1 2013-11-13 Christophe Lyon <christophe.lyon@linaro.org>
3 * config/aarch64/aarch64.h (FRAME_GROWS_DOWNWARD): Define to 1.
4 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5 Update offset calculations.
7 2013-11-13 Eric Botcazou <ebotcazou@adacore.com>
10 * dwarf2out.c (add_byte_size_attribute): Also use int_size_in_bytes
11 for fields. Do not add the attribute if the size is negative.
13 2013-11-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15 * config/arm/arm.c: Include aarch-cost-tables.h.
16 (generic_extra_costs): Move from here...
17 * config/arm/aarch-cost-tables.h: ... To here. New file.
19 2013-11-13 Alexander Ivchenko <alexander.ivchenko@intel.com>
20 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
21 Sergey Lega <sergey.s.lega@intel.com>
22 Anna Tikhonova <anna.tikhonova@intel.com>
23 Ilya Tocar <ilya.tocar@intel.com>
24 Andrey Turetskiy <andrey.turetskiy@intel.com>
25 Ilya Verbin <ilya.verbin@intel.com>
26 Kirill Yukhin <kirill.yukhin@intel.com>
27 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
29 * config/i386/i386.c (ix86_print_operand): Support z-masking
30 * config/i386/predicate.md (const_0_to_4_operand): New.
31 (const_0_to_5_operand): Ditto.
32 * config/i386/sse.md (UNSPEC_COMPRESS): New.
33 (UNSPEC_COMPRESS_STORE): Ditto.
34 (UNSPEC_EXPAND): Ditto.
35 (UNSPEC_EMBEDDED_ROUNDING): Ditto.
36 (define_mode_attr ssescalarsize): Ditto.
37 (avx512f_load<mode>_mask): Ditto.
38 (avx512f_store<mode>_mask): Ditto.
39 (avx512f_storedqu<mode>_mask): Ditto.
40 (avx512f_vmcmp<mode>3_mask): Ditto.
41 (avx512f_fmadd_<mode>_mask): Ditto.
42 (avx512f_fmadd_<mode>_mask3): Ditto.
43 (avx512f_fmsub_<mode>_mask): Ditto.
44 (avx512f_fmsub_<mode>_mask3): Ditto.
45 (avx512f_fnmadd_<mode>_mask): Ditto.
46 (avx512f_fnmadd_<mode>_mask3): Ditto.
47 (avx512f_fnmsub_<mode>_mask): Ditto.
48 (avx512f_fnmsub_<mode>_mask3): Ditto.
49 (avx512f_fmaddsub_<mode>_mask): Ditto.
50 (avx512f_fmaddsub_<mode>_mask3): Ditto.
51 (avx512f_fmsubadd_<mode>_mask): Ditto.
52 (avx512f_fmsubadd_<mode>_mask3): Ditto.
53 (vec_unpacku_float_lo_v16si): Ditto.
54 (avx512f_vextract<shuffletype>32x4_mask): Ditto.
55 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
56 (avx512f_vextract<shuffletype>64x4_mask): Ditto.
57 (vec_extract_lo_<mode>_maskm): Ditto.
58 (vec_extract_hi_<mode>_maskm): Ditto.
59 (avx512f_vternlog<mode>_mask): Ditto.
60 (avx512f_shufps512_mask): Ditto.
61 (avx512f_fixupimm<mode>_mask): Ditto.
62 (avx512f_shufpd512_mask): Ditto.
63 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
64 (avx512f_<code>v8div16qi2_mask/trunc): Ditto.
65 (*avx512f_<code>v8div16qi2_store_mask): Ditto.
66 (ashr<mode>3<mask_name>): Ditto.
67 (avx512f_vinsert<shuffletype>32x4_mask): Ditto.
68 (avx512f_vinsert<shuffletype>64x4_mask): Ditto.
69 (avx512f_shuf_<shuffletype>64x2_mask): Ditto.
70 (avx512f_shuf_<shuffletype>32x4_mask): Ditto.
71 (avx512f_pshufdv3_mask): Ditto.
72 (avx512f_perm<mode>_mask): Ditto.
73 (avx512f_vpermi2var<mode>3_mask): Ditto.
74 (avx512f_vpermt2var<mode>3_mask): Ditto.
75 (avx512f_compress<mode>_mask): Ditto.
76 (avx512f_compressstore<mode>_mask): Ditto.
77 (avx512f_expand<mode>_mask): Ditto.
78 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>): Extend
80 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
81 (<plusminus_insn><mode>3<mask_name>): Ditto.
82 (*<plusminus_insn><mode>3<mask_name>): Ditto.
83 (mul<mode>3<mask_name>): Ditto.
84 (*mul<mode>3<mask_name>): Ditto.
85 (<sse>_div<mode>3<mask_name>): Ditto.
86 (<mask_codefor>rcp14<mode><mask_name>): Ditto.
87 (<sse>_sqrt<mode>2<mask_name>): Ditto.
88 (<mask_codefor>rsqrt14<mode><mask_name>): Ditto.
89 (<code><mode>3<mask_name>/smaxmin): Ditto.
90 (*<code><mode>3_finite<mask_name>/smaxmin): Ditto.
91 (*<code><mode>3<mask_name>/smaxmin): Ditto.
92 (float<sseintvecmodelower><mode>2<mask_name>): Ditto.
93 (ufloatv16siv16sf2<mask_name>): Ditto.
94 (<mask_codefor>avx512f_fix_notruncv16sfv16si<mask_name>): Ditto.
95 (<mask_codefor>avx512f_ufix_notruncv16sfv16si<mask_name>): Ditto.
96 (<fixsuffix>fix_truncv16sfv16si2<mask_name>): Ditto.
97 (float<si2dfmodelower><mode>2<mask_name>): Ditto.
98 (ufloatv8siv8df<mask_name>): Ditto.
99 (<mask_codefor>avx512f_cvtpd2dq512<mask_name>): Ditto.
100 (avx512f_ufix_notruncv8dfv8si<mask_name>): Ditto.
101 (<fixsuffix>fix_truncv8dfv8si2<mask_name>): Ditto.
102 (<mask_codefor>avx512f_cvtpd2ps512<mask_name>): Ditto.
103 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix><mask_name>): Ditto.
104 (<mask_codefor>avx512f_unpckhps512<mask_name>): Ditto.
105 (<mask_codefor>avx512f_unpcklps512<mask_name>): Ditto.
106 (<mask_codefor>avx512f_movshdup512<mask_name>): Ditto.
107 (<mask_codefor>avx512f_movsldup512<mask_name>): Ditto.
108 (<mask_codefor>avx512f_vextract<shuffletype>32x4_1<mask_name>): Ditto.
109 (vec_extract_lo_<mode><mask_name>): Ditto.
110 (vec_extract_hi_<mode><mask_name>): Ditto.
111 (<mask_codefor>avx512f_unpckhpd512<mask_name>): Ditto.
112 (avx512f_movddup512<mask_name>): Ditto.
113 (avx512f_unpcklpd512<mask_name>): Ditto.
114 (*avx512f_unpcklpd512<mask_name>): Ditto.
115 (*avx512f_vmscalef<mode>): Ditto.
116 (avx512f_scalef<mode><mask_name>): Ditto.
117 (avx512f_getexp<mode><mask_name>): Ditto.
118 (<mask_codefor>avx512f_align<mode><mask_name>): Ditto.
119 (avx512f_rndscale<mode><mask_name>): Ditto.
120 (avx512f_shufps512_1<mask_name>): Ditto.
121 (avx512f_shufpd512_1<mask_name>): Ditto.
122 (<plusminus_insn><mode>3<mask_name>): Ditto.
123 (*<plusminus_insn><mode>3<mask_name>): Ditto.
124 (vec_widen_umult_even_v16si<mask_name>): Ditto.
125 (*vec_widen_umult_even_v16si<mask_name>): Ditto.
126 (vec_widen_smult_even_v16si<mask_name>): Ditto.
127 (*vec_widen_smult_even_v16si<mask_name>): Ditto.
128 (mul<mode>3<mask_name>): Ditto.
129 (*<sse4_1_avx2>_mul<mode>3<mask_name>): Ditto.
130 (<shift_insn><mode>3<mask_name>): Ditto.
131 (avx512f_<rotate>v<mode><mask_name>/rotate): Ditto.
132 (avx512f_<rotate><mode><mask_name>): Ditto.
133 (<code><mode>3<mask_name>/maxmin): Ditto.
134 (*avx2_<code><mode>3<mask_name>/maxmin): Ditto.
135 (<sse2_avx2>_andnot<mode>3<mask_name>): Ditto.
136 (*andnot<mode>3<mask_name>): Ditto.
137 (<mask_codefor><code><mode>3<mask_name>/any_logic): Ditto.
138 (<mask_codefor>avx512f_interleave_highv16si<mask_name>): Ditto.
139 (<mask_codefor>avx512f_interleave_lowv16si<mask_name>): Ditto.
140 (<mask_codefor>avx512f_vinsert<shuffletype>32x4_1<mask_name>): Ditto.
141 (vec_set_lo_<mode><mask_name>): Ditto.
142 (vec_set_hi_<mode><mask_name>): Ditto.
143 (avx512f_shuf_<shuffletype>64x2_1<mask_name>): Ditto.
144 (avx512f_shuf_<shuffletype>32x4_1<mask_name>): Ditto.
145 (avx512f_pshufd_1<mask_name>): Ditto.
146 (<mask_codefor>abs<mode>2<mask_name>): Ditto.
147 (<mask_codefor>avx512f_<code>v16qiv16si2<mask_name>): Ditto.
148 (avx512f_<code>v16hiv16si2<mask_name>/any_extend): Ditto.
149 (avx512f_<code>v8qiv8di2<mask_name>/any_extend): Ditto.
150 (avx512f_<code>v8hiv8di2<mask_name>/any_extend): Ditto.
151 (avx512f_<code>v8siv8di2<mask_name>/any_extend): Ditto.
152 (avx512er_exp2<mode><mask_name>): Ditto.
153 (<mask_codefor>avx512er_rcp28<mode><mask_name>): Ditto.
154 (<mask_codefor>avx512er_rsqrt28<mode><mask_name>): Ditto.
155 (<avx2_avx512f>_permvar<mode><mask_name>): Ditto.
156 (<avx2_avx512f>_perm<mode>_1<mask_name>): Ditto.
157 (<mask_codefor>avx512f_vec_dup<mode><mask_name>): Ditto.
158 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V16FI): Ditto.
159 (<mask_codefor>avx512f_broadcast<mode><mask_name>/V8FI): Ditto.
160 (<mask_codefor>avx512f_vec_dup_gpr<mode><mask_name>): Ditto.
161 (<mask_codefor>avx512f_vec_dup_mem<mode><mask_name>): Ditto.
162 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF2): Ditto.
163 (<sse2_avx_avx512f>_vpermil<mode><mask_name>/VF1): Ditto.
164 (*<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto.
165 (<sse2_avx_avx512f>_vpermilvar<mode>3<mask_name>): Ditto.
166 (<avx2_avx512f>_ashrv<mode><mask_name>): Ditto.
167 (<avx2_avx512f>_<shift_insn>v<mode><mask_name>): Ditto.
168 (<mask_codefor>avx512f_vcvtph2ps512<mask_name>): Ditto.
169 (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Ditto.
170 (avx512f_getmant<mode><mask_name>): Ditto.
171 (clz<mode>2<mask_name>): Ditto.
172 (<mask_codefor>conflict<mode><mask_name>): Ditto.
173 (*srcp14<mode>): Remove visibility.
174 (*rsqrt14<mode>): Ditto.
175 (*fma_fmsub_<mode>): Ditto.
176 (*fma_fnmadd_<mode>): Ditto.
177 (*avx512f_rndscale<mode>): Ditto.
178 * config/i386/subst.md: New file.
180 2013-11-13 Joseph Myers <joseph@codesourcery.com>
182 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
183 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
184 (atomic_load_explicit, atomic_exchange_explicit)
185 (atomic_compare_exchange_strong_explicit)
186 (atomic_compare_exchange_weak_explicit): Use __auto_type to
187 declare variable initialized with PTR argument.
189 2013-11-12 Jeff Law <law@redhat.com>
191 * tree-ssa-threadedge.c (thread_around_empty_blocks): New
192 argument backedge_seen_p. Set, use and pass it to children
194 (thread_through_normal_block): Similarly.
195 (thread_across_edge): Similarly.
197 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
198 memory references as volatile.
199 (insert_trap_and_remove_trailing_statements): Fix comment.
201 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
204 * ira-costs.c (record_operand_costs): Check operands number for
207 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
210 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
211 constraints that would allow DImode into the traditional Altivec
212 registers, but cause undesirable code generation when loading 0 as
214 (movdi_internal64): Likewise.
215 (cmp<mode>_fpr): Do not use %x for CR register output.
216 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
217 -mallow-upper-sf debug switches are used.
219 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
221 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
222 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
223 Relocate prototypes from gimple.h.
224 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
225 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
227 * gimple.h: Move some prototypes to gimplify.h.
228 (gimple_predicate, enum fallback, enum gimplify_status): Move to
230 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
231 Relocate from gimpify.c.
232 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
233 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
234 is_gimple_reg_rhs) Relocate from gimplify.c.
235 * gimplify.c (mark_addressable): Move to gimple-expr.c.
236 (gimple_seq_add_stmt_without_update): Move to gimple.c.
237 (remove_suffix, tmp_var_id_num, create_tmp_var_name, create_tmp_var_raw,
238 create_tmp_var, create_tmp_reg, is_gimple_reg_rhs): Move to
240 (should_carry_location_p): Move to gimple.c.
241 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
243 (annotate_one_with_location, annotate_all_with_location_after,
244 annotate_all_with_location): Move to gimple.c.
245 (compare_case_labels, sort_case_labels,
246 preprocess_case_label_vec_for_gimple): Move to gimple.c.
247 (rhs_predicate_for): Make static.
248 (gimplify_assign): Relocate from gimple.c.
249 * gimple.c (gimplify_assign): Move to gimplify.c.
250 (gimple_seq_add_stmt_without_update, should_carry_location_p,
251 annotate_one_with_location, annotate_all_with_location_after,
252 annotate_all_with_location, compare_case_labels, sort_case_labels,
253 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
254 * tree.h (unshare_expr, unshare_expr_without_location,
255 mark_addressable): Move prototypes to gimplify.h.
256 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
258 * asan.c: Include gimplify.h rather than gimple.h.
259 * cfgloopmanip.c: Likewise.
260 * cgraphunit.c: Likewise.
261 * cilk-common.c: Likewise.
262 * dwarf2out.c: Dont include gimple.h.
263 * fold-const.c: Include gimplify.h rather than gimple.h.
264 * function.c: Likewise.
265 * gimple-fold.c: Likewise.
266 * gimple-ssa-strength-reduction.c: Likewise.
267 * graphite-clast-to-gimple.c: Likewise.
268 * graphite-sese-to-poly.c: Likewise.
269 * ipa-prop.c: Likewise.
270 * ipa-split.c: Likewise.
272 * langhooks.c: Dont include gimple.h.
273 * loop-init.c: Include gimplify.h rather than gimple.h.
274 * omp-low.c: Likewise.
276 * stor-layout.c: Likewise.
277 * targhooks.c: Likewise.
278 * trans-mem.c: Likewise.
279 * tree-affine.c: Likewise.
280 * tree-cfg.c: Likewise.
281 * tree-cfgcleanup.c: Likewise.
282 * tree-complex.c: Likewise.
283 * tree-if-conv.c: Likewise.
284 * tree-inline.c: Likewise.
285 * tree-iterator.c: Likewise.
286 * tree-loop-distribution.c: Likewise.
287 * tree-nested.c: Likewise.
288 * tree-parloops.c: Likewise.
289 * tree-predcom.c: Likewise.
290 * tree-profile.c: Likewise.
291 * tree-scalar-evolution.c: Likewise.
292 * tree-sra.c: Likewise.
293 * tree-ssa-address.c: Likewise.
294 * tree-ssa-ccp.c: Likewise.
295 * tree-ssa-dce.c: Likewise.
296 * tree-ssa-forwprop.c: Likewise.
297 * tree-ssa-ifcombine.c: Likewise.
298 * tree-ssa-loop-im.c: Likewise.
299 * tree-ssa-loop-ivopts.c: Likewise.
300 * tree-ssa-loop-manip.c: Likewise.
301 * tree-ssa-loop-niter.c: Likewise.
302 * tree-ssa-loop-prefetch.c: Likewise.
303 * tree-ssa-loop-unswitch.c: Likewise.
304 * tree-ssa-math-opts.c: Likewise.
305 * tree-ssa-phiopt.c: Likewise.
306 * tree-ssa-phiprop.c: Likewise.
307 * tree-ssa-pre.c: Likewise.
308 * tree-ssa-propagate.c: Likewise.
309 * tree-ssa-reassoc.c: Likewise.
310 * tree-ssa-sccvn.c: Likewise.
311 * tree-ssa-strlen.c: Likewise.
312 * tree-ssa.c: Likewise.
313 * tree-switch-conversio: Likewise.n.c
314 * tree-tailcall.c: Likewise.
315 * tree-vect-data-refs.c: Likewise.
316 * tree-vect-generic.c: Likewise.
317 * tree-vect-loop-manip.c: Likewise.
318 * tree-vect-loop.c: Likewise.
319 * tree-vect-patterns.c: Likewise.
320 * tree-vect-stmts.c: Likewise.
322 * value-prof.c: Likewise.
323 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
324 * config/alpha/alpha.c: Likewise.
325 * config/darwin.c: Likewise.
326 * config/i386/i386.c: Likewise.
327 * config/ia64/ia64.c: Likewise.
328 * config/mep/mep.c: Likewise.
329 * config/mips/mips.c: Likewise.
330 * config/rs6000/rs6000.c: Likewise.
331 * config/s390/s390.c: Likewise.
332 * config/sh/sh.c: Likewise.
333 * config/sparc/sparc.c: Likewise.
334 * config/spu/spu.c: Likewise.
335 * config/stormy16/stormy16.c: Likewise.
336 * config/tilegx/tilegx.c: Likewise.
337 * config/tilepro/tilepro.c: Likewise.
338 * config/xtensa/xtensa.c: Likewise.
340 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
342 * tree.c (grow_tree_vec_stat): New function ...
343 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
346 2013-11-12 Marek Polacek <polacek@redhat.com>
348 * final.c (update_alignments): Initialize label to NULL_RTX.
350 2013-11-12 Jeff Law <law@redhat.com>
352 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
353 (insert_trap_and_remove_trailing_statements): New argument OP which
354 is the NULL pointer. Emit the trap after the load/store through
355 the NULL pointer. Simplify the RHS of a store through a NULL pointer
356 when trivial to do so.
357 (isolate_path): Corresponding changes.
358 (gimple_ssa_isolate_erroneous_path): Likewise.
360 2013-11-12 Teresa Johnson <tejohnson@google.com>
361 Jan Hubicka <jh@suse.cz>
363 * predict.c (drop_profile): New function.
364 (handle_missing_profiles): Ditto.
365 (counts_to_freqs): Don't overwrite estimated frequencies
366 when function has no profile counts.
367 * predict.h (handle_missing_profiles): Declare.
368 * tree-inline.c (freqs_to_counts): New function.
369 (copy_cfg_body): Invoke freqs_to_counts as needed.
370 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
372 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
375 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
377 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
379 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
382 * config/i386/i386.c (ix86_option_override_internal): Check
383 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
384 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
385 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
386 MASK_AVX256_SPLIT_UNALIGNED_STORE.
388 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
389 Clear m_COREI7_AVX and update comments.
390 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
392 2013-11-12 Martin Jambor <mjambor@suse.cz>
394 PR rtl-optimization/10474
395 * ira.c (interesting_dest_for_shprep): New function.
396 (split_live_ranges_for_shrink_wrap): Likewise.
397 (find_moveable_pseudos): Move calculation of dominance info,
398 df_analysios and the final anlyses to...
399 (ira): ...here, call split_live_ranges_for_shrink_wrap.
401 2013-11-12 Bin Cheng <bin.cheng@arm.com>
403 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
404 Handle type conversion.
406 2013-11-11 Martin Liska <marxin.liska@gmail.com>
407 Jan Hubicka <jh@suse.cz>
409 * cgraph.c (dump_cgraph_node): Profile dump added.
410 * cgraph.h (struct cgraph_node): New time profile variable added.
411 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
412 * gcov-io.h (gcov_type): New profiler type introduced.
413 * ipa-profile.c (lto_output_node): Streaming for time profile added.
414 (input_node): Time profiler is read from LTO stream.
415 * predict.c (maybe_hot_count_p): Hot prediction changed.
416 * profile.c (instrument_values): New case for time profiler added.
417 (compute_value_histograms): Read of time profile.
418 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
419 * tree-profile.c (init_ic_make_global_vars): Time profiler function added.
420 (gimple_init_edge_profiler): TP function instrumentation.
421 (gimple_gen_time_profiler): New.
422 * value-prof.c (gimple_add_histogram_value): Support for time profiler
424 (dump_histogram_value): TP type added to dumps.
425 (visit_hist): More sensitive check that takes TP into account.
426 (gimple_find_values_to_profile): TP instrumentation.
427 * value-prof.h (hist_type): New histogram type added.
428 (struct histogram_value_t): Pointer to struct function added.
429 * libgcc/Makefile.in: New GCOV merge function for TP added.
430 * libgcov.c: function_counter variable introduced.
431 (_gcov_merge_time_profile): New.
432 (_gcov_time_profiler): New.
434 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
435 Jeff Law <law@redhat.com>
437 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
438 ao_ref_init_from_ptr_and_size for builtins.
440 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
441 H.J. Lu <hongjiu.lu@intel.com>
444 * config/i386/x86-tune.def
445 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
446 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
448 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
449 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
450 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
451 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
452 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
453 misaligned_prologue_used. Check that
454 desired_aling <= epilogue_size_needed.
456 2013-11-11 Cong Hou <congh@google.com>
458 PR tree-optimization/59050
459 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
461 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
464 * expmed.c (emit_store_flag): Fail for const-const comparison.
466 2013-11-11 Tristan Gingold <gingold@adacore.com>
467 Eric Botcazou <ebotcazou@adacore.com>
469 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
470 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
471 * tree.def (CONSTRUCTOR): Likewise.
472 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
473 * gimplify.c (gimplify_init_constructor): Do not clear the object when
474 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
476 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
478 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
481 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
483 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
486 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
488 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
490 2013-11-08 Jeff Law <law@redhat.com>
492 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
493 threading paths first, then perform PHI node checks if applicable.
495 2013-11-10 Karlson2k <k2k@narod.ru>
496 Kai Tietz <ktietz@redhat.com>
499 * configure.ac: Adding for exported symbols check
500 and for rdynamic-check executable-extension.
501 * configure: Regenerated.
503 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
505 * mode-switching.c (optimize_mode_switching): Mark block as
506 nontransparent, if last_mode at block exit is different from no_mode.
508 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
510 * function.c (NAME__MAIN): Move to...
511 * cfgexpand.c (NAME__MAIN): ...here.
513 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
515 * target.def (can_use_doloop_p): New hook.
516 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
517 * doc/tm.texi: Regenerate.
518 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
519 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
520 * hooks.c (hook_bool_dint_dint_uint_true): New function.
521 * targhooks.h (can_use_doloop_if_innermost): Declare.
522 * targhooks.c (can_use_doloop_if_innermost): New function.
523 * target.h: Include double-int.h.
524 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
525 Remove iteration count, maximum iteration count, loop depth and
526 enter-at-top inputs from doloop_begin and doloop_end.
527 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
529 * config/arc/arc.c (arc_can_use_doloop_p): New function.
530 (TARGET_CAN_USE_DOLOOP_P): Define.
531 * config/arm/thumb2.md (doloop_end): Update for new interface.
532 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
533 * config/bfin/bfin.md (doloop_end): Update for new interface.
534 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
535 (TARGET_CAN_USE_DOLOOP_P): Define.
536 * config/c6x/c6x.md (doloop_end): Update for new interface.
537 * config/ia64/ia64.md (doloop_end): Update for new interface.
538 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
539 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
541 * config/mep/mep.c (mep_emit_doloop): Likewise.
542 (TARGET_CAN_USE_DOLOOP_P): Define.
543 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
544 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
545 * config/s390/s390.md (doloop_end): Update for new interface.
546 * config/sh/sh.md (doloop_end): Likewise.
547 * config/spu/spu.md (doloop_end): Likewise.
548 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
549 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
550 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
551 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
552 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
553 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
555 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
557 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
560 * doc/extend.texi: Move Cilk Plus Builtins node before Other
563 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
564 Joseph Myers <joseph@codesourcery.com>
566 * ginclude/stdatomic.h: New file.
567 * Makefile.in (USER_H): Add stdatomic.h.
569 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
571 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
574 2013-11-08 Jeff Law <law@redhat.com>
576 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
577 * tree-ssa-threadupdate.c (delete_thread_path): New function.
578 (ssa_redirect_edges, thread_block_1): Use it.
579 (thread_through_loop_header, mark_threaded_blocks): Likewise.
580 (thread_through_all_blocks, register_jump_thread): Likewise.
581 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
583 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
585 * config/arm/aarch-common.c
586 (search_term): New typedef.
587 (shift_rtx_costs): New array.
588 (arm_rtx_shift_left_p): New.
589 (arm_find_sub_rtx_with_search_term): Likewise.
590 (arm_find_sub_rtx_with_code): Likewise.
591 (arm_early_load_addr_dep): Add sanity checking.
592 (arm_no_early_alu_shift_dep): Likewise.
593 (arm_no_early_alu_shift_value_dep): Likewise.
594 (arm_no_early_mul_dep): Likewise.
595 (arm_no_early_store_addr_dep): Likewise.
597 2013-11-08 Richard Biener <rguenther@suse.de>
599 PR tree-optimization/59047
600 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
602 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
604 * common.opt (fcheck-pointer-bounds): Move to ...
605 * c-family/c.opt: ... here.
606 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
607 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
608 * langhooks.h (lang_hooks): Remove chkp_supported field.
609 * toplev.c (process_options): Remove chkp_supported check.
611 2013-11-08 Richard Biener <rguenther@suse.de>
613 PR tree-optimization/59038
614 PR tree-optimization/58955
615 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
616 previous change. Handle known dependences correctly.
618 2013-11-08 Tom de Vries <tom@codesourcery.com>
620 * config/rs6000/t-xilinx: Remove duplicate contents.
622 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
623 Joseph Myers <joseph@codesourcery.com>
625 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
626 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
627 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
628 (struct tree_base): Add atomic_flag field.
629 * tree.h (TYPE_ATOMIC): New accessor macro.
630 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
631 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
632 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
633 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
634 * tree.c (set_type_quals): Set TYPE_ATOMIC.
635 (find_atomic_core_type): New function.
636 (build_qualified_type): Adjust alignment for qualified types.
637 (build_atomic_base): New function
638 (build_common_tree_nodes): Build atomicQI_type_node,
639 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
641 * print-tree.c (print_node): Print atomic qualifier.
642 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
643 * target.def (atomic_assign_expand_fenv): New hook.
644 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
645 * doc/tm.texi: Regenerate.
646 * targhooks.c (default_atomic_assign_expand_fenv): New function.
647 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
648 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
649 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
651 * config/i386/i386.c (enum ix86_builtins): Add
652 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
654 (bdesc_special_args): Add __builtin_ia32_fnstenv,
655 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
656 (ix86_expand_builtin): Handle the new built-in functions.
657 (ix86_atomic_assign_expand_fenv): New function.
658 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
659 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
660 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
661 (fnstenv, fldenv, fnstsw, fnclex): New insns.
663 2013-11-07 Steve Ellcey <sellcey@mips.com>
665 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
666 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
667 (MULTILIB_DIRNAMES): Add fp64 directory.
668 (MULTILIB_EXCEPTIONS): Add new exclusions.
670 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
672 * gimplify.c (gimple_regimplify_operands): Do not set
674 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
675 (rewrite_close_phi_out_of_ssa): Same.
676 (rewrite_phi_out_of_ssa): Same.
677 (rewrite_degenerate_phi): Same.
678 (handle_scalar_deps_crossing_scop_limits): Same.
679 * tree-if-conv.c (predicate_scalar_phi): Same.
680 * tree-parloops.c (create_loads_for_reductions): Same.
681 (create_final_loads_for_reduction): Same.
682 (create_loads_and_stores_for_name): Same.
683 (transform_to_exit_first_loop): Same.
684 (create_parallel_loop): Same.
686 (move_computations_dom_walker::before_dom_children): Same.
687 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
688 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
689 * tree-ssa-propagate.c (substitute_and_fold): Same.
690 * tree-vect-loop.c (vect_finalize_reduction): Same.
691 * tree-vect-stmts.c (vectorizable_call): Same.
693 2013-11-07 Mike Stump <mikestump@comcast.net>
695 * config/pdp11/pdp11.c: Include dbxout.h.
696 * config/picochip/picochip.c: Likewise.
698 2013-11-07 Cong Hou <congh@google.com>
700 PR tree-optimization/56764
701 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
702 Combine alias checks if it is possible to amortize the runtime
703 overhead. Return the number of alias checks after merging.
704 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
705 Use the function vect_create_cond_for_alias_checks () to check
706 the number of alias checks.
708 2013-11-07 Jeff Law <law@redhat.com>
710 * varpool.c (ctor_for_folding): Fix typo in comment.
712 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
714 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
715 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
717 2013-11-07 Jeff Law <law@redhat.com>
719 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
721 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
722 No longer check if we have __builtin_trap, assume it's available.
724 2013-11-07 Diego Novillo <dnovillo@google.com>
726 * attribs.c (lookup_scoped_attribute_spec): Make static.
727 (get_attribute_namespace): Likewise.
728 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
729 (validate_arglist): Move earlier in the file. Make static.
730 (expand_stack_restore): Move from stmt.c
731 (expand_stack_save): Move from stmt.c
732 (rewrite_call_expr_array): Move earlier in the file.
733 (rewrite_call_expr_valist): Likewise.
734 * cfgexpand.c: Include hard-reg-set.h before tree.h
737 (expand_asm_loc): Move from stmt.c.
738 (n_occurrences): Move from stmt.c.
739 (check_operand_nalternatives): Move from stmt.c.
740 (tree_conflicts_with_clobbers_p): Move from stmt.c.
