bit instructions are enabled.
9986 (wv): New constraint if power8 vector instructions are enabled.
9987 (wQ): New constraint for quad word memory locations.
9988
9989 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
9990 constraint for 0..15 for crypto instructions.
9991 (gpc_reg_operand): If VSX allow registers in VSX registers as well
9992 as GPR and floating point registers.
9993 (int_reg_operand): New predicate to match only GPR registers.
9994 (base_reg_operand): New predicate to match base registers.
9995 (quad_int_reg_operand): New predicate to match even GPR registers
9996 for quad memory operations.
9997 (vsx_reg_or_cint_operand): New predicate to allow vector logical
9998 operations in both GPR and VSX registers.
9999 (quad_memory_operand): New predicate for quad memory operations.
10000 (reg_or_indexed_operand): New predicate for direct move support.
10001
10002 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
10003 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
10004 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
10005 (POWERPC_MASKS): Add power8 options.
10006 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
10007 various options.
10008
10009 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10010 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
10011
10012 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
10013 (-mpower8-fusion): New power8 options.
10014 (-mpower8-fusion-sign): Likewise.
10015 (-mpower8-vector): Likewise.
10016 (-mcrypto): Likewise.
10017 (-mdirect-move): Likewise.
10018 (-mquad-memory): Likewise.
10019
10020 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
10021 power8.
10022 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
10023 registers.
70cc1536 10024 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM
10025 (rs6000_debug_vector_unit): Add p8_vector.
10026 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
10027 definitions. Also print fusion state.
10028 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
10029 (rs6000_builtin_mask_calculate): Add power8 builtin support.
10030 (rs6000_option_override_internal): Add support for power8.
10031 (rs6000_common_init_builtins): Add debugging for skipped builtins
10032 if -mdebug=builtin.
10033 (rs6000_adjust_cost): Add power8 support.
10034 (rs6000_issue_rate): Likewise.
10035 (insn_must_be_first_in_group): Likewise.
10036 (insn_must_be_last_in_group): Likewise.
10037 (force_new_group): Likewise.
10038 (rs6000_register_move_cost): Likewise.
10039 (rs6000_opt_masks): Likewise.
10040
10041 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
10042 power8 capable assembler, default to power7 options.
10043 (TARGET_DIRECT_MOVE): Likewise.
10044 (TARGET_CRYPTO): Likewise.
10045 (TARGET_P8_VECTOR): Likewise.
10046 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
10047 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
10048 (VECTOR_MEM_P8_VECTOR_P): Likewise.
10049 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
10050 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
10051 (TARGET_XSCVDPSPN): Likewise.
10052 (TARGET_XSCVSPDPN): Likewsie.
10053 (TARGET_SYNC_HI_QI): Likewise.
10054 (TARGET_SYNC_TI): Likewise.
10055 (MASK_CRYPTO): Likewise.
10056 (MASK_DIRECT_MOVE): Likewise.
10057 (MASK_P8_FUSION): Likewise.
10058 (MASK_P8_VECTOR): Likewise.
70cc1536
UB
10059 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
10060 temporary used by some of the direct move instructions to get two FP
10061 temporary registers does not force creation of a stack frame.
f62511da
MM
10062 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
10063 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
10064 that any VSX registers are tieable, even if they are also an
10065 Altivec vector mode.
10066 (r6000_reg_class_enum): Add wm, wr, wv constraints.
10067 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
10068 (RS6000_BTM_CRYPTO): Likewise.
10069 (RS6000_BTM_COMMON): Likewise.
10070
10071 * config/rs6000/rs6000.md (cpu attribute): Add power8.
10072 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
10073 (enum rs6000_vector): Add power8 vector support.
10074
73a1a707
RR
100752013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10076
10077 PR target/19599
10078 PR target/57340
10079 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
10080 (any_sibcall_could_use_r3): this and handle indirect calls.
10081 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
10082
d28073d4
BS
100832013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10084
10085 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
10086
bee0b10c
RB
100872013-05-22 Richard Biener <rguenther@suse.de>
10088
10089 PR middle-end/57349
10090 * profile.c (branch_prob): Do not split blocks that are
10091 abnormally receiving from ECF_RETURNS_TWICE functions.
10092
98409b51
RS
100932013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10094
10095 * recog.c (offsettable_address_addr_space_p): Fix calculation of
10096 address mode. Move pointer mode initialization to the same place.
10097
c0602ab8
MZ
100982013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
10099
10100 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
10101 while it has any effect.
10102
4f7a634e
ER
101032013-05-21 Easwaran Raman <eraman@google.com>
10104
10105 PR tree-optimization/57322
9daf6dbb
ER
10106 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
10107 UID of the statement added to the BB to be 1.
4f7a634e 10108
a4ce1258
JJ
101092013-05-21 Jakub Jelinek <jakub@redhat.com>
10110
10111 PR tree-optimization/57331
70cc1536
UB
10112 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
10113 of conversion from pointer type to integral type with integer.
a4ce1258 10114
1b14621a
MJ
101152013-05-21 Martin Jambor <mjambor@suse.cz>
10116
10117 PR lto/57289
10118 * ipa-prop.c (ipa_read_node_info): Process param_used and
10119 controlled_uses in the same order as when writing.
10120
e60661f0
MG
101212013-05-21 Magnus Granberg <baldrick@free.fr>
10122
10123 PR plugins/56754
ee49aa34 10124 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 10125
4a61be9a
RB
101262013-05-21 Richard Biener <rguenther@suse.de>
10127
10128 PR tree-optimization/57318
10129 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
10130 estimate stmts with side-effects as likely eliminated.
10131
c52da5f7
RB
101322013-05-21 Richard Biener <rguenther@suse.de>
10133
10134 PR tree-optimization/57330
10135 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
10136 preserve the call stmts fntype.
10137
7ec67e2a
RB
101382013-05-21 Richard Biener <rguenther@suse.de>
10139
10140 PR tree-optimization/57303
10141 * tree-ssa-sink.c (statement_sink_location): Improve killing
10142 stmt detection and properly handle self-assignments.
10143
b112d513
CB
101442013-05-21 Christian Bruel <christian.bruel@st.com>
10145
55805e54
YR
10146 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
10147 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB
10148 registers. Set register size out of the PARALLEL loop.
10149
14c2ec26
OE
101502013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
10151
10152 PR target/56547
10153 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
10154 (*fmasf4, *fmasf4_media): New insns.
10155
da734fa1
RS
101562013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10157
10158 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
10159 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
10160 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
10161 (mips_idiv_insns): Update the comments to say that the returned
10162 instruction counts are in units of BASE_INSN_LENGTH.
10163 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
10164 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
10165 using 2 rather than 4 as the length of indirect MIPS16 and
10166 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
10167 length of a NOP. Don't divide MIPS16 lengths by 2.
10168 (mips16_split_long_branches): Assume a branch is long if the
10169 length is greater than 4 rather than 8.
10170 * config/mips/mips.md (length): Give MIPS16 lengths directly,
10171 rather than multiplying them by 2. Multiply instruction counts
10172 by BASE_INSN_LENGTH rather than 4.
10173 (*jump_mips16, tls_get_tp_mips16_<mode>)
10174 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
10175
13719e8b
RS
101762013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10177
10178 * config/mips/mips.md (extended_mips16): Remove branch case.
10179 (length): Remove duplicated extended_mips16 test.
10180
c3850d14
RS
101812013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
10182
10183 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
10184
8da2e059
RS
101852013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
10186
10187 * recog.h (Recog_data): Rename to...
10188 (recog_data_d): ...this.
10189 (recog_data): Update accordingly.
10190 * recog.c (recog_data): Likewise.
10191 * reload.c (save_recog_data): Likewise.
10192 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
10193 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
10194
2df013f3
JB
101952013-05-17 Julian Brown <julian@codesourcery.com>
10196
10197 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
10198 found in a REG_EQUAL note, invalidate it.
10199
371e77e3 102002013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER
10201
10202 * tree-ssa-reassoc.c (find_insert_point): New function.
10203 (insert_stmt_after): Likewise.
10204 (get_def_stmt): Likewise.
10205 (ensure_ops_are_available): Likewise.
10206 (not_dominated_by): Likewise.
10207 (rewrite_expr_tree): Do not move statements beyond what is
10208 necessary. Remove call to swap_ops_for_binary_stmt...
10209 (reassociate_bb): ... and move it here.
10210 (build_and_add_sum): Assign UIDs for new statements.
10211 (linearize_expr): Likewise.
10212 (do_reassoc): Renumber gimple statement UIDs.
10213
e01c7cca
JH
102142013-05-17 Jan Hubicka <jh@suse.cz>
10215
10216 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
10217 weakrefs.
10218 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB
10219 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
10220 weakrefs.
e01c7cca
JH
10221 (output_weakrefs): Update.
10222
c3272a92
PCC
102232013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
10224 Martin Jambor <mjambor@suse.cz>
10225
10226 PR middle-end/57276
10227 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
10228 value that corresponds to the given aggval is found in values vector.
10229
11c2aa39
UB
102302013-05-17 Uros Bizjak <ubizjak@gmail.com>
10231
10232 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
10233 sse, sse2, sse3, ssse3 and sse4a flags to options.
10234
abecc8c6
DM
102352013-05-17 David Malcolm <dmalcolm@redhat.com>
10236
10237 * gengtype-state.c: (s_expr_writer): New class, to handle
10238 prettifying of output layout of s-expressions.
10239 (state_writer): New class, to write out gtype.state.
10240 (state_written_type_count): Move this variable into member data of
10241 state_writer.
10242 (s_expr_writer::s_expr_writer): New code: constructor for new class
10243 (state_writer::state_writer(): ditto
10244 (s_expr_writer::write_new_line): New function
10245 (s_expr_writer::write_any_indent): ditto
10246 (s_expr_writer::begin_s_expr): ditto
10247 (s_expr_writer::end_s_expr): ditto
10248 (write_state_fileloc): convert to method of state_writer...
10249 (state_writer:: write_state_fileloc): ...and use methods of
10250 s_expr_writer to write indentation into the gtype.state output file
10251 to visually represent the hierarchical structure of the list
10252 structures
10253 (write_state_fields): ditto, renaming to...
10254 (state_writer::write_state_fields)
10255 (write_state_a_string): ditto, renaming to...
10256 (state_writer::write_state_a_string)
10257 (write_state_string_option): ditto, renaming to...
10258 (state_writer::write_state_string_option)
10259 (write_state_type_option): ditto, renaming to...
10260 (state_writer::write_state_type_option)
10261 (write_state_nested_option): ditto, renaming to...
10262 (state_writer::write_state_nested_option)
10263 (write_state_option): ditto, renaming to...
10264 (state_writer::write_state_option)
10265 (write_state_options): ditto, renaming to...
10266 (state_writer::write_state_options)
10267 (write_state_lang_bitmap): ditto, renaming to...
10268 (state_writer::write_state_lang_bitmap)
10269 (write_state_version): ditto, renaming to...
10270 (state_writer::write_state_version)
10271 (write_state_scalar_type): ditto, renaming to...
10272 (state_writer::write_state_scalar_type)
10273 (write_state_string_type): ditto, renaming to...
10274 (state_writer::write_state_string_type)
10275 (write_state_undefined_type): ditto, renaming to...
10276 (state_writer::write_state_undefined_type)
10277 (write_state_struct_union_type): ditto, renaming to...
10278 (state_writer::write_state_struct_union_type)
10279 (write_state_struct_type): ditto, renaming to...
10280 (state_writer::write_state_struct_type)
10281 (write_state_user_struct_type): ditto, renaming to...
10282 (state_writer::write_state_user_struct_type)
10283 (write_state_lang_struct_type): ditto, renaming to...
10284 (state_writer::write_state_lang_struct_type)
10285 (write_state_param_struct_type): ditto, renaming to...
10286 (state_writer::write_state_param_struct_type)
10287 (write_state_pointer_type): ditto, renaming to...
10288 (state_writer::write_state_pointer_type)
10289 (write_state_array_type): ditto, renaming to...
10290 (state_writer::write_state_array_type)
10291 (write_state_gc_used): ditto, renaming to...
10292 (state_writer::write_state_gc_used)
10293 (write_state_common_type_content): ditto, renaming to...
10294 (state_writer::write_state_common_type_content)
10295 (write_state_type): ditto, renaming to...
10296 (state_writer::write_state_type)
10297 (write_state_pair_list): ditto, renaming to...
10298 (state_writer::write_state_pair_list)
10299 (write_state_pair): ditto, renaming to...
10300 (state_writer::write_state_pair)
10301 (write_state_typedefs): ditto, renaming to...
10302 (state_writer::write_state_typedefs)
10303 (write_state_structures): ditto, renaming to...
10304 (state_writer::write_state_structures)
10305 (write_state_param_structs): ditto, renaming to...
10306 (state_writer::write_state_param_structs)
10307 (write_state_variables): ditto, renaming to...
10308 (state_writer::write_state_variables)
10309 (write_state_srcdir): ditto, renaming to...
10310 (state_writer::write_state_srcdir)
10311 (write_state_files_list): ditto, renaming to...
10312 (state_writer::write_state_files_list)
10313 (write_state_languages): ditto, renaming to...
10314 (state_writer::write_state_languages)
10315 (write_state): create a state_writer instance and use it when
10316 writing out the state file
10317
d6545f29
MS
103182013-05-17 Mike Stump <mikestump@comcast.net>
10319
816a3f73 10320 PR rtl-optimization/57304
d6545f29
MS
10321 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
10322 accessing DF_REF_REAL_LOC.
10323
38c821cf
JJ
103242013-05-17 Jakub Jelinek <jakub@redhat.com>
10325
10326 PR rtl-optimization/57281
10327 PR rtl-optimization/57300
10328 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
10329 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
10330 what the other splitter did if the registers are dead.
10331
2a293391
RB
103322013-05-17 Richard Biener <rguenther@suse.de>
10333
10334 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
10335 MEM_REF offsets.
10336
17042d2b
JJ
103372013-05-17 Jakub Jelinek <jakub@redhat.com>
10338
10339 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
10340 linking.
10341
5b115c1f
MP
103422013-05-17 Marek Polacek <polacek@redhat.com>
10343
11c2aa39
UB
10344 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
10345 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 10346
52d84413
JJ
103472013-05-17 Jakub Jelinek <jakub@redhat.com>
10348
68119618
JJ
10349 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
10350 vect_external_def oprnd1 with loop_vinfo, try to emit
10351 optional cast, negation and and stmts on the loop preheader
10352 edge instead of into the pattern def seq.
10353
52d84413
JJ
10354 PR tree-optimization/57051
10355 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
10356 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
10357
17b962bd
NC
103582013-05-16 Nick Clifton <nickc@redhat.com>
10359
10360 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
10361 (rl78_is_naked_func): New function.
56aefbf7
UB
10362 (rl78_expand_prologue): Skip prologue generation for naked functions.
10363 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC
10364 * doc/extend.texi (naked): Add RL78 to the list of processors
10365 that supports this attribute.
10366
b1a0f84e
JL
103672013-05-16 Jeff Law <law@redhat.com>
10368
10369 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
10370
af0e415b
UB
103712013-05-16 Uros Bizjak <ubizjak@gmail.com>
10372
10373 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
10374 cache parameters using detect_caches_amd also for CYRIX,
10375 NSC and TM2 signatures.
10376
19db293a
UB
103772013-05-16 Uros Bizjak <ubizjak@gmail.com>
10378 Dzianis Kahanovich <mahatma@eu.by>
10379
10380 PR target/45359
10381 PR target/46396
10382 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
10383 VIA/Centaur processors and determine their cache parameters
10384 using detect_caches_amd.
10385
251a41b9
TJ
103862013-05-16 Teresa Johnson <tejohnson@google.com>
10387
10388 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
10389 (rtl_verify_edges): New function.
10390 (rtl_verify_bb_insns): Ditto.
10391 (rtl_verify_bb_pointers): Ditto.
10392 (rtl_verify_bb_insn_chain): Ditto.
10393 (rtl_verify_fallthru): Ditto.
10394 (rtl_verify_bb_layout): Ditto.
10395 (rtl_verify_flow_info_1): Outline checks into new functions.
10396 (rtl_verify_flow_info): Ditto.
10397
f14540b6
SE
103982013-05-16 Steve Ellcey <sellcey@imgtec.com>
10399
10400 * cfghooks.c (copy_bbs): Add update_dominance argument.
10401 * cfghooks.h (copy_bbs): Update prototype.
10402 * tree-cfg.c (gimple_duplicate_sese_region):
10403 Add update_dominance argument.
10404 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
10405 * tree-ssa-loop-ch.c (copy_loop_headers): Update
10406 gimple_duplicate_sese_region call.
10407 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
10408 Update copy_bbs call.
10409 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
10410 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
10411
7e9a3abb
JJ
104122013-05-16 Jakub Jelinek <jakub@redhat.com>
10413
10414 * tree-vectorizer.h (NUM_PATTERNS): Increment.
10415 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
10416 vect_recog_rotate_pattern.
10417 (vect_recog_rotate_pattern): New function.
10418
427b248d
JM
104192013-05-16 Jason Merrill <jason@redhat.com>
10420
10421 * Makefile.in (LLINKER): New variable.
10422 (mostlyclean): Remove link mutex.
10423 * configure.ac: Handle --enable-link-mutex.
10424 * lock-and-run.sh: New script.
10425
b871e3d2
RR
104262013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10427
10428 PR target/19599
10429 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
10430 for NULL decl.
10431
ce858126
RO
104322013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10433
10434 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
10435
798d3d04
GY
104362013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
10437
10438 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
10439 * config/arm/arm.c (next_consecutive_mem): New function.
10440 (gen_movmem_ldrd_strd): Likewise.
10441 * config/arm/arm.md (movmemqi): Update condition and code.
10442 (unaligned_loaddi, unaligned_storedi): New patterns.
10443
0baddc45
RO
104442013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10445
10446 * config.gcc: Obsolete *-*-solaris2.9*.
10447 * doc/install.texi (Specific, *-*-solaris2*): Document it.
10448
aa9d5bcf
RB
104492013-05-16 Richard Biener <rguenther@suse.de>
10450
10451 * passes.c (init_optimization_passes): Move pass_parallelize_loops
10452 earlier, after GRAPHITE transforms and IV canonicalization.
10453
5a0f4dd3
JJ
104542013-05-16 Jakub Jelinek <jakub@redhat.com>
10455
10456 * omp-low.c (extract_omp_for_data): For collapsed loops,
10457 if at least one of the loops is known at compile time to
10458 iterate zero times, set count to 0.
10459 (expand_omp_regimplify_p): New function.
10460 (expand_omp_for_generic): For collapsed loops, if at least
10461 one of the loops isn't known to iterate at least once,
10462 add runtime check with setting count to 0.
10463 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
10464 For unsigned types if it isn't known at compile time that
10465 the loop will iterate at least once, add runtime check to bypass
10466 the whole loop if initial condition isn't true.
10467
e3753785
NS
104682013-05-16 Nathan Sidwell <nathan@codesourcery.com>
10469
10470 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
10471
43bb4dd1
MG
104722013-05-16 Marc Glisse <marc.glisse@inria.fr>
10473
10474 PR middle-end/57286
10475 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
10476 transformations to avoid an infinite loop.
10477
3571dde6
MP
104782013-05-16 Marek Polacek <polacek@redhat.com>
10479
10480 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
10481
b8b3f0ca
LE
104822013-05-15 Leif Ekblad <leif@rdos.net>
10483
10484 * config/i386/i386.c (ix86_decompose_address): Use
10485 DEFAULT_TLS_SEG_REG to access TLS segment register.
10486 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
10487 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
10488 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
10489
e299a383
RS
104902013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
10491
10492 PR target/57260
10493 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
10494 sibling calls to functions that would normally be lazily bound,
10495 unless $gp is call-clobbered.
10496
e7208ea3
UB
104972013-05-15 Uros Bizjak <ubizjak@gmail.com>
10498
19db293a 10499 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 10500 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB
10501 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
10502 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB
10503 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
10504 of TARGET_3DNOW.
10505 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
10506
e7413f3d
AS
105072013-05-15 Andreas Schwab <schwab@suse.de>
10508
10509 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
10510 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
10511 third operand.
10512
ad4db775
TJ
105132013-05-15 Teresa Johnson <tejohnson@google.com>
10514
10515 * loop-unroll.c (report_unroll_peel): Check decision before
10516 emitting unroll/peel message.
10517
af205f67
TJ
105182013-05-15 Teresa Johnson <tejohnson@google.com>
10519
10520 * function.h (has_bb_partition): New rtl_data flag.
10521 (bb_reorder_complete): Ditto.
10522 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
10523 instead of flag_reorder_blocks_and_partition.
10524 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
10525 with some enhancements.
10526 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
10527 * bb-reorder.c (connect_traces): Check for has_bb_partition
10528 instead of flag_reorder_blocks_and_partition.
10529 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
10530 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
10531 verify_hot_cold_block_grouping.
10532 (partition_hot_cold_basic_blocks): Set has_bb_partition.
10533
9adcfa3c
RR
105342013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
10535
10536 PR target/19599
10537 * config/arm/predicates.md (call_insn_operand): New predicate.
10538 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
10539 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
10540 if insn is not a tail call.
10541 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
10542 registers.
10543 * config/arm/arm.h (enum reg_class): New caller save register class.
10544 (REG_CLASS_NAMES): Likewise.
10545 (REG_CLASS_CONTENTS): Likewise.
10546 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
10547 without decls.
10548
ccb3ad87
RB
105492013-05-15 Richard Biener <rguenther@suse.de>
10550
10551 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
10552 of MSG_OPTIMIZED_LOCATIONS.
10553 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
10554 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
10555 message.
10556 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
10557 of MSG_OPTIMIZED_LOCATIONS.
10558 (execute_vect_slp): Likewise.
10559 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
10560 (vect_create_cond_for_alias_checks): Likewise.
10561 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
10562 (vect_recog_widen_mult_pattern): Likewise.
10563 (vect_recog_widen_sum_pattern): Likewise.
10564 (vect_recog_over_widening_pattern): Likewise.
10565 (vect_recog_widen_shift_pattern): Likewise.
10566 (vect_recog_vector_vector_shift_pattern): Likewise.
10567 (vect_recog_divmod_pattern): Likewise.
10568 (vect_recog_mixed_size_cond_pattern): Likewise.
10569 (vect_recog_bool_pattern): Likewise.
10570 (vect_pattern_recog_1): Likewise.
10571
48b1474e
MJ
105722013-05-15 Martin Jambor <mjambor@suse.cz>
10573
10574 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
10575 non-functions to builtin_unreachable.
10576 * ipa-inline-transform.c (inline_call): Do not assert estimates were
10577 correct when new direct edges were discovered.
10578
9de04252
MJ
105792013-05-15 Martin Jambor <mjambor@suse.cz>
10580
10581 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
10582 header, print symbol order instead of node uid, print more information
10583 about indirect edge targets.
10584 (ipa_make_edge_direct_to_target): Print symbol order instead of node
10585 uids.
10586 (ipa_make_edge_direct_to_target): Likewise.
10587 (remove_described_reference): Likewise.
10588 (propagate_controlled_uses): Likewise.
10589 (ipa_print_node_params): Also print symbol order.
10590 (ipcp_transform_function): Print symbol order instead of node uids.
10591 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
10592 (cgraph_get_create_real_symbol_node): Likewise.
10593 * ipa-cp.c (print_lattice): Likewise.
10594 (print_all_lattices): Likewise.
10595 (determine_versionability): Likewise.
10596 (initialize_node_lattices): Likewise.
10597 (estimate_local_effects): Likewise.
10598 (update_profiling_info): Likewise.
10599 (create_specialized_node): Likewise.
10600 (perhaps_add_new_callers): Likewise.
10601 (decide_about_value): Likewise.
10602 (decide_whether_version_node): Likewise.
10603 (identify_dead_nodes): Likewise.
10604 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
10605 (dump_inline_summary): Likewise.
10606 (estimate_node_size_and_time): Likewise.
10607 (inline_analyze_function): Likewise.
10608 * ipa-inline.c (report_inline_failed_reason): Likewise.
10609 (want_early_inline_function_p): Likewise.
10610 (edge_badness): Likewise.
10611 (update_edge_key): Likewise.
10612 (inline_small_functions): Likewise. Add dumping of order to two other
10613 dumps.
10614 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
10615 instead of node uids.
10616 (propagate_pure_const): Likewise.
10617 (propagate_pure_const): Likewise.
10618 * ipa-utils.c (dump_cgraph_node_set): Likewise.
10619 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
10620 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
10621 of node uids.
10622 * tree-pretty-print.c (dump_function_header): Likewise.
10623 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
10624 Print symbol order instead of node uids.
10625
1dd03b91
AK
106262013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10627
10628 * config/s390/s390.c (s390_register_move_cost): Don't impose the
10629 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
10630
fdf6a7b9
RB
106312013-05-15 Richard Biener <rguenther@suse.de>
10632
10633 PR tree-optimization/57275
10634 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
10635 return value for fail to do runtime alias checks for gather loads.
10636
2d6e4603
JH
106372013-05-15 Jan Hubicka <jh@suse.cz>
10638
10639 PR lto/57038
10640 PR lto/47375
e7208ea3
UB
10641 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
10642 weakrefs are not external.
10643 (lto_symtab_merge_decls): Fix thinko when dealing with
10644 non-lto_symtab decls.
2d6e4603
JH
10645 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
10646 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
10647 * varpool.c (dump_varpool_node): Dump more flags.
10648
83f44b39
GG
106492013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10650
10651 * config/i386/i386.c (processor_alias_table): Add instruction
10652 FSGSBASE for AMD bdver3 architecture.
10653
2e55d062
JJ
106542013-05-14 Jakub Jelinek <jakub@redhat.com>
10655
10656 * tree.c (warn_deprecated_use): Print file:line using locus color.
10657 * diagnostic.c (diagnostic_report_current_module): Print file:line
10658 and file:line:column using locus color.
10659
0bfdb81e
MS
106602013-05-14 Mike Stump <mikestump@comcast.net>
10661
10662 * gdbinit.in: Add __null.
10663
a508ef22
MS
106642013-05-14 Mike Stump <mikestump@comcast.net>
10665
10666 * recog.h: Rename struct recog_data to Recog_data.
10667 * recog.c: Likewise.
10668 * reload.c (can_reload_into): Likewise.
10669 * config/picochip/picochip.c: Likewise.
10670
e7180acb
MS
106712013-05-14 Mike Stump <mikestump@comcast.net>
10672
10673 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
10674
61216c26
SB
106752013-05-14 Steven Bosscher <steven@gcc.gnu.org>
10676
fd6beed4
SB
10677 * resource.h (struct resources): Remove unch_memory member.
10678 (CLEAR_RESOURCE): Don't clear unch_memory.
10679 * resource.c (mark_referenced_resources): Don't set it.
10680 (mark_set_resources): Likewise.
10681 (mark_target_live_regs): Don't clear it.
10682 (init_resource_info): Likewise.
10683 * reorg.c (resource_conflicts_p): Don't compare it.
10684 (redundant_insn): Don't set it.
10685
61216c26
SB
10686 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
10687 Remove prototypes.
10688 * emit-rtl.c (next_label): Remove unused function.
10689 (skip_consecutive_labels, link_cc0_insns): Move to ...
10690 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
10691 only place where these functions are used, and make them static.
10692
418d1b87
MG
106932013-05-14 Marc Glisse <marc.glisse@inria.fr>
10694
10695 * fold-const.c (fold_negate_expr): Handle vectors.
10696 (fold_truth_not_expr): Make it static.
10697 (fold_invert_truthvalue): New static function.
10698 (invert_truthvalue_loc): Handle vectors. Do not call
10699 fold_truth_not_expr directly.
10700 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
10701 <TRUTH_NOT_EXPR>: Do not cast to boolean.
10702 (fold_comparison): Handle vector constants.
10703 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
10704 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
10705 * tree.h (fold_truth_not_expr): Remove declaration.
10706
fc21784d
JG
107072013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
10708
10709 * config/aarch64/aarch64-simd.md
10710 (aarch64_vcond_internal<mode>): Rename to...
10711 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
10712 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
10713 float modes. Clarify all iterator modes.
10714 (vcond<mode><mode>): Use new name for vcond expanders.
10715 (vcond<v_cmp_result><mode>): Likewise.
10716 (vcondu<mode><mode>: Likewise.
10717 * config/aarch64/iterators.md (VDQF_COND): New.
10718
d4c52634
MG
107192013-05-14 Marc Glisse <marc.glisse@inria.fr>
10720
10721 PR bootstrap/57266
10722 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
10723 variable for the shift amount. Check that we shift by non-negative
10724 amounts.
10725
2b261262
CLT
107262013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
10727
10728 PR target/42017
10729 * config/arm/arm.h (EPILOGUE_USES): Only return true
10730 for LR_REGNUM after epilogue_completed.
10731
12211b99 107322013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR
10733
10734 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 10735 is error_mark_node.
6e022d7b 10736
56cf7859
RO
107372013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10738
10739 PR target/57261
10740 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
10741 and Solaris 11+/x86 with gld.
10742 * configure: Regenerate.
10743
75776c6d
JJ
107442013-05-14 Jakub Jelinek <jakub@redhat.com>
10745
10746 * expmed.c (expand_shift_1): Canonicalize rotates by
10747 constant bitsize / 2 to bitsize - 1.
7f998021 10748 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ
10749 case ROTATERT>: Likewise.
10750
10751 Revert:
10752 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10753
10754 * config/i386/i386.md (rotateinv): New code attr.
10755 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10756 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10757 roll $31, %eax, etc.
10758
df35498a
RB
107592013-05-14 Richard Biener <rguenther@suse.de>
10760
10761 PR middle-end/57235
10762 * tree-eh.c (sink_clobbers): Give up for successors with
10763 multiple predecessors and no virtual uses.
10764
cc6e7ece
EB
107652013-05-14 Eric Botcazou <ebotcazou@adacore.com>
10766
10767 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
10768 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
10769
f778dd4d
JJ
107702013-05-14 Jakub Jelinek <jakub@redhat.com>
10771
10772 PR middle-end/57251
10773 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
10774 the case when both op0 and op1 have VOIDmode.
10775
bad4df9b
KP
107762013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10777
10778 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
10779 in multiply-accumulate mode.
10780
56f3e9ac
GW
107812013-05-13 Guozhi Wei <carrot@google.com>
10782
10783 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
10784
a3d7ab92
KT
107852013-05-13 Kai Tietz <ktietz@redhat.com>
10786
10787 PR target/56975
10788 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 10789 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 10790 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 10791 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT
10792 (ix86_expand_prologue): Likewise.
10793 (ix86_expand_split_stack_prologue): Likewise.
10794 (legitimate_pic_address_disp_p): Likewise.
10795 (legitimize_pic_address): Likewise.
10796 (legitimize_tls_address): Likewise.
10797 (legitimize_pe_coff_symbol): Likewise.
10798 (output_pic_addr_const): Likewise.
10799 (construct_plt_address): Likewise.
10800 (ix86_expand_call): Likewise.
10801 (x86_output_mi_thunk): Likewise.
10802 (x86_function_profiler): Likewise.
10803
c59b7e28
SN
108042013-05-13 Sofiane Naci <sofiane.naci@arm.com>
10805
10806 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
10807 similar switch cases.
10808 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
10809 (aarch64_simd_mov_to_<mode>low): Delete.
10810 (aarch64_simd_mov_to_<mode>high): Delete.
10811 (move_lo_quad_<mode>): Add w<-r alternative.
10812 (aarch64_simd_move_hi_quad_<mode>): Likewise.
10813 (aarch64_simd_mov_from_*): Update type attribute.
10814 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
10815 statement.
10816
5f28524a
JH
108172013-05-13 Jan Hubicka <jh@suse.cz>
10818
10819 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
10820 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB
10821 ix86_expand_epilogue, emit_i387_cw_initialization,
10822 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
10823 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 10824
ae6fa899
JJ
108252013-05-13 Jakub Jelinek <jakub@redhat.com>
10826
10827 PR tree-optimization/45216
10828 PR tree-optimization/57157
10829 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
10830 the (-Y) & (B - 1) variant if OP is |.
10831 * expmed.c (expand_shift_1): For rotations by const0_rtx just
10832 return shifted. Use (-op1) & (prec - 1) as other_amount
10833 instead of prec - op1.
10834
4502fe8d
MJ
108352013-05-13 Martin Jambor <mjambor@suse.cz>
10836
10837 PR middle-end/42371
10838 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
10839 (ipa_constant_data): New type.
10840 (ipa_jump_func): Use ipa_constant_data to hold information about
10841 constant jump functions.
10842 (ipa_get_jf_constant): Adjust to jump function type changes.
10843 (ipa_get_jf_constant_rdesc): New function.
10844 (ipa_param_descriptor): New field controlled_uses.
10845 (ipa_get_controlled_uses): New function.
10846 (ipa_set_controlled_uses): Likewise.
10847 * ipa-ref.h (ipa_find_reference): Declare.
10848 * ipa-prop.c (ipa_cst_ref_desc): New type.
10849 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
10850 changes.
10851 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
10852 New parameter cs. Adjust all callers.
10853 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
10854 (remove_described_reference): New function.
10855 (jfunc_rdesc_usable): Likewise.
10856 (try_make_edge_direct_simple_call): Decrement controlled use count,
10857 attempt to remove reference if it hits zero.
10858 (combine_controlled_uses_counters): New function.
10859 (propagate_controlled_uses): Likewise.
10860 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
10861 (ipa_edge_duplication_hook): Duplicate reference descriptions.
10862 (ipa_print_node_params): Print described use counter.
10863 (ipa_write_jump_function): Adjust to jump function type changes.
10864 (ipa_read_jump_function): New parameter CS, pass it to
10865 ipa_set_jf_constant. Adjust caller.
10866 (ipa_write_node_info): Stream controlled use count
10867 (ipa_read_node_info): Likewise.
10868 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
10869 asserting.
10870 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
10871 count. Remove cloning-added reference if it reaches zero.
10872 * ipa-ref.c (ipa_find_reference): New function.
10873
0864bfc2
GG
108742013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10875
e7208ea3 10876 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG
10877 alignment values for AMD BD and BT architectures.
10878
640bfeb2
MG
108792013-05-13 Marc Glisse <marc.glisse@inria.fr>
10880
10881 * tree-vect-generic.c (uniform_vector_p): Move ...
10882 * tree.c (uniform_vector_p): ... here.
10883 * tree.h (uniform_vector_p): Declare it.
10884 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
10885 into a scalar.
10886
3a60f32b
JJ
108872013-05-13 Jakub Jelinek <jakub@redhat.com>
10888
198fe1bf
JJ
10889 PR tree-optimization/57230
10890 * tree-ssa-strlen.c (handle_char_store): Record length for
10891 array store from STRING_CST.
10892
3a60f32b
JJ
10893 PR tree-optimization/57230
10894 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
10895 check.
10896
566be57c
JR
108972013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
10898
10899 * config/epiphany/epiphany.c (epiphany_init): Check size of
10900 NUM_MODES_FOR_MODE_SWITCHING.
10901 (epiphany_expand_prologue):
10902 Remove CONFIG_REGNUM initial value handling code.
10903 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
10904 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 10905 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR
10906 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
10907 Don't return 1 for FP_MODE_NONE.
10908 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
10909 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
10910 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
10911 * config/epiphany/epiphany.md (save_config): New pattern.
10912
0f2c2331
UB
109132013-05-12 Uros Bizjak <ubizjak@gmail.com>
10914
10915 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
10916
5b3f93c7
UB
109172013-05-10 Uros Bizjak <ubizjak@gmail.com>
10918
10919 * config/i386/i386.md (memory): Handle sseishft1.
10920 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
10921 (*vec_extractv2di_1): Ditto.
10922
1f873f0e
VM
109232013-05-10 Vladimir Makarov <vmakarov@redhat.com>
10924
10925 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
10926 saved registers.
10927
341427fa 109282013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH
10929
10930 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
10931 Add mthumb/march=armv7-a multilib.
10932 Add mthumb/march=armv7-r multilib.
10933 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
10934
9e69bdde
RC
109352013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
10936
10937 * config/v850/t-rtems: Add more multilibs.
10938
9ff09a22
RB
109392013-05-10 Richard Biener <rguenther@suse.de>
10940
10941 PR tree-optimization/57214
10942 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
10943 not propagate from SSA names that occur in abnormal PHI nodes.
10944
a5e0cd1d
MG
109452013-05-10 Marc Glisse <marc.glisse@inria.fr>
10946
10947 * stor-layout.c (element_precision): New function.
10948 * machmode.h (element_precision): Declare it.
10949 * tree.c (build_minus_one_cst): New function.
10950 (element_precision): Likewise.
10951 * tree.h (build_minus_one_cst): Declare new function.
10952 (element_precision): Likewise.
10953 * fold-const.c (operand_equal_p): Use element_precision.
10954 (fold_binary_loc): Handle vector types.
10955 * convert.c (convert_to_integer): Use element_precision.
10956 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
10957 separately.
10958
cb2558bc
RS
109592013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
10960
10961 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
10962 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
10963 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
10964 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10965 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
10966 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
10967 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
10968 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10969 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
10970 (Uuw8): New constraints.
10971 (Usb4): Move into alphabetical order.
10972 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
10973 (sd8_operand, ub8_operand, uw8_operand): New predicates.
10974 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
10975 previously unnamed patterns.
10976 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
10977 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
10978 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
10979 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
10980 of set_attr_alternative/if_then_else. Use extended_mips16 instead
10981 of specific lengths.
10982
cb3b8d33
JJ
109832013-05-10 Jakub Jelinek <jakub@redhat.com>
10984
6f93c008
JJ
10985 * config/i386/i386.md (rotateinv): New code attr.
10986 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10987 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10988 roll $31, %eax, etc.
10989
cb3b8d33
JJ
10990 PR tree-optimization/45216
10991 PR tree-optimization/57157
10992 * tree-ssa-forwprop.c (simplify_rotate): New function.
10993 (ssa_forward_propagate_and_combine): Call it.
10994
afb119be
RB
109952013-05-10 Richard Biener <rguenther@suse.de>
10996
10997 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
10998 disable peeling when we version for aliasing.
10999 (vector_alignment_reachable_p): Honor explicit user alignment.
11000 (vect_supportable_dr_alignment): Likewise.
11001 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
11002 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
11003 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
11004 then peeling to arrange for the cost-model check to come first.
11005
01ae4861
AM
110062013-05-10 Alan Modra <amodra@gmail.com>
11007
11008 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
11009 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
11010 * configure: Regenerate.
11011
ebc9a431
AM
110122013-05-10 Alan Modra <amodra@gmail.com>
11013
11014 PR target/55033
11015 * varasm.c (default_elf_select_section): Move !DECL_P check..
11016 (get_named_section): ..to here before calling get_section_name.
11017 Adjust assertion.
11018 (default_section_type_flags): Add DECL_P check.
11019 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
11020 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
11021
d4bca93c
JR
110222013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
11023
11024 * config/epiphany/epiphany.c (epiphany_expand_prologue):
11025 When using gen_stack_adjust_str with a register offset, add a
11026 REG_FRAME_RELATED_EXPR note.
11027
60ca9a65
UB
110282013-05-09 Uros Bizjak <ubizjak@gmail.com>
11029
11030 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
11031 (*vec_extractv4si_zext_mem): Ditto.
11032 (*vec_extractv2di): Add 0->x and x->x alternatives.
11033 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
11034 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
11035
a3409c02
JM
110362013-05-09 Jason Merrill <jason@redhat.com>
11037
0138d6b2
JM
11038 N3639 C++1y VLA support
11039 * gimplify.c (gimplify_vla_decl): Don't touch an existing
11040 DECL_VALUE_EXPR.
11041
a3409c02
JM
11042 * tree.c (build_constructor_va): New.
11043 * tree.h: Declare it.
11044
66e6b990
MJ
110452013-05-09 Martin Jambor <mjambor@suse.cz>
11046
11047 PR lto/57084
11048 * gimple-fold.c (canonicalize_constructor_val): Call
11049 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
11050
64cfa6c0
JH
110512013-05-09 Jan Hubicka <jh@suse.cz>
11052 Richard Biener <rguenther@suse.de>
11053
11054 PR lto/54095
11055 * symtab.c (symtab_make_decl_local): Do not add private names.
11056
c3167b00
JH
110572013-05-09 Jan Hubicka <jh@suse.cz>
11058
11059 PR lto/54095
11060 * symtab.c (insert_to_assembler_name_hash): Handle clones.
11061 (unlink_from_assembler_name_hash): Likewise.
11062 (symtab_prevail_in_asm_name_hash, symtab_register_node,
11063 symtab_unregister_node, symtab_initialize_asm_name_hash,
11064 change_decl_assembler_name): Update.
11065
12dc6974
SN
110662013-05-09 Sofiane Naci <sofiane.naci@arm.com>
11067
11068 * config/aarch64/aarch64.md: New movtf split.
11069 (*movtf_aarch64): Update.
11070 (aarch64_movdi_tilow): Handle TF modes and rename to
11071 aarch64_movdi_<mode>low.
11072 (aarch64_movdi_tihigh): Handle TF modes and rename to
11073 aarch64_movdi_<mode>high
11074 (aarch64_movtihigh_di): Handle TF modes and rename to
11075 aarch64_mov<mode>high_di
11076 (aarch64_movtilow_di): Handle TF modes and rename to
11077 aarch64_mov<mode>low_di
11078 (aarch64_movtilow_tilow): Remove spurious whitespace.
11079 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
11080 splits.
11081 (aarch64_print_operand): Update.
11082
227eb343
AM
110832013-05-09 Alan Modra <amodra@gmail.com>
11084
11085 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
11086 powerpc64le.
11087 * configure: Regenerate.
11088
0b013847
UB
110892013-05-08 Uros Bizjak <ubizjak@gmail.com>
11090
11091 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
11092 splitter preparation statements.
11093 * config/i386/sse.md (*vec_extract* splitters): Ditto.
11094 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
11095 adjust_address_nv.
11096
1dc3d6e9
BS
110972013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11098
11099 * gimple-ssa-strength-reduction.c (count_candidates): Change
11100 return value to int.
11101 (analyze_candidates_and_replace): Change type of length to int.
11102
e61e7d28
UB
111032013-05-08 Uros Bizjak <ubizjak@gmail.com>
11104
11105 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
11106 (*vec_extract<mode>): Use VI12_128 mode iterator.
11107 (*vec_extract<mode>_mem): Ditto.
11108 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
11109 attribute.
11110
4fbfcf44
DN
111112013-05-08 Diego Novillo <dnovillo@google.com>
11112
11113 PR bootstrap/54659
11114
11115 Revert:
4fbfcf44
DN
11116 2012-08-17 Diego Novillo <dnovillo@google.com>
11117
e61e7d28
UB
11118 PR bootstrap/54281
11119 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
11120 * config.in: Regenerate.
11121 * configure: Regenerate.
11122 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 11123
702d8703
JH
111242013-05-08 Jan Hubicka <jh@suse.cz>
11125
11126 PR lto/54095
11127 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
11128 * cgraph.h (symtab_node_base): Add unique_name.
11129 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11130 input_overwrite_node, input_varpool_node): Stream unique_name.
11131 * cgraphclones.c (cgraph_create_virtual_clone,
11132 cgraph_function_versioning): Set unique_name.
11133 * ipa.c (function_and_variable_visibility): Set unique_name.
11134
8b28cf47
BS
111352013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11136
11137 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
11138 (alloc_cand_and_find_basis): Restrict conditional candidate
11139 processing to CAND_MULTs.
11140
e86074fd
JH
111412013-05-08 Jan Hubicka <jh@suse.cz>
11142
11143 PR lto/54095
11144 lto-symtab.c (lto_symtab_symbol_p): New function.
11145 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
11146 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
11147 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
11148 Skip static symbols.
11149
44398cbe
PC
111502013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
11151
11152 PR tree-optimization/57200
11153 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
11154 Only call inform if the preceding warning_at returns true.
11155
f6bc1c4a
HS
111562013-05-07 Han Shen <shenhan@google.com>
11157
11158 * cfgexpand.c (record_or_union_type_has_array_p): New function.
11159 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
11160 * common.opt (fstack-protector-strong): New option.
11161 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
11162 * doc/invoke.texi (Optimization Options): Document
11163 "-fstack-protector-strong".
11164 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
11165
4ffecb1f
SB
111662013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11167
11168 * config/mips/mips.c (mips_machine_reorg2): Return 0.
11169
5a107a0f
VM
111702013-05-07 Vladimir Makarov <vmakarov@redhat.com>
11171
11172 * ira.c (update_equiv_regs): Add insn having equiv memory even if
11173 it is not lhs of the insn.
11174 (setup_reg_equiv): Remove insn having equiv memory which it is not
11175 lhs of the insn.
11176 * lra-constraints.c (process_address): Try to improve generation
11177 code for address base + disp.
11178 (lra_constraints): Make correct the code for checking insn setting
11179 up backward equivalence. Remove insn only if it is in the init
11180 insn list.
11181 * lra-eliminations.c (update_reg_eliminate): Change return value.
11182 (lra_eliminate): Use the result.
11183
3f5783ea
UB
111842013-05-07 Uros Bizjak <ubizjak@gmail.com>
11185
11186 * config/i386/sse.md (ssescalarnummask): New mode attribute.
11187 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
11188 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
11189 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
11190 register target operands.
11191 (*vec_extractv8hi_sse2): New pattern.
11192 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
11193 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
11194 (*vec_extract<mode>_mem): New insn and split pattern.
11195
8a5800b8
CL
111962013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
11197
11198 * config/arm/arm.c (arm_asan_shadow_offset): New function.
11199 (TARGET_ASAN_SHADOW_OFFSET): Define.
11200 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
11201 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
11202
7bf55a70
BS
112032013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11204
11205 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
11206 (incr_vec_index): Return -1 if increment not found.
11207 (create_add_on_incoming_edge): Assert if increment not found.
11208 (record_increment): Limit number of increments recorded.
11209 (all_phi_incrs_profitable): Return false if an increment not found.
11210 (replace_profitable_candidates): Don't process increments that were
11211 not recorded.
11212 (analyze_candidates_and_replace): Limit size of incr_vec.
11213
3f8825c0
RB
112142013-05-07 Richard Biener <rguenther@suse.de>
11215
11216 * calls.c (special_function_p): setjmp-like functions are leaf.
11217 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
11218 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
11219
fd4842cd
SN
112202013-05-07 Sofiane Naci <sofiane.naci@arm.com>
11221
11222 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
11223 (aarch64_simd_mov<mode>): New expander.
11224 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
11225 (aarch64_simd_mov_to_<mode>high): Likewise.
11226 (aarch64_simd_mov_from_<mode>low): Likewise.
11227 (aarch64_simd_mov_from_<mode>high): Likewise.
11228 (aarch64_dup_lane<mode>): Update.
11229 (aarch64_dup_lanedi): New instruction pattern.
11230 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
11231 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
11232
a7a7d10e
BS
112332013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11234
11235 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
11236 (replace_mult_candidate): Remove unnecessary argument; remove
11237 unnecessary parameter from call to introduce_cast_before_cand.
11238 (replace_unconditional_candidate): Remove unnecessary parameter
11239 from call to replace_mult_candidate.
11240 (replace_conditional_candidate): Likewise.
11241 (insert_initializers): Use make_temp_ssa_name.
11242 (introduce_cast_before_cand): Remove unnecessary argument; use
11243 make_temp_ssa_name.
11244 (replace_one_candidate): Remove unnecessary argument; remove
11245 unnecessary parameter from calls to introduce_cast_before_cand.
11246 (replace_profitable_candidates): Remove unnecessary parameters
11247 from calls to replace_one_candidate.
11248
29105868
BS
112492013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11250
11251 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
11252 phi def as possibly hiding a basis for a CAND_ADD whose operands
11253 have been commuted in the analysis.
11254 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
11255
4095f9fa
N
112562013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11257
11258 * config/aarch64/aarch64.md
11259 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
11260 shift value between 0-4.
11261
39e843e8
RB
112622013-05-07 Richard Biener <rguenther@suse.de>
11263
11264 * double-int.h (rshift): New overload.
11265 * double-int.c (rshift): New function.
11266 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
11267 (create_reference_ops_from_ref): Remove.
11268 (vn_reference_insert): Use shared ops for constructing the
11269 reference and copy it.
11270
0a1a83cb
RB
112712013-05-07 Richard Biener <rguenther@suse.de>
11272
11273 PR middle-end/57190
11274 * tree-eh.c (sink_clobbers): Properly propagate
11275 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
11276
43303d6f
JJ
112772013-05-07 Jakub Jelinek <jakub@redhat.com>
11278
ba7e83f8
JJ
11279 PR tree-optimization/57149
11280 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
11281 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
11282 collect_phi_def_edges, execute_late_warn_uninitialized): Use
11283 uninit_undefined_value_p instead of ssa_undefined_value_p.
11284
43303d6f
JJ
11285 PR debug/57184
11286 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
11287 for modifier == EXPAND_INITIALIZER.