741 (expand_asm_operands): Move from stmt.c
742 (expand_asm_stmt): Move from stmt.c
743 (expand_computed_goto): Move from stmt.c
744 (expand_goto): Move from stmt.c
745 (expand_null_return_1): Move from stmt.c
746 (expand_null_return): Move from stmt.c
747 (expand_value_return): Move from stmt.c
748 (expand_return): Move from stmt.c
749 (expand_main_function): Move from function.c
750 (stack_protect_prologue): Move from function.c
751 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
752 (build_function_decl_skip_args): Move from tree.c.
753 * explow.c (tree_expr_size): Move from tree.c.
754 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
755 (fields_length): Move from tree.c.
756 * fold-const.c (size_low_cst): Move from tree.c.
757 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
758 (tree_expr_nonzero_p): Make static. Move earlier in the file.
759 (fold_build3_initializer_loc): Remove.
760 (tree_invalid_nonnegative_warnv_p): Make static.
761 * function.c (expand_main_function): Move to cfgexpand.c.
762 (stack_protect_prologue): Move to cfgexpand.c.
763 (set_insn_locations): Move earlier in the file.
764 * gimple-fold.c: Include langhooks.h.
765 (truth_type_for): Move from tree.c.
766 * print-tree.c (print_vec_tree): Remove.
767 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
768 (expand_goto): Move to cfgexpand.c.
769 (n_occurrences): Move to cfgexpand.c.
770 (expand_asm_loc): Move to cfgexpand.c
771 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
772 (expand_asm_operands): Move to cfgexpand.c.
773 (expand_asm_stmt): Move to cfgexpand.c.
774 (check_operand_nalternatives): Move to cfgexpand.c
775 (expand_null_return): Move to cfgexpand.c.
776 (expand_value_return): Move to cfgexpand.c.
777 (expand_null_return_1): Move to cfgexpand.c.
778 (expand_return): Move to cfgexpand.c.
779 (expand_stack_save): Move to builtins.c.
780 (expand_stack_restore): Move to builtins.c
781 * symtab.c: Include output.h.
782 (decl_assembler_name_hash): Move from tree.c.
783 (decl_assembler_name_equal): Move from tree.c.
784 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
785 * tree-eh.c (in_array_bounds_p): Move from tree.c.
786 (range_in_array_bounds_p): Move from tree.c.
787 * tree-object-size.c (fini_object_sizes): Make static.
788 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
789 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
790 * tree.c (decl_assembler_name_equal): Move to symtab.c.
791 (tree_expr_size): Move to explow.c.
792 (decl_assembler_name_hash): Move to symtab.c.
794 (tree_expr_size): Move to explow.c.
795 (stabilize_reference_1): Move earlier in the file. Make static.
796 (omp_remove_redundant_declare_simd_attrs): Remove.
797 (simple_cst_list_equal): Move earlier in the file. Make static.
798 (size_low_cst): Move to fold-const.c.
799 (build_type_no_quals): Remove.
800 (build_function_type_skip_args): Move to cgraphclones.c.
801 (build_function_decl_skip_args): Move to cgraphclones.c.
802 (in_array_bounds_p): Move to tree-eh.c.
803 (range_in_array_bounds_p): Move to tree-eh.c.
804 (truth_type_for): Move to gimple-fold.c.
805 (list_equal_p): Remove.
806 * tree.h (decl_assembler_name_equal): Remove.
807 (decl_assembler_name_hash): Remove.
808 (truth_type_for): Remove.
809 (build_type_no_quals): Remove.
810 (build_function_decl_skip_args): Remove.
811 (in_array_bounds_p): Remove.
812 (range_in_array_bounds_p): Remove.
813 (size_low_cst): Remove.
814 (omp_remove_redundant_declare_simd_attrs): Remove.
815 (tree_expr_size): Remove.
816 (fields_length): Remove.
817 (stabilize_reference_1): Remove.
818 (expand_goto): Remove.
819 (expand_stack_save): Remove.
820 (expand_stack_restore): Remove.
821 (expand_return): Remove.
822 (fold_build3_initializer_loc): Remove.
823 (tree_expr_nonzero_p): Remove.
824 (tree_invalid_nonnegative_warnv_p): Remove.
825 (tree_expr_nonzero_warnv_p): Remove.
826 (fold_builtin_snprintf_chk): Remove.
827 (validate_arglist): Remove.
828 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
829 (simple_cst_list_equal): Remove.
831 (expand_main_function): Remove.
832 (stack_protect_prologue): Remove.
833 (print_vec_tree): Remove.
834 (lookup_scoped_attribute_spec): Remove.
835 (get_attribute_namespace): Remove.
836 (expand_computed_goto): Remove.
837 (expand_asm_stmt): Remove.
838 (list_equal_p): Remove.
839 (ssa_name_nonnegative_p): Remove.
840 (fini_object_sizes): Remove.
841 (addr_expr_of_non_mem_decl_p): Remove.
842 (is_tm_safe_or_pure): Move to trans-mem.c.
843 (more_const_call_expr_args_p): Remove.
844 (save_vtable_map_decl): Remove.
846 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
848 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
851 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
852 link to libgomp and its dependencies.
853 * config/ia64/hpux.h (LIB_SPEC): Likewise.
854 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
855 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
856 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
858 2013-11-07 Jakub Jelinek <jakub@redhat.com>
860 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
861 (determine_value_range): Add loop argument. Use get_range_info to
863 (bound_difference): Adjust caller.
865 2013-11-07 Richard Biener <rguenther@suse.de>
866 Jakub Jelinek <jakub@redhat.com>
868 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
869 latches from header PHI arguments from the latch edge.
871 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
874 * varasm.c (output_constant): Handle NULLPTR_TYPE.
876 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
878 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
879 misaligned_prologue_used when it has been set.
881 2013-11-07 Yury Gribov <y.gribov@samsung.com>
882 Jakub Jelinek <jakub@redhat.com>
885 * asan.c (get_mem_refs_of_builtin_call): Allow
886 integer literals as addresses in instrumented builtins.
888 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
890 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
891 Explain why plus_constant is not used.
893 2013-11-07 Richard Biener <rguenther@suse.de>
895 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
896 (canonicalize_value): ... this. Also handle stripping of
898 (get_value, set_lattice_value, get_value_for_expr): Adjust.
899 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
900 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
902 2013-11-07 Richard Biener <rguenther@suse.de>
904 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
906 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
909 * config/i386/i386.md (push peepholer/splitter): Use Pmode
910 with stack_pointer_rtx.
912 2013-11-07 Bin Cheng <bin.cheng@arm.com>
914 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
915 using operand_equal_p.
917 2013-11-07 Bin Cheng <bin.cheng@arm.com>
919 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
920 * tree-affine.c (get_inner_reference_aff): Return base.
921 * tree-affine.h (get_inner_reference_aff): Change prototype.
923 2013-11-06 Tobias Burnus <burnus@net-b.de>
925 * doc/invoke.texi (Wdate-time): Fix typo.
927 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
929 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
930 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
931 fp_arith_reg_operand instead of arith_reg_operand.
933 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
935 * config/sh/sh.md (adddi3): Remove empty constraints.
936 Remove can_create_pseudo_p and arith_reg_operand check.
937 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
940 2013-11-06 Jeff Law <law@redhat.com>
941 Tom Tromey <tromey@redhat.com>
943 * gdbinit.in: Disable strict type checking.
945 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
947 * tree-pass.h (make_pass_live_range_shrinkage): New external.
948 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
949 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
950 (rest_of_handle_live_range_shrinkage): Ditto
951 (class pass_live_range_shrinkage): Ditto.
952 (pass_data_live_range_shrinkage): Ditto.
953 (make_pass_live_range_shrinkage): Ditto.
954 * sched-int.h (initialize_live_range_shrinkage): New prototype.
955 (finish_live_range_shrinkage): Ditto.
956 * sched-deps.c (create_insn_reg_set): Make void return value.
957 * passes.def: Add pass_live_range_shrinkage.
958 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
959 * haifa-sched.c (live_range_shrinkage_p): New.
960 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
962 (rank_for_schedule): Add code for pressure relief through live
964 (schedule_insn): Print more debug info.
965 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
966 through live range shrinkage.
967 * doc/invoke.texi (-flive-range-shrinkage): New.
968 * common.opt (flive-range-shrinkage): New.
970 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
973 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
974 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
975 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
976 RTXes that return in AVX256 register.
978 2013-11-06 Richard Biener <rguenther@suse.de>
980 PR tree-optimization/58653
981 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
982 (prepare_initializers_chain): Adjust.
984 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
986 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
987 (gimple_in_transaction): Use bb_in_transaction.
988 * basic-block.h (bb_in_transaction): Relocate here and rename.
989 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
991 2013-11-06 Richard Biener <rguenther@suse.de>
993 * tree.c (drop_tree_overflow): New function.
994 * tree.h (drop_tree_overflow): Declare.
995 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
996 * tree-vrp.c (range_int_cst_singleton_p): Use
997 is_overflow_infinity instead of testing TREE_OVERFLOW.
998 (extract_range_from_assert): Likewise.
999 (zero_nonzero_bits_from_vr): Likewise.
1000 (extract_range_basic): Likewise.
1001 (register_new_assert_for): Use drop_tree_overflow.
1002 (vrp_visit_phi_node): Likewise.
1004 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
1006 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
1007 checking for leaf functions without dynamic stack allocation.
1008 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
1009 (ia64_expand_prologue): Likewise.
1010 * config/mips/mips.c (mips_expand_prologue): Likewise.
1011 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
1012 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
1013 (sparc_flat_expand_prologue): Likewise.
1015 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
1017 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
1018 (__ST3_LANE_FUNC): Likewise.
1019 (__ST4_LANE_FUNC): Likewise.
1021 2013-11-06 Nick Clifton <nickc@redhat.com>
1023 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
1024 name returned by msp430_mcu_name.
1025 (LIB_SPEC): If a -T option has not been specified then set a
1026 default, mcu-specific, linker script.
1027 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
1028 * config/msp430/msp430.c (msp430x_names): Likewise.
1029 Alpha sort the names for ease of comparison.
1030 (msp430_mcu_name): New function: Returns a string suitable for
1031 use as a C preprocessor symbol based upon the name of the MCU
1033 (msp430_option_override): Accept msp430x and msp430xv2 as generic
1035 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
1037 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
1038 unless it is the last part of a spec substring.
1039 * doc/invoke.texi (Spec Files): Document space insertion
1042 2013-11-06 Christian Bruel <christian.bruel@st.com>
1044 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
1045 Factorize probabilities, Use adjust_address instead of
1046 adjust_automodify_address when possible. Enable for optimize.
1047 (sh_expand_strlen): New function.
1048 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
1049 * config/sh/sh.md (strlensi): New pattern.
1050 (UNSPEC_BUILTIN_STRLEN): Define.
1052 2013-11-06 Jakub Jelinek <jakub@redhat.com>
1055 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
1056 (expand_assignment): If *bitpos is negative, set *offset
1057 and adjust *bitpos, so that it is not negative.
1059 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
1061 * config/i386/bdver3.md : Added two additional decoder units
1062 to support issue rate of 4 and remodeled vector unit.
1063 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
1064 architectures is set to 4.
1065 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
1066 lookahead is set to 4 for BD architectures.
1068 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1070 * config/rs6000/rs6000.c (rs6000_option_override_internal):
1071 Remove restriction against use of VSX instructions when generating
1072 code for little endian mode.
1074 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1076 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
1077 for both big and little endian.
1078 (mulv8hi3): Swap input operands for merge high and merge low
1079 instructions for little endian.
1081 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1083 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
1084 define_insn to define_expand that uses even patterns for big
1085 endian and odd patterns for little endian.
1086 (vec_widen_smult_even_v16qi): Likewise.
1087 (vec_widen_umult_even_v8hi): Likewise.
1088 (vec_widen_smult_even_v8hi): Likewise.
1089 (vec_widen_umult_odd_v16qi): Likewise.
1090 (vec_widen_smult_odd_v16qi): Likewise.
1091 (vec_widen_umult_odd_v8hi): Likewise.
1092 (vec_widen_smult_odd_v8hi): Likewise.
1093 (altivec_vmuleub): New define_insn.
1094 (altivec_vmuloub): Likewise.
1095 (altivec_vmulesb): Likewise.
1096 (altivec_vmulosb): Likewise.
1097 (altivec_vmuleuh): Likewise.
1098 (altivec_vmulouh): Likewise.
1099 (altivec_vmulesh): Likewise.
1100 (altivec_vmulosh): Likewise.
1102 2013-11-05 Mike Stump <mikestump@comcast.net>
1104 * Makefile.in (mostlyclean): Remove c-family objects.
1106 2013-11-05 Ian Lance Taylor <iant@google.com>
1108 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
1109 If possible, add .cfi directives to record change to bx.
1110 * config/i386/i386.c (ix86_emit_cfi): New function.
1111 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
1113 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
1115 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
1116 immediately after a label for a tablejump pattern.
1118 * config/arm/arm.c (is_jump_table): Remove.
1119 (create_fix_barrier): Use tablejump_p instead.
1120 (arm_reorg): Likewise.
1121 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
1122 (thumb2_output_casesi): Likewise.
1123 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
1124 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
1125 casesi_shift_media, casesi_load_media): Likewise.
1126 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
1127 * config/microblaze/microblaze.md: Likewise.
1129 2013-11-05 Tobias Burnus <burnus@net-b.de>
1131 * doc/invoke.texi (-Wdate-time): Document.
1133 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
1135 * double-int.c (lshift_double, rshift_double): Remove
1136 SHIFT_COUNT_TRUNCATED handling.
1138 2013-11-05 Jeff Law <law@redhat.com>
1140 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
1141 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
1142 * gimple-ssa-isolate-paths.c: New file.
1143 * gimple.c (check_loadstore): New function.
1144 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
1145 Verify OP is in the argument list and the argument corresponding
1146 to OP is a pointer type. Use operand_equal_p rather than
1147 pointer equality when testing if OP is on the nonnull list.
1148 Use check_loadstore rather than count_ptr_derefs. Handle
1149 GIMPLE_RETURN statements.
1150 * tree-vrp.c (infer_nonnull_range): Remove.
1151 * gimple.h (infer_nonnull_range): Declare.
1152 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
1153 * passes.def: Add pass_isolate_erroneous_paths.
1154 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
1155 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
1156 * tree-ssa.c (struct count_ptr_d): Remove.
1157 (count_ptr_derefs, count_uses_and_derefs): Remove.
1158 * tree-ssa.h (count_uses_and_derefs): Remove.
1160 2013-11-05 Jakub Jelinek <jakub@redhat.com>
1162 PR rtl-optimization/58997
1163 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
1164 get_iv_value to be in iv->mode rather than iv->extend_mode.
1165 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
1166 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
1167 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
1170 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
1172 * gimple.h: Move some prototypes to gimple-expr.h and add to include
1174 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
1175 Move to gimple-expr.h.
1176 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
1177 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
1178 is_gimple_id, virtual_operand_p, is_gimple_addressable,
1179 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
1181 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
1182 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
1183 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
1184 is_gimple_address, is_gimple_invariant_address,
1185 is_gimple_ip_invariant_address, is_gimple_min_invariant,
1186 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
1187 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
1188 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
1189 gimple_decl_printable_name, useless_type_conversion_p,
1190 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
1192 * gimple-expr.c: New File.
1193 (useless_type_conversion_p, gimple_set_body, gimple_body,
1194 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
1195 gimple_can_coalesce_p, extract_ops_from_tree_1,
1196 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
1197 is_gimple_address, is_gimple_invariant_address,
1198 is_gimple_ip_invariant_address, is_gimple_min_invariant,
1199 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
1200 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
1201 is_gimple_mem_ref_addr): Relocate here.
1202 * Makefile.in (OBJS): Add gimple-expr.o.
1204 2013-11-05 David Malcolm <dmalcolm@redhat.com>
1206 * gengtype-parse.c (struct_field_seq): Support empty structs.
1208 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1210 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
1212 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1214 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
1216 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
1218 (LONG_DOUBLE_TYPE_SIZE): New define.
1219 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
1221 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
1224 * doc/extend.texi [visibility ("visibility_type")]: Add example
1225 about visibility attribute on namespace declaration.
1227 2013-11-05 Richard Biener <rguenther@suse.de>
1230 * passes.def (all_passes): Start with pass_fixup_cfg again.
1232 2013-11-05 Richard Biener <rguenther@suse.de>
1234 PR tree-optimization/58955
1235 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
1238 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1240 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
1242 (vec_pack_ufix_trunc_v2df): Likewise.
1244 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
1247 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
1248 pattern, instead of word_mode.
1250 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
1252 (set_storage_via_setmem): Likewise.
1254 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
1256 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
1257 gimple.h and the rest of the condition in eliminate_build.
1258 (eliminate_build): Call new routine.
1259 * gimple.h (phi_ssa_name_p): Delete.
1261 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
1263 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
1264 case of no prefix and reserving zero slots, because when that's the
1265 case we'll never get here.
1266 * vec.h (va_heap::reserve): Don't try and handle
1267 vec_prefix::calculate_allocation returning zero because that should
1270 2013-11-05 Richard Biener <rguenther@suse.de>
1273 * tree-dfa.c (get_ref_base_and_extent): Merge common code
1274 in MEM_REF and TARGET_MEM_REF handling. Make sure to
1275 process trailing array detection before diving into the
1276 view-converted object (and possibly apply some extra offset).
1278 2013-11-05 Joseph Myers <joseph@codesourcery.com>
1280 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
1282 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
1284 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1286 PR tree-optimization/58958
1287 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
1288 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
1290 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1292 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
1293 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
1295 2013-11-05 Jakub Jelinek <jakub@redhat.com>
1297 PR tree-optimization/58984
1298 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
1299 set *SIZE_P if non-NULL on success.
1300 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
1302 (ipcp_transform_function): Likewise. Punt if size of access
1303 is different from TYPE_SIZE on v->value's type.
1305 2013-11-05 Tobias Burnus <burnus@net-b.de>
1307 * doc/invoke.texi (-fopenmp-simd): Document new option.
1308 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
1309 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
1310 * tree.c (attribute_value_equal): Ditto.
1312 2013-11-04 Wei Mi <wmi@google.com>
1314 * sched-rgn.c (add_branch_dependences): Keep insns in
1315 a SCHED_GROUP at the end of BB to remain their location.
1317 2013-11-04 Wei Mi <wmi@google.com>
1319 * config/i386/i386.c (memory_address_length): Extract a part
1320 of code to rip_relative_addr_p.
1321 (rip_relative_addr_p): New Function.
1322 (ix86_macro_fusion_p): Ditto.
1323 (ix86_macro_fusion_pair_p): Ditto.
1324 * config/i386/i386.h: Add new tune features about macro-fusion.
1325 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1326 * doc/tm.texi: Generated.
1327 * doc/tm.texi.in: Ditto.
1328 * haifa-sched.c (try_group_insn): New Function.
1329 (group_insns_for_macro_fusion): Ditto.
1330 (sched_init): Call group_insns_for_macro_fusion.
1331 * target.def: Add two hooks: macro_fusion_p and
1332 macro_fusion_pair_p.
1334 2013-11-04 Kostya Serebryany <kcc@google.com>
1336 Update to match the changed asan API.
1337 * asan.c (asan_function_start): New function.
1338 (asan_emit_stack_protection): Update the string stored in the
1339 stack red zone to match new API. Store the PC of the current
1340 function in the red zone.
1341 (asan_global_struct): Update the __asan_global definition to match
1343 (asan_add_global): Ditto.
1344 * asan.h (asan_function_start): New prototype.
1345 * final.c (final_start_function): Call asan_function_start.
1346 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
1349 2013-11-04 Wei Mi <wmi@google.com>
1351 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
1352 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
1353 * config/i386/i386.c (ix86_option_override_internal): Ditto.
1354 (ix86_issue_rate): Ditto.
1355 (ix86_adjust_cost): Ditto.
1356 (ia32_multipass_dfa_lookahead): Ditto.
1357 (ix86_sched_init_global): Ditto.
1358 (get_builtin_code_for_version): Ditto.
1359 * config/i386/i386.h (enum target_cpu_default): Ditto.
1360 (enum processor_type): Ditto.
1361 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1363 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1365 PR rtl-optimization/58967
1366 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
1367 !lra_in_progress for mode sizes bigger word.
1369 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1371 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
1372 arguments to merge instruction for little endian.
1373 (vec_widen_umult_lo_v16qi): Likewise.
1374 (vec_widen_smult_hi_v16qi): Likewise.
1375 (vec_widen_smult_lo_v16qi): Likewise.
1376 (vec_widen_umult_hi_v8hi): Likewise.
1377 (vec_widen_umult_lo_v8hi): Likewise.
1378 (vec_widen_smult_hi_v8hi): Likewise.
1379 (vec_widen_smult_lo_v8hi): Likewise.
1381 2013-11-04 Ian Lance Taylor <iant@google.com>
1383 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
1384 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
1385 builtins that take pointers.
1386 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
1387 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
1388 (run_gcc): Pass -fnon-call-exceptions.
1390 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1392 * optabs.c (expand_vec_perm): Revert one incorrect line from
1395 PR tree-optimization/58978
1396 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
1397 use_stmt by single_imm_use directly. Only call single_imm_use
1400 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1402 PR rtl-optimization/58968
1403 * lra-spills.c (return_regno_p): New function.
1404 (lra_final_code_change): Use it.
1406 2013-11-04 Joseph Myers <joseph@codesourcery.com>
1408 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
1409 * target.def (float_exceptions_rounding_supported_p): New hook.
1410 * targhooks.c (default_float_exceptions_rounding_supported_p): New
1412 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
1413 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
1415 * doc/tm.texi: Regenerate.
1416 * config.gcc (powerpc*-*-linux*): Set extra_objs.
1417 * config/rs6000/rs6000-linux.c: New file.
1418 * config/rs6000/rs6000-protos.h
1419 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
1420 * config/rs6000/linux.h
1421 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
1422 * config/rs6000/linux64.h
1423 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
1424 * config/rs6000/t-linux (rs6000-linux.o): New rule.
1425 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
1427 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1429 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
1430 Replace the define_insn_and_split with a define_insn and two
1431 define_splits, with the split after reload re-permuting the source
1432 register to its original value.
1433 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
1434 (*vsx_le_perm_store_v8hi): Likewise.
1435 (*vsx_le_perm_store_v16qi): Likewise.
1437 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1439 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
1442 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1444 PR tree-optimization/58946
1445 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
1446 bbs with bbinfo[idx].op != NULL before all blocks with
1447 bbinfo[idx].op == NULL.
1449 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
1451 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
1452 (avr_dump_double_int_hex): Likewise.
1453 (avr_log_vadump): Remove %D and %X handling.
1454 * config/avr/avr.c (avr_double_int_push_digit): Delete.
1455 (avr_map_op_t): Change map from double_int to unsigned int.
1456 (avr_map_op): Update accordingly.
1457 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
1458 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
1459 unsigned ints rather than double_ints.
1461 2013-11-03 Marek Polacek <polacek@redhat.com>
1463 Implement -fsanitize=vla-bound.
1464 * opts.c (common_handle_option): Handle vla-bound.
1465 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
1466 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
1467 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
1469 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1471 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
1474 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
1476 * config/i386/constraints.md (Ts, Tv): New address constrains.
1477 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
1478 constraint for address_no_seg_operand.
1479 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
1480 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
1481 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
1482 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
1483 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
1484 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
1485 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
1486 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
1487 constraint for vsib_address_operand.
1489 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
1491 * gcse.c (pre_delete): Remove references to regmove from comments.
1492 * recog.c: (validate_replace_rtx_1): Likewise.
1493 * config/rl78/rl78.c: Likewise.
1494 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
1495 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
1496 * common/config/mmix/mmix-common.c: Likewise.
1498 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
1500 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
1501 * gimplify.c (gimplify_compound_lval): Likewise.
1502 * graphite-clast-to-gimple.c (gloog): Likewise.
1503 * graphite-dependences.c (loop_is_parallel_p): Likewise.
1504 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
1505 (limit_scop); Likewise.
1506 (build_scops): Likewise.
1507 (dot_scop): Likewise.
1508 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1509 (build_scop_drs): Likewise.
1510 (insert_stmts): Likewise.
1511 (insert_out_of_ssa_copy): Likewise.
1512 (remove_phi): Likewise.
1513 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
1514 * hw-doloop.c (discover_loop): Likewise.
1515 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
1516 * tree-dfa.c (dump_enumerated_decls): Likewise.
1517 * tree-if-conv.c (if_convertable_loop_p): Likewise.
1518 * tree-inline.c (tree_function_versioning): Likewise.
1519 * tree-loop-distribution.c (build_rdg): Likewise.
1520 (rdg_flag_vertex_and_dependent): Likewise.
1521 (distribute_loop): Likewise.
1522 * tree-parloops.c (loop_parallel_p): Likewise.
1523 (eliminate_local_variables): Likewise.
1524 (separate_decls_in_region): Likewise.
1525 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
1526 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
1527 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1528 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
1529 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
1530 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
1531 (vectorizable_condition): Likewise.
1533 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
1535 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
1536 * configure: Regenerate.
1537 * config/i386/i386.md (*movdi_internal): Change
1538 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
1539 (*movdf_internal): Ditto.
1540 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
1541 * config/i386/sse.md (vec_concatv2di): Output interunit movq
1542 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
1544 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
1546 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
1549 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1551 Automated part of renaming of symtab_node_base to symtab_node.
1553 Patch autogenerated by rename_symtab.py from
1554 https://github.com/davidmalcolm/gcc-refactoring-scripts
1555 revision 58bb219cc090b2f4516a9297d868c245495ee622
1556 with ChangeLog entry fixed up by hand.
1558 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
1560 (cgraph_node_for_asm): Likewise.
1561 * cgraph.h (symtab_node_base): Likewise.
1562 (cgraph_node): Likewise.
1563 (varpool_node): Likewise.
1564 (is_a_helper <cgraph_node>::test): Likewise.
1565 (is_a_helper <varpool_node>::test): Likewise.
1566 (symtab_nodes): Likewise.
1567 (symtab_register_node): Likewise.
1568 (symtab_unregister_node): Likewise.
1569 (symtab_remove_node): Likewise.
1570 (symtab_get_node): Likewise.
1571 (symtab_node_for_asm): Likewise.
1572 (symtab_node_asm_name): Likewise.
1573 (symtab_node_name): Likewise.
1574 (symtab_insert_node_to_hashtable): Likewise.
1575 (symtab_add_to_same_comdat_group): Likewise.
1576 (symtab_dissolve_same_comdat_group_list): Likewise.
1577 (dump_symtab_node): Likewise.
1578 (debug_symtab_node): Likewise.
1579 (dump_symtab_base): Likewise.
1580 (verify_symtab_node): Likewise.
1581 (verify_symtab_base): Likewise.
1582 (symtab_used_from_object_file_p): Likewise.
1583 (symtab_alias_ultimate_target): Likewise.
1584 (symtab_resolve_alias): Likewise.
1585 (fixup_same_cpp_alias_visibility): Likewise.
1586 (symtab_for_node_and_aliases): Likewise.
1587 (symtab_nonoverwritable_alias): Likewise.
1588 (availability symtab_node_availability): Likewise.
1589 (symtab_semantically_equivalent_p): Likewise.
1590 (fixup_same_cpp_alias_visibility): Likewise.
1591 (symtab_prevail_in_asm_name_hash): Likewise.
1593 (varpool): Likewise.
1594 (varpool_first_variable): Likewise.
1595 (varpool_next_variable): Likewise.
1596 (varpool_first_static_initializer): Likewise.
1597 (varpool_next_static_initializer): Likewise.
1598 (varpool_first_defined_variable): Likewise.
1599 (varpool_next_defined_variable): Likewise.
1600 (cgraph_first_defined_function): Likewise.
1601 (cgraph_next_defined_function): Likewise.
1602 (cgraph_first_function): Likewise.
1603 (cgraph_next_function): Likewise.
1604 (cgraph_first_function_with_gimple_body): Likewise.
1605 (cgraph_next_function_with_gimple_body): Likewise.
1606 (symtab_alias_target): Likewise.
1607 (symtab_real_symbol_p): Likewise.
1608 (symtab_can_be_discarded): Likewise.
1609 * cgraphbuild.c (mark_address): Likewise.
1610 (mark_load): Likewise.
1611 (mark_store): Likewise.
1612 * cgraphunit.c (decide_is_symbol_needed): Likewise.
1614 (enqueue_node): Likewise.
1615 (referred_to_p): Likewise.
1616 (cgraph_process_same_body_aliases): Likewise.
1617 (analyze_functions): Likewise.
1618 (handle_alias_pairs): Likewise.
1619 (output_weakrefs): Likewise.
1620 (compile): Likewise.
1621 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1622 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1623 * ipa-prop.c (remove_described_reference): Likewise.
1624 (try_decrement_rdesc_refcount): Likewise.
1625 (ipa_edge_duplication_hook): Likewise.
1626 * ipa-ref.c (ipa_record_reference): Likewise.
1627 (ipa_maybe_record_reference): Likewise.
1628 (ipa_clone_ref): Likewise.
1629 (ipa_clone_references): Likewise.
1630 (ipa_clone_referring): Likewise.
1631 (ipa_find_reference): Likewise.
1632 (ipa_remove_stmt_references): Likewise.
1633 (ipa_clear_stmts_in_references): Likewise.
1634 * ipa-ref.h (symtab_node_base): Likewise.
1635 (ipa_ref): Likewise.
1636 (ipa_record_reference): Likewise.
1637 (ipa_maybe_record_reference): Likewise.
1638 (ipa_clone_references): Likewise.
1639 (ipa_clone_referring): Likewise.
1640 (ipa_clone_ref): Likewise.
1641 (ipa_find_reference): Likewise.
1642 (ipa_remove_stmt_references): Likewise.
1643 (ipa_clear_stmts_in_references): Likewise.
1644 * ipa-reference.c (ipa_reference_write_optimization_summary):
1646 * ipa.c (enqueue_node): Likewise.
1647 (process_references): Likewise.
1648 (walk_polymorphic_call_targets): Likewise.
1649 (symtab_remove_unreachable_nodes): Likewise.
1650 (address_taken_from_non_vtable_p): Likewise.
1651 (comdat_can_be_unshared_p_1): Likewise.
1652 (comdat_can_be_unshared_p): Likewise.
1653 (can_replace_by_local_alias): Likewise.
1654 (function_and_variable_visibility): Likewise.
1655 * is-a.h: Likewise (within example in comment).
1656 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
1657 (lto_symtab_encoder_encode): Likewise.
1658 (lto_symtab_encoder_delete_node): Likewise.
1659 (lto_symtab_encoder_in_partition_p): Likewise.
1660 (lto_set_symtab_encoder_in_partition): Likewise.
1661 (output_refs): Likewise.
1662 (compute_ltrans_boundary): Likewise.
1663 (output_symtab): Likewise.
1664 (input_node): Likewise.
1665 (input_ref): Likewise.
1666 (input_edge): Likewise.
1667 (input_cgraph_1): Likewise.
1668 (input_refs): Likewise.
1669 (output_cgraph_opt_summary): Likewise.
1670 (input_node_opt_summary): Likewise.
1671 (input_cgraph_opt_section): Likewise.
1672 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1674 * lto-streamer-out.c (lto_output): Likewise.
1675 (output_symbol_p): Likewise.
1676 (produce_symtab): Likewise.
1677 * lto-streamer.h (lto_encoder_entry): Likewise.
1678 (lto_free_function_in_decl_state_for_node): Likewise.
1679 (lto_symtab_encoder_encode): Likewise.
1680 (lto_symtab_encoder_delete_node): Likewise.
1681 (lto_symtab_encoder_in_partition_p): Likewise.
1682 (lto_set_symtab_encoder_in_partition): Likewise.
1683 (lto_symtab_encoder_lookup): Likewise.
1684 (lsei_node): Likewise.
1685 (lto_symtab_encoder_deref): Likewise.
1686 * symtab.c (symtab_hash): Likewise.
1687 (assembler_name_hash): Likewise.
1688 (symtab_nodes): Likewise.
1689 (hash_node): Likewise.
1690 (eq_node): Likewise.
1691 (hash_node_by_assembler_name): Likewise.
1692 (eq_assembler_name): Likewise.
1693 (insert_to_assembler_name_hash): Likewise.
1694 (unlink_from_assembler_name_hash): Likewise.
1695 (symtab_prevail_in_asm_name_hash): Likewise.
1696 (symtab_register_node): Likewise.
1697 (symtab_insert_node_to_hashtable): Likewise.
1698 (symtab_unregister_node): Likewise.
1699 (symtab_get_node): Likewise.
1700 (symtab_remove_node): Likewise.
1701 (symtab_initialize_asm_name_hash): Likewise.
1702 (symtab_node_for_asm): Likewise.
1703 (symtab_add_to_same_comdat_group): Likewise.
1704 (symtab_dissolve_same_comdat_group_list): Likewise.
1705 (symtab_node_asm_name): Likewise.
1706 (symtab_node_name): Likewise.
1707 (dump_symtab_base): Likewise.
1708 (dump_symtab_node): Likewise.
1709 (dump_symtab): Likewise.
1710 (debug_symtab_node): Likewise.
1711 (verify_symtab_base): Likewise.
1712 (verify_symtab_node): Likewise.
1713 (verify_symtab): Likewise.
1714 (symtab_used_from_object_file_p): Likewise.
1715 (symtab_node_availability): Likewise.
1716 (symtab_alias_ultimate_target): Likewise.
1717 (fixup_same_cpp_alias_visibility): Likewise.
1718 (symtab_resolve_alias): Likewise.
1719 (symtab_for_node_and_aliases): Likewise.
1720 (symtab_for_node_and_aliases): Likewise.
1721 (symtab_nonoverwritable_alias_1): Likewise.
1722 (symtab_nonoverwritable_alias): Likewise.
1723 (symtab_semantically_equivalent_p): Likewise.
1724 * value-prof.c (init_node_map): Likewise.
1725 * varasm.c (find_decl): Likewise.
1726 * varpool.c (varpool_node_for_asm): Likewise.
1727 (varpool_remove_unreferenced_decls): Likewise.
1729 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1731 Manual part of renaming of symtab_node_base to symtab_node.
1733 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1734 clashes with the preferred name for the base class.
1735 (const_symtab_node): Remove redundant typedef.
1737 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1739 * optabs.c (expand_vec_perm): Avoid vector mode punning
1740 SUBREGs in SET_DEST.
1741 * expmed.c (store_bit_field_1): Likewise.
1742 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1743 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1744 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1745 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1746 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1747 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1748 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1749 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1750 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1751 expand_vec_perm_vpshufb2_vpermq,
1752 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1753 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1754 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1755 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1758 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1760 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1761 Use gcc_unreachable for unhandled modes. Do not check results of
1762 expand_simple_binop. If not expanded to target, move the result.
1764 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1765 Shiva Chen <shiva0217@gmail.com>
1767 * config.gcc (nds32*-*-*): Add nds32 target.
1768 * config/nds32/nds32.c: New file.
1769 * config/nds32/nds32.h: New file.
1770 * config/nds32/nds32.md: New file.
1771 * config/nds32/constants.md: New file.
1772 * config/nds32/constraints.md: New file.
1773 * config/nds32/iterators.md: New file.
1774 * config/nds32/nds32-doubleword.md: New file.
1775 * config/nds32/nds32-intrinsic.md: New file.
1776 * config/nds32/nds32_intrinsic.h: New file.
1777 * config/nds32/nds32-modes.def: New file.
1778 * config/nds32/nds32-multiple.md: New file.
1779 * config/nds32/nds32.opt: New file.
1780 * config/nds32/nds32-opts.h: New file.
1781 * config/nds32/nds32-protos.h: New file.
1782 * config/nds32/nds32-peephole2.md: New file.
1783 * config/nds32/pipelines.md: New file.
1784 * config/nds32/predicates.md: New file.
1785 * config/nds32/t-mlibs: New file.
1786 * common/config/nds32: New directory and files.
1788 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1789 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1790 * doc/install.texi (Cross-Compiler-Specific Options): Document
1791 --with-nds32-lib for nds32 target.
1792 * doc/extend.texi (Function Attributes, Target Builtins): Document
1793 nds32 specific attributes.
1795 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1797 * lra-constraints (process_alt_operands): Use the result
1798 elimination register for operand when matching constraints.
1800 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1802 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1803 (remove_range_assertions): Call it.
1805 * tree.c (tree_ctz): New function.
1806 * tree.h (tree_ctz): New prototype.
1807 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1808 first argument from tree to const_tree.
1809 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1810 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1811 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1813 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1814 * tree-vect-loop.c (vect_analyze_loop_operations,
1815 vect_transform_loop): Don't force scalar loop for bound just because
1816 number of iterations is unknown, only do it if it is not known to be
1817 a multiple of vectorization_factor.
1818 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1820 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1821 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1822 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1823 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1824 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1825 a default def isn't partially constant.
1826 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1827 is known to be partially zero.
1828 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1829 to see if a default def isn't partially constant.
1830 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1831 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1833 (set_nonzero_bits, get_nonzero_bits): New functions.
1835 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1836 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1837 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1838 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1839 uses but in the condition and ASSERT_EXPR and the other successor of
1840 the predecessor bb is __builtin_unreachable (), set_range_info of the
1841 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1843 2013-10-31 Martin Jambor <mjambor@suse.cz>
1845 PR rtl-optimization/58934
1847 2013-10-30 Martin Jambor <mjambor@suse.cz>
1848 PR rtl-optimization/10474
1849 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1851 (interesting_dest_for_shprep): New function.
1852 (split_live_ranges_for_shrink_wrap): Likewise.
1853 (ira): Calculate dominance info and df analysis. Call
1854 split_live_ranges_for_shrink_wrap.
1856 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1857 Yury Gribov <y.gribov@samsung.com>
1860 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1861 shadow pointer to avoid clobbering the main one.
1863 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1865 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1867 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1870 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1871 it for calling update_costs_from_allocno.
1872 (assign_hard_reg): Call restore_costs_from_copies only for
1873 !retry_p. Pass new argument to update_costs_from_copies.
1874 (color_pass): Pass new argument to update_costs_from_copies.
1875 (ira_mark_allocation_change): Ditto.
1877 2013-10-30 Sharad Singhai <singhai@google.com>
1880 * opts.c (common_handle_option): Remove deprecated option
1881 -ftree-vectorizer-verbose.
1882 * doc/invoke.texi (Debugging Options): Ditto.
1883 * opts-global.c (handle_common_deferred_options): Ditto.
1884 (dump_remap_tree_vectorizer_verbose): Delete.
1885 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1887 2013-10-30 DJ Delorie <dj@redhat.com>
1889 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1890 that take no arguments.
1892 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1895 * reload1.c (update_eliminables_and_spill): New function, broken
1897 (reload): Use it. Check for frame size change after frame size
1898 alignment, and call update_eliminables_and_spill first if continue-ing.
1900 2013-10-30 Cong Hou <congh@google.com>
1903 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1904 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1905 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1907 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1909 PR rtl-optimization/58369
1910 * reload1.c (compute_reload_subreg_offset): New function.
1911 (choose_reload_regs): Use it to pass endian-correct
1912 offset to subreg_regno_offset.
1914 2013-10-30 Tobias Burnus <burnus@net-b.de>
1917 * tree-cfg.c (replace_loop_annotate): Replace warning by
1920 2013-10-30 Jason Merrill <jason@redhat.com>
1922 * configure.ac (loose_warn): Add -Wno-format if
1923 --disable-build-format-warnings.
1925 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1927 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1928 onto multiple lines to make things easier for rename_symtab.py.
1930 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1931 (symtab_semantically_equivalent_p): Likewise.
1933 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1936 * lra.c (check_rtl): Remove address check before LRA work.
1938 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1940 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1941 POINTER_PLUS_EXPR in the defining statement.
1943 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1945 * regmove.c: Remove.
1946 * tree-pass.h (make_pass_regmove): Remove.
1947 * timevar.def (TV_REGMOVE): Remove.
1948 * passes.def (pass_regmove): Remove.
1949 * opts.c (default_options_table): Remove entry for regmove.
1950 * doc/passes.texi: Remove regmove pass description.
1951 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1953 (-fdump-rtl-regmove): Ditto.
1954 * common.opt (foptimize-register-move, fregmove): Ignore.
1955 * Makefile.in (OBJS): Remove regmove.o.
1956 * regmove.c: Remove.
1957 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1959 (struct ira_allocno) New member allocno_prefs.
1960 (ALLOCNO_PREFS): New macro.
1961 (ira_prefs, ira_prefs_num): New external vars.
1962 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1963 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1964 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1966 (ira_add_allocno_copy_to_list): Remove prototype.
1967 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1968 (ira_pref_iterator): New type.
1969 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1970 (FOR_EACH_PREF): New macro.
1971 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1972 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1974 (ira_setup_alts): New function.
1975 (decrease_live_ranges_number): New function.
1976 (ira): Call the above function.
1977 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1978 (ira_create_allocno): Initialize allocno prefs.
1979 (pref_pool, pref_vec): New static vars.
1980 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1982 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1983 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1984 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1985 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1986 (ira_add_allocno_copy_to_list): Make static. Rename to
1987 add_allocno_copy_to_list.
1988 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1989 swap_allocno_copy_ends_if_necessary.
1990 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1991 ira_remove_allocno_prefs.
1992 (ira_flattening): Ditto.
1993 (ira_build): Call initiate_prefs, print_prefs.
1994 (ira_destroy): Call finish_prefs.
1995 * ira-color.c (struct update_cost_record): New.
1996 (struct allocno_color_data): Add new member update_cost_records.
1997 (update_cost_record_pool): New static var.
1998 (init_update_cost_records, get_update_cost_record): New functions.
1999 (free_update_cost_record_list, finish_update_cost_records): Ditto.
2000 (struct update_cost_queue_elem): Add member from.
2001 (initiate_cost_update): Call init_update_cost_records.
2002 (finish_cost_update): Call finish_update_cost_records.
2003 (queue_update_cost, get_next_update_cost): Add new param from.
2004 (Update_allocno_cost, update_costs_from_allocno): New functions.
2005 (update_costs_from_prefs): Ditto.
2006 (update_copy_costs): Rename to update_costs_from_copies.
2007 (restore_costs_from_copies): New function.
2008 (update_conflict_hard_regno_costs): Don't go back.
2009 (assign_hard_reg): Call restore_costs_from_copies. Add printing
2011 (pop_allocnos): Add priniting more debug info.
2012 (color_allocnos): Remove prefs for conflicting hard regs.
2013 Call update_costs_from_prefs.
2014 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
2015 (get_dup_num): Rename, modify, and move to ira.c
2016 (process_regs_for_copy): Add prefs.
2017 (add_insn_allocno_copies): Put src as first arg of
2018 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
2019 * ira-costs.c (record_reg_classes): Modify and move code into
2020 record_operands_costs.
2021 (find_costs_and_classes): Create prefs for the hard reg of small
2023 (process_bb_node_for_hard_reg_moves): Add prefs.
2025 2013-10-30 Richard Biener <rguenther@suse.de>
2028 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
2029 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
2031 (pre_and_rev_post_order_compute): ... which now wraps it.
2032 * graph.c (draw_cfg_nodes_no_loops): Use
2033 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
2036 2013-10-30 Christian Bruel <christian.bruel@st.com>
2038 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
2039 (sh_expand_cmpstr): Handle known align and schedule improvements.
2040 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
2041 * config/sh/sh.md (cmpstrnsi): New pattern.
2043 2013-10-30 Martin Jambor <mjambor@suse.cz>
2045 PR rtl-optimization/10474
2046 * ira.c (find_moveable_pseudos): Do not calculate dominance info
2048 (interesting_dest_for_shprep): New function.
2049 (split_live_ranges_for_shrink_wrap): Likewise.
2050 (ira): Calculate dominance info and df analysis. Call
2051 split_live_ranges_for_shrink_wrap.
2053 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2056 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
2058 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2060 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
2061 * tree.h (POINTER_BOUNDS_P): New.
2062 (BOUNDED_TYPE_P): New.
2064 (pointer_bounds_type_node): New.
2065 * tree.c (build_common_tree_nodes): Initialize
2066 pointer_bounds_type_node.
2067 * gimple.h (gimple_call_get_nobnd_arg_index): New.
2068 (gimple_call_num_nobnd_args): New.
2069 (gimple_call_nobnd_arg): New.
2070 (gimple_return_retbnd): New.
2071 (gimple_return_set_retbnd): New
2072 * gimple.c (gimple_build_return): Increase number of ops
2073 for return statement.
2074 (gimple_call_get_nobnd_arg_index): New.
2075 * gimple-pretty-print.c (dump_gimple_return): Print second op.
2077 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2079 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
2080 with "chkp ctor" and "bnd_legacy" attributes.
2081 * gimplify.c (gimplify_init_constructor): Avoid infinite
2082 loop during gimplification of bounds initializer.
2084 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2086 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
2087 (handle_bnd_legacy): New.
2088 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
2089 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
2092 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
2094 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
2095 (BT_FN_PTR_CONST_PTR): New.
2096 (BT_FN_CONST_PTR_CONST_PTR): New.
2097 (BT_FN_PTR_CONST_PTR_SIZE): New.
2098 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
2099 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
2100 (BT_FN_VOID_CONST_PTR_SIZE): New.
2101 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
2102 * chkp-builtins.def: New.
2103 * builtins.def: include chkp-builtins.def.
2104 (DEF_CHKP_BUILTIN): New.
2105 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
2106 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
2107 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
2108 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
2109 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
2110 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
2111 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
2112 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
2113 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
2114 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
2115 * common.opt (fcheck-pointer-bounds): New.
2116 * toplev.c (process_options): Check Pointer Bounds Checker is
2118 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
2120 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
2122 * target.def (builtin_chkp_function): New.
2123 (chkp_bound_type): New.
2124 (chkp_bound_mode): New.
2125 (fn_abi_va_list_bounds_size): New.
2126 (load_bounds_for_arg): New.
2127 (store_bounds_for_arg): New.
2128 * targhooks.h (default_load_bounds_for_arg): New.
2129 (default_store_bounds_for_arg): New.
2130 (default_fn_abi_va_list_bounds_size): New.
2131 (default_chkp_bound_type): New.
2132 (default_chkp_bound_mode): New.
2133 (default_builtin_chkp_function): New.
2134 * targhooks.c (default_load_bounds_for_arg): New.
2135 (default_store_bounds_for_arg): New.
2136 (default_fn_abi_va_list_bounds_size): New.
2137 (default_chkp_bound_type): New.
2138 (default_chkp_bound_mode); New.
2139 (default_builtin_chkp_function): New.
2140 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
2141 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
2142 (TARGET_STORE_BOUNDS_FOR_ARG): New.
2143 (TARGET_BUILTIN_CHKP_FUNCTION): New.
2144 (TARGET_CHKP_BOUND_TYPE): New.
2145 (TARGET_CHKP_BOUND_MODE): New.
2146 * doc/tm.texi: Regenerated.
2147 * langhooks.h (lang_hooks): Add chkp_supported field.
2148 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
2149 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
2151 2013-10-29 Andrew Pinski <apinski@cavium.com>
2153 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
2154 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
2155 fails, combining the branches anyways.
2156 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
2157 increases the number of combinings.
2158 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
2160 2013-10-29 Mike Stump <mikestump@comcast.net>
2162 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
2163 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
2164 (make_vector_mode): Increase namebuf to 16.
2165 (emit_insn_modes_h): When processing BImode, don't
2166 also match partial int modes.
2167 (emit_class_narrowest_mode): Likewise.
2169 * config/bfin/bfin-modes.def: Add precision to PDI.
2170 * config/m32c/m32c-modes.def: Add precision to PSI.
2171 * config/msp430/msp430-modes.def: Add precision to PSI.
2172 * config/rs6000/rs6000-modes.def: Add precision to PTI.
2173 * config/sh/sh-modes.def: Add precision to PSI and PDI.
2175 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
2178 * config/sh/sh.md (*addc): Rename existing variations to ...
2179 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
2180 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
2181 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
2182 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
2184 2013-10-29 Teresa Johnson <tejohnson@google.com>
2187 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
2188 insanities when updating probabilities.
2190 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2192 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
2193 prettyprinter for cgraph_node to reflect the conversion of the
2194 symtable types to a C++ class hierarchy: it now *is* a
2195 symtab_node_base, rather than having one (named "symbol").
2197 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
2199 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
2200 __cilkrts_pop_frame. If matched, then return true for built-in
2202 (expand_builtin): Added BUILT_IN_CILK_DETACH and
2203 BUILT_IN_CILK_POP_FRAME case.
2204 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
2205 (lhs_cilk_detect_spawn): Likewise.
2206 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
2207 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
2208 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
2209 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
2210 (LANG_HOOKS_CILKPLUS): Likewise.
2211 * tree.h (CILK_SPAWN_FN): Likewise.
2212 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
2213 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
2214 (OBJS): Added cilk-common.o.
2215 (BUILTINS_DEF): Added cilk-builtins.def.
2216 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
2217 (lhd_cilk_detect_spawn): Likewise.
2218 * langhooks.h (lang_hooks_for_cilkplus): New struct.
2219 (struct lang_hooks): Added new field called "cilkplus."
2220 * cilk-common.c: New file.
2222 * cilk-builtins.def: Likewise.
2223 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
2224 "__cilk" macro and set it to 200.
2225 * function.h (struct function::cilk_frame_decl): New field.
2226 (struct function::is_cilk_function): Likewise.
2227 (struct function::calls_cilk_spawn): Likewise.
2228 * gimplify.c (gimplify_call_expr): Added a check if the function call
2229 being gimplified is a spawn detach point. If so, then add pop_frame
2230 and detach function calls.
2231 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
2232 for gimplifying _Cilk_spawn and _Cilk_sync statements.
2233 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
2234 function. If so, added a _Cilk_sync and gimplified it.
2235 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
2236 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
2237 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
2239 (can_inline_edge_p): Prevent inling of spawnee function.
2240 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
2241 for functions that use Cilk keywords.
2242 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
2243 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
2244 CILK_SYNC_STMT cases.
2245 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
2247 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
2248 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
2249 * passes.texi (Cilk Keywords): New section that describes the compiler
2250 code changes for handling Cilk Keywords.
2252 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2254 Patch autogenerated by refactor_symtab.py from
2255 https://github.com/davidmalcolm/gcc-refactoring-scripts
2256 revision 58bb219cc090b2f4516a9297d868c245495ee622
2258 * asan.c (asan_finish_file): Update for conversion of symtab types to
2259 a true class hierarchy.
2260 * cfgexpand.c (estimated_stack_frame_size): Likewise.
2261 * cgraph.c (cgraph_get_body): Likewise.
2262 (cgraph_get_create_real_symbol_node): Likewise.
2263 (verify_cgraph_node): Likewise.
2264 (verify_edge_corresponds_to_fndecl): Likewise.
2265 (verify_edge_count_and_frequency): Likewise.
2266 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
2267 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2268 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
2269 (cgraph_node_cannot_return): Likewise.
2270 (cgraph_set_pure_flag_1): Likewise.
2271 (cgraph_set_const_flag_1): Likewise.
2272 (cgraph_set_nothrow_flag_1): Likewise.
2273 (cgraph_make_node_local_1): Likewise.
2274 (cgraph_for_node_and_aliases): Likewise.
2275 (cgraph_for_node_thunks_and_aliases): Likewise.
2276 (cgraph_node_can_be_local_p): Likewise.
2277 (cgraph_node_cannot_be_local_p_1): Likewise.
2278 (cgraph_function_body_availability): Likewise.
2279 (dump_cgraph_node): Likewise.
2280 (cgraph_rtl_info): Likewise.
2281 (cgraph_mark_address_taken_node): Likewise.
2282 (cgraph_remove_node): Likewise.
2283 (cgraph_release_function_body): Likewise.
2284 (cgraph_update_edges_for_call_stmt_node): Likewise.
2285 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
2286 (cgraph_make_edge_direct): Likewise.
2287 (cgraph_resolve_speculation): Likewise.
2288 (cgraph_speculative_call_info): Likewise.
2289 (cgraph_turn_edge_to_speculative): Likewise.
2290 (cgraph_create_edge_1): Likewise.
2291 (cgraph_set_call_stmt): Likewise.
2292 (cgraph_node_for_asm): Likewise.
2293 (cgraph_add_thunk): Likewise.
2294 (cgraph_same_body_alias): Likewise.
2295 (cgraph_create_function_alias): Likewise.
2296 (cgraph_create_node): Likewise.
2297 (cgraph_create_empty_node): Likewise.
2298 (record_function_versions): Likewise.
2299 (used_from_object_file_p): Likewise.
2300 * cgraph.h (symtab_can_be_discarded): Likewise.
2301 (symtab_real_symbol_p): Likewise.
2302 (cgraph_mark_force_output_node): Likewise.
2303 (cgraph_edge_recursive_p): Likewise.
2304 (symtab_alias_target): Likewise.
2305 (varpool_all_refs_explicit_p): Likewise.
2306 (varpool_can_remove_if_no_refs): Likewise.
2307 (cgraph_only_called_directly_or_aliased_p): Likewise.
2308 (cgraph_next_function_with_gimple_body): Likewise.
2309 (cgraph_first_function_with_gimple_body): Likewise.
2310 (cgraph_function_with_gimple_body_p): Likewise.
2311 (cgraph_next_function): Likewise.
2312 (cgraph_first_function): Likewise.
2313 (cgraph_next_defined_function): Likewise.
2314 (cgraph_first_defined_function): Likewise.
2315 (varpool_next_defined_variable): Likewise.
2316 (varpool_first_defined_variable): Likewise.
2317 (varpool_next_static_initializer): Likewise.
2318 (varpool_first_static_initializer): Likewise.
2319 (varpool_next_variable): Likewise.
2320 (varpool_first_variable): Likewise.
2321 (varpool_node_name): Likewise.
2322 (varpool): Likewise.
2324 (is_a_helper <varpool_node>::test): Likewise.
2325 (is_a_helper <cgraph_node>::test): Likewise.
2326 (varpool_variable_node): Likewise.
2327 (cgraph_function_or_thunk_node): Likewise.
2328 (varpool_alias_target): Likewise.
2329 (cgraph_alias_target): Likewise.
2330 (cgraph_node_name): Likewise.
2331 (varpool_node_asm_name): Likewise.
2332 (cgraph_node_asm_name): Likewise.
2333 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
2334 (cgraph_rebuild_references): Likewise.
2335 (rebuild_cgraph_edges): Likewise.
2336 (record_eh_tables): Likewise.
2337 (build_cgraph_edges): Likewise.
2338 (mark_store): Likewise.
2339 (mark_load): Likewise.
2340 (mark_address): Likewise.
2341 (record_type_list): Likewise.
2342 (record_reference): Likewise.
2343 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
2344 (cgraph_materialize_clone): Likewise.
2345 (cgraph_function_versioning): Likewise.
2346 (cgraph_copy_node_for_versioning): Likewise.
2347 (update_call_expr): Likewise.
2348 (cgraph_find_replacement_node): Likewise.
2349 (cgraph_create_virtual_clone): Likewise.
2350 (cgraph_clone_node): Likewise.
2351 * cgraphunit.c (compile): Likewise.
2352 (output_weakrefs): Likewise.
2353 (output_in_order): Likewise.
2354 (expand_function): Likewise.
2355 (assemble_thunks_and_aliases): Likewise.
2356 (expand_thunk): Likewise.
2357 (mark_functions_to_output): Likewise.
2358 (handle_alias_pairs): Likewise.
2359 (analyze_functions): Likewise.
2360 (walk_polymorphic_call_targets): Likewise.
2361 (varpool_finalize_decl): Likewise.
2362 (process_function_and_variable_attributes): Likewise.
2363 (cgraph_process_same_body_aliases): Likewise.
2364 (analyze_function): Likewise.
2365 (cgraph_add_new_function): Likewise.
2366 (cgraph_finalize_function): Likewise.
2367 (referred_to_p): Likewise.
2368 (cgraph_reset_node): Likewise.
2369 (cgraph_process_new_functions): Likewise.
2370 (enqueue_node): Likewise.
2371 (decide_is_symbol_needed): Likewise.
2372 * coverage.c (coverage_compute_profile_id): Likewise.
2373 * dbxout.c (dbxout_expand_expr): Likewise.
2374 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
2375 (reference_to_unused): Likewise.
2376 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2377 * gimplify.c (unvisit_body): Likewise.
2378 (unshare_body): Likewise.
2379 * ipa-cp.c (ipcp_generate_summary): Likewise.
2380 (ipcp_decision_stage): Likewise.
2381 (identify_dead_nodes): Likewise.
2382 (decide_whether_version_node): Likewise.
2383 (decide_about_value): Likewise.
2384 (perhaps_add_new_callers): Likewise.
2385 (create_specialized_node): Likewise.
2386 (update_profiling_info): Likewise.
2387 (ipcp_propagate_stage): Likewise.
2388 (estimate_local_effects): Likewise.
2389 (good_cloning_opportunity_p): Likewise.
2390 (devirtualization_time_bonus): Likewise.
2391 (propagate_constants_accross_call): Likewise.
2392 (initialize_node_lattices): Likewise.
2393 (ipcp_cloning_candidate_p): Likewise.
2394 (determine_versionability): Likewise.
2395 (print_all_lattices): Likewise.
2396 (print_lattice): Likewise.
2397 (ipcp_discover_new_direct_edges): Likewise.
2398 * ipa-devirt.c (ipa_devirt): Likewise.
2399 (likely_target_p): Likewise.
2400 (update_type_inheritance_graph): Likewise.
2401 (possible_polymorphic_call_target_p): Likewise.
2402 (dump_possible_polymorphic_call_targets): Likewise.
2403 (devirt_variable_node_removal_hook): Likewise.
2404 (record_binfo): Likewise.
2405 (maybe_record_node): Likewise.
2406 (build_type_inheritance_graph): Likewise.
2407 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2408 (inline_generate_summary): Likewise.
2409 (inline_analyze_function): Likewise.
2410 (do_estimate_growth): Likewise.
2411 (simple_edge_hints): Likewise.
2412 (estimate_node_size_and_time): Likewise.
2413 (estimate_edge_devirt_benefit): Likewise.
2414 (compute_inline_parameters): Likewise.
2415 (estimate_function_body_sizes): Likewise.
2416 (compute_bb_predicates): Likewise.
2417 (initialize_inline_failed): Likewise.
2418 (dump_inline_summary): Likewise.
2419 (dump_inline_edge_summary): Likewise.
2420 * ipa-inline-transform.c (inline_transform): Likewise.
2421 (preserve_function_body_p): Likewise.
2422 (save_inline_function_body): Likewise.
2423 (inline_call): Likewise.
2424 (clone_inlined_nodes): Likewise.
2425 (can_remove_node_now_p): Likewise.
2426 (can_remove_node_now_p_1): Likewise.
2427 * ipa-inline.c (early_inliner): Likewise.
2428 (early_inline_small_functions): Likewise.
2429 (inline_always_inline_functions): Likewise.
2430 (ipa_inline): Likewise.
2431 (flatten_function): Likewise.
2432 (inline_small_functions): Likewise.
2433 (speculation_useful_p): Likewise.
2434 (recursive_inlining): Likewise.
2435 (update_caller_keys): Likewise.
2436 (reset_edge_caches): Likewise.
2437 (update_edge_key): Likewise.
2438 (edge_badness): Likewise.
2439 (relative_time_benefit): Likewise.
2440 (want_inline_self_recursive_call_p): Likewise.
2441 (want_inline_small_function_p): Likewise.
2442 (want_early_inline_function_p): Likewise.
2443 (num_calls): Likewise.
2444 (can_early_inline_edge_p): Likewise.
2445 (can_inline_edge_p): Likewise.
2446 (report_inline_failed_reason): Likewise.
2447 * ipa-profile.c (ipa_profile): Likewise.
2448 (ipa_propagate_frequency): Likewise.
2449 (ipa_propagate_frequency_1): Likewise.
2450 (ipa_profile_generate_summary): Likewise.
2451 * ipa-prop.c (ipcp_transform_function): Likewise.
2452 (read_replacements_section): Likewise.
2453 (ipa_prop_read_section): Likewise.
2454 (ipa_modify_call_arguments): Likewise.
2455 (ipa_print_node_params): Likewise.
2456 (propagate_controlled_uses): Likewise.
2457 (update_indirect_edges_after_inlining): Likewise.
2458 (remove_described_reference): Likewise.