11288
14523c25
AB
112892013-05-07 Anton Blanchard <anton@samba.org>
11290
11291 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
11292 for powerpc64 little endian.
11293 * configure: Regenerate.
11294
cb7c8be9
GS
112952013-05-06 Graham Stott <grahams@btinternet.com>
11296
11297 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
11298 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
11299 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
11300 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
11301
113022013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS
11303
11304 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
11305 codes which allow non-lvalues.
11306
9a0ee7b0
MG
113072013-05-06 Marc Glisse <marc.glisse@inria.fr>
11308
11309 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
11310 components are all 1s.
11311 (integer_minus_onep): New function.
11312 * tree.h (integer_minus_onep): Declare it.
11313 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
11314 integer_minus_onep instead of integer_all_onesp.
11315
f2c17ea9
OE
113162013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11317
11318 PR target/52933
11319 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
11320 variations of these patterns.
11321
f75e6a51
UB
113222013-05-06 Uros Bizjak <ubizjak@gmail.com>
11323
11324 * config/i386/i386.md (isa): Add x64_sse4 member.
11325 (enabled): Handle x64_sse4.
11326 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
11327 instruction for 64bit SSE4_1 targets. Update insn attributes.
11328 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
11329 instruction for SSE4_1 targets. Update insn attributes.
11330 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11331 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
11332 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
11333 const_1 selector.
11334 (*vec_extractv4si): Rename from *sse4_1_pextrd.
11335 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
11336 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
11337
a986d468
OE
113382013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11339
11340 PR target/57108
11341 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
11342
382522cb
MK
113432013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
11344
11345 * final.c (do_assembler_dialects): Don't handle curly braces and
11346 vertical bar escaped by % as dialect delimiters.
11347 (output_asm_insn): Print curly braces and vertical bar if escaped
11348 by % and ASSEMBLER_DIALECT defined.
11349 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
11350 * doc/tm.texi: Regenerated.
11351
fb0d5c60
SB
113522013-05-06 Steven Bosscher <steven@gcc.gnu.org>
11353
fb0d5c60
SB
11354 * config/mips/mips.c: Include tree-pass.h.
11355 (mips_reorg): Split in pre- and post-dbr_schedule parts.
11356 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
11357 (pass_mips_machine_reorg2): New machine specific pass.
11358 (insert_pass_mips_machine_reorg2): New pass plugin definition.
11359 (mips_option_override): Register the new pass.
11360 * rtl.h (cleanup_barriers): Remove prototype.
11361 (dbr_schedule): Likewise.
11362 * jump.c (cleanup_barriers): Make static.
11363 * reorg.c (dbr_schedule): Likewise.
11364
aa06a978
RB
113652013-05-06 Richard Biener <rguenther@suse.de>
11366
11367 PR tree-optimization/57185
11368 * tree-parloops.c (add_field_for_reduction): Handle anonymous
11369 SSA names properly.
11370
0b953bec
UB
113712013-05-06 Uros Bizjak <ubizjak@gmail.com>
11372
11373 PR target/57106
11374 * config/i386/i386.c (add_parameter_dependencies): Add dependence
11375 between "first_arg" and "insn", not "last" and "insn".
11376
28708525
WS
113772013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11378
11379 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
11380 (find_candidates_in_block): Re-enable slsr_process_phi.
11381 (create_phi_basis): Fix double counting of candidate adjustment.
11382
0107dca2
RB
113832013-05-06 Richard Biener <rguenther@suse.de>
11384
11385 PR middle-end/57147
11386 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
11387 the edge is also fallthru, preserve it and just clear the
11388 abnormal flag.
11389 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
11390 also complex, preserve that and just clear the fallthru flag.
11391 * tree-inline.c (update_ssa_across_abnormal_edges): Also
11392 update virtual operands.
11393
470d4d13
AM
113942013-05-06 Alan Modra <amodra@gmail.com>
11395
11396 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
11397 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
11398 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
11399 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
11400 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11401 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
11402
8f1dbf8d
AM
114032013-05-06 Alan Modra <amodra@gmail.com>
11404
11405 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
11406 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
11407 (DEFAULT_ASM_ENDIAN): Define.
11408 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
11409 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
11410 Update -K PIC clause from sysv4.h.
11411 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
11412 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
11413
54890767
AM
114142013-05-06 Alan Modra <amodra@gmail.com>
11415
11416 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
11417 twice for little-endian.
11418 (ashrdi3_no_power, ashrdi3): Support little-endian.
11419
2353515d
OE
114202013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
11421
11422 PR target/55303
11423 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
11424 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
11425 related expanders.
11426 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
11427 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
11428 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
11429 New predicates.
11430
33e67557
SB
114312013-05-05 Steven Bosscher <steven@gcc.gnu.org>
11432 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
11433
11434 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
11435 * config/pa/pa.opt: Make mbig-switch a no-op.
11436 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
11437 (CASE_VECTOR_MODE): Always return SImode.
11438 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
11439 for the !TARGET_BIG_SWITCH case.
11440 * config/pa/pa-linux.h: Likewise.
11441 * config/pa/pa-openbsd.h: Likewise.
11442 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
11443 * config/pa/pa.md (short_jump): Remove define_insn.
11444 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
11445 (casesi0): Remove define_insn.
11446 (type): Remove btable_branch.
11447 (pa_combine_type): Likewise.
11448 (in_nullified_branch_delay): Likewise.
11449 (in_call_delay): Likewise.
11450 (define_delay): Likewise.
11451 (define_insn_reservation "Z3"): Likewise.
11452 (define_insn_reservation "Z4"): Likewise.
11453 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
11454 (pa_adjust_insn_length): Remove adjustment for btable branches.
11455 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
11456 and mno-big-switch
11457
3095685e
UB
114582013-05-05 Uros Bizjak <ubizjak@gmail.com>
11459
11460 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
11461 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
11462 Add m->r,x alternatives.
11463 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
11464 splitters using SWI48x mode iterator.
11465 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
11466 TARGET_64BIT. Add m->x alternative.
11467 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
11468 Add o->x alternative. Enable for TARGET_SSE.
11469 (sse_storeq): Remove expander.
11470 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
11471 with memory input operand.
11472 (*vec_extractv2di_1 splitter): New.
11473 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
11474 * config/i386/i386.md (ssevecmodelower): New mode attribute.
11475
4b36ae28
SB
114762013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11477
11478 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
11479 (INT_LOWPART): Delete.
11480 (extract_MB): Adjust.
11481 (extract_ME): Adjust.
11482 (print_operand): Adjust.
11483
da226db2
SB
114842013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
11485
11486 * config/rs6000/predicates.md (reg_or_add_cint_operand,
11487 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11488 (reg_or_logical_cint_operand, easy_fp_constant,
11489 logical_const_operand): Delete "CONST_DOUBLE" case.
11490 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
11491 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 11492 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB
11493 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
11494 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
11495 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
11496 test.
11497 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
11498 CONST_DOUBLE DImode/VOIDmode case.
11499 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
11500 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11501 CONST_DOUBLE VOIDmode case.
11502 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
11503 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
11504 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
11505 Delete CONST_DOUBLE case.
11506 (splitters for mov FMOVE64 const_double): Delete
11507 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
11508 "HOST_BITS_PER_WIDE_INT >= 64" test.
11509 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
11510 case.
11511 (mov DI const_double): Delete.
11512
40de22d6
JJ
115132013-05-04 Jakub Jelinek <jakub@redhat.com>
11514
11515 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
11516 on op shows all bits zero in mode of a lowpart subreg, return zero.
11517
5ec6aff2
MM
115182013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
11519
11520 PR target/57150
11521 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
11522 to save TFmode registers and DImode to save TImode registers for
11523 caller save operations.
11524 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
11525 mark being partially clobbered since they only use the first
11526 double word.
11527
11528 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
11529 and TDmode only use the upper 64-bits of each VSX register.
11530
2cefad90
BS
115312013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11532
11533 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
11534 (find_candidates_in_block): Disable slsr_process_phi.
11535
d6d7eee1
GW
115362013-05-03 Guozhi Wei <carrot@google.com>
11537
11538 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB
11539 constructor to ...
11540 (build_init_ctor): ... here.
d6d7eee1 11541
9b92d12b
BS
115422013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11543
11544 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
11545 (slsr_cand_d): Redefine def_phi.
11546 (stride_status, phi_adjust_status, count_phis_status): New enums.
11547 (find_phi_def): New.
11548 (find_basis_for_base_expr): New.
11549 (find_basis_for_candidate): Handle hidden bases.
11550 (alloc_cand_and_find_basis): Handle phi candidates.
11551 (slsr_process_phi): New.
11552 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
11553 (create_mul_imm_cand): Likewise.
11554 (create_add_ssa_cand): Exclude phi base candidates.
11555 (create_add_imm_cand): Likewise.
11556 (slsr_process_cast): Likewise.
11557 (slsr_process_copy): Likewise.
11558 (find_candidates_in_block): Handle phi candidates.
11559 (dump_candidate): Likewise.
11560 (unconditional_cands): Delete.
11561 (unconditional_cands_with_known_stride_p): Delete.
11562 (phi_dependent_cand_p): New.
11563 (cand_increment): Handle phi-dependent candidates.
11564 (replace_dependent): Delete.
11565 (replace_mult_candidate): New.
11566 (replace_unconditional_candidate): New.
11567 (incr_vec_index): Move to avoid forward reference.
11568 (create_add_on_incoming_edge): New.
11569 (create_phi_basis): New.
11570 (replace_dependents): Delete.
11571 (replace_conditional_candidate): New.
11572 (phi_add_costs): New.
11573 (replace_uncond_cands_and_profitable_phis): New.
11574 (record_increment): Handle phi adjustments.
11575 (record_phi_increments): New.
11576 (record_increments): Handle phi adjustments.
11577 (phi_incr_cost): New.
11578 (lowest_cost_path): Handle phis.
11579 (total_savings): Likewise.
11580 (analyze_increments): Likewise.
11581 (ncd_with_phi): New.
11582 (ncd_of_cand_and_phis): New.
11583 (nearest_common_dominator_for_cands): Handle phi increments.
11584 (all_phi_incrs_profitable): New.
11585 (replace_profitable_candidates): Handle phi-dependent candidates.
11586 (analyze_candidates_and_replace): Likewise.
11587
68f073d4
TJ
115882013-05-03 Teresa Johnson <tejohnson@google.com>
11589
11590 PR bootstrap/57154
11591 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
11592 do not exceed REG_BR_PROB_BASE.
11593
a4ee7cb9
JL
115942013-05-03 Jeff Law <law@redhat.com>
11595
ade67f70 11596 PR tree-optimization/57144
a4ee7cb9
JL
11597 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
11598 operand of the condition will bit into the new type when eliminating
11599 a cast feeding a condition.
11600
47954c4e
JJ
116012013-05-03 Jakub Jelinek <jakub@redhat.com>
11602
11603 PR rtl-optimization/57130
3095685e
UB
11604 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
11605 of COMPARE as in_code to the recursive call if needed.
47954c4e 11606
3c21604f
UB
116072013-05-03 Uros Bizjak <ubizjak@gmail.com>
11608
11609 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
11610 (enabled): Handle new members.
11611 * config/i386/sse.md (*vec_concatv2si): Merge from
11612 *vec_concatv2si_sse2 and vec_concatv2si_sse.
11613 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
11614
12211b99 116152013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR
11616
11617 PR tree-optimization/57027
11618 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
11619 for fnms opportunity, check we got the prerequisite kind
11620 of tree / gimple before using accessor functions.
11621
07bfc9ec
RB
116222013-05-03 Richard Biener <rguenther@suse.de>
11623
11624 * double-int.h (lshift): New overload without precision
11625 and arith argument.
11626 (operator *=, operator +=, operator -=): Move ...
11627 * double-int.c (operator *=, operator +=, operator -=): ... here
11628 and implement more efficiently.
11629 (mul_double_with_sign): Remove.
11630 (lshift_double): Adjust to take unsinged shift argument, push
11631 dispatching code to callers.
11632 (mul_double_wide_with_sign): Add early out for callers that
11633 are not interested in high parts or overflow.
11634 (lshift): New function.
11635 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
11636 dispatch code here.
11637 (lrotate, rrotate): Use logical shifts.
11638 * expr.c (get_inner_reference): Use lshift.
11639 * fixed-value.c (do_fixed_divide): Likewise.
11640 * tree-dfa.c (get_ref_base_and_extent): Likewise.
11641 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
11642 (indirect_refs_may_alias_p): Likewise.
11643 (stmt_kills_ref_p_1): Likewise.
11644
7769bb64
VP
116452013-05-03 Vidya Praveen <vidyapraveen@arm.com>
11646
11647 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
11648
1d0c8e5c
VP
116492013-05-03 Vidya Praveen <vidyapraveen@arm.com>
11650
11651 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
11652 scalar form of FABD instruction.
11653
f15feaf9
VM
116542013-05-02 Vladimir Makarov <vmakarov@redhat.com>
11655
11656 * lra-constraints.c (process_alt_operands): Add checking alt
11657 number to choose the best alternative.
11658
d90e76d4
RB
116592013-05-02 Richard Biener <rguenther@suse.de>
11660
11661 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
11662 bitmap and its handling.
11663 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
11664
9f8e7a96
RB
116652013-05-02 Richard Biener <rguenther@suse.de>
11666
11667 PR middle-end/57140
11668 * tree-inline.c (copy_loops): Properly handle removed loops.
11669 (copy_cfg_body): Mark destination loops for fixup if source
11670 loops needed fixup.
11671
f3a81b39
GY
116722013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
11673
11674 PR target/56732
11675 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
11676 generating simple_return for naked functions.
11677
7b920a9a
MJ
116782013-05-02 Martin Jambor <mjambor@suse.cz>
11679
11680 PR middle-end/56988
11681 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
11682 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
11683 flags match.
11684 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
11685 ipa_agg_replacement_value structures.
11686 (known_aggs_to_agg_replacement_list): Likewise.
11687 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
11688 (read_agg_replacement_chain): Likewise.
11689 (ipcp_transform_function): Also check that by_ref flags match.
11690
2c41c19d
RB
116912013-05-02 Richard Biener <rguenther@suse.de>
11692
11693 * graphds.h (struct graph): Add obstack member.
11694 * graphds.c (new_graph): Initialize obstack and allocate
11695 vertices from it.
11696 (add_edge): Allocate edge from the obstack.
3c21604f 11697 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 11698
8b47039c
TJ
116992013-05-02 Teresa Johnson <tejohnson@google.com>
11700
11701 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
11702 divides.
11703 * cfg.c (update_bb_profile_for_threading): Ditto.
11704 * tree-inline.c (copy_bb): Ditto.
11705 (copy_edges_for_bb): Ditto.
11706 (initialize_cfun): Ditto.
11707 (copy_cfg_body): Ditto.
11708 (expand_call_inline): Ditto.
11709 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
11710 (estimate_node_size_and_time): Ditto.
11711 (inline_merge_summary): Ditto.
11712 * cgraphclones.c (cgraph_clone_edge): Ditto.
11713 (cgraph_clone_node): Ditto.
11714 * sched-rgn.c (compute_dom_prob_ps): Ditto.
11715 (compute_trg_info): Ditto.
11716
da65928c
IB
117172013-05-02 Ian Bolton <ian.bolton@arm.com>
11718
11719 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
11720 S reg when fp attribute set.
11721 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
11722
add3c965
IB
117232013-05-02 Ian Bolton <ian.bolton@arm.com>
11724
11725 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
11726 New pattern.
11727 (*and_one_cmplsi3_compare0_uxtw): Likewise.
11728 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
11729 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
11730
fdd43ac4
RB
117312013-05-02 Richard Biener <rguenther@suse.de>
11732
11733 * tree-scalar-evolution.c (scev_info_hasher): Remove.
11734 (struct instantiate_cache_entry): New type.
11735 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
11736 (struct instantiate_cache_type): New type.
11737 (set_instantiated_value, get_instantiated_value): Remove.
11738 (get_instantiated_value_entry): New function.
11739 (instantiate_scev_name): Use the new cache and adjust.
11740 (instantiate_scev_poly): Adjust.
11741 (instantiate_scev_binary): Likewise.
11742 (instantiate_array_ref): Likewise.
11743 (instantiate_scev_convert): Likewise.
11744 (instantiate_scev_not): Likewise.
11745 (instantiate_scev_3): Likewise.
11746 (instantiate_scev_2): Likewise.
11747 (instantiate_scev_r): Likewise.
11748 (instantiate_scev): Likewise.
11749 (resolve_mixers): Likewise.
11750
36ff9dfb
VM
117512013-05-01 Vladimir Makarov <vmakarov@redhat.com>
11752
11753 PR target/57091
11754 * lra-constraints.c (best_small_class_operands_num): Remove.
11755 (process_alt_operands): Remove small_class_operands_num. Take
11756 small classes operands into losers and only if the operand is not
11757 matched. Modify debugging output.
11758 (curr_insn_transform): Remove best_small_class_operands_num.
11759 Print insn name.
11760
36054fab
JG
117612013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11762
11763 * config/aarch64/aarch64-builtins.c
11764 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
11765 * config/aarch64/aarch64-simd-builtins.def
11766 (reduc_splus_): Add new modes.
11767 (reduc_uplus_): New.
11768 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
11769 (reduc_uplus_v4sf): Likewise.
11770 (reduc_splus_v4sf): Likewise.
11771 (aarch64_addv<mode>): Likewise.
11772 (reduc_uplus_<mode>): Likewise.
11773 (reduc_splus_<mode>): Likewise.
11774 (aarch64_addvv2di): Likewise.
11775 (reduc_uplus_v2di): Likewise.
11776 (reduc_splus_v2di): Likewise.
11777 (aarch64_addvv2si): Likewise.
11778 (reduc_uplus_v2si): Likewise.
11779 (reduc_splus_v2si): Likewise.
11780 (reduc_<sur>plus_<mode>): New.
11781 (reduc_<sur>plus_v2di): Likewise.
11782 (reduc_<sur>plus_v2si): Likewise.
11783 (reduc_<sur>plus_v4sf): Likewise.
11784 (aarch64_addpv4sf): Likewise.
11785 * config/aarch64/arm_neon.h
11786 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
11787 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
11788 add UNSPEC_SADDV, UNSPEC_UADDV.
11789 (SUADDV): New.
11790 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
11791
6dce23a8
JG
117922013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11793
11794 * config/aarch64/arm_neon.h
11795 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
11796
1598945b
JG
117972013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11798
11799 * config/aarch64/aarch64-builtins
11800 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
11801
998eaf97
JG
118022013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11803
11804 * config/aarch64/aarch64-simd-builtins.def
11805 (reduc_smax_): New.
11806 (reduc_smin_): Likewise.
11807 (reduc_umax_): Likewise.
11808 (reduc_umin_): Likewise.
11809 (reduc_smax_nan_): Likewise.
11810 (reduc_smin_nan_): Likewise.
11811 (fmax): Remove.
11812 (fmin): Likewise.
11813 (smax): Update for V2SF, V4SF and V2DF modes.
11814 (smin): Likewise.
11815 (smax_nan): New.
11816 (smin_nan): Likewise.
11817 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
11818 (<su><maxmin><mode>3): ...This, refactor.
11819 (s<maxmin><mode>3): New.
11820 (<maxmin_uns><mode>3): Likewise.
11821 (reduc_<maxmin_uns>_<mode>): Refactor.
11822 (reduc_<maxmin_uns>_v4sf): Likewise.
11823 (reduc_<maxmin_uns>_v2si): Likewise.
11824 (aarch64_<fmaxmin><mode>: Remove.
11825 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
11826 new builtin names.
11827 (vmin<q>_f<32,64>): Likewise.
11828 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
11829 (FMAXMIN): New.
11830 (su): Add mappings for smax, smin, umax, umin.
11831 (maxmin): New.
11832 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
11833 (FMAXMIN): Rename as...
11834 (FMAXMIN_UNS): ...This.
11835 (maxminv): Remove.
11836 (fmaxminv): Likewise.
11837 (fmaxmin): Likewise.
11838 (maxmin_uns): New.
11839 (maxmin_uns_op): Likewise.
11840
bd11644e
JG
118412013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11842
11843 * config/aarch64/arm_neon.h
11844 (vac<ge, gt><sd>_f<32, 64>): Rename to...
11845 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
11846 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
11847
75dd5ace
JG
118482013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11849
11850 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
11851 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
11852
7c19979f
JG
118532013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11854
11855 * config/aarch64/aarch64-simd.md
11856 (vcond<mode>_internal): Handle special cases for constant masks.
11857 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
11858 (vcondu<mode><mode>): Likewise.
11859 (vcond<v_cmp_result><mode>): New.
11860
bb60efd9
JG
118612013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11862
11863 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
11864 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
11865 * config/aarch64/aarch64-simd-builtins.def
11866 (cmeq): Update to BUILTIN_VALLDI.
11867 (cmgt): Likewise.
11868 (cmge): Likewise.
11869 (cmle): Likewise.
11870 (cmlt): Likewise.
11871 * config/aarch64/arm_neon.h
11872 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
11873 to builtins or C as appropriate.
11874
889b9412
JG
118752013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
11876
11877 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
11878 (cmgeu): ...This.
11879 (cmhi): Rename to...
11880 (cmgtu): ...This.
11881 * config/aarch64/aarch64-simd.md
11882 (simd_mode): Add SF.
11883 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
11884 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
11885 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
11886 (cstore<mode>_neg): ...This.
11887 * config/aarch64/iterators.md
11888 (VALLF): new.
11889 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
11890 (COMPARISONS): New.
11891 (UCOMPARISONS): Likewise.
11892 (optab): Add missing comparisons.
11893 (n_optab): New.
11894 (cmp_1): Likewise.
11895 (cmp_2): Likewise.
11896 (CMP): Likewise.
11897 (cmp): Remove.
11898 (VCMP_S): Likewise.
11899 (VCMP_U): Likewise.
11900 (V_cmp_result): Add DF, SF modes.
11901 (v_cmp_result): Likewise.
11902 (v): Likewise.
11903 (vmtype): Likewise.
11904 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
11905
0a7dbb76
GY
119062013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
11907
11908 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
11909 define_insn to define_insn_and_split.
11910 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
11911 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
11912 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
11913 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
11914 (thumb2_negscc): Likewise.
11915
fb614ca6
GY
119162013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11917
11918 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
11919
9e64a0bf
GY
119202013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11921
11922 * config/arm/thumb2.md: Remove trailing whitespaces.
11923
d6b28156
RS
119242013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11925
11926 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
11927 Use gen_int_mode rather than GEN_INT.
11928
f91674c3
L
119292013-04-30 H.J. Lu <hongjiu.lu@intel.com>
11930
b0dec607 11931 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L
11932 debug_gimple_stmt.
11933
3551257c
RB
119342013-04-30 Richard Biener <rguenther@suse.de>
11935
11936 PR middle-end/57122
3c21604f 11937 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 11938
4c1aff1c
RB
119392013-04-30 Richard Biener <rguenther@suse.de>
11940
11941 PR middle-end/57107
11942 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
11943
54b8379a
AB
119442013-04-30 Andrey Belevantsev <abel@ispras.ru>
11945
11946 PR rtl-optimization/56957
11947 PR rtl-optimization/57105
54b8379a
AB
11948 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
11949 variable. Use just INSN_UID for determining whether an insn
11950 should be only disconnected from the insn stream.
11951 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
11952
abc27962
JJ
119532013-04-30 Jakub Jelinek <jakub@redhat.com>
11954
11955 PR tree-optimization/57104
11956 * tsan.c (instrument_expr): Don't instrument accesses to
11957 DECL_HARD_REGISTER VAR_DECLs.
11958
0fc822d0
RB
119592013-04-30 Richard Biener <rguenther@suse.de>
11960
11961 * function.h (loops_for_fn): New inline function.
11962 (set_loops_for_fn): Likewise.
11963 * cfgloop.h (place_new_loop): Add struct function parameter.
11964 (get_loop): Likewise.
11965 (get_loops): Likewise.
11966 (number_of_loops): Likewise.
11967 (fel_next): Adjust.
11968 (fel_init): Likewise.
11969 * cfg.c (get_loop_copy): Adjust.
11970 * cfgloop.c (flow_loops_dump): Likewise.
11971 (record_loop_exits): Likewise.
11972 (verify_loop_structure): Likewise.
11973 * cfgloopanal.c (mark_irreducible_loops): Likewise.
11974 (estimate_reg_pressure_cost): Likewise.
11975 (mark_loop_exit_edges): Likewise.
11976 * cfgloopmanip.c (place_new_loop): Likewise.
11977 (add_loop): Likewise.
11978 (duplicate_loop): Likewise.
11979 * graph.c (draw_cfg_nodes): Likewise.
11980 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
11981 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
11982 (extract_affine_chrec): Likewise.
11983 (build_scop_iteration_domain): Likewise.
11984 * graphite.c (graphite_initialize): Likewise.
11985 * ira-build.c (create_loop_tree_nodes): Likewise.
11986 (more_one_region_p): Likewise.
11987 (rebuild_regno_allocno_maps): Likewise.
11988 (mark_loops_for_removal): Likewise.
11989 (mark_all_loops_for_removal): Likewise.
11990 (remove_unnecessary_regions): Likewise.
11991 (ira_build): Likewise.
11992 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
11993 * loop-init.c (fix_loop_structure): Likewise.
11994 (gate_rtl_move_loop_invariants): Likewise.
11995 (gate_rtl_unswitch): Likewise.
11996 (gate_rtl_unroll_and_peel_loops): Likewise.
11997 (rtl_doloop): Likewise.
11998 * lto-streamer-in.c (input_cfg): Likewise.
11999 * lto-streamer-out.c (output_cfg): Likewise.
12000 * modulo-sched.c (sms_schedule): Likewise.
12001 * predict.c (tree_estimate_probability): Likewise.
12002 (tree_estimate_probability_driver): Likewise.
12003 (estimate_loops): Likewise.
12004 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
12005 (move_sese_region_to_fn): Likewise.
12006 (debug_loop_num): Likewise.
12007 * tree-chrec.c (chrec_evaluate): Likewise.
12008 (hide_evolution_in_other_loops_than_loop): Likewise.
12009 (chrec_component_in_loop_num): Likewise.
12010 (reset_evolution_in_loop): Likewise.
12011 (evolution_function_is_invariant_rec_p): Likewise.
12012 * tree-if-conv.c (main_tree_if_conversion): Likewise.
12013 * tree-inline.c (copy_loops): Likewise.
12014 (copy_cfg_body): Likewise.
12015 (tree_function_versioning): Likewise.
12016 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
12017 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
12018 Likewise.
12019 (add_to_evolution_1): Likewise.
12020 (scev_const_prop): Likewise.
12021 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
12022 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
12023 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
12024 (tree_ssa_lim_initialize): Likewise.
12025 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
12026 (verify_loop_closed_ssa): Likewise.
12027 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
12028 (tree_ssa_loop_im): Likewise.
12029 (tree_ssa_loop_unswitch): Likewise.
12030 (tree_vectorize): Likewise.
12031 (check_data_deps): Likewise.
12032 (tree_ssa_loop_ivcanon): Likewise.
12033 (tree_ssa_loop_bounds): Likewise.
12034 (tree_complete_unroll): Likewise.
12035 (tree_complete_unroll_inner): Likewise.
12036 (tree_parallelize_loops): Likewise.
12037 (tree_ssa_loop_prefetch): Likewise.
12038 (tree_ssa_loop_ivopts): Likewise.
12039 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
12040 * tree-vectorizer.c (vectorize_loops): Likewise.
12041
37953bd3
MF
120422013-04-29 Mike Frysinger <vapier@gentoo.org>
12043
12044 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
12045 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
12046 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
12047 with EABI_LINK_SPEC.
12048
f9ed28db
UB
120492013-04-29 Uros Bizjak <ubizjak@gmail.com>
12050
12051 PR target/44578
12052 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
12053 alternative.
12054
deca73f5
VM
120552013-04-29 Vladimir Makarov <vmakarov@redhat.com>
12056
12057 PR target/57097
37953bd3 12058 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM
12059 using memory for pseudos. Print cost dump for alternatives.
12060 Modify cost values for conflicts with early clobbers.
12061 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
12062
fdca7d03
UB
120632013-04-29 Uros Bizjak <ubizjak@gmail.com>
12064
12065 PR target/57098
12066 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
12067
0b064172
IB
120682013-04-29 Ian Bolton <ian.bolton@arm.com>
12069
12070 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
12071 from/to S register.
12072 (movdi_aarch64): Support LDR/STR from/to D register.
12073
473cec55
IB
120742013-04-29 Ian Bolton <ian.bolton@arm.com>
12075
12076 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
12077 or higher by default.
12078
a6f30e66
RB
120792013-04-29 Richard Biener <rguenther@suse.de>
12080
12081 PR middle-end/57075
12082 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
12083 even if not adding abnormal edges for calls that can make
12084 abnormal gotos.
12085
0c2b2040
RB
120862013-04-29 Richard Biener <rguenther@suse.de>
12087
12088 PR middle-end/57103
12089 * tree-cfg.c (move_stmt_op): Fix condition under which to update
12090 TREE_BLOCK.
12091 (move_stmt_r): Remove redundant checking.
12092
f41f80f9
TJ
120932013-04-29 Teresa Johnson <tejohnson@google.com>
12094
12095 PR bootstrap/57077
12096 * basic-block.h (apply_scale): New function.
12097 (apply_probability): Use apply_scale.
12098 * gimple-streamer-in.c (input_bb): Ditto.
12099 * lto-streamer-in.c (input_cfg): Ditto.
12100 * lto-cgraph.c (merge_profile_summaries): Ditto.
12101 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 12102 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ
12103 (copy_edges_for_bb): Ditto.
12104 (copy_cfg_body): Ditto.
12105
315bbd2e
TV
121062013-04-29 Tom de Vries <tom@codesourcery.com>
12107
12108 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
12109 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
12110 (tail_merge_optimize): Handle current_loops == NULL.
12111
ebbd90d8
JL
121122013-04-26 Jeff Law <law@redhat.com>
12113
12114 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
12115 (simplify_cond_using_ranges): Generalize code to simplify
12116 COND_EXPRs where one argument is a constant and the other
12117 is an SSA_NAME created by an integral type conversion.
12118
8b9b57eb
KT
121192013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12120
12121 * config/arm/arm.md (store_minmaxsi): Use only when
12122 optimize_insn_for_size_p.
12123
9498e5dc
CB
121242013-04-29 Christian Bruel <christian.bruel@st.com>
12125
12126 PR target/57108
12127 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
12128
6093bc06
RB
121292013-04-29 Richard Biener <rguenther@suse.de>
12130
12131 PR middle-end/57089
fdca7d03
UB
12132 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
12133 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB
12134 (expand_omp_for_generic): Properly add loops.
12135 (expand_omp_for_static_nochunk): Likewise.
12136 (expand_omp_for_static_chunk): Likewise.
12137 (expand_omp_for): For the degenerate case fixup loops.
12138 (expand_omp_sections): Fix default bb placement in loops.
12139 (expand_omp_atomic_pipeline): Properly add loops.
12140
84aacbfd
KT
121412013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12142
12143 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
12144
8dee4479
TV
121452013-04-29 Tom de Vries <tom@codesourcery.com>
12146
12147 * tree-ssa-tail-merge.c: Update header comment.
12148
47934dc4
JG
121492013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12150
12151 * config/aarch64/arm_neon.h
12152 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
12153 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
12154 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
12155 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
12156 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
12157 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
12158 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
12159 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
12160
384be29f
JG
121612013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12162
12163 * config/aarch64/aarch64-simd.md
12164 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
12165 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
12166 fix_trunc, fixuns_trunc.
12167 (ftrunc<VDQF:mode>2): New.
12168 * config/aarch64/iterators.md (optab): Add fix, fixuns.
12169 (fix_trunc_optab): New.
12170
0386b123
JG
121712013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12172
12173 * config/aarch64/aarch64-builtins.c
12174 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
12175 iceilf, lround, iroundf.
12176
00fcb892
UB
121772013-04-29 Uros Bizjak <ubizjak@gmail.com>
12178
12179 PR target/54349
12180 * config/i386/i386.h (enum ix86_tune_indices)
12181 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
12182 New, split from X86_TUNE_INTER_UNIT_MOVES.
12183 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
12184 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
12185 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
12186 (TARGET_INTER_UNIT_MOVES): Remove.
12187 * config/i386/i386.c (initial_ix86_tune_features): Update.
12188 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
12189 (ix86_expand_convert_uns_didf_sse): Use
12190 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12191 (ix86_expand_vector_init_one_nonzero): Ditto.
12192 (ix86_expand_vector_init_interleave): Ditto.
12193 (inline_secondary_memory_needed): Return true for moves from SSE class
12194 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
12195 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
12196 * config/i386/constraints.md (Yi, Ym): Depend on
12197 TARGET_INTER_UNIT_MOVES_TO_VEC.
12198 (Yj, Yn): New constraints.
12199 * config/i386/i386.md (*movdi_internal): Change constraints of
12200 operand 1 from Yi to Yj and from Ym to Yn.
12201 (*movsi_internal): Ditto.
12202 (*movdf_internal): Ditto.
12203 (*movsf_internal): Ditto.
12204 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
12205 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
12206 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
12207 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
12208 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
12209 * config/i386/sse.md (movdi_to_sse): Ditto.
12210 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
12211 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
12212 TARGET_INTER_UNIT_MOVES.
12213 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
12214 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
12215 instead of TARGET_INTER_UNIT_MOVES.
12216 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
12217 operand 1 from Yi to Yj and from Ym to Yn.
12218
4c871069
JG
122192013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12220
12221 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
12222 (float_truncate_hi_): Likewise.
12223 (float_extend_lo_): Likewise.
12224 (float_truncate_lo_): Likewise.
12225 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
12226 (aarch64_float_extend_lo_v2df): Likewise.
12227 (vec_unpacks_hi_v4sf): Likewise.
12228 (aarch64_float_truncate_lo_v2sf): Likewise.
12229 (aarch64_float_truncate_hi_v4sf): Likewise.
12230 (vec_pack_trunc_v2df): Likewise.
12231 (vec_pack_trunc_df): Likewise.
12232
1709ff9b
JG
122332013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12234
12235 * config/aarch64/aarch64-builtins.c
12236 (aarch64_fold_builtin): Fold float conversions.
12237 * config/aarch64/aarch64-simd-builtins.def
12238 (floatv2si, floatv4si, floatv2di): New.
12239 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
12240 * config/aarch64/aarch64-simd.md
12241 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
12242 * config/aarch64/iterators.md (FLOATUORS): New.
12243 (optab): Add float, floatuns.
12244 (su_optab): Likewise.
12245
ce966824
JG
122462013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12247
12248 * config/aarch64/aarch64-builtins.c
12249 (aarch64_builtin_vectorized_function): Use new names for
12250 fcvt builtins.
12251 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
12252 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
12253 (fcvtzu): Split as...
12254 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
12255 (fcvtas): Split as...
12256 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
12257 (fcvtau): Split as...
12258 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
12259 (fcvtps): Split as...
12260 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
12261 (fcvtpu): Split as...
12262 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
12263 (fcvtms): Split as...
12264 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
12265 (fcvtmu): Split as...
12266 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
12267 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
12268 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
12269 (lfrintnusf, lfrintnudf): Likewise.
12270 * config/aarch64/aarch64-simd.md
12271 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
12272 define_insn.
12273 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
12274 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
12275 (fcvt_pattern): Likewise.
12276
b9de24fe
JG
122772013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12278
12279 * config/aarch64/aarch64-simd.md
12280 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
12281 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
12282
77a205be
JG
122832013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12284
12285 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
12286 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
12287 (vrnd<a,m,n,p>_f32): Implement using builtins.
12288 (vrnd<i,x><q>_f<32, 64>): New.
12289
0659ce6f
JG
122902013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
12291
12292 * config/aarch64/aarch64-builtins.c
12293 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
12294 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
12295 (frintz): Rename to...
12296 (btrunc): ...this.
12297 (frintp): Rename to...
12298 (ceil): ...this.
12299 (frintm): Rename to...
12300 (floor): ...this.
12301 (frinti): Rename to...
12302 (nearbyint): ...this.
12303 (frintx): Rename to...
12304 (rint): ...this.
12305 (frinta): Rename to...
12306 (round): ...this.
12307 * config/aarch64/aarch64-simd.md
12308 (aarch64_frint<frint_suffix><mode>): Delete.
12309 (<frint_pattern><mode>2): Convert to insn.
12310 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
12311 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
12312 (frint_pattern): Likewise.
12313 (frint_suffix): Likewise.
12314
ea78906a
RB
123152013-04-29 Richard Biener <rguenther@suse.de>
12316
12317 PR tree-optimization/57081
12318 * loop-init.c: Include tree-flow.h.
12319 (loop_optimizer_finalize): Free number of iteration estimates.
12320 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
12321
baee1763
JJ
123222013-04-29 Jakub Jelinek <jakub@redhat.com>
12323
94dc5332
JJ
12324 PR tree-optimization/57083
12325 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
12326 non-singleton shift count range, zero extend low_bound for uns case.
12327
baee1763
JJ
12328 * config/i386/predicates.md (general_vector_operand): New predicate.
12329 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
12330 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
12331 if they aren't nonimmediate operands. If their original values
12332 satisfy const_vector_equal_evenodd_p, don't shift them.
12333 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
12334 predicates. For the SSE4.1 case force operands[{1,2}] into registers
12335 if not nonimmediate_operand.
12336 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
12337 instead of register_operand.
12338 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
12339
a9073727 123402013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB
12341
12342 * stor-layout.c (finalize_size_functions): Allocate a structure and
12343 reset cfun before dumping the functions.
12344
ba8011e6
JJ
123452013-04-27 Jakub Jelinek <jakub@redhat.com>
12346
d6fde69e
JJ
12347 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
12348
ba8011e6
JJ
12349 PR target/56866
12350 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
12351 use xop_pmacsdqh if uns_p.
12352 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
12353 the immediate rotate count.
12354
2c62cbaa
VM
123552013-04-26 Vladimir Makarov <vmakarov@redhat.com>
12356
12357 * rtl.h (struct rtx_def): Add comment for field jump.
12358 (LRA_SUBREG_P): New macro.
12359 * recog.c (register_operand): Check LRA_SUBREG_P.
12360 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12361 stack frame.
12362 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12363 (lra_final_code_change): Skip subreg change for operators.
12364 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12365 if there are no operand changes.
12366 * lra-constraints.c (curr_insn_set): New.
12367 (match_reload): Set LRA_SUBREG_P.
12368 (emit_spill_move): Ditto.
12369 (check_and_process_move): Use curr_insn_set. Process only single
12370 set insns. Don't initialize sec_mem_p and change_p.
12371 (simplify_operand_subreg): Use LRA_SUBREG_P.
12372 (reg_in_class_p): New function.
12373 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12374 of #ifdef. Add code to remove cycling.
12375 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12376 non-null disp. Reload inner instead of disp when base and index
12377 are null. Try to put lo_sum into register.
12378 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12379 (check_and_process_move): Move code for move cost check to
12380 simple_move_p. Remove equiv_substitution.
12381 (simple_move_p): New function.
12382 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12383 curr_insn_set. Call check_and_process_move only for single set
12384 insns. Use the new function. Move call of check_and_process_move
12385 after operand equiv substitution and address process.
12386
e7d764f3
JJ
123872013-04-26 Jakub Jelinek <jakub@redhat.com>
12388
12389 PR go/57045
12390 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
12391 with nonlocal goto receivers or returns twice calls, ignore
12392 unininitialized values from abnormal edges to nl goto receiver
12393 or returns twice call.
12394
41e10689
JJ
123952013-04-26 Jakub Jelinek <jakub@redhat.com>
12396
12397 PR tree-optimization/57051
12398 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
12399 and VEC_RSHIFT_EXPR if shift count is a multiple of element
12400 bitsize.
12401
d7ed20db
RB
124022013-04-26 Richard Biener <rguenther@suse.de>
12403
12404 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
12405 (expand_omp_taskreg): Likewise. Mark loops for fixup.
12406 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
12407 (fixup_loop_arrays_after_move): New function.
12408 (move_sese_region_to_fn): Properly outline the loop tree parts
12409 of the SESE region.
12410
df93505e
UB
124112013-04-26 Uros Bizjak <ubizjak@gmail.com>
12412
12413 * config/i386/i386.md (type, unit): Fix long lines.
12414
dd366ec3
RB
124152013-04-26 Richard Biener <rguenther@suse.de>
12416
12417 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
12418 (lto-streamer-out.o): Likewise.
12419 * cfgloop.c (init_loops_structure): Export, add struct function
12420 argument and adjust.
12421 (flow_loops_find): Adjust.
12422 * cfgloop.h (enum loop_estimation): Add EST_LAST.
12423 (init_loops_structure): Declare.
12424 * lto-streamer-in.c: Include cfgloop.h.
12425 (input_cfg): Input the loop tree.
12426 * lto-streamer-out.c: Include cfgloop.h.
12427 (output_cfg): Output the loop tree.
12428 (output_struct_function_base): Do not drop PROP_loops.
12429
a9e0d843
RB
124302013-03-26 Richard Biener <rguenther@suse.de>
12431
12432 * tree-cfg.c (execute_build_cfg): Build the loop tree.
12433 (pass_build_cfg): Provide PROP_loops.
12434 (move_sese_region_to_fn): Remove loops that are outlined into fn
12435 for now.
12436 * tree-inline.c: Include cfgloop.h.
12437 (initialize_cfun): Do not drop PROP_loops.
12438 (copy_loops): New function.
12439 (copy_cfg_body): Copy loop structure.
12440 (tree_function_versioning): Initialize destination loop tree.
12441 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
12442 (pass_parallelize_loops): Do IL verification.
12443 * loop-init.c (loop_optimizer_init): Fixup loops if required.
12444 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
12445 the CFG make sure we fixup loops as well.
12446 * tree-ssa-tail-merge.c: Include cfgloop.h.
12447 (replace_block_by): When merging loop latches mark loops for fixup.
12448 * lto-streamer-out.c (output_struct_function_base): Drop
12449 PROP_loops for now.
12450 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
12451 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
12452 * ipa-split.c: Include cfgloop.h.
12453 (split_function): Add the new return block to the loop tree root.
12454 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
12455 whether we have removed the forwarder block.
12456 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
12457 * cfgloop.h (place_new_loop): Declare.
12458 * cfgloopmanip.c (place_new_loop): Export.
12459 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
12460 (tree-switch-conversion.o): Likewise.
12461 (tree-complex.o): Likewise.
12462 (tree-inline.o): Likewise.
12463 (tree-ssa-tailmerge.o): Likewise.
12464 (ipa-split.o): Likewise.
12465 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
12466 (tree-ssa-copy.o): Likewise.
12467 * tree-switch-conversion.c: Include cfgloop.h
12468 (process_switch): If we emit a bit-test cascade, schedule loops
12469 for fixup.
12470 * tree-complex.c: Include cfgloop.h.
12471 (expand_complex_div_wide): Properly add new basic-blocks to loops.
12472 * asan.c: Include cfgloop.h.
12473 (create_cond_insert_point): Properly add new basic-blocks to
12474 loops, schedule loop fixup.
12475 * cfgloop.c (verify_loop_structure): Check that looks are not
12476 marked for fixup.
12477 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
12478 to loops.
12479 (expand_omp_for_generic): Likewise.
12480 (expand_omp_sections): Likewise.
12481 (expand_omp_atomic_pipeline): Schedule loops for fixup.
12482 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
12483 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
12484 is initialized, not when loops are present.
12485 * tree-parloops.c (parallelize_loops): Remove checking here.
12486 * passes.c (init_optimization_passes): Schedule a copy-propagation
12487 pass before complete unrolling of inner loops.
12488
e78e8a0b
JJ
124892013-04-26 Jakub Jelinek <jakub@redhat.com>
12490
a2e836b2
JJ
12491 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
12492 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
12493 (colorize_init): Add argument to _WIN32 version.
12494 * toplev.c: Include diagnostic-color.h.
12495 (process_options): Default to -fdiagnostics-color=auto if
12496 GCC_COLORS env var is in the environment.
12497 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB
12498 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
12499 env var is in the environment, the default is auto rather than never.
a2e836b2 12500
e78e8a0b
JJ
12501 * diagnostic.h (file_name_as_prefix): Add context argument.
12502 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
12503 the string as locus.
12504 * langhooks.c (lhd_print_error_function): Adjust caller.
12505
013e5ef9
LC
125062013-04-25 Lawrence Crowl <crowl@google.com>
12507
12508 * var-tracking.c (shared_hash_def::htab):
12509 Change type to hash_table. Update dependent calls and types.
12510
4a8fb1a1
LC
125112013-04-25 Lawrence Crowl <crowl@google.com>
12512
12513 * Makefile.in: Update as needed below.
12514
12515 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
12516 Move declaration to after the type's method definitons.
12517
12518 * attribs.c (htab_t scoped_attributes::attribute_hash):
12519 Change type to hash_table. Update dependent calls and types.
12520
12521 * bitmap.c (htab_t bitmap_desc_hash):
12522 Change type to hash_table. Update dependent calls and types.
12523
12524 * cselib.c (htab_t cselib_hash_table):
12525 Change type to hash_table. Update dependent calls and types.
12526
12527 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
12528 (hash_string_slot_node): Move implementation into lto-streamer.h
12529 struct string_slot_hasher.
12530 (eq_string_slot_node): Likewise.
12531
12532 * data-streamer-out.c: Update output_block::string_hash_table
12533 dependent calls and types.
12534
12535 * dwarf2cfi.c (htab_t trace_index):
12536 Change type to hash_table. Update dependent calls and types.
12537
12538 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
12539 Change type to hash_table. Update dependent calls and types.
12540 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
12541 (htab_t optimize_external_refs::map): Likewise.
12542 (htab_t output_comp_unit::extern_map): Likewise.
12543 (htab_t output_comdat_type_unit::extern_map): Likewise.
12544 (htab_t output_macinfo::macinfo_htab): Likewise.
12545 (htab_t optimize_location_lists::htab): Likewise.
12546 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
12547
12548 * except.c (htab_t ehspec_hash_type):
12549 Change type to hash_table. Update dependent calls and types.
12550 (assign_filter_values::ttypes): Likewise.
12551 (assign_filter_values::ehspec): Likewise.
12552 (sjlj_assign_call_site_values::ar_hash): Likewise.
12553 (convert_to_eh_region_ranges::ar_hash): Likewise.
12554
12555 * gcse.c (htab_t pre_ldst_table):
12556 Change type to hash_table. Update dependent calls and types.
12557
12558 * ggc-common.c (htab_t saving_htab):
12559 Change type to hash_table. Update dependent calls and types.
12560 (htab_t loc_hash): Likewise.
12561 (htab_t ptr_hash): Likewise.
12562 (call_count): Rename ggc_call_count.
12563 (call_alloc): Rename ggc_call_alloc.
12564 (loc_descriptor): Rename make_loc_descriptor.
12565 (add_statistics): Rename ggc_add_statistics.
12566
12567 * ggc-common.c (saving_htab):
12568 Change type to hash_table. Update dependent calls and types.
12569
12570 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
12571 (push_gimplify_context): Likewise.
12572 (pop_gimplify_context): Likewise.
12573 (struct gimple_temp_hash_elt): Added.
12574 (struct gimplify_hasher): Likewise.
12575 (struct gimplify_ctx.temp_htab):
12576 Change type to hash_table. Update dependent calls and types.
12577
12578 * gimple-fold.c: Include gimplify-ctx.h.
12579
12580 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
12581 Change type to hash_table. Update dependent calls and types.
12582 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
12583 avoid potential global name collision.
12584
12585 * gimplify.c: Include gimplify-ctx.h.
12586 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
12587 (htab_t gimplify_ctx::temp_htab):
12588 Update dependent calls and types for new type hash_table.
12589 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
12590 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
12591
12592 * gimplify-ctx.h: New.
12593 (struct gimple_temp_hash_elt): Move from gimplify.c.
12594 (class gimplify_hasher): New.
12595 (struct gimplify_ctx): Move from gimple.h.
12596 (htab_t gimplify_ctx::temp_htab):
12597 Change type to hash_table. Update dependent calls and types.
12598
12599 * graphite-clast-to-gimple.c: Include graphite-htab.h.
12600 (htab_t ivs_params::newivs_index):
12601 Change type to hash_table. Update dependent calls and types.
12602 (htab_t ivs_params::params_index): Likewise.
12603 (htab_t print_generated_program::params_index): Likewise.
12604 (htab_t gloog::newivs_index): Likewise.
12605 (htab_t gloog::params_index): Likewise.
12606
12607 * graphite.c: Include graphite-htab.h.
12608 4htab_t graphite_transform_loops::bb_pbb_mapping):
12609 Change type to hash_table. Update dependent calls and types.
12610
12611 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
12612 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
12613 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
12614
12615 * graphite-dependences.c: Include graphite-htab.h.
12616 (loop_is_parallel_p): Change hash table type of parameter.
12617
12618 * graphite-htab.h: New.
12619 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
12620 (extern find_pbb_via_hash): Move from graphite-poly.h.