2459 (ipa_make_edge_direct_to_target): Likewise.
2460 (ipa_analyze_node): Likewise.
2461 (ipa_analyze_params_uses): Likewise.
2462 (ipa_compute_jump_functions): Likewise.
2463 (ipa_get_callee_param_type): Likewise.
2464 (ipa_print_node_jump_functions): Likewise.
2465 (ipa_initialize_node_params): Likewise.
2466 (ipa_populate_param_decls): Likewise.
2467 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
2468 (write_agg_replacement_chain): Likewise.
2469 (ipa_write_node_info): Likewise.
2470 (ipa_edge_duplication_hook): Likewise.
2471 (try_decrement_rdesc_refcount): Likewise.
2472 * ipa-pure-const.c (propagate_nothrow): Likewise.
2473 (propagate_pure_const): Likewise.
2474 (pure_const_read_summary): Likewise.
2475 (pure_const_write_summary): Likewise.
2476 (analyze_function): Likewise.
2477 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
2478 (ipa_ref_referring_ref_list): Likewise.
2479 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
2480 (ipa_remove_stmt_references): Likewise.
2481 (ipa_find_reference): Likewise.
2482 (ipa_dump_referring): Likewise.
2483 (ipa_dump_references): Likewise.
2484 (ipa_record_reference): Likewise.
2485 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
2486 (ipa_reference_write_optimization_summary): Likewise.
2487 (write_node_summary_p): Likewise.
2488 (propagate): Likewise.
2489 (read_write_all_from_decl): Likewise.
2490 (generate_summary): Likewise.
2491 (analyze_function): Likewise.
2492 (propagate_bits): Likewise.
2493 (ipa_reference_get_not_written_global): Likewise.
2494 (ipa_reference_get_not_read_global): Likewise.
2495 * ipa-split.c (execute_split_functions): Likewise.
2496 (split_function): Likewise.
2497 * ipa-utils.c (ipa_merge_profiles): Likewise.
2498 (dump_cgraph_node_set): Likewise.
2499 (ipa_reverse_postorder): Likewise.
2500 (ipa_edge_within_scc): Likewise.
2501 (ipa_get_nodes_in_cycle): Likewise.
2502 (ipa_free_postorder_info): Likewise.
2503 (ipa_reduced_postorder): Likewise.
2504 (searchc): Likewise.
2505 (recursive_call_p): Likewise.
2506 * ipa.c (ipa_cdtor_merge): Likewise.
2507 (record_cdtor_fn): Likewise.
2508 (function_and_variable_visibility): Likewise.
2509 (varpool_externally_visible_p): Likewise.
2510 (cgraph_externally_visible_p): Likewise.
2511 (comdat_can_be_unshared_p): Likewise.
2512 (comdat_can_be_unshared_p_1): Likewise.
2513 (address_taken_from_non_vtable_p): Likewise.
2514 (ipa_discover_readonly_nonaddressable_vars): Likewise.
2515 (symtab_remove_unreachable_nodes): Likewise.
2516 (walk_polymorphic_call_targets): Likewise.
2517 (process_references): Likewise.
2518 (enqueue_node): Likewise.
2519 (has_addr_references_p): Likewise.
2520 (cgraph_non_local_node_p_1): Likewise.
2521 * is-a.h (varpool_analyze_node): Likewise.
2522 * lto-cgraph.c (input_symtab): Likewise.
2523 (merge_profile_summaries): Likewise.
2524 (input_cgraph_1): Likewise.
2525 (input_edge): Likewise.
2526 (input_varpool_node): Likewise.
2527 (input_node): Likewise.
2528 (input_overwrite_node): Likewise.
2529 (compute_ltrans_boundary): Likewise.
2530 (output_refs): Likewise.
2531 (lto_output_varpool_node): Likewise.
2532 (lto_output_node): Likewise.
2533 (reachable_from_other_partition_p): Likewise.
2534 (referenced_from_other_partition_p): Likewise.
2535 (lto_output_edge): Likewise.
2536 (output_node_opt_summary): Likewise.
2537 (add_node_to): Likewise.
2538 (reachable_from_this_partition_p): Likewise.
2539 (lto_set_symtab_encoder_in_partition): Likewise.
2540 (lto_symtab_encoder_in_partition_p): Likewise.
2541 (lto_set_symtab_encoder_encode_initializer): Likewise.
2542 (lto_symtab_encoder_encode_initializer_p): Likewise.
2543 (lto_set_symtab_encoder_encode_body): Likewise.
2544 (lto_symtab_encoder_encode_body_p): Likewise.
2545 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2547 * lto-streamer-in.c (lto_read_body): Likewise.
2548 (fixup_call_stmt_edges): Likewise.
2549 (fixup_call_stmt_edges_1): Likewise.
2550 * lto-streamer-out.c (produce_symtab): Likewise.
2551 (output_symbol_p): Likewise.
2552 (write_symbol): Likewise.
2553 (lto_output): Likewise.
2554 (copy_function): Likewise.
2555 (output_function): Likewise.
2556 * passes.c (function_called_by_processed_nodes_p): Likewise.
2557 (ipa_write_optimization_summaries): Likewise.
2558 (ipa_write_summaries): Likewise.
2559 (do_per_function_toporder): Likewise.
2560 (do_per_function): Likewise.
2561 (dump_passes): Likewise.
2562 * symtab.c (symtab_semantically_equivalent_p): Likewise.
2563 (symtab_nonoverwritable_alias): Likewise.
2564 (symtab_nonoverwritable_alias_1): Likewise.
2565 (symtab_for_node_and_aliases): Likewise.
2566 (symtab_resolve_alias): Likewise.
2567 (fixup_same_cpp_alias_visibility): Likewise.
2568 (symtab_alias_ultimate_target): Likewise.
2569 (symtab_used_from_object_file_p): Likewise.
2570 (verify_symtab_base): Likewise.
2571 (dump_symtab_base): Likewise.
2572 (symtab_node_name): Likewise.
2573 (symtab_node_asm_name): Likewise.
2574 (symtab_dissolve_same_comdat_group_list): Likewise.
2575 (symtab_add_to_same_comdat_group): Likewise.
2576 (symtab_unregister_node): Likewise.
2577 (symtab_insert_node_to_hashtable): Likewise.
2578 (symtab_register_node): Likewise.
2579 (unlink_from_assembler_name_hash): Likewise.
2580 (insert_to_assembler_name_hash): Likewise.
2581 (eq_assembler_name): Likewise.
2582 (hash_node_by_assembler_name): Likewise.
2583 (eq_node): Likewise.
2584 (hash_node): Likewise.
2585 * toplev.c (wrapup_global_declaration_2): Likewise.
2586 * trans-mem.c (ipa_tm_execute): Likewise.
2587 (ipa_tm_transform_clone): Likewise.
2588 (ipa_tm_transform_transaction): Likewise.
2589 (ipa_tm_transform_calls_redirect): Likewise.
2590 (ipa_tm_insert_gettmclone_call): Likewise.
2591 (ipa_tm_insert_irr_call): Likewise.
2592 (ipa_tm_create_version): Likewise.
2593 (ipa_tm_create_version_alias): Likewise.
2594 (ipa_tm_mark_forced_by_abi_node): Likewise.
2595 (ipa_tm_mark_force_output_node): Likewise.
2596 (ipa_tm_diagnose_tm_safe): Likewise.
2597 (ipa_tm_mayenterirr_function): Likewise.
2598 (ipa_tm_scan_irr_function): Likewise.
2599 (ipa_tm_note_irrevocable): Likewise.
2600 (ipa_tm_scan_calls_clone): Likewise.
2601 (get_cg_data): Likewise.
2602 * tree-eh.c (tree_could_trap_p): Likewise.
2603 * tree-emutls.c (ipa_lower_emutls): Likewise.
2604 (create_emultls_var): Likewise.
2605 (lower_emutls_function_body): Likewise.
2606 (gen_emutls_addr): Likewise.
2607 (emutls_decl): Likewise.
2608 (new_emutls_decl): Likewise.
2609 * tree-inline.c (tree_function_versioning): Likewise.
2610 (optimize_inline_calls): Likewise.
2611 (expand_call_inline): Likewise.
2612 (estimate_num_insns): Likewise.
2613 (copy_bb): Likewise.
2614 (delete_unreachable_blocks_update_callgraph): Likewise.
2615 * tree-nested.c (gimplify_all_functions): Likewise.
2616 (create_nesting_tree): Likewise.
2617 (check_for_nested_with_variably_modified): Likewise.
2618 * tree-pretty-print.c (dump_function_header): Likewise.
2619 * tree-profile.c (tree_profiling): Likewise.
2620 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
2621 (modify_function): Likewise.
2622 (convert_callers): Likewise.
2623 (convert_callers_for_node): Likewise.
2624 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
2625 (associate_varinfo_to_alias): Likewise.
2626 (create_variable_info_for): Likewise.
2627 (get_constraint_for_ssa_var): Likewise.
2628 * tree-vectorizer.c (increase_alignment): Likewise.
2629 * tree.c (find_decls_types_in_var): Likewise.
2630 (find_decls_types_in_node): Likewise.
2631 (free_lang_data_in_decl): Likewise.
2632 * value-prof.c (gimple_ic_transform): Likewise.
2633 (gimple_ic): Likewise.
2634 (check_ic_target): Likewise.
2635 (init_node_map): Likewise.
2636 * varasm.c (decl_binds_to_current_def_p): Likewise.
2637 (default_binds_local_p_1): Likewise.
2638 (dump_tm_clone_pairs): Likewise.
2639 (assemble_alias): Likewise.
2640 (find_decl): Likewise.
2641 (mark_decl_referenced): Likewise.
2642 * varpool.c (varpool_for_node_and_aliases): Likewise.
2643 (varpool_extra_name_alias): Likewise.
2644 (varpool_create_variable_alias): Likewise.
2645 (add_new_static_var): Likewise.
2646 (varpool_finalize_named_section_flags): Likewise.
2647 (varpool_remove_unreferenced_decls): Likewise.
2648 (enqueue_node): Likewise.
2649 (varpool_assemble_decl): Likewise.
2650 (assemble_aliases): Likewise.
2651 (varpool_analyze_node): Likewise.
2652 (cgraph_variable_initializer_availability): Likewise.
2653 (varpool_add_new_variable): Likewise.
2654 (ctor_for_folding): Likewise.
2655 (dump_varpool_node): Likewise.
2656 (varpool_remove_initializer): Likewise.
2657 (varpool_remove_node): Likewise.
2658 (varpool_node_for_decl): Likewise.
2659 (varpool_create_empty_node): Likewise.
2660 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
2661 (ix86_get_function_versions_dispatcher): Likewise.
2663 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2665 * cgraph.h (symtab_node_base): Convert to a class;
2666 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
2667 chain_next/prev from symtab_node_def.
2668 (cgraph_node): Inherit from symtab_node; add GTY option
2669 tag ("SYMTAB_FUNCTION").
2670 (varpool_node): Inherit from symtab_node; add GTY option
2671 tag ("SYMTAB_VARIABLE").
2672 (symtab_node_def): Remove.
2673 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
2674 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
2675 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
2676 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
2678 * ipa-ref.h (symtab_node_def): Drop.
2679 (symtab_node): Change underlying type from symtab_node_def to
2681 (const_symtab_node): Likwise.
2683 * is-a.h: Update examples in comment.
2685 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
2686 (assembler_name_hash): Likewise.
2688 2013-10-29 Martin Liska <marxin.liska@gmail.com>
2690 * doc/tree-ssa.texi (gimple_phi_result): Document.
2691 (gimple_phi_num_args, gimple_phi_arg): Likewise.
2692 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
2693 (PHI_RESULT, PHI_NUM_ARGS): Remove.
2694 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
2696 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2698 * expr.h: Revert change and include tree-core.h.
2699 * rtl.h: Revert change and don't include tree-core.h.
2701 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2703 * config/darwin.c: Include gimple.h.
2704 * config/i386/winnt.c: Likewise.
2706 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
2708 PR tree-optimization/19831
2709 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
2711 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2713 * tree-outof-ssa.h: Remove include files.
2714 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
2716 * tree-ssa-coalesce.c: Likewise.
2717 * cfgexpand.c: Likewise.
2718 * tree-ssa-ter.c: Likewise.
2719 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
2720 * lto-streamer.h: Likewise.
2721 * cgraphbuild.c: Add gimple.h to include list.
2722 * data-streamer-in.c: Likewise.
2723 * ipa-cp.c: Likewise.
2724 * tree-streamer.c: Likewise.
2725 * lto-compress.c: Likewise.
2726 * ipa-reference.c: Likewise.
2727 * data-streamer-out.c: Likewise.
2728 * lto-cgraph.c: Likewise.
2729 * cgraphclones.c: Likewise.
2730 * ipa-utils.c: Likewise.
2731 * data-streamer.c: Likewise.
2732 * ipa-split.c: Likewise.
2733 * lto-section-in.c: Likewise.
2734 * tree-streamer-out.c: Likewise.
2735 * ipa-prop.c: Likewise.
2736 * tree-streamer-in.c: Likewise.
2737 * symtab.c: Likewise.
2738 * opts-global.c: Likewise.
2739 * lto-opts.c: Likewise.
2740 * lto-section-out.c: Likewise.
2741 * lto-streamer.c: Likewise.
2742 * rtl.h: Add tree-core.h to include list.
2743 * expr.h: Remove tree-core.h from include list.
2744 * gimple.h: Likewise.
2745 * ipa-utils.h: Likewise.
2746 * streamer-hooks.h: Likewise.
2747 * streamer-hooks.c: Include input.h.
2749 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2751 * config/arm/arm.c (cortexa7_extra_costs): New table.
2752 (arm_cortex_a7_tune): New.
2753 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2755 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2757 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2759 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2761 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2762 to use autogenerated markers for a class-hierarchy, every class
2763 must have a GTY marker.
2764 * gengtype.h (struct type): Add linked list of subclasses to
2765 the "s" member of the union.
2766 (add_subclass): New decl.
2767 * gengtype-state.c (read_state_struct_type): Set up subclass
2769 * gengtype.c (get_ultimate_base_class): New.
2770 (add_subclass): New.
2771 (new_structure): Set up subclass linked list.
2772 (set_gc_used_type): Propagate usage information to subclasses.
2773 (output_mangled_typename): Use get_ultimate_base_class.
2774 (walk_subclasses): Use the subclass linked list, avoiding an
2775 O(N^2) when writing out all types.
2776 (walk_type): Issue an error if the base class is missing a tag,
2777 rather than generating bogus C code. Add a gcc_unreachable
2778 default case, in case people omit tags from concrete subclasses,
2779 or get the values wrong.
2780 (write_func_for_structure): Issue an error for subclasses for
2781 which the base doesn't have a "desc", since otherwise the
2782 autogenerated routines for the base would silently fail to visit
2783 any subclass fields.
2784 (write_root): Use get_ultimate_base_class, tweaking constness of
2785 tp to match that function's signature.
2787 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2789 * doc/gty.texi (GTY Options): Add note about inheritance to
2790 description of desc and tag.
2791 (Inheritance and GTY): New.
2793 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2795 * gengtype-parse.c (opts_have): Drop "static" so that
2796 we can use this from gengtype.c.
2797 * gengtype.c (set_gc_used_type): Mark any base class as used;
2798 update field traversal to visit inherited fields.
2799 (output_mangled_typename): Convert references to classes within
2800 an inheritance hierarchy to reference the ultimate base class,
2801 since only it will have gt_ functions.
2802 (get_string_option): New.
2803 (walk_subclasses): New.
2804 (walk_type): Treat GTY structs that have a "desc" as being the
2805 root of an inheritance hierarchy. Generate a switch on it
2806 within the marking function which walks all subclasses, adding
2807 cases for them via walk_subclasses. For subclasses, visit all
2808 fields of the type (including inherited ones).
2809 (write_func_for_structure): Don't write fns for subclasses, only
2810 for the ultimate base class within an inheritance hierarchy.
2811 Subclasses-marking will be handled by the base class marking functions.
2812 (write_types): Likewise.
2813 (write_local_func_for_structure): Likewise.
2814 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2815 a "tag" option (and are thus concrete subclasses).
2816 (write_root): Use the marker function for the ultimate base class.
2817 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2818 (opts_have): Add declaration.
2820 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2822 * lra-spills.c (lra_final_code_change): Remove useless move insns
2823 originated from moves of pseudos.
2825 2013-10-28 Jeff Law <law@redhat.com>
2827 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2829 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2831 * df-scan.c (df_collection_rec): Adjust.
2832 (copy_defs): New constant.
2833 (copy_uses): Likewise.
2834 (copy_eq_uses): Likewise.
2835 (copy_mw): Likewise.
2836 (copy_all): Likewise.
2837 (df_insn_rescan): Adjust.
2838 (df_notes_rescan): Likewise.
2839 (df_swap_refs): Likewise.
2840 (df_sort_and_compress_refs): Likewise.
2841 (df_sort_and_compress_mws): Likewise.
2842 (df_install_refs): Likewise.
2843 (df_install_mws): Likewise.
2844 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2846 (df_bb_refs_record): Adjust.
2847 (df_record_entry_block_defs): Likewise.
2848 (df_record_exit_block_defs): Likewise.
2849 (df_refs_verify): Likewise.
2850 (df_mws_verify): Likewise.
2851 (df_insn_refs_verify): Likewise.
2852 (df_bb_verify): Likewise.
2853 * ipa-pure-const.c (finish_state): Remove.
2854 (propagate): Adjust.
2855 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2856 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2857 var-tracking.c: Adjust.
2858 * vec.c (stack_vecs): Remove.
2859 (register_stack_vec): Likewise.
2860 (stack_vec_register_index): Likewise.
2861 (unregister_stack_vec): Likewise.
2862 * vec.h (struct va_stack): Remove.
2863 (struct vec<T, A, vl_ptr>): Specialize as
2864 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2865 allocation strategy compatable with the vl_ptr layout.
2866 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2867 specialization anyway.
2868 (class stack_vec): New class.
2869 (vec_stack_alloc): Remove.
2870 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2872 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2873 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2874 Sergey Lega <sergey.s.lega@intel.com>
2875 Anna Tikhonova <anna.tikhonova@intel.com>
2876 Ilya Tocar <ilya.tocar@intel.com>
2877 Andrey Turetskiy <andrey.turetskiy@intel.com>
2878 Ilya Verbin <ilya.verbin@intel.com>
2879 Kirill Yukhin <kirill.yukhin@intel.com>
2880 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2882 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2883 (*prefetch_avx512pf_<mode>): New.
2884 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2885 (avx512f_maskcmp<mode>3): Ditto.
2886 (vashrv16si3): Ditto.
2888 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2889 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2890 Sergey Lega <sergey.s.lega@intel.com>
2891 Anna Tikhonova <anna.tikhonova@intel.com>
2892 Ilya Tocar <ilya.tocar@intel.com>
2893 Andrey Turetskiy <andrey.turetskiy@intel.com>
2894 Ilya Verbin <ilya.verbin@intel.com>
2895 Kirill Yukhin <kirill.yukhin@intel.com>
2896 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2898 * config/i386/i386.md (any_truncate): New.
2899 (trunsuffix): Ditto.
2900 * config/i386/predicates.md (const_8_to_9_operand): New.
2901 (const_10_to_11_operand): Ditto.
2902 (const_12_to_13_operand): Ditto.
2903 (const_14_to_15_operand): Ditto.
2904 (const_16_to_19_operand): Ditto.
2905 (const_20_to_23_operand): Ditto.
2906 (const_24_to_27_operand): Ditto.
2907 (const_28_to_31_operand): Ditto.
2908 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2909 (cvtusi2<ssescalarmodesuffix>32): New.
2910 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2911 (ufloatv16siv16sf2): Ditto.
2912 (avx512f_fix_notruncv16sfv16si): Ditto.
2913 (avx512f_ufix_notruncv16sfv16si): Ditto.
2914 (avx512f_vcvtss2usi): Ditto.
2915 (avx512f_vcvtss2usiq): Ditto.
2916 (avx512f_vcvttss2usi): Ditto.
2917 (avx512f_vcvttss2usiq): Ditto.
2918 (avx512f_vcvtsd2usi): Ditto.
2919 (avx512f_vcvtsd2usiq): Ditto.
2920 (avx512f_vcvttsd2usi): Ditto.
2921 (avx512f_vcvttsd2usiq): Ditto.
2922 (ufloatv8siv8df): Ditto.
2923 (avx512f_cvtdq2pd512_2): Ditto.
2924 (avx512f_cvtpd2dq512): Ditto.
2925 (avx512f_ufix_notruncv8dfv8si): Ditto.
2926 (avx512f_cvtpd2ps512): Ditto.
2927 (vec_unpacks_lo_v16sf): Ditto.
2928 (vec_unpacks_hi_v16sf): Ditto.
2929 (vec_unpacks_float_hi_v16si): Ditto.
2930 (vec_unpacks_float_lo_v16si): Ditto.
2931 (avx512f_unpckhps512): Ditto.
2932 (avx512f_unpcklps512): Ditto.
2933 (avx512f_movshdup512): Ditto.
2934 (avx512f_movsldup512): Ditto.
2935 (vec_extract_lo_v32hi): Ditto.
2936 (vec_extract_hi_v32hi): Ditto.
2937 (vec_extract_lo_v64qi): Ditto.
2938 (vec_extract_hi_v64qi): Ditto.
2939 (avx512f_unpckhpd512): Ditto.
2940 (avx512f_movddup512): Ditto.
2941 (avx512f_unpcklpd512): Ditto.
2942 (*avx512f_unpcklpd512): Ditto.
2943 (avx512f_shufps512_1): Ditto.
2944 (avx512f_shufpd512_1): Ditto.
2945 (avx512f_interleave_highv8di): Ditto.
2946 (avx512f_interleave_lowv8di): Ditto.
2947 (PMOV_DST_MODE): Ditto.
2948 (pmov_src_mode): Ditto.
2949 (pmov_src_lower): Ditto.
2951 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2952 (*avx512f_<code>v8div16qi2): Ditto.
2953 (*avx512f_<code>v8div16qi2_store): Ditto.
2954 (vec_widen_umult_even_v16si): Ditto.
2955 (*vec_widen_umult_even_v16si): Ditto.
2956 (vec_widen_smult_even_v16si): Ditto.
2957 (*vec_widen_smult_even_v16si): Ditto.
2958 (avx512f_interleave_highv16si): Ditto.
2959 (avx512f_interleave_lowv16si): Ditto.
2960 (avx512f_<code>v16qiv16si2): Ditto.
2961 (avx512f_<code>v16hiv16si2): Ditto.
2962 (avx512f_<code>v8qiv8di2): Ditto.
2963 (avx512f_<code>v8hiv8di2): Ditto.
2964 (avx512f_<code>v8siv8di2): Ditto.
2965 (avx512cd_maskb_vec_dupv8di): Ditto.
2966 (avx512cd_maskw_vec_dupv16si): Ditto.
2967 (avx512f_vcvtph2ps512): Ditto.
2968 (avx512f_vcvtps2ph512): Ditto.
2969 (VEC_EXTRACT_MODE): Extened with wider modes.
2970 (VEC_PERM_AVX2): Ditto.
2971 (VEC_PERM_CONST): Ditto.
2973 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2975 * config/arc/arc.c (arc_ccfsm_post_advance):
2976 Add comment about TYPE_RETURN.
2978 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2980 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2981 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2982 (strip_offset): Convert offset to unsigned number.
2984 2013-10-27 Tom de Vries <tom@codesourcery.com>
2986 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2987 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2989 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2991 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2992 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2993 (TARGET_BIG_ENDIAN): New macro.
2994 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2995 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2996 * config/sh/sh.c: Likewise.
2997 * config/sh/sh.h: Likewise.
2999 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
3001 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
3002 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
3004 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
3006 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
3008 (cmpstr_t, cmpstrsi): Fix formatting.
3010 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
3013 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
3014 addressing, do not use general_operand for memory operands.
3016 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
3019 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
3020 * lra-spills.c (lra_final_code_change): Remove useless move insns.
3022 2013-10-26 Jeff Law <law@redhat.com>
3024 * predict.c (PRED_MUDFLAP): Remove.
3025 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
3027 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
3028 (OBJS): Remove tree-nomudflap.o
3029 (GTFILES): Remove tree-mudflap.c
3030 * builtins.c (expand_builtin_alloc): Remove mudflap support.
3031 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
3032 (mfwrap_spec, mflib_spec): Likewise.
3033 (cpp_unique_options, cc1_options, static_specs): Likewise.
3034 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
3035 * passes.def: Likewise.
3036 * toplev.c (compile_file, process_options): Likewise.
3037 * tree-inline.c (copy_tree_r): Likewise.
3038 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
3039 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
3040 (build_constant_desc, output_constant_def_contents): Likewise.
3041 (categorize_decl_for_section): Likewise.
3042 * tree-mudflap.c: Removed.
3043 * tree-mudflap.h: Removed.
3044 * tree-nomudflap.c: Removed.
3045 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
3046 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
3047 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
3048 * config/sol2.h (MFLIB_SPEC): Likewise.
3049 * doc/install.texi: Remove mudflap references.
3050 * doc/passes.texi: Similarly.
3051 * doc/sourcebuild.texi: Similarly.
3052 * doc/invoke.texi: Remove mudlfap related options.
3054 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
3056 PR rtl-optimization/58759
3057 * lra-constraints.c (lra_constraints): Remove wrong condition to
3058 remove insn setting up an equivalent pseudo.
3060 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
3062 * config/rs6000/rs6000-protos.h
3063 (rs6000_secondary_memory_needed_mode): New prototype.
3064 * config/rs6000/rs6000.c: Include ira.h.
3065 (TARGET_LRA_P): Redefine.
3066 (rs6000_legitimate_offset_address_p): Call
3067 legitimate_constant_pool_address_p in strict mode for LRA.
3068 (rs6000_legitimate_address_p): Ditto.
3069 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
3070 (rs6000_emit_move): Add LRA version of code to generate load/store
3072 (rs6000_secondary_memory_needed_mode): New.
3073 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
3074 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
3075 constants, memory, and FP registers.
3076 (rs6000_lra_p): New.
3077 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
3078 * config/rs6000/rs6000.opt (mlra): New option.
3079 * lra-spills.c (lra_final_code_change): Remove useless move insns.
3081 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
3083 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
3084 has_single_use () and not do the conversion if has_single_use ()
3085 returns false for the multiplication result.
3087 2013-10-25 David Malcolm <dmalcolm@redhat.com>
3089 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
3090 to "NODE", since this works on a "tree", not an
3092 (CONSTANT_CLASS_P): Likewise.
3095 (INDIRECT_REF_P): Likewise.
3096 (REFERENCE_CLASS_P): Likewise.
3097 (COMPARISON_CLASS_P): Likewise.
3098 (UNARY_CLASS_P): Likewise.
3099 (BINARY_CLASS_P): Likewise.
3100 (STATEMENT_CLASS_P): Likewise.
3101 (VL_EXP_CLASS_P): Likewise.
3102 (EXPRESSION_CLASS_P): Likewise.
3103 (IS_TYPE_OR_DECL_P): Likewise.
3105 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
3107 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
3108 ADDR_EXPR in the defining statement.
3110 2013-10-25 Richard Biener <rguenther@suse.de>
3112 PR tree-optimization/58626
3113 * tree-loop-distribution.c (enum rdg_dep_type): Remove
3114 anti_dd, output_dd and input_dd.
3115 (struct rdg_edge): Remove level and relation members.
3116 (RDGE_LEVEL, RDGE_RELATION): Remove.
3117 (dot_rdg_1): Adjust.
3118 (create_rdg_edge_for_ddr): Remove.
3119 (create_rdg_edges_for_scalar): Adjust.
3120 (create_edge_for_control_dependence): Likewise.
3121 (create_rdg_edges): Split into ...
3122 (create_rdg_flow_edges): ... this
3123 (create_rdg_cd_edges): ... and this.
3125 (build_rdg): Likewise, do not compute data dependences or
3127 (pg_add_dependence_edges): New function.
3129 (distribute_loop): First apply all non-dependence based
3130 partition mergings. Then compute dependences between partitions
3131 and merge and order partitions according to them.
3133 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
3135 PR rtl-optimization/58831
3136 * alias.c (init_alias_analysis): At the beginning of each iteration,
3137 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
3139 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
3141 * recog.c (search_ofs): New static variable moved from...
3142 (peep2_find_free_register): ...here.
3143 (peephole2_optimize): Initialize it.
3145 2013-10-25 Tobias Burnus <burnus@net-b.de>
3147 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
3149 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
3151 * config/i386/i386.h (TARGET_MPX): New define.
3152 (TARGET_MPX_P): Ditto.
3154 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
3156 * config/i386/constraints.md (B): New.
3159 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
3160 * config/i386/i386-modes.def (BND32): New.
3162 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
3163 * config/i386/i386.c (isa_opts): Add mmpx.
3164 (regclass_map): Add bound registers.
3165 (dbx_register_map): Likewise.
3166 (dbx64_register_map): Likewise.
3167 (svr4_dbx_register_map): Likewise.
3169 (ix86_option_override_internal): Support MPX ISA.
3170 (ix86_conditional_register_usage): Support bound registers.
3171 (print_reg): Likewise.
3172 (ix86_code_end): Add MPX bnd prefix.
3173 (output_set_got): Likewise.
3174 (ix86_output_call_insn): Likewise.
3175 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
3176 (ix86_print_operand_punct_valid_p): Likewise.
3177 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
3179 (ix86_class_likely_spilled_p): Add bound regs support.
3180 (ix86_hard_regno_mode_ok): Likewise.
3181 (x86_order_regs_for_local_alloc): Likewise.
3182 (ix86_bnd_prefixed_insn_p): New.
3183 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
3184 (FIXED_REGISTERS): Add bound registers.
3185 (CALL_USED_REGISTERS): Likewise.
3186 (REG_ALLOC_ORDER): Likewise.
3187 (HARD_REGNO_NREGS): Likewise.
3189 (VALID_BND_REG_MODE): New.
3190 (FIRST_BND_REG): New.
3191 (LAST_BND_REG): New.
3192 (reg_class): Add BND_REGS.
3193 (REG_CLASS_NAMES): Likewise.
3194 (REG_CLASS_CONTENTS): Likewise.
3196 (ANY_BND_REG_P): New.