12621 (extern loop_is_parallel_p): Move from graphite-poly.h.
12622 (extern get_loop_body_pbbs): Move from graphite-poly.h.
12623
12624 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
12625 (extern loop_is_parallel_p): Move to graphite-htab.h.
12626 (extern get_loop_body_pbbs): Move to graphite-htab.h.
12627
12628 * haifa-sched.c (htab_t delay_htab):
12629 Change type to hash_table. Update dependent calls and types.
12630 (htab_t delay_htab_i2): Likewise.
12631
12632 * ira-color.c (htab_t allocno_hard_regs_htab):
12633 Change type to hash_table. Update dependent calls and types.
12634
12635 * ira-costs.c (htab_t cost_classes_htab):
12636 Change type to hash_table. Update dependent calls and types.
12637
12638 * loop-invariant.c (htab_t merge_identical_invariants::eq):
12639 Change type to hash_table. Update dependent calls and types.
12640
12641 * loop-iv.c (htab_t bivs):
12642 Change type to hash_table. Update dependent calls and types.
12643
12644 * loop-unroll.c (htab_t opt_info::insns_to_split):
12645 Change type to hash_table. Update dependent calls and types.
12646 (htab_t opt_info::insns_with_var_to_expand): Likewise.
12647
12648 * lto-streamer.h (struct string_slot): Move from data-streamer.h
12649 (struct string_slot_hasher): New.
12650 (htab_t output_block::string_hash_table):
12651 Change type to hash_table. Update dependent calls and types.
12652
12653 * lto-streamer-in.c (freeing_string_slot_hasher): New.
12654 (htab_t file_name_hash_table):
12655 Change type to hash_table. Update dependent calls and types.
12656
12657 * lto-streamer-out.c: Update output_block::string_hash_table dependent
12658 calls and types.
12659
12660 * lto-streamer.c (htab_t tree_htab):
12661 Change type to hash_table. Update dependent calls and types.
12662
12663 * omp-low.c: Include gimplify-ctx.h.
12664
12665 * passes.c (htab_t name_to_pass_map):
12666 Change type to hash_table. Update dependent calls and types.
12667 (pass_traverse): Rename to passes_pass_traverse.
12668
12669 * plugin.c (htab_t event_tab):
12670 Change type to hash_table. Update dependent calls and types.
12671
12672 * postreload-gcse.c (htab_t expr_table):
12673 Change type to hash_table. Update dependent calls and types.
12674 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
12675
12676 * sese.c (debug_rename_map_1): Make extern.
12677 (htab_t copy_bb_and_scalar_dependences::rename_map):
12678 Change type to hash_table. Update dependent calls and types.
12679
12680 * sese.h (extern debug_rename_map): Move to .c file.
12681
12682 * store-motion.c (htab_t store_motion_mems_table):
12683 Change type to hash_table. Update dependent calls and types.
12684
12685 * trans-mem.c (htab_t tm_new_mem_hash):
12686 Change type to hash_table. Update dependent calls and types.
12687
12688 * tree-browser.c (htab_t TB_up_ht):
12689 Change type to hash_table. Update dependent calls and types.
12690
12691 * tree-cfg.c (htab_t discriminator_per_locus):
12692 Change type to hash_table. Update dependent calls and types.
12693
12694 * tree-complex.c: Include tree-hasher.h
12695 (htab_t complex_variable_components):
12696 Change type to hash_table. Update dependent calls and types.
12697
12698 * tree-eh.c (htab_t finally_tree):
12699 Change type to hash_table. Update dependent calls and types.
12700
12701 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
12702 struct int_tree_hasher.
12703 (extern int_tree_map_eq): Likewise.
12704 (uid_decl_map_hash): Removed.
12705 (extern decl_tree_map_eq): Likewise.
12706
12707 * tree-hasher.h: New.
12708 (struct int_tree_hasher): New.
12709 (typedef int_tree_htab_type): New.
12710
12711 * tree-inline.c: Include gimplify-ctx.h.
12712
12713 * tree-mudflap.c: Include gimplify-ctx.h.
12714
12715 * tree-parloops.c: Include tree-hasher.h.
12716 (htab_t eliminate_local_variables_stmt::decl_address):
12717 Change type to hash_table. Update dependent calls and types.
12718 (htab_t separate_decls_in_region::decl_copies): Likewise.
12719
12720 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
12721 Change type to hash_table. Update dependent calls and types.
12722
12723 * tree-sra.c (candidates):
12724 Change type to hash_table. Update dependent calls and types.
12725
12726 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
12727 in tree-flow.h.
12728 (int_tree_map_hash): Likewise.
12729
12730 * tree-ssa-dom.c (htab_t avail_exprs):
12731 Change type to hash_table. Update dependent calls and types.
12732
12733 * tree-ssa-live.c (var_map_base_init::tree_to_index):
12734 Change type to hash_table. Update dependent calls and types.
12735
12736 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
12737 Change type to hash_table. Update dependent calls and types.
12738
12739 * tree-ssa-phiopt.c (seen_ssa_names):
12740 Change type to hash_table. Update dependent calls and types.
12741
12742 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
12743 Change type to hash_table. Update dependent calls and types.
12744
12745 * tree-ssa-uncprop.c (equiv):
12746 Change type to hash_table. Update dependent calls and types.
12747
c5a44004
JJ
127482013-04-25 Jakub Jelinek <jakub@redhat.com>
12749
12750 PR rtl-optimization/57003
12751 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
12752 call note_stores with kill_clobbered_value callback again after
12753 killing regs_invalidated_by_call.
12754
09962a4a
JG
127552013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12756
12757 * config/aarch64/aarch64-simd.md
12758 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
12759 (aarch64_simd_bsl<mode>): Likewise.
12760 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
12761
ea28bb0b
MP
127622013-04-25 Marek Polacek <polacek@redhat.com>
12763
12764 PR tree-optimization/57066
3c21604f 12765 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 12766
96659611
JG
127672013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
12768
12769 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
12770
9697e620
JG
127712013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12772
12773 * config/aarch64/aarch64-builtins.c
12774 (aarch64_fold_builtin): New.
12775 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
12776 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
12777 * config/aarch64/aarch64-simd-builtins.def (abs): New.
12778 * config/aarch64/arm_neon.h
12779 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
12780
0ac198d3
JG
127812013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12782 Tejas Belagod <tejas.belagod@arm.com>
12783
12784 * config/aarch64/aarch64-builtins.c
12785 (aarch64_gimple_fold_builtin): New.
12786 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
12787 * config/aarch64/aarch64-simd-builtins.def (addv): New.
12788 * config/aarch64/aarch64-simd.md (addpv4sf): New.
12789 (addvv4sf): Update.
12790 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
12791
58cff58c
N
127922013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12793
df93505e 12794 * config/aarch64/aarch64.md
58cff58c
N
12795 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
12796
7e0228bf
N
127972013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12798
12799 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
12800 (*ngcsi_uxtw): New pattern.
12801
5819f96f 128022013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 12803 Julian Brown <julian@codesourcery.com>
5819f96f
KT
12804
12805 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
12806 (TB_DREG): Add T_V4HF.
12807 (v4hf_UP): New macro.
12808 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 12809 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
12810 Handle initialisation of V4HF. Adjust initialisation of reinterpret
12811 built-ins.
df93505e 12812 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
12813 (arm_vector_mode_supported_p): Handle V4HF.
12814 (arm_mangle_map): Handle V4HFmode.
12815 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
12816 * config/arm/arm_neon_builtins.def: Add entries for
12817 vcvtv4hfv4sf, vcvtv4sfv4hf.
12818 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
12819 (neon_vcvtv4hfv4sf): Likewise.
12820 * config/arm/neon-gen.ml: Handle half-precision floating point
12821 features.
12822 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
12823 * config/arm/arm_neon.h: Regenerate.
12824 * config/arm/neon.ml (type elts): Add F16.
12825 (type vectype): Add T_float16x4, T_floatHF.
12826 (type vecmode): Add V4HF.
12827 (type features): Add Requires_FP_bit feature.
12828 (elt_width): Handle F16.
12829 (elt_class): Likewise.
12830 (elt_of_class_width): Likewise.
12831 (mode_of_elt): Refactor.
12832 (type_for_elt): Handle F16, fix error messages.
12833 (vectype_size): Handle T_float16x4.
12834 (vcvt_sh): New function.
12835 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
12836 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
12837 (string_of_mode): Handle V4HF.
12838 * doc/arm-neon-intrinsics.texi: Regenerate.
12839
1ef395e4
JG
128402013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
12841
12842 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
12843 format specifier in 'X' case.
12844
41c34e94
AM
128452013-04-25 Alan Modra <amodra@gmail.com>
12846
12847 PR target/57052
12848 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
12849 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
12850 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
12851 Repeat for many other rotate/shift and mask patterns using subregs.
12852 Name lshiftrt insns.
12853 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
12854 on WORDS_BIG_ENDIAN.
12855
b9a7eb5d
AM
128562013-04-25 Alan Modra <amodra@gmail.com>
12857
12858 * config.gcc: Support little-endian powerpc-linux targets.
12859 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
12860 (LINK_OS_LINUX_SPEC): Define.
12861 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
12862 Preserve MASK_LITTLE_ENDIAN.
12863 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
12864 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
12865 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
12866 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
12867 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
12868 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
12869 Correct fp word order for little-endian. Don't shift toc entries
12870 smaller than a word for little-endian.
12871 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
12872 (bswapdi2 splits): Correct low-part subreg for little-endian.
12873 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
12874 low/high where such is correct only for be.
12875 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
12876 little-endian for -mcall-aixdesc.
12877
87f73374
AM
128782013-04-25 Alan Modra <amodra@gmail.com>
12879
12880 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
12881 replace_equiv_address_nv.
12882
cabf91cd
AM
128832013-04-25 Alan Modra <amodra@gmail.com>
12884
12885 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
12886
0ae24cc8
VM
128872013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12888
12889 Revert:
12890 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12891 * rtl.h (struct rtx_def): ...
cabf91cd 12892
77bce07c
VM
128932013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12894
12895 PR rtl-optimizations/57046
12896 * lra-constraints (split_reg): Set up lra_risky_transformations_p
12897 for multi-reg splits.
12898
0db63e7f
L
128992013-04-24 H.J. Lu <hongjiu.lu@intel.com>
12900
12901 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
12902
3362b6b6
SA
129032013-04-24 Sterling Augustine <saugustine@google.com>
12904
12905 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
12906 (comp_dir_string, debug_str_dwo_section): New.
12907 (DEBUG_STR_DWO_SECTION): Rename to ...
12908 (DEBUG_DWO_STR_SECTION): ... this.
12909 (DEBUG_NORM_STR_SECTION): Delete.
12910 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
12911 (DEBUG_STR_DWO_SECTION_FLAGS): New.
12912 (find_AT_string): Move most logic to ...
12913 (find_AT_string_in_table): ... here. New.
12914 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
12915 add_skeleton_AT_string. Delete logic.
12916 (output_skeleton_debug_sections): Remove call to
12917 add_top_level_skeleton_die_attrs.
12918 (add_comp_dir_attribute): Move logic to comp_dir_string.
12919 (dwarf2out_init): Initialize debug_str_dwo_section.
12920 (output_indirect_string): Call find_string_form.
12921 (output_indirect_strings): Rewrite.
12922 (prune_unused_types): Empty skeleton_debug_str_hash.
12923 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
12924 (dwarf2out_finish): Call output_indirect_strings.
12925
e93e18e9
PC
129262013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12927
12928 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
12929
f6ce35ac
VM
129302013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12931
cabf91cd 12932 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM
12933 (LRA_SUBREG_P): New macro.
12934 * recog.c (register_operand): Check LRA_SUBREG_P.
12935 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12936 stack frame.
12937 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12938 (lra_final_code_change): Skip subreg change for operators.
12939 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12940 if there are no operand changes.
12941 * lra-constraints.c (curr_insn_set): New.
12942 (match_reload): Set LRA_SUBREG_P.
12943 (emit_spill_move): Ditto.
12944 (check_and_process_move): Use curr_insn_set. Process only single
12945 set insns. Don't initialize sec_mem_p and change_p.
12946 (simplify_operand_subreg): Use LRA_SUBREG_P.
12947 (reg_in_class_p): New function.
12948 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12949 of #ifdef. Add code to remove cycling.
12950 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12951 non-null disp. Reload inner instead of disp when base and index
12952 are null. Try to put lo_sum into register.
12953 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 12954 (check_and_process_move): Move code for move cost check to
f6ce35ac 12955 simple_move_p. Remove equiv_substitution.
cabf91cd 12956 (simple_move_p): New function.
f6ce35ac
VM
12957 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12958 curr_insn_set. Call check_and_process_move only for single set
12959 insns. Use the new function. Move call of check_and_process_move
12960 after operand equiv substitution and address process.
12961
38047d90
JG
129622013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
12963
12964 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
12965 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
12966 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
12967
13f39b2e
PC
129682013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12969
12970 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
12971
97c116dc
MP
129722013-04-24 Marek Polacek <polacek@redhat.com>
12973
12974 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
12975 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
12976 (select_loops_exit_conditions): Likewise.
12977 (number_of_iterations_for_all_loops): Likewise.
12978 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
12979 (scev_analysis): Likewise.
12980
83082391 129812013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 12982 Chao-ying Fu <fu@mips.com>
83082391 12983
cabf91cd
AM
12984 * config/mips/micromips.md (jraddiusp): New pattern.
12985 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
12986 instruction if possible.
83082391 12987
19e34aa2
AM
129882013-04-24 Alan Modra <amodra@gmail.com>
12989
12990 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
12991
fdb6603c
JB
129922013-04-24 Julian Brown <julian@codesourcery.com>
12993 Chung-Lin Tang <cltang@codesourcery.com>
12994
12995 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB
12996 dependency behavior in enumeration type DIE generation. Add TODO note
12997 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 12998
bf190e8d
LC
129992013-04-23 Lawrence Crowl <crowl@google.com>
13000
13001 * Makefile.in: Update as needed below.
13002
13003 * hash-table.h (class hash_table):
13004 Correct many methods with parameter types compare_type to the correct
13005 value_type. (Correct code was unlikely to notice the change.)
13006 (hash_table::elements_with_deleted) New.
13007 (class hashtable::iterator): New.
13008 (hashtable::begin()): New.
13009 (hashtable::end()): New.
13010 (FOR_EACH_HASH_TABLE_ELEMENT): New.
13011
13012 * statistics.c (statistics_hashes):
13013 Change type to hash_table. Update dependent calls and types.
13014
13015 * tree-into-ssa.c (var_infos):
13016 Change type to hash_table. Update dependent calls and types.
13017
13018 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
13019 Change type to hash_table. Update dependent calls and types.
13020
13021 * tree-ssa-loop-im.c (struct mem_ref.refs):
13022 Change type to hash_table. Update dependent calls and types.
13023
13024 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
13025 Change type to hash_table. Update dependent calls and types.
13026
13027 * tree-ssa-sccvn.c (vn_tables_s::nary):
13028 Change type to hash_table. Update dependent calls and types.
13029 (vn_tables_s::phis): Likewise.
13030 (vn_tables_s::references): Likewise.
13031
13032 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
13033 (vn_reference_eq): Update parameter and return types.
13034
13035 * tree-ssa-structalias.c (pointer_equiv_class_table):
13036 Change type to hash_table. Update dependent calls and types.
13037 (location_equiv_class_table): Likewise.
13038
13039 * tree-vect-data-refs.c: Consequential changes for making
13040 peeling a hash_table.
13041
13042 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
13043 (destroy_loop_vec_info): Dependent hash_table update.
13044
13045 * tree-vectorizer.h (peeling_htab):
13046 Change type to hash_table. Update dependent calls and types.
13047
d70a81dd
SC
130482013-04-23 Shiva Chen <shiva0217@gmail.com>
13049
cabf91cd
AM
13050 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
13051 to check the register content is equal or not.
13052 * lra-constraints.c (match_reload): Use lra_assign_reg_val
13053 to assign register content record.
13054 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 13055 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM
13056 * lra-int.h (struct lra_reg): Add offset member.
13057 (lra_reg_val_equal_p): New static inline function.
13058 (lra_update_reg_val_offset): New static inline function.
13059 (lra_assign_reg_val): New static inline function.
13060 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
13061 to assign register content record.
13062 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 13063
b894a1f3
CM
130642013-04-23 Catherine Moore <clm@codesourcery.com>
13065
13066 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
13067 operands. Record compression.
13068
ecd14de9
XDL
130692013-04-23 Xinliang David Li <davidxl@google.com>
13070
13071 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
13072
92e776e9
RB
130732013-04-23 Richard Biener <rguenther@suse.de>
13074
13075 PR middle-end/57036
13076 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
13077 parameter, only add abnormal goto edges from the copied body
13078 if the call could perform abnormal gotos.
13079 (copy_cfg_body): Adjust.
13080
a15ee567
SN
130812013-04-23 Sofiane Naci <sofiane.naci@arm.com>
13082
13083 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
13084
08c52234
AS
130852013-04-23 Andreas Schwab <schwab@linux-m68k.org>
13086
13087 * coretypes.h (gimple_stmt_iterator): Add struct to make
13088 compatible with C.
13089
999c1171
RB
130902013-04-23 Richard Biener <rguenther@suse.de>
13091
13092 PR tree-optimization/57026
13093 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
13094 from SSA names occuring in abnormal PHI nodes.
13095
53669259
AK
130962013-04-22 Andi Kleen <ak@linux.intel.com>
13097
13098 * lto/lto.c (print_lto_report_1): Fix LTO report names.
13099
1a0ad150
AK
131002013-04-22 Andi Kleen <ak@linux.intel.com>
13101
13102 * lto/lto.c (print_lto_report_1): Declare early.
13103 (read_cgraph_and_symbols): Call print_lto_report_1 early.
13104
057f8f20
AK
131052013-04-22 Andi Kleen <ak@linux.intel.com>
13106
13107 * common.opt (-flto-report-wpa): Add.
13108 * doc/invoke.texi (-flto-report-wpa): Add.
13109 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
13110 (lto_main): dito.
13111
473b1e05
XDL
131122013-04-22 Xinliang David Li <davidxl@google.com>
13113
13114 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
13115 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
13116 * Makefile.in: New dependency
13117
cabf91cd 13118 David Daney <ddaney.cavm@gmail.com>
b1485a33 13119
cabf91cd
AM
13120 * configure.ac (gcc_cv_as_micromips_support): Use the
13121 --fatal-warnings option.
13122 * configure: Regenerate.
b1485a33 13123
829d0168
MP
131242013-04-22 Marek Polacek <polacek@redhat.com>
13125
13126 PR sanitizer/56990
13127 * tsan.c (instrument_expr): Don't instrument expression
13128 in case its size is zero.
13129
6d9b7208
UB
131302013-04-22 Uros Bizjak <ubizjak@gmail.com>
13131
13132 PR target/57032
13133 Revert:
13134 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
13135
13136 * config/alpha/alpha.c (TARGET_LRA_P): New define.
13137
ea679d55
JG
131382013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13139
13140 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
13141 (gimple_stmt_iterator): New typedef.
13142 * gimple.h (gimple_stmt_iterator): Rename to...
13143 (gimple_stmt_iterator_d): ... This.
13144 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
13145 trees be valid for GIMPLE and GENERIC.
13146 (TARGET_GIMPLE_FOLD_BUILTIN): New.
13147 * gimple-fold.c (gimple_fold_call): Call target hook
13148 gimple_fold_builtin.
13149 * hooks.c (hook_bool_gsiptr_false): New.
13150 * hooks.h (hook_bool_gsiptr_false): New.
13151 * target.def (fold_stmt): New.
13152 * doc/tm.texi: Regenerate.
13153
88a581da
VM
131542013-04-22 Vladimir Makarov <vmakarov@redhat.com>
13155
13156 PR target/57018
13157 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
13158 a set sp if no stack realignment.
13159
92be22dc
NC
131602013-04-22 Nick Clifton <nickc@redhat.com>
13161
13162 * config.gcc (tilegx-linux): Extend extra_objs rather than
13163 overwriting it.
13164 (tilepro-linux): Likewise.
13165
0ddec79f
JG
131662013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13167
13168 * config/aarch64/aarch64-builtins.c
13169 (CF): Remove.
13170 (CF0, CF1, CF2, CF3, CF4, CF10): New.
13171 (VAR<1-12>): Add MAP parameter.
13172 (BUILTIN_*): Likewise.
13173 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
13174 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
13175 (aarch64_ushl_n<mode>): Likewise.
13176 (aarch64_sshr_n<mode>): Likewise.
13177 (aarch64_ushr_n<mode>): Likewise.
13178 (aarch64_<maxmin><mode>): Likewise.
13179 (aarch64_sqrt<mode>): Likewise.
13180 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
13181 (vshr<q>_n_*): Likewise.
13182
0050faf8
JG
131832013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
13184
13185 * config/aarch64/aarch64-builtins.c
13186 (aarch64_simd_builtin_type_mode): Handle SF types.
13187 (sf_UP): Define.
13188 (BUILTIN_GPF): Define.
13189 (aarch64_init_simd_builtins): Handle SF types.
13190 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
13191 (frecps): Likewise.
13192 (frecpx): Likewise.
13193 * config/aarch64/aarch64-simd.md
13194 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
13195 (aarch64_frecpe<mode>): New.
13196 (aarch64_frecps<mode>): Likewise.
13197 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
13198 (v8type): Add frecp<esx>.
13199 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
13200 (aarch64_frecps<mode>): Likewise.
13201 * config/aarch64/iterators.md (FRECP): New.
13202 (frecp_suffix): Likewise.
13203 * config/aarch64/arm_neon.h
13204 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
13205
0fad3dbc 132062013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB
13207
13208 PR target/56995
13209 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
13210 (REG_CLASS_NAMES): Idem.
13211 (REG_CLASS_CONTENTS): Idem.
13212 (REGCLASS_HAS_FP_REG): Idem.
13213 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
13214 (sh_conditional_register_usage): Idem.
13215
3e8a33f9
JL
132162013-04-21 Jeff Law <law@redhat.com>
13217
13218 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
13219 (ssa_forward_propagate_and_combine): Use it.
13220
f38e1b0a
VM
132212013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13222
13223 * lra.c: Update the flow chart diagram.
13224
682303da
VM
132252013-04-19 Vladimir Makarov <vmakarov@redhat.com>
13226
13227 PR rtl-optimization/56847
13228 * lra-constraints.c (process_alt_operands): Discourage alternative
13229 with non-matche doffsettable memory constraint fro memory with
13230 known offset.
13231
f6b64c35
RB
132322013-04-19 Richard Biener <rguenther@suse.de>
13233
13234 PR tree-optimization/56982
13235 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
13236 function.
13237 * gimplify.c (gimplify_call_expr): Notice special calls.
13238 (gimplify_modify_expr): Likewise.
13239 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
13240 abnormal control flow receivers.
13241 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
13242 in the same way as cfun->has_nonlocal_labels.
13243 (gimple_purge_dead_abnormal_call_edges): Likewise.
13244 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
13245 receivers start a basic-block.
13246
01d8bf07
RB
132472013-04-19 Richard Biener <rguenther@suse.de>
13248
13249 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
13250 member ...
13251 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
13252 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
13253 (SLP_TREE_LOAD_PERMUTATION): Add.
13254 (vect_transform_slp_perm_load): Adjust prototype.
13255 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
13256 (vect_free_slp_instance): Likewise.
13257 (vect_create_new_slp_node): Likewise.
13258 (vect_supported_slp_permutation_p): Remove.
13259 (vect_slp_rearrange_stmts): Adjust.
13260 (vect_supported_load_permutation_p): Likewise. Inline
13261 vect_supported_slp_permutation_p here.
13262 (vect_analyze_slp_instance): Compute load permutations per
13263 slp node instead of per instance.
13264 (vect_get_slp_defs): Adjust.
13265 (vect_transform_slp_perm_load): Likewise.
13266 (vect_schedule_slp_instance): Remove redundant code.
13267 (vect_schedule_slp): Remove hack for PR56270, add it ...
13268 * tree-vect-stmts.c (vectorizable_load): ... here, do not
13269 CSE loads for SLP. Adjust.
13270
ede22fc3
GY
132712013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13272
13273 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
13274 spelling in two comments.
13275
67bc84fb
GY
132762013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
13277
13278 PR target/56797
13279 * config/arm/arm.c (load_multiple_sequence): Require SP
13280 as base register for loads if SP is in the register list.
13281
e248d83f
MJ
132822013-04-19 Martin Jambor <mjambor@suse.cz>
13283
13284 PR tree-optimization/56718
13285 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
13286 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
13287 and made public. Adjusted all callers.
13288 (ipa_intraprocedural_devirtualization): New function.
13289 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
13290 (ipa_intraprocedural_devirtualization): Likewise.
13291 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
13292
4891e8f8
RB
132932013-04-19 Richard Biener <rguenther@suse.de>
13294
13295 PR tree-optimization/57000
13296 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
13297
dad89f7c
TG
132982013-04-19 Terry Guo <terry.guo@arm.com>
13299
13300 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
13301 Replace with ...
13302 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
13303 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
13304 (cortex_m4_fmacs): Use new reservations.
13305 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
13306
72ea0d47
VM
133072013-04-18 Vladimir Makarov <vmakarov@redhat.com>
13308
f1e6512c 13309 PR rtl-optimization/56999
72ea0d47
VM
13310 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
13311 related code.
13312 (lra_coalesce): Remove split_origin_bitmap and related code.
13313 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
13314 ranges if necessary.
13315
780a5b71
UB
133162013-04-18 Uros Bizjak <ubizjak@gmail.com>
13317
13318 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
13319 New array.
13320 (ix86_expand_call): Remove clobbered_registers array and use
13321 x86_64_ms_sysv_extra_clobbered_registers instead.
13322 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
13323 Declare here.
13324 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
13325 predicate.
13326 * config/i386/i386.md (*call_rex64_ms_sysv): Use
13327 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
13328 (*call_value_rex64_ms_sysv): Ditto.
13329
6f5a366a
CC
133302013-04-18 Cary Coutant <ccoutant@google.com>
13331
13332 * dwarf2out.c (output_pubnames): Check die_perennial_p of
13333 parent instead of die_mark.
13334
475b8f37
DN
133352013-04-18 Diego Novillo <dnovillo@google.com>
13336
13337 * gimple.c (create_gimple_tmp): New.
13338 (get_expr_type): New.
13339 (build_assign): New.
13340 (build_type_cast): New.
13341 * gimple.h (enum ssa_mode): Define.
13342 (gimple_seq_set_location): New.
13343 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 13344 to use build_assign and build_type_cast.
475b8f37 13345
08940f33
RB
133462013-04-18 Richard Biener <rguenther@suse.de>
13347
13348 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
13349 handle negative step. Remove redundant checks.
13350 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
13351 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
13352 for negative step and grouped loads fail to vectorize.
13353
0e0f87d4
SB
133542013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13355
13356 * emit-rtl.c (reset_insn_used_flags): New function.
13357 (reset_all_used_flags): Use it.
13358 (verify_insn_sharing): New function.
13359 (verify_rtl_sharing): Fix verification for SEQUENCEs.
13360
4c445590
JJ
133612013-04-18 Jakub Jelinek <jakub@redhat.com>
13362
13363 PR tree-optimization/56984
13364 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
13365 and (x >> M) >= N don't register any assertion if N << M is the
13366 minimum value.
13367
6873ecab
SB
133682013-04-18 Steven Bosscher <steven@gcc.gnu.org>
13369
13370 * lower-subreg.c (resolve_simple_move): If called self-recursive,
13371 do not delete_insn insns that have not yet been emitted, only
13372 unlink them with remove_insn.
13373 * df-scan.c (df_insn_delete): Revert r197492.
13374
3ccb989e
SB
133752013-04-17 Steven Bosscher <steven@gcc.gnu.org>
13376
13377 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
13378 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
13379
133802013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY
13381
13382 * config/arm/arm.md (movsicc_insn): Convert define_insn into
13383 define_insn_and_split.
13384 (and_scc,ior_scc,negscc): Likewise.
13385 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
13386
3ccb989e 133872013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY
13388
13389 * config/arm/arm.c (use_return_insn): Return 0 for targets that
13390 can benefit from using a sequence of LDRD instructions in epilogue
13391 instead of a single LDM instruction.
13392
6d10a203
MLI
133932013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
13394
13395 PR 45688
13396 * doc/extend.texi: Fix typo.
13397
6983e6b5
RB
133982013-04-17 Richard Biener <rguenther@suse.de>
13399
13400 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
13401 (vect_build_slp_tree): ... here.
13402 (vect_build_slp_tree_1): Compute which stmts of the SLP group
13403 match. Remove special-casing of mismatched complex loads.
13404 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
13405 re-try the match with swapped commutative operands.
13406 (vect_supported_load_permutation_p): Remove special-casing of
13407 mismatched complex loads.
13408 (vect_analyze_slp_instance): Adjust.
13409
ef23e6a2
RB
134102013-04-17 Richard Biener <rguenther@suse.de>
13411
13412 PR rtl-optimization/56921
13413 * cfgloop.h (struct loop): Add simple_loop_desc member.
13414 (struct niter_desc): Mark with GTY(()).
13415 (simple_loop_desc): Do not use aux field but simple_loop_desc.
13416 * loop-iv.c (get_simple_loop_desc): Likewise.
13417 (free_simple_loop_desc): Likewise.
13418
13419 Revert
13420 2013-04-16 Richard Biener <rguenther@suse.de>
13421
13422 PR rtl-optimization/56921
13423 * loop-init.c (pass_rtl_move_loop_invariants): Add
13424 TODO_do_not_ggc_collect to todo_flags_finish.
13425 (pass_rtl_unswitch): Same.
13426 (pass_rtl_unroll_and_peel_loops): Same.
13427 (pass_rtl_doloop): Same.
13428
fc6f94f5
EB
134292013-04-17 Eric Botcazou <ebotcazou@adacore.com>
13430
13431 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
13432 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
13433 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
13434 references.
13435 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
13436 * tree-streamer.c (record_common_node): Adjust reference in comment.
13437
10a88311
TG
134382013-04-17 Terry Guo <terry.guo@arm.com>
13439
13440 * config/arm/cortex-m4.md: Add a new bypass.
13441
6d9b7208 134422013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N
13443
13444 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
13445 New pattern.
13446 (*subs_<optab><mode>_multp2): New pattern.
13447 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13448 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
13449
6d9b7208 134502013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N
13451
13452 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
13453 (*subs_mul_imm_<mode>): New pattern.
13454
18a6701e
DE
134552013-04-16 David Edelsohn <dje.gcc@gmail.com>
13456
13457 PR target/56948
13458 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
13459 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
13460 (vsx_movti_32bit): Change j->wa to O->wa.
13461
07c37b2f
RB
134622013-04-16 Richard Biener <rguenther@suse.de>
13463
13464 PR rtl-optimization/56921
13465 * loop-init.c (pass_rtl_move_loop_invariants): Add
13466 TODO_do_not_ggc_collect to todo_flags_finish.
13467 (pass_rtl_unswitch): Same.
13468 (pass_rtl_unroll_and_peel_loops): Same.
13469 (pass_rtl_doloop): Same.
13470
0e0f87d4 134712013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY
13472
13473 * config/arm/arm.c (emit_multi_reg_push): New declaration
13474 for an existing function.
13475 (arm_emit_strd_push): New function.
13476 (arm_expand_prologue): Used here.
13477 (arm_emit_ldrd_pop): New function.
13478 (arm_expand_epilogue): Used here.
13479 (arm_get_frame_offsets): Update condition.
13480 (arm_emit_multi_reg_pop): Add a special case for load of a single
13481 register with writeback.
13482
5e8e2af4
UB
134832013-04-16 Uros Bizjak <ubizjak@gmail.com>
13484
13485 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
13486 description.
13487
9fd9ccf7
RB
134882013-04-16 Richard Biener <rguenther@suse.de>
13489
13490 PR tree-optimization/56756
13491 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
13492 (first_mem_ref_loc): New.
13493 (execute_sm): Place the load temporarily before a previous
13494 access instead of in the latch edge to ensure its SSA dependencies
13495 are defined at points dominating the load.
13496
96fba521
SB
134972013-04-16 Steven Bosscher <steven@gcc.gnu.org>
13498
4c8af858
SB
13499 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
13500 correct fix by moving header and footer insn to the footer of
13501 the merged basic block. Clear BB_END of the merged-away block.
13502
96fba521
SB
13503 PR middle-end/43631
13504 * emit-rtl.c (make_note_raw): New function.
13505 (link_insn_into_chain): New static inline function.
13506 (add_insn): Use it.
13507 (add_insn_before, add_insn_after): Factor insn chain linking code...
13508 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
13509 using link_insn_into_chain.
13510 (note_outside_basic_block_p): New helper function for emit_note_after
13511 and emit_note_before.
13512 (emit_note_after): Use nobb variant of add_insn_after if the note
13513 should not be contained in a basic block.
13514 (emit_note_before): Use nobb variant of add_insn_before if the note
13515 should not be contained in a basic block.
13516 (emit_note_copy): Use make_note_raw.
13517 (emit_note): Likewise.
13518 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
13519 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
13520 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
13521 the moved barrier the tail of the basic block it follows.
13522 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
13523
7b8265ba
JJ
135242013-04-15 Jakub Jelinek <jakub@redhat.com>
13525
13526 PR tree-optimization/56962
13527 * gimple-ssa-strength-reduction.c (record_increment): Only set
13528 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
13529 either rhs1 or rhs2 is equal to c->base_expr.
13530
5185d248
RB
135312013-04-15 Richard Biener <rguenther@suse.de>
13532
13533 PR tree-optimization/56933
13534 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
13535 member.
13536 (GROUP_READ_WRITE_DEPENDENCE): Remove.
13537 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
13538 * tree-vect-data-refs.c (vect_analyze_group_access): Move
13539 dependence check ...
13540 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
13541 ... here.
13542 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
13543 GROUP_READ_WRITE_DEPENDENCE.
13544
a24243a0
AK
135452013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13546
13547 * emit-rtl.c (reset_all_used_flags): New function.
13548 (verify_rtl_sharing): Call reset_all_used_flags before and after
13549 performing the checks.
13550
1c50eada
KT
135512013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13552
13553 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
13554 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
13555 * config/arm/constraints.md (De): New constraint.
13556 * config/arm/neon.md (anddi3_neon): Delete.
13557 (neon_vand<mode>): Expand to standard anddi3 pattern.
13558 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
13559 Move earlier in the file.
13560 (neon_inv_logic_op2): Likewise.
13561 (arm_anddi_operand_neon): New predicate.
13562
e927b6ad
RO
135632013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13564
13565 * configure.ac (gcc_cv_ld_as_needed): Set
13566 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
13567 Use -z ignore, -z record on *-*-solaris2*.
13568 (HAVE_LD_AS_NEEDED): Update comment.
13569 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
13570 * configure: Regenerate.
13571 * config.in: Regenerate.
13572 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
13573 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
13574 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
13575 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
13576 equivalents. Fix markup.
13577 * doc/tm.texi: Regenerate.
13578
e0ea8797
AH
135792013-04-15 Andrew Hsieh <andrewhsieh.google.com>
13580
13581 * config/i386/i386.opt: New option mstack-protector-guard=.
13582 * config/i386/i386-opts.h: Add enum stack_protector_guard.
13583 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
13584 TARGET_SSP_TLS_GUARD.
13585 * config/i386/i386.c (ix86_option_override_internal): Set
13586 ix86_stack_protector_guard.
13587 * config/i386/i386.md (stack_protect_set): Enable for
13588 TARGET_SSP_TLS_GUARD only.
13589 (stack_protect_set_<mode>): Ditto.
13590 (stack_protect_test): Ditto.
13591 (stack_protect_test_<mode>): Ditto.
13592 * doc/invoke.texi (i386 Option): Document.
13593
811b72f9
EB
135942013-04-15 Eric Botcazou <ebotcazou@adacore.com>
13595
13596 PR target/56890
13597 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
13598 (S_MODES): Set H_MODE bit.
13599 (SF_MODES): Set only S_MODE and SF_MODE bits.
13600 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
13601 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
13602 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
13603 <MODE_FLOAT>: Likewise.
13604
5529fdd6
JY
136052013-04-15 Joey Ye <joey.ye@arm.com>
13606
13607 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
13608
517b1da2
JY
136092013-04-15 Joey Ye <joey.ye@arm.com>
13610
13611 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
13612 for real far jump.
13613 (thumb_far_jump_used_p): Count instruction size and set
13614 far_jump_used.
13615
01007ae0
EB
136162013-04-14 Eric Botcazou <ebotcazou@adacore.com>
13617
13618 * reorg.c (fill_simple_delay_slots): Reindent block of code.
13619 * resource.c (mark_target_live_regs): Reformat conditional block.
13620
c46f6580
SB
136212013-04-13 Steven Bosscher <steven@gcc.gnu.org>
13622
13623 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
13624 notes, they are emitted only just before final.
13625 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
13626
1f397f45
SB
136272013-04-13 Steven Bosscher <steven@gcc.gnu.org>
13628
13629 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
13630 * cfgrtl.c (delete_insn): Call it here instead.
13631 * lra-spills.c (lra_final_code_change): Use delete_insn.
13632 * haifa-sched.c (sched_remove_insn): Likewise.
13633 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
13634 returning to the nop pool.
13635 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
13636 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
13637
58a51369
SB
136382013-04-12 Steven Bosscher <steven@gcc.gnu.org>
13639
13640 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
13641 * doc/tm.texi: Regenerated.
13642
33159866
UB
136432013-04-12 Uros Bizjak <ubizjak@gmail.com>
13644
13645 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
13646 QImode checks.
13647
226e378f
SB
136482013-04-12 Steven Bosscher <steven@gcc.gnu.org>
13649
13650 * df-core.c (df_find_def): Compare register numbers.
13651 (df_find_use): Likewise.
13652
fafb9b18
VM
136532013-04-12 Vladimir Makarov <vmakarov@redhat.com>
13654
13655 PR target/56903
13656 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
13657 lra_in_progress for return.
13658
9a946fd6
GY
136592013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
13660
13661 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
13662 define_insn into define_insn_and_split and emit movsicc patterns.
13663
41b83758
GY
136642013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
13665
13666 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
13667
d3afd9aa
RB
136682013-04-12 Richard Biener <rguenther@suse.de>
13669
13670 * tree-pass.h (TODO_do_not_ggc_collect): New.
13671 * passes.c (execute_one_ipa_transform_pass): Honor
13672 TODO_do_not_ggc_collect.
13673 (execute_one_pass): Likewise.
13674
13675 Revert
13676 2013-04-10 Richard Biener <rguenther@suse.de>
13677
13678 * passes.c (init_optimization_passes): Remove reload pass.
13679 * ira.c (do_reload): Merge into ...
13680 (ira): ... this.
13681 (rest_of_handle_reload): Remove.
13682 (pass_reload): Likewise.
13683 * config/i386/i386.c (ix86_option_override): Refer to ira instead
13684 of reload for vzeroupper pass placement.
13685
06f9b387
JJ
136862013-04-12 Jakub Jelinek <jakub@redhat.com>
13687
13688 PR tree-optimization/56918
13689 PR tree-optimization/56920
13690 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
13691 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
13692 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
13693 use wide_mul_with_sign method.
13694
953094d2
RB
136952013-04-12 Richard Biener <rguenther@suse.de>
13696
13697 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
13698 not be considered a gimple constant.
13699
bb506982
MG
137002013-04-12 Marc Glisse <marc.glisse@inria.fr>
13701
13702 * fold-const.c (const_binop): Handle vector shifts by a scalar.
13703 (fold_binary_loc): Call const_binop also for mixed vector-scalar
13704 operations.
13705
4b84d650
JJ
137062013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
13707 Jakub Jelinek <jakub@redhat.com>
13708
13709 * opts.c: Include diagnostic-color.h.
13710 (common_handle_option): Handle OPT_fdiagnostics_color_.
13711 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
13712 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
13713 (diagnostic-color.o): New.
13714 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
13715 (diagnostic_color_rule): New enum.
13716 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
13717 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
13718 the location string.
13719 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
13720 either NULL, or color kind.
13721 * diagnostic-color.c: New file.
13722 * diagnostic-color.h: New file.
13723 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
13724 arguments.
13725 * doc/invoke.texi (-fdiagnostics-color): Document.
13726 * pretty-print.h (pp_show_color): Define.
13727 (struct pretty_print_info): Add show_color field.
13728 * diagnostic.c: Include diagnostic-color.h.
13729 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
13730 macros. Colorize error:, warning: etc. strings and also the location
13731 string.
13732 (diagnostic_show_locus): Colorize the caret line.
13733 * pretty-print.c: Include diagnostic-color.h.
13734 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
13735 inside of %< %> quotes or quoted through q format modifier.
13736
067a1e71
AK
137372013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13738
33159866 13739 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 13740
33159866 137412013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N
13742
13743 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
13744 code in CC_NZ mode.
13745 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
13746 pattern.
13747
7b55f98f
MP
137482013-04-11 Marek Polacek <polacek@redhat.com>
13749
13750 PR tree-optimization/48184
33159866 13751 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 13752
966b587e
EB
137532013-04-11 Eric Botcazou <ebotcazou@adacore.com>
13754
13755 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
13756 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
13757 (skip_simple_arithmetic): Tidy up.
13758 * tree.h (skip_simple_constant_arithmetic): Declare.
13759
33159866 137602013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N
13761
13762 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
13763
1d42c1ec
RB
137642013-04-11 Richard Biener <rguenther@suse.de>
13765
13766 * tree-vect-loop.c (get_initial_def_for_induction): Properly
13767 generate vector constants.
13768
4ba5ea11
RB
137692013-04-11 Richard Biener <rguenther@suse.de>
13770
13771 PR tree-optimization/56878
13772 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
13773 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
13774 New function.
13775 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
13776 Prefer to align the DR with the most invariant base address.
13777
f0defe58
SKS
137782013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
13779
13780 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
13781 comment.
13782
d07458be
JG
137832013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
13784
13785 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
13786 floating-point vector comparisons against 0.
13787
146b8692
JJ
137882013-04-11 Jakub Jelinek <jakub@redhat.com>
13789
13790 PR tree-optimization/56899
13791 * fold-const.c (extract_muldiv_1): Apply distributive law
13792 only if TYPE_OVERFLOW_WRAPS (ctype).
13793
b8578ff7
BC
137942013-04-11 Bin Cheng <bin.cheng@arm.com>
13795
13796 PR target/56124
13797 * ira-costs.c (scan_one_insn): Check whether the source rtx of
13798 loading has side effect.
13799
0ea8a6f9
SB
138002013-04-10 Steven Bosscher <steven@gcc.gnu.org>
13801
13802 * config/sparc/sparc.c: Include tree-pass.h.
13803 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
13804 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
13805 head of file. Change return type. Split off gate function.
13806 (sparc_gate_work_around_errata): New function.
13807 (pass_work_around_errata): New pass definition.
13808 (insert_pass_work_around_errata) New pass insert definition to
13809 insert pass_work_around_errata just after delayed-branch scheduling.
13810 (sparc_option_override): Insert the pass.
13811 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
13812
42e37616
DM
138132013-04-10 David S. Miller <davem@davemloft.net>
13814
89deeb3b
DM
13815 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
13816 or -mcpu=hypersparc.
13817
42e37616
DM
13818 * target.def (cstore_mode): New hook.
13819 * target.h: Include insn-codes.h
13820 * targhooks.c: Likewise.
13821 (default_cstore_mode): New function.
13822 * targhooks.h: Declare it.
13823 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
13824 * doc/tm.texi: Rebuild.
13825 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
13826 target hook, rather than inspecting the insn_data.
13827 * config/sparc/sparc.c (sparc_cstore_mode): New function.
13828 (TARGET_CSTORE_MODE): Redefine.
13829 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
13830 result patterns.
13831 * config/sparc/predicates.md (cstore_result_operand): New special
13832 predicate.
13833 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
13834 Use it for operand 0.
13835 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
13836 (*snesi_special): Likewise.
13837 (*snesi_zero): Likewise.
13838 (*seqsi_zero): Likewise.
13839 (*sltu_insn): Likewise.
13840 (*sgeu_insn): Likewise.
13841 (*seqdi_special): Make operand 0 and comparison operation be of
13842 DImode.
13843 (*snedi_special): Likewise.
13844 (*snedi_special_vis3): Likewise.
13845 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
13846 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
13847 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
13848 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
13849 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
13850 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
13851 (*sltu_extend_sp64): Likewise.
13852 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
13853 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
13854 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
13855 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
13856 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
13857
95ca411e
YZ
138582013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
13859
13860 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
13861 (aarch64_start_file): Use the new function.
13862
6782438d 138632013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 13864 Jason Merrill <jason@redhat.com>
6782438d
SKS
13865
13866 * common.opt: Add -gdwarf.
13867 * opts.c (common_handle_option): Handle it.
13868 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
13869
bb313b93
RB
138702013-04-10 Richard Biener <rguenther@suse.de>
13871
13872 * passes.c (execute_todo): Do not call ggc_collect conditional here.
13873 (execute_one_ipa_transform_pass): But unconditionally here.
13874 (execute_one_pass): And here.
13875 (init_optimization_passes): Remove reload pass.
13876 * tree-pass.h (TODO_ggc_collect): Remove.
13877 (pass_reload): Likewise.
13878 * ira.c (do_reload): Merge into ...
13879 (ira): ... this.
13880 (rest_of_handle_reload): Remove.
13881 (pass_reload): Likewise.
13882 * config/i386/i386.c (ix86_option_override): Refer to ira instead
13883 of reload for vzeroupper pass placement.
13884 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
13885 and todo_flags_finish of all passes.
13886
793d9a16
RB
138872013-04-10 Richard Biener <rguenther@suse.de>
13888
13889 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
13890 first_const_oprnd field, rename first_def_type to first_op_type.
13891 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
13892 (vect_get_and_check_slp_defs): Always use the type of the
13893 operand. Allow mixed vect_external_def, vect_constant_def types.
13894 (vect_get_constant_vectors): Handle mixed vect_external_def,
13895 vect_constant_def types.
13896
12211b99 138972013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR
13898
13899 PR tree-optimization/55524
13900 * tree-ssa-math-opts.c
13901 (convert_mult_to_fma): Don't use an fms construct
13902 when we don't have an fms operation, but fnma, and it looks
13903 likely that we'll be able to use the latter.
13904
12211b99 139052013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ
13906
13907 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
13908 function.
13909 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
13910 inline fail caused by overwritable functions.
13911
34ab4a5b
CJW
139122013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
13913
13914 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
13915 unnecessary bits in the constant power of two case.
13916
abf9bfbc
RB
139172013-04-10 Richard Biener <rguenther@suse.de>
13918
13919 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
13920 broken code swapping operands.
13921 (vect_build_slp_tree): Do not compute load permutations here.
13922 (vect_analyze_slp_instance): Compute load permutations here,
13923 after building the SLP tree.
13924
f408477e
CB
139252013-04-09 Christian Bruel <christian.bruel@st.com>
13926
13927 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
13928 of next/prev_real_insn.
13929
5ac42672
JH
139302013-04-09 Jan Hubicka <jh@suse.cz>
13931
abf9bfbc
RB
13932 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13933 Drop aliased parameter.
5ac42672
JH
13934 (function_and_variable_visibility): Do not handle alias pairs.
13935 * cgraph.c (varpool_externally_visible_p): Update prototype.
13936 * varpool.c (varpool_add_new_variable): Update.
13937
5017f1d2
KT
139382013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13939
13940 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
13941
48eecbee
SB
139422013-04-09 Steven Bosscher <steven@gcc.gnu.org>
13943
a949cf1c
SB
13944 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
13945
48eecbee
SB
13946 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
13947
75ef8e3d
MP
139482013-04-09 Marek Polacek <polacek@redhat.com>
13949
13950 PR tree-optimization/48762
33159866 13951 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 13952
23847df4
RB
139532013-04-09 Richard Biener <rguenther@suse.de>
13954
13955 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
13956 dealing with cost.
13957 (vect_build_slp_tree): Likewise.
13958 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
13959 calculating the cost of a SLP instance.
13960 (vect_analyze_slp_instance): Use it from here, after building
13961 the SLP tree.
13962
ea3a0fde
JJ
139632013-04-09 Jakub Jelinek <jakub@redhat.com>
13964
13965 PR middle-end/56883
13966 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
13967 expand_omp_for_static_chunk): Use simple_p = true in
13968 force_gimple_operand_gsi calls when assigning to addressable decls.
13969
a32dfe9d
JL
139702013-04-09 Jeff Law <law@redhat.com>
13971
13972 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
13973 when the boolean was created by converting a wider object which
13974 had a boolean range.
13975
d755c7ef
RB
139762013-04-09 Richard Biener <rguenther@suse.de>
13977
13978 * tree-vectorizer.h (slp_void_p): Remove.
13979 (slp_tree): Typedef before _slp_tree declaration.
13980 (struct _slp_tree): Use a vector of slp_tree as children.
13981 (vect_get_place_in_interleaving_chain): Remove.
13982 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
13983 Move ...
13984 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
13985 and make static.