3198 (HI_REGISTER_NAMES): Add bound registers.
3199 * config/i386/i386.md (UNSPEC_BNDMK): New.
3200 (UNSPEC_BNDMK_ADDR): New.
3201 (UNSPEC_BNDSTX): New.
3202 (UNSPEC_BNDLDX): New.
3203 (UNSPEC_BNDLDX_ADDR): New.
3204 (UNSPEC_BNDCL): New.
3205 (UNSPEC_BNDCU): New.
3206 (UNSPEC_BNDCN): New.
3207 (UNSPEC_MPX_FENCE): New.
3210 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
3211 (length_immediate): Likewise.
3212 (prefix_0f): Likewise.
3214 (prefix_rep): Check for bnd prefix.
3215 (length_nobnd): New.
3216 (length): Use length_nobnd if specified.
3221 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
3224 (simple_return_internal): Likewise.
3225 (simple_return_pop_internal): Likewise.
3226 (*indirect_jump): Add MPX bnd prefix.
3227 (*tablejump_1): Likewise.
3228 (simple_return_internal_long): Likewise.
3229 (simple_return_indirect_internal): Likewise.
3233 (*mov<mode>_internal_mpx): New.
3234 (<mode>_<bndcheck>): New.
3235 (*<mode>_<bndcheck>): New.
3240 * config/i386/predicates.md (lea_address_operand): Rename to...
3241 (address_no_seg_operand): ... this.
3242 (address_mpx_no_base_operand): New.
3243 (address_mpx_no_index_operand): New.
3244 (bnd_mem_operator): New.
3245 * config/i386/i386.opt (mmpx): New.
3246 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
3247 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
3249 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
3251 * mode-classes.def (MODE_POINTER_BOUNDS): New.
3252 * tree.def (POINTER_BOUNDS_TYPE): New.
3253 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
3254 (POINTER_BOUNDS_MODE): New.
3255 (make_pointer_bounds_mode): New.
3256 * machmode.h (POINTER_BOUNDS_MODE_P): New.
3257 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
3258 (layout_type): Support POINTER_BOUNDS_TYPE.
3259 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
3260 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
3261 (type_contains_placeholder_1): Likewise.
3262 * tree.h (POINTER_BOUNDS_TYPE_P): New.
3263 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
3264 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
3266 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3268 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
3269 mode of address computation when calling memory_address_addr_space.
3271 2013-08-24 Richard Henderson <rth@twiddle.net>
3274 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
3275 instead of create_convert_operand_to.
3276 (maybe_emit_sync_lock_test_and_set): Likewise.
3277 (expand_atomic_compare_and_swap): Likewise.
3278 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
3280 2013-08-24 Sriraman Tallam <tmsriram@google.com>
3282 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
3283 Change param_is to use the struct and not the pointer to the struct.
3285 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
3287 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
3288 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3289 Move to targhooks.c.
3290 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
3291 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3292 Relocate from gimplify.c.
3293 * targhooks.h: Add 2 prototypes.
3294 * tree.h: Delete 2 prototypes.
3296 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3298 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
3299 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
3301 2013-10-24 Cong Hou <congh@google.com>
3303 * convert.c (convert_to_real): Guard those unsafe math function
3304 convertions with flag_unsafe_math_optimizations. Handle sqrt()
3307 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
3310 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
3312 (cgraph_create_edge): Use the new argument.
3313 (cgraph_create_indirect_edge): Likewise.
3315 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
3317 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
3319 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
3320 changes statep->state.
3322 2013-10-24 Tobias Burnus <burnus@net-b.de>
3325 * tree-cfg.c (replace_loop_annotate): New function.
3326 (execute_build_cfg): Call it.
3327 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
3328 * internal-fn.c (expand_ANNOTATE): New function.
3329 * internal-fn.def (ANNOTATE): Define as new internal function.
3330 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
3331 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
3332 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
3333 * doc/extend.texi (Pragmas): Document #pragma ivdep.
3334 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
3336 2013-10-17 Ian Bolton <ian.bolton@arm.com>
3337 Marcus Shawcroft <marcus.shawcroft@arm.com>
3339 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
3340 Special case reload SP+C into none GENERAL_REGS.
3342 2013-10-24 Michael Matz <matz@suse.de>
3344 * gengtype.c (is_file_equal): Check that files will be same length.
3346 2013-10-25 Christian Bruel <christian.bruel@st.com>
3348 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
3349 * config/sh/t-sh (sh-mem.o): New rule.
3350 * config/sh/sh-mem.cc (expand_block_move): Moved here.
3351 (sh_expand_cmpstr): New function.
3352 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
3353 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
3354 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
3355 (rotlhi3_8): Rename.
3357 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3359 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
3360 * Makefile.in (CXXDEPMODE): Assign and change users.
3361 (CCDEPMODE): Delete.
3362 * configure: Regenerate.
3364 2013-10-23 David Malcolm <dmalcolm@redhat.com>
3366 * gengtype-parse.c (require_without_advance): New.
3367 (type): For GTY-marked types that are not GTY((user)), parse any
3368 base classes, requiring them to be single-inheritance, and not
3369 be templates. For non-GTY-marked types and GTY((user)),
3370 continue to skip over any C++ inheritance specification.
3371 * gengtype-state.c (state_writer::write_state_struct_type):
3372 Write base class of type (if any).
3373 (read_state_struct_type): Read base class of type (if any).
3374 * gengtype.c (new_structure): Add a "base_class" parameter.
3375 (create_optional_field_): Update for new parameter to new_structure.
3376 (adjust_field_rtx_def): Likewise.
3377 (adjust_field_tree_exp): Likewise.
3378 * gengtype.h (struct type): Add "base_class" field to the s
3380 (new_structure): Add "base" parameter.
3382 2013-10-23 Sriraman Tallam <tmsriram@google.com>
3385 * config/i386/i386.c (ix86_option_override_internal):
3386 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
3387 (ix86_valid_target_attribute_tree):
3388 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
3389 Change TARGET_SSE to TARGET_SSE_P (opts->...)
3391 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3393 * tree-ssa-loop.h: Remove include files.
3394 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3395 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
3396 * graphite-clast-to-gimple.c: Likewise.
3397 * graphite-scop-detection.c: Likewise.
3398 * graphite-sese-to-poly.c: Likewise.
3399 * ipa-inline-analysis.c: Likewise.
3400 * ipa-pure-const.c: Likewise.
3401 * loop-init.c: Likewise.
3402 * passes.c: Likewise.
3403 * predict.c: Likewise.
3404 * tree-cfg.c: Likewise.
3405 * tree-cfgcleanup.c: Likewise.
3406 * tree-chrec.c: Likewise.
3407 * tree-data-ref.c: Likewise.
3408 * tree-loop-distribution.c: Likewise.
3409 * tree-parloops.c: Likewise.
3410 * tree-predcom.c: Likewise.
3411 * tree-scalar-evolution.c: Likewise.
3412 * tree-ssa-address.c: Likewise.
3413 * tree-ssa.c: Likewise.
3414 * tree-ssa-dce.c: Likewise.
3415 * tree-ssa-loop.c: Likewise.
3416 * tree-ssa-loop-im.c: Likewise.
3417 * tree-ssa-loop-ivcanon.c: Likewise.
3418 * tree-ssa-loop-ivopts.c: Likewise.
3419 * tree-ssa-loop-manip.c: Likewise.
3420 * tree-ssa-loop-niter.c: Likewise.
3421 * tree-ssa-loop-prefetch.c: Likewise.
3422 * tree-ssa-loop-unswitch.c: Likewise.
3423 * tree-ssa-reassoc.c: Likewise.
3424 * tree-vect-data-refs.c: Likewise.
3425 * tree-vect-loop.c: Likewise.
3426 * tree-vect-loop-manip.c: Likewise.
3427 * tree-vectorizer.c: Likewise.
3428 * tree-vect-stmts.c: Likewise.
3429 * tree-vrp.c: Likewise.
3431 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3433 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
3435 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3437 PR tree-optimization/58775
3438 PR tree-optimization/58791
3439 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
3440 (insert_stmt_after): Rewritten, don't move the stmt, but really
3442 (get_stmt_uid_with_default): Remove.
3443 (build_and_add_sum): Use insert_stmt_after and
3444 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
3445 (update_range_test): Set uid on stmts added by
3446 force_gimple_operand_gsi. Don't immediately modify statements
3447 in inter-bb optimization, just update oe->op values.
3448 (optimize_range_tests): Return bool whether any changed have been made.
3449 (update_ops): New function.
3450 (struct inter_bb_range_test_entry): New type.
3451 (maybe_optimize_range_tests): Perform statement changes here.
3452 (not_dominated_by, appears_later_in_bb, get_def_stmt,
3453 ensure_ops_are_available): Remove.
3454 (find_insert_point): Rewritten.
3455 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
3456 return LHS of the (new resp. old) stmt. Don't call
3457 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
3458 instead of last, move new stmt at the right place.
3459 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
3460 (negate_value): Likewise. Set uids.
3461 (break_up_subtract_bb): Initialize uids.
3462 (reassociate_bb): Adjust rewrite_expr_tree caller.
3463 (do_reassoc): Don't call renumber_gimple_stmt_uids.
3465 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
3468 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
3469 TARGET_32BIT final condition.
3470 (mulsi3_internal2 and splitter): Same.
3472 2013-10-23 Jeff Law <law@redhat.com>
3474 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
3475 through joiner blocks with abnormal outgoing edges.
3477 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
3478 Add parameter JOINERS, to allow/disallow threading through joiner
3480 (thread_block): New. Call thread_block_1.
3481 (mark_threaded_blocks): Remove code to filter out certain cases
3482 of threading through joiner blocks.
3483 (thread_through_all_blocks): Document how we can have a dangling
3484 edge AUX field and clear it.
3486 2013-10-23 Ian Lance Taylor <iant@google.com>
3488 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
3489 (C++ Dialect Options): Likewise.
3490 (Optimize Options): Likewise.
3492 2013-10-23 Tom de Vries <tom@codesourcery.com>
3494 PR tree-optimization/58805
3495 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
3497 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3499 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
3500 sequence based on get_range_info returned range.
3502 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3504 * tree-ssa.h: Remove all #include's
3505 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3506 * alias.c: Move required includes from tree-ssa.h.
3508 * builtins.c: Likewise.
3509 * calls.c: Likewise.
3510 * cfgexpand.c: Likewise.
3511 * cfghooks.c: Likewise.
3512 * cfgloop.c: Likewise.
3513 * cfgloopmanip.c: Likewise.
3514 * cgraph.c: Likewise.
3515 * cgraphbuild.c: Likewise.
3516 * cgraphclones.c: Likewise.
3517 * cgraphunit.c: Likewise.
3519 * except.c: Likewise.
3521 * final.c: Likewise.
3522 * fold-const.c: Likewise.
3523 * ggc-page.c: Likewise.
3524 * gimple-builder.c: Likewise.
3525 * gimple-fold.c: Likewise.
3526 * gimple-iterator.c: Likewise.
3527 * gimple-low.c: Likewise.
3528 * gimple-pretty-print.c: Likewise.
3529 * gimple-ssa-strength-reduction.c: Likewise.
3530 * gimple-streamer-in.c: Likewise.
3531 * gimple-streamer-out.c: Likewise.
3532 * gimplify.c: Likewise.
3533 * graphite-blocking.c: Likewise.
3534 * graphite-clast-to-gimple.c: Likewise.
3535 * graphite-dependences.c: Likewise.
3536 * graphite-interchange.c: Likewise.
3537 * graphite-optimize-isl.c: Likewise.
3538 * graphite-poly.c: Likewise.
3539 * graphite-scop-detection.c: Likewise.
3540 * graphite-sese-to-poly.c: Likewise.
3541 * graphite.c: Likewise.
3542 * ipa-cp.c: Likewise.
3543 * ipa-inline-analysis.c: Likewise.
3544 * ipa-inline-transform.c: Likewise.
3545 * ipa-inline.c: Likewise.
3546 * ipa-prop.c: Likewise.
3547 * ipa-pure-const.c: Likewise.
3548 * ipa-reference.c: Likewise.
3549 * ipa-split.c: Likewise.
3550 * ipa-utils.c: Likewise.
3551 * loop-init.c: Likewise.
3552 * lto-cgraph.c: Likewise.
3553 * lto-section-in.c: Likewise.
3554 * lto-section-out.c: Likewise.
3555 * lto-streamer-in.c: Likewise.
3556 * lto-streamer-out.c: Likewise.
3557 * lto-streamer.c: Likewise.
3558 * omp-low.c: Likewise.
3559 * passes.c: Likewise.
3560 * predict.c: Likewise.
3561 * print-tree.c: Likewise.
3562 * profile.c: Likewise.
3564 * targhooks.c: Likewise.
3565 * tracer.c: Likewise.
3566 * trans-mem.c: Likewise.
3567 * tree-call-cdce.c: Likewise.
3568 * tree-cfg.c: Likewise.
3569 * tree-cfgcleanup.c: Likewise.
3570 * tree-chrec.c: Likewise.
3571 * tree-complex.c: Likewise.
3572 * tree-data-ref.c: Likewise.
3573 * tree-dfa.c: Likewise.
3574 * tree-eh.c: Likewise.
3575 * tree-emutls.c: Likewise.
3576 * tree-if-conv.c: Likewise.
3577 * tree-inline.c: Likewise.
3578 * tree-into-ssa.c: Likewise.
3579 * tree-loop-distribution.c: Likewise.
3580 * tree-mudflap.c: Likewise.
3581 * tree-nested.c: Likewise.
3582 * tree-nrv.c: Likewise.
3583 * tree-object-size.c: Likewise.
3584 * tree-outof-ssa.c: Likewise.
3585 * tree-parloops.c: Likewise.
3586 * tree-phinodes.c: Likewise.
3587 * tree-predcom.c: Likewise.
3588 * tree-pretty-print.c: Likewise.
3589 * tree-profile.c: Likewise.
3590 * tree-scalar-evolution.c: Likewise.
3591 * tree-sra.c: Likewise.
3592 * tree-ssa-address.c: Likewise.
3593 * tree-ssa-alias.c: Likewise.
3594 * tree-ssa-ccp.c: Likewise.
3595 * tree-ssa-coalesce.c: Likewise.
3596 * tree-ssa-copy.c: Likewise.
3597 * tree-ssa-copyrename.c: Likewise.
3598 * tree-ssa-dce.c: Likewise.
3599 * tree-ssa-dom.c: Likewise.
3600 * tree-ssa-dse.c: Likewise.
3601 * tree-ssa-forwprop.c: Likewise.
3602 * tree-ssa-ifcombine.c: Likewise.
3603 * tree-ssa-live.c: Likewise.
3604 * tree-ssa-loop-ch.c: Likewise.
3605 * tree-ssa-loop-im.c: Likewise.
3606 * tree-ssa-loop-ivcanon.c: Likewise.
3607 * tree-ssa-loop-ivopts.c: Likewise.
3608 * tree-ssa-loop-manip.c: Likewise.
3609 * tree-ssa-loop-niter.c: Likewise.
3610 * tree-ssa-loop-prefetch.c: Likewise.
3611 * tree-ssa-loop-unswitch.c: Likewise.
3612 * tree-ssa-loop.c: Likewise.
3613 * tree-ssa-math-opts.c: Likewise.
3614 * tree-ssa-operands.c: Likewise.
3615 * tree-ssa-phiopt.c: Likewise.
3616 * tree-ssa-phiprop.c: Likewise.
3617 * tree-ssa-pre.c: Likewise.
3618 * tree-ssa-propagate.c: Likewise.
3619 * tree-ssa-reassoc.c: Likewise.
3620 * tree-ssa-sccvn.c: Likewise.
3621 * tree-ssa-sink.c: Likewise.
3622 * tree-ssa-strlen.c: Likewise.
3623 * tree-ssa-structalias.c: Likewise.
3624 * tree-ssa-tail-merge.c: Likewise.
3625 * tree-ssa-ter.c: Likewise.
3626 * tree-ssa-threadedge.c: Likewise.
3627 * tree-ssa-threadupdate.c: Likewise.
3628 * tree-ssa-uncprop.c: Likewise.
3629 * tree-ssa-uninit.c: Likewise.
3630 * tree-ssa.c: Likewise.
3631 * tree-ssanames.c: Likewise.
3632 * tree-stdarg.c: Likewise.
3633 * tree-streamer-in.c: Likewise.
3634 * tree-switch-conversion.c: Likewise.
3635 * tree-tailcall.c: Likewise.
3636 * tree-vect-data-refs.c: Likewise.
3637 * tree-vect-generic.c: Likewise.
3638 * tree-vect-loop-manip.c: Likewise.
3639 * tree-vect-loop.c: Likewise.
3640 * tree-vect-patterns.c: Likewise.
3641 * tree-vect-slp.c: Likewise.
3642 * tree-vect-stmts.c: Likewise.
3643 * tree-vectorizer.c: Likewise.
3644 * tree-vrp.c: Likewise.
3647 * value-prof.c: Likewise.
3648 * var-tracking.c: Likewise.
3649 * varpool.c: Likewise.
3650 * vtable-verify.c: Likewise.
3652 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3654 * config/tilegx/tilegx.c: Include "tree.h".
3656 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3658 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
3659 the info, not after it.
3660 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
3661 dump_ssaname_info call.
3662 (pp_gimple_stmt_1): Adjust caller.
3663 (dump_phi_nodes): Likewise. Don't print "# " here.
3665 2013-10-22 Jan Hubicka <jh@suse.cz>
3667 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
3669 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
3670 * i386.c (expand_small_movmem_or_setmem): New function.
3671 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
3673 (alg_usable_p): Add support for value ranges; cleanup.
3674 (ix86_expand_set_or_movmem): Add support for misaligned moves.
3676 2013-10-22 Sterling Augustine <saugustine@google.com>
3678 * doc/invoke.texi: Document -ggnu-pubnames.
3679 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
3681 * dwarf2out.c: Include gdb/gdb-index.h.
3682 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
3683 debug_generate_pub_sections.
3684 (is_java, output_pubtables, output_pubname): New functions.
3685 (include_pubname_in_output): Handle debug_generate_pub_sections at
3687 (size_of_pubnames): Use new local space_for_flags based on
3688 debug_generate_pub_sections.
3689 (output_pubnames): Unify pubnames and pubtypes output logic.
3690 Genericize comments. Call output_pubname.
3691 (dwarf2out_finish): Move logic to output_pubnames and call it.
3693 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
3696 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
3697 Remove CCCmode handling.
3698 <case LTU>: Return 'c' suffix for CCCmode.
3699 <case GEU>: Return 'nc' suffix for CCCmode.
3700 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
3701 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
3702 (*sub<mode>3_cc_overflow): Ditto.
3703 (*subsi3_zext_cc_overflow): Ditto.
3705 2013-10-22 Steve Ellcey <sellcey@mips.com>
3707 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
3708 (AND (NOT OP1) (NOT OP2)).
3710 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
3712 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3713 meaning of merge-high and merge-low masks for little endian; avoid
3714 use of vector-pack masks for little endian for mismatched modes.
3716 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3718 * config/tilepro/tilepro.c: Include "tree.h".
3720 2013-10-22 Andreas Schwab <schwab@suse.de>
3722 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
3725 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
3727 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
3728 [Ed Smith-Rowland]): New entries.
3729 ([Stephen M. Webb]): Update.
3731 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3733 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3735 2013-10-21 Marek Polacek <polacek@redhat.com>
3738 * fold-const.c (fold_range_test): Return 0 if the type is not
3741 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3743 * system.h: Move hwint.h include further down.
3744 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3745 gcc_checking_asserts.
3746 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3748 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3750 Fix volatile issues in optimize_bit_field_compare.
3751 * fold-const.c (optimize_bit_field_compare): Bail out if
3752 lvolatilep or rvolatilep.
3754 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3756 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3757 and get_inner_reference returning different pmode for non-volatile
3758 bit-field members dependent on flag_strict_volatile_bitfields.
3759 * stor-layout.c (layout_decl): Remove special handling of
3760 flag_strict_volatile_bitfields.
3761 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3762 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3764 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3766 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3769 2013-10-21 Jeff Law <law@redhat.com>
3771 * tree-ssa-threadedge.c (thread_through_normal_block): New
3772 argument VISITED. Remove VISISTED as a local variable. When we
3773 have a threadable jump, verify the destination of the jump has not
3775 (thread_across_edge): Allocate VISITED bitmap once at function
3776 scope and use it throughout. Make sure to set appropriate bits in
3777 VISITED for E (start of jump thread path).
3778 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3779 through a joiner if any edge on the path has a recorded jump thread.
3781 2013-10-21 Ian Lance Taylor <iant@google.com>
3783 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3784 don't imply that attributes can solve all problems.
3785 (Directory Options): Fix typo.
3787 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3789 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3790 extend and extend_add.
3792 2013-10-21 Richard Biener <rguenther@suse.de>
3794 PR tree-optimization/58794
3795 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3796 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3798 2013-10-21 Richard Biener <rguenther@suse.de>
3801 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3802 to (T) X for sign-changing conversions (or no conversion).
3804 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3806 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3808 2013-10-20 Jan Hubicka <jh@suse.cz>
3810 * config/i386/i386-tune.def: Add comment; organize into categories
3812 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3814 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3815 argument. Update function comment.
3816 (expand_set_or_movmem_via_rep): New function combining
3817 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3818 (expand_movmem_via_rep_mov): Remove.
3819 expand_setmem_via_rep_stos): Remove.
3820 (expand_movmem_epilogue): Update calls correspondingly.
3821 (expand_setmem_epilogue_via_loop): Likewise.
3823 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3824 (expand_set_or_movmem_prologue): New function combining
3825 expand_movmem_prologue and expand_setmem_prologue.
3826 (expand_movmem_prologue): Remove.
3827 (expand_setmem_prologue): Remove.
3828 (expand_set_or_movmem_constant_prologue): New function combining
3829 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3830 (expand_constant_movmem_prologue): Remove.
3831 (expand_constant_setmem_prologue): Remove.
3832 (promote_duplicated_reg): Allow vector-const0 value.
3833 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3834 and ix86_expand_setmem.
3835 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3836 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3838 2013-10-21 Diego Novillo <dnovillo@google.com>
3840 * asan.c: Include tree.h
3841 * bb-reorder.c: Likewise.
3842 * cfgcleanup.c: Likewise.
3843 * cfgloopmanip.c: Likewise.
3844 * data-streamer-in.c: Likewise.
3845 * data-streamer-out.c: Likewise.
3846 * data-streamer.c: Likewise.
3847 * dwarf2cfi.c: Likewise.
3848 * graphite-blocking.c: Likewise.
3849 * graphite-clast-to-gimple.c: Likewise.
3850 * graphite-dependences.c: Likewise.
3851 * graphite-interchange.c: Likewise.
3852 * graphite-optimize-isl.c: Likewise.
3853 * graphite-poly.c: Likewise.
3854 * graphite-scop-detection.c: Likewise.
3855 * graphite-sese-to-poly.c: Likewise.
3856 * graphite.c: Likewise.
3857 * ipa-devirt.c: Likewise.
3858 * ipa-profile.c: Likewise.
3861 * loop-init.c: Likewise.
3862 * loop-unroll.c: Likewise.
3863 * lower-subreg.c: Likewise.
3864 * lto/lto-object.c: Likewise.
3865 * recog.c: Likewise.
3866 * reginfo.c: Likewise.
3867 * tree-loop-distribution.c: Likewise.
3868 * tree-parloops.c: Likewise.
3869 * tree-ssa-strlen.c: Likewise.
3870 * tree-streamer.c: Likewise.
3871 * value-prof.c: Likewise.
3872 * target-globals.c: Likewise.
3873 * expr.h: Include tree-core.h instead of tree.h.
3874 * gimple.h: Likewise.
3875 * ipa-prop.h: Likewise.
3876 * ipa-utils.h: Likewise.
3877 * lto-streamer.h: Likewise.
3878 * streamer-hooks.h: Likewise.
3879 * ipa-reference.h: Include cgraph.h instead of tree.h.
3880 * cgraph.h: Include basic-block.h instead of tree.h.
3881 * tree-streamer.h: Do not include tree.h.
3882 * genattrtab.c (write_header): Generate inclusion of
3884 * genautomata.c (main): Likewise.
3885 * genemit.c: Likewise.
3886 * genopinit.c: Likewise.
3887 * genoutput.c (output_prologue): Likewise.
3888 * genpeep.c: Likewise.
3890 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3892 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3894 (vec_unpacku_hi_v8hi): Likewise.
3895 (vec_unpacku_lo_v16qi): Likewise.
3896 (vec_unpacku_lo_v8hi): Likewise.
3898 2013-10-20 Jan Hubicka <jh@suse.cz>
3900 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3901 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3902 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3904 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3906 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3907 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3909 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3911 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3914 * system.h: Undef m_slot.
3916 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3918 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3919 all elements for both endian flavors.
3921 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3924 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3925 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3926 and SI_REG for 64bit SYSV ABI targets.
3928 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3930 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3931 to multi_reg_return. Clarify that we are skipping USEs of multiple
3932 return registers. Use bool type where appropriate.
3934 2013-10-18 Jan Hubicka <jh@suse.cz>
3936 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3938 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3939 (X86_TUNE_PUSH_MEMORY): Likewise.
3940 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3941 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3942 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3944 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3945 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3947 (ix86_option_override_internal): Update to use tune features instead
3950 2013-10-18 Cong Hou <congh@google.com>
3952 PR tree-optimization/58508
3953 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3954 statement that contains data refs with zero-step.
3956 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3958 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3959 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3960 * gimple-low.c (gimple_check_call_arg,
3961 gimple_check_call_matching_types): Move to cgraph.c.
3962 * gimple-low.h: Remove prototype.
3963 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3964 Relocate from gimple-low.c.
3965 * cgraph.h: Add prototype. Don't include basic-block.h.
3966 * gimplify.c: Add gimple-low to include list.
3967 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3968 * tree-eh.c: Add gimple-low to include list.
3969 * tree-nested.c: Likewise.
3970 * cfgexpand.c: Add tree-ssa-address.h to include list.
3972 * gimple-fold.c: Likewise.
3973 * gimple-ssa-strength-reduction.c: Likewise.
3974 * trans-mem.c: Likewise.
3975 * tree-mudflap.c: Likewise.
3976 * tree-ssa-loop-ivopts.c: Likewise.
3977 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3978 (degenerate_phi_result): Move to tree-phinodes.c.
3979 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3980 * tree-ssa-threadedge.c: Likewise.
3981 * tree-vrp.c: Likewise.
3982 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3983 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3984 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3985 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3986 * tree-ssa-forwprop.c: Likewise.
3987 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3988 pass_data_cleanup_cfg_post_optimizing,
3989 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3990 * tree-optimize.c: Delete File.
3991 * graphite.c: Include tree-cfgcleanup.h.
3992 * passes.c: Likewise.
3993 * tree-cfg.c: Likewise.
3994 * tree-profile.c: Likewise.
3995 * tree-ssa-dse.c: Likewise.
3996 * tree-ssa-loop-ivcanon.c: Likewise.
3997 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3998 * tree-outof-ssa.c: Include sbitmap.h.
3999 * tree-ssa-live.c: Likewise.
4000 * tree-ssa-propagate.c: Likewise.
4001 * tree-ssa-structalias.c: Likewise.
4002 * tree-stdarg.c: Likewise.
4003 * Makefile.in (OBJS): Delete tree-optimize.o.
4004 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
4005 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
4006 * varasm.c: Include basic-block.h.
4007 * cfgloop.h: Include function.h instead of basic-block.h
4008 (bb_loop_depth): Move to cfgloop.c.
4009 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
4011 2013-10-18 Teresa Johnson <tejohnson@google.com>
4013 * predict.c (probably_never_executed): Compare frequency-based
4014 count to number of training runs.
4015 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
4017 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4019 * config/arm/arm.c (cortexa9_extra_costs): New table.
4020 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
4022 2013-10-18 Jeff Law <law@redhat.com>
4024 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
4026 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
4027 (register_jump_thread): Add "registered_jump_thread" debug
4029 * dbgcnt.def (registered_jump_thread): New debug counter.
4031 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4033 * config/rs6000/rs6000.c: Include cgraph.h.
4035 2013-10-18 Teresa Johnson <tejohnson@google.com>
4037 * tree-ssa-tail-merge.c (replace_block_by): Update edge
4038 weights during merging.
4040 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
4042 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
4043 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
4044 * tree-ssa-operands.h: Remove prototype.
4045 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
4046 * gimple.c (virtual_operand_p): Relocate from gimple.c.
4047 * gimple.h: Add prototype.
4048 * gimple-ssa.h: Include tree-ssa-operands.h.
4049 * tree-dump.c: Add tree-cfg.h to include list.
4050 * tree-ssa-alias.c: Add ipa-reference.h to include list.
4051 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
4052 * config/i386/i386.c: Don't include tree-flow.h.
4053 * config/rs6000/rs6000.c: Likewise.
4055 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4057 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
4059 2013-10-18 Richard Biener <rguenther@suse.de>
4061 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
4062 or TYPE_UNSIGNED of integral types.
4063 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
4064 NULL_TREE for zero-precision integral types.
4066 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
4068 * config/aarch64/arm_neon.h
4069 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
4071 2013-10-17 Sriraman Tallam <tmsriram@google.com>
4074 * opth-gen.awk: Define target_flags_explicit.
4076 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4078 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
4079 fields to the reg_addr array that describes the valid addressing
4080 mode for any register, general purpose registers, floating point
4081 registers, and Altivec registers.
4082 (FIRST_RELOAD_REG_CLASS): Likewise.
4083 (LAST_RELOAD_REG_CLASS): Likewise.
4084 (struct reload_reg_map_type): Likewise.
4085 (reload_reg_map_type): Likewise.
4086 (RELOAD_REG_VALID): Likewise.
4087 (RELOAD_REG_MULTIPLE): Likewise.
4088 (RELOAD_REG_INDEXED): Likewise.
4089 (RELOAD_REG_OFFSET): Likewise.
4090 (RELOAD_REG_PRE_INCDEC): Likewise.
4091 (RELOAD_REG_PRE_MODIFY): Likewise.
4092 (reg_addr): Likewise.
4093 (mode_supports_pre_incdec_p): New helper functions to say whether
4094 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
4095 (mode_supports_pre_modify_p): Likewise.
4096 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
4097 print the valid address mode bits for each mode.