13986 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
13987 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
13988 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
13989 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
13990 Use slp_node instead of slp_void_p and adjust.
13991
3d741091
RB
139922013-04-09 Richard Biener <rguenther@suse.de>
13993
13994 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
13995 work that is not necessary.
13996
39307ba7
JJ
139972013-04-09 Jakub Jelinek <jakub@redhat.com>
13998
13999 PR tree-optimization/56854
14000 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
14001 forward into clobber stmts if it would change MEM_REF lhs into
14002 non-MEM_REF.
14003
343881fd
MK
140042013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
14005
14006 * tree.c (type_hash_lookup, type_hash_add): Make static.
14007 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
14008
3922658a
RB
140092013-04-09 Richard Biener <rguenther@suse.de>
14010
14011 * tree.h (unsave_expr_now): Remove.
14012 * tree-inline.c (mark_local_for_remap_r): Remove.
14013 (unsave_expr_1): Likewise.
14014 (unsave_r): Likewise.
14015 (unsave_expr_now): Likewise.
14016 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
14017 (propagate_tree_value): Likewise.
14018
9fb6b620
SB
140192013-04-08 Steven Bosscher <steven@gcc.gnu.org>
14020
14021 * doc/rtl.texi (sequence): Rewrite documentation to match the
14022 current use of SEQUENCE rtl objects.
14023 * rtl.def (SEQUENCE): Likewise.
14024
14025 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
14026 Update documentation.
14027 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
14028 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
14029
14030 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
14031
8ddb5a29
TJ
140322013-04-08 Teresa Johnson <tejohnson@google.com>
14033
14034 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
14035 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 14036 methods.
8ddb5a29
TJ
14037 (estimate_edge_size_and_time): Add comment to suggest using rounding
14038 methods.
14039 (estimate_node_size_and_time): Ditto.
14040 (remap_edge_change_prob): Use helper rounding divide methods.
14041 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
14042 (gimple_mod_pow2_value_transform): Ditto.
14043 (gimple_mod_subtract_transform): Ditto.
14044 (gimple_ic_transform): Ditto.
14045 (gimple_stringops_transform): Ditto.
14046 * stmt.c (conditional_probability): Ditto.
14047 (emit_case_dispatch_table): Ditto.
14048 * lto-cgraph.c (merge_profile_summaries): Ditto.
14049 * tree-optimize.c (execute_fixup_cfg): Ditto.
14050 * cfgcleanup.c (try_forward_edges): Ditto.
14051 * cfgloopmanip.c (scale_loop_profile): Ditto.
14052 (loopify): Ditto.
14053 (duplicate_loop_to_header_edge): Ditto.
14054 (lv_adjust_loop_entry_edge): Ditto.
14055 * tree-vect-loop.c (vect_transform_loop): Ditto.
14056 * profile.c (compute_branch_probabilities): Ditto.
14057 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
14058 * lto-streamer-in.c (input_cfg): Ditto.
14059 * gimple-streamer-in.c (input_bb): Ditto.
14060 * ipa-cp.c (update_profiling_info): Ditto.
14061 (update_specialized_profile): Ditto.
14062 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
14063 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 14064 rounding methods.
8ddb5a29
TJ
14065 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14066 (compute_trg_info): Ditto.
14067 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
14068 (purge_dead_edges): Ditto.
14069 * loop-unswitch.c (unswitch_loop): Ditto.
14070 * cgraphclones.c (cgraph_clone_edge): Ditto.
14071 (cgraph_clone_node): Ditto.
14072 * tree-inline.c (copy_bb): Ditto.
14073 (copy_edges_for_bb): Ditto.
14074 (initialize_cfun): Ditto.
14075 (copy_cfg_body): Ditto.
14076 (expand_call_inline): Ditto.
14077
661e6bd7
KT
140782013-04-08 Kai Tietz <ktietz@redhat.com>
14079
14080 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
14081 TARGET_CYGWIN64 by TARGET_64BIT.
14082
105766f3
JR
140832013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14084
14085 * config/epiphany/epiphany.md (GPR_1): New constant.
14086 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
14087 * config/epiphany/epiphany.c (gen_compare_reg):
14088 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
14089 is already in place.
14090 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
14091 Don't require being called during rtl expansion; If y operlaps r0,
14092 return 0.
14093 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
14094 (epiphany_expand_epilogue): Likewise.
14095
8afab237
JR
14096 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
14097 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 14098 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 14099
fa7707d6
JR
14100 * config/epiphany/constraints.md (CnL): New constraint.
14101 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
14102 * config/epiphany/predicates.md (add_operand): Allow 1024.
14103
5138e18d
JR
14104 * config/epiphany/epiphany.md (logical_op): New code iterator.
14105 (op_mnc): New code attribute.
14106 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
14107 (mov_f+1, mov_f+2): New peephole2 patterns.
14108
2ccc703d
JR
14109 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
14110 (cstoresi4): Also allow re-use of zero result when doing a NE
14111 comparison to a non-zero operand.
aefb0819 14112 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 14113
093ac0a5
JR
14114 * config/epiphany/epiphany.md (<insn_opname>v2si3):
14115 Use gen_addsi3_i / gen_subsi3_i.
14116
f223bb13
JJ
141172013-04-08 Jakub Jelinek <jakub@redhat.com>
14118
14119 PR c++/34949
14120 PR c++/50243
14121 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
14122 contain anything but clobbers, at most one __builtin_stack_restore,
14123 optionally debug stmts and final resx, and if it has at least one
14124 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
14125 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
14126 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
14127 which isn't defaut definition, remove them.
14128 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
14129 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
14130 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
14131 with MEM_REF LHS with SSA_NAME address.
14132
4481581f
JL
141332013-04-08 Jeff Law <law@redhat.com>
14134
14135 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 14136
451dabda
RB
141372013-04-08 Richard Biener <rguenther@suse.de>
14138
14139 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
14140 extra newline.
14141 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
14142 determined vector type.
14143 (vect_analyze_data_refs): Likewise.
14144 (vect_get_new_vect_var): Adjust.
14145 (vect_create_destination_var): Preserve SSA name versions.
14146 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
14147 not dump anything here.
14148
3b088b47
JR
141492013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
14150
14151 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
14152 Add member lr_slot_known.
14153 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
14154 if necessary.
14155 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14156 Remove code that sets lr_slot_offset according to what a previous
14157 version of epiphany_emit_save_restore used to do.
14158 (epiphany_emit_save_restore): When doing an lr save or restore,
14159 set/verify lr_slot_known and lr_slot_offset.
14160
d8484d41
XQ
141612013-04-08 Xinyu Qi <xyqi@marvell.com>
14162
33159866
UB
14163 PR target/54338
14164 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ
14165 in ALL_REGS.
14166
481be1c4
RB
141672013-04-08 Richard Biener <rguenther@suse.de>
14168
14169 * alias.c (find_base_term): Fix thinko in previous change.
14170
401f3a81
JJ
141712013-04-08 Jakub Jelinek <jakub@redhat.com>
14172
14173 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
14174 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
14175 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
14176 if possible to compute val.
14177 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
14178 For QImode integers don't require anything about precision. Use
14179 const_with_all_bytes_same to find out if the constant doesn't have
14180 repeated bytes in it.
14181
03ed99a8
AK
141822013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14183
14184 * config/s390/s390.c (s390_expand_insv): Only accept insertions
14185 within mode size.
14186
781b2e62
MP
141872013-04-08 Marek Polacek <polacek@redhat.com>
14188
14189 PR rtl-optimization/48182
14190 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
14191 value to 1.
14192
27e430a2
JDA
141932013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14194
14195 PR target/55487
14196 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
14197 nuses, make sure we have a label.
14198
4902aa64
BS
141992013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14200
14201 PR target/56843
14202 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
14203 (rs6000_emit_swdiv_low_precision): Remove.
14204 (rs6000_emit_swdiv): Rewrite to handle between one and four
14205 iterations of Newton-Raphson generally; modify required number of
14206 iterations for some cases.
14207 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
14208
7bca81dc
SB
142092013-04-05 Steven Bosscher <steven@gcc.gnu.org>
14210
14211 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
14212 set-but-unused variable.
14213
14214 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
14215 basic blocks of released function bodies garbage-collectable.
14216
14217 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
14218 (struct rtl_opt_pass): Add TODO_df_finish.
14219
14220 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
14221
4542a38a
GY
142222013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14223
14224 * config/arm/constraints.md (q): New constraint.
14225 * config/arm/ldrdstrd.md: New file.
14226 * config/arm/arm.md (ldrdstrd.md) New include.
14227 (arm_movdi): Use "q" instead of "r" constraint
14228 for double-word memory access.
14229 (movdf_soft_insn): Likewise.
14230 * config/arm/vfp.md (movdi_vfp): Likewise.
14231 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 14232 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY
14233 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
14234 (mem_ok_for_ldrd_strd): Likewise.
14235 (output_move_double): Update assertion.
14236
2385b218
GY
142372013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14238
14239 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
14240
75fe1cb5
GY
142412013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14242
14243 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
14244 define_insn_and_split.
14245 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
14246
dd660e8e
GY
142472013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14248
14249 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
14250 define_insn_and_split.
33159866 14251 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY
14252 (shiftsi3_compare): New pattern.
14253 (rrx): New pattern.
14254 * config/arm/unspecs.md (UNSPEC_RRX): New.
14255
045e472c
GY
142562013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14257
14258 * config/arm/arm.md (negdi_extendsidi): New pattern.
14259 (negdi_zero_extendsidi): Likewise.
14260
3f3bf1a8
GY
142612013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14262
14263 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
14264 define_insn_and_split.
14265 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
14266 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
14267
b0b49556
GY
142682013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14269
14270 * config/arm/arm.md (arm_subdi3): Convert define_insn into
14271 define_insn_and_split.
14272 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
14273 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
14274
d633dd84
GY
142752013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14276
14277 * config/arm/arm.md (subsi3_carryin): New pattern.
14278 (subsi3_carryin_const): Likewise.
14279 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
14280 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
14281
f4499066
GY
142822013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14283
14284 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
14285
ceef6fd9
GY
142862013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
14287
14288 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 14289 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 14290
ddbdd8a7
KT
142912013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14292
14293 * config/arm/arm.c (arm_expand_builtin): Change fcode
14294 type to unsigned int.
14295
8456d78a
RR
142962013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14297
14298 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
14299
526c230f
ILT
143002013-04-04 Ian Lance Taylor <iant@google.com>
14301
33159866
UB
14302 * doc/standards.texi (Standards): The Go frontend supports the Go 1
14303 language standard.
526c230f 14304
3cfbe04d
SB
143052013-04-04 Steven Bosscher <steven@gcc.gnu.org>
14306
14307 PR middle-end/56729
14308 * df-scan.c (df_insn_delete): Disable failing assert.
14309
dfa3f8d0
KT
143102013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14311
14312 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
14313 New function prototype.
14314 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
14315 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
14316 (arm_builtin_vectorized_function): New function.
14317
f7837758
KT
143182013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14319
14320 * config/arm/arm_neon_builtins.def: New file.
14321 * config/arm/arm.c (neon_builtin_data): Move contents to
14322 arm_neon_builtins.def.
14323 (enum arm_builtins): Include neon builtin definitions.
14324 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 14325 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 14326
39c1b6db
MP
143272013-04-04 Marek Polacek <polacek@redhat.com>
14328
14329 PR tree-optimization/48186
14330 * predict.c (maybe_hot_frequency_p): Return false if
14331 HOT_BB_FREQUENCY_FRACTION is 0.
14332 (cgraph_maybe_hot_edge_p): Likewise.
14333
314f64eb
RB
143342013-04-04 Richard Biener <rguenther@suse.de>
14335
14336 PR tree-optimization/56826
14337 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
14338 more accurately.
14339
51a905b2
RB
143402013-04-04 Richard Biener <rguenther@suse.de>
14341
14342 PR tree-optimization/56213
14343 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 14344 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 14345
f1bf4f3a
RB
143462013-04-04 Richard Biener <rguenther@suse.de>
14347
14348 PR tree-optimization/56837
14349 * tree-loop-distribution.c (classify_partition): For non-zero
14350 values require that the value has the same precision as its
14351 mode to be useful as memset value.
14352
0bca7ded
NC
143532013-04-03 Nick Clifton <nickc@redhat.com>
14354
33159866 14355 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC
14356 (fmssf4): Use fmsf.s on E3V5 architectures.
14357 (fnmasf4): Use fnmaf.s on E3V5 architectures.
14358 (fnmssf4): Use fnmsf.s on E3V5 architectures.
14359
b4019227
JL
143602013-04-03 Jeff Law <law@redhat.com>
14361
14362 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
14363 (lra-eliminations.o): Likewise.
14364
f57ddb5b
TJ
143652013-04-03 Teresa Johnson <tejohnson@google.com>
14366
14367 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 14368 compute_working_sets here from profile.c.
f57ddb5b 14369 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB
14370 (gcov_working_set_t): Moved typedef here from basic-block.h
14371 (compute_working_set): Declare.
f57ddb5b
TJ
14372 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
14373 (get_working_sets): Renamed from compute_working_set,
33159866 14374 replace most of body with call to new compute_working_sets.
f57ddb5b 14375 (get_exec_counts): Replace call to compute_working_sets
33159866
UB
14376 to get_working_sets.
14377 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 14378 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 14379 to get_working_sets.
f57ddb5b
TJ
14380 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
14381 * gcov-dump.c (dump_working_sets): New function.
14382
12211b99 143832013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ
14384
14385 * hwint.c (sext_hwi, zext_hwi): New functions.
14386 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
14387 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
14388 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
14389 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
14390 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
14391 (sext_hwi, zext_hwi): New functions.
0bca7ded 14392
be672e08
JL
143932013-04-03 Jeff Law <law@redhat.com>
14394
14395 PR tree-optimization/56799
33159866
UB
14396 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
14397 back test for widening conversion erroneously dropped in prior change.
be672e08 14398
9d821fa5
KT
143992013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14400
14401 PR target/56809
14402 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
14403 instead of next_real_insn.
14404
71c581e7
MP
144052013-04-03 Marek Polacek <polacek@redhat.com>
14406
14407 PR sanitizer/55702
33159866 14408 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 14409
4a32ef80
KT
144102013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14411
14412 PR target/56809
14413 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
14414 next_real_insn.
14415 (thumb1_output_casesi): Likewise.
14416 (thumb2_output_casesi): Likewise.
14417
1b2253d4
RB
144182013-04-03 Richard Biener <rguenther@suse.de>
14419
14420 PR tree-optimization/56817
14421 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
14422 Split out ...
14423 (tree_unroll_loops_completely_1): ... new function to manually
14424 walk the loop tree, properly defering outer loops of unrolled
14425 loops to later iterations.
14426
38000232
MG
144272013-04-03 Marc Glisse <marc.glisse@inria.fr>
14428
14429 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
14430 (vectorizable_load): Likewise.
14431 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
14432 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
14433
3fa35298
MG
144342013-04-03 Marc Glisse <marc.glisse@inria.fr>
14435
14436 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
14437 BIT_FIELD_REF.
14438
b3d45ff0
UW
144392013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
14440
14441 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
14442
ec9202a8
BC
144432013-04-03 Bin Cheng <bin.cheng@arm.com>
14444
14445 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
14446
6805bd36
MG
144472013-04-03 Marc Glisse <marc.glisse@inria.fr>
14448
14449 PR tree-optimization/56790
33159866
UB
14450 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
14451 folding.
6805bd36 14452
da694a77
MG
144532013-04-03 Marc Glisse <marc.glisse@inria.fr>
14454
14455 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
14456 Handle VEC_MERGE.
14457 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
14458 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
14459 equal arguments.
14460
4e7d7b3d
JJ
144612013-04-03 Jakub Jelinek <jakub@redhat.com>
14462
14463 PR c/19449
14464 * tree.h (force_folding_builtin_constant_p): New decl.
14465 * builtins.c (force_folding_builtin_constant_p): New variable.
14466 (fold_builtin_constant_p): Fold immediately also if
14467 force_folding_builtin_constant_p.
14468
e6c9d234
RB
144692013-04-03 Richard Biener <rguenther@suse.de>
14470
14471 PR tree-optimization/56812
14472 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
14473 DRs of the same interleaving chain are independent.
14474
984d07dd
JM
144752013-04-02 Jason Merrill <jason@redhat.com>
14476
14477 * gdbinit.in (pbb): Use debug fn.
14478
622849c9
LC
144792013-04-02 Lawrence Crowl <crowl@google.com>
14480
14481 * sese.h (struct ivtype_map_elt_s): Remove unused.
14482 (extern debug_ivtype_map): Remove unused.
14483 (extern eq_ivtype_map_elts): Remove unused.
14484 * sese.c (debug_ivtype_map): Removed unused.
14485 (debug_ivtype_map_1): Removed unused.
14486 (debug_ivtype_elt): Remove unused.
14487 (eq_ivtype_map_elts): Remove unused.
14488
14489
82c0e1a0
KT
144902013-04-02 Kai Tietz <ktietz@redhat.com>
14491
14492 PR target/52790
14493 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
14494 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB
14495 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
14496 function.
82c0e1a0
KT
14497 (legitimize_pe_coff_symbol): Likewise.
14498 (is_imported_p): New helper-function.
14499 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
14500 for Windows x64 targets.
14501 (ix86_expand_prologue): Optimize for pe-coff targets.
14502 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
14503 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
14504 medium/large code-model.
14505 (legitimize_pic_address): Likewise.
14506 (legitimize_tls_address): Likewise.
14507 (ix86_expand_call): Likewise.
14508 (x86_output_mi_thunk): Likewise.
14509 (get_dllimport_decl): Add new beimport argument.
14510 (construct_plt_address): Don't assert for x64 pe-coff targets.
14511 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
14512 targets.
14513 (SYMBOL_FLAG_STUBVAR): New macro.
14514 (SYMBOL_REF_STUBVAR_P): Likewise.
14515 * config/i386/winnt.c (stub_list): New structure.
14516 (stub_head): New local variable.
14517 (i386_pe_record_stub): New function.
14518 (i386_pe_file_end): Emit refptr-stubs.
14519
5d751b0c
JJ
145202013-04-02 Jakub Jelinek <jakub@redhat.com>
14521
09bb4c99
JJ
14522 PR rtl-optimization/56745
14523 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
14524 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
14525
a1d8947a
JJ
14526 PR c++/34949
14527 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
14528 and both of them are MEM_REFs, just compare first argument for
14529 equality and attempt to deal even with differing offsets.
14530
5d751b0c
JJ
14531 PR c++/34949
14532 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
14533 of gimple_clobber_p to be MEM_REF.
14534 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
14535 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
14536 after gimplification.
14537 * asan.c (get_mem_ref_of_assignment): Don't instrument
14538 gimple_clobber_p stmts.
14539 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
14540 gimple_clobber_p stmt if they have MEM_REF lhs and
14541 are dead because of another gimple_clobber_p stmt.
14542 * tree-ssa-live.c (clear_unused_block_pointer): Treat
14543 gimple_clobber_p stmts like debug stmts.
14544 (remove_unused_locals): Remove clobbers with MEM_REF lhs
14545 that refer to unused VAR_DECLs or uninitialized values.
14546 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
14547 gimple_clobber_p stmts if they refer to removed parameters.
14548 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
14549 formatting.
14550
e594716a
UB
145512013-04-02 Uros Bizjak <ubizjak@gmail.com>
14552
14553 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
14554 using SWI48 mode attribute.
14555
7a80735b
WM
145562013-04-02 Wei Mi <wmi@google.com>
14557
14558 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
14559 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
14560 *<rotate_insn><mode>3_mask in i386.md.
14561
f423a9e4
AI
145622013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
14563
14564 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
14565
90eb75f2
RB
145662013-04-02 Richard Biener <rguenther@suse.de>
14567
14568 PR tree-optimization/56778
14569 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
14570 Runtime alias tests are not supported for gather loads.
14571 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
14572 stmts referenced from SSA operands before updating SSA form.
14573
d8c69a92
IC
145742013-04-02 Ian Caulfield <ian.caulfield@arm.com>
14575 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14576
14577 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB
14578 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
14579 * config/arm/cortex-a53.md: New file.
14580 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
14581 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
14582 * config/arm/arm.c (arm_issue_rate): Likewise.
14583 * config/arm/arm-tune.md: Regenerate
14584 * config/arm/arm-tables.opt: Regenerate.
14585 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 14586
239eb04c
ZC
145872013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
14588
14589 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
14590 non-static link.
14591
c902d3c8
SN
145922013-04-02 Sofiane Naci <sofiane.naci@arm.com>
14593
d8c69a92
IC
14594 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
14595 scalar load/store operations using B/H registers.
c902d3c8
SN
14596 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
14597
051d0e2f
SN
145982013-04-02 Sofiane Naci <sofiane.naci@arm.com>
14599
14600 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
14601 scalar move.
14602 * config/aarch64/aarch64.c
14603 (aarch64_simd_scalar_immediate_valid_for_move): New.
14604 * config/aarch64/aarch64-protos.h
14605 (aarch64_simd_scalar_immediate_valid_for_move): New.
14606 * config/aarch64/constraints.md (Dh, Dq): New.
14607 * config/aarch64/iterators.md (hq): New.
14608
0ee1e3d9
EB
146092013-04-02 Eric Botcazou <ebotcazou@adacore.com>
14610
14611 * reorg.c (get_branch_condition): Deal with conditional returns.
14612 (fill_simple_delay_slots): Remove dead code dealing with jumps.
14613
136fb3f7
WM
146142013-04-01 Wei Mi <wmi@google.com>
14615
14616 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
14617 Truncate operand 2 using %b asm operand modifier.
14618 (*<shift_insn><mode>3_mask): Ditto.
14619 (*<rotate_insn><mode>3_mask): Ditto.
14620
6388c738
SB
146212013-04-01 Steven Bosscher <steven@gcc.gnu.org>
14622
14623 PR middle-end/56798
14624 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
14625
f7a4d826
KK
146262013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
14627
14628 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
14629 of next_real_insn.
14630 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
14631
4b943a49
LC
146322013-03-30 Lawrence Crowl <crowl@google.com>
14633
14634 * dse.c (clear_alias_sets): Remove never set.
14635 (disqualified_clear_alias_sets): Remove never set.
14636 (clear_alias_mode_pool): Remove never set.
14637 (dse_step0): Remove condition that is never true.
14638 (canon_address): Remove condition that is never true.
14639 (dse_step7): Remove condition that is never true.
14640 (rest_of_handle_dse): Remove condition that is never true.
14641 (rest_of_handle_dse::did_global): Remove never read from above.
14642 (dse_step2_spill): Remove never called from above.
14643 (dse_step5_spill): Remove never called from above.
14644
39718607
SB
146452013-03-30 Steven Bosscher <steven@gcc.gnu.org>
14646
da5c6bde
SB
14647 * doc/md.texi (Standard Names) <casesi>: Update documentation for
14648 JUMP_TABLE_DATA changes.
14649 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
14650 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
14651 (Insns) <jump_table_data>: New entry.
14652 * doc/tm.texi: Regenerate.
14653
39718607
SB
14654 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
14655
14656 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
14657 for table jump at the end of a basic block using tablejump_p.
14658 * targhooks.c (default_invalid_within_doloop): Likewise.
14659 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
14660 target hook implementation that is identical to the default hook.
14661 (rs6000_invalid_within_doloop): Remove.
14662
14663 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
14664 unused variable from tablejump_p call.
14665
14666 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
14667 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
14668 (INSN_DELETED_P): Likewise.
14669 (emit_jump_table_data): New prototype.
14670 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
14671 after 4th as unused.
14672 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
14673 * sched-vis.c (print_insn): Likewise.
14674 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
14675 insn for compatibility with back ends that use next_active_insn to
14676 identify jump table data.
14677 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
14678 (remove_insn): Likewise.
14679 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
14680 to be emitted.
14681 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
14682 (emit_jump_table_data): New function.
14683
14684 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
14685 basic block, a JUMP_TABLE_DATA never is.
14686 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
14687 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
14688 off from code handling real insns.
14689 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
14690 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
14691 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
14692 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
14693 is not a NONDEBUG_INSN_P.
14694 * ira-costs.c (scan_one_insn): Likewise.
14695 * jump.c (mark_all_labels): Likewise.
14696 (mark_jump_label_1): Likewise.
14697 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
14698 * lra.c (get_insn_freq): Expect all insns reaching here to be in
14699 a basic block.
14700 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
14701 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
14702 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
14703 JUMP_TABLE_DATA_P insns.
14704 (calculate_elim_costs_all_insns): Likewise.
14705 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
14706 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
14707 (delete_output_reload): Code style fixups.
14708 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
14709 insn flags on this non-insn.
14710 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
14711 as scheduling barriers, for pre-change compatibility.
14712 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
14713 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
14714
14715 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
14716 redundant JUMP_TABLE_DATA_P test.
14717 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
14718 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
14719 (frv_for_each_packet): Likewise.
14720 * config/i386/i386.c (min_insn_size): Likewise.
14721 (ix86_avoid_jump_mispredicts): Likewise.
14722 * config/m32r/m32r.c (m32r_is_insn): Likewise.
14723 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
14724 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
14725 (mips16_insn_length): Robustify.
14726 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
14727 (mips16_split_long_branches): Likewise.
14728 * config/pa/pa.c (pa_combine_instructions): Likewise.
14729 * config/rs6000/rs6000.c (get_next_active_insn): Treat
14730 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
14731 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
14732 as contributing to pool range lengths.
14733 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
14734 Remove redundant JUMP_TABLE_DATA_P test.
14735 (sh_loop_align): Likewise.
14736 (split_branches): Likewise.
14737 (sh_insn_length_adjustment): Likewise.
14738 * config/spu/spu.c (get_branch_target): Likewise.
14739
0208f7da
JH
147402013-03-29 Jan Hubicka <jh@suse.cz>
14741
14742 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
14743 gcov streaming; stream hot bb threshold to ltrans.
14744 * predict.c (get_hot_bb_threshold): Break out from ....
14745 (maybe_hot_count_p): ... here.
14746 (set_hot_bb_threshold): New function.
14747 * lto-section-in.c (lto_section_name): Add profile.
14748 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
14749 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
14750 and data-streamer.h
14751 (histogram_entry): New structure.
14752 (histogram, histogram_pool): New global vars.
14753 (histogram_hash): New structure.
14754 (histogram_hash::hash): New method.
14755 (histogram_hash::equal): Likewise.
14756 (account_time_size): New function.
14757 (cmp_counts): New function.
14758 (dump_histogram): New function.
14759 (ipa_profile_generate_summary): New function.
14760 (ipa_profile_write_summary): New function.
14761 (ipa_profile_read_summary): New function.
14762 (ipa_profile): Decide on threshold.
e594716a
UB
14763 (pass_ipa_profile): Add ipa_profile_write_summary and
14764 ipa_profile_read_summary.
0208f7da
JH
14765 * Makefile.in (ipa.o): Update dependencies.
14766 * lto-streamer.h (LTO_section_ipa_profile): New section.
14767
5a6ccc94
GDR
147682013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
14769
14770 * tree.h (VAR_P): New.
14771
39385fa6
PC
147722013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
14773
14774 PR lto/56777
14775 * doc/invoke.texi ([-fwhole-program]): Fix typo.
14776
34f0d87a
SB
147772013-03-29 Steven Bosscher <steven@gcc.gnu.org>
14778
14779 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
14780 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
14781 (control_flow_insn_p): Likewise.
14782 * cfgrtl.c (duplicate_insn_chain): Likewise.
14783 * final.c (get_attr_length_1): Likewise.
14784 (shorten_branches): Likewise.
14785 (final_scan_insn): Likewise.
14786 * function.c (instantiate_virtual_regs): Likewise.
14787 * gcse.c (insert_insn_end_basic_block): Likewise.
14788 * ira-costs.c (scan_one_insn): Likewise.
14789 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
14790 * lra.c (check_rtl): Likewise.
14791 * reload1.c (elimination_costs_in_insn): Likewise.
14792 * reorg.c (follow_jumps): Likewise.
14793
14794 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
14795 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
14796 (thumb_far_jump_used_p): Likewise.
14797 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
14798 (workaround_speculation): Likewise.
14799 (add_sched_insns_for_speculation): Likewise.
14800 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
14801 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
14802 (frv_for_each_packet): Likewise.
14803 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
14804 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
14805 (final_emit_insn_group_barriers): Likewise.
14806 * config/m32r/m32r.c (m32r_is_insn): Likewise.
14807 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
14808 (mips16_insn_length): Likewise.
14809 * config/pa/pa.c (pa_reorg): Likewise.
14810 (pa_combine_instructions): Likewise.
14811 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
14812 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
14813 (sh_reorg): Likewise.
14814 (split_branches): Likewise.
14815 * config/spu/spu.c (get_branch_target): Likewise.
14816
14817 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
14818 JUMP_TABLE_DATA_P.
14819
4ac761b0
KY
148202013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
14821
39385fa6 14822 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY
14823 Fix declaration name.
14824
58a49189
LC
148252013-03-28 Lawrence Crowl <crowl@google.com>
14826
14827 * graphds.h (struct graph.indicies): Remove unused.
14828 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
14829 (SCOP_ORIGINAL_PDDRS): Remove unused.
14830 * sese.h (extern insert_loop_close_phis): Removed unused.
14831 (extern insert_guard_phis): Removed unused.
14832 (extern ivtype_map_elt_info): Removed unused.
14833 (new_ivtype_map_elt): Removed unused.
14834 * sese.c (ivtype_map_elt_info): Removed unused.
14835
7b3b6ae4
LC
148362013-03-28 Lawrence Crowl <crowl@google.com>
14837
14838 * Makefile.in: Add several missing include dependences.
14839 (DUMPFILE_H): New.
14840 (test-dump.o): New. This object is not added to any executable,
14841 but is present for ad-hoc testing.
14842 * bitmap.c
14843 (debug (const bitmap_head_def &)): New.
14844 (debug (const bitmap_head_def *)): New.
14845 * bitmap.h
14846 (extern debug (const bitmap_head_def &)): New.
14847 (extern debug (const bitmap_head_def *)): New.
14848 * cfg.c
14849 (debug (edge_def &)): New.
14850 (debug (edge_def *)): New.
14851 * cfghooks.c
14852 (debug (basic_block_def &)): New.
14853 (debug (basic_block_def *)): New.
14854 * dumpfile.h
14855 (dump_node (const_tree, int, FILE *)): Correct source file.
14856 * dwarf2out.c
14857 (debug (die_struct &)): New.
14858 (debug (die_struct *)): New.
14859 * dwarf2out.h
14860 (extern debug (die_struct &)): New.
14861 (extern debug (die_struct *)): New.
14862 * gimple-pretty-print.c
14863 (debug (gimple_statement_d &)): New.
14864 (debug (gimple_statement_d *)): New.
14865 * gimple-pretty-print.h
14866 (extern debug (gimple_statement_d &)): New.
14867 (extern debug (gimple_statement_d *)): New.
14868 * ira-build.c
14869 (debug (ira_allocno_copy &)): New.
14870 (debug (ira_allocno_copy *)): New.
14871 (debug (ira_allocno &)): New.
14872 (debug (ira_allocno *)): New.
14873 * ira-int.h
14874 (extern debug (ira_allocno_copy &)): New.
14875 (extern debug (ira_allocno_copy *)): New.
14876 (extern debug (ira_allocno &)): New.
14877 (extern debug (ira_allocno *)): New.
14878 * ira-lives.c
14879 (debug (live_range &)): New.
14880 (debug (live_range *)): New.
14881 * lra-int.h
14882 (debug (lra_live_range &)): New.
14883 (debug (lra_live_range *)): New.
14884 * lra-lives.c
14885 (debug (lra_live_range &)): New.
14886 (debug (lra_live_range *)): New.
14887 * omega.c
14888 (debug (omega_pb_d &)): New.
14889 (debug (omega_pb_d *)): New.
14890 * omega.h
14891 (extern debug (omega_pb_d &)): New.
14892 (extern debug (omega_pb_d *)): New.
14893 * print-rtl.c
14894 (debug (const rtx_def &)): New.
14895 (debug (const rtx_def *)): New.
14896 * print-tree.c
14897 (debug_tree (tree): Move within file.
14898 (debug_raw (const tree_node &)): New.
14899 (debug_raw (const tree_node *)): New.
14900 (dump_tree_via_hooks (const tree_node *, int)): New.
14901 (debug (const tree_node &)): New.
14902 (debug (const tree_node *)): New.
14903 (debug_verbose (const tree_node &)): New.
14904 (debug_verbose (const tree_node *)): New.
14905 (debug_head (const tree_node &)): New.
14906 (debug_head (const tree_node *)): New.
14907 (debug_body (const tree_node &)): New.
14908 (debug_body (const tree_node *)): New.
14909 (debug_vec_tree (tree): Move and reimplement in terms of dump.
14910 (debug (vec<tree, va_gc> &)): New.
14911 (debug (vec<tree, va_gc> *)): New.
14912 * rtl.h
14913 (extern debug (const rtx_def &)): New.
14914 (extern debug (const rtx_def *)): New.
14915 * sbitmap.c
14916 (debug_raw (simple_bitmap_def &)): New.
14917 (debug_raw (simple_bitmap_def *)): New.
14918 (debug (simple_bitmap_def &)): New.
14919 (debug (simple_bitmap_def *)): New.
14920 * sbitmap.h
14921 (extern debug (simple_bitmap_def &)): New.
14922 (extern debug (simple_bitmap_def *)): New.
14923 (extern debug_raw (simple_bitmap_def &)): New.
14924 (extern debug_raw (simple_bitmap_def *)): New.
14925 * sel-sched-dump.c
14926 (debug (vinsn_def &)): New.
14927 (debug (vinsn_def *)): New.
14928 (debug_verbose (vinsn_def &)): New.
14929 (debug_verbose (vinsn_def *)): New.
14930 (debug (expr_def &)): New.
14931 (debug (expr_def *)): New.
14932 (debug_verbose (expr_def &)): New.
14933 (debug_verbose (expr_def *)): New.
14934 (debug (vec<rtx> &)): New.
14935 (debug (vec<rtx> *)): New.
14936 * sel-sched-dump.h
14937 (extern debug (vinsn_def &)): New.
14938 (extern debug (vinsn_def *)): New.
14939 (extern debug_verbose (vinsn_def &)): New.
14940 (extern debug_verbose (vinsn_def *)): New.
14941 (extern debug (expr_def &)): New.
14942 (extern debug (expr_def *)): New.
14943 (extern debug_verbose (expr_def &)): New.
14944 (extern debug_verbose (expr_def *)): New.
14945 (extern debug (vec<rtx> &)): New.
14946 (extern debug (vec<rtx> *)): New.
14947 * sel-sched-ir.h
14948 (_list_iter_cond_expr): Make inline instead of static.
14949 * sreal.c
14950 (debug (sreal &)): New.
14951 (debug (sreal *)): New.
14952 * sreal.h
14953 (extern debug (sreal &)): New.
14954 (extern debug (sreal *)): New.
14955 * tree.h
14956 (extern debug_raw (const tree_node &)): New.
14957 (extern debug_raw (const tree_node *)): New.
14958 (extern debug (const tree_node &)): New.
14959 (extern debug (const tree_node *)): New.
14960 (extern debug_verbose (const tree_node &)): New.
14961 (extern debug_verbose (const tree_node *)): New.
14962 (extern debug_head (const tree_node &)): New.
14963 (extern debug_head (const tree_node *)): New.
14964 (extern debug_body (const tree_node &)): New.
14965 (extern debug_body (const tree_node *)): New.
14966 (extern debug (vec<tree, va_gc> &)): New.
14967 (extern debug (vec<tree, va_gc> *)): New.
14968 * tree-cfg.c
14969 (debug (struct loop &)): New.
14970 (debug (struct loop *)): New.
14971 (debug_verbose (struct loop &)): New.
14972 (debug_verbose (struct loop *)): New.
14973 * tree-dump.c: Add header dependence.
14974 * tree-flow.h
14975 (extern debug (struct loop &)): New.
14976 (extern debug (struct loop *)): New.
14977 (extern debug_verbose (struct loop &)): New.
14978 (extern debug_verbose (struct loop *)): New.
14979 * tree-data-ref.c
14980 (debug (data_reference &)): New.
14981 (debug (data_reference *)): New.
14982 (debug (vec<data_reference_p> &)): New.
14983 (debug (vec<data_reference_p> *)): New.
14984 (debug (vec<ddr_p> &)): New.
14985 (debug (vec<ddr_p> *)): New.
14986 * tree-data-ref.h
14987 (extern debug (data_reference &)): New.
14988 (extern debug (data_reference *)): New.
14989 (extern debug (vec<data_reference_p> &)): New.
14990 (extern debug (vec<data_reference_p> *)): New.
14991 (extern debug (vec<ddr_p> &)): New.
14992 (extern debug (vec<ddr_p> *)): New.
14993 * tree-ssa-alias.c
14994 (debug (pt_solution &)): New.
14995 (debug (pt_solution *)): New.
14996 * tree-ssa-alias.h
14997 (extern debug (pt_solution &)): New.
14998 (extern debug (pt_solution *)): New.
14999 * tree-ssa-alias.c
15000 (debug (_var_map &)): New.
15001 (debug (_var_map *)): New.
15002 (debug (tree_live_info_d &)): New.
15003 (debug (tree_live_info_d *)): New.
15004 * tree-ssa-alias.h
15005 (extern debug (_var_map &)): New.
15006 (extern debug (_var_map *)): New.
15007 (extern debug (tree_live_info_d &)): New.
15008 (extern debug (tree_live_info_d *)): New.
15009
be77e1e5
JH
150102013-03-28 Jan Hubicka <jh@suse.cz>
15011
15012 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
15013
777e6976
IB
150142013-03-28 Ian Bolton <ian.bolton@arm.com>
15015
15016 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
15017 record only when desired or required.
15018
9cd347ae
UB
150192013-03-28 Uros Bizjak <ubizjak@gmail.com>
15020
15021 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
15022 *vec_extractv2di_1_rex64. Use x64 isa attribute.
15023
33159866 150242013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N
15025
15026 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
15027 (*andsi3_compare0_uxtw): New pattern.
15028 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
15029 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
15030
89ab31c1
JH
150312013-03-28 Jan Hubicka <jh@suse.cz>
15032
15033 * data-streamer-in.c (streamer_read_gcov_count): New function.
15034 * gimple-streamer-out.c: Include value-prof.h.
15035 (output_gimple_stmt): Output histogram.
15036 (output_bb): Use streamer_write_gcov_count.
15037 * value-prof.c: Include data-streamer.h
15038 (dump_histogram_value): Add HIST_TYPE_MAX.
15039 (stream_out_histogram_value): New function.
15040 (stream_in_histogram_value): New function.
15041 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
15042 (stream_out_histogram_value, stream_in_histogram_value): Declare.
15043 * data-streamer-out.c (streamer_write_gcov_count): New function.
15044 (streamer_write_gcov_count_stream): New function.
15045 * lto-cgraph.c (lto_output_edge): Update counter streaming.
15046 (lto_output_node): Likewise.
15047 (input_node, input_edge): Likewise.
15048 * lto-streamer-out.c (output_cfg): Update streaming.
15049 * lto-streamer-in.c (input_cfg): Likewise.
15050 * data-streamer.h (streamer_write_gcov_count,
15051 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
15052 * gimple-streamer-in.c: Include value-prof.h
15053 (input_gimple_stmt): Input histograms.
15054 (input_bb): Update profile streaming.
15055
e594716a 150562013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 15057
33159866 15058 * genmodes.c (emit_max_int): New function.
8697be17 15059 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC
15060 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
15061 Added doc.
8697be17 15062 * machmode.def: Fixed comment.
89ab31c1 15063
e594716a 150642013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ
15065
15066 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ
15067 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
15068
74946978
MP
150692013-03-28 Marek Polacek <polacek@redhat.com>
15070 Richard Biener <rguenther@suse.de>
15071
15072 PR tree-optimization/56695
15073 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
15074 build signed result of a vector comparison.
15075 * tree-cfg.c (verify_gimple_comparison): Check that a result
15076 of a vector comparison has signed type.
15077
a64b9c26
RB
150782013-03-28 Richard Biener <rguenther@suse.de>
15079
15080 PR tree-optimization/37021
15081 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
15082 do not restrict gaps between groups.
15083 * tree-vect-stmts.c (vectorizable_load): Properly account for
15084 a gap between groups.
15085
a9dc2a2f
EB
150862013-03-28 Eric Botcazou <ebotcazou@adacore.com>
15087
15088 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
15089 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
15090 is not enabled.
15091
53cb97f9
GP
150922013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
15093
15094 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
15095 * doc/extend.texi (Named Address Spaces): Ditto.
15096 (Variable Attributes): Ditto.
15097
b802ae5c
KT
150982013-03-27 Kai Tietz <ktietz@redhat.com>
15099
eddae10a
KT
15100 * config.build: Add support for cygwin x64 target.
15101 * config.gcc: Likewise.
15102 * config.host: Likewise.
15103 * configure.ac: Likewise
15104 * configure: Regenerated.
b802ae5c 15105
371e77e3 151062013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT
15107
15108 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
15109 * config/i386/t-cygwin-w64: New file.
15110 * config/i386/cygwin-w64.h: New file.
15111 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
15112 and add support for x64-cygwin target.
15113 (CPP_SPEC): Likewise.
15114 (CXX_WRAP_SPEC_LIST): Undefine before define.
15115 (LIBGCJ_SONAME): Use 15 as version.
15116
f49b33cb
RB
151172013-03-27 Richard Biener <rguenther@suse.de>
15118
15119 PR tree-optimization/56716
15120 * tree-ssa-structalias.c (perform_var_substitution): Adjust
15121 dumping for ref nodes.
15122
b37a6ce5
MJ
151232013-03-27 Martin Jambor <mjambor@suse.cz>
15124
15125 PR tree-optimization/55334
15126 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
15127 restricted pointers to arrays.
15128
9469b9b2
GDR
151292013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
15130
15131 * Makefile.in (.SUFFIXES): Add .cc.
15132 (.c.o): Apply same recipe for implicit rule .cc.o.
15133
7d24f650
RB
151342013-03-27 Richard Biener <rguenther@suse.de>
15135
15136 PR tree-optimization/37021
15137 * tree-vect-data-refs.c (vect_check_strided_load): Allow
15138 REALPART/IMAGPART_EXPRs around the supported refs.
15139 * tree-ssa-structalias.c (find_func_aliases): Assume that
15140 floating-point values are not used to transfer pointers.
15141
2f251a05
AI
151422013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
15143
d33d9e47
AI
15144 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
15145 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
15146 * gcc/doc/tm.texi: Regenerate.
15147 * gcc/targhooks.h (default_has_ifunc_p): New.
15148 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
15149 * gcc/config/linux-protos.h: New file.
15150 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
2f251a05
AI
15151 this hook for linux which disables support of indirect functions in
15152 android.
d33d9e47
AI
15153 * gcc/config/linux-android.c: New file.
15154 * gcc/config/t-linux-android.c: Ditto.
15155 * gcc/config.gcc: Added new object file linux-android.o.
15156 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 15157 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
d33d9e47 15158 * gcc/varasm.c (do_assemble_alias): Likewise.
2f251a05
AI
15159 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
15160 doesn't support indirect functions.
15161 * configure: Regenerate.
15162
78b4e425
BC
151632013-03-27 Bin Cheng <bin.cheng@arm.com>
15164
15165 PR target/56102
15166 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
15167 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
15168 mult-word mode.
15169
a5ba7b92
AK
151702013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15171
15172 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
15173
be624986
TG
151742013-03-27 Terry Guo <terry.guo@arm.com>
15175
15176 * config/arm/arm-cores.def: Added core cortex-r7.
15177 * config/arm/arm-tune.md: Regenerated.
15178 * config/arm/arm-tables.opt: Regenerated.
15179 * doc/invoke.texi: Added entry for core cortex-r7.
15180
0a514f47
WL
151812013-03-27 Walter Lee <walt@tilera.com>
15182
15183 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
15184 double-decrement of next_scratch_regno.
15185
9332b0d2
WL
151862013-03-27 Walter Lee <walt@tilera.com>
15187
801d9b2a 15188 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL
15189 input operands.
15190 (insn_v1mulus): Ditto.
15191 (insn_v2muls): Ditto.
15192
f54ea5dd
WL
151932013-03-27 Walter Lee <walt@tilera.com>
15194
e594716a 15195 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL
15196 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
15197
bf60f4ca
WL
151982013-03-27 Walter Lee <walt@tilera.com>
15199
e594716a 15200 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL
15201 (*sibcall_value): Ditto.
15202
e3b51eeb
WL
152032013-03-27 Walter Lee <walt@tilera.com>
15204
15205 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
15206 (insn_mnz_v8qi): ... this ...
15207 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
15208 vector equivalent.
15209 (insn_v<n>mnz): Replaced by ...
15210 (insn_v1mnz): ... this ...
15211 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
15212 equivalent.
15213 (insn_mz_<mode>): Replaced by ...
15214 (insn_mz_v8qi): ... this ...
15215 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
15216 vector equivalent.
15217 (insn_v<n>mz): Replaced by ...
15218 (insn_v1mz): ... this ...
15219 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
15220 equivalent.
15221
4fc7b145
EB
152222013-03-26 Eric Botcazou <ebotcazou@adacore.com>
15223
15224 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
15225
6f33817e
RM
152262013-03-26 Roland McGrath <mcgrathr@google.com>
15227
15228 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
15229 than fprintf with a non-constant, non-format string.
15230
93a4145b
UB
152312013-03-26 Uros Bizjak <ubizjak@gmail.com>
15232
15233 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
15234 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
15235 operand 0 predicate.
15236 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
15237 attribute. Use general_x64nomem_operand as operand 1 predicate.
15238 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
15239 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15240 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
15241 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
15242 (mov<mode>_insv_1): Remove expander. Merge insn with
15243 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
15244 Use general_x64nomem_operand as operand 1 predicate.
15245 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
15246 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
15247 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15248 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
15249 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
15250 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
15251 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
15252 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
15253 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
15254 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
15255 (general_x64nomem_operand): Ditto.
15256
c6a9ed5a 152572013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH
15258
15259 * config/rtems.opt: Add -pthread option.
15260
31b0a960
RB
152612013-03-26 Richard Biener <rguenther@suse.de>
15262
93a4145b
UB
15263 * alias.c (find_base_term): Avoid redundant and not used recursion.
15264 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB
15265 (true_dependence_1): Compute and pass base terms to base_alias_check.
15266 (write_dependence_p): Likewise.
15267 (may_alias_p): Likewise.
15268
79517551
SN
152692013-03-26 Sofiane Naci <sofiane.naci@arm.com>
15270
15271 * config/aarch64/aarch64.c (aarch64_classify_address): Support
15272 PC-relative load in SI modes and above only.
15273
a76213b9
XQ
152742013-03-26 Xinyu Qi <xyqi@marvell.com>
15275
15276 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB
15277 * config/arm/iwmmxt.md (WCGR0): Update.
15278 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 15279
37ff9355
UB
152802013-03-26 Uros Bizjak <ubizjak@gmail.com>
15281
15282 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
15283 Use x64 and nox64 isa attributes.
15284
30b0317c
RB
152852013-03-26 Richard Biener <rguenther@suse.de>
15286
15287 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
15288 alignment computations and rely on get_object_alignment_1
15289 for the !TYPE_P case.
15290 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
15291
cbcdb140
WL
152922013-03-26 Walter Lee <walt@tilera.com>
15293
15294 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
15295 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
15296
dd884525 152972013-03-25 Jeff Law <law@redhat.com>
be672e08 15298
33159866
UB
15299 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
15300 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL
15301 wrong version of prior patch.
15302
0051d3ec
WL
153032013-03-25 Walter Lee <walt@tilera.com>
15304
15305 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
15306 TILEGX_INSN_SHUFFLEBYTES1.
15307 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
15308 shufflebytes1.
15309 (tilegx_builtins): Ditto.
15310 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
15311
065a3605
WL
153122013-03-25 Walter Lee <walt@tilera.com>
15313
15314 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL
15315 (floatunssisf2): New pattern.
15316 (floatsidf2): New pattern.
15317 (floatunssidf2): New pattern.
065a3605 15318
5b2a3c82
WL
153192013-03-25 Walter Lee <walt@tilera.com>
15320
15321 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
15322 tests for constraint J, K, N, P.
15323
192ea533
WL
153242013-03-25 Walter Lee <walt@tilera.com>
15325
15326 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
15327 Use indirect/pcrel encoding.
15328 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
15329 Ditto.
15330
855e0d0b
SE
153312013-03-25 Steve Ellcey <sellcey@mips.com>
15332
15333 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
15334 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
15335 * config/mips/mips.c (mips_option_override): Set IMADD default.
15336 * config/mips/mips.h (PTF_AVOID_IMADD): New.
15337 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
15338 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
15339 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
15340 * doc/invoke.texi (-mimadd/-mno-imadd): New.