4098 (rs6000_debug_print_mode): Likewise.
4099 (rs6000_debug_reg_global): Likewise.
4100 (rs6000_setup_reg_addr_masks): New function to set up the address
4101 mask bits for each type.
4102 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
4103 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
4104 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
4105 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
4106 PRE_MODIFY are supported.
4107 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
4108 registers, instead of {src,dest}_av_p.
4109 (rs6000_print_options_internal): Tweak the debug output slightly.
4111 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
4113 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
4116 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
4118 * tree-flow.h (struct omp_region): Move to omp-low.c.
4119 Remove omp_ prototypes and variables.
4120 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
4121 (copy_var_decl): Relocate prototype from tree-flow.h.
4122 * gimple.c (copy_var_decl): Relocate from omp-low.c.
4123 * tree.h: Move prototype to omp-low.h.
4124 * omp-low.h: New File. Relocate prototypes here.
4125 * omp-low.c (struct omp_region): Make local here.
4126 (root_omp_region): Make static.
4127 (copy_var_decl) Move to gimple.c.
4128 (new_omp_region): Make static.
4129 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
4130 * tree-cfg.c: Include omp-low.h.
4131 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
4132 * gimplify.c: Include omp-low.h.
4133 * tree-parloops.c: Likewise.
4135 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
4137 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
4138 operands are in memory, prefer to force non-matched operand 1 to
4141 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
4144 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
4145 restrict TImode addresses to single indirect registers if both
4146 -mquad-memory and -mvsx-timode are used.
4147 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
4148 we should emit load/store quad. Remove using %y for quad memory
4151 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
4152 constraints to allow load/store quad on machines where TImode is
4153 not allowed in VSX registers. Use 'n' instead of 'F' constraint
4154 for TImode to load integer constants.
4156 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4158 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
4160 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
4162 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
4163 handling of STACK_REG.
4165 2013-10-17 Richard Biener <rguenther@suse.de>
4167 PR tree-optimization/58143
4168 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
4170 (rewrite_to_defined_overflow): Likewise.
4171 (move_computations_dom_walker::before_dom): Rewrite stmts
4172 with undefined signed overflow that are not always executed
4173 into unsigned arithmetic.
4175 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
4178 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
4179 explicit isa flag to be an options variable, instead of using
4180 global_options_set. Remove define from rs6000.h.
4181 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
4183 * config/rs6000/rs6000.c (rs6000_option_override_internal):
4184 Initialize rs6000_isa_flags_explicit.
4185 (rs6000_function_specific_save): Add gcc_options* parameter, so
4186 that the powerpc builds after the 2013-10-15 changes.
4187 (rs6000_function_specific_restore): Likewise.
4189 2013-10-16 DJ Delorie <dj@redhat.com>
4191 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
4192 op is a REG before checking REGNO.
4193 (rl78_alloc_physical_registers): Verify pattern is a SET before
4196 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4198 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
4200 (vec_unpacks_lo_v4sf): Likewise.
4201 (vec_unpacks_float_hi_v4si): Likewise.
4202 (vec_unpacks_float_lo_v4si): Likewise.
4203 (vec_unpacku_float_hi_v4si): Likewise.
4204 (vec_unpacku_float_lo_v4si): Likewise.
4206 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4208 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
4209 (vsx_concat_v2sf): Likewise.
4211 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
4213 * config/aarch64/aarch64.md
4214 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
4216 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
4218 PR tree-optimization/58697
4219 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
4220 estimated_loop_iterations_int.
4221 (max_stmt_executions_int): Call get_max_loop_iterations_int.
4222 (get_max_loop_iterations_int): New. HWINT version of
4223 get_max_loop_iterations.
4224 * cfgloop.h: Add prototypes.
4225 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
4226 * loop-unroll.c (decide_peel_once_rolling): Call
4227 get_estimated_loop_iterations_int.
4228 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
4229 * tree-ssa-loop-niter.h: Tweak prototypes.
4231 2013-10-16 David Malcolm <dmalcolm@redhat.com>
4233 * gengtype-parse.c (struct_field_seq): Ignore access-control
4234 keywords ("public:" etc).
4236 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
4238 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
4239 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
4241 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
4243 * config/arm/arm.opt (mlra): New option.
4244 * config/arm/arm.c (arm_lra_p): New function.
4245 (TARGET_LRA_P): Define.
4247 2013-10-16 Paulo Matos <pmatos@broadcom.com>
4249 * tree-core.h (tree_code_name): Remove.
4250 * tree.h (get_tree_code_name): New prototype.
4251 * tree.c (tree_code_name): Make static.
4252 (get_tree_code_name): New function.
4253 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
4254 tree_class_check_failed, tree_range_check_failed,
4255 tree_not_class_check_failed, omp_clause_check_failed,
4256 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
4257 wrapper get_tree_code_name instead of calling tree_code_name directly.
4258 * tree-vrp.c (dump_asserts_for): Likewise.
4259 * tree-dump.c (dequeue_and_dump): Likewise.
4260 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
4261 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
4262 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
4263 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
4264 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
4265 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
4266 dump_gimple_omp_for): Likewise.
4267 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
4268 * tree-ssa-pre.c (print_pre_expr): Likewise.
4269 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
4270 * print-tree.c (print_node_brief, print_node): Likewise.
4271 * gimple.c (gimple_check_failed): Likewise.
4272 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
4273 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
4274 * config/mep/mep.c (mep_validate_vliw): Likewise.
4275 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
4276 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
4278 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
4280 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
4283 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
4285 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
4286 (MULTILIB_MATCHES): Add multilib for -march=v8.
4288 2013-10-15 Sriraman Tallam <tmsriram@google.com>
4291 * optc-save-gen.awk: Add extra parameter to the save and restore
4293 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
4294 * tree.c (build_optimization_node): New parameter. Add extra parameter
4295 to call to cl_optimization_save.
4296 (build_target_option_node): New parameter. Add extra parameter
4297 to call to cl_target_option_save.
4298 * tree.h (build_optimization_node): New parameter.
4299 (build_target_option_node): New parameter.
4300 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
4301 build_optimization_node and build_target_option_node.
4302 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
4303 (handle_pragma_push_options): Ditto.
4304 * toplev.c (process_options): Ditto.
4305 * opts.c (init_options_struct): Check for opts_set non-null.
4306 * target.def (target_option.save): New parameter.
4307 (target_option.restore): New parameter.
4308 * tm.texi: Generate.
4309 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
4310 (ix86_pragma_target_parse): Ditto.
4311 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
4313 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
4314 to build_optimization_node and build_target_option_node.
4315 (rs6000_valid_attribute_p): Ditto.
4316 (rs6000_pragma_target_parse): Ditto.
4317 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
4319 * config/i386/i386.h:
4320 TARGET_64BIT_P: New Macro
4321 TARGET_MMX_P: New Macro.
4322 TARGET_3DNOW_P: New Macro.
4323 TARGET_3DNOW_A_P: New Macro.
4324 TARGET_SSE_P: New Macro.
4325 TARGET_SSE2_P: New Macro.
4326 TARGET_SSE3_P: New Macro.
4327 TARGET_SSSE3_P: New Macro.
4328 TARGET_SSE4_1_P: New Macro.
4329 TARGET_SSE4_2_P: New Macro.
4330 TARGET_AVX_P: New Macro.
4331 TARGET_AVX2_P: New Macro.
4332 TARGET_AVX512F_P: New Macro.
4333 TARGET_AVX512PF_P: New Macro.
4334 TARGET_AVX512ER_P: New Macro.
4335 TARGET_AVX512CD_P: New Macro.
4336 TARGET_FMA_P: New Macro.
4337 TARGET_SSE4A_P: New Macro.
4338 TARGET_FMA4_P: New Macro.
4339 TARGET_XOP_P: New Macro.
4340 TARGET_LWP_P: New Macro.
4341 TARGET_ABM_P: New Macro.
4342 TARGET_BMI_P: New Macro.
4343 TARGET_BMI2_P: New Macro.
4344 TARGET_LZCNT_P: New Macro.
4345 TARGET_TBM_P: New Macro.
4346 TARGET_POPCNT_P: New Macro.
4347 TARGET_SAHF_P: New Macro.
4348 TARGET_MOVBE_P: New Macro.
4349 TARGET_CRC32_P: New Macro.
4350 TARGET_AES_P: New Macro.
4351 TARGET_PCLMUL_P: New Macro.
4352 TARGET_CMPXCHG16B_P: New Macro.
4353 TARGET_FSGSBASE_P: New Macro.
4354 TARGET_RDRND_P: New Macro.
4355 TARGET_F16C_P: New Macro.
4356 TARGET_RTM_P: New Macro.
4357 TARGET_HLE_P: New Macro.
4358 TARGET_RDSEED_P: New Macro.
4359 TARGET_PRFCHW_P: New Macro.
4360 TARGET_ADX_P: New Macro.
4361 TARGET_FXSR_P: New Macro.
4362 TARGET_XSAVE_P: New Macro.
4363 TARGET_XSAVEOPT_P: New Macro.
4364 TARGET_LP64_P: New Macro.
4365 TARGET_X32_P: New Macro.
4366 TARGET_FPMATH_DEFAULT_P: New Macro.
4367 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
4368 * config/i386/i386.c (ix86_option_override_internal): New parameters.
4370 Change ix86_tune_string to access opts->x_ix86_tune_string.
4371 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
4372 Change ix86_arch_string to access opts->x_ix86_arch_string.
4373 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
4374 Change ix86_pmode to access opts->x_ix86_pmode.
4375 Change ix86_abi to access opts->x_ix86_abi.
4376 Change ix86_cmodel to access opts->x_ix86_cmodel.
4377 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
4378 Change ix86_isa_flags_explicit to access
4379 opts->x_ix86_isa_flags_explicit.
4380 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
4381 Change ix86_regparm to access opts->x_ix86_regparm.
4382 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
4383 Change ix86_preferred_stack_boundary_arg to access
4384 opts->x_ix86_preferred_stack_boundary_arg.
4385 Change ix86_force_align_arg_pointer to access
4386 opts->x_ix86_force_align_arg_pointer.
4387 Change ix86_incoming_stack_boundar_arg to access
4388 opts->x_ix86_incoming_stack_boundar_arg.
4389 Change ix86_fpmath to access opts->x_ix86_fpmath.
4390 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
4391 Change ix86_recip_name to access opts->x_ix86_recip_name.
4392 Change ix86_stack_protector_guard to access
4393 opts->x_ix86_stack_protector_guard.
4394 Change ix86_tune_memcpy_strategy to access
4395 opts->x_ix86_tune_memcpy_strategy.
4396 Change ix86_tune_memset_strategy to access
4397 opts->x_ix86_tune_memset_strategy.
4398 Change global_options to access opts.
4399 Change global_options_set to access opts_set.
4400 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
4401 Change TARGET_MMX to TARGET_MMX_P (opts->...).
4402 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
4403 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
4404 Change TARGET_SSE to TARGET_SSE_P (opts->...).
4405 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
4406 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
4407 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
4408 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
4409 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
4410 Change TARGET_AVX to TARGET_AVX_P (opts->...).
4411 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
4412 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
4413 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
4414 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
4415 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
4416 Change TARGET_FMA to TARGET_FMA_P (opts->...).
4417 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
4418 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
4419 Change TARGET_XOP to TARGET_XOP_P (opts->...).
4420 Change TARGET_LWP to TARGET_LWP_P (opts->...).
4421 Change TARGET_ABM to TARGET_ABM_P (opts->...).
4422 Change TARGET_BMI to TARGET_BMI_P (opts->...).
4423 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
4424 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
4425 Change TARGET_TBM to TARGET_TBM_P (opts->...).
4426 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
4427 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
4428 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
4429 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
4430 Change TARGET_AES to TARGET_AES_P (opts->...).
4431 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
4432 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
4433 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
4434 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
4435 Change TARGET_F16C to TARGET_F16C_P (opts->...).
4436 Change TARGET_RTM to TARGET_RTM_P (opts->...).
4437 Change TARGET_HLE to TARGET_HLE_P (opts->...).
4438 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
4439 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
4440 Change TARGET_ADX to TARGET_ADX_P (opts->...).
4441 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
4442 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
4443 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
4444 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
4445 Change TARGET_X32 to TARGET_X32_P (opts->...).
4446 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
4447 Change TARGET_FLOAT_RETURNS_IN_80387 to
4448 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
4449 (ix86_function_specific_save): New parameter. Use opts-> fields
4450 to replace global fields.
4451 (ix86_function_specific_restore): Ditto.
4452 (ix86_valid_target_attribute_inner_p): New parameters.
4454 Fix call to ix86_handle_option and set_option.
4455 (ix86_valid_target_attribute_tree): New parameters.
4456 Change global_options to access opts.
4457 Change global_options_set to access opts_set.
4458 Fix call to ix86_valid_target_attribute_inner_p.
4459 Change ix86_tune_string to access opts->x_ix86_tune_string.
4460 Change ix86_arch_string to access opts->x_ix86_arch_string.
4461 Change ix86_fpmath to access opts->x_ix86_fpmath
4462 Fix call to ix86_option_override_internal.
4463 Fix call to ix86_add_new_builtins.
4464 Fix calls to build_optimization_node and build_target_option_node.
4465 (ix86_valid_target_attribute_p): Remove access to global_options.
4466 Use new gcc_options structure func_options.
4467 Fix call to ix86_valid_target_attribute_tree.
4468 Fix call to build_optimization_node.
4469 (get_builtin_code_for_version): Fix call to
4470 ix86_valid_target_attribute_tree.
4472 2013-10-15 David Malcolm <dmalcolm@redhat.com>
4474 * Makefile.in (PICFLAG): New.
4475 (enable_host_shared): New.
4476 (INTERNAL_CFLAGS): Use PICFLAG.
4477 (LIBIBERTY): Use pic build of libiberty.a if configured with
4478 --enable-host-shared.
4479 * configure.ac: Add --enable-host-shared, setting up new
4481 * configure: Regenerate.
4482 * doc/install.texi (--enable-shared): Add note contrasting it with ...
4483 (--enable-host-shared): New option.
4485 2013-10-15 Richard Biener <rguenther@suse.de>
4487 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
4488 for built-in functions.
4490 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
4492 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
4493 (optimize_range_tests_1): New function,
4494 extracted from optimize_range_tests.
4495 (optimize_range_tests_xor): Similarly.
4496 (optimize_range_tests_diff): New function.
4497 (optimize_range_tests): Use optimize_range_tests_1.
4499 2013-10-15 Cong Hou <congh@google.com>
4501 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
4502 requirement of the reduction pattern so that one operand of the
4503 reduction operation can come from outside of the loop.
4505 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4507 * config/arm/neon-schedgen.ml: Remove.
4508 * config/arm/cortex-a9-neon.md: Remove comment regarding
4511 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4513 * config/arm/types: Remove old neon types.
4515 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4517 * config/arm/cortex-a7.md
4518 (cortex_a7_neon_type): New.
4519 (cortex_a7_neon_mul): Update for new types.
4520 (cortex_a7_neon_mla): Likewise.
4521 (cortex_a7_neon): Likewise.
4523 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4525 * config/arm/cortex-a15-neon.md
4526 (cortex_a15_neon_type): New,
4528 (cortex_a15_neon_int_1): Remove.
4529 (cortex_a15_neon_int_2): Likewise.
4530 (cortex_a15_neon_int_3): Likewise.
4531 (cortex_a15_neon_int_4): Likewise.
4532 (cortex_a15_neon_int_5): Likewise.
4533 (cortex_a15_neon_vqneg_vqabs): Likewise.
4534 (cortex_a15_neon_vmov): Likewise.
4535 (cortex_a15_neon_vaba): Likewise.
4536 (cortex_a15_neon_vaba_qqq): Likewise.
4537 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4538 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4539 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4541 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4542 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4543 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
4544 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4545 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4546 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4547 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4548 (cortex_a15_neon_shift_1): Likewise.
4549 (cortex_a15_neon_shift_2): Likewise.
4550 (cortex_a15_neon_shift_3): Likewise.
4551 (cortex_a15_neon_vshl_ddd): Likewise.
4552 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4553 (cortex_a15_neon_vsra_vrsra): Likewise.
4554 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4555 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4556 (cortex_a15_neon_bp_3cycle): Likewise.
4557 (cortex_a15_neon_ldm_2): Likewise.
4558 (cortex_a15_neon_stm_2): Likewise.
4559 (cortex_a15_neon_mcr): Likewise.
4560 (cortex_a15_neon_mrc): Likewise.
4561 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4562 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4563 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4564 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4565 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4566 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4567 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4568 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4569 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4570 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4571 (cortex_a15_neon_bp_simple): Likewise.
4572 (cortex_a15_neon_bp_2cycle): Likewise.
4573 (cortex_a15_neon_bp_3cycle): Likewise.
4574 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4575 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4576 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4577 (cortex_a15_neon_vld2_4_regs): Likewise.
4578 (cortex_a15_neon_vld3_vld4): Likewise.
4579 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4580 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4581 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
4582 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
4583 (cortex_a15_neon_vst3_vst4): Rename to...
4584 (cortex_a15_neon_vst4): This, update for new attributes.
4585 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
4586 (cortex_a15_neon_vld3_vld4_lane): Likewise.
4587 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4588 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4589 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4590 (cortex_a15_neon_ldm_2): Likewise.
4591 (cortex_a15_neon_stm_2): Likewise.
4592 (cortex_a15_neon_mcr): Likewise.
4593 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4594 (cortex_a15_neon_mrc): Likewise.
4595 (cortex_a15_neon_mrrc): Likewise.
4597 (cortex_a15_neon_abd): New.
4598 (cortex_a15_neon_abd_q): Likewise.
4599 (cortex_a15_neon_aba): Likewise.
4600 (cortex_a15_neon_aba_q): Likewise.
4601 (cortex_a15_neon_acc): Likewise.
4602 (cortex_a15_neon_acc_q): Likewise.
4603 (cortex_a15_neon_arith_basic): Likewise.
4604 (cortex_a15_neon_arith_complex): Likewise.
4605 (cortex_a15_neon_multiply): Likewise.
4606 (cortex_a15_neon_multiply_q): Likewise.
4607 (cortex_a15_neon_mla): Likewise.
4608 (cortex_a15_neon_mla_q): Likewise.
4609 (cortex_a15_neon_sat_mla_long): Likewise.
4610 (cortex_a15_neon_shift_acc): Likewise.
4611 (cortex_a15_neon_shift_imm_basic): Likewise.
4612 (cortex_a15_neon_shift_imm_complex): Likewise.
4613 (cortex_a15_neon_shift_reg_basic): Likewise.
4614 (cortex_a15_neon_shift_reg_basic_q): Likewise.
4615 (cortex_a15_neon_shift_reg_complex): Likewise.
4616 (cortex_a15_neon_shift_reg_complex_q): Likewise.
4617 (cortex_a15_neon_fp_negabs): Likewise
4618 (cortex_a15_neon_fp_arith): Likewise
4619 (cortex_a15_neon_fp_arith_q): Likewise
4620 (cortex_a15_neon_fp_cvt_int): Likewise
4621 (cortex_a15_neon_fp_cvt_int_q): Likewise
4622 (cortex_a15_neon_fp_cvt_16): Likewise
4623 (cortex_a15_neon_fp_mul): Likewise
4624 (cortex_a15_neon_fp_mul_q): Likewise
4625 (cortex_a15_neon_fp_mla): Likewise
4626 (cortex_a15_neon_fp_mla_q): Likewise
4627 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
4628 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
4629 (cortex_a15_neon_bitops): Likewise.
4630 (cortex_a15_neon_bitops_q): Likewise.
4631 (cortex_a15_neon_from_gp): Likewise.
4632 (cortex_a15_neon_from_gp_q): Likewise.
4633 (cortex_a15_neon_tbl3_tbl4): Likewise.
4634 (cortex_a15_neon_zip_q): Likewise.
4635 (cortex_a15_neon_to_gp): Likewise.
4636 (cortex_a15_neon_load_a): Likewise.
4637 (cortex_a15_neon_load_b): Likewise.
4638 (cortex_a15_neon_load_c): Likewise.
4639 (cortex_a15_neon_load_d): Likewise.
4640 (cortex_a15_neon_load_e): Likewise.
4641 (cortex_a15_neon_load_f): Likewise.
4642 (cortex_a15_neon_store_a): Likewise.
4643 (cortex_a15_neon_store_b): Likewise.
4644 (cortex_a15_neon_store_c): Likewise.
4645 (cortex_a15_neon_store_d): Likewise.
4646 (cortex_a15_neon_store_e): Likewise.
4647 (cortex_a15_neon_store_f): Likewise.
4648 (cortex_a15_neon_store_g): Likewise.
4649 (cortex_a15_neon_store_h): Likewise.
4650 (cortex_a15_vfp_to_from_gp): Likewise.
4652 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4654 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
4656 (cortex_a9_neon_vshl_ddd): Remove.
4657 (cortex_a9_neon_vst3_vst4): Likewise.
4658 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4660 (cortex_a9_neon_bit_ops_q): New.
4662 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
4663 (cortex_a9_neon_int_2): Likewise.
4664 (cortex_a9_neon_int_3): Likewise.
4665 (cortex_a9_neon_int_4): Likewise.
4666 (cortex_a9_neon_int_5): Likewise.
4667 (cortex_a9_neon_vqneg_vqabs): Likewise.
4668 (cortex_a9_neon_vmov): Likewise.
4669 (cortex_a9_neon_vaba): Likewise.
4670 (cortex_a9_neon_vaba_qqq): Likewise.
4671 (cortex_a9_neon_shift_1): Likewise.
4672 (cortex_a9_neon_shift_2): Likewise.
4673 (cortex_a9_neon_shift_3): Likewise.
4674 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4675 (cortex_a9_neon_vsra_vrsra): Likewise.
4676 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4677 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4678 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4680 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4681 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4682 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4684 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4685 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4686 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4687 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4688 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4689 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4690 (cortex_a9_neon_fp_vsum): Likewise.
4691 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4692 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4693 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4694 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4695 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4696 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4697 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4698 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4699 (cortex_a9_neon_bp_simple): Likewise.
4700 (cortex_a9_neon_bp_2cycle): Likewise.
4701 (cortex_a9_neon_bp_3cycle): Likewise.
4702 (cortex_a9_neon_ldr): Likewise.
4703 (cortex_a9_neon_str): Likewise.
4704 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4705 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4706 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4707 (cortex_a9_neon_vld2_4_regs): Likewise.
4708 (cortex_a9_neon_vld3_vld4): Likewise.
4709 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4710 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4711 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4712 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4713 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4714 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4715 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4716 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4717 (cortex_a9_neon_mcr): Likewise.
4718 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4719 (cortex_a9_neon_mrc): Likewise.
4720 (cortex_a9_neon_mrrc): Likewise.
4722 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4724 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
4726 (cortex_a8_neon_vshl_ddd): Remove.
4727 (cortex_a8_neon_vst3_vst4): Likewise.
4728 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4730 (cortex_a8_neon_bit_ops_q): New.
4732 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4733 (cortex_a8_neon_int_2): Likewise..
4734 (cortex_a8_neon_int_3): Likewise.
4735 (cortex_a8_neon_int_5): Likewise.
4736 (cortex_a8_neon_vqneg_vqabs): Likewise.
4737 (cortex_a8_neon_int_4): Likewise.
4738 (cortex_a8_neon_vaba): Likewise.
4739 (cortex_a8_neon_vaba_qqq): Likewise.
4740 (cortex_a8_neon_shift_1): Likewise.
4741 (cortex_a8_neon_shift_2): Likewise.
4742 (cortex_a8_neon_shift_3): Likewise.
4743 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4744 (cortex_a8_neon_vsra_vrsra): Likewise.
4745 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4746 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4747 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4749 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4750 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4751 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4753 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4754 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4755 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4756 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4757 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4758 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4759 (cortex_a8_neon_fp_vsum): Likewise.
4760 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4761 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4762 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4763 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4764 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4765 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4766 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4767 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4768 (cortex_a8_neon_bp_simple): Likewise.
4769 (cortex_a8_neon_bp_2cycle): Likewise.
4770 (cortex_a8_neon_bp_3cycle): Likewise.
4771 (cortex_a8_neon_ldr): Likewise.
4772 (cortex_a8_neon_str): Likewise.
4773 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4774 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4775 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4776 (cortex_a8_neon_vld2_4_regs): Likewise.
4777 (cortex_a8_neon_vld3_vld4): Likewise.
4778 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4779 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4780 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4781 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4782 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4783 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4784 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4785 (cortex_a8_neon_mcr): Likewise.
4786 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4787 (cortex_a8_neon_mrc): Likewise.
4788 (cortex_a8_neon_mrrc): Likewise.
4790 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4792 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4794 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4795 (aarch64_simd_dup<mode>): Add "type" attribute.
4796 (aarch64_dup_lane<mode>): Likewise.
4797 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4798 (*aarch64_simd_mov<mode>): Likewise.
4799 (aarch64_simd_mov_from_<mode>low): Likewise.
4800 (aarch64_simd_mov_from_<mode>high): Likewise.
4801 (orn<mode>3): Likewise.
4802 (bic<mode>3): Likewise.
4803 (add<mode>3): Likewise.
4804 (sub<mode>3): Likewise.
4805 (mul<mode>3): Likewise.
4806 (*aarch64_mul3_elt<mode>): Likewise.
4807 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4808 (*aarch64_mul3_elt_to_128df): Likewise.
4809 (*aarch64_mul3_elt_to_64v2df): Likewise.
4810 (neg<mode>2): Likewise.
4811 (abs<mode>2): Likewise.
4812 (abd<mode>_3): Likewise.
4813 (aba<mode>_3): Likewise.
4814 (fabd<mode>_3): Likewise.
4815 (*fabd_scalar<mode>3): Likewise.
4816 (and<mode>3): Likewise.
4817 (ior<mode>3): Likewise.
4818 (xor<mode>3): Likewise.
4819 (one_cmpl<mode>2): Likewise.
4820 (aarch64_simd_vec_set<mode>): Likewise.
4821 (aarch64_simd_lshr<mode>): Likewise.
4822 (aarch64_simd_ashr<mode>): Likewise.
4823 (aarch64_simd_imm_shl<mode>): Likewise.
4824 (aarch64_simd_reg_sshl<mode): Likewise.
4825 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4826 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4827 (aarch64_simd_vec_setv2di): Likewise.
4828 (aarch64_simd_vec_set<mode>): Likewise.
4829 (aarch64_mla<mode>): Likewise.
4830 (*aarch64_mla_elt<mode>): Likewise.
4831 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4832 (aarch64_mls<mode>): Likewise.
4833 (*aarch64_mls_elt<mode>): Likewise.
4834 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4835 (<su><maxmin><mode>3): Likewise.
4836 (move_lo_quad_<mode>): Likewise.
4837 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4838 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4839 (vec_pack_trunc_<mode>): Likewise.
4840 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4841 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4842 (*aarch64_<su>mlal_lo<mode>): Likewise.
4843 (*aarch64_<su>mlal_hi<mode>): Likewise.
4844 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4845 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4846 (*aarch64_<su>mlal<mode>): Likewise.
4847 (*aarch64_<su>mlsl<mode>): Likewise.
4848 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4849 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4850 (add<mode>3): Likewise.
4851 (sub<mode>3): Likewise.
4852 (mul<mode>3): Likewise.
4853 (div<mode>3): Likewise.
4854 (neg<mode>2): Likewise.
4855 (abs<mode>2): Likewise.
4856 (fma<mode>4): Likewise.
4857 (*aarch64_fma4_elt<mode>): Likewise.
4858 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4859 (*aarch64_fma4_elt_to_128df): Likewise.
4860 (*aarch64_fma4_elt_to_64v2df): Likewise.
4861 (fnma<mode>4): Likewise.
4862 (*aarch64_fnma4_elt<mode>): Likewise.
4863 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4864 (*aarch64_fnma4_elt_to_128df): Likewise.
4865 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4866 (<frint_pattern><mode>2): Likewise.
4867 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4868 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4869 (vec_unpacks_lo_v4sf): Likewise.
4870 (aarch64_float_extend_lo_v2df): Likewise.
4871 (vec_unpacks_hi_v4sf): Likewise.
4872 (aarch64_float_truncate_lo_v2sf): Likewise.
4873 (aarch64_float_truncate_hi_v4sf): Likewise.
4874 (aarch64_vmls<mode>): Likewise.
4875 (<su><maxmin><mode>3): Likewise.
4876 (<maxmin_uns><mode>3): Likewise.
4877 (reduc_<sur>plus_<mode>): Likewise.
4878 (reduc_<sur>plus_v2di): Likewise.
4879 (reduc_<sur>plus_v2si): Likewise.
4880 (reduc_<sur>plus_<mode>): Likewise.
4881 (aarch64_addpv4sf): Likewise.
4882 (clz<mode>2): Likewise.
4883 (reduc_<maxmin_uns>_<mode>): Likewise.
4884 (reduc_<maxmin_uns>_v2di): Likewise.
4885 (reduc_<maxmin_uns>_v2si): Likewise.
4886 (reduc_<maxmin_uns>_<mode>): Likewise.
4887 (reduc_<maxmin_uns>_v4sf): Likewise.
4888 (aarch64_simd_bsl<mode>_internal): Likewise.
4889 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4890 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4891 (aarch64_get_lane<mode>): Likewise.
4892 (*aarch64_combinez<mode>): Likewise.
4893 (aarch64_combine<mode>): Likewise.
4894 (aarch64_simd_combine<mode>): Likewise.
4895 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4896 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4897 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4898 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4899 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4900 (aarch64_<sur>h<addsub><mode>): Likewise.
4901 (aarch64_<sur><addsub>hn<mode>): Likewise.
4902 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4903 (aarch64_pmul<mode>): Likewise.
4904 (aarch64_<su_optab><optab><mode>): Likewise.
4905 (aarch64_<sur>qadd<mode>): Likewise.
4906 (aarch64_sqmovun<mode>): Likewise.
4907 (aarch64_<sur>qmovn<mode>): Likewise.
4908 (aarch64_s<optab><mode>): Likewise.
4909 (aarch64_sq<r>dmulh<mode>): Likewise.
4910 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4911 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4912 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4913 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4914 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4915 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4916 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4917 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4918 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4919 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4920 (aarch64_sqdmull<mode>): Likewise.