15341
39e45653
JL
153422013-03-25 Jeff Law <law@redhat.com>
15343
15344 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
15345 slightly to avoid creating and folding useless trees. Simplify
15346 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
15347
7482c470
UB
153482013-03-25 Uros Bizjak <ubizjak@gmail.com>
15349
15350 * config/i386/i386.md (*zero_extendsidi2): Merge with
15351 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
15352 * config/i386/predicates.md (x86_64_zext_operand): Rename from
15353 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
15354 targets. Clarify comment.
15355
4a53743e
MJ
153562013-03-25 Martin Jambor <mjambor@suse.cz>
15357
15358 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
15359 pass-through jump functions differently.
15360 (ipa_read_jump_function): Likewise. Also use setter functions to set
15361 up jump functions.
15362
162712de
MJ
153632013-03-25 Martin Jambor <mjambor@suse.cz>
15364
15365 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
15366 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
15367 process it.
15368 (ipa_get_indirect_edge_target): New function.
15369 (devirtualization_time_bonus): New parameter known_aggs, pass it to
15370 ipa_get_indirect_edge_target. Update all callers.
15371 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
15372 ipa_get_indirect_edge_target_1 instead of calling
15373 ipa_get_indirect_edge_target.
15374 (create_specialized_node): Pass aggvlas to
15375 ipcp_discover_new_direct_edges.
15376
a5a4c20a
KT
153772013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15378
15379 * config/arm/arm.md (f_sels, f_seld): New types.
15380 (*cmov<mode>): New pattern.
15381 * config/arm/predicates.md (arm_vsel_comparison_operator): New
15382 predicate.
15383
f992b9fc
KT
153842013-03-25 Kai Tietz <ktietz@redhat.com>
15385
15386 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
15387 POSIX-printf for mingw-hosted builds.
15388
c16fd676
RB
153892013-03-25 Richard Biener <rguenther@suse.de>
15390
15391 PR middle-end/56694
15392 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
15393 must-not-throw stmt location.
15394
5ad29f12
KT
153952013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15396
15397 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
15398 Emit load-acquire versions when acq is true.
15399 (arm_emit_store_exclusive): Add rel parameter.
15400 Emit store-release versions when rel is true.
15401 (arm_split_compare_and_swap): Use acquire-release instructions
15402 instead.
15403 of barriers when appropriate.
15404 (arm_split_atomic_op): Likewise.
15405 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
15406 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
15407 (VUNSPEC_SLX): Likewise.
15408 (VUNSPEC_LDA): Likewise.
15409 (VUNSPEC_STL): Likewise.
15410 * config/arm/sync.md (atomic_load<mode>): New pattern.
15411 (atomic_store<mode>): Likewise.
15412 (arm_load_acquire_exclusive<mode>): Likewise.
15413 (arm_load_acquire_exclusivesi): Likewise.
15414 (arm_load_acquire_exclusivedi): Likewise.
15415 (arm_store_release_exclusive<mode>): Likewise.
15416
03a7dddb
CM
154172013-03-25 Catherine Moore <clm@codesourcery.com>
15418
15419 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
15420 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
15421 * config/mip/predicates.md (lwsp_swsp_operand,
15422 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
15423 sb16_operand, db4_operand, db7_operand, ib3_operand,
15424 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
15425 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
15426 andi16_operand): New predicates.
15427 * config/mips/mips.md (compression): New attribute.
15428 (enabled): New attribute.
15429 (length): Consider compression in computing length.
15430 (shift_compression): New code attribute.
15431 (*add<mode>3): New operands. Record compression.
15432 (sub<mode>3): Likewise.
15433 (one_cmpl<mode>2): Likewise.
15434 (*and<mode>3): Likewise.
15435 (*ior<mode>3): Likewise.
15436 (unnamed pattern for xor): Likewise.
15437 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
15438 (*<optab><mode>3): Likewise.
15439 (*mov<mode>_internal: Likewise.
15440 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
15441 (mips_unsigned_immediate_p): New.
15442 (umips_lwsp_swsp_address_p): New.
15443 (m16_based_address_p): New.
15444 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
15445 (mips_unsigned_immediate_p): New prototype.
15446 (lwsp_swsp_address_p): New prototype.
15447 (m16_based_address_p): New prototype.
15448 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
15449 (mips_signed_immediate_p): New function.
15450 (m16_based_address_p): New function.
15451 (lwsp_swsp_address_p): New function.
15452 (mips_print_operand_punctuation): Recognize short delay slot insns
15453 for microMIPS.add<mode>3"
15454
f35c297f 154552013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 15456
f35c297f
KT
15457 PR target/56720
15458 * config/arm/iterators.md (v_cmp_result): New mode attribute.
15459 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
15460
051b9446
RB
154612013-03-25 Richard Biener <rguenther@suse.de>
15462
15463 PR tree-optimization/56689
15464 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
15465 any edge.
15466
374001cb
RB
154672013-03-25 Richard Biener <rguenther@suse.de>
15468
15469 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
15470 of bitmap.
15471 (memory_references): Likewise.
15472 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
15473 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
15474 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
15475 (gather_mem_refs_in_loops): Fold into ...
15476 (analyze_memory_references): ... this. Move initialization
15477 to tree_ssa_lim_initialize.
15478 (fill_always_executed_in): Rename to ...
15479 (fill_always_executed_in_1): ... this.
15480 (fill_always_executed_in): Move contains_call computation to
15481 this new function from ...
15482 (tree_ssa_lim_initialize): ... here.
15483 (tree_ssa_lim): Call fill_always_executed_in.
15484
57895947
EB
154852013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15486
15487 * postreload.c (reload_combine): Fix code detecting returns.
15488
1e1b18c1
EB
154892013-03-25 Eric Botcazou <ebotcazou@adacore.com>
15490
15491 * function.c (emit_use_return_register_into_block): On cc0 targets,
15492 do not emit the sequence between cc0 setter and user.
15493
9216baf1
KT
154942013-03-25 Kai Tietz <ktietz@redhat.com>
15495
1e1b18c1
EB
15496 * config/i386/predicates.md (local_symbolic_operand): Interpret
15497 dll-imported symbols as none-local.
9216baf1 15498
76421b44
RB
154992013-03-25 Richard Biener <rguenther@suse.de>
15500
15501 * tree-ssa-loop-im.c (struct depend): Remove.
15502 (struct lim_aux_data): Make depends a vec of gimples.
15503 (free_lim_aux_data): Adjust.
15504 (add_dependency): Likewise.
15505 (set_level): Likewise.
15506
d154bfa2
RB
155072013-03-25 Richard Biener <rguenther@suse.de>
15508
15509 PR middle-end/56434
15510 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
15511 the pointer returned by calls with ECF_MALLOC set.
15512
2cd9804e
UB
155132013-03-24 Uros Bizjak <ubizjak@gmail.com>
15514
acef5fe0
UB
15515 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
15516
155172013-03-24 Uros Bizjak <ubizjak@gmail.com>
15518
15519 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB
15520 using MMXMODE mode iterator.
15521 (*move<mode>_internal): Merge with *movv2sf_internal and
15522 *movv2sf_internal_rex64 using MMXMODE mode iterator.
15523
225ccc68
SB
155242013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15525
85c0f02d
SB
15526 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
15527 (record_last_mem_set_info): Likewise.
15528
225ccc68
SB
15529 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
15530 of XNEWVEC followed by memset.
15531 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
15532
b64925dc
SB
155332013-03-23 Steven Bosscher <steven@gcc.gnu.org>
15534
15535 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
15536 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
15537 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
15538 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
15539 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
15540 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
15541 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
15542 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
15543 BARRIER_P instead of GET_CODE.
15544
b9aaf52e
EB
155452013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15546
15547 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
15548 inaccuracy in the probing code.
15549
15550 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
15551 (ctrapdi4): Likewise.
15552
66de4d7c
EB
155532013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15554
15555 * calls.c (expand_call): Add missing guard to code handling return
15556 of non-BLKmode structures in MSB.
15557 * function.c (expand_function_end): Likewise.
15558
4f25c8fb
EB
155592013-03-23 Eric Botcazou <ebotcazou@adacore.com>
15560
15561 * combine.c (try_combine): Adjust comment. Do not add the set of
15562 insn #0 if the destination indirectly is set or dies in insn #2.
15563 Tidy up code to distribute a new note.
15564
23b7850d
UB
155652013-03-22 Uros Bizjak <ubizjak@gmail.com>
15566
15567 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
15568 also for alternatives 16 and 17.
15569
bed852cf
UB
155702013-03-22 Uros Bizjak <ubizjak@gmail.com>
15571
15572 * config/i386/sse.md (*mov<mode>_internal): Merge with
15573 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
15574 Emit insn template depending on type attribute. Use
15575 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
15576 movd instead of movq mnemonic for interunit moves. Rewrite mode
15577 attribute calculation. Remove unit attribute calculation.
15578 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
15579 Set prefix_data16 attribute for DImode ssemov types.
15580 Use Ym instead of y for SSE-MMX conversion alternatives.
15581 Reorder operand constraints.
15582
1e8a7937
SB
155832013-03-22 Steven Bosscher <steven@gcc.gnu.org>
15584
15585 * df.h (df_insn_delete): Adjust prototype.
15586 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
15587 and let it decide whether mark the basic block dirty.
15588 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
15589 * df-scan.c (df_insn_info_delete): New helper function, split
15590 off from df_insn_delete.
15591 (df_scan_free_bb_info): Use it.
15592 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
15593 Likewise.
15594 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
15595 that the insn is actually an insn and it has a non-NULL basic block.
15596 Do not mark basic block dirty if only deleting a DEBUG_INSN.
15597
e14d094c
RB
155982013-03-22 Richard Biener <rguenther@suse.de>
15599
15600 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
15601 dep_ref members.
15602 (mem_ref_alloc): Do not allocate them.
15603 (refs_independent_p): Do not query or maintain a cache.
15604
e9cf7316
RB
156052013-03-22 Richard Biener <rguenther@suse.de>
15606
15607 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
15608 (gather_mem_refs_in_loops): Do not compute it.
15609 (analyze_memory_references): Do not allocate it.
15610 (tree_ssa_lim_finalize): Do not free it.
15611 (for_all_locs_in_loop): Do not query all_refs_in_loop.
15612
f046e81b
RB
156132013-03-22 Richard Biener <rguenther@suse.de>
15614
15615 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 15616
a05c0ddf
IB
156172013-03-22 Ian Bolton <ian.bolton@arm.com>
15618
15619 * config/aarch64/aarch64.c (aarch64_print_operand): New
15620 format specifier for printing a constant in hex.
15621 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
15622 format specifier for printing second operand.
15623
c00217fc
RB
156242013-03-22 Richard Biener <rguenther@suse.de>
15625
15626 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
15627 bitmaps.
15628 (gather_mem_refs_in_loops): Perform store accumulation here.
15629 (create_vop_ref_mapping_loop): Remove.
15630 (create_vop_ref_mapping): Likewise.
15631 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 15632 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB
15633 (record_indep_loop): Remove.
15634 (record_dep_loop): New function.
15635 (ref_indep_loop_p_1): Adjust to only walk over references
15636 in the loop, not its subloops.
15637 (ref_indep_loop_p): Rename to ...
15638 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
15639 maintaining a more fine-grained cache.
15640 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
15641 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
15642
15d19bf8
RB
156432013-03-22 Richard Biener <rguenther@suse.de>
15644
15645 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
15646 (struct mem_ref): Make accesses_in_loop a vec of a vec of
15647 aggregate mem_ref_loc.
15648 (free_mem_ref_locs): Inline into ...
15649 (memref_free): ... this and adjust.
15650 (mem_ref_alloc): Adjust.
15651 (mem_ref_locs_alloc): Remove.
15652 (record_mem_ref_loc): Adjust.
15653 (get_all_locs_in_loop): Rewrite into ...
15654 (for_all_locs_in_loop): ... this iterator.
15655 (rewrite_mem_ref_loc): New functor.
15656 (rewrite_mem_refs): Use for_all_locs_in_loop.
15657 (sm_set_flag_if_changed): New functor.
15658 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
15659 (ref_always_accessed): New functor.
15660 (ref_always_accessed_p): Use for_all_locs_in_loop.
15661
6f37411d
MG
156622013-03-21 Marc Glisse <marc.glisse@inria.fr>
15663
15664 * tree-pass.h (PROP_gimple_lvec): New.
15665 * passes.c (dump_properties): Handle PROP_gimple_lvec.
15666 (init_optimization_passes): Move pass_lower_vector.
15667 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
15668 PROP_gimple_lvec.
15669 (pass_lower_vector): Provide PROP_gimple_lvec.
15670 (pass_lower_vector_ssa): Likewise.
15671 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
15672
511ef689
MW
156732013-03-21 Mark Wielaard <mjw@redhat.com>
15674
15675 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
15676
cc1df30b
UB
156772013-03-21 Uros Bizjak <ubizjak@gmail.com>
15678
15679 * config/i386/i386.md (*movdi_internal): Disparage slightly
15680 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
15681 conversion alternatives.
15682
c845cfe1
JJ
156832013-03-21 Jakub Jelinek <jakub@redhat.com>
15684
37e99116
JJ
15685 PR middle-end/48087
15686 * diagnostic.def (DK_WERROR): New kind.
15687 * diagnostic.h (werrorcount): Define.
15688 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
15689 promoted to DK_ERROR, increment DK_WERROR counter instead of
15690 DK_ERROR counter.
15691 * toplev.c (toplev_main): Call print_ignored_options even if
15692 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
15693 even if just werrorcount is non-zero.
15694
c845cfe1
JJ
15695 PR debug/55608
15696 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
15697 on failure.
15698 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
15699 (string_cst_pool_decl): New function.
15700 (optimize_one_addr_into_implicit_ptr): New function.
15701 (resolve_addr_in_expr): Optimize DWARF location expression
15702 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
15703 which doesn't live in memory, but has DW_AT_location or
15704 DW_AT_const_value, or refers to a string literal, into
15705 DW_OP_GNU_implicit_pointer.
15706 (optimize_location_into_implicit_ptr): New function.
15707 (resolve_addr): If removing DW_AT_location of a variable because
15708 it was DW_OP_addr of address of the variable, but the variable doesn't
15709 live in memory, try to emit const value attribute for the initializer.
15710
08e0cda6
MG
157112013-03-21 Marc Glisse <marc.glisse@inria.fr>
15712
15713 * tree.h (VECTOR_TYPE_P): New macro.
15714 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
15715 TYPE_MODE): Use it.
15716 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
15717 VEC_COND_EXPR cannot be lvalues.
15718 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
15719
d08633b4
MG
157202013-03-21 Marc Glisse <marc.glisse@inria.fr>
15721
15722 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
15723 Restrict the transformation to equal modes.
15724
e6647190
RB
157252013-03-21 Richard Biener <rguenther@suse.de>
15726
15727 PR tree-optimization/39326
15728 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
15729 (MEM_ANALYZABLE): Adjust.
15730 (record_mem_ref_loc): Move bitmap ops ...
15731 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
15732 unanalyzable refs, do not record locations for it.
15733 (analyze_memory_references): Allocate ref zero as shared
15734 unanalyzable ref.
15735 (refs_independent_p): Do not test for unanalyzed mems here.
15736 (ref_indep_loop_p_1): Special-case disambiguation against
15737 the unanalyzed ref.
cc1df30b 15738 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 15739
65074f54
CL
157402013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
15741
15742 * config/arm/arm-protos.h (tune_params): Add
15743 prefer_neon_for_64bits field.
15744 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
15745 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
15746 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
15747 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
15748 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
15749 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
15750 (arm_option_override): Handle -mneon-for-64bits new option.
15751 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
15752 (prefer_neon_for_64bits): Declare new variable.
15753 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 15754 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL
15755 (arch_enabled): Handle new arch types. Remove support for onlya8
15756 and nota8.
15757 (one_cmpldi2): Use new arch names.
bc5faa5b 15758 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL
15759 * config/arm/arm.opt (mneon-for-64bits): Add option.
15760 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
15761 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
15762 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
15763 of onlya8.
15764 * doc/invoke.texi (-mneon-for-64bits): Document.
15765
5a2d2a79
RB
157662013-03-21 Richard Biener <rguenther@suse.de>
15767
15768 PR tree-optimization/39326
15769 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
15770 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 15771 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 15772
5abe1e05
RB
157732013-03-21 Richard Biener <rguenther@suse.de>
15774
15775 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
15776 (vect_insert_into_interleaving_chain): Likewise.
15777 (vect_drs_dependent_in_basic_block): Inline ...
15778 (vect_slp_analyze_data_ref_dependence): ... here. New function,
15779 split out from ...
15780 (vect_analyze_data_ref_dependence): ... here. Simplify.
15781 (vect_check_interleaving): Simplify.
15782 (vect_analyze_data_ref_dependences): Likewise. Split out ...
15783 (vect_slp_analyze_data_ref_dependences): ... this new function.
15784 (dr_group_sort_cmp): New function.
15785 (vect_analyze_data_ref_accesses): Compute data-reference groups
15786 here instead of in vect_analyze_data_ref_dependence. Use
15787 a more efficient algorithm.
15788 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
15789 vect_slp_analyze_data_ref_dependences. Call
15790 vect_analyze_data_ref_accesses earlier.
15791 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
15792 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
15793 (vect_slp_analyze_data_ref_dependences): New prototype.
15794
cad1735b
RB
157952013-03-21 Richard Biener <rguenther@suse.de>
15796
15797 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
15798 ref is stored in the loop.
15799 (find_refs_for_sm): Walk only over all stores.
15800 (store_motion_loop): Allocate from lim_bitmap_obstack.
15801 (store_motion): Likewise.
15802
141310ef
RB
158032013-03-21 Richard Biener <rguenther@suse.de>
15804
15805 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
15806 Update virtual SSA form.
15807
5022315a
RO
158082013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15809
15810 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
15811 * configure: Regenerate.
15812 * config.in: Regenerate.
15813 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
15814 if !HAVE_LD_EH_FRAME_CIEV3.
15815
50d4421c
RB
158162013-03-21 Richard Biener <rguenther@suse.de>
15817
15818 * tree-cfg.c (verify_expr_no_block): New function.
15819 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
15820 nor DECL_VALUE_EXPR have locations with associated blocks.
15821 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
15822 (clear_unused_block_pointer): Remove code dealing with
15823 blocks in DECL_DEBUG_EXPR locations.
15824
839b422f
RB
158252013-03-21 Richard Biener <rguenther@suse.de>
15826
15827 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
15828 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
15829 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
15830 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
15831 instead of DECL_DEBUG_EXPR_IS_FROM.
15832 * gimplify.c (gimplify_modify_expr): Likewise.
15833 * tree-cfg.c (verify_expr_location_1): Likewise.
15834 * tree-complex.c (create_one_component_var): Likewise.
15835 * tree-sra.c (create_access_replacement): Likewise.
15836 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
15837 (clear_unused_block_pointer): Likewise.
15838 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
15839 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
15840 * var-tracking.c (var_debug_decl): Likewise.
15841 (track_expr_p): Likewise.
15842 * tree-inline.c (add_local_variables): Likewise. Set
15843 DECL_HAS_DEBUG_EXPR_P after copying it.
15844 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
15845 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
15846
fe04878d
UB
158472013-03-21 Uros Bizjak <ubizjak@gmail.com>
15848
15849 PR bootstrap/56656
15850 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
15851 * configure: Regenerate.
15852 * config.in: Regenerate.
15853 * config/i386/i386.md (*movdf_internal): Use
15854 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
15855 movd instead of movq mnemonic for interunit moves.
15856 (*movdi_internal): Ditto.
15857
bd059b26 158582013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N
15859
15860 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
15861 (abd<mode>_3): New pattern.
15862 (aba<mode>_3): New pattern.
15863 (fabd<mode>_3): New pattern.
15864
bd059b26 158652013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N
15866
15867 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
15868 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
15869 occurrence of REGISTER_PREFIX as its empty string.
15870
4f1f78b9
JL
158712013-03-20 Jeff Law <law@redhat.com>
15872
15873 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
15874 addititional equivalences for equality comparisons between an SSA_NAME
15875 and a constant where the SSA_NAME was set from a widening conversion.
15876
327a1118
WL
158772013-03-20 Walter Lee <walt@tilera.com>
15878
15879 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
15880
813e0036
UB
158812013-03-20 Uros Bizjak <ubizjak@gmail.com>
15882
15883 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
15884 depending on type attribute.
15885 (*movti_internal): Ditto.
15886 (*movtf_internal): Ditto.
15887 (*movxf_internal): Ditto.
15888 (*movdf_internal): Ditto.
15889 (*movsf_internal): Ditto.
15890
7cf34aae
UB
158912013-03-20 Uros Bizjak <ubizjak@gmail.com>
15892
15893 * config/i386/i386.md (*movti_internal): Set prefix attribute to
15894 maybe_vex for sselog1 and ssemov types.
15895 (*movdi_internal): Reorder operand constraints.
15896 (*movsi_internal): Ditto. Set prefix attribute to
15897 maybe_vex for sselog1 and ssemov types.
15898 (*movtf_internal): Set prefix attribute to maybe_vex
15899 for sselog1 and ssemov types.
15900 (*movdf_internal): Ditto. Set prefix_data16 attribute for
15901 DImode ssemov types. Reorder operand constraints.
15902 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
15903 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
15904 attribute for SImode ssemov types. Reorder operand constraints.
15905
19321415
MJ
159062013-03-20 Martin Jambor <mjambor@suse.cz>
15907
15908 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
15909 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
15910
12211b99 159112013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM
15912
15913 * config/rs6000/predicates.md (indexed_address, update_address_mem
15914 update_indexed_address_mem): New predicates.
15915 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
15916 attribute for load/store instructions.
15917 * config/rs6000/dfp.md (movsd_store): Likewise.
15918 (movsd_load): Likewise.
15919 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
15920 (unnamed HI->DI extend define_insn): Likewise.
15921 (unnamed SI->DI extend define_insn): Likewise.
15922 (unnamed QI->SI extend define_insn): Likewise.
15923 (unnamed QI->HI extend define_insn): Likewise.
15924 (unnamed HI->SI extend define_insn): Likewise.
15925 (unnamed HI->SI extend define_insn): Likewise.
15926 (extendsfdf2_fpr): Likewise.
15927 (movsi_internal1): Likewise.
15928 (movsi_internal1_single): Likewise.
15929 (movhi_internal): Likewise.
15930 (movqi_internal): Likewise.
15931 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
15932 attribute for load/store instructions.
15933 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
15934 instructions.
15935 (mov<mode>_softfloat): Likewise.
15936 (mov<mode>_hardfloat32): Likewise.
15937 (mov<mode>_hardfloat64): Likewise.
15938 (mov<mode>_softfloat64): Likewise.
15939 (movdi_internal32): Likewise.
15940 (movdi_internal64): Likewise.
15941 (probe_stack_<mode>): Likewise.
15942
159432013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
15944
15945 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
15946 floating point, and decimal floating point to reload iterator.
15947
15948 * config/rs6000/constraints.md (wl constraint): New constraints to
15949 return FLOAT_REGS if certain options are used to reduce the number
15950 of separate patterns that exist in the file.
15951 (wx constraint): Likewise.
15952 (wz constraint): Likewise.
15953
15954 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15955 -mdebug=reg, print wg, wl, wx, and wz constraints.
15956 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15957 Initialize the reload functions for 64-bit binary/decimal floating
15958 point types.
15959 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15960 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15961 create the buffer on the stack to overcome not having a 32-bit
15962 load and store.
15963 (rs6000_emit_move): Likewise.
15964 (rs6000_secondary_memory_needed_rtx): Likewise.
15965 (rs6000_alloc_sdmode_stack_slot): Likewise.
15966 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15967 via xxlxor, just like DFmode 0.0.
15968
15969 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
15970 define as 1 if we are running on a power7 or newer.
15971 (enum r6000_reg_class_enum): Add new constraints.
15972
15973 * config/rs6000/dfp.md (movsd): Delete, combine with binary
15974 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15975 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 15976 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM
15977 (movsd splitter): Likewise.
15978 (movsd_hardfloat): Likewise.
15979 (movsd_softfloat): Likewise.
15980
15981 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
15982 binary and decimal floating point moves.
15983 (fmove_ok): New attributes to combine binary and decimal floating
15984 point moves, and to combine power6x (mfpgpr) moves along normal
15985 floating moves.
15986 (real_value_to_target): Likewise.
15987 (f32_lr): Likewise.
15988 (f32_lm): Likewise.
15989 (f32_li): Likewise.
15990 (f32_sr): Likewise.
15991 (f32_sm): Likewise.
15992 (f32_si): Likewise.
15993 (movsf): Combine binary and decimal floating point moves. Combine
15994 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 15995 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM
15996 (mov<mode> for SFmode/SDmode); Likewise.
15997 (SFmode/SDmode splitters): Likewise.
15998 (movsf_hardfloat): Likewise.
15999 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
16000 (movsf_softfloat): Likewise.
16001 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
16002
16003 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
16004 wx and wz constraints.
16005
16006 * config/rs6000/constraints.md (wg constraint): New constraint to
16007 return FLOAT_REGS if -mmfpgpr (power6x) was used.
16008
16009 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
16010 constraint.
16011
16012 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
16013 -mdebug=reg, print wg, wl, wx, and wz constraints.
16014 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
16015 Initialize the reload functions for 64-bit binary/decimal floating
16016 point types.
16017 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
16018 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
16019 create the buffer on the stack to overcome not having a 32-bit
16020 load and store.
16021 (rs6000_emit_move): Likewise.
16022 (rs6000_secondary_memory_needed_rtx): Likewise.
16023 (rs6000_alloc_sdmode_stack_slot): Likewise.
16024 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
16025 via xxlxor, just like DFmode 0.0.
16026
c6d5ff83
MM
16027 * config/rs6000/dfp.md (movdd): Delete, combine with binary
16028 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
16029 with other moves by using conditional constraits (wg). Use LFIWZX
16030 and STFIWX for loading SDmode on power7.
16031 (movdd splitters): Likewise.
16032 (movdd_hardfloat32): Likewise.
16033 (movdd_softfloat32): Likewise.
16034 (movdd_hardfloat64_mfpgpr): Likewise.
16035 (movdd_hardfloat64): Likewise.
16036 (movdd_softfloat64): Likewise.
16037
16038 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
16039 64-bit binary and decimal floating point moves.
16040 (FMOVE64X): Likewise.
16041 (movdf): Combine 64-bit binary and decimal floating point moves.
16042 Combine power6x (mfpgpr) moves with other moves by using
16043 conditional constraits (wg).
16044 (mov<mode> for DFmode/DDmode): Likewise.
16045 (DFmode/DDmode splitters): Likewise.
16046 (movdf_hardfloat32): Likewise.
16047 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
16048 (movdf_softfloat32): Likewise.
16049 (movdf_hardfloat64_mfpgpr): Likewise.
16050 (movdf_hardfloat64): Likewise.
16051 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
16052 (movdf_softfloat64): Likewise.
16053 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
16054 (reload_<mode>_load): Move to later in the file so they aren't in
16055 the middle of the floating point move insns.
16056 (reload_<mode>_store): Likewise.
16057
16058 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
16059 constraint.
16060
16061 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
16062 constraint if -mdebug=reg.
bd059b26
UB
16063 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
16064 Enable using dd reload support if needed.
c6d5ff83
MM
16065
16066 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
16067 binary and decimal floating point moves in rs6000.md.
16068 (movtd_internal): Likewise.
16069
16070 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
16071 decimal floating point moves.
16072 (movtf): Likewise.
16073 (movtf_internal): Likewise.
16074 (mov<mode>_internal, TDmode/TFmode): Likewise.
16075 (movtf_softfloat): Likewise.
16076 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
16077
16078 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
16079 movdi_internal64, using wg constraint for move direct operations.
16080 (movdi_internal64): Likewise.
16081
16082 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
16083 MODES_TIEABLE_P for selected modes. Print the numerical value of
16084 the various virtual registers. Use GPR/FPR first/last values,
16085 instead of hard coding the register numbers. Print which modes
16086 have reload functions registered.
bd059b26
UB
16087 (rs6000_option_override_internal): If -mdebug=reg, trace the options
16088 settings before/after setting cpu, target and subtarget settings.
16089 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
16090 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM
16091 (rs6000_secondary_reload_fail): Likewise.
16092 (rs6000_secondary_reload_inner): Likewise.
16093
16094 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
16095 macros for first/last GPR and FPR registers.
16096 (LAST_GPR_REGNO): Likewise.
16097 (FIRST_FPR_REGNO): Likewise.
16098 (LAST_FPR_REGNO): Likewise.
16099
16100 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
16101 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
16102 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
16103 (vcond<mode><mode>): Likewise.
16104 (vcondu<mode><mode>): Likewise.
16105 (vector_gtu<mode>): Likewise.
16106 (vector_gte<mode>): Likewise.
16107 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 16108 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM
16109 (ior<mode>3): Likewise.
16110 (and<mode>3): Likewise.
16111 (one_cmpl<mode>2): Likewise.
16112 (nor<mode>3): Likewise.
16113 (andc<mode>3): Likewise.
16114
16115 * config/rs6000/constraints.md (wt constraint): New constraint
16116 that returns VSX_REGS if TImode is allowed in VSX registers.
16117
16118 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
16119 constant under VSX.
16120
16121 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
16122 similar to TImode, but it is restricted to being in the GPRs.
16123
16124 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
16125 TImode to occupy a single VSX register.
16126
16127 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
16128 -mvsx-timode for power7/power8.
16129 (power7 cpu): Likewise.
16130 (power8 cpu): Likewise.
16131
16132 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
16133 sure that TFmode/TDmode take up two registers if they are ever
16134 allowed in the upper VSX registers.
16135 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
16136 registers.
16137 (rs6000_init_hard_regno_mode_ok): Likewise.
16138 (rs6000_debug_reg_global): Add debugging for PTImode and wt
16139 constraint. Print if LRA is turned on.
16140 (rs6000_option_override_internal): Give an error if -mvsx-timode
16141 and VSX is not enabled.
16142 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
16143 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
16144 to reg+offset addressing. Use PTImode when checking offset
16145 addresses for validity.
16146 (reg_offset_addressing_ok_p): Likewise.
16147 (rs6000_legitimate_offset_address_p): Likewise.
16148 (rs6000_legitimize_address): Likewise.
16149 (rs6000_legitimize_reload_address): Likewise.
16150 (rs6000_legitimate_address_p): Likewise.
16151 (rs6000_eliminate_indexed_memrefs): Likewise.
16152 (rs6000_emit_move): Likewise.
16153 (rs6000_secondary_reload): Likewise.
16154 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
16155 reloads to fpr registers to continue to use reg+offset addressing,
16156 but 64-bit reloads to altivec registers need reg+reg addressing.
16157 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
16158 it. Treat LO_SUM like a PLUS operation.
16159 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 16160 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM
16161 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
16162 registers to share a register with a smaller sized type, since VSX
16163 puts scalars in the upper 64-bits.
16164 (print_operand): Add support for PTImode.
16165 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
16166 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
16167 registers, but don't have arithmetic support.
16168 (rs6000_memory_move_cost): Add test for VSX.
16169 (rs6000_opt_masks): Add -mvsx-timode.
16170
16171 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
16172 for TImode.
16173 (VSs): Likewise.
16174 (VSr): Use wt constraint for TImode.
16175 (VSv): Drop TImode support.
16176 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
16177 (vsx_movti_64bit): Likewise.
16178 (vsx_movti_32bit): Likewise.
16179 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
16180 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
16181 one '?' on the appropriate output constraint. Do not allow TImode
16182 logical operations on 32-bit systems.
16183 (vsx_ior<mode>3): Likewise.
16184 (vsx_xor<mode>3): Likewise.
16185 (vsx_one_cmpl<mode>2): Likewise.
16186 (vsx_nor<mode>3): Likewise.
16187 (vsx_andc<mode>3): Likewise.
16188 (vsx_concat_<mode>): Likewise.
16189 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
16190
16191 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
16192 OPTION_MASK_VSX_TIMODE.
16193 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
16194 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
16195
16196 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
16197 (TI2 iterator): New iterator for TImode, PTImode.
16198 (wd mode attribute): Add values for vector types.
bd059b26
UB
16199 (movti_string): Replace TI move operations with operations for TImode
16200 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM
16201 (mov<mode>_string, TImode/PTImode): Likewise.
16202 (movti_ppc64): Likewise.
16203 (mov<mode>_ppc64, TImode/PTImode): Likewise.
16204 (TI mode splitters): Likewise.
16205
16206 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
16207 constraint.
16208
1fc5eced
MG
162092013-03-20 Marc Glisse <marc.glisse@inria.fr>
16210
16211 PR tree-optimization/56355
16212 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
16213 Also handle integers with undefined overflow.
16214
22c4c869
CM
162152013-03-20 Catherine Moore <clm@codesourcery.com>
16216 Maciej W. Rozycki <macro@codesourcery.com>
16217 Tom de Vries <tom@codesourcery.com>
12211b99 16218 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM
16219 Iain Sandoe <iain@codesourcery.com>
16220 Nathan Froyd <froydnj@codesourcery.com>
12211b99 16221 Chao-ying Fu <fu@mips.com>
22c4c869
CM
16222
16223 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 16224 Document new function attributes.
22c4c869
CM
16225 * doc/invoke.texi (minterlink-compressed, mmicromips,
16226 m14k, m14ke, m14kec): Document new options.
16227 (minterlink-mips16): Update documentation.
16228 * doc/md.texi (ZC, ZD): Document new constraints.
16229 * configure.ac (gcc_cv_as_micromips): Check if linker
16230 supports the .set micromips directive.
16231 * configure: Regenerate.
16232 * config.in: Regenerate.
16233 * config/mips/mips-tables.opt: Regenerate.
16234 * config/mips/micromips.md: New file.
16235 * constraints.md (ZC, ZD): New constraints.
16236 * config/mips/predicates.md (movep_src_register): New predicate.
16237 (movep_src_operand): New predicate.
16238 (non_volatile_mem_operand): New predicate.
16239 * config/mips/mips.md (multimem): New type.
16240 (length): Differentiate between 17-bit and 18-bit branch offsets.
16241 (MOVEP1, MOVEP2): New mode iterator.
33159866 16242 (mov_<load>l): Use ZC constraint.
22c4c869
CM
16243 (mov_<load>r): Likewise.
16244 (mov_<store>l): Likewise.
16245 (mov_<store>r): Likewise.
16246 (*branch_equality<mode>_inverted): Add microMIPS support.
16247 (*branch_equality<mode>): Likewise.
16248 (*jump_absolute): Likewise.
16249 (indirect_jump_<mode>): Likewise.
16250 (tablejump_<mode>): Likewise.
16251 (<optab>_internal): Likewise.
16252 (sibcall_internal): Likewise.
16253 (sibcall_value_internal): Likewise.
16254 (prefetch): Use constraint ZD.
16255 * config/mips/mips.opt (minterlink-compressed): New option.
16256 (minterlink-mips16): Now an alias for minterlink-compressed.
16257 (mmicromips): New option.
16258 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16259 (compare_and_swap_12): Likewise.
16260 (sync_add<mode>): Likewise.
16261 (sync_<optab>_12): Likewise.
16262 (sync_old_<optab>_12): Likewise.
16263 (sync_new_<optab>_12): Likewise.
16264 (sync_nand_12): Likewise.
16265 (sync_old_nand_12): Likewise.
16266 (sync_new_nand_12): Likewise.
16267 (sync_sub<mode>): Likewise.
16268 (sync_old_add<mode>): Likewise.
16269 (sync_old_sub<mode>): Likewise.
16270 (sync_new_add<mode>): Likewise.
16271 (sync_new_sub<mode>): Likewise.
16272 (sync_<optab><mode>): Likewise.
16273 (sync_old_<optab><mode>): Likewise.
16274 (sync_new_<optab><mode>): Likewise.
16275 (sync_nand<mode>): Likewise.
16276 (sync_old_nand<mode>): Likewise.
16277 (sync_new_nand<mode>): Likewise.
16278 (sync_lock_test_and_set<mode>): Likewise.
16279 (test_and_set_12): Likewise.
16280 (atomic_compare_and_swap<mode>): Likewise.
16281 (atomic_exchange<mode>_llsc): Likewise.
16282 (atomic_fetch_add<mode>_llsc): Likewise.
16283 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16284 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16285 (umips_save_restore_pattern_p): Likewise.
16286 (umips_load_store_pair_p): Likewise.
16287 (umips_output_load_store_pair): Likewise.
16288 (umips_movep_target_p): Likewise.
16289 (umips_12bit_offset_address_p): Likewise.
16290 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16291 (mips_base_mips16): Rename this...
16292 (mips_base_compression_flags): ...to this. Update all uses.
16293 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16294 (mips_mips16_decl_p): Delete.
16295 (mips_nomips16_decl_p): Delete.
33159866 16296 (mips_get_compress_on_flags): New function.
22c4c869
CM
16297 (mips_get_compress_off_flags): New function.
16298 (mips_get_compress_mode): New function.
16299 (mips_get_compress_on_name): New function.
16300 (mips_get_compress_off_name): New function.
16301 (mips_insert_attributes): Support multiple compression types.
16302 (mips_merge_decl_attributes): Likewise.
16303 (umips_12bit_offset_address_p): New function.
16304 (mips_start_function_definition): Emit .set micromips directive.
16305 (mips_call_may_need_jalx_p): New function.
16306 (mips_function_ok_for_sibcall): Add microMIPS support.
16307 (mips_print_operand_punctuation): Support short delay slots and
16308 compact jumps.
16309 (umips_swm_mask, umips_swm_encoding): New.
16310 (umips_build_save_restore): New function.
16311 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16312 (was_mips16_p): Remove.
16313 (old_compression_mode): New.
16314 (mips_set_compression_mode): New function.
16315 (mips_set_current_function): Add microMIPS support.
16316 (mips_option_override): Likewise.
16317 (umips_save_restore_pattern_p): New function.
16318 (umips_output_save_restore): New function.
16319 (umips_load_store_pair_p_1): New function.
16320 (umips_load_store_pair_p): New function.
16321 (umips_output_load_store_pair_1): New function.
16322 (umips_output_load_store_pair): New function.
16323 (umips_movep_target_p) New function.
16324 (mips_prepare_pch_save): Add microMIPS support.
16325 * config/mips/mips.h (TARGET_COMPRESSION): New.
16326 (TARGET_CPU_CPP_BUILTINS): Update macro
16327 to use new compression flags and to support microMIPS.
16328 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16329 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16330 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16331 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16332 (ASM_SPEC): Support mmicromips and mno-micromips.
16333 (M16STORE_REG_P): New macro.
16334 (MIPS_CALL): Support TARGET_MICROMIPS.
16335 (MICROMIPS_J): New macro.
16336 (mips_base_mips16): Rename this...
16337 (mips_base_compression_flags): ...to this.
16338 (UMIPS_12BIT_OFFSET_P): New macro.
16339 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16340 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB
163412013-03-20 Richard Biener <rguenther@suse.de>
16342
16343 PR tree-optimization/56661
16344 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
16345 the result does not have to be distinct.
16346
54714c68
RB
163472013-03-20 Richard Biener <rguenther@suse.de>
16348
16349 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
16350 remap_gimple_op_r.
16351
cca1130d 163522013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 16353 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS
16354
16355 PR rtl-optimization/56605
16356 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
16357
2a930377
UB
163582013-03-20 Uros Bizjak <ubizjak@gmail.com>
16359
16360 PR bootstrap/56656
16361 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
16362 that require movd instead of movq.
16363
d6d305fe
RB
163642013-03-20 Richard Biener <rguenther@suse.de>
16365
16366 * tree-ssa-structalias.c (struct variable_info): Add pointer
16367 to the first field of an aggregate with sub-vars. Make
16368 this and the pointer to the next subfield its ID.
16369 (vi_next): New function.
16370 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
16371 storedanything_id, integer_id): Increment by one.
16372 (new_var_info, get_call_vi, lookup_call_clobber_vi,
16373 get_call_clobber_vi): Adjust.
16374 (solution_set_expand): Simplify and speedup.
16375 (solution_set_add): Inline into ...
16376 (set_union_with_increment): ... this. Adjust accordingly.
16377 (do_sd_constraint): Likewise.
16378 (do_ds_constraint): Likewise.
16379 (do_complex_constraint): Simplify.
16380 (build_pred_graph): Adjust.
16381 (solve_graph): Likewise. Simplify and speedup.
16382 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
16383 get_constraint_for_component_ref, get_constraint_for_1,
16384 first_vi_for_offset, first_or_preceding_vi_for_offset,
16385 create_function_info_for, create_variable_info_for_1,
16386 create_variable_info_for, intra_create_variable_infos): Adjust.
16387 (init_base_vars): Push NULL for ID zero.
16388 (compute_points_to_sets): Adjust.
16389
a271b42d
RB
163902013-03-20 Richard Biener <rguenther@suse.de>
16391
16392 * cfgloop.c (verify_loop_structure): Streamline and avoid
16393 ICEing on corrupt loop tree.
16394 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
16395 loop tree.
16396
ebd65954
RB
163972013-03-20 Richard Biener <rguenther@suse.de>
16398
16399 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
16400 check whether an SSA update is needed.
16401
4547b7ee
RS
164022013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
16403
16404 * config/mips/constraints.md (T): Rename to...
16405 (Yf): ...this.
16406 (U): Rename to...
16407 (Yd): ...this.
16408 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
16409 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
16410
0b8cdc58
IB
164112013-03-19 Ian Bolton <ian.bolton@arm.com>
16412
16413 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
16414 (*subsi3_carryin_uxtw): Likewise.
16415
5977a10d
IB
164162013-03-19 Ian Bolton <ian.bolton@arm.com>
16417
16418 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
16419 (*rorsi3_insn_uxtw): Likewise.
16420
bd83ff2c
IB
164212013-03-19 Ian Bolton <ian.bolton@arm.com>
16422
16423 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
16424 (*extrsi5_insn_uxtw): Likewise.
16425
ba65123c
RB
164262013-03-19 Richard Biener <rguenther@suse.de>
16427
16428 PR tree-optimization/56273
16429 * passes.c (init_optimization_passes): Move second VRP after DOM.
16430
2eac0476
UB
164312013-03-19 Uros Bizjak <ubizjak@gmail.com>
16432
16433 * config/i386/i386.md (*movti_internal): Merge from
16434 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
16435 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
16436 nox64 isa attributes.
16437
6e55eda7
RB
164382013-03-18 Richard Biener <rguenther@suse.de>
16439
16440 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
16441 (unite): Likewise.
16442 (merge_node_constraints): Likewise.
16443 (build_succ_graph): Likewise.
16444 (valid_graph_edge): Inline into single caller.
16445 (unify_nodes): Likewise. Use bitmap_set_bit return value
16446 and cache varinfo.
16447 (scc_visit): Fix formatting and variable use.
16448 (do_sd_constraint): Use gcc_checking_assert.
16449 (do_ds_constraint): Likewise.
16450 (do_complex_constraint): Likewise.
16451 (condense_visit): Likewise. Cleanup.
16452 (dump_pred_graph): New function.
16453 (perform_var_substitution): Dump the pred-graph before
16454 variable substitution.
16455 (find_equivalent_node): Use gcc_checking_assert.
16456 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
16457
4bdd44c4
RB
164582013-03-18 Richard Biener <rguenther@suse.de>
16459
16460 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
16461 Remove cond_expr_stmt_list argument and do not gimplify the
16462 built expression.
16463 (vect_loop_versioning): Adjust.
16464 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
16465 Cleanup to use less temporaries.
16466 (vect_create_data_ref_ptr): Cleanup.
16467
38c56a5b
JJ
164682013-03-18 Jakub Jelinek <jakub@redhat.com>
16469
16470 PR tree-optimization/56635
16471 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
16472 require types_compatible_p types.
16473
20b2e6a0
NC
164742013-03-18 Nick Clifton <nickc@redhat.com>
16475
a6178a25
NC
16476 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
16477 spurious backslash.
16478
20b2e6a0
NC
16479 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
16480 Add missing line to comment describing function.
16481
92608d0e
RB
164822013-03-18 Richard Biener <rguenther@suse.de>
16483
16484 PR tree-optimization/56210
16485 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
16486 Handle string / character search functions.
16487 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
16488
31348d52
RB
164892013-03-18 Richard Biener <rguenther@suse.de>
16490
16491 PR middle-end/56483
16492 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
16493 and implement properly.
16494 * gimple.h (gimple_cond_single_var_p): Remove.
16495
fcac74a1
RB
164962013-03-18 Richard Biener <rguenther@suse.de>
16497
16498 * tree-data-ref.h (find_data_references_in_loop): Declare.
16499 * tree-data-ref.c (get_references_in_stmt): Use a stack
16500 vector pre-allocated in the callers.
16501 (find_data_references_in_stmt): Adjust.
16502 (graphite_find_data_references_in_stmt): Likewise.
16503 (create_rdg_vertices): Likewise.
16504 (find_data_references_in_loop): Export.
16505 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
16506 Compute dependences here...
16507 (vect_analyze_data_refs): ...not here. When we encounter
16508 a non-vectorizable data reference in basic-block vectorization
16509 truncate the data reference vector. Do not bother to
16510 fixup data-dependence information for gather loads.
16511 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
16512 of data references, as reported.
16513
0d5a1b56
RB
165142013-03-18 Richard Biener <rguenther@suse.de>
16515
16516 PR tree-optimization/3713
16517 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
16518 has_constants and expr.
16519 (stmt_has_constants): Properly valueize SSA names when deciding
16520 whether the stmt has constants.
16521
789c34e3
RB
165222013-03-18 Richard Biener <rguenther@suse.de>
16523
16524 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
16525 whole function when there is nothing to do.
16526 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
16527 * tree-vectorizer.c (vectorize_loops): Update virtual and
16528 loop-closed SSA once.
16529 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
16530
076b4605
RB
165312013-03-18 Richard Biener <rguenther@suse.de>
16532
16533 PR middle-end/56113
16534 * domwalk.c (bb_postorder): New global static.
16535 (cmp_bb_postorder): New function.
16536 (walk_dominator_tree): Replace scheme imposing an order for
16537 visiting dominator sons by one sorting them at the time they
16538 are pushed on the stack.
16539
bdb01696
RB
165402013-03-18 Richard Biener <rguenther@suse.de>
16541
16542 PR tree-optimization/39326
16543 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
16544 (struct mem_ref): Replace mem member with ao_ref typed member.
16545 (MEM_ANALYZABLE): Adjust.
16546 (memref_eq): Likewise.
16547 (mem_ref_alloc): Likewise.
16548 (gather_mem_refs_stmt): Likewise.
16549 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
16550 (execute_sm_if_changed_flag_set): Adjust.
16551 (execute_sm): Likewise.
16552 (ref_always_accessed_p): Likewise.
16553 (refs_independent_p): Likewise.
16554 (can_sm_ref_p): Likewise.
16555
12d2dc5e
JJ
165562013-03-18 Jakub Jelinek <jakub@redhat.com>
16557
16558 PR c/56566
16559 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
16560 return 1 even for !unsignedp.
16561
286fb677
UB
165622013-03-17 Uros Bizjak <ubizjak@gmail.com>
16563
16564 * config/i386/i386.md (isa): Add x64 and nox64.
16565 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
16566 (*pushtf): Enable *roF alternative for x64 isa only.
16567 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
16568 mode attribute of integer alternatives to DImode for TARGET_64BIT.
16569 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
16570 (*movtf_internal): Merge from *movtf_internal_rex64 and
16571 *movtf_internal_sse. Use x64 and nox64 isa attributes.
16572 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
16573 nox64 isa attributes.
16574 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
16575 nox64 isa attributes.
16576 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
16577
88b97037
UB
165782013-03-17 Uros Bizjak <ubizjak@gmail.com>
16579
16580 * config/alpha/alpha.c (TARGET_LRA_P): New define.
16581
9f4f1735
JJ
165822013-03-17 Jakub Jelinek <jakub@redhat.com>
16583
16584 PR target/56640
16585 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
16586 class names. Remove trailing comma after "ALL_REGS".
16587
48f4a6fa
JH
165882013-03-16 Jan Hubicka <jh@suse.cz>
16589
16590 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
16591 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
16592 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
16593 of cgraph_get_create_node.
16594 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
16595
98e81123
JM
165962013-03-16 Jason Merrill <jason@redhat.com>
16597
16598 PR debug/49090
16599 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
16600 with DW_AT_default_value.
16601
8a298c04
JJ
166022013-03-16 Jakub Jelinek <jakub@redhat.com>
16603
16604 * BASE-VER: Set to 4.9.0.
16605
4323afa0
AK
166062013-03-14 Andi Kleen <ak@linux.intel.com>
16607
16608 PR target/56619
16609 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
16610 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
16611 Document _x* TSX intrinsics.
16612
b3c0d469
JJ
166132013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
16614 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI
16615
16616 * configure.ac: Add MicroBlaze TLS support detection.
16617 * configure: Regenerate.
88b97037
UB
16618 * config/microblaze/microblaze-protos.h
16619 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
16620 symbol_mentioned_p, label_mentioned_p): Add prototypes.
16621 * config/microblaze/microblaze.c (microblaze_address_type): Add
16622 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI
16623 (microblaze_address_info): Add tls_reloc.