4921 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4922 (aarch64_sqdmull_n<mode>): Likewise.
4923 (aarch64_sqdmull2<mode>_internal): Likewise.
4924 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4925 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4926 (aarch64_<sur>shl<mode>): Likewise.
4927 (aarch64_<sur>q<r>shl<mode>
4928 (aarch64_<sur>shll_n<mode>): Likewise.
4929 (aarch64_<sur>shll2_n<mode>): Likewise.
4930 (aarch64_<sur>shr_n<mode>): Likewise.
4931 (aarch64_<sur>sra_n<mode>): Likewise.
4932 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4933 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4934 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4935 (aarch64_cm<optab><mode>): Likewise.
4936 (aarch64_cm<optab>di): Likewise.
4937 (aarch64_cm<optab><mode>): Likewise.
4938 (aarch64_cm<optab>di): Likewise.
4939 (aarch64_cmtst<mode>): Likewise.
4940 (aarch64_cmtstdi): Likewise.
4941 (aarch64_cm<optab><mode>): Likewise.
4942 (*aarch64_fac<optab><mode>): Likewise.
4943 (aarch64_addp<mode>): Likewise.
4944 (aarch64_addpdi): Likewise.
4945 (sqrt<mode>2): Likewise.
4946 (vec_load_lanesoi<mode>): Likewise.
4947 (vec_store_lanesoi<mode>): Likewise.
4948 (vec_load_lanesci<mode>): Likewise.
4949 (vec_store_lanesci<mode>): Likewise.
4950 (vec_load_lanesxi<mode>): Likewise.
4951 (vec_store_lanesxi<mode>): Likewise.
4952 (*aarch64_mov<mode>): Likewise.
4953 (aarch64_ld2<mode>_dreg): Likewise.
4954 (aarch64_ld2<mode>_dreg): Likewise.
4955 (aarch64_ld3<mode>_dreg): Likewise.
4956 (aarch64_ld3<mode>_dreg): Likewise.
4957 (aarch64_ld4<mode>_dreg): Likewise.
4958 (aarch64_ld4<mode>_dreg): Likewise.
4959 (aarch64_tbl1<mode>): Likewise.
4960 (aarch64_tbl2v16qi): Likewise.
4961 (aarch64_combinev16qi): Likewise.
4962 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4963 (aarch64_st2<mode>_dreg): Likewise.
4964 (aarch64_st2<mode>_dreg): Likewise.
4965 (aarch64_st3<mode>_dreg): Likewise.
4966 (aarch64_st3<mode>_dreg): Likewise.
4967 (aarch64_st4<mode>_dreg): Likewise.
4968 (aarch64_st4<mode>_dreg): Likewise.
4969 (*aarch64_simd_ld1r<mode>): Likewise.
4970 (aarch64_frecpe<mode>): Likewise.
4971 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4972 (aarch64_frecps<mode>): Likewise.
4974 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4976 * config/arm/iterators.md (V_elem_ch): New.
4978 (VQH_type): Likewise.
4979 * config/arm/arm.md (is_neon_type): New.
4980 (conds): Use is_neon_type.
4981 (anddi3_insn): Update type attribute.
4982 (xordi3_insn): Likewise.
4983 (one_cmpldi2): Likewise.
4984 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4985 * config/arm/neon.md (neon_mov): Update type attribute.
4986 (*movmisalign<mode>_neon_store): Likewise.
4987 (*movmisalign<mode>_neon_load): Likewise.
4988 (vec_set<mode>_internal): Likewise.
4989 (vec_set<mode>_internal): Likewise.
4990 (vec_setv2di_internal): Likewise.
4991 (vec_extract<mode>): Likewise.
4992 (vec_extract<mode>): Likewise.
4993 (vec_extractv2di): Likewise.
4994 (*add<mode>3_neon): Likewise.
4995 (adddi3_neon): Likewise.
4996 (*sub<mode>3_neon): Likewise.
4997 (subdi3_neon): Likewise.
4998 (fma<VCVTF:mode>4): Likewise.
4999 (fma<VCVTF:mode>4_intrinsic): Likewise.
5000 (*fmsub<VCVTF:mode>4): Likewise.
5001 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
5002 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
5003 (ior<mode>3): Likewise.
5004 (and<mode>3): Likewise.
5005 (orn<mode>3_neon): Likewise.
5006 (orndi3_neon): Likewise.
5007 (bic<mode>3_neon): Likewise.
5008 (bicdi3_neon): Likewise.
5009 (xor<mode>3): Likewise.
5010 (one_cmpl<mode>2): Likewise.
5011 (abs<mode>2): Likewise.
5012 (neg<mode>2): Likewise.
5013 (negdi2_neon): Likewise.
5014 (*umin<mode>3_neon): Likewise.
5015 (*umax<mode>3_neon): Likewise.
5016 (*smin<mode>3_neon): Likewise.
5017 (*smax<mode>3_neon): Likewise.
5018 (vashl<mode>3): Likewise.
5019 (vashr<mode>3_imm): Likewise.
5020 (vlshr<mode>3_imm): Likewise.
5021 (ashl<mode>3_signed): Likewise.
5022 (ashl<mode>3_unsigned): Likewise.
5023 (neon_load_count): Likewise.
5024 (ashldi3_neon_noclobber): Likewise.
5025 (ashldi3_neon): Likewise.
5026 (signed_shift_di3_neon): Likewise.
5027 (unsigned_shift_di3_neon): Likewise.
5028 (ashrdi3_neon_imm_noclobber): Likewise.
5029 (lshrdi3_neon_imm_noclobber): Likewise.
5030 (<shift>di3_neon): Likewise.
5031 (widen_ssum<mode>3): Likewise.
5032 (widen_usum<mode>3): Likewise.
5033 (quad_halves_<code>v4si): Likewise.
5034 (quad_halves_<code>v4sf): Likewise.
5035 (quad_halves_<code>v8hi): Likewise.
5036 (quad_halves_<code>v16qi): Likewise.
5037 (reduc_splus_v2di): Likewise.
5038 (neon_vpadd_internal<mode>): Likewise.
5039 (neon_vpsmin<mode>): Likewise.
5040 (neon_vpsmax<mode>): Likewise.
5041 (neon_vpumin<mode>): Likewise.
5042 (neon_vpumax<mode>): Likewise.
5043 (*ss_add<mode>_neon): Likewise.
5044 (*us_add<mode>_neon): Likewise.
5045 (*ss_sub<mode>_neon): Likewise.
5046 (*us_sub<mode>_neon): Likewise.
5047 (neon_vadd<mode>_unspec): Likewise.
5048 (neon_vaddl<mode>): Likewise.
5049 (neon_vaddw<mode>): Likewise.
5050 (neon_vhadd<mode>): Likewise.
5051 (neon_vqadd<mode>): Likewise.
5052 (neon_vaddhn<mode>): Likewise.
5053 (neon_vmul<mode>): Likewise.
5054 (neon_vfms<VCVTF:mode>): Likewise.
5055 (neon_vmlal<mode>): Likewise.
5056 (neon_vmls<mode>): Likewise.
5057 (neon_vmlsl<mode>): Likewise.
5058 (neon_vqdmulh<mode>): Likewise.
5059 (neon_vqdmlal<mode>): Likewise.
5060 (neon_vqdmlsl<mode>): Likewise.
5061 (neon_vmull<mode>): Likewise.
5062 (neon_vqdmull<mode>): Likewise.
5063 (neon_vsub<mode>_unspec): Likewise.
5064 (neon_vsubl<mode>): Likewise.
5065 (neon_vsubw<mode>): Likewise.
5066 (neon_vqsub<mode>): Likewise.
5067 (neon_vhsub<mode>): Likewise.
5068 (neon_vsubhn<mode>): Likewise.
5069 (neon_vceq<mode>): Likewise.
5070 (neon_vcge<mode>): Likewise.
5071 (neon_vcgeu<mode>): Likewise.
5072 (neon_vcgt<mode>): Likewise.
5073 (neon_vcgtu<mode>): Likewise.
5074 (neon_vcle<mode>): Likewise.
5075 (neon_vclt<mode>): Likewise.
5076 (neon_vcage<mode>): Likewise.
5077 (neon_vcagt<mode>): Likewise.
5078 (neon_vtst<mode>): Likewise.
5079 (neon_vabd<mode>): Likewise.
5080 (neon_vabdl<mode>): Likewise.
5081 (neon_vaba<mode>): Likewise.
5082 (neon_vabal<mode>): Likewise.
5083 (neon_vmax<mode>): Likewise.
5084 (neon_vmin<mode>): Likewise.
5085 (neon_vpaddl<mode>): Likewise.
5086 (neon_vpadal<mode>): Likewise.
5087 (neon_vpmax<mode>): Likewise.
5088 (neon_vpmin<mode>): Likewise.
5089 (neon_vrecps<mode>): Likewise.
5090 (neon_vrsqrts<mode>): Likewise.
5091 (neon_vqabs<mode>): Likewise.
5092 (neon_vqneg<mode>): Likewise.
5093 (neon_vcls<mode>): Likewise.
5094 (clz<mode>2): Likewise.
5095 (popcount<mode>2): Likewise.
5096 (neon_vrecpe<mode>): Likewise.
5097 (neon_vrsqrte<mode>): Likewise.
5098 (neon_vget_lane<mode>_sext_internal): Likewise.
5099 (neon_vget_lane<mode>_zext_internal): Likewise.
5100 (neon_vdup_n<mode>): Likewise.
5101 (neon_vdup_n<mode>): Likewise.
5102 (neon_vdup_nv2di): Likewise.
5103 (neon_vdup_lane<mode>_interal): Likewise.
5104 (*neon_vswp<mode>): Likewise.
5105 (neon_vcombine<mode>): Likewise.
5106 (float<mode><V_cvtto>2): Likewise.
5107 (floatuns<mode><V_cvtto>2): Likewise.
5108 (fix_trunc<mode><V_cvtto>2): Likewise.
5109 (fixuns_trunc<mode><V_cvtto>2
5110 (neon_vcvt<mode>): Likewise.
5111 (neon_vcvt<mode>): Likewise.
5112 (neon_vcvtv4sfv4hf): Likewise.
5113 (neon_vcvtv4hfv4sf): Likewise.
5114 (neon_vcvt_n<mode>): Likewise.
5115 (neon_vcvt_n<mode>): Likewise.
5116 (neon_vmovn<mode>): Likewise.
5117 (neon_vqmovn<mode>): Likewise.
5118 (neon_vqmovun<mode>): Likewise.
5119 (neon_vmovl<mode>): Likewise.
5120 (neon_vmul_lane<mode>): Likewise.
5121 (neon_vmul_lane<mode>): Likewise.
5122 (neon_vmull_lane<mode>): Likewise.
5123 (neon_vqdmull_lane<mode>): Likewise.
5124 (neon_vqdmulh_lane<mode>): Likewise.
5125 (neon_vqdmulh_lane<mode>): Likewise.
5126 (neon_vmla_lane<mode>): Likewise.
5127 (neon_vmla_lane<mode>): Likewise.
5128 (neon_vmlal_lane<mode>): Likewise.
5129 (neon_vqdmlal_lane<mode>): Likewise.
5130 (neon_vmls_lane<mode>): Likewise.
5131 (neon_vmls_lane<mode>): Likewise.
5132 (neon_vmlsl_lane<mode>): Likewise.
5133 (neon_vqdmlsl_lane<mode>): Likewise.
5134 (neon_vext<mode>): Likewise.
5135 (neon_vrev64<mode>): Likewise.
5136 (neon_vrev32<mode>): Likewise.
5137 (neon_vrev16<mode>): Likewise.
5138 (neon_vbsl<mode>_internal): Likewise.
5139 (neon_vshl<mode>): Likewise.
5140 (neon_vqshl<mode>): Likewise.
5141 (neon_vshr_n<mode>): Likewise.
5142 (neon_vshrn_n<mode>): Likewise.
5143 (neon_vqshrn_n<mode>): Likewise.
5144 (neon_vqshrun_n<mode>): Likewise.
5145 (neon_vshl_n<mode>): Likewise.
5146 (neon_vqshl_n<mode>): Likewise.
5147 (neon_vqshlu_n<mode>): Likewise.
5148 (neon_vshll_n<mode>): Likewise.
5149 (neon_vsra_n<mode>): Likewise.
5150 (neon_vsri_n<mode>): Likewise.
5151 (neon_vsli_n<mode>): Likewise.
5152 (neon_vtbl1v8qi): Likewise.
5153 (neon_vtbl2v8qi): Likewise.
5154 (neon_vtbl3v8qi): Likewise.
5155 (neon_vtbl4v8qi): Likewise.
5156 (neon_vtbl1v16qi): Likewise.
5157 (neon_vtbl2v16qi): Likewise.
5158 (neon_vcombinev16qi): Likewise.
5159 (neon_vtbx1v8qi): Likewise.
5160 (neon_vtbx2v8qi): Likewise.
5161 (neon_vtbx3v8qi): Likewise.
5162 (neon_vtbx4v8qi): Likewise.
5163 (*neon_vtrn<mode>_insn): Likewise.
5164 (*neon_vzip<mode>_insn): Likewise.
5165 (*neon_vuzp<mode>_insn): Likewise.
5166 (neon_vld1<mode>): Likewise.
5167 (neon_vld1_lane<mode>): Likewise.
5168 (neon_vld1_lane<mode>): Likewise.
5169 (neon_vld1_dup<mode>): Likewise.
5170 (neon_vld1_dup<mode>): Likewise.
5171 (neon_vld1_dupv2di): Likewise.
5172 (neon_vst1<mode>): Likewise.
5173 (neon_vst1_lane<mode>): Likewise.
5174 (neon_vst1_lane<mode>): Likewise.
5175 (neon_vld2<mode>): Likewise.
5176 (neon_vld2<mode>): Likewise.
5177 (neon_vld2_lane<mode>): Likewise.
5178 (neon_vld2_lane<mode>): Likewise.
5179 (neon_vld2_dup<mode>): Likewise.
5180 (neon_vst2<mode>): Likewise.
5181 (neon_vst2<mode>): Likewise.
5182 (neon_vst2_lane<mode>): Likewise.
5183 (neon_vst2_lane<mode>): Likewise.
5184 (neon_vld3<mode>): Likewise.
5185 (neon_vld3qa<mode>): Likewise.
5186 (neon_vld3qb<mode>): Likewise.
5187 (neon_vld3_lane<mode>): Likewise.
5188 (neon_vld3_lane<mode>): Likewise.
5189 (neon_vld3_dup<mode>): Likewise.
5190 (neon_vst3<mode>): Likewise.
5191 (neon_vst3qa<mode>): Likewise.
5192 (neon_vst3qb<mode>): Likewise.
5193 (neon_vst3_lane<mode>): Likewise.
5194 (neon_vst3_lane<mode>): Likewise.
5195 (neon_vld4<mode>): Likewise.
5196 (neon_vld4qa<mode>): Likewise.
5197 (neon_vld4qb<mode>): Likewise.
5198 (neon_vld4_lane<mode>): Likewise.
5199 (neon_vld4_lane<mode>): Likewise.
5200 (neon_vld4_dup<mode>): Likewise.
5201 (neon_vst4<mode>): Likewise.
5202 (neon_vst4qa<mode>): Likewise.
5203 (neon_vst4qb<mode>): Likewise.
5204 (neon_vst4_lane<mode>): Likewise.
5205 (neon_vst4_lane<mode>): Likewise.
5206 (neon_vec_unpack<US>_lo_<mode>): Likewise.
5207 (neon_vec_unpack<US>_hi_<mode>): Likewise.
5208 (neon_vec_<US>mult_lo_<mode>): Likewise.
5209 (neon_vec_<US>mult_hi_<mode>): Likewise.
5210 (neon_vec_<US>shiftl_<mode>): Likewise.
5211 (neon_unpack<US>_<mode>): Likewise.
5212 (neon_vec_<US>mult_<mode>): Likewise.
5213 (vec_pack_trunc_<mode>): Likewise.
5214 (neon_vec_pack_trunc_<mode>): Likewise.
5215 (neon_vabd<mode>_2): Likewise.
5216 (neon_vabd<mode>_3): Likewise.
5218 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5220 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
5221 (load_pair): Update type attribute.
5222 (store_pair): Update type attribute.
5223 * config/aarch64/iterators.md (q): New.
5225 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5227 * config/arm/types.md: Add new types for Neon insns.
5229 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5230 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5231 Sergey Lega <sergey.s.lega@intel.com>
5232 Anna Tikhonova <anna.tikhonova@intel.com>
5233 Ilya Tocar <ilya.tocar@intel.com>
5234 Andrey Turetskiy <andrey.turetskiy@intel.com>
5235 Ilya Verbin <ilya.verbin@intel.com>
5236 Kirill Yukhin <kirill.yukhin@intel.com>
5237 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5239 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
5240 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
5241 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
5243 (srcp14<mode>): Ditto.
5244 (rsqrt14<mode>): Ditto.
5245 (rsqrt14<mode>): Ditto.
5246 (avx512f_vmscalef<mode>): Ditto.
5247 (avx512f_scalef<mode>): Ditto.
5248 (avx512f_getexp<mode>): Ditto.
5249 (avx512f_sgetexp<mode>): Ditto.
5250 (avx512f_fixupimm<mode>): Ditto.
5251 (avx512f_sfixupimm<mode>): Ditto.
5252 (avx512f_rndscale<mode>): Ditto.
5253 (*avx512er_exp2<mode>): Ditto.
5254 (*avx512er_rcp28<mode>): Ditto.
5255 (avx512er_rsqrt28<mode>): Ditto.
5256 (avx512f_getmant<mode>): Ditto.
5257 (avx512f_getmant<mode>): Ditto.
5258 (avx512f_rndscale<mode>): Fix formatting.
5260 2013-10-15 Martin Jambor <mjambor@suse.cz>
5262 * ipa-utils.h (ipa_edge_within_scc): Declare.
5263 * ipa-cp.c (edge_within_scc): Moved...
5264 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
5266 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5267 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5268 Sergey Lega <sergey.s.lega@intel.com>
5269 Anna Tikhonova <anna.tikhonova@intel.com>
5270 Ilya Tocar <ilya.tocar@intel.com>
5271 Andrey Turetskiy <andrey.turetskiy@intel.com>
5272 Ilya Verbin <ilya.verbin@intel.com>
5273 Kirill Yukhin <kirill.yukhin@intel.com>
5274 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5276 * config/i386/predicates.md (const_8_to_15_operand): New.
5277 (const_16_to_31_operand): Ditto.
5278 * config/i386/sse.md (V8FI): New.
5280 (reduc_splus_v8df): Ditto.
5281 (reduc_splus_v16sf): Ditto.
5282 (avx512f_vextract<shuffletype>32x4_1): Ditto.
5283 (vec_extract_hi_<mode>): Ditto.
5284 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
5285 (vec_set_lo_<mode>): Ditto.
5286 (vec_set_hi_<mode>): Ditto.
5287 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
5288 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
5289 (avx512f_pshufd_1): Ditto.
5290 (avx512f_broadcast<mode>): Ditto.
5291 (avx512f_broadcast<mode>): Ditto.
5292 (define_split): Split vec_extract_lo into move.
5293 (ssequartermode): Ditto.
5294 (ssedoublemode): Extened with wider modes.
5295 (vec_extract_lo_<mode>): Ditto.
5297 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5298 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5299 Sergey Lega <sergey.s.lega@intel.com>
5300 Anna Tikhonova <anna.tikhonova@intel.com>
5301 Ilya Tocar <ilya.tocar@intel.com>
5302 Andrey Turetskiy <andrey.turetskiy@intel.com>
5303 Ilya Verbin <ilya.verbin@intel.com>
5304 Kirill Yukhin <kirill.yukhin@intel.com>
5305 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5307 * config/i386/predicates.md (register_or_constm1_operand): New.
5308 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
5309 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
5310 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
5311 UNSPEC_SCATTER_PREFETCH
5313 (avx512f_ucmp<mode>3): Ditto.
5314 (avx512f_vternlog<mode>): Ditto.
5315 (avx512f_align<mode>): Ditto.
5316 (<shift_insn><mode>3): Ditto.
5317 (avx512f_<rotate>v<mode>): Ditto.
5318 (avx512f_<rotate><mode>): Ditto.
5319 (avx512f_eq<mode>3): Ditto.
5320 (avx512f_eq<mode>3_1): Ditto.
5321 (avx512f_gt<mode>3): Ditto.
5322 (avx512f_testm<mode>3): Ditto.
5323 (avx512f_testnm<mode>3): Ditto.
5324 (avx512pf_gatherpf<mode>): Ditto.
5325 (*avx512pf_gatherpf<mode>_mask): Ditto.
5326 (*avx512pf_gatherpf<mode>): Ditto.
5327 (avx512pf_scatterpf<mode>): Ditto.
5328 (*avx512pf_scatterpf<mode>_mask): Ditto.
5329 (*avx512pf_scatterpf<mode>): Ditto.
5330 (avx512f_vec_dup_gpr<mode>): Ditto.
5331 (clz<mode>2): Ditto.
5332 (conflict<mode>): Ditto.
5333 (REDUC_SMINMAX_MODE): Extened with wider modes.
5334 (reduc_<code>_<mode>): Ditto.
5335 (vlshr<mode>3): Ditto.
5336 (vashl<mode>3): Ditto.
5338 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5339 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5340 Sergey Lega <sergey.s.lega@intel.com>
5341 Anna Tikhonova <anna.tikhonova@intel.com>
5342 Ilya Tocar <ilya.tocar@intel.com>
5343 Andrey Turetskiy <andrey.turetskiy@intel.com>
5344 Ilya Verbin <ilya.verbin@intel.com>
5345 Kirill Yukhin <kirill.yukhin@intel.com>
5346 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5348 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
5351 (avx512fmaskmode): Ditto.
5352 (bcstscalarsuff): Ditto.
5353 (avx512f_blendm<mode>): Ditto.
5354 (cmp_imm_predicate): Ditto.
5355 (avx512f_cmp<mode>3): Ditto.
5356 (avx512f_vec_dup<mode>): Ditto.
5357 (avx512f_vec_dup_mem<mode>): Ditto.
5358 (avx512f_vpermi2var<mode>3): Ditto.
5359 (avx512f_vpermt2var<mode>3): Ditto.
5360 (vec_init<mode>): Ditto.
5361 (avx512f_gathersi<mode>): Ditto.
5362 (*avx512f_gathersi<mode>): Ditto.
5363 (*avx512f_gathersi<mode>_2): Ditto.
5364 (avx512f_gatherdi<mode>): Ditto.
5365 (*avx512f_gatherdi<mode>): Ditto.
5366 (*avx512f_gatherdi<mode>_2): Ditto.
5367 (avx512f_scattersi<mode>): Ditto.
5368 (*avx512f_scattersi<mode>): Ditto.
5369 (avx512f_scatterdi<mode>): Ditto.
5370 (*avx512f_scatterdi<mode>): Ditto.
5371 (sseintprefix): Extened with wider modes.
5372 (VEC_GATHER_IDXSI): Ditto.
5373 (VEC_GATHER_IDXDI): Ditto.
5374 (VEC_GATHER_SRCDI): Ditto.
5376 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
5377 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5379 * config/arm/t-aprofile: New file.
5380 * config.gcc: Handle --with-multilib-list option.
5382 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
5384 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
5387 2013-10-15 Richard Biener <rguenther@suse.de>
5389 * tree-loop-distribution.c (build_empty_rdg): Inline into
5391 (rdg_flag_vertex): Inline into single user.
5392 (rdg_flag_vertex_and_dependent): Likewise.
5393 (build_rdg_partition_for_vertex): Remove processed bitmap.
5394 (rdg_build_partitions): Simplify.
5396 2013-10-15 Richard Biener <rguenther@suse.de>
5398 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
5399 Restructure forwarding through conversions and copies to
5400 avoid performing copy-propagation the wrong way. Adjust
5401 recursion invocations.
5402 (forward_propagate_addr_expr): Add argument stating if we
5403 are recursing from a single-use.
5404 (ssa_forward_propagate_and_combine): Adjust.
5406 2013-10-14 David Malcolm <dmalcolm@redhat.com>
5408 * dumpfile.h (gcc::dump_manager): New class, to hold state
5409 relating to dumpfile management.
5410 (get_dump_file_name): Remove in favor of method of dump_manager.
5411 (dump_initialized_p): Likewise.
5412 (dump_start): Likewise.
5413 (dump_finish): Likewise.
5414 (dump_switch_p): Likewise.
5415 (dump_register): Likewise.
5416 (get_dump_file_info): Likewise.
5417 * context.c (gcc::context::context): Construct the dump_manager
5419 * context.h (gcc::context::get_dumps): New.
5420 (gcc::context::m_dumps): New.
5421 * coverage.c (coverage_init): Port to dump_manager API.
5422 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
5423 (extra_dump_files_in_use): Likewise.
5424 (extra_dump_files_alloced): Likewise.
5425 (gcc::dump_manager::dump_manager): New.
5426 (dump_register): Convert to...
5427 (gcc::dump_manager::dump_register): ...method, replacing
5428 function-static next_dump with m_next_dump field.
5429 (get_dump_file_info): Convert to...
5430 (gcc::dump_manager::get_dump_file_info): ...method.
5431 (get_dump_file_name): Convert to...
5432 (gcc::dump_manager::get_dump_file_name): ...method.
5433 (dump_start): Convert to...
5434 (gcc::dump_manager::dump_start): ...method.
5435 (dump_finish): Convert to...
5436 (gcc::dump_manager::dump_finish): ...method.
5437 (dump_begin): Replace body with...
5438 (gcc::dump_manager::dump_begin): ...new method.
5439 (dump_phase_enabled_p): Convert to...
5440 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5441 (dump_phase_enabled_p): Convert to...
5442 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5443 (dump_initialized_p): Convert to...
5444 (gcc::dump_manager::dump_initialized_p): ...method.
5445 (dump_flag_name): Replace body with...
5446 (gcc::dump_manager::dump_flag_name): ...new method.
5447 (dump_enable_all): Convert to...
5448 (gcc::dump_manager::dump_enable_all): ...new method.
5449 (opt_info_enable_passes): Convert to...
5450 (gcc::dump_manager::opt_info_enable_passes): ...new method.
5451 (dump_switch_p_1): Convert to...
5452 (gcc::dump_manager::dump_switch_p_1): ...new method.
5453 (dump_switch_p): Convert to...
5454 (gcc::dump_manager::dump_switch_p): ...new method.
5455 (opt_info_switch_p): Port to dump_manager API.
5456 (enable_rtl_dump_file): Likewise.
5457 * opts-global.c (handle_common_deferred_options): Port to new
5459 * passes.c (pass_manager::finish_optimization_passes): Likewise.
5460 (pass_manager::register_one_dump_file): Likewise.
5461 (pass_manager::register_pass): Likewise.
5462 (pass_init_dump_file): Likewise.
5463 (pass_fini_dump_file): Likewise.
5464 * statistics.c (statistics_early_init): Likewise.
5466 2013-10-14 Richard Biener <rguenther@suse.de>
5468 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
5469 iterative_hash_canonical_type, gimple_canonical_type_hash,
5470 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
5471 gimple_register_canonical_type, print_gimple_types_stats,
5472 free_gimple_type_tables): Move to lto/lto.c
5473 (gt-gimple.h): Do not include.
5474 * gimple.h (gimple_register_canonical_type,
5475 print_gimple_types_stats, free_gimple_type_tables): Remove.
5476 * Makefile.in (GTFILES): Remove gimple.c.
5478 2013-10-14 Travis Snoozy <quandary@remstate.com>
5481 * config/msp430/msp430.c (msp430_option_override): Correct thinko
5482 scanning for msp430x targets.
5484 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
5487 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
5488 (registers_ok_for_ldd_peep): Move around.
5489 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
5490 * config/sparc/sparc.md (widening peepholes): Use it.
5492 2013-10-14 Richard Biener <rguenther@suse.de>
5496 * gimple.c (iterative_hash_canonical_type): Make sure to
5497 record the hash into the correct hashtable slot.
5499 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
5501 PR rtl-optimization/58662
5502 * combine.c (try_combine): Take into account death nodes on I2 when
5503 splitting a PARALLEL of two independent SETs. Fix dump message.
5505 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
5508 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
5509 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
5510 * config/sh/t-sh (sh_treg_combine.o): New entry.
5511 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
5512 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
5513 (register_sh_passes): New function. Register sh_treg_combine pass.
5514 (sh_option_override): Invoke it.
5515 (sh_canonicalize_comparison): Handle op0_preserve_value.
5516 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
5517 opportunities. Canonicalize branch condition.
5518 (nott): Allow only if pseudos can be created for non-SH2A.
5520 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
5523 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
5524 (ix86_expand_movmem): Replace copy_addr_to_reg with
5525 ix86_copy_addr_to_reg.
5526 (ix86_expand_setmem): Likewise.
5528 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
5530 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
5531 provided by ix86_fp_compare_mode instead of CCFPUmode.
5533 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
5535 * config/aarch64/arm_neon.h
5536 (vtbx<1,3>_<psu>8): Fix register constriants.
5538 2013-10-11 Jeff Law <law@redhat.com>
5540 PR tree-optimization/58640
5541 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
5542 threading paths that cross over two loop entry points.
5544 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5546 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
5547 handle vector float as well.
5548 (*vsx_le_perm_load_v4si): Likewise.
5549 (*vsx_le_perm_store_v2di): Likewise.
5550 (*vsx_le_perm_store_v4si): Likewise.
5552 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5554 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
5555 directly to circumvent subtract from splat{31} workaround.
5556 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
5558 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
5559 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
5560 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
5561 (altivec_vperm_<mode>): Convert to define_insn_and_split to
5562 separate big and little endian logic.
5563 (*altivec_vperm_<mode>_internal): New define_insn.
5564 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
5565 separate big and little endian logic.
5566 (*altivec_vperm_<mode>_uns_internal): New define_insn.
5567 (vec_permv16qi): Add little endian logic.
5569 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
5571 * doc/extend.texi (returns_nonnull): Remove arguments.
5573 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5574 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5575 Sergey Lega <sergey.s.lega@intel.com>
5576 Anna Tikhonova <anna.tikhonova@intel.com>
5577 Ilya Tocar <ilya.tocar@intel.com>
5578 Andrey Turetskiy <andrey.turetskiy@intel.com>
5579 Ilya Verbin <ilya.verbin@intel.com>
5580 Kirill Yukhin <kirill.yukhin@intel.com>
5581 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5583 * config/i386/sse.md (VI48F_256_512): New.