16624 (TARGET_HAVE_TLS): Define.
16625 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
16626 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB
16627 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
16628 load_tls_operand, microblaze_call_tls_get_addr,
16629 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI
16630 (microblaze_classify_unspec): Handle UNSPEC_TLS.
16631 (get_base_reg): Use microblaze_tls_symbol_p.
16632 (microblaze_classify_address): Handle TLS.
88b97037
UB
16633 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
16634 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI
16635 (microblaze_legitimize_address): Handle TLS.
16636 (microblaze_address_insns): Handle ADDRESS_TLS.
16637 (pic_address_needs_scratch): Handle TLS.
16638 (print_operand_address): Handle TLS.
16639 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
16640 (microblaze_expand_move): Handle TLS.
88b97037
UB
16641 (microblaze_legitimate_constant_p): Check
16642 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI
16643 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
16644 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
16645 (PIC_OFFSET_TABLE_REGNUM): Set.
16646 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
16647 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
16648 (addsi3, movsi_internal2, movdf_internal): Update constraints
16649 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB
16650 (move_operand): Redefine as move_src_operand,
16651 check microblaze_tls_referenced_p.
8cc9a5a5 16652
d803a491
IB
166532013-03-14 Ian Bolton <ian.bolton@arm.com>
16654
16655 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
16656 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
16657
a8504f22
IB
166582013-03-14 Ian Bolton <ian.bolton@arm.com>
16659
16660 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
16661 CC mode for AND.
16662
df2dfaea
JJ
166632013-03-14 Jakub Jelinek <jakub@redhat.com>
16664
fbd28bc3
JJ
16665 PR tree-optimization/53265
16666 * common.opt (Waggressive-loop-optimizations): New option.
16667 * tree-ssa-loop-niter.c: Include tree-pass.h.
16668 (do_warn_aggressive_loop_optimizations): New function.
16669 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
16670 if number_of_latch_executions returned constant.
16671 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
16672 early. If number_of_latch_executions returned constant, set
16673 nb_iterations_upper_bound back to it.
16674 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
16675 field.
16676 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
16677 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
16678
df2dfaea
JJ
16679 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
16680 (MULTILIB_OSDIRNAMES): Set.
16681 * genmultilib: If defaultosdirname doesn't start with :: , set
16682 defaultosdirname2 instead, clear it and emit two . multilib_raw
16683 entries instead of just one.
16684
ee0d2b68
KK
166852013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
16686
16687 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
16688 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
16689 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
16690 (SUBTARGET_OVERRIDE_OPTIONS): New.
16691
decc676e
OE
166922013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
16693
16694 PR target/49880
16695 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
16696 (musermode): Convert to Var(TARGET_USERMODE).
16697 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
16698 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
16699 * config/sh/sh.c (sh_option_override): Use
16700 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
16701 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
16702 condition.
16703 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
16704 TARGET_SH4.
16705 (udivsi3_i4_single, divsi3_i4_single): Use
16706 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
16707
f4b719c7
DK
167082013-03-13 Dave Korn <dave.korn.cygwin@....>
16709
16710 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
16711 default setting.
16712
c40eced0
RB
167132013-03-13 Richard Biener <rguenther@suse.de>
16714
16715 PR tree-optimization/56608
16716 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
16717 calls when vectorizing basic-blocks.
16718
1bfa5973
JJ
167192013-03-13 Jakub Jelinek <jakub@redhat.com>
16720
16721 PR plugins/45078
16722 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
16723 tm_file.
16724
6fcf5434
JJ
167252013-03-12 Jakub Jelinek <jakub@redhat.com>
16726
16727 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
16728
4f38fa8c
JH
167292013-03-11 Jan Hubicka <jh@suse.cz>
16730
16731 PR lto/56557
16732 * lto-streamer-out.c (output_symbol_p): Skip references from
16733 constructors of external variables.
16734
c5c5ba89
JH
167352013-03-11 Jan Hubicka <jh@suse.cz>
16736
16737 PR middle-end/56571
16738 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
16739 from pseudos.
16740 * emit-rtl.c (verify_rtx_sharing): Likewise.
16741 (copy_insn_1): Likewise.
16742 * rtl.c (copy_rtx): Likewise.
16743
c2a939b1
GJL
167442013-03-11 Georg-Johann Lay <avr@gjlay.de>
16745
16746 PR target/56591
16747 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
16748 output_operand_lossage message.
88b97037 16749
3c3279fb
RE
167502013-03-11 Richard Earnshaw <rearnsha@arm.com>
16751
16752 PR target/56470
16753 * arm.c (shift_op): Validate RTL pattern on the fly.
16754 (arm_print_operand, case 'S'): Don't use shift_operator to validate
16755 the RTL.
16756
aef5ef9d
JDA
167572013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16758
16759 PR target/56347
16760 * config/pa/pa.md (call_value): Check for calls to powf and direct to
16761 new call patterns that clobber %fr12.
16762 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
16763 split and postreload patterns.
16764 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
16765 registers %fr12 and %fr12R as call used.
16766
456610d3
SB
167672013-03-09 Steven Bosscher <steven@gcc.gnu.org>
16768
16769 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
16770 (canon_address, record_store, replace_read, check_mem_read_rtx,
16771 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
16772 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
16773 rest_of_handle_dse): Likewise.
16774
4b1baac8
RS
167752013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
16776
16777 PR middle-end/56524
16778 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
16779 Add base_optabs.
16780 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
16781 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
16782 (save_optabs_if_changed): Replace with...
16783 (init_tree_optimization_optabs): ...this.
16784 * optabs.c (save_optabs_if_changed): Rename to...
16785 (init_tree_optimization_optabs): ...this. Take the optimization node
16786 as argument. Do nothing if the base optabs are already correct.
16787 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
16788 to recompute optabs.
16789 * function.h (function): Remove optabs field.
16790 * function.c (invoke_set_current_function_hook): Call
16791 init_tree_optimization_optabs. Use the result to initialize
16792 this_fn_optabs.
16793
b7a78683
AH
167942013-02-27 Aldy Hernandez <aldyh@redhat.com>
16795
16796 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
16797 if GTMA_HAS_NO_INSTRUMENTATION.
16798 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
16799 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
16800 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
16801 * gimple-pretty-print.c (dump_gimple_transaction): Handle
16802 GTMA_HAS_NO_INSTRUMENTATION.
16803
6384c29b
JJ
168042013-03-08 Jakub Jelinek <jakub@redhat.com>
16805
16806 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
16807 libasan_preinit.o.
16808
ecd4f20a
MP
168092013-03-08 Marek Polacek <polacek@redhat.com>
16810 Jakub Jelinek <jakub@redhat.com>
16811
16812 PR tree-optimization/56478
16813 * predict.c (is_comparison_with_loop_invariant_p): Change the
16814 type of loop_step to tree.
16815 (predict_loops): Adjust.
16816 (predict_iv_comparison): Perform the computations on double_ints.
16817
64366d35
RB
168182013-03-08 Richard Biener <rguenther@suse.de>
16819
16820 PR tree-optimization/56570
16821 * tree-cfg.c (verify_expr_location_1): Verify locations for
16822 DECL_DEBUG_EXPR.
16823 * tree-sra.c (create_access_replacement): Strip locations
16824 from DECL_DEBUG_EXPRs.
16825
a9d5a059
RB
168262013-03-08 Richard Biener <rguenther@suse.de>
16827
16828 * tree-inline.c (expand_call_inline): Do not associate
16829 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
16830 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
16831
b7aa4e9a
RB
168322013-03-08 Richard Biener <rguenther@suse.de>
16833
16834 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
16835 or block changes with -Og. Fix for location / block encoding
16836 changes and PHI arguments with locations.
16837
c4c2f9fa
SB
168382013-03-07 Steven Bosscher <steven@gcc.gnu.org>
16839
16840 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
16841 for all counters.
16842 (struct output_info): Likewise.
16843 (register_overhead): Remove bad gcc_assert.
16844 (bitmap_find_bit): If there is only a single bitmap element, do not
16845 count a miss as a search.
16846 (print_statistics): Update for counter type changes.
16847 (dump_bitmap_statistics): Likewise. Print headers such that they
16848 are properly lined up with the printed counters.
16849
5bf6606a
JJ
168502013-03-07 Jakub Jelinek <jakub@redhat.com>
16851
16852 PR tree-optimization/56559
16853 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
16854 check that it has only a single use.
16855
2c653d46
RB
168562013-03-07 Richard Biener <rguenther@suse.de>
16857
16858 * doc/invoke.texi (fwhole-program): Discourage use in combination
16859 with -flto.
16860
a72d8780
JJ
168612013-03-06 Jakub Jelinek <jakub@redhat.com>
16862
01a454df
JJ
16863 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
16864
c1781047
JJ
16865 PR tree-optimization/56539
16866 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
16867 instead of GSI_CONTINUE_LINKING as last argument to
16868 force_gimple_operand_gsi. Adjust function comment.
16869
9772c47a
JJ
16870 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
16871 aarch64-cores.def.
16872
a72d8780
JJ
16873 PR middle-end/56548
16874 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
16875 promoted mode, convert the result back to the original mode.
16876
fa5556de
RB
168772013-03-06 Richard Biener <rguenther@suse.de>
16878
16879 PR middle-end/56294
16880 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
16881 (insert_updated_phi_nodes_compare_uids): New function.
16882 (update_ssa): Sort symbols_to_rename after UID before
16883 traversing it to insert PHI nodes.
16884
010403d1
RB
168852013-03-06 Richard Biener <rguenther@suse.de>
16886
16887 PR middle-end/50494
16888 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
16889 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
16890
16891 Revert
16892 2013-02-13 Richard Biener <rguenther@suse.de>
16893
16894 PR lto/50494
16895 * varasm.c (output_constant_def_1): Get the decl representing
16896 the constant as argument.
16897 (output_constant_def): Wrap output_constant_def_1.
16898 (make_decl_rtl): Use output_constant_def_1 with the decl
16899 representing the constant.
16900 (build_constant_desc): Optionally re-use a decl already
16901 representing the constant.
16902 (tree_output_constant_def): Adjust.
16903
3c27ce4c
JY
169042013-03-06 Joey Ye <joey.ye@arm.com>
16905
16906 PR lto/50293
16907 * gcc.c (convert_white_space): New function.
16908 (main): Handles white space in function name.
16909
8f6d1c86
OE
169102013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
16911
16912 PR target/56529
16913 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
16914 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
16915 to SH_DIV_CALL_TABLE for TARGET_SH2.
16916 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
16917 list.
16918 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
16919 call-table options.
16920
798a209f
SA
169212013-03-05 Sterling Augustine <saugustine@google.com>
16922 Cary Coutant <ccoutant@google.com>
16923
16924 PR debug/55364
16925 * dwarf2out.c (resolve_addr): Don't call
16926 remove_loc_list_addr_table_entries a second time for the same
16927 expression.
16928
6cfa417f 169292013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 16930
6cfa417f
JJ
16931 PR debug/56510
16932 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
16933 (avoid_complex_debug_insns): New function.
16934 (expand_debug_locations): Call it.
16935
16936 PR rtl-optimization/56484
16937 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
16938 lifetimes of hard registers on small register class machines.
16939
169402013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16941
16942 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 16943 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 16944 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH
16945 fast_interrupt.
16946 (microblaze_fast_interrupt_function_p): New function.
16947 (microblaze_is_interrupt_handler): Rename to
16948 microblaze_is_interrupt_variant and add fast_interrupt check.
16949 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
16950 (save_restore_insns): Likewise.
16951 (compute_frame_size): Likewise.
16952 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
16953 (microblaze_globalize_label): Likewise.
6cfa417f
JJ
16954 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
16955 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH
16956 microblaze_is_interrupt_variant.
16957
731aefac
KT
169582013-03-05 Kai Tietz <ktietz@redhat.com>
16959
16960 * sdbout.c (sdbout_one_type): Switch to current function's section
16961 supporting cold/hot.
16962
a72d8780 169632013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH
16964
16965 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
16966 -mxl-reorder.
16967
334e71e8
JJ
169682013-03-05 Jakub Jelinek <jakub@redhat.com>
16969
0b50e654
JJ
16970 PR middle-end/56461
16971 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
16972 if VALGRIND_GET_VBITS is defined, temporarily make object
16973 memory all defined, and restore previous valgrind addressability
16974 and definability afterwards. Free this_object at the end.
16975
4ccf8f43
JJ
16976 PR middle-end/56461
16977 * lra.c (lra): Call lra_clear_live_ranges if live_p,
16978 right before calling lra_create_live_ranges, also call it
16979 when clearing live_p. Only call lra_clear_live_ranges
16980 at the end if live_p.
16981
334e71e8
JJ
16982 PR middle-end/56461
16983 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
16984
9ca966ca
RB
169852013-03-05 Richard Biener <rguenther@suse.de>
16986
16987 PR tree-optimization/56521
16988 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
16989 value-id.
16990
d273b176
SB
169912013-03-05 Steven Bosscher <steven@gcc.gnu.org>
16992
16993 PR c++/55135
16994 * except.h (remove_unreachable_eh_regions): New prototype.
16995 * except.c (remove_eh_handler_splicer): New function, split out
16996 of remove_eh_handler.
16997 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
16998 warning about running it on many EH regions one at a time.
16999 (remove_unreachable_eh_regions_worker): New function, walk the
17000 EH tree in depth-first order and remove non-marked regions.
17001 (remove_unreachable_eh_regions): New function.
17002 * tree-eh.c (mark_reachable_handlers): New function, split out
17003 from remove_unreachable_handlers.
17004 (remove_unreachable_handlers): Use mark_reachable_handlers and
17005 remove_unreachable_eh_regions.
17006 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
17007 and remove_unreachable_eh_regions.
17008
5e7f1aef
RB
170092013-03-05 Richard Biener <rguenther@suse.de>
17010
17011 PR middle-end/56525
17012 * loop-init.c (fix_loop_structure): Remove loops in two stages,
17013 not freeing them until the end.
17014
f276b762
AK
170152013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17016
17017 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
17018
12a43ab8
RB
170192013-03-05 Richard Biener <rguenther@suse.de>
17020
17021 PR tree-optimization/56270
17022 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
17023 of loads after scheduling an SLP instance.
17024
93675444
JJ
170252013-03-05 Jakub Jelinek <jakub@redhat.com>
17026
db4138e3
JJ
17027 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
17028 tic6x.exp.
17029 (check_gcc_parallelize): Run guality.exp as a separate job from
17030 vect.exp with unsorted.exp and $(dg_target_exps) separately from
17031 struct-layout-1.exp with stackalign.exp.
17032
dd3d1ec0
JJ
17033 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
17034
b8d381a3
JJ
17035 PR middle-end/56461
17036 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
17037 load_index sbitmap even if some bit in it isn't set.
17038
b4f9786b
JJ
17039 PR middle-end/56461
17040 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
17041 (discover_iteration_bound_by_body_walk): Change queues to
17042 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
17043 spelling in comment. Call safe_push on queues[bound_index] directly.
17044 Release queues[queue_index] in every iteration unconditionally.
17045 Release bounds vector.
17046
93675444
JJ
17047 PR middle-end/56461
17048 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
17049 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
17050 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
17051 inner_phis vector.
17052
3e492e9c
RB
170532013-03-05 Richard Biener <rguenther@suse.de>
17054
17055 PR lto/56515
17056 * tree-inline.c (remap_blocks_to_null): New function.
17057 (expand_call_inline): When expanding a call stmt without
17058 an associated block inline remap all callee blocks to NULL.
17059
a5d1569a
JJ
170602013-03-05 Jakub Jelinek <jakub@redhat.com>
17061
86efb5cd
JJ
17062 PR rtl-optimization/56494
17063 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
17064 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
17065 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
17066
85f5dbea
JJ
17067 PR middle-end/56461
17068 * sel-sched-ir.c (free_sched_pools): Release
17069 succs_info_pool.stack[succs_info_pool.max_top] vectors too
17070 if succs_info_pool.max_top isn't -1.
17071
a5d1569a
JJ
17072 PR bootstrap/56509
17073 * opts.c (opts_obstack, opts_concat): Moved to...
17074 * opts-common.c (opts_obstack, opts_concat): ... here.
17075
4432aa6c
JJ
170762013-03-04 Jakub Jelinek <jakub@redhat.com>
17077
17078 PR middle-end/56461
17079 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
17080
f3a1fb91
MJ
170812013-03-04 Martin Jambor <mjambor@suse.cz>
17082
17083 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
17084 all appropriate places.
17085
22110e6c
EB
170862013-01-04 Eric Botcazou <ebotcazou@adacore.com>
17087
17088 PR tree-optimization/56424
17089 * ipa-split.c (split_function): Do not set the RSO flag if result is
17090 not by reference and its type is a register type.
17091
a72d8780 170922013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 17093
88b97037 17094 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 17095 (microblaze_legitimate_pic_operand): Likewise
88b97037 17096 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH
17097 new function microblaze_legitimate_pic_operand
17098 * config/microblaze/microblaze-protos.h
17099 (microblaze_legitimate_pic_operand): Declare.
17100
a72d8780 171012013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 17102
a72d8780 17103 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 17104 New predicate for supported rtx code types.
a72d8780 17105 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI
17106 call_insn_simple_operand predicate.
17107
541d9ac8
JJ
171082013-03-04 Jakub Jelinek <jakub@redhat.com>
17109
5eb010bc
JJ
17110 PR middle-end/56461
17111 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
17112 partitions.ordered_remove.
17113
30862efc
JJ
17114 PR middle-end/56461
17115 * tree-vect-stmts.c (vectorizable_conversion): Don't call
17116 vec_oprnds0.create (1) for modifier == NONE.
17117
8930f723
JJ
17118 PR middle-end/56461
17119 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
17120 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
17121 vec_oprnds1 right before pushing anything to it for
17122 scalar_shift_arg.
17123
541d9ac8
JJ
17124 PR middle-end/56461
17125 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
17126 set nbbs to 0 instead of having separate code path.
17127 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
17128 instead of false as last argument if returning NULL.
17129
7aa7f2e3
SL
171302013-03-03 Sandra Loosemore <sandra@codesourcery.com>
17131
17132 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
17133 the attribute is now called "target" instead of "option".
17134 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
17135 * doc/tm.texi.in (Target Attributes): Likewise document the correct
17136 attribute/pragma name for TARGET_OPTION_VALID_P and
17137 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
17138 * doc/tm.texi: Regenerated.
17139
8930f723 171402013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 17141
8930f723 17142 * config/microblaze/microblaze.c:
cb8a1637 17143 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ
17144 * config/microblaze/microblaze.h: Add -mxl-reorder to
17145 DRIVER_SELF_SPECS.
17146 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
17147 instructions emitted if TARGET_REORDER.
88b97037
UB
17148 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
17149 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 17150 separately.
cb8a1637 17151
4dc1d68c
XDL
171522013-03-01 Xinliang David Li <davidxl@google.com>
17153
17154 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
17155 walk length.
17156
689eaba3
JJ
171572013-03-01 Jakub Jelinek <jakub@redhat.com>
17158
9d676bf6
JJ
17159 PR middle-end/56461
17160 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
17161 vector even when returning true. Fix up function comment formatting.
17162
caff7edf
JJ
17163 PR middle-end/56461
17164 * ira-build.c (ira_loop_nodes_count): New variable.
17165 (create_loop_tree_nodes): Initialize it.
17166 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
17167
b6b9227d
JJ
17168 PR middle-end/56461
17169 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
17170 method on dr_chain and result_chain.
17171 * tree-vect-stmts.c (vectorizable_store): Only call
17172 result_chain.create if j == 0.
17173
689eaba3
JJ
17174 PR middle-end/56461
17175 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
17176 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
17177 before overwriting it.
17178
40bf31ed
TB
171792013-03-01 Tobias Burnus <burnus@net-b.de>
17180
17181 * doc/extended.texi (C Extensions): Change order in @menu
17182 to match @node.
17183 (Other MIPS Built-in Functions): Move last MIPS entry before
17184 "picoChip Built-in Functions".
17185 (SH Built-in Functions): Move after RX Built-in Functions.
17186 * doc/gcc.texi (Introduction): Change order in @menu
17187 to match @node.
17188 * doc/md.texi (Constraints): Ditto.
17189 * gty.texi (Type Information): Ditto.
17190 (User-provided marking routines for template types): Make
17191 subsection.
17192 * doc/invoke.texi (AArch64 Options): Move before
17193 "Adapteva Epiphany Options".
17194
e664c61c
KS
171952013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
17196 Jakub Jelinek <jakub@redhat.com>
17197
17198 PR sanitizer/56454
17199 * asan.c (gate_asan): Lookup no_sanitize_address instead of
17200 no_address_safety_analysis attribute.
17201 * doc/extend.texi (no_address_safety_attribute): Rename to
17202 no_sanitize_address attribute, mention no_address_safety_analysis
17203 attribute as deprecated alias.
17204
37b5ec8f
JJ
172052013-02-28 Jakub Jelinek <jakub@redhat.com>
17206
17207 PR middle-end/56461
17208 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
17209 type to vec<vec<tree> > *.
17210 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
17211 to be vec<tree> instead of vec<tree> *, set vec_defs
17212 to vNULL and call vec_defs.create (number_of_vects), adjust other
17213 uses of vec_defs.
17214 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
17215 vectorizable_condition): Adjust vect_get_slp_defs callers.
17216
ba96cdfb
JG
172172013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17218
17219 * config/aarch64/aarch64.c
17220 (aarch64_float_const_representable): Remove unused variable.
17221
6f549691
JG
172222013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17223
17224 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
17225
af55e82d
JG
172262013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17227
17228 * config/aarch64/aarch64-builtins.c
17229 (aarch64_init_simd_builtins): Make static.
17230
1df3f464
JG
172312013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
17232
17233 * config/aarch64/aarch64.c
17234 (aarch64_simd_make_constant): Make static.
17235
f8f42513
MJ
172362013-02-28 Martin Jambor <mjambor@suse.cz>
17237
17238 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
17239 with no initialization to the RHS of debug statements.
17240
b48b3fc4
MJ
172412013-02-28 Martin Jambor <mjambor@suse.cz>
17242
17243 PR tree-optimization/56294
17244 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
17245 Adjust dumping.
17246 (get_access_replacement): Do not call create_access_replacement.
17247 Assert a replacement exists.
17248 (get_repl_default_def_ssa_name): Create the replacement declaration
17249 itself.
17250
c3ae224c
RR
172512013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17252
17253 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
17254 final_end_function.
17255
45fa0eef
MP
172562013-02-28 Marek Polacek <polacek@redhat.com>
17257
17258 PR rtl-optimization/56466
17259 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
17260 if we're changing a loop.
17261 (peel_loops_completely): Likewise.
17262
502c067d
PC
172632013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
17264
17265 PR c++/55813
17266 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
17267
ff24afc8
GJL
172682013-02-28 Georg-Johann Lay <avr@gjlay.de>
17269
17270 PR target/56445
17271 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
17272 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
17273 INTX_FTYPE_FX, FX_FTYPE_INTX.
17274 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
17275
c7b1fc1b
GJL
172762013-02-28 Georg-Johann Lay <avr@gjlay.de>
17277
17278 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
17279 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
17280 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
17281 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
17282 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
17283 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
17284 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
17285 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
17286 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
17287 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
17288 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
17289 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
17290 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
17291 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
17292 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
17293 (avrxmega6): Increase max flash segments from 5 to 6.
17294 * config/avr/t-multilib: Regenerate.
17295 * config/avr/avr-tables.opt: Regenerate.
17296 * doc/avr-mmcu.texi: Regenerate.
17297
4a0e3cfe
GJL
172982013-02-28 Georg-Johann Lay <avr@gjlay.de>
17299
17300 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
17301 (avr_device_to_arch): Rename to avr_device_to_ld.
17302 (avr_device_to_as): New prototype.
17303 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
17304 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
17305 * config/avr/driver-avr.c (avr_device_to_as): New.
17306 (avr_device_to_arch): Rename to avr_device_to_ld.
17307
97785e52
JJ
173082013-02-27 Jakub Jelinek <jakub@redhat.com>
17309
3f292312
JJ
17310 PR middle-end/56461
17311 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
17312 method on dr_chain and result_chain.
17313
a344216b
JJ
17314 PR middle-end/56461
17315 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
17316 pointer_set_destroy on not_executed_last_iteration.
17317
f121ad02 17318 PR middle-end/56461
88b97037 17319 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 17320
307f83a3
JJ
17321 PR middle-end/56461
17322 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
17323 FOR_EACH_DEFINED_FUNCTION when freeing state.
17324
e19624ee
JJ
17325 PR middle-end/56461
17326 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
17327 pool_free.
17328 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
17329 overwriting it.
17330
90e709fd
JJ
17331 PR middle-end/56461
17332 * ipa-cp.c (decide_whether_version_node): Call vec_free on
17333 known_aggs[i].items and release known_aggs vector.
17334
97785e52
JJ
17335 PR middle-end/56461
17336 * ipa-reference.c (propagate): Free node_info even for alias nodes.
17337
173382013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 17339
97785e52
JJ
17340 * config/microblaze/microblaze.c (microblaze_emit_compare):
17341 Use xor for EQ/NE comparisions.
17342 * config/microblaze/microblaze.md (cstoresf4): Add constraints
17343 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 17344
6fa5e0ed
JJ
173452013-02-27 Jakub Jelinek <jakub@redhat.com>
17346
17347 PR middle-end/56461
17348 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
17349 vector.
17350 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
17351 vec_safe_push, always update *slot.
17352 (redirect_edge_var_map_clear): Use vec_free.
17353 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
17354 (free_var_map_entry): Use vec_free.
17355 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
17356 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
17357
436a956a
AB
173582013-02-27 Andrey Belevantsev <abel@ispras.ru>
17359
17360 PR middle-end/45472
436a956a
AB
17361 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
17362 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB
17363 Reorder tests for speculativeness in the logical and operator.
17364
f5c2caca
JJ
173652013-02-27 Jakub Jelinek <jakub@redhat.com>
17366
0fcb564b
JJ
17367 * incpath.c (add_standard_paths): Use reconcat instead of concat
17368 where appropriate and avoid leaking memory.
17369
dc357798
JJ
17370 * opts.h: Include obstack.h.
17371 (opts_concat): New prototype.
17372 (opts_obstack): New declaration.
17373 * opts.c (opts_concat): New function.
17374 (opts_obstack): New variable.
17375 (init_options_struct): Call gcc_init_obstack on opts_obstack.
17376 (finish_options): Use opts_concat instead of concat
17377 and XOBNEWVEC instead of XNEWVEC.
17378 * opts-common.c (generate_canonical_option, decode_cmdline_option,
17379 generate_option): Likewise.
17380 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
17381 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
17382
f5c2caca
JJ
17383 PR target/56455
17384 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
17385 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
17386
d0163673
JJ
173872013-02-26 Jakub Jelinek <jakub@redhat.com>
17388
17389 PR middle-end/56461
17390 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
17391
b5ad2b8e
JR
173922013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17393
17394 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
17395 (arm_block_move_unaligned_straight): Likewise.
17396 (arm_adjust_block_mem): Likewise.
17397
9b639e2c
JR
173982013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
17399
17400 PR target/48901
17401 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
17402 temp, cond and label.
17403 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
17404
e13a0ccb
JR
17405 PR target/52500
17406 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
17407 * config/c6x/c6x.h (dbx_register_map): Update declaration.
17408
fbe4f171
JR
17409 PR target/52501
17410 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
17411 of prologue/epilogue functions.
17412
ae006543
JR
17413 PR target/52550
17414 * config/tilegx/tilegx.c (tilegx_expand_prologue):
17415 Remove unused variable cfa_offset.
17416 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
17417
c81369fa
JR
17418 PR target/54639
17419 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
17420 type promotion to unsigned.
17421
f8a8fea7
JR
17422 PR target/54640
17423 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
17424 for HOST_WIDE_INT of 32 bit / same size as int.
17425 (arm_block_move_unaligned_straight): Likewise.
17426 (arm_adjust_block_mem): Likewise.
17427
f8be5169
JR
17428 PR target/54662
17429 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
17430 ALL_CFLAGS.
17431
f1ad3354
MP
174322013-02-26 Marek Polacek <polacek@redhat.com>
17433
17434 PR tree-optimization/56426
88b97037 17435 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 17436
a6af0f42
RB
174372013-02-26 Richard Biener <rguenther@suse.de>
17438
17439 PR target/56444
17440 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
17441 unused variable loops.
17442
aca43c6c
JJ
174432013-02-26 Jakub Jelinek <jakub@redhat.com>
17444
cecbe5d9
JJ
17445 PR tree-optimization/56448
17446 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
17447 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
17448 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
17449 later operands of the references, or even first operand for
17450 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
17451
aca43c6c
JJ
17452 PR tree-optimization/56443
17453 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
17454 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
17455 to type_for_mode langhook.
17456
53e2e141
MT
174572013-02-25 Matt Turner <mattst88@gmail.com>
17458
17459 * doc/invoke.texi: Document r4700.
17460
259ee451
RB
174612013-02-25 Richard Biener <rguenther@suse.de>
17462
17463 PR tree-optimization/56175
17464 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
17465 split out from ...
17466 (simplify_bitwise_binary): ... here. Also guard the conversion
17467 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
17468
6d65e8f1
CM
174692013-02-25 Catherine Moore <clm@codesourcery.com>
17470
17471 Revert:
17472 2013-02-24 Catherine Moore <clm@codesourcery.com>
17473 Maciej W. Rozycki <macro@codesourcery.com>
17474 Tom de Vries <tom@codesourcery.com>
a72d8780 17475 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM
17476 Iain Sandoe <iain@codesourcery.com>
17477 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 17478 Chao-ying Fu <fu@mips.com>
88b97037 17479
6d65e8f1 17480 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 17481 Document new function attributes.
6d65e8f1
CM
17482 * doc/invoke.texi (minterlink-compressed, mmicromips,
17483 m14k, m14ke, m14kec): Document new options.
17484 (minterlink-mips16): Update documentation.
17485 * doc/md.texi (ZC, ZD): Document new constraints.
17486 * configure.ac (gcc_cv_as_micromips): Check if linker
17487 supports the .set micromips directive.
17488 * configure: Regenerate.
17489 * config.in: Regenerate.
17490 * config/mips/mips-tables.opt: Regenerate.
17491 * config/mips/micromips.md: New file.
17492 * constraints.md (ZC, AD): New constraints.
17493 * config/mips/predicates.md (movep_src_register): New predicate.
17494 (movep_src_operand): New predicate.
17495 (non_volatile_mem_operand): New predicate.
17496 * config/mips/mips.md (multimem): New type.
17497 (length): Differentiate between 17-bit and 18-bit branch offsets.
17498 (MOVEP1, MOVEP2): New mode iterator.
17499 (mov_<load>l): Use ZC constraint.
17500 (mov_<load>r): Likewise.
17501 (mov_<store>l): Likewise.
17502 (mov_<store>r): Likewise.
17503 (*branch_equality<mode>_inverted): Add microMIPS support.
17504 (*branch_equality<mode>): Likewise.
17505 (*jump_absolute): Likewise.
17506 (indirect_jump_<mode>): Likewise.
17507 (tablejump_<mode>): Likewise.
17508 (<optab>_internal): Likewise.
17509 (sibcall_internal): Likewise.
17510 (sibcall_value_internal): Likewise.
17511 (prefetch): Use constraint ZD.
17512 * config/mips/mips.opt (minterlink-compressed): New option.
17513 (minterlink-mips16): Now an alias for minterlink-compressed.
17514 (mmicromips): New option.
17515 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17516 (compare_and_swap_12): Likewise.
17517 (sync_add<mode>): Likewise.
17518 (sync_<optab>_12): Likewise.
17519 (sync_old_<optab>_12): Likewise.
17520 (sync_new_<optab>_12): Likewise.
17521 (sync_nand_12): Likewise.
17522 (sync_old_nand_12): Likewise.
17523 (sync_new_nand_12): Likewise.
17524 (sync_sub<mode>): Likewise.
17525 (sync_old_add<mode>): Likewise.
17526 (sync_old_sub<mode>): Likewise.
17527 (sync_new_add<mode>): Likewise.
17528 (sync_new_sub<mode>): Likewise.
17529 (sync_<optab><mode>): Likewise.
17530 (sync_old_<optab><mode>): Likewise.
17531 (sync_new_<optab><mode>): Likewise.
17532 (sync_nand<mode>): Likewise.
17533 (sync_old_nand<mode>): Likewise.
17534 (sync_new_nand<mode>): Likewise.
17535 (sync_lock_test_and_set<mode>): Likewise.
17536 (test_and_set_12): Likewise.
17537 (atomic_compare_and_swap<mode>): Likewise.
17538 (atomic_exchange<mode>_llsc): Likewise.
17539 (atomic_fetch_add<mode>_llsc): Likewise.
17540 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17541 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17542 (umips_save_restore_pattern_p): Likewise.
17543 (umips_load_store_pair_p): Likewise.
17544 (umips_output_load_store_pair): Likewise.
17545 (umips_movep_target_p): Likewise.
17546 (umips_12bit_offset_address_p): Likewise.
17547 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17548 (mips_base_mips16): Rename this...
17549 (mips_base_compression_flags): ...to this. Update all uses.
17550 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17551 (mips_mips16_decl_p): Delete.
17552 (mips_nomips16_decl_p): Delete.
17553 (mips_get_compress_on_flags): New function.
17554 (mips_get_compress_off_flags): New function.
17555 (mips_get_compress_mode): New function.
17556 (mips_get_compress_on_name): New function.
17557 (mips_get_compress_off_name): New function.
17558 (mips_insert_attributes): Support multiple compression types.
17559 (mips_merge_decl_attributes): Likewise.
17560 (umips_12bit_offset_address_p): New function.
17561 (mips_start_function_definition): Emit .set micromips directive.
17562 (mips_call_may_need_jalx_p): New function.
17563 (mips_function_ok_for_sibcall): Add microMIPS support.
17564 (mips_print_operand_punctuation): Support short delay slots and
17565 compact jumps.
17566 (umips_swm_mask, umips_swm_encoding): New.
17567 (umips_build_save_restore): New function.
17568 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17569 (was_mips16_p): Remove.
17570 (old_compression_mode): New.
17571 (mips_set_compression_mode): New function.
17572 (mips_set_current_function): Add microMIPS support.
17573 (mips_option_override): Likewise.
17574 (umips_save_restore_pattern_p): New function.
17575 (umips_output_save_restore): New function.
17576 (umips_load_store_pair_p_1): New function.
17577 (umips_load_store_pair_p): New function.
17578 (umips_output_load_store_pair_1): New function.
17579 (umips_output_load_store_pair): New function.
17580 (umips_movep_target_p) New function.
17581 (mips_prepare_pch_save): Add microMIPS support.
17582 * config/mips/mips.h (TARGET_COMPRESSION): New.
17583 (TARGET_CPU_CPP_BUILTINS): Update macro
17584 to use new compression flags and to support microMIPS.
17585 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17586 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17587 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17588 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17589 (ASM_SPEC): Support mmicromips and mno-micromips.
17590 (M16STORE_REG_P): New macro.
17591 (MIPS_CALL): Support TARGET_MICROMIPS.
17592 (MICROMIPS_J): New macro.
17593 (mips_base_mips16): Rename this...
17594 (mips_base_compression_flags): ...to this.
17595 (UMIPS_12BIT_OFFSET_P): New macro.
17596 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17597 (MULTILIB_DIRNAMES): Likewise.
17598
5e5df392
TV
175992013-02-25 Tom de Vries <tom@codesourcery.com>
17600
17601 PR rtl-optimization/56131
17602 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
17603 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
17604 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
17605
3fdb53c1
TB
176062013-02-25 Tobias Burnus <burnus@net-b.de>
17607
17608 * doc/invoke.texi (-fsanitize=): Move from optimization
17609 to debugging options.
17610
ed358aea
AB
176112013-02-25 Andrey Belevantsev <abel@ispras.ru>
17612
17613 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
17614
f45e9053
AB
176152013-02-25 Andrey Belevantsev <abel@ispras.ru>
17616 Alexander Monakov <amonakov@ispras.ru>
17617
17618 PR middle-end/56077
17619 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
17620 flush pending lists also on non-jumps. Adjust comment.
17621
6941b508
CM
176222013-02-24 Catherine Moore <clm@codesourcery.com>
17623 Maciej W. Rozycki <macro@codesourcery.com>
17624 Tom de Vries <tom@codesourcery.com>
a72d8780 17625 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM
17626 Iain Sandoe <iain@codesourcery.com>
17627 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 17628 Chao-ying Fu <fu@mips.com>
6941b508
CM
17629
17630 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 17631 Document new function attributes.
6941b508
CM
17632 * doc/invoke.texi (minterlink-compressed, mmicromips,
17633 m14k, m14ke, m14kec): Document new options.
17634 (minterlink-mips16): Update documentation.
17635 * doc/md.texi (ZC, ZD): Document new constraints.
17636 * configure.ac (gcc_cv_as_micromips): Check if linker
17637 supports the .set micromips directive.
17638 * configure: Regenerate.
17639 * config.in: Regenerate.
17640 * config/mips/mips-tables.opt: Regenerate.
17641 * config/mips/micromips.md: New file.
17642 * constraints.md (ZC, AD): New constraints.
17643 * config/mips/predicates.md (movep_src_register): New predicate.
17644 (movep_src_operand): New predicate.
17645 (non_volatile_mem_operand): New predicate.
17646 * config/mips/mips.md (multimem): New type.
17647 (length): Differentiate between 17-bit and 18-bit branch offsets.
17648 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 17649 (mov_<load>l): Use ZC constraint.
6941b508
CM
17650 (mov_<load>r): Likewise.
17651 (mov_<store>l): Likewise.
17652 (mov_<store>r): Likewise.
17653 (*branch_equality<mode>_inverted): Add microMIPS support.
17654 (*branch_equality<mode>): Likewise.
17655 (*jump_absolute): Likewise.
17656 (indirect_jump_<mode>): Likewise.
17657 (tablejump_<mode>): Likewise.
17658 (<optab>_internal): Likewise.
17659 (sibcall_internal): Likewise.
17660 (sibcall_value_internal): Likewise.
17661 (prefetch): Use constraint ZD.
17662 * config/mips/mips.opt (minterlink-compressed): New option.
17663 (minterlink-mips16): Now an alias for minterlink-compressed.
17664 (mmicromips): New option.
17665 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17666 (compare_and_swap_12): Likewise.
17667 (sync_add<mode>): Likewise.
17668 (sync_<optab>_12): Likewise.
17669 (sync_old_<optab>_12): Likewise.
17670 (sync_new_<optab>_12): Likewise.
17671 (sync_nand_12): Likewise.
17672 (sync_old_nand_12): Likewise.
17673 (sync_new_nand_12): Likewise.
17674 (sync_sub<mode>): Likewise.
17675 (sync_old_add<mode>): Likewise.
17676 (sync_old_sub<mode>): Likewise.
17677 (sync_new_add<mode>): Likewise.
17678 (sync_new_sub<mode>): Likewise.
17679 (sync_<optab><mode>): Likewise.
17680 (sync_old_<optab><mode>): Likewise.
17681 (sync_new_<optab><mode>): Likewise.
17682 (sync_nand<mode>): Likewise.
17683 (sync_old_nand<mode>): Likewise.
17684 (sync_new_nand<mode>): Likewise.
17685 (sync_lock_test_and_set<mode>): Likewise.
17686 (test_and_set_12): Likewise.
17687 (atomic_compare_and_swap<mode>): Likewise.
17688 (atomic_exchange<mode>_llsc): Likewise.
17689 (atomic_fetch_add<mode>_llsc): Likewise.
17690 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17691 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17692 (umips_save_restore_pattern_p): Likewise.
17693 (umips_load_store_pair_p): Likewise.
17694 (umips_output_load_store_pair): Likewise.
17695 (umips_movep_target_p): Likewise.
17696 (umips_12bit_offset_address_p): Likewise.
17697 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17698 (mips_base_mips16): Rename this...
17699 (mips_base_compression_flags): ...to this. Update all uses.
17700 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17701 (mips_mips16_decl_p): Delete.
17702 (mips_nomips16_decl_p): Delete.
3fdb53c1 17703 (mips_get_compress_on_flags): New function.
6941b508
CM
17704 (mips_get_compress_off_flags): New function.
17705 (mips_get_compress_mode): New function.
17706 (mips_get_compress_on_name): New function.
17707 (mips_get_compress_off_name): New function.
17708 (mips_insert_attributes): Support multiple compression types.
17709 (mips_merge_decl_attributes): Likewise.
17710 (umips_12bit_offset_address_p): New function.
17711 (mips_start_function_definition): Emit .set micromips directive.
17712 (mips_call_may_need_jalx_p): New function.
17713 (mips_function_ok_for_sibcall): Add microMIPS support.
17714 (mips_print_operand_punctuation): Support short delay slots and
17715 compact jumps.
17716 (umips_swm_mask, umips_swm_encoding): New.
17717 (umips_build_save_restore): New function.
17718 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17719 (was_mips16_p): Remove.
17720 (old_compression_mode): New.
17721 (mips_set_compression_mode): New function.
17722 (mips_set_current_function): Add microMIPS support.
17723 (mips_option_override): Likewise.
17724 (umips_save_restore_pattern_p): New function.
17725 (umips_output_save_restore): New function.
17726 (umips_load_store_pair_p_1): New function.
17727 (umips_load_store_pair_p): New function.
17728 (umips_output_load_store_pair_1): New function.
17729 (umips_output_load_store_pair): New function.
17730 (umips_movep_target_p) New function.
17731 (mips_prepare_pch_save): Add microMIPS support.
17732 * config/mips/mips.h (TARGET_COMPRESSION): New.
17733 (TARGET_CPU_CPP_BUILTINS): Update macro
17734 to use new compression flags and to support microMIPS.
17735 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17736 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17737 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17738 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17739 (ASM_SPEC): Support mmicromips and mno-micromips.
17740 (M16STORE_REG_P): New macro.
17741 (MIPS_CALL): Support TARGET_MICROMIPS.
17742 (MICROMIPS_J): New macro.
17743 (mips_base_mips16): Rename this...
17744 (mips_base_compression_flags): ...to this.
17745 (UMIPS_12BIT_OFFSET_P): New macro.
17746 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17747 (MULTILIB_DIRNAMES): Likewise.
17748
cdebbc6d
JJ
177492013-02-24 Jakub Jelinek <jakub@redhat.com>
17750
17751 PR target/52555
17752 * target-globals.c (save_target_globals): For init_reg_sets and
17753 target_reinit remporarily set this_fn_optabs to this_target_optabs.
17754
18c63565
JG
177552013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
17756
17757 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
17758 * config/aarch64/t-aarch64
17759 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
17760
2194f7a2
VM
177612013-02-22 Vladimir Makarov <vmakarov@redhat.com>
17762
17763 PR inline-asm/56148
17764 * lra-constraints.c (process_alt_operands): Reload operand
17765 conflicting with earlier clobber only if no more other conflicting
17766 operands.
17767
7d613735
JJ
177682013-02-22 Jakub Jelinek <jakub@redhat.com>
17769
17770 PR sanitizer/56393
17771 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
17772 if not linking a shared library.
17773
ac8d93a7
SL
177742013-02-22 Seth LaForge <sethml@google.com>
17775
17776 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
17777
e0237780
GY
177782013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
17779
3fdb53c1
TB
17780 * config/arm/arm.md (split for extendsidi): Update condition.
17781 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
17782 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
17783 (qhs_zextenddi_cstr): Likewise.
e0237780 17784
d7fde18c
JJ
177852013-02-21 Jakub Jelinek <jakub@redhat.com>
17786
be63b77d
JJ
17787 PR middle-end/56420
17788 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
17789 avoid signed wrapping.
17790 (expand_mult): Handle properly multiplication by
17791 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
17792 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
17793 in the compiler if coeff is HOST_WIDE_INT_MIN.
17794 (expand_divmod): Don't make ext_op1 static, change it's type to
17795 uhwi. Avoid undefined behavior in -INTVAL (op1).
17796
d7fde18c
JJ
17797 PR rtl-optimization/50339
17798 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
17799 field.
17800 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
17801 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
17802 into splitting_ashiftrt field.
17803 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
17804 ASHIFTRT.
17805 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
17806 choices.
17807
6aad4455
AH
178082013-02-20 Aldy Hernandez <aldyh@redhat.com>
17809
17810 PR middle-end/56108
17811 * trans-mem.c (execute_tm_mark): Do not expand transactions that
17812 are sure to go irrevocable.
17813
38fe784d
HPN
178142013-02-21 Hans-Peter Nilsson <hp@axis.com>
17815
17816 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
17817 scalars are valid operands.
17818
0fd44da3
MJ
178192013-02-21 Martin Jambor <mjambor@suse.cz>
17820
17821 PR tree-optimization/56310
17822 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
17823 only matching indices and non-negative final offsets.
17824 (intersect_aggregates_with_edge): Pass src_idx to
17825 agg_replacements_to_vector. Pass src_idx insstead of index to
17826 intersect_with_agg_replacements.
17827
7a92038b
MJ
178282013-02-21 Martin Jambor <mjambor@suse.cz>
17829
17830 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
17831 instead of hard-wired defaults.
17832
c0da9c37
MR
178332013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
17834
17835 * doc/invoke.texi (MIPS Options): Update documentation of the
17836 floating-point multiply-accumulate instruction restrictions.
17837
d247ea0c 178382013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS
17839
17840 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
17841 asan_shadow_offset on x86_64 linux.
17842
22deefcb
RB
178432013-02-21 Richard Biener <rguenther@suse.de>
17844
17845 PR tree-optimization/56415
17846 Revert
17847 2013-02-11 Richard Biener <rguenther@suse.de>
17848
17849 PR tree-optimization/56273
17850 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
17851 first VRP run.
17852
7bcc6e75
JJ
178532013-02-21 Jakub Jelinek <jakub@redhat.com>
17854
a0ad148f
JJ
17855 PR bootstrap/56258
17856 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
17857 instead of @itemx.
17858
7bcc6e75
JJ
17859 PR inline-asm/56405
17860 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
17861 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
17862
a0a7b611
JH
178632013-02-20 Jan Hubicka <jh@suse.cz>
17864
17865 PR tree-optimization/56265
88b97037
UB
17866 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
17867 when target is referenced for first time.
a0a7b611 17868
c0e50f72
RB
178692013-02-20 Richard Biener <rguenther@suse.de>
17870
17871 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
17872 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
17873 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
17874 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
17875 not return anything.
17876 (rename_ssa_copies): Do not remove unused locals.
17877 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 17878 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB
17879 * passes.c (execute_function_todo): Do not schedule unused locals
17880 removal if cleanup_tree_cfg did something.
17881 * tree-ssa-live.c (remove_unused_locals): Dump statistics
17882 about the number of removed locals.
17883
a52ca739
RB
178842013-02-20 Richard Biener <rguenther@suse.de>
17885
17886 PR tree-optimization/56398
88b97037 17887 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 17888
ec9c9d1b
MJ
178892013-02-20 Martin Jambor <mjambor@suse.cz>
17890
17891 PR tree-optimization/55334
17892 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
17893 restricted pointers to arrays.
17894
e91c8ed6 178952013-02-20 Richard Biener <rguenther@suse.de>
88b97037 17896 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB
17897
17898 PR tree-optimization/56396
17899 * tree-ssa-ccp.c (n_const_val): New static variable.
17900 (get_value): Return NULL for SSA names we don't have a lattice
17901 entry for.
17902 (ccp_initialize): Initialize n_const_val.
17903 * tree-ssa-copy.c (n_copy_of): New static variable.
17904 (init_copy_prop): Initialize n_copy_of.
17905 (get_value): Return NULL_TREE for SSA names we don't have a
17906 lattice entry for.
17907
3d916479
MJ
179082013-02-20 Martin Jambor <mjambor@suse.cz>
17909
17910 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
17911
71a86758
RB
179122013-02-20 Richard Biener <rguenther@suse.de>
17913
17914 * genpreds.c (write_lookup_constraint): Do not compare first
17915 letter of the constraint again.
17916
79836a12
RB
179172013-02-20 Richard Biener <rguenther@suse.de>
17918
17919 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
17920 and ceil_log2.
17921 (get_use_iv_cost): Terminate hashtable walk when coming across
17922 an empty entry.
17923
bbe4fb2c
IZ
179242013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
17925
17926 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
17927 reassociation for avx2 targets.
17928
a72d8780 179292012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 17930
c77f83d5
EI
17931 * config/microblaze/microblaze.c: microblaze_has_clz = 0
17932 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 17933 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI
17934 version and TARGET_PATTERN_COMPARE check
17935 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 17936
a72d8780 179372012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 17938
a72d8780 17939 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI
17940 function before branching.
17941
242387fa
AB
179422012-02-19 Andrey Belevantsev <abel@ispras.ru>
17943
17944 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
17945 DUMP_INSN_RTX_UID.
17946 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
17947
b4979ab9
AB
179482012-02-19 Andrey Belevantsev <abel@ispras.ru>
17949
17950 PR middle-end/55889
b4979ab9
AB
17951 * sel-sched.c: Include ira.h.
17952 (implicit_clobber_conflict_p): New function.
17953 (moveup_expr): Use it.
88b97037 17954 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 17955
24d63016
RB
179562013-02-19 Richard Biener <rguenther@suse.de>
17957
17958 PR tree-optimization/56384
17959 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
17960 (vn_hash_type): Split out from ...
17961 (vn_hash_constant_with_type): ... here.
17962 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
17963 (vn_phi_eq): Compare types from vn_phi_s structure.
17964 (vn_phi_lookup): Populate vn_phi_s type.
17965 (vn_phi_insert): Likewise.
17966
a475fd3d
JJ
179672013-02-19 Jakub Jelinek <jakub@redhat.com>
17968
47cc28f5
JJ
17969 PR tree-optimization/56350
17970 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
17971 if haven't found reduction or nested cycle operand, rather than
17972 asserting we must find it.
17973
a475fd3d
JJ
17974 PR tree-optimization/56381
17975 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
17976 to fold_build3.
17977
179782013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH
17979 Jakub Jelinek <jakub@redhat.com>
17980
17981 PR target/52555
17982 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
17983 (swap_optab_enable): Same.
17984 (init_all_optabs): Use argument instead of global.
88b97037 17985 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH
17986 * expr.h (init_all_optabs): Add argument to prototype.
17987 (TREE_OPTIMIZATION_OPTABS): New.
17988 (save_optabs_if_changed): Protoize.
17989 * optabs.h: Declare this_fn_optabs.
17990 * optabs.c (save_optabs_if_changed): New.
17991 Declare this_fn_optabs.
17992 (init_optabs): Add argument to init_all_optabs() call.
17993 * function.c (invoke_set_current_function_hook): Handle per
17994 function optabs.
17995 * function.h (struct function): New field optabs.
17996 * config/mips/mips.c (mips_set_mips16_mode): Handle when
17997 optimization_current_node has changed.
17998 * target-globals.h (save_target_globals_default_opts): Protoize.
17999 * target-globals.c (save_target_globals_default_opts): New.
18000
3f587ca3
JDA
180012013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18002
18003 PR target/56347
18004 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
18005 registers %fr12 and %fr12R as call used.
ceaca33e
JDA
18006
18007 PR target/56214
18008 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB
18009 and HImode, require all displacements to be an integer multiple of
18010 their mode size.
ceaca33e
JDA
18011 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
18012 only allow QImode and HImode when reload is in progress and strict is
18013 true. Likewise for symbolic addresses. Use base14_operand to check
18014 displacements in REG+BASE addresses.
18015
fe0b4796
RB
180162013-02-18 Richard Biener <rguenther@suse.de>
18017
18018 PR tree-optimization/56366
18019 * tree-vect-loop.c (get_initial_def_for_induction): Properly
18020 handle sign-conversion of outer-loop initial induction value.
18021
6aaf596b
RB
180222013-02-18 Richard Biener <rguenther@suse.de>
18023
73db8ff1 18024 PR middle-end/56349
6aaf596b
RB
18025 * cfghooks.c (merge_blocks): If we merge a latch into another
18026 block adjust references to it.
18027 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
18028 (verify_loop_structure): Verify that a recorded latch is in fact
18029 a latch.
18030
5e97dfb6
RB
180312013-02-18 Richard Biener <rguenther@suse.de>
18032
18033 PR tree-optimization/56321
18034 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
18035 order SSA name release and virtual operand unlinking.
18036
825527e8
EI
180372013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18038
18039 * config/microblaze/microblaze.md (save_stack_block): Define.
18040 (restore_stack_block): Likewise.
18041
debd11d9
EI
180422013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18043
18044 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
18045 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
18046 * config/microblaze/microblaze.c (microblaze_option_override):
18047 Bail out early for PIC modes when target does not support PIC.
18048
8ec77be0
EI
180492013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18050
88b97037 18051 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI
18052 Replace with a microblaze version.
18053 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 18054 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI
18055 microblaze.
18056
b41288b3
JJ
180572013-02-16 Jakub Jelinek <jakub@redhat.com>
18058 Dodji Seketeli <dodji@redhat.com>
18059
18060 PR asan/56330
88b97037 18061 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ
18062 (instrument_mem_region_access): Do not forget to always put
18063 instrumentation of the of 'base' and 'base + len' in a "if (len !=
18064 0) statement, even for cases where either 'base' or 'base + len'
18065 are not instrumented -- because they have been previously
18066 instrumented. Simplify the logic by putting all the statements
18067 instrument 'base + len' inside a sequence, and then insert that
18068 sequence right before the current insertion point. Then, to
18069 instrument 'base + len', just get an iterator on that statement.
18070 And do not forget to update the pointer to iterator the function
18071 received as argument.
18072
47918951
VM
180732013-02-15 Vladimir Makarov <vmakarov@redhat.com>
18074
18075 PR rtl-optimization/56348
18076 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
18077
3fb7c699
SB
180782013-02-15 Steven Bosscher <steven@gcc.gnu.org>
18079
18080 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
18081 (clean_graph_dump_file): Pass base to start_graph_dump.
18082
576fe41a
RH
180832013-02-14 Richard Henderson <rth@redhat.com>
18084
18085 PR target/55941
18086 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
18087
bb3accfa
SB
180882013-02-14 Steven Bosscher <steven@gcc.gnu.org>
18089
18090 * collect2-aix.h: Define F_LOADONLY.
18091
03143140
RB
180922013-02-14 Richard Biener <rguenther@suse.de>
18093
18094 PR lto/50494
18095 * varasm.c (output_constant_def_1): Get the decl representing
18096 the constant as argument.
18097 (output_constant_def): Wrap output_constant_def_1.
18098 (make_decl_rtl): Use output_constant_def_1 with the decl
18099 representing the constant.
18100 (build_constant_desc): Optionally re-use a decl already
18101 representing the constant.
18102 (tree_output_constant_def): Adjust.
18103
8fb06726
DS
181042013-02-14 Dodji Seketeli <dodji@redhat.com>
18105
18106 Fix an asan crash
18107 * asan.c (instrument_builtin_call): Really put the length of the
18108 second source argument into src1_len.
18109
c4bfe8bf
JJ
181102013-02-13 Jakub Jelinek <jakub@redhat.com>
18111
18112 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
18113 argument. If it is false, don't create edge from then_bb to
18114 fallthru_bb.
18115 (insert_if_then_before_iter): Pass true to it.
18116 (build_check_stmt): Pass false to it.
18117 (transform_statements): Flush hash table only on extended basic
18118 block boundaries, rather than at the beginning of every bb.
18119 Don't flush hash table on nonfreeing_call_p calls.
18120 * tree-flow.h (nonfreeing_call_p): New prototype.
18121 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
18122
7afe2801
DM
181232013-02-13 David S. Miller <davem@davemloft.net>
18124
18125 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
18126
70cc3288
VM
181272013-02-13 Vladimir Makarov <vmakarov@redhat.com>
18128
18129 PR target/56184
18130 * ira.c (max_regno_before_ira): Move from ...
18131 (ira): ... here.
18132 (fix_reg_equiv_init): Use max_regno_before_ira instead of
18133 vec_safe_length.
18134
6422242b
JJ
181352013-02-13 Jakub Jelinek <jakub@redhat.com>
18136
18137 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
18138
976a81ee
RB
181392013-02-13 Richard Biener <rguenther@suse.de>
18140
18141 PR lto/56295
18142 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
18143 globals in MEM_REFs.
18144
c1874a87
RB
181452013-02-13 Richard Biener <rguenther@suse.de>
18146
18147 * loop-init.c (loop_optimizer_init): Clear loop state when
18148 re-initializing preserved loops.
18149 * loop-unswitch.c (unswitch_single_loop): Return whether
18150 we unswitched the loop. Do not verify loop state here.
88b97037 18151 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 18152
b4ab7d34
KS
181532013-02-13 Kostya Serebryany <kcc@google.com>
18154
88b97037
UB
18155 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
18156 on x86_64 linux.
b4ab7d34
KS
18157 * sanitizer.def: Rename __asan_init to __asan_init_v1.
18158
bdcbe80c
DS
181592013-02-12 Dodji Seketeli <dodji@redhat.com>
18160
18161 Avoid instrumenting duplicated memory access in the same basic block
18162 * Makefile.in (asan.o): Add new dependency on hash-table.h
18163 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
18164 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
18165 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
18166 (free_mem_ref_resources, has_mem_ref_been_instrumented)
18167 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
18168 (get_mem_ref_of_assignment): New functions.
18169 (get_mem_refs_of_builtin_call): Extract from
18170 instrument_builtin_call and tweak a little bit to make it fit with
18171 the new signature.
18172 (instrument_builtin_call): Use the new
18173 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
18174 of is_gimple_builtin_call.
18175 (instrument_derefs, instrument_mem_region_access): Insert the
18176 instrumented memory reference into the hash table.
18177 (maybe_instrument_assignment): Renamed instrument_assignment into
18178 this, and change it to advance the iterator when instrumentation
18179 actually happened and return true in that case. This makes it
18180 homogeneous with maybe_instrument_assignment, and thus give a
18181 chance to callers to be more 'regular'.
18182 (transform_statements): Clear the memory reference hash table
18183 whenever we enter a new BB, when we cross a function call, or when
18184 we are done transforming statements. Use
18185 maybe_instrument_assignment instead of instrumentation. No more
18186 need to special case maybe_instrument_assignment and advance the
18187 iterator after calling it; it's now handled just like
18188 maybe_instrument_call. Update comment.
18189
4861a1f7
RB
181902013-02-13 Richard Biener <rguenther@suse.de>
18191
18192 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
18193 Fix loop discovery code.
18194
1c86bd80
VM
181952013-02-12 Vladimir Makarov <vmakarov@redhat.com>
18196
18197 PR inline-asm/56148
18198 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 18199 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM
18200 the operand is not reloaded. Prefer to reload conflicting operand
18201 if earlyclobber and matching operands are the same.
18202
7cbda518
RB
182032013-02-12 Richard Biener <rguenther@suse.de>
18204
18205 PR lto/56297
18206 * lto-streamer-out.c (write_symbol): Do not output symbols
18207 for hard register variables.
18208
e68a4ef6
GJL
182092013-02-12 Georg-Johann Lay <avr@gjlay.de>
18210
18211 PR target/54222
18212 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
18213 (umulsidi3_insn, mulsidi3_insn): New insns.
18214
a72d8780 182152013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL
18216
18217 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
18218 (struct tune_params): Add vec_costs field.
18219 * config/arm/arm.c (arm_builtin_vectorization_cost)
18220 (arm_add_stmt_cost): New functions.
18221 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
18222 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
18223 (arm_default_vec_cost): New struct of type cpu_vec_costs.
18224 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
18225 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
18226 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
18227 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
18228
43320568
RB
182292013-02-12 Richard Biener <rguenther@suse.de>
18230
18231 PR lto/56295
18232 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
18233 decls again if possible.
18234
a011aa39
RB
182352013-02-12 Richard Biener <rguenther@suse.de>
18236
18237 PR middle-end/56288
18238 * tree-ssa.c (verify_ssa_name): Fix check, move
18239 SSA_NAME_IN_FREE_LIST check up.
18240
6da26889
JJ
182412013-02-12 Jakub Jelinek <jakub@redhat.com>
18242 Steven Bosscher <steven@gcc.gnu.org>
18243
18244 PR rtl-optimization/56151
18245 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
18246 equal to op0 or op1, and last_insn pattern is CODE operation
18247 with MEM dest and one of the operands matches that MEM.
18248
f80e0faf
ST
182492013-02-11 Sriraman Tallam <tmsriramgoogle.com>
18250
18251 * doc/extend.texi: Document Function Multiversioning and "default"
18252 parameter string to target attribute.
18253 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
18254 target attribute parameter is "default".
18255 (ix86_compare_version_priority): Remove checks for target attribute.
18256 (ix86_mangle_function_version_assembler_name): Change error to sorry.
18257 Remove check for target attribute equal to NULL. Add assert.
18258 (ix86_generate_version_dispatcher_body): Change error to sorry.
18259
6c59ffd1
IS
182602013-02-11 Iain Sandoe <iain@codesourcery.com>
18261 Jack Howarth <howarth@bromo.med.uc.edu>
18262 Patrick Marlier <patrick.marlier@gmail.com>
18263
18264 PR libitm/55693
18265 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
18266 define ENDFILE_SPEC as TM_DESTRUCTOR.
18267 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
18268
a72d8780 182692013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP
18270 Jack Howarth <howarth@bromo.med.uc.edu>
18271 Jakub Jelinek <jakub@redhat.com>
18272
18273 PR sanitizer/55617
18274 * config/darwin.c (cdtor_record): Rename ctor_record.
18275 (sort_cdtor_records): Rename sort_ctor_records.
18276 (finalize_dtors): New routine to sort destructors by
18277 priority before use in assemble_integer.
18278 (machopic_asm_out_destructor): Use finalize_dtors if needed.
18279
b63fe007
UB
182802013-02-11 Uros Bizjak <ubizjak@gmail.com>
18281
18282 PR rtl-optimization/56275
18283 * simplify-rtx.c (avoid_constant_pool_reference): Check that
18284 offset is non-negative and less than cmode size before
18285 calling simplify_subreg.
18286
8e89b5b5
RB
182872013-02-11 Richard Biener <rguenther@suse.de>
18288
18289 PR tree-optimization/56264
18290 * cfgloop.h (fix_loop_structure): Adjust prototype.
18291 * loop-init.c (fix_loop_structure): Return the number of
18292 newly discovered loops.
18293 * tree-cfgcleanup.c (repair_loop_structures): When new loops
18294 are discovered, do a full loop-closed SSA rewrite.
18295
b4a4b56d
RB
182962013-02-11 Richard Biener <rguenther@suse.de>
18297
18298 PR tree-optimization/56273
18299 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
18300 first VRP run.
18301 (check_array_ref): Fix missing newline in dumps.
18302 (search_for_addr_array): Likewise.
18303
0c885229
DE
183042013-02-09 David Edelsohn <dje.gcc@gmail.com>
18305
18306 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
18307
59ac9a55
JJ
183082013-02-09 Jakub Jelinek <jakub@redhat.com>
18309
18310 PR target/56256
18311 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
18312
25bb0bb5
VM
183132013-02-08 Vladimir Makarov <vmakarov@redhat.com>
18314
18315 PR rtl-optimization/56246
0c885229 18316 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM
18317 reload pseudo.
18318 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
18319 constraints are satisfied.
18320
a698cc03
JL
183212013-02-08 Jeff Law <law@redhat.com>
18322
18323 PR debug/53948
18324 * emit-rtl.c (reg_is_parm_p): New function.
18325 * regs.h (reg_is_parm_p): New prototype.
18326 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
18327 callee-clobbered registers.
18328
e1122ddd
MM
183292013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
18330
18331 PR target/56043
18332 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
18333 If there is no implicit builtin declaration, just return NULL.
18334
19c5f6e6
UB
183352013-02-08 Uros Bizjak <ubizjak@gmail.com>
18336
18337 * config/i386/sse.md (FMAMODEM): New mode iterator.
18338 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
18339 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
18340
2480f2ca 183412013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 18342
2480f2ca
UB
18343 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
18344 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
18345 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
18346
183472013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18348
18349 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
18350 (microblaze*-*-elf): Likewise.
18351 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
18352 LINK_SPEC.
18353 * config/microblaze/microblaze-c.c: Add builtin defines for
18354 _LITTLE_ENDIAN and _BIG_ENDIAN.
18355 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
18356 add to TARGET_DEFAULT flags.
76ef61fb 18357 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB
18358 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
18359 * config/microblaze/microblaze.md: Update extendsidi2 and
18360 movdi_internal instructions to use low-order / high-order reg
18361 print_operands.
18362 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
18363 options and inversemask / mask of LITTLE_ENDIAN.
18364 * config/microblaze/t-microblaze: Expand multilib options to
18365 include mlittle-endian (le) and update exceptions patterns.
18366
600a5961
JJ
183672013-02-08 Jakub Jelinek <jakub@redhat.com>
18368
5df81313
JJ
18369 PR rtl-optimization/56195
18370 * lra-constraints.c (get_reload_reg): Don't reuse regs
18371 if they have smaller mode than requested, if they have
18372 wider mode than requested, try to return a SUBREG.
18373
600a5961
JJ
18374 PR tree-optimization/56250
18375 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
18376 if type is unsigned and code isn't MULT_EXPR.
18377
ff544649
GJL
183782013-02-08 Georg-Johann Lay <avr@gjlay.de>
18379
18380 PR tree-optimization/56064
18381 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
18382 bits according to mode.
18383 * fixed-value.h (fixed_from_double_int)
18384 (const_fixed_from_double_int): Adjust comments.
18385
e45cde98
RB
183862013-02-08 Richard Biener <rguenther@suse.de>
18387
18388 PR lto/56231
18389 * lto-streamer.h (struct data_in): Remove current_file, current_line
18390 and current_col members.
18391 * lto-streamer-out.c (lto_output_location): Stream changed bits
18392 en-block for efficiency.
18393 * lto-streamer-in.c (clear_line_info): Remove.
18394 (lto_input_location): Cache current file, line and column
18395 globally via local statics. Read changed bits en-block.
18396 (input_function): Do not call clear_line_info.
18397 (lto_read_body): Likewise.
18398 (lto_input_toplevel_asms): Likewise.
18399
c1ca73d8
MM
184002013-02-08 Michael Matz <matz@suse.de>
18401
18402 PR tree-optimization/52448
18403 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
18404 (nt_call_phase): New static.
18405 (add_or_mark_expr): Only mark accesses with newer phase than any
18406 call seen.
18407 (nonfreeing_call_p): New.
18408 (nt_init_block): Update nt_call_phase, mark blocks as visited.
18409 (nt_fini_block): Keep blocks marked as visited.
18410 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
18411
57548aa2
RB
184122013-02-08 Richard Biener <rguenther@suse.de>
18413
18414 * ira.c (ira): Free broken dominator information.
18415
8e10366f
UB
184162013-02-08 Uros Bizjak <ubizjak@gmail.com>
18417
18418 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
18419
f64fb0fa
MP
184202013-02-08 Marek Polacek <polacek@redhat.com>
18421
8e10366f 18422 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 18423
0375167b
RB
184242013-02-08 Richard Biener <rguenther@suse.de>
18425
18426 PR middle-end/56181
18427 * cfgloop.h (flow_loops_find): Adjust.
18428 (bb_loop_header_p): Declare.
18429 * cfgloop.c (bb_loop_header_p): New function split out from ...
18430 (flow_loops_find): ... here. Adjust function signature,
18431 support incremental loop structure update.
18432 (verify_loop_structure): Cleanup. Verify a loop is a loop.
18433 * cfgloopmanip.c (fix_loop_structure): Move ...
18434 * loop-init.c (fix_loop_structure): ... here.
18435 (apply_loop_flags): Split out from ...
18436 (loop_optimizer_init): ... here.
18437 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
18438 in incremental mode, only remove dead loops here.
18439
85d768f3
GJL
184402013-02-08 Georg-Johann Lay <avr@gjlay.de>
18441
18442 PR target/54222
18443 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
18444 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
18445 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
18446 (*round<mode>3.libgcc): New insns for fixed-modes.
18447 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
18448 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
18449 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
18450 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
18451 implementations. Define to __builtin_avr_absFX,
18452 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
18453 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
18454 __builtin_avr_countlsFX, respectively.
18455 * config/avr/avr-c.c (target.h): Include it.
18456 (enum avr_builtin_id): New enum.
18457 (avr_resolve_overloaded_builtin): New static function.
18458 (avr_register_target_pragmas): Use it to set
18459 targetm.resolve_overloaded_builtin.
18460 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
18461 tree nodes used by DEF_BUILTIN.
18462 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
18463 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
18464 <AVR_BUILTIN_xxBITS>: Same.
18465
661bc682
RB
184662013-02-08 Richard Biener <rguenther@suse.de>
18467
18468 * cfgloop.c (verify_loop_structure): Properly handle
18469 a loop exiting to another loop header.
18470 * ira-int.h (ira_loops): Remove.
18471 * ira.c (ira_loops): Remove.
18472 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
18473 (do_reload): Use loop_optimizer_finalize.
18474 * ira-build.c (create_loop_tree_nodes): Use get_loops and
18475 number_of_loops to access the loop tree.
18476 (more_one_region_p): Likewise.
18477 (finish_loop_tree_nodes): Likewise.
18478 (rebuild_regno_allocno_maps): Likewise.
18479 (mark_loops_for_removal): Likewise.
18480 (mark_all_loops_for_removal): Likewise.
18481 (remove_unnecessary_regions): Likewise.
18482 (ira_build): Likewise.
18483 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18484
0d5049b2
RB
184852013-02-08 Richard Biener <rguenther@suse.de>
18486
18487 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
18488 * ipa-pure-const.c (analyze_function): Avoid calling
18489 mark_irreducible_loops twice.
8e10366f 18490 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 18491
499f32e8
DM
184922013-02-07 David S. Miller <davem@davemloft.net>
18493
18494 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
18495 on 'reg'.
18496 * var-tracking.c (vt_add_function_parameter): Test the presence of
18497 HAVE_window_save properly and do not remap argument registers when
18498 we have a leaf function.
18499
6edc3e32
UB
185002013-02-07 Uros Bizjak <ubizjak@gmail.com>
18501
18502 PR bootstrap/56227
18503 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
18504 instead of "ll".
18505 * config/i386/i386.c (ix86_print_operand): Ditto.
18506
5306401f
VM
185072013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18508
6edc3e32 18509 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 18510
027ece11
VM
185112013-02-07 Vladimir Makarov <vmakarov@redhat.com>
18512
18513 PR rtl-optimization/56225
18514 * lra-constraints.c (process_alt_operands): Check that reload hard
18515 reg can hold value for strict_low_part.
18516
f980dfdb
JJ
185172013-02-07 Jakub Jelinek <jakub@redhat.com>
18518
18519 PR debug/56154
18520 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
18521 dwarf2out_end_function.
18522 (in_first_function_p, maybe_at_text_label_p,
18523 first_loclabel_num_not_at_text_label): New variables.
18524 (dwarf2out_var_location): In the first function find out
6edc3e32 18525 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ
18526 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
18527 functions.
18528
73dd3123
EB
185292013-02-07 Eric Botcazou <ebotcazou@adacore.com>
18530
18531 PR rtl-optimization/56178
18532 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
18533 SUBREG of a register. Tidy up related block of code.
18534 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
18535 note if the source is a register or a SUBREG of a register.
18536
e3936f47
JJ
185372013-02-07 Jakub Jelinek <jakub@redhat.com>
18538
18539 PR target/56228
18540 * config/rs6000/rs6000.md (ptrm): New mode attr.
18541 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
18542 call_value_indirect_aix<pttrsize>,
18543 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
18544 m in constraints.
18545
d96d674b
MH
185462013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18547
18548 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
18549 if -bnortl. Convert to strcmp and strncmp.
18550
921f2dee
AM
185512013-02-07 Alan Modra <amodra@gmail.com>
18552
18553 PR target/54009
18554 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
18555 addresses won't wrap when offsetting.
18556 (rs6000_secondary_reload): Provide secondary reloads needed for
18557 wrapping LO_SUM addresses.
18558
d09c7dba
TS
185592013-02-06 Thomas Schwinge <thomas@codesourcery.com>
18560
18561 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
18562 MACH, just __MACH__.
18563
a44bbd48
RB
185642013-02-06 Richard Biener <rguenther@suse.de>
18565
18566 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
18567 instead of calling fix_loop_structure.
18568
6d840d99
JJ
185692013-02-06 Jakub Jelinek <jakub@redhat.com>
18570
18571 PR middle-end/56217
18572 * omp-low.c (use_pointer_for_field): Return false if
18573 lower_send_shared_vars doesn't generate any copy-out code.
18574
0f33baa9
TV
185752013-02-06 Tom de Vries <tom@codesourcery.com>
18576
18577 PR rtl-optimization/56131
18578 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
18579 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
18580 of the label is NULL. Add comment.
18581
11f1e3ab
JJ
185822013-02-05 Jakub Jelinek <jakub@redhat.com>
18583
31502f9f
JJ
18584 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
18585
a74db9bd
JJ
18586 PR sanitizer/55374
18587 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
18588 (STATIC_LIBTSAN_LIBS): Likewise.
18589 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
18590 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
18591 is defined, don't add anything else beyond that.
18592 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
18593 (LINK_COMMAND_SPEC): Use them.
18594
11f1e3ab
JJ
18595 PR tree-optimization/56205
18596 * tree-stdarg.c (check_all_va_list_escapes): Return true if
18597 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
18598 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
18599
14ac6aa2
RB
186002013-02-05 Richard Biener <rguenther@suse.de>
18601
18602 PR tree-optimization/53342
18603 PR tree-optimization/53185
18604 * tree-vectorizer.h (vect_check_strided_load): Remove.
18605 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
18606 not disallow peeling for vectorized strided loads.
18607 (vect_check_strided_load): Make static and simplify.
18608 (vect_analyze_data_refs): Adjust.
18609 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
18610 correctly when vectorizing strided loads.
18611
6f22445a
RB
186122013-02-05 Richard Biener <rguenther@suse.de>
18613
18614 * doc/install.texi: Refer to ISL, not PPL.
18615
39f9719e
JH
186162013-02-05 Jan Hubicka <jh@suse.cz>
18617
ec4224ac
JH
18618 PR tree-optimization/55789
18619 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
18620
186212013-02-05 Jan Hubicka <jh@suse.cz>
18622
18623 PR tree-optimization/55789
39f9719e
JH
18624 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
18625 the dead call anyway.
18626
956d3b33
EB
186272013-02-05 Eric Botcazou <ebotcazou@adacore.com>
18628
18629 PR sanitizer/55374
18630 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
18631
186322013-02-04 Alexander Potapenko <glider@google.com>
18633 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP
18634 Jakub Jelinek <jakub@redhat.com>
18635
18636 PR sanitizer/55617
18637 * config/darwin.c (sort_ctor_records): Stabilized qsort
18638 on constructor priority by using original position.
18639 (finalize_ctors): New routine to sort constructors by
18640 priority before use in assemble_integer.
18641 (machopic_asm_out_constructor): Use finalize_ctors if needed.
18642
7ac3af38
JJ
186432013-02-04 Jakub Jelinek <jakub@redhat.com>
18644
18645 PR libstdc++/54314
18646 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
18647 about visibility on artificial decls.
18648 * config/sol2.c (solaris_assemble_visibility): Likewise.
18649
152689dc
KT
186502013-02-04 Kai Tietz <ktietz@redhat.com>
18651
18652 PR target/56186
18653 * config/i386/i386.c (function_value_ms_64): Add additional valtype
18654 argument and improve checking of return-argument types for 16-byte
18655 modes.
18656 (ix86_function_value_1): Add additional valtype argument on call
18657 of function_value_64.
18658 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
18659 handling infunction_value_64 function.
18660
2b5987b5
MGD
186612013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
18662
7ac3af38 18663 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 18664
14597080
RB
186652013-02-04 Richard Biener <rguenther@suse.de>
18666
18667 PR tree-optimization/56188
18668 * tree-ssa-structalias.c (label_visit): Consider case with
18669 initially non-empty points-to set.
18670 (perform_var_substitution): Dump node mapping and clean up.
18671
ed73881e
RG
186722013-02-04 Richard Guenther <rguenther@suse.de>
18673
18674 PR lto/56168
18675 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
18676 node prevail as last resort.
18677 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 18678 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 18679
139a0707
RB
186802013-02-04 Richard Biener <rguenther@suse.de>
18681
18682 PR tree-optimization/56113
18683 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
18684 Merge into ...
18685 (equiv_class_lookup_or_add): ... this.
18686 (label_visit): Adjust and fix error in previous patch.
18687 (perform_var_substitution): Adjust.
18688
50fe8924
OE
186892013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
18690
18691 * config/sh/divtab.c: Fix formatting and comments throughout the file.
18692 * config/sh/sh4-300.md: Likewise.
18693 * config/sh/sh4a.md: Likewise.
18694 * config/sh/constraints.md: Likewise.
18695 * config/sh/sh.md: Likewise.
18696 * config/sh/netbsd-elf.h: Likewise.
18697 * config/sh/predicates.md: Likewise.
18698 * config/sh/sh-protos.h: Likewise.
18699 * config/sh/ushmedia.h: Likewise.
18700 * config/sh/linux.h: Likewise.
18701 * config/sh/sh.c: Likewise.
18702 * config/sh/superh.h: Likewise.
18703 * config/sh/elf.h: Likewise.
18704 * config/sh/sh4.md: Likewise.
18705 * config/sh/sh.h: Likewise.
18706
1a04ac2b
JDA
187072013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18708
18709 * config/pa/constraints.md: Adjust unused letters. Change "T"
18710 constraint to match_test floating_point_store_memory_operand().
18711 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
18712 (base14_operand): New.
18713 (floating_point_store_memory_operand): New.
18714 (integer_store_memory_operand): Revise to use base14_operand and
18715 reg_plus_base_memory_operand.
18716 (move_dest_operand): Allow symbolic_memory_operands.
18717 (symbolic_memory_operand): Check for LO_SOM.
18718 (symbolic_operand): Change default case to break.
18719 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
18720 CONST_DOUBLE values to be reloaded by putting them into memory when
18721 the destination is a floating point register.
18722 (movdf): Remove code to handle CONST_DOUBLE.
18723 (movsf): Likewise.
18724 (reload_indf_r1): New.
18725 (reload_insf_r1): New.
18726 Consistently use "Q" and "T" constraints with integer and floating
18727 point move instructions, respectively.
18728 (movdi): Remove FAIL.
18729 Change predicate for source operand unamed DImode move from
18730 general_operand to move_src_operand.
18731 (umulsidi3): Change predicate for destination operand to
18732 register_operand.
18733 Likewise for similar unamed patterns.
18734 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
18735 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
18736 (hppa_legitimize_address): Simplify mask calculation.
18737 (pa_emit_move_sequence): Revised handling of secondary reloads from
18738 REG+D addresses for floating point loads and stores. Directly handle
18739 loading CONST0_RTX (mode) to a floating point register.
18740 (pa_secondary_reload): Handle reloading DF and SFmode constant values
18741 to floating point registers. Don't restrict secondary reloads to
18742 floating point registers to integer modes. Revise some comments and
18743 cleanup some code.
18744 (TARGET_LEGITIMATE_ADDRESS_P): Define.
18745 (pa_legitimate_address_p): New.
18746 (pa_legitimize_reload_address): New.
18747 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
18748 (STRICT_REG_OK_FOR_BASE_P): New.
18749 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
18750 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
18751
3bf65591
DE
187522013-02-03 David Edelsohn <dje.gcc@gmail.com>
18753 Andrew Dixie <andrewd@gentrack.com>
18754
18755 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
18756 flag set.
18757
c0a8a3e6
RS
187582013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
18759
18760 * expmed.c (extract_bit_field_1): Pass the full width of the
18761 structure to get_best_reg_extraction_insn.
18762
99113dff
DE
187632013-02-01 David Edelsohn <dje.gcc@gmail.com>
18764
18765 PR target/54601
18766 * configure.ac (use_cxa_atexit): Add AIX.
18767 * configure: Regenerate.
18768
18769 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
18770
ddd84654
JJ
187712013-02-01 Jakub Jelinek <jakub@redhat.com>
18772
18773 PR debug/54793
18774 * final.c (need_profile_function): New variable.
18775 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
18776 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
18777 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
18778 notes, targetm.asm_out.function_prologue doesn't emit anything,
18779 HAVE_prologue and profiler should be emitted before prologue,
18780 set need_profile_function instead of emitting it.
18781 (final_scan_insn): If need_profile_function, emit
18782 profile_function on the first NOTE_INSN_BASIC_BLOCK or
18783 NOTE_INSN_FUNCTION_BEG note.
18784
187852013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH
18786
18787 * config/rs6000/rs6000.md (smulditi3): New.
18788 (umulditi3): New.
18789
ff2a9d88
RH
18790 * config/alpha/alpha.md (umulditi3): New.
18791
14d52b90
DE
187922013-02-01 David Edelsohn <dje.gcc@gmail.com>
18793
18794 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
18795 (ASM_OUTPUT_ALIGNED_LOCAL): New.
18796
8c7ca45c
RB
187972013-02-01 Richard Biener <rguenther@suse.de>
18798
18799 PR tree-optimization/56113
18800 * tree-ssa-structalias.c (label_visit): Reduce work for
18801 single-predecessor nodes.
18802
9f419393
EB
188032013-02-01 Eric Botcazou <ebotcazou@adacore.com>
18804
18805 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
18806 range isn't testing for zero.
18807
c5f4be84
SB
188082013-01-31 Steven Bosscher <steven@gcc.gnu.org>
18809
18810 PR middle-end/56113
18811 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
18812
dbdbd982
NC
188132013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
18814 Nick Clifton <nickc@redhat.com>
18815
18816 * config/v850/constraints.md (Q): Define as a memory constraint.
18817 * config/v850/predicates.md (label_ref_operand): New predicate.
18818 (e3v5_shift_operand): New predicate.
18819 (ior_operator): New predicate.
18820 * config/v850/t-v850: Add e3v5 multilib.
18821 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
18822 (v850_gen_movdi): Prototype.
18823 * config/v850/v850.c: Add support for e3v5 architecture.
18824 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
18825 TARGET_V850E_UP.
18826 (construct_save_jarl): Add e3v5 long JARL support.
18827 (v850_adjust_insn_length): New function. Adjust length of call
18828 insns when using e3v5 instructions.
18829 (v850_gen_movdi): New function: Generate instructions to move a
18830 DImode value.
18831 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
18832 (CPP_SPEC): Define __v850e3v5__ as appropriate.
18833 (TARGET_USE_FPU): Enable for e3v5.
18834 (CONST_OK_FOR_W): New macro.
18835 (ADJUST_INSN_LENGTH): Define.
18836 * config/v850/v850.md (UNSPEC_LOOP): Define.
18837 (attr cpu): Add v850e3v5.
18838 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
18839 (movdi): New pattern.
18840 (movdi_internal): New pattern.
18841 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
18842 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
18843 (cstoresf4): Likewise.
18844 (cstoredf4): Likewise.
18845 (insv): New pattern.
18846 (rotlso3_a): New pattern.
18847 (rotlsi3_b): New pattern
18848 (rotlsi3_v850e3v5): New pattern.
18849 (doloop_begin): New pattern.
18850 (fix_loop_counter): New pattern.
18851 (doloop_end): New pattern.
18852 (branch_normal): Add e3v5 long branch support.
18853 (branch_invert): Likewise.
18854 (branch_z_normal): Likewise.
18855 (branch_z_invert): Likewise.
18856 (branch_nz_normal): Likewise.
18857 (branch_nz_invert): Likewise.
18858 (call_internal_short): Add e3v5 register-indirect JARL support.
18859 (call_internal_long): Likewise.
18860 (call_value_internal_short): Likewise.
18861 (call_value_internal_long): Likewise.
18862 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
18863 (mloop): New option.
18864 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 18865 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 18866
73861a41
PK
188672013-01-31 Paul Koning <ni1d@arrl.net>
18868
18869 PR debug/55059
18870 PR debug/54508
18871 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
18872 children if parent is a class.
18873 (prune_unused_types_prune): Don't add DW_AT_declaration.
18874
e44978dc
RB
188752013-01-31 Richard Biener <rguenther@suse.de>
18876
18877 PR tree-optimization/56157
18878 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
18879 match up operand with SLP child.
18880
90b10dec
JM
188812013-01-31 Jason Merrill <jason@redhat.com>
18882
04d2dadd 18883 PR debug/54410
90b10dec
JM
18884 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
18885 parameters the first time.
18886 (gen_scheduled_generic_parms_dies): Check completeness here.
18887
6e616110
RB
188882013-01-31 Richard Biener <rguenther@suse.de>
18889
18890 PR middle-end/53073
18891 * common.opt (faggressive-loop-optimizations): New flag,
18892 enabled by default.
18893 * doc/invoke.texi (faggressive-loop-optimizations): Document.
18894 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
18895 infer_loop_bounds_from_undefined by it.
18896
636f59cf
RB
188972013-01-31 Richard Biener <rguenther@suse.de>
18898
18899 PR tree-optimization/56150
18900 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
18901 visit virtual operands.
18902 (find_uses_to_rename_bb): Likewise.
18903
189042013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB
18905
18906 PR tree-optimization/56150
18907 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
18908 mixed store non-store stmts.
18909
32887460
JJ
189102013-01-30 Jakub Jelinek <jakub@redhat.com>
18911
e60e09a0
JJ
18912 PR sanitizer/55374
18913 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
18914 LIBASAN_EARLY_SPEC is defined.
18915 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
18916 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
18917 before %o.
18918 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
18919
32887460
JJ
18920 PR c++/55742
18921 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
18922 invalid args instead of ICEing on it.
18923 (ix86_valid_target_attribute_tree): Return error_mark_node if
18924 ix86_valid_target_attribute_inner_p failed.
18925 (ix86_valid_target_attribute_p): Return false only if
18926 ix86_valid_target_attribute_tree returned error_mark_node. Allow
18927 target("default") attribute.
18928 (sorted_attr_string): Change argument from const char * to tree,
18929 merge in all target attribute arguments rather than just one.
18930 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
18931 instead of free. Avoid using strcat.
18932 (ix86_mangle_function_version_assembler_name): Mangle
18933 target("default") as if no target attribute is present. Adjust
18934 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
18935 instead of xmalloc and XDELETEVEC instead of free.
18936 (ix86_function_versions): Don't return true if one of the decls
18937 doesn't have target attribute. If they don't and one of the decls
18938 is DECL_FUNCTION_VERSIONED, report an error. Adjust
18939 sorted_attr_string caller. Use XDELETEVEC instead of free.
18940 (ix86_supports_function_versions): Remove.
18941 (make_name): Fix up formatting.
18942 (make_dispatcher_decl): Remove resolver_name and its initialization.
18943 Avoid leaking memory.
18944 (is_function_default_version): Return true if there is
18945 target("default") attribute rather than no target attribute at all.
18946 (make_resolver_func): Avoid leaking memory.
18947 (ix86_generate_version_dispatcher_body): Likewise.
18948 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
18949 * target.def (supports_function_versions): Remove.
18950 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
18951 * doc/tm.texi: Regenerated.
18952
73cca0cc
VM
189532013-01-30 Vladimir Makarov <vmakarov@redhat.com>
18954
18955 PR rtl-optimization/56144
18956 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
18957 for values with side effects.
18958
11452e7b
RB
189592013-01-30 Richard Biener <rguenther@suse.de>
18960
18961 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
18962 (sparseset_pop): Likewise.
18963 * cfganal.c (compute_idf): Likewise. Increase work-stack size
18964 to be able to use quick_push in the worker loop.
18965
01cb1ef5
MP
189662013-01-30 Marek Polacek <polacek@redhat.com>
18967
18968 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
18969
e892936e
RB
189702013-01-30 Richard Biener <rguenther@suse.de>
18971
18972 PR lto/56147
6edc3e32 18973 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 18974
cc06c01d
GJL
189752013-01-30 Georg-Johann Lay <avr@gjlay.de>
18976
18977 PR tree-optimization/56064
18978 * fixed-value.c (fixed_from_double_int): New function.
18979 * fixed-value.h (fixed_from_double_int): New prototype.
18980 (const_fixed_from_double_int): New static inline function.
18981 * fold-const.c (native_interpret_fixed): New static function.
18982 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
18983 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
18984 (native_encode_fixed): New static function.
18985 (native_encode_expr) <FIXED_CST>: Use it.
18986 (native_interpret_int): Move double_int worker code to...
18987 * double-int.c (double_int::from_buffer): ...this new static method.
18988 * double-int.h (double_int::from_buffer): Prototype it.
18989
d394a308
RB
189902013-01-30 Richard Biener <rguenther@suse.de>
18991
18992 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
18993 New pointer-map and obstack.
18994 (init_alias_vars): Allocate pointer-map and obstack.
18995 (delete_points_to_sets): Free them.
18996 (find_what_var_points_to): Cache result.
18997 (find_what_p_points_to): Adjust for changed interface of
18998 find_what_var_points_to.
18999 (compute_points_to_sets): Likewise.
19000 (ipa_pta_execute): Likewise.
19001
20804d96
RO
190022013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19003
19004 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
19005 * configure: Regenerate.
19006 * config.in: Regenerate.
19007 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
19008 #nobits/#progbits if supported.
19009
a7ad88a2
OE
190102013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
19011
19012 PR target/56121
19013 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
19014 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
19015 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
19016
7af79f92
GY
190172013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19018
91bfca59
OE
19019 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
19020 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 19021
753bcf7b
GY
190222013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19023
91bfca59
OE
19024 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
19025 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 19026
8cbc2ea8
GY
190272013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19028
91bfca59
OE
19029 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
19030 declaration.
19031 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
19032 * config/arm/cortex-a7.md: New bypasses using
19033 arm_mac_accumulator_is_result.
8cbc2ea8 19034
697a3325
GY
190352013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19036
91bfca59 19037 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 19038 (cortex_a7_neon_mla): Likewise.
91bfca59
OE
19039 (cortex_a7_fpfmad): New reservation.
19040 (cortex_a7_fpmacs): Use ffmas and update required units.
19041 (cortex_a7_fpmuld): Update required units and latency.
19042 (cortex_a7_fpmacd): Likewise.
19043 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
19044 (cortex_a7_neon). Likewise.
19045 (bypass) Update participating units.
697a3325 19046
29637783
GY
190472013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
19048
91bfca59
OE
19049 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
19050 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
19051 from fmac to ffma.
19052 * config/arm/vfp11.md (vfp_farith): Use ffmas.
19053 (vfp_fmul): Use ffmad.
19054 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
19055 (cortex_r4_fmacd): Use ffmad.
19056 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
19057 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
19058 (cortex_a9_fmacd): Use ffmad.
19059 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
19060 (cortex_a8_vfp_macd): Use ffmad.
19061 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
19062 (cortex_a5_fpmacd): Use ffmad.
19063 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
19064 (cortex_a15_vfp_macd): Use ffmad.
19065 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 19066
2ee8a2d5
JM
190672013-01-29 Jason Merrill <jason@redhat.com>
19068
19069 PR libstdc++/54314
19070 * varasm.c (default_assemble_visibility): Don't warn about
19071 visibility on artificial decls.
19072
78d087bc
RB
190732013-01-29 Richard Biener <rguenther@suse.de>
19074
19075 PR tree-optimization/56113
19076 * tree-ssa-structalias.c (equiv_class_lookup): Also return
19077 the bitmap leader.
19078 (label_visit): Free duplicate bitmaps and record the leader instead.
19079 (perform_var_substitution): Adjust.
19080
83ba4d6f
RB
190812013-01-29 Richard Biener <rguenther@suse.de>
19082
19083 PR tree-optimization/55270
19084 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
19085 the CFG, schedule loops for fixup.
19086
66dfe13f
NC
190872013-01-29 Nick Clifton <nickc@redhat.com>
19088
19089 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
19090 SP_REG.
19091
5a579c3b
LE
190922013-01-28 Leif Ekblad <leif@rdos.net>
19093
19094 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
19095 * config/i386/i386.h (TARGET_RDOS): New macro.
19096 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
19097 * config/i386/i386.c (ix86_option_override_internal): For 64bit
19098 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
19099 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
19100 DEFAULT_LARGE_SECTION_THRESHOLD.
19101 * config/i386/i386.md (R14_REG, R15_REG): New constants.
19102 * config/i386/rdos.h: New file.
19103 * config/i386/rdos64.h: New file.
19104
51e44392
BS
191052013-01-28 Bernd Schmidt <bernds@codesourcery.com>
19106
19107 PR other/54814
19108 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
19109 TEST_HARD_REG_BIT.
19110
db1fb332
JJ
191112013-01-28 Jakub Jelinek <jakub@redhat.com>
19112
19113 PR rtl-optimization/56117
19114 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
19115 call cselib_lookup_from_insn on the MEM before calling
19116 add_insn_mem_dependence.
19117
16917761
RB
191182013-01-28 Richard Biener <rguenther@suse.de>
19119
19120 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
19121 to a stmt that didn't have one.
19122 (copy_phis_for_bb): Likewise for PHI arguments.
19123 (copy_debug_stmt): Likewise for debug stmts.
19124
b9fc0497
RB
191252013-01-28 Richard Biener <rguenther@suse.de>
19126
19127 PR tree-optimization/56034
6edc3e32 19128 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB
19129 (partition_builtin_p): Adjust.
19130 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
19131 it is the last partition.
19132 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
19133 up the vertex for the definition.
19134 (classify_partition): Classify whether a partition is a
19135 PKIND_REDUCTION, thus has uses outside of the loop.
19136 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
19137 Merge all PKIND_REDUCTION partitions into the last partition.
19138 (tree_loop_distribution): Seed partitions from reductions as well.
19139
aa710d25
JJ
191402013-01-28 Jakub Jelinek <jakub@redhat.com>
19141
0bfbca58
JJ
19142 PR tree-optimization/56125
19143 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
19144 pow(x,c) into sqrt(x) * powi(x, n/2) or
19145 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
19146 optimizing for size.
19147 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
19148 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
19149 integer.
19150
aa710d25
JJ
19151 PR tree-optimization/56094
19152 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
19153 to UNKNOWN_LOCATION while gimplifying expr.
19154
77dc5297
UB
191552013-01-27 Uros Bizjak <ubizjak@gmail.com>
19156
19157 PR target/56114
19158 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
19159 operand 0 in movabs insn template for -masm=intel asm alternative.
19160 (*movabs<mode>_2): Ditto for operand 1.
19161
0bfbca58 191622013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH
19163
19164 PR target/54663
19165 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
19166 of microblaze-c.o
19167
0bfbca58 191682013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI
19169
19170 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
19171 tm_file.
19172
0bfbca58 191732013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N
19174
19175 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
19176 Undef to avoid warning.
19177
478f60f9
MH
191782013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19179
19180 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
19181 * configure: Regenerate.
19182
d7fa6ee2
JJ
191832013-01-25 Jakub Jelinek <jakub@redhat.com>
19184
19185 PR tree-optimization/56098
19186 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
19187 for stmts with volatile ops.
19188 (cond_store_replacement): Don't optimize if assign has volatile ops.
19189 (cond_if_else_store_replacement_1): Don't optimize if either
19190 then_assign or else_assign have volatile ops.
19191 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
19192 volatile ops.
19193
f8fe87bd
GJL
191942013-01-25 Georg-Johann Lay <avr@gjlay.de>
19195
19196 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
19197
28398d0d
GJL
191982013-01-25 Georg-Johann Lay <avr@gjlay.de>
19199
19200 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
19201 missing ':' in asm example.
19202
b7d7d917
TB
192032013-01-25 Tejas Belagod <tejas.belagod@arm.com>
19204
19205 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
19206 entries into lane and laneq entries.
77dc5297
UB
19207 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
19208 Remove AdvSIMD scalar modes.
b7d7d917
TB
19209 (aarch64_sq<r>dmulh_laneq<mode>): New.
19210 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
19211 modes.
19212 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
19213 builtin implementations to relfect changes in RTL in aarch64-simd.md.
19214 * config/aarch64/iterators.md (VCOND): New.
19215 (VCONQ): New.
19216
556f9906
GJL
192172013-01-25 Georg-Johann Lay <avr@gjlay.de>
19218
19219 PR target/54222
19220 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
19221 Add NULL LIBNAME argument to existing definitions.
19222 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
19223 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
19224 * config/avr/avr.c (DEF_BUILTIN): Same.
19225 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
19226 (avr_expand_builtin): Expand to a vanilla call if a libgcc
19227 implementation is available (DECL_ASSEMBLER_NAME is set).
19228 (avr_fold_absfx): New static function.
19229 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
19230 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
19231 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
19232 AVR_BUILTIN_ABSLLK.
19233 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
19234 (abshk, absk, abslk, absllk): Provide as static inline functions.
19235
1f546bbb
MP
192362013-01-25 Marek Polacek <polacek@redhat.com>
19237
19238 PR tree-optimization/56035
19239 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
19240
7e184bd7
UB
192412012-01-24 Uros Bizjak <ubizjak@gmail.com>
19242
19243 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
19244 (*movtf_internal_rex64): Add (!o,C) alternative
19245 (*movxf_internal_rex64): Ditto.
19246 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
19247
f5ba49ac
SM
192482013-01-24 Shenghou Ma <minux.ma@gmail.com>
19249
19250 * doc/invoke.texi: fix typo.
19251 * doc/objc.texi: fix typo.
19252
a6343728
RS
192532013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
19254
19255 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
19256 for the first two alternatives.
19257
cd030c07
DN
192582013-01-24 Diego Novillo <dnovillo@google.com>
19259
77dc5297 19260 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN
19261 (ggc-zone.o): Remove.
19262 * configure.ac: Remove option --with-gc.
19263 * configure: Re-generate.
19264 * doc/install.texi: Remove documentation for --with-gc.
19265 * gengtype.c (write_enum_defn): Remove. Update all users.
19266 (write_Types_process_field): Remove generation of gt_e_* argument.
19267 (output_type_enum): Remove. Update all users.
19268 (write_enum_defn): Remove. Update all users.
19269 (enum alloc_zone): Remove. Update all users.
77dc5297 19270 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN
19271 * ggc-common.c (ggc_splay_alloc): Remove first argument.
19272 Update all callers.
19273 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 19274 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN
19275 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
19276 Update all users.
aaf1e810 19277 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN
19278 (struct alloc_zone): Remove.
19279 (ggc_internal_alloc_zone_stat): Remove.
19280 (ggc_internal_cleared_alloc_zone_stat): Remove.
19281 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB
19282 (ggc_pch_count_object): Remove last argument. Update all users.
19283 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN
19284 (struct alloc_zone): Remove.
19285 * ggc-zone.c: Remove.
77dc5297 19286 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN
19287 (struct alloc_zone): Remove.
19288 (ggc_alloc_typed_stat): Remove.
19289 (ggc_alloc_typed): Remove.
19290 (ggc_splay_alloc): Remove first argument.
19291 (rtl_zone): Remove. Update all users.
19292 (tree_zone): Remove. Update all users.
19293 (tree_id_zone): Remove. Update all users.
19294 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 19295 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 19296 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 19297 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 19298
a861ffa4
GJL
192992013-01-24 Georg-Johann Lay <avr@gjlay.de>
19300
19301 * config/avr/avr.c (avr_out_fract): Make register numbers that
19302 might be outside of source operand signed.
19303
593c0ddd
UB
193042013-01-24 Uros Bizjak <ubizjak@gmail.com>
19305
19306 * config/i386/constraints.md (Yf): New constraint.
19307 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
19308 of f constraint to conditionaly disable x87 register preferences.
19309 (*movdf_internal): Ditto.
19310 (*movsf_internal): Ditto.
19311
e86c0101
SB
193122013-01-24 Steven Bosscher <steven@gcc.gnu.org>
19313
19314 PR inline-asm/55934
19315 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
19316 that have operands with impossible constraints.
19317 Add a FIXME for a speed-up opportunity.
19318 * lra-constraints.c (process_alt_operands): Verify that a class
19319 selected from constraints on asms is valid for the operand mode.
19320 (curr_insn_transform): Remove incorrect comment.
19321
f6fee35f
DE
193222013-01-23 David Edelsohn <dje.gcc@gmail.com>
19323
19324 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
19325 TOC operand is a valid symbol ref in the constant pool.
19326
193272013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 19328
aaf1e810 19329 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 19330
dc62d7d1
GJL
193312013-01-23 Georg-Johann Lay <avr@gjlay.de>
19332
19333 PR target/54222
19334 * config/avr/stdfix.h: New file.
19335 * t-avr (stdfix-gcc.h): New rule to build it.
19336 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
19337
2660d12d
KS
193382013-01-23 Kostya Serebryany <kcc@google.com>
19339
77dc5297
UB
19340 * config/darwin.h: remove dependency on
19341 CoreFoundation (asan on Mac OS).
2660d12d 19342
a70418fc
JJ
193432013-01-23 Jakub Jelinek <jakub@redhat.com>
19344
19345 PR target/49069
19346 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
19347 instead of cmpdi_operand for first comparison operand.
19348 Don't assert that comparison operands aren't both constants.
19349
47876a2a
JW
193502013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
19351
19352 * doc/install.texi (Downloading the Source): Update references to
19353 downloading separate components.
19354
193552013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW
19356
19357 * doc/extend.texi (__int128): Improve grammar.
19358
47876a2a 193592013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB
19360
19361 PR target/56028
19362 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
19363 alternative to (o,r).
19364 (*movdi_internal_rex64): Remove (!o,n) alternative.
19365 (DImode immediate->memory splitter): Remove.
19366 (DImode immediate->memory peephole2): Remove.
19367 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
19368 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
19369 alternative to (!o,*r).
19370 (*movtf_internal_sse): New pattern.
19371 (*movxf_internal_rex64): New pattern.
19372 (*movxf_internal): Disable for TARGET_64BIT.
19373 (*movdf_internal_rex64): Remove (!o,F) alternative.
19374
3a984f10
JJ
193752013-01-22 Jakub Jelinek <jakub@redhat.com>
19376
502498d5
JJ
19377 PR middle-end/56074
19378 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
19379 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
19380 * tree-vect-loop-manip.c (find_loop_location): Also ignore
19381 stmt locations where LOCATION_LOCUS of the stmt location is
19382 UNKNOWN_LOCATION or BUILTINS_LOCATION.
19383
3a984f10
JJ
19384 PR target/55686
19385 * config/i386/i386.md (UNSPEC_STOS): New.
19386 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
19387 *strsetqi_1): Add UNSPEC_STOS.
19388
fa817f7f
PC
193892013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
19390
19391 PR c++/56067
19392 * doc/invoke.texi: Remove left over -Wsynth example.
19393
8f498c1b
JJ
193942013-01-21 Jakub Jelinek <jakub@redhat.com>
19395
19396 PR tree-optimization/56051
19397 * fold-const.c (fold_binary_loc): Don't fold
19398 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
19399 a narrowing conversion, or widening conversion from signed
19400 to unsigned.
19401
47876a2a 194022013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB
19403
19404 PR rtl-optimization/56023
19405 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
19406 dependent on debug instruction.
19407
5b9db1bc
MJ
194082013-01-21 Martin Jambor <mjambor@suse.cz>
19409
19410 PR middle-end/56022
19411 * function.c (allocate_struct_function): Call
19412 invoke_set_current_function_hook earlier.
19413
e8bb7d68
JJ
194142013-01-21 Jakub Jelinek <jakub@redhat.com>
19415
19416 * reload1.c (init_reload): Only initialize reload_obstack
19417 during the first call.
19418
616a4e32
MP
194192013-01-21 Marek Polacek <polacek@redhat.com>
19420
19421 * cfgloop.c (verify_loop_structure): Fix up grammar.
19422
4401981b
YHH
194232013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
19424
19425 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
19426 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
19427
8e87740b
RR
194282013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19429
19430 PR target/56058
19431 * config/arm/marvell-pj4.md: Update copyright year.
19432 Fix up use of alu to alu_reg and simple_alu_imm.
19433
47876a2a 194342013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB
19435
19436 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
19437
89d56d79
VM
194382013-01-20 Vladimir Makarov <vmakarov@redhat.com>
19439
19440 PR target/55433
19441 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 19442 insn for secondary memory move when memory mode should be different.
89d56d79 19443
fe603553
JDA
194442013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19445
19446 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
19447 atomic_storedi_1): New patterns.
19448
01284895
VK
194492013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19450
19451 btver2 pipeline descriptions.
19452 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
19453 descriptions.
19454 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 19455 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK
19456 type attributes.
19457 * config/i386/btver2.md: New file describing btver2 pipelines.
19458
5630e3e1
JL
194592013-01-19 Andrew Pinski <apinski@cavium.com>
19460
19461 PR tree-optimization/52631
19462 * tree-ssa-sccvn (visit_use): Before looking up the original
19463 statement, try looking up the simplified expression.
19464
650ae806
AG
194652013-01-19 Anthony Green <green@moxielogic.com>
19466
19467 * config/moxie/moxie.c (moxie_expand_prologue): Set
19468 current_function_static_stack_size.
19469
e300ec2d
JJ
194702013-01-18 Jakub Jelinek <jakub@redhat.com>
19471
19472 PR tree-optimization/56029
19473 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
19474 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
19475
a3d7af04
SS
194762013-01-18 Sharad Singhai <singhai@google.com>
19477
19478 PR tree-optimization/55995
19479 * dumpfile.c (dump_loc): Print location only if available.
19480 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
19481
66aa7879
VM
194822013-01-18 Vladimir Makarov <vmakarov@redhat.com>
19483
19484 PR target/55433
19485 * lra-constraints.c (curr_insn_transform): Reuse original insn for
19486 secondary memory move.
19487 (inherit_reload_reg): Use rclass instead of cl for
19488 check_secondary_memory_needed_p.
19489
3f0fee7b
JJ
194902013-01-18 Jakub Jelinek <jakub@redhat.com>
19491
19492 PR middle-end/56015
19493 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 19494 the case where writing real complex part of target modifies op1.
3f0fee7b 19495
70c67693
JG
194962013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
19497
19498 * config/aarch64/aarch64-simd.md
19499 (aarch64_vcond_internal<mode>): Handle unordered cases.
19500 * config/aarch64/iterators.md (v_cmp_result): New.
19501
df8de9b3
YHH
195022013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
19503 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19504
19505 * config/arm/marvell-pj4.md: New file.
19506 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
19507 * config/arm/arm.md (generic_sched): Add marvell_pj4.
19508 (generic_vfp): Likewise.
19509 * config/arm/arm-cores.def: Add marvell-pj4.
19510 * config/arm/arm-tune.md: Regenerate.
19511 * config/arm/arm-tables.opt: Regenerate.
19512 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
19513 * doc/invoke.texi: Document marvell-pj4.
19514
be30c356
TB
195152013-01-18 Tejas Belagod <tejas.belagod@arm.com>
19516
19517 * config/aarch64/arm_neon.h: Map scalar types to standard types.
19518
0bfbca58 195192013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO
19520
19521 PR debug/54114
19522 PR debug/54402
19523 PR debug/49888
19524 * var-tracking.c (negative_power_of_two_p): New.
19525 (global_get_addr_cache, local_get_addr_cache): New.
19526 (get_addr_from_global_cache, get_addr_from_local_cache): New.
19527 (vt_canonicalize_addr): Rewrite using the above. Adjust the
19528 heading comment.
19529 (vt_stack_offset_p): Remove.
19530 (vt_canon_true_dep): Always canonicalize loc's address.
19531 (clobber_overlapping_mems): Make sure we have a MEM.
19532 (local_get_addr_clear_given_value): New.
19533 (val_reset): Clear local cached entries.
19534 (compute_bb_dataflow): Create and release the local cache.
19535 Disable duplicate MEMs clobbering.
19536 (emit_notes_in_bb): Clobber MEMs likewise.
19537 (vt_emit_notes): Create and release the local cache.
19538 (vt_initialize, vt_finalize): Create and release the global
19539 cache, respectively.
1f6bc337 19540 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 19541
0bfbca58 195422013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO
19543
19544 PR libmudflap/53359
19545 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
19546 not found in the symtab.
19547
0bfbca58 195482013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 19549
c350ba53 19550 PR debug/56006
3aa03517
AO
19551 PR rtl-optimization/55547
19552 PR rtl-optimization/53827
19553 PR debug/53671
19554 PR debug/49888
19555 * alias.c (offset_overlap_p): New, factored out of...
19556 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
19557 the conservative special case for symbolic constants. Don't
19558 adjust zero sizes on alignment.
19559
c664546f
JL
195602013-01-18 Bernd Schmidt <bernds@codesourcery.com>
19561
19562 PR rtl-optimization/52573
19563 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
19564 REG_UNUSED for the same register.
19565
1bd3f750
MP
195662013-01-17 Richard Biener <rguenther@suse.de>
19567 Marek Polacek <polacek@redhat.com>
19568
19569 PR rtl-optimization/55833
19570 * loop-unswitch.c (unswitch_loops): Move loop verification...
19571 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
19572 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
19573 Set it to true when we're removing a loop from hierarchy tree in
19574 an irreducible region.
19575 (fix_bb_placements): Adjust caller.
19576 (fix_loop_placements): Likewise.
19577
e52a8b71
GJL
195782013-01-17 Georg-Johann Lay <avr@gjlay.de>
19579
19580 * config/avr/builtins.def (DEF_BUILTIN): Factor out
19581 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
19582 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
19583 Remove ID. Adjust comments.
19584 * config/avr/avr-c.c (avr_builtin_name): Remove.
19585 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
19586 * config/avr/avr.c (avr_tolower): New static function.
19587 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
19588 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
19589 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
19590 default expansion.
19591
8386a7ea
JH
195922013-01-17 Jan Hubicka <jh@suse.cz>
19593
610fb637 19594 PR tree-optimization/55273
8386a7ea
JH
19595 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
19596
47876a2a 195972013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB
19598
19599 PR target/55981
19600 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
19601 store through atomic_store<mode>_1.
19602 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
19603
8222c37e
MJ
196042013-01-17 Martin Jambor <mjambor@suse.cz>
19605
19606 PR tree-optimizations/55264
19607 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
19608 for virtual methods.
19609 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
19610 virtual methods before inlining is over.
19611 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
19612 virtual functions.
19613 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
19614 non-virtual.
19615
79f01c76
VM
196162013-01-16 Vladimir Makarov <vmakarov@redhat.com>
19617
19618 PR rtl-optimization/56005
19619 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
19620 pending reads for prefetch.
19621
d0b6bb1b
IB
196222013-01-16 Ian Bolton <ian.bolton@arm.com>
19623
aaf1e810 19624 * config/aarch64/aarch64.md
d0b6bb1b
IB
19625 (*cstoresi_neg_uxtw): New pattern.
19626 (*cmovsi_insn_uxtw): New pattern.
19627 (*<optab>si3_uxtw): New pattern.
19628 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
19629 (*<optab>si3_insn_uxtw): New pattern.
19630 (*bswapsi2_uxtw): New pattern.
19631
cb9cf03b
RB
196322013-01-16 Richard Biener <rguenther@suse.de>
19633
19634 * tree-inline.c (tree_function_versioning): Remove set but
19635 never used variable.
19636
2cfc56b9
RB
196372013-01-16 Richard Biener <rguenther@suse.de>
19638
19639 PR tree-optimization/55964
19640 * tree-flow.h (rename_variables_in_loop): Remove.
19641 (rename_variables_in_bb): Likewise.
19642 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
19643 (copy_loop_before): Adjust and delete update-ssa status.
19644 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 19645 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB
19646 (rename_variables_in_loop): Remove.
19647 (slpeel_update_phis_for_duplicate_loop): Likewise.
19648 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
19649 use available cfg machinery instead of duplicating it.
19650 Update PHI nodes and perform poor-mans SSA update here.
19651 (slpeel_tree_peel_loop_to_edge): Adjust.
19652
c25a0c60
RB
196532013-01-16 Richard Biener <rguenther@suse.de>
19654
19655 PR tree-optimization/54767
19656 PR tree-optimization/53465
19657 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
19658 (vrp_visit_phi_node): For PHI arguments coming via backedges
19659 drop all symbolical range information.
19660 (execute_vrp): Compute backedges.
19661
04b535af
RB
196622013-01-16 Richard Biener <rguenther@suse.de>
19663
19664 * doc/install.texi: Update CLooG and ISL requirements to
19665 0.18.0 and 0.11.1.
19666
8b0a1e0b
CB
196672013-01-16 Christian Bruel <christian.bruel@st.com>
19668
19669 PR target/55301
19670 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
19671 (broken_move): Handle UNSPECV_SP_SWITCH_B.
19672 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
19673
196742013-01-16 DJ Delorie <dj@redhat.com>
19675
19676 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
19677 (UNSPECV_SP_SWITCH_E): New.
19678 (sp_switch_1): Change to an unspec.
19679 (sp_switch_2): Change to an unspec. Don't use post-inc when we
19680 replace $r15.
19681
196822013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB
19683
19684 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
19685 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
19686 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
19687 (expand_mem_thread_fence): Ditto.
19688 (expand_mem_signal_fence): Ditto.
19689 (expand_atomic_load): Ditto.
19690 (expand_atomic_store): Ditto.
19691
0bfbca58 196922013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO
19693
19694 PR rtl-optimization/55547
19695 PR rtl-optimization/53827
19696 PR debug/53671
19697 PR debug/49888
19698 * alias.c (memrefs_conflict_p): Set sizes to negative after
19699 AND adjustments.
19700
305e3ac1
JJ
197012013-01-15 Jakub Jelinek <jakub@redhat.com>
19702
19703 PR target/55940
19704 * function.c (thread_prologue_and_epilogue_insns): Always
19705 add crtl->drap_reg to set_up_by_prologue.set, even if
19706 stack_realign_drap is false.
19707
f78ac4f2
JBG
197082013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
19709
19710 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
19711 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
19712 *call): Fix indention.
19713
a78a8cc4
TV
197142013-01-15 Tom de Vries <tom@codesourcery.com>
19715
19716 PR target/55876
19717 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
19718 Update comment.
19719
0e80383f
VM
197202013-01-15 Vladimir Makarov <vmakarov@redhat.com>
19721
305e3ac1 19722 PR rtl-optimization/55153
0e80383f
VM
19723 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
19724
207b5956
MJ
197252013-01-15 Martin Jambor <mjambor@suse.cz>
19726
19727 PR tree-optimization/55920
19728 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
19729 accesses as grp_to_be_debug_replaced.
19730
a7818b54
JJ
197312013-01-15 Jakub Jelinek <jakub@redhat.com>
19732
19733 PR tree-optimization/55920
19734 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
19735 there is non-useless type conversion needed from debug rhs to lhs,
19736 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
19737
b0fe107e
JM
197382013-01-15 Joseph Myers <joseph@codesourcery.com>
19739 Mikael Pettersson <mikpe@it.uu.se>
19740
19741 PR target/43961
19742 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
19743 Thumb.
19744 (ASM_OUTPUT_CASE_LABEL): Remove.
19745 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
19746 * final.c (shorten_branches): Update alignment of labels before
19747 jump tables if CASE_VECTOR_SHORTEN_MODE.
19748
34ab62ee
RB
197492013-01-15 Richard Biener <rguenther@suse.de>
19750
19751 PR bootstrap/55961
19752 * system.h: Do not include gmp.h for building host tools.
19753
783a3a05
RB
197542013-01-15 Richard Biener <rguenther@suse.de>
19755
19756 PR middle-end/55882
19757 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
19758 account for bitpos when computing alignment.
19759
3a579e09
VY
197602013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
19761
19762 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
19763 (ix86_target_macros_internal): Likewise.
19764
19765 * config/i386/i386.c (m_CORE2I7): Removed.
19766 (m_CORE_HASWELL): New macro.
19767 (m_CORE_ALL): Likewise.
19768 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
19769 (initial_ix86_arch_features): Likewise.
19770 (processor_target_table): Initializations for Core avx2.
19771 (cpu_names): New names "core-avx2".
19772 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
19773 PROCESSOR_CORE_HASWELL.
19774 (ix86_issue_rate): New case.
19775 (ia32_multipass_dfa_lookahead): Likewise.
19776 (ix86_sched_init_global): Likewise.
19777
19778 * config/i386/i386.h (TARGET_HASWELL): New macro.
19779 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
19780 (processor_type): New PROCESSOR_HASWELL.
19781
955f5a07
JJ
197822013-01-15 Jakub Jelinek <jakub@redhat.com>
19783
ff784829
JJ
19784 PR tree-optimization/55955
19785 * tree-vect-loop.c (vectorizable_reduction): Give up early on
19786 *SHIFT_EXPR and *ROTATE_EXPR codes.
19787
955f5a07
JJ
19788 PR tree-optimization/48766
19789 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
19790 -ftrapv disable -fwrapv.
19791
aeb8b4e9
GJL
197922013-01-14 Georg-Johann Lay <avr@gjlay.de>
19793
19794 PR target/55974
19795 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
19796 etc. to 1 and not to __flash.
19797 Use LL suffix for __INT24_MAX__ with -mint8.
19798 Use ULL suffix for __UINT24_MAX__ with -mint8.
19799
1c494c6a
GJL
198002013-01-14 Georg-Johann Lay <avr@gjlay.de>
19801
19802 * config/avr/avr-arch.h
19803 (struct base_arch_s): Use typedef avr_arch_t instead.
19804 (struct arch_info_s): Use typedef avr_arch_info_t instead.
19805 (struct mcu_type_s): Use typedef avr_mcu_t instead.
19806 * config/avr/avr.c: Same.
19807 * config/avr/avr-devices.c: Same.
19808 * config/avr/driver-avr.c: Same.
19809 * config/avr/gen-avr-mmcu-texi.c: Same.
19810 * config/avr/avr-mcus.def: Adjust comment.
19811
a50344cb
TB
198122013-01-14 Tejas Belagod <tejas.belagod@arm.com>
19813
88e784e6
UB
19814 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
19815 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 19816
47876a2a 198172013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB
19818 Andi Kleen <ak@linux.intel.com>
19819
19820 PR target/55948
19821 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
19822 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
19823 memmodel flag.
19824
00892272
GJL
198252013-01-14 Georg-Johann Lay <avr@gjlay.de>
19826
19827 * config/avr/avr-stdint.h: Remove trailing blanks.
19828 * config/avr/avr-log.h: Same.
19829 * config/avr/avr-arch.h: Same.
19830 * config/avr/avr-devices.c: Same.
19831 * config/avr/avr-dimode.md: Same.
19832 * config/avr/predicates.md: Same.
19833 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 19834
00892272
GJL
19835 * config/avr/avr-protos.h: Same. And:
19836 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
19837 (init_cumulative_args): Rename to avr_init_cumulative_args.
19838 (expand_prologue): Rename to avr_expand_prologue.
19839 (expand_epilogue): Rename to avr_expand_epilogue.
19840 (adjust_insn_length): Rename to avr_adjust_insn_length.
19841 (notice_update_cc): Rename to avr_notice_update_cc.
19842 (final_prescan_insn): Rename to avr_final_prescan_insn.
19843 * config/avr/avr.c: Same.
19844 * config/avr/avr.h: Same.
19845 * config/avr/avr.md: Remove trailing blanks.
19846 (prologue): Use avr_expand_prologue.
19847 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
19848
7bb01996
RB
198492013-01-14 Richard Biener <rguenther@suse.de>
19850
19851 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
19852 verify_location, collect_subblocks): New functions.
19853 (verify_gimple_in_cfg): Verify that locations only reference
19854 BLOCKs in the functions BLOCK tree.
19855
2724573f
RB
198562013-01-14 Richard Biener <rguenther@suse.de>
19857
19858 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
19859 PHI argument.
19860 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
19861 unshare reference.
19862 (insert_out_of_ssa_copy_on_edge): Likewise.
19863 (rewrite_close_phi_out_of_ssa): Likewise.
19864 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
19865 debug expressions.
19866 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
19867 propagated constants.
19868 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
19869 can not be shared.
19870
9a0bbab6
GJL
198712013-01-14 Georg-Johann Lay <avr@gjlay.de>
19872
19873 * config/avr/avr-modes.def: Add GPL copyright notice.
19874
45805f17
UB
198752013-01-13 Uros Bizjak <ubizjak@gmail.com>
19876
19877 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
19878 MEMMODEL_MASK to determine memory model.
19879 (atomic_store<mode>): Ditto from operands[2].
19880 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
19881
9d60be38
JJ
198822013-01-13 Jakub Jelinek <jakub@redhat.com>
19883
19884 PR fortran/55935
45805f17 19885 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ
19886 (fold_gimple_assign): Don't call unshare_expr here.
19887 (fold_ctor_reference): Call unshare_expr.
19888
e7f49d92
TG
198892013-01-13 Terry Guo <terry.guo@arm.com>
19890
aaf1e810
EB
19891 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
19892 * doc/fragments.texi: Document MULTILIB_REUSE.
19893 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 19894 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 19895 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG
19896 (tmpmultilib4): Ditto.
19897 (multilib_reuse): New multilib argument.
19898
fbd03a27
RS
198992013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
19900
19901 * Makefile.in: Update copyright.
19902
b3681f13
TV
199032013-01-12 Tom de Vries <tom@codesourcery.com>
19904
19905 PR middle-end/55890
19906 * calls.c (expand_call): Check if arg_nr is valid.
19907
3f287e4b
MM
199082013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
19909
19910 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
19911 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM
19912 documentation. Add missing '__' in front of
19913 __builtin_ia32_packssdw256.
3f287e4b 19914
1abcd5eb
AK
199152013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19916
19917 PR target/55719
19918 * config/s390/s390.c (s390_preferred_reload_class): Do not return
19919 NO_REGS for larl operands.
19920 (s390_reload_larl_operand): Use s390_load_address instead of
19921 emit_move_insn.
19922
980d0812
RB
199232013-01-11 Richard Biener <rguenther@suse.de>
19924
19925 * tree-cfg.c (verify_node_sharing_1): Split out from ...
19926 (verify_node_sharing): ... here.
19927 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
19928
7cb9fd07
EB
199292013-01-11 Eric Botcazou <ebotcazou@adacore.com>
19930
19931 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
19932 Substitute TREECHECKING.
19933 * configure: Regenerate.
19934 * Makefile.in (TREECHECKING): New.
19935
47876a2a 199362013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG
19937
19938 PR tree-optimization/44061
19939 * tree-vrp.c (extract_range_basic): Compute zero as
19940 value-range for __builtin_constant_p of function parameters.
19941
d1e082c2
RS
199422013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
19943
45805f17 19944 Update copyright years.
d1e082c2 19945
f85021b0
VM
199462013-01-10 Vladimir Makarov <vmakarov@redhat.com>
19947
9d60be38 19948 PR rtl-optimization/55672
0160442c
VM
19949 * lra-eliminations.c (mark_not_eliminable): Permit addition with
19950 const to be eliminable.
f85021b0 19951
7a8b1ec4
DE
199522013-01-10 David Edelsohn <dje.gcc@gmail.com>
19953
19954 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
19955 * configure: Regenerate.
19956
ad2e5b71
RB
199572013-01-10 Richard Biener <rguenther@suse.de>
19958
ddf9322d 19959 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 19960
ffc5b2bb
RB
199612013-01-10 Richard Biener <rguenther@suse.de>
19962
19963 PR bootstrap/55792
19964 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
19965 locations for virtual PHI arguments.
19966 (rewrite_update_phi_arguments): Likewise.
19967
e1f674e4
JS
199682013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
19969
19970 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
19971 on to assembler.
19972
a57fc743
JJ
199732013-01-10 Jakub Jelinek <jakub@redhat.com>
19974
19975 PR tree-optimization/55921
19976 * tree-complex.c (expand_complex_asm): New function.
19977 (expand_complex_operations_1): Call it for GIMPLE_ASM.
19978
0ff4390d
AK
199792013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19980
19981 PR target/55718
19982 * config/s390/s390.c (s390_symref_operand_p)
19983 (s390_loadrelative_operand_p): Merge the two functions.
19984 (s390_check_qrst_address, print_operand_address): Add parameters
19985 to s390_loadrelative_operand_p invokation.
19986 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
19987 (s390_reload_larl_operand, s390_secondary_reload): Use
19988 s390_loadrelative_operand_p instead of s390_symref_operand_p.
19989 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
19990
6f557e0e
MS
199912013-01-09 Mike Stump <mikestump@comcast.net>
19992
19993 * dse.c (record_store): Remove unnecessary assert.
19994
7770c9e9
JH
199952013-01-09 Jan Hubicka <jh@suse.cz>
19996
19997 PR tree-optimization/55569
19998 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
19999 * cfgloop.h (scale_loop_profile): Likewise.
20000
a19b1432
JH
200012013-01-09 Jan Hubicka <jh@suse.cz>
20002
20003 PR lto/45375
ddf9322d
UB
20004 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
20005 functions.
a19b1432
JH
20006 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
20007
9a002da8
RS
200082013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
20009
20010 PR middle-end/55114
20011 * expr.h (maybe_emit_group_store): Declare.
20012 * expr.c (maybe_emit_group_store): New function.
20013 * builtins.c (expand_builtin_int_roundingfn): Call it.
20014 (expand_builtin_int_roundingfn_2): Likewise.
20015
511dcace
VM
200162013-01-09 Vladimir Makarov <vmakarov@redhat.com>
20017
e1f2b729 20018 PR rtl-optimization/55829
511dcace
VM
20019 * lra-constraints.c (match_reload): Add code for absent output.
20020 (curr_insn_transform): Add code for reloads of matched inputs
20021 without output.
20022
7b0fe4f4
UB
200232013-01-09 Uros Bizjak <ubizjak@gmail.com>
20024
20025 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
20026 attribute of movddup insn to DF.
20027 (*vec_interleave_lowv2df): Ditto.
20028 (vec_dupv2df): Ditto.
20029
870ca331
JH
200302013-01-09 Jan Hubicka <jh@suse.cz>
20031
20032 PR tree-optimiation/55875
20033 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
20034 EVERY_ITERATION parameter.
7b0fe4f4 20035 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH
20036 (idx_infer_loop_bounds): Similarly here.
20037 (n_of_executions_at_most): Simplify
20038 to only test for cases where statement is dominated by the
7b0fe4f4 20039 particular bound; handle correctly the "postdominance" test.
870ca331
JH
20040 (scev_probably_wraps_p): Use max loop iterations info
20041 as a global bound first.
20042
6f575fe4 200432013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD
20044 Nick Clifton <nickc@redhat.com>
20045
20046 * config/v850/v850.md (cbranchsf4): New pattern.
20047 (cstoresf4): New pattern.
20048 (cbranchdf4): New pattern.
20049 (cstoredf4): New pattern.
20050 (movsicc): Disallow floating point comparisons.
20051 (cmpsf_le_insn): Fix order of operators.
20052 (cmpsf_lt_insn): Likewise.
20053 (cmpsf_eq_insn): Likewise.
20054 (cmpdf_le_insn): Likewise.
20055 (cmpdf_lt_insn): Likewise.
20056 (cmpdf_eq_insn): Likewise.
20057 (cmpsf_ge_insn): Use LE comparison.
20058 (cmpdf_ge_insn): Likewise.
20059 (cmpsf_gt_insn): Use LT comparison.
20060 (cmpdf_gt_insn): Likewise.
20061 (cmpsf_ne_insn): Delete pattern.
20062 (cmpdf_ne_insn): Delete pattern.
20063 * config/v850/v850.c (v850_gen_float_compare): Use
20064 gen_cmpdf_eq_insn for NE comparison.
20065 (v850_float_z_comparison_operator)
20066 (v850_float_nz_comparison_operator): Move from here ...
20067 * config/v850/predicates.md: ... to here. Move GT and GE
20068 comparisons into v850_float_z_comparison_operator.
20069 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
20070 Delete prototype.
20071 (v850_float_nz_comparison_operator): Likewise.
20072
f0d54148
JDA
200732013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20074
20075 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
20076 with calls to gen_insvsi/gen_insvdi.
20077
8f01beca
VK
200782013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20079
20080 * config/i386/i386.c (initial_ix86_tune_features): Set up
20081 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
20082
2aa579ad
JJ
200832013-01-09 Steven Bosscher <steven@gcc.gnu.org>
20084 Jakub Jelinek <jakub@redhat.com>
20085
20086 PR tree-optimization/48189
20087 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
20088 If nitercst is 0, don't predict the exit edge.
20089
6edc3e32 200902013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N
20091
20092 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
20093 in asm_fprintf with reg_names.
20094 (aarch64_print_operand_address): Likewise.
20095 (aarch64_return_addr): Likewise.
20096 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
20097
f6f94d94
JDA
200982013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20099
20100 * config/pa/pa.h (VAL_U6_BITS_P): Define.
20101 (INT_U6_BITS): Likewise.
20102 * config/pa/predicates.md (uint6_operand): New predicate.
20103 (shift5_operand, shift6_operand): Likewise.
20104 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
20105 arith32_operand.
20106 (lshrdi3): Use shift6_operand.
20107 (shrpsi4, shrpdi4): New insn patterns.
20108 (extzv): Delete expander.
20109 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
20110 predicates in unamed zero extract patterns. Tighten common constraint.
20111 (extv): Delete expander.
20112 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
20113 predicates in unamed sign extract patterns. Tighten common constraint.
20114 (insv): Delete expander.
20115 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
20116 predicates in unamed insert patterns. Tighten common constraint.
20117 Change uint32_operand predicate to uint6_operand predicate in unamed
20118 DImode pattern to insert constant values of type 1...1xxxx.
20119
36b72910
JH
201202013-01-04 Jan Hubicka <jh@suse.cz>
20121
20122 PR tree-optimization/55823
7b0fe4f4
UB
20123 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
20124 issue.
36b72910 20125
47876a2a 201262013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ
20127 Uros Bizjak <ubizjak@gmail.com>
20128
20129 PR rtl-optimization/55845
20130 * df-problems.c (can_move_insns_across): Stop scanning at
20131 volatile_insn_p source instruction or give up if
20132 across_from .. across_to range contains any volatile_insn_p
20133 instructions.
20134
4369c11e
TB
201352013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20136
7b0fe4f4
UB
20137 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
20138 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
20139 Declare.
4369c11e 20140 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 20141 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 20142
aeb7e7c1
JJ
201432013-01-08 Jakub Jelinek <jakub@redhat.com>
20144
20145 PR fortran/55341
20146 * asan.c (asan_clear_shadow): New function.
20147 (asan_emit_stack_protection): Use it.
20148
a02ad1aa
TB
201492013-01-08 Tejas Belagod <tejas.belagod@arm.com>
20150
20151 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
20152 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
20153 with tab instead of space.
20154
f90d87f5
NC
201552013-01-08 Nick Clifton <nickc@redhat.com>
20156
20157 * config/rl78/rl78.c (rl78_expand_prologue): Always select
20158 register bank 0 at the start of an interrupt handler.
83ffd964
NC
20159 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
20160 MDBH registers.
f90d87f5 20161
385eb93d
JG
201622013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20163
20164 * config/aarch64/aarch64-simd.md
20165 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
20166 (aarch64_simd_bsl): Likewise.
20167 (aarch64_vcond_internal<mode>): Likewise.
20168 (vcond<mode><mode>): Likewise.
20169 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
20170 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
20171
4dcd1054
JG
201722013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
20173
20174 * config/aarch64/aarch64-builtins.c
20175 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
20176
4267a4a6
MJ
201772013-01-08 Martin Jambor <mjambor@suse.cz>
20178
20179 PR debug/55579
20180 * tree-sra.c (analyze_access_subtree): Return true also after
20181 potentially creating a debug-only replacement.
20182
5f4e6de3
JJ
201832013-01-08 Jakub Jelinek <jakub@redhat.com>
20184
3138f224
JJ
20185 PR middle-end/55890
20186 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
20187
5f4e6de3
JJ
20188 PR tree-optimization/54120
20189 * tree-vrp.c (range_fits_type_p): Don't allow
20190 src_precision < precision from signed vr to unsigned_p
20191 if vr->min or vr->max is negative.
20192 (simplify_float_conversion_using_ranges): Test can_float_p
20193 against CODE_FOR_nothing.
20194
f3ef18ff
JJ
201952013-01-08 Jakub Jelinek <jakub@redhat.com>
20196 Richard Biener <rguenther@suse.de>
20197
20198 PR middle-end/55851
20199 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
20200 types instead of just INTEGER_TYPE types.
20201
25c210f9
MK
202022013-01-07 Mark Kettenis <kettenis@openbsd.org>
20203
20204 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
20205 TF_SIZE): Define.
f90d87f5 20206
81826a7b
SE
202072013-01-07 Steve Ellcey <sellcey@mips.com>
20208
20209 PR target/42661
20210 * config/mips/mips.opt: Change mad to mmad to match documentation.
20211
35678514
GJL
202122013-01-07 Georg-Johann Lay <avr@gjlay.de>
20213
20214 PR target/55897
20215 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
20216 .progmemx.data now.
20217
e5669488
GJL
202182013-01-07 Georg-Johann Lay <avr@gjlay.de>
20219
20220 PR target/55897
20221 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
20222 (avr_addrspace_t): Add .section_name field.
20223 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
20224 array size.
20225 (avr_addrspace): Same. Initialize .section_name. Remove last
20226 NULL entry. Put __memx into .progmemx.data.
20227 (progmem_section_prefix): Remove.
20228 (avr_asm_init_sections): No need to initialize progmem_section.
20229 (avr_asm_named_section): Use avr_addrspace[].section_name to get
20230 section name prefix.
20231 (avr_asm_select_section): Ditto. And use get_unnamed_section to
20232 retrieve the progmem section.
20233 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
20234 boundary to run over avr_addrspace[].
20235 (avr_register_target_pragmas): Ditto.
20236
656e6f37
JJ
202372013-01-06 Jakub Jelinek <jakub@redhat.com>
20238
20239 * varasm.c (output_constant_def_contents): For asan_protect_global
20240 protected strings, adjust DECL_ALIGN if needed, before testing for
20241 anchored symbols.
20242 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
20243 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
20244 normal decls.
20245 (output_object_block): For asan protected decls, emit asan padding
20246 after their contents.
20247 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
20248 (asan_finish_file): Test it here instead.
20249
6edc3e32
UB
202502013-01-07 Nick Clifton <nickc@redhat.com>
20251 Matthias Klose <doko@debian.org>
20252 Doug Kwan <dougkwan@google.com>
20253 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC
20254
20255 PR driver/55470
20256 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
20257
20258 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
20259
20260 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
20261
7b0fe4f4 20262 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC
20263
20264 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
20265
2a095093
GJL
202662013-01-07 Georg-Johann Lay <avr@gjlay.de>
20267
f3b503f4 20268 PR target/54461
2a095093
GJL
20269 * doc/install.texi (Cross-Compiler-Specific Options): Document
20270 --with-avrlibc.
20271
383f9b34
TB
202722013-01-07 Tejas Belagod <tejas.belagod@arm.com>
20273
20274 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
20275 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
20276 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
20277 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
20278 vqmovun_high_s64): Fix source operand number and update copyright.
20279
3626621a
RB
202802013-01-07 Richard Biener <rguenther@suse.de>
20281
20282 PR middle-end/55890
20283 * gimple.h (gimple_call_builtin_p): New overload.
20284 * gimple.c (validate_call): New function.
20285 (gimple_call_builtin_p): Likewise.
20286 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20287 Use gimple_call_builtin_p.
20288 (find_func_clobbers): Likewise.
20289 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
20290 (strlen_optimize_stmt): Likewise.
20291
8b2ea410
JG
202922013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20293
20294 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
20295 (vld1q_dup_*): Likewise.
20296 (vld1_*): Likewise.
20297 (vld1q_*): Likewise.
20298 (vld1_lane_*): Likewise.
20299 (vld1q_lane_*): Likewise.
20300
9713d329
RB
203012013-01-07 Richard Biener <rguenther@suse.de>
20302
20303 * lto-streamer.h (LTO_minor_version): Bump to 2.
20304
3520f7cc
JG
203052013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
20306
20307 * config/aarch64/aarch64-protos.h
20308 (aarch64_const_double_zero_rtx_p): Rename to...
20309 (aarch64_float_const_zero_rtx_p): ...this.
20310 (aarch64_float_const_representable_p): New.
20311 (aarch64_output_simd_mov_immediate): Likewise.
20312 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
20313 move immediate case.
20314 * config/aarch64/aarch64.c
20315 (aarch64_const_double_zero_rtx_p): Rename to...
20316 (aarch64_float_const_zero_rtx_p): ...this.
20317 (aarch64_print_operand): Allow printing of new constants.
20318 (aarch64_valid_floating_const): New.
20319 (aarch64_legitimate_constant_p): Check for valid floating-point
20320 constants.
20321 (aarch64_simd_valid_immediate): Likewise.
20322 (aarch64_vect_float_const_representable_p): New.
20323 (aarch64_float_const_representable_p): Likewise.
20324 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
20325 (aarch64_output_simd_mov_immediate): New.
20326 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
20327 (*movdf_aarch64): Likewise.
20328 * config/aarch64/constraints.md (Ufc): New.
20329 (Y): call aarch64_float_const_zero_rtx.
20330 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
20331
e167c04d
RB
203322013-01-07 Richard Biener <rguenther@suse.de>
20333
20334 PR tree-optimization/55888
20335 PR tree-optimization/55862
20336 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
20337 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
20338 not if it is contained therein.
20339
0139aaab
GJL
203402013-01-07 Georg-Johann Lay <avr@gjlay.de>
20341
20342 * config/avr/t-avr: Typo.
20343
4a176b23
GJL
203442013-01-07 Georg-Johann Lay <avr@gjlay.de>
20345
20346 PR55243
20347 * config/avr/t-avr: Don't automatically rebuild
20348 $(srcdir)/config/avr/t-multilib
20349 $(srcdir)/config/avr/avr-tables.opt
20350 $(srcdir)/doc/avr-mmcu.texi
20351 (avr-mcus): New phony target to build them on request.
20352 (s-avr-mlib, s-avr-mmcu-texi): Remove.
20353 * avr/avr-mcus.def: Adjust comments.
20354
c7afdc98
UB
203552013-01-07 Uros Bizjak <ubizjak@gmail.com>
20356
20357 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
20358
1ab05c31
RS
203592013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20360
20361 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
20362
488e3acc
RS
203632013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
20364
20365 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
20366
a89599d2
DE
203672013-01-05 David Edelsohn <dje.gcc@gmail.com>
20368
20369 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
20370 to generate profiling.
20371 * config/rs6000/aix64.h (LIB_SPEC): Same.
20372
70f09188
AP
203732013-01-04 Andrew Pinski <apinski@cavium.com>
20374
20375 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
20376 New function.
20377 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
20378
918d445f
UB
203792013-01-04 Uros Bizjak <ubizjak@gmail.com>
20380
20381 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
20382 unconditionally.
20383 (ix86_expand_move): Ditto.
20384 (ix86_zero_extend_to_Pmode): Ditto.
20385 (ix86_expand_call): Ditto.
20386 (ix86_expand_special_args_builtin): Ditto.
20387 (ix86_expand_builtin): Ditto.
20388
361618ec
RB
203892013-01-04 Richard Biener <rguenther@suse.de>
20390
20391 PR tree-optimization/55862
20392 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
20393 translating them through PHI nodes.
20394
87eab554
MJ
203952013-01-04 Martin Jambor <mjambor@suse.cz>
20396
20397 PR tree-optimization/55755
20398 * tree-sra.c (sra_modify_assign): Do not check that an access has no
20399 children when trying to avoid producing a VIEW_CONVERT_EXPR.
20400
33879b9f
MP
204012013-01-04 Marek Polacek <polacek@redhat.com>
20402
20403 PR middle-end/55859
20404 * opts.c (default_options_optimization): Clarify error message.
20405
3068819a
RB
204062013-01-04 Richard Biener <rguenther@suse.de>
20407
20408 PR middle-end/55863
20409 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
20410 reassociation.
20411
0e4ae794
JDA
204122013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20413
20414 PR target/53789
20415 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA
20416 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
20417 references.
0e4ae794 20418
65c88cef
RH
204192013-01-03 Richard Henderson <rth@redhat.com>
20420
20421 * config/i386/i386.c (ix86_expand_move): Always assign to op1
20422 after eliminating TLS symbols.
20423
8ac16127
MG
204242013-01-03 Marc Glisse <marc.glisse@inria.fr>
20425
20426 PR bootstrap/50167
20427 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
20428 * graphite-poly.c (debug_gmp_value): Likewise.
20429
bb664f09
UB
204302013-01-03 Uros Bizjak <ubizjak@gmail.com>
20431
20432 PR target/55712
20433 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
20434 selected code model, define __code_mode_small__, __code_model_medium__,
20435 __code_model_large__, __code_model_32__ or __code_model_kernel__.
20436 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
20437 xchg temporary register with %k. Declare temporary register as
20438 early clobbered.
20439 [__x86_64__]: For medium and large code models, preserve %rbx register.
20440
bcf1ef00
RB
204412013-01-03 Richard Biener <rguenther@suse.de>
20442
0506634a 20443 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB
20444 (dump_subscript): Adjust.
20445 (finalize_ddr_dependent): Do not dump redundant info.
20446 (analyze_siv_subscript): Adjust.
20447 (subscript_dependence_tester): Likewise.
20448 (compute_affine_dependence): Likewise.
20449
59fd17e3
RB
204502013-01-03 Richard Biener <rguenther@suse.de>
20451
20452 Revert
20453 2013-01-03 Richard Biener <rguenther@suse.de>
20454
20455 PR tree-optimization/55857
20456 * tree-vect-stmts.c (vectorizable_load): Do not setup
20457 re-alignment for invariant loads.
20458
20459 2013-01-02 Richard Biener <rguenther@suse.de>
20460
20461 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 20462 invariant load do not generate a vector load from the scalar location.
59fd17e3 20463
595c2679
RB
204642013-01-03 Richard Biener <rguenther@suse.de>
20465
20466 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
20467 for not vectorizing.
20468 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
20469 not build INDIRECT_REFs, call get_name once only.
20470 (vect_create_data_ref_ptr): Likewise. Dump base object kind
20471 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
20472
90bb7d7a
RB
204732013-01-03 Richard Biener <rguenther@suse.de>
20474
20475 PR tree-optimization/55857
20476 * tree-vect-stmts.c (vectorizable_load): Do not setup
20477 re-alignment for invariant loads.
20478
f09b77ca
RB
204792013-01-03 Richard Biener <rguenther@suse.de>
20480
20481 PR lto/55848
20482 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
20483 prefer a built-in decl.
20484
3461a16e
JJ
204852013-01-03 Jakub Jelinek <jakub@redhat.com>
20486
df8e2b4f
JJ
20487 * gcc.c (process_command): Update copyright notice dates.
20488 * gcov.c (print_version): Likewise.
20489 * gcov-dump.c (print_version): Likewise.
20490
3461a16e
JJ
20491 PR rtl-optimization/55838
20492 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
20493 iv0.step, iv1.step and step.
20494
8b5546d6
JJ
204952013-01-03 Jakub Jelinek <jakub@redhat.com>
20496 Marc Glisse <marc.glisse@inria.fr>
20497
20498 PR tree-optimization/55832
20499 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
20500 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
20501 integer_{one,zero}_node.
20502
8ab1d2e9
JJ
205032013-01-03 Jakub Jelinek <jakub@redhat.com>
20504
20505 PR debug/54402
20506 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
20507 * var-tracking.c (reverse_op): Don't add reverse ops to
20508 VALUEs that have already
0506634a 20509 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 20510
5b9ad1d4
GP
205112013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
20512
20513 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
20514
e25a6711
TJ
205152013-01-02 Teresa Johnson <tejohnson@google.com>
20516
20517 * dumpfile.c (dump_loc): Print filename with location.
20518 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
20519 new location_t parameter to emit complete unroll message with
20520 new dump framework.
20521 (canonicalize_loop_induction_variables): Compute loops location
20522 and pass to try_unroll_loop_completely.
20523 * loop-unroll.c (report_unroll_peel): New function.
20524 (peel_loops_completely): Use new dump format with location
20525 for main dumpfile message, and invoke report_unroll_peel on success.
20526 (decide_unrolling_and_peeling): Ditto.
20527 (decide_peel_once_rolling): Remove old dumpfile message subsumed
20528 by report_unroll_peel.
20529 (decide_peel_completely): Ditto.
20530 (decide_unroll_constant_iterations): Ditto.
20531 (decide_unroll_runtime_iterations): Ditto.
20532 (decide_peel_simple): Ditto.
20533 (decide_unroll_stupid): Ditto.
20534 * cfgloop.c (get_loop_location): New function.
20535 * cfgloop.h (get_loop_location): Declare.
20536
77878621
ST
205372013-01-02 Sriraman Tallam <tmsriram@google.com>
20538
20539 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
20540 NULL.
20541
9e65d03e
JDA
205422013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20543
20544 PR middle-end/55198
20545 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
20546 BLKmode objects when EXPAND_MEMORY is specified.
20547
6a7da30f
ST
205482013-01-02 Sriraman Tallam <tmsriram@google.com>
20549
20550 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
20551 in loop predicate.
20552 (fold_builtin_cpu): Do not share cpu model decls across statements.
20553
e78167f2
JM
205542013-01-02 Jason Merrill <jason@redhat.com>
20555
20556 PR c++/55804
20557 * tree.c (build_array_type_1): Revert earlier change.
20558
8c075fb4
YZ
205592013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
20560
20561 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
20562 "cortex-a57".
20563 * config/aarch64/aarch64-tune.md: Re-generate.
20564
0682ed3e
RB
205652013-01-02 Richard Biener <rguenther@suse.de>
20566
20567 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 20568 invariant load do not generate a vector load from the scalar location.
0682ed3e 20569
b8f6a302
RB
205702013-01-02 Richard Biener <rguenther@suse.de>
20571
20572 PR bootstrap/55784
20573 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
20574 * configure: Regenerate.
20575
04b80dbb
RS
205762013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
20577
20578 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
20579 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
20580 (expand_builtin_int_roundingfn_2): Keep the original target around
20581 for the fallback case.
20582
635b0b0c
RS
205832013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
20584
20585 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
20586 to be clear for sign changes.
20587
42d57399
JH
205882013-01-01 Jan Hubicka <jh@suse.cz>
20589
20590 * ipa-inline-analysis.c: Fix formatting.
20591
5bb6669d
JJ
205922013-01-01 Jakub Jelinek <jakub@redhat.com>
20593
20594 PR tree-optimization/55831
20595 * tree-vect-loop.c (get_initial_def_for_induction): Use
20596 gsi_after_labels instead of gsi_start_bb.
ad41bd84 20597\f
86a2db33 20598Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM
20599
20600Copying and distribution of this file, with or without modification,
20601are permitted in any medium without royalty provided the copyright
20602notice and this notice are preserved.