5584 (avx2_permvar<mode>): Change to ...
5585 (<avx2_avx512f>_permvar<mode>): This.
5587 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5588 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5589 Sergey Lega <sergey.s.lega@intel.com>
5590 Anna Tikhonova <anna.tikhonova@intel.com>
5591 Ilya Tocar <ilya.tocar@intel.com>
5592 Andrey Turetskiy <andrey.turetskiy@intel.com>
5593 Ilya Verbin <ilya.verbin@intel.com>
5594 Kirill Yukhin <kirill.yukhin@intel.com>
5595 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5597 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
5598 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
5599 * config/i386/sse.md (VI4_AVX): New.
5600 (sf2simodelower): Ditto.
5601 (sse2_cvtps2dq): Change to ...
5602 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
5604 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5605 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5606 Sergey Lega <sergey.s.lega@intel.com>
5607 Anna Tikhonova <anna.tikhonova@intel.com>
5608 Ilya Tocar <ilya.tocar@intel.com>
5609 Andrey Turetskiy <andrey.turetskiy@intel.com>
5610 Ilya Verbin <ilya.verbin@intel.com>
5611 Kirill Yukhin <kirill.yukhin@intel.com>
5612 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5614 * config/i386/sse.md (V_512): New.
5616 (vcond<V_512:mode><VF_512:mode>): Ditto.
5617 (vcond<V_512:mode><VI_512:mode>): Ditto.
5618 (vcondu<V_512:mode><VI_512:mode>): Ditto.
5620 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5621 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5622 Sergey Lega <sergey.s.lega@intel.com>
5623 Anna Tikhonova <anna.tikhonova@intel.com>
5624 Ilya Tocar <ilya.tocar@intel.com>
5625 Andrey Turetskiy <andrey.turetskiy@intel.com>
5626 Ilya Verbin <ilya.verbin@intel.com>
5627 Kirill Yukhin <kirill.yukhin@intel.com>
5628 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5630 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
5631 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
5633 (fma<mode>4): Removed condition.
5634 (fms<mode>4): Ditto.
5635 (fnma<mode>4): Ditto.
5636 (fnms<mode>4): Ditto.
5637 (fma4i_fmadd_<mode>): Ditto.
5638 (*fma_fmadd_<mode>): Ditto.
5639 (*fma_fmsub_<mode>): Ditto.
5640 (*fma_fnmadd_<mode>): Ditto.
5641 (*fma_fnmsub_<mode>): Ditto.
5642 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
5643 (*fma_fmaddsub_<mode>): Ditto.
5644 (*fma_fmsubadd_<mode>): Ditto.
5645 (*fmai_fmadd_<mode>): Ditto.
5646 (*fmai_fmsub_<mode>): Ditto.
5647 (*fmai_fnmadd_<mode>): Ditto.
5648 (*fmai_fnmsub_<mode>): Ditto.
5650 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5651 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5652 Sergey Lega <sergey.s.lega@intel.com>
5653 Anna Tikhonova <anna.tikhonova@intel.com>
5654 Ilya Tocar <ilya.tocar@intel.com>
5655 Andrey Turetskiy <andrey.turetskiy@intel.com>
5656 Ilya Verbin <ilya.verbin@intel.com>
5657 Kirill Yukhin <kirill.yukhin@intel.com>
5658 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5660 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
5661 (VI124_256): Changed to ...
5662 (VI124_256_48_512): This.
5663 (ssepackmode): Extended with wider modes.
5664 (<code><mode>3): Changed iterator.
5665 (*avx2_<code><mode>3): Ditto.
5666 (vec_pack_trunc_<mode>): Ditto.
5668 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5669 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5670 Sergey Lega <sergey.s.lega@intel.com>
5671 Anna Tikhonova <anna.tikhonova@intel.com>
5672 Ilya Tocar <ilya.tocar@intel.com>
5673 Andrey Turetskiy <andrey.turetskiy@intel.com>
5674 Ilya Verbin <ilya.verbin@intel.com>
5675 Kirill Yukhin <kirill.yukhin@intel.com>
5676 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5678 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
5679 (VI8F_256_512): Ditto.
5680 (abs<mode>2): Changed iterator.
5681 (avx2_perm<mode>): Changed to ...
5682 (<avx2_avx512f>_perm<mode>): This.
5683 (avx2_perm<mode>_1): Changed to ...
5684 (<avx2_avx512f>_perm<mode>_1): This.
5686 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5687 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5688 Sergey Lega <sergey.s.lega@intel.com>
5689 Anna Tikhonova <anna.tikhonova@intel.com>
5690 Ilya Tocar <ilya.tocar@intel.com>
5691 Andrey Turetskiy <andrey.turetskiy@intel.com>
5692 Ilya Verbin <ilya.verbin@intel.com>
5693 Kirill Yukhin <kirill.yukhin@intel.com>
5694 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5696 * config/i386/sse.md (VI48_AVX512F): New.
5697 (VI48_AVX2): Changed to ...
5698 (VI48_AVX2_48_AVX512F): This.
5699 (avx2_ashrv<mode>): Changed to ...
5700 (<avx2_avx512f>_ashrv<mode>): This.
5701 (avx2_<shift_insn>v<mode>): Changed to ...
5702 (<avx2_avx512f>_<shift_insn>v<mode>): This.
5704 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5705 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5706 Sergey Lega <sergey.s.lega@intel.com>
5707 Anna Tikhonova <anna.tikhonova@intel.com>
5708 Ilya Tocar <ilya.tocar@intel.com>
5709 Andrey Turetskiy <andrey.turetskiy@intel.com>
5710 Ilya Verbin <ilya.verbin@intel.com>
5711 Kirill Yukhin <kirill.yukhin@intel.com>
5712 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5714 * config/i386/sse.md (VI4_AVX512F): New.
5715 (VI8_AVX2_AVX512F): Ditto.
5716 (mul<mode>3): Extended with wider modes.
5717 (*<sse4_1_avx2>_mul<mode>3): Ditto.
5718 (mul<mode>3): Ditto.
5719 (vec_widen_<s>mult_odd_<mode>): Ditto.
5721 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5722 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5723 Sergey Lega <sergey.s.lega@intel.com>
5724 Anna Tikhonova <anna.tikhonova@intel.com>
5725 Ilya Tocar <ilya.tocar@intel.com>
5726 Andrey Turetskiy <andrey.turetskiy@intel.com>
5727 Ilya Verbin <ilya.verbin@intel.com>
5728 Kirill Yukhin <kirill.yukhin@intel.com>
5729 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5731 * config/i386/sse.md (VI2_AVX512F): New.
5732 (VI124_AVX512F): Ditto.
5733 (sseunpackmode): Extended with wider modes.
5734 (sseunpackfltmode): Ditto.
5735 (vec_unpacks_float_hi_<mode>): Ditto.
5736 (vec_unpacks_float_lo_<mode>): Ditto.
5737 (vec_unpacku_float_hi_<mode>): Ditto.
5738 (vec_unpacku_float_lo_<mode>): Ditto.
5739 (vec_unpacks_lo_<mode>): Ditto.
5740 (vec_unpacks_hi_<mode>): Ditto.
5741 (vec_unpacku_lo_<mode>): Ditto.
5742 (vec_unpacku_hi_<mode>): Ditto.
5744 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5745 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5746 Sergey Lega <sergey.s.lega@intel.com>
5747 Anna Tikhonova <anna.tikhonova@intel.com>
5748 Ilya Tocar <ilya.tocar@intel.com>
5749 Andrey Turetskiy <andrey.turetskiy@intel.com>
5750 Ilya Verbin <ilya.verbin@intel.com>
5751 Kirill Yukhin <kirill.yukhin@intel.com>
5752 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5754 * config/i386/i386.md (multdiv): New.
5755 (multdiv_mnemonic): Ditto.
5756 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5757 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5758 (<sse>_vmdiv<mode>3): Removed.
5760 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5761 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5762 Sergey Lega <sergey.s.lega@intel.com>
5763 Anna Tikhonova <anna.tikhonova@intel.com>
5764 Ilya Tocar <ilya.tocar@intel.com>
5765 Andrey Turetskiy <andrey.turetskiy@intel.com>
5766 Ilya Verbin <ilya.verbin@intel.com>
5767 Kirill Yukhin <kirill.yukhin@intel.com>
5768 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5770 * config/i386/sse.md (V): Extended with wider modes.
5772 (ssehalfvecmode): Ditto.
5774 (ssepackfltmode): Ditto.
5775 (avx_vec_concat<mode>): Ditto.
5776 (V_256_512): New iterator.
5777 (VF2_512_256): Ditto.
5778 (si2dfmode): New attribute.
5779 (si2dfmodelower): Ditto.
5781 (concat_tg_mode): Ditto.
5782 (floatv4siv4df2): Changed to ...
5783 (float<si2dfmodelower><mode>2): This.
5784 (avx_cvtps2pd256): Changed to ...
5785 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5786 (vec_pack_trunc_v4df): Changed to ...
5787 (vec_pack_trunc_<mode>): This.
5788 (avx_vpermil<mode>): Changed to ...
5789 (<sse2_avx_avx512f>_vpermil<mode>): This.
5790 (<fixsuffix>fix_truncv8dfv8si2): New.
5791 (vec_pack_sfix_trunc_v8df): Ditto.
5792 (avx512f_rndscale<mode>): Ditto.
5793 (avx512f_roundpd512): Ditto.
5794 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5796 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5797 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5798 Sergey Lega <sergey.s.lega@intel.com>
5799 Anna Tikhonova <anna.tikhonova@intel.com>
5800 Ilya Tocar <ilya.tocar@intel.com>
5801 Andrey Turetskiy <andrey.turetskiy@intel.com>
5802 Ilya Verbin <ilya.verbin@intel.com>
5803 Kirill Yukhin <kirill.yukhin@intel.com>
5804 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5806 * config/i386/i386.md (any_fix): New iterator.
5807 (fixsuffix): New attribute.
5808 * config/i386/sse.md (VF1): Extened with wider modes.
5812 (sseintvecmodelower): Ditto.
5813 (ssescalarmode): Ditto.
5814 (ssescalarnum): Ditto.
5816 (ssexmmmode): Ditto.
5817 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5818 (<sse>_rcp<mode>2): Change iterator.
5819 (rsqrt<mode>2): Ditto.
5820 (<sse>_rsqrt<mode>2): Ditto.
5821 (avx2_vec_dup<mode>): Ditto.
5822 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5823 (round<mode>2_sfix): Ditto.
5824 (avx2_pbroadcast<mode>): Ditto.
5825 (*andnot<mode>3): Handle XI mode.
5826 (*<code><mode>3): Ditto.
5827 (AVXTOSSEMODE): Removed.
5828 (avx_vpermil<mode>): Changed to ...
5829 (<sse2_avx_avx512f>_vpermil<mode>): This.
5831 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5832 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5833 Sergey Lega <sergey.s.lega@intel.com>
5834 Anna Tikhonova <anna.tikhonova@intel.com>
5835 Ilya Tocar <ilya.tocar@intel.com>
5836 Andrey Turetskiy <andrey.turetskiy@intel.com>
5837 Ilya Verbin <ilya.verbin@intel.com>
5838 Kirill Yukhin <kirill.yukhin@intel.com>
5839 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5841 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5842 (<sse>_comi): Ditto.
5843 (<sse>_ucomi): Ditto.
5844 (sse_cvtss2siq_2): Ditto.
5845 (sse2_cvtsd2si): Ditto.
5846 (sse2_cvtsd2siq): Ditto.
5847 (sse2_cvttsd2si): Ditto.
5848 (sse2_cvttsd2siq): Ditto.
5849 (<shift_insn><mode>3): Ditto.
5850 (sse2_cvtsi2sdq): Update constraint and prefix.
5851 (sse_cvtsi2ss): Update prefix.
5852 (sse_cvtsi2ssq): Ditto.
5854 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5856 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5857 ignore internal calls.
5859 2013-10-11 Richard Biener <rguenther@suse.de>
5861 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5862 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5863 TREE_OVERFLOW is set.
5865 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5867 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5869 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5870 * gimple.h: Likewise.
5872 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5874 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5877 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5879 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5880 * configure: Regenerate.
5882 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5884 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5885 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5886 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5887 similarly pp_right_paren.
5888 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5889 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5890 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5891 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5893 (omp_declare_simd_clauses_equal,
5894 omp_remove_redundant_declare_simd_attrs): New functions.
5895 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5896 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5897 * tree.h (OMP_LOOP_CHECK): Define.
5898 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5899 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5900 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5901 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5902 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5903 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5904 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5905 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5906 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5907 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5908 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5909 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5910 (omp_remove_redundant_declare_simd_attrs): New prototype.
5911 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5912 GIMPLE_OMP_TEAMS): New codes.
5913 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5914 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5916 (target_nesting_level): New variable.
5917 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5918 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5919 for collapse > 1 static schedule unless ordered.
5920 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5921 (determine_parallel_type): Adjust get_ws_args_for caller.
5922 (install_var_field): Handle mask & 4 for double indirection.
5923 (scan_sharing_clauses): Ignore shared clause on teams construct.
5924 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5925 (create_omp_child_function): If inside target or declare target
5926 constructs, set "omp declare target" attribute on the child function.
5927 (find_combined_for): New function.
5928 (scan_omp_parallel): Handle combined loops.
5929 (scan_omp_target, scan_omp_teams): New functions.
5930 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5931 restrictions and set ctx->cancellable for cancellable constructs.
5932 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5933 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5934 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5935 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5936 (omp_clause_aligned_alignment): New function.
5937 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5938 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5939 on teams constructs. Handle user defined reductions and new
5941 (lower_reduction_clauses): Don't set placeholder to address of ref
5942 if it has already the right type.
5943 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5944 (expand_parallel_call): Use the new non-_start suffixed builtins,
5945 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5946 and GOMP_parallel_end after the call.
5947 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5948 (expand_omp_for_init_counts): Handle combined loops.
5949 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5951 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5952 end of cancellable loops.
5953 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5954 Likewise. Handle collapse > 1 loops.
5955 (expand_omp_simd): Handle combined loops.
5956 (expand_omp_for): Add inner_stmt argument, adjust callers of
5957 expand_omp_for* functions, use expand_omp_for_static*chunk even
5958 for collapse > 1 unless ordered.
5959 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5960 of cancellable sections.
5961 (expand_omp_single): Remove need_barrier variable, just rely on
5962 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5963 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5964 (expand_omp_atomic_load, expand_omp_atomic_store,
5965 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5966 (expand_omp_target): New function.
5967 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5968 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5969 (build_omp_regions_1): Immediately close region for
5970 GF_OMP_TARGET_KIND_UPDATE.
5971 (maybe_add_implicit_barrier_cancel): New function.
5972 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5974 (lower_omp_single): Likewise. Add clobber after the barrier.
5975 (lower_omp_taskgroup): New function.
5976 (lower_omp_for): Handle combined loops. Adjust
5977 lower_rec_input_clauses caller. Handle cancellation.
5978 (lower_depend_clauses): New function.
5979 (lower_omp_taskreg): Lower depend clauses. Adjust
5980 lower_rec_input_clauses caller. Add clobber after the call. Handle
5982 (lower_omp_target, lower_omp_teams): New functions.
5983 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5984 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5985 and GOMP_cancellation_point calls.
5986 (lower_omp): Fold stmts inside of target region.
5987 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5988 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5989 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5990 (BT_FN_VOID_OMPFN_PTR_UINT,
5991 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5992 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5993 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5994 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5995 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5996 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5997 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5998 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5999 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
6000 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
6001 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
6002 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
6003 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
6004 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
6005 BUILT_IN_GOMP_PARALLEL_END.
6006 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
6007 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
6008 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
6009 GF_OMP_FOR_KIND_DISTRIBUTE.
6010 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
6011 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
6012 (dump_gimple_omp_return): Print lhs if it has any.
6013 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
6014 gimple_omp_atomic_seq_cst_p.
6015 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
6016 and GIMPLE_OMP_TEAMS.
6017 * langhooks.c (lhd_omp_mappable_type): New function.
6018 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
6019 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
6021 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
6022 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
6023 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
6024 (struct gimplify_omp_ctx): Add combined_loop field.
6025 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
6026 on stmts inside of target region.
6027 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
6028 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
6029 ORT_TARGET and ORT_TARGET_DATA.
6030 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
6032 (omp_notice_threadprivate_variable): Complain about threadprivate
6033 variables in target region.
6034 (omp_notice_variable): Complain about vars with non-mappable type
6035 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
6036 (omp_check_private): Ignore ORT_TARGET* regions.
6037 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
6038 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
6039 (find_combined_omp_for): New function.
6040 (gimplify_omp_for): Handle gimplification of combined loops.
6041 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
6043 (gimplify_omp_target_update): New function.
6044 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
6045 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
6046 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
6047 (gimplify_body): If fndecl has "omp declare target" attribute, add
6048 implicit ORT_TARGET context around it.
6049 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
6050 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
6051 * tree-nested.c (convert_nonlocal_reference_stmt,
6052 convert_local_reference_stmt, convert_gimple_call): Handle
6053 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6054 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
6055 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
6056 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
6057 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
6058 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
6059 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
6060 BUILT_IN_GOMP_LOOP_END_CANCEL,
6061 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
6062 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
6063 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
6064 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
6065 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
6066 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
6067 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
6068 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
6069 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
6070 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
6071 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
6072 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
6073 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
6074 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
6075 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
6076 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6077 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
6078 gimple_build_omp_teams): New functions.
6079 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
6080 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
6081 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
6082 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6083 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
6084 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
6085 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
6086 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
6087 GF_OMP_ATOMIC_SEQ_CST): New.
6088 (gimple_build_omp_taskgroup, gimple_build_omp_target,
6089 gimple_build_omp_teams): New prototypes.
6090 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
6091 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6092 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
6093 GIMPLE_OMP_SINGLE as end of range.
6094 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
6095 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
6096 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
6097 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
6098 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
6099 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
6100 gimple_omp_target_kind, gimple_omp_target_set_kind,
6101 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
6102 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
6103 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
6104 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
6105 gimple_omp_teams_set_clauses): New inlines.
6106 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
6107 and GIMPLE_OMP_TASKGROUP.
6108 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
6109 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
6110 enum omp_clause_proc_bind_kind): New.
6111 (union omp_clause_subcode): Add depend_kind, map_kind and
6112 proc_bind_kind fields.
6113 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
6114 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
6115 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
6116 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
6117 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
6119 2013-10-10 Teresa Johnson <tejohnson@google.com>
6121 * predict.c (tree_estimate_probability): Add new parameter
6122 for estimate_bb_frequencies.
6123 (estimate_bb_frequencies): Add new parameter to force estimation.
6124 (rebuild_frequencies): When max frequency in function is small,
6125 recompute counts from frequencies.
6126 * predict.h (estimate_bb_frequencies): New parameter.
6128 2013-10-10 David Malcolm <dmalcolm@redhat.com>
6130 * ipa-inline.c (ipa_inline): Fix leak of "order" when
6131 optimizations are disabled.
6133 2013-10-10 David Malcolm <dmalcolm@redhat.com>
6135 * coverage.c (coverage_finish): Fix leak of da_file_name.
6137 2013-10-10 Jan Hubicka <jh@suse.cz>
6139 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
6140 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
6143 2013-10-10 Jakub Jelinek <jakub@redhat.com>
6146 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
6147 if any labels are in FALLTHRU_BB, use a special label emitted
6148 immediately after the asm goto insn rather than label_rtx
6150 (expand_asm_stmt): Adjust caller.
6151 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
6152 edge if the last insn in predecessor is a jump with single successor,
6153 but it isn't simplejump_p.
6155 2013-10-10 Richard Biener <rguenther@suse.de>
6157 PR tree-optimization/58656
6158 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
6160 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
6162 * gimplify.c: Include expr.h and tm_p.h for targets with special
6163 va-arg padding requirements.
6165 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
6167 * tree-flow.h: Move some prototypes to gimple.h.
6168 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
6169 * gimple.h: Relocate some prototypes from tree-flow.h
6170 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
6172 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
6173 (build_va_arg_indirect_ref): Relocate and make static.
6174 (std_gimplify_va_arg_expr): Relocate here.
6175 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
6176 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
6178 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6180 * doc/md.texi: Document the mnemonic attribute.
6182 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6185 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
6186 (set_attr_alternative x ...) when searching for user defined
6189 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
6191 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
6193 2013-10-09 Easwaran Raman <eraman@google.com>
6195 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
6196 * cfgexpand.c (defer_stack_allocation): ...use here
6197 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
6199 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
6201 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
6202 (operand_equal_for_value_replacement): New function, extracted from
6203 value_replacement and enhanced to catch more cases.
6204 (value_replacement): Use operand_equal_for_value_replacement.
6206 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6208 * loop-doloop.c (doloop_modify, doloop_optimize): Use
6209 get_max_loop_iterations.
6211 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6213 * config/arm/aarch-common.c (arm_early_load_addr_dep):
6214 Place comment above function.
6216 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6218 * tree-flow.h: Remove all remaining prototypes, enums and structs that
6219 are not related to tree-cfg.c.
6220 * tree-ssa-address.h: New file. Relocate prototypes.
6221 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
6222 (addr_for_mem_ref): New. Combine call to get_address_description and
6223 return addr_for_mem_ref.
6224 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
6225 * tree-ssa-live.h: Adjust prototypes.
6226 * passes.c: Include tree-ssa-live.h.
6227 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
6228 * graphite.c (graphite_transform_loops): Make static.
6229 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6230 make_pass_graphite, pass_data_graphite_transforms,
6231 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
6232 * ipa-pure-const.c (warn_function_noreturn): Make static.
6233 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6234 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6235 Relocate from tree-cfg.c
6236 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
6238 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6239 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6240 Move to ipa-pure-const.c.
6241 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
6242 Relocate from tree-optimize.c.
6243 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
6244 make_pass_fixup_cfg): Move to tree-cfg.c.
6245 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
6246 Relocate some prototypes.
6247 * tree-data-ref.h (tree_check_data_deps) Add prototype.
6248 * tree-dump.c (dump_function_to_file): Remove prototype.
6249 Add tree-flow.h to the include file.
6250 * tree-dump.h: Remove prototype.
6251 * tree-parloops.h: New File. Add prototypes.
6252 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
6253 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
6254 from tree-ssa-loop.c.
6255 * tree-predcom.c (run_tree_predictive_commoning,
6256 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
6257 Relocate here from tree-ssa-loop.c.
6258 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
6259 ssa_name_values.release ().
6260 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
6261 (ssa_name_values): Relocate from tree-flow.h.
6262 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
6263 * tree-ssa-loop.c (run_tree_predictive_commoning,
6264 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
6265 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6266 make_pass_graphite, pass_data_graphite_transforms,
6267 make_pass_graphite_transforms, gate_tree_parallelize_loops,
6268 tree_parallelize_loops, pass_data_parallelize_loops,
6269 make_pass_parallelize_loops): Move to other files.
6270 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
6272 * tree.h: Remove prototypes from tree-address.c.
6274 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6276 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
6277 (struct int_tree_map): Move to tree-hasher.h
6278 (SCALE, LABEL, PERCENT): Move to gimple.h
6279 * tree-flow-inline.h: Delete. Move functions to other files.
6280 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
6281 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
6282 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
6283 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
6284 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
6285 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
6286 * tree-hasher.h: Don't include tree-flow.h.
6287 (struct int_tree_map): Relocate from tree-flow.h.
6288 * tree-sra.c (contains_view_convert_expr_p): Relocate from
6289 tree-flow-inline.h and make static.
6290 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
6292 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
6293 tree-flow-inline.h and make static.
6294 * tree.h (is_global_var, may_be_aliased): Relocate from
6296 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
6297 * value-prof.c: No longer include tree-flow-inline.h.
6298 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
6300 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6302 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
6303 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
6304 (enum move_pos): Move to tree-ssa-loop-im.h
6305 * cfgloop.h: Move some prototypes.
6306 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
6307 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
6308 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
6309 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
6310 (loop_containing_stmt): Relocate from tree-flow-inline.h.
6311 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
6312 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
6313 (enum move_pos): Relocate here.
6314 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
6316 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
6317 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
6318 make_pass_lim): Relocate here from tree-ssa-loop.c.
6319 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
6321 (loop_edge_to_cancel, unloop_loops): Make static.
6322 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6323 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
6324 (tree_complete_unroll, gate_tree_complete_unroll,
6325 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
6326 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6327 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
6328 * tree-ssa-loop-ivopts.c: Remove local prototypes.
6329 (stmt_invariant_in_loop_p): Remove unused function.
6330 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
6331 * tree-ssa-loop-manip.h: New file. Add prototypes.
6332 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
6333 (gcov_type_to_double_int): Move to cfgloop.h.
6334 (double_int_cmp, bound_index,
6335 estimate_numbers_of_iterations_loop): Make static.
6336 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
6337 (max_loop_iterations): Factor out get_max_loop_iterations.
6338 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
6340 * tree-ssa-loop-niter.h: New file. Add prototypes.
6341 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
6342 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
6343 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
6344 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
6345 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
6346 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
6347 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
6348 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
6349 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
6350 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
6351 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6352 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
6353 pass_data_complete_unroll, make_pass_complete_unroll,
6354 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6355 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
6356 tree-ssa-loop-ivcanon.c.
6357 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
6358 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
6359 tree-ssa-loop-prefetch.c.
6360 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
6362 (get_lsm_tmp_name): Relocate and add suffix parameter.
6363 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
6364 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
6365 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
6366 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
6367 (get_estimated_loop_iterations): Factor out accessor from
6368 estimated_loop_iterations in tree-ssa-loop-niter.c.
6369 (get_max_loop_iterations): Factor out accessor from
6370 _max_loop_iterations in tree-ssa-niter.c.
6371 * loop-unroll.c (decide_unroll_constant_iterations,
6372 decide_unroll_runtime_iterations, decide_peel_simple,
6373 decide_unroll_stupid): Use new get_* accessors.
6375 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
6377 PR tree-optimization/20318
6378 * doc/extend.texi (returns_nonnull): New function attribute.
6379 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
6381 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
6382 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
6384 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
6387 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
6388 false if both components are bitfields.
6390 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6392 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
6393 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
6394 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
6395 * config/aarch64/aarch64.h
6396 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
6397 * config/aarch64/aarch64-simd-builtins.def
6398 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
6400 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6402 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
6404 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6406 * config/aarch64/arm_neon.h (vdiv_f64): Added.
6408 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6410 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
6411 (vneg_f64): New intrinsic.
6412 (vneg_s8): Asm replaced with C.
6413 (vneg_s16): Likewise.
6414 (vneg_s32): Likewise.
6415 (vneg_s64): New intrinsic.
6416 (vnegq_f32): Asm replaced with C.
6417 (vnegq_f64): Likewise.
6418 (vnegq_s8): Likewise.
6419 (vnegq_s16): Likewise.
6420 (vnegq_s32): Likewise.
6421 (vnegq_s64): Likewise.
6423 2013-10-09 Renlin Li <Renlin.Li@arm.com>
6425 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
6427 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6429 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
6430 packed stack special handling.
6431 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
6432 back to fixed stack slots for FPRs saved due to stdarg.
6434 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6436 * config/s390/s390.c (s390_frame_info): Restructure function.
6438 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6440 * config/s390/s390.c (struct s390_frame_layout): New field
6442 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
6443 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
6444 regs_ever_clobbered to char*.
6445 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
6446 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6447 (s390_register_info): Enable FPR save slots. Move/Copy some
6448 functionality into ...
6449 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
6450 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
6452 (s390_frame_info): Do gpr slot allocation here now. stdarg does
6453 not imply a stack frame.
6454 (s390_init_frame_layout): Remove variable clobbered_regs.
6455 (s390_update_register_info): Remove function.
6456 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
6457 cannot be used for register renaming.
6458 (s390_hard_regno_scratch_ok): New function.
6459 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
6460 (s390_initial_elimination_offset): Change offset calculation of
6461 the return address pointer.
6462 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6463 (restore_gprs): Set frame related flag.
6464 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6465 (s390_emit_prologue): Call s390_register_info instead of
6466 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
6467 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
6468 (s390_optimize_prologue): Call s390_optimize_register_info.
6469 Try to remove also FPR slot save/restore INSNs. Remove frame
6470 related flags from restore INSNs.
6472 2013-10-08 DJ Delorie <dj@redhat.com>
6474 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
6477 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
6478 avoid conflict with the MI use of %c.
6479 * config/rl78/rl78-real.md: change %c to %C throughout.
6480 * config/rl78/rl78-virt.md: Likewise.
6482 2013-10-08 Jan Hubicka <jh@suse.cz>
6484 * config/i386/i386.c (ix86_option_override_internal): Switch
6485 to SSE math for -ffast-math when target ISA supports SSE2.
6487 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
6489 * tree-flow.h: Remove some prototypes.
6490 * tree.h: Remove some protypes, add a couple.
6491 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
6492 using_eh_for_cleanups_p): Add interface routines for front ends.
6493 * tree-eh.h: New file. Add protoptyes.
6494 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
6495 (add_stmt_to_eh_lp_fn): Make static.
6496 (lower_try_finally): Use new using_eh_for_cleanups_p.
6497 * emit-rtl.c: Include tree-eh.h.
6498 * gimple.h: Include tree-eh.h.
6500 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
6502 PR tree-optimization/58480
6503 * tree-vrp.c (infer_nonnull_range): New function.
6504 (infer_value_range): Call infer_nonnull_range.
6506 2013-10-08 Dehao Chen <dehao@google.com>
6508 PR tree-optimization/58619
6509 * tree-inline.c (copy_phis_for_bb): Combine location data
6512 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6515 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
6516 RTX_FRAME_RELATED_P on INSN.
6518 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6520 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
6521 (altivec_expand_vec_perm_const): Call it.
6523 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6525 * config/rs6000/vector.md (mov<mode>): Emit permuted move
6526 sequences for LE VSX loads and stores at expand time.
6527 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
6529 * config/rs6000/rs6000.c (rs6000_const_vec): New.
6530 (rs6000_gen_le_vsx_permute): New.
6531 (rs6000_gen_le_vsx_load): New.
6532 (rs6000_gen_le_vsx_store): New.
6533 (rs6000_gen_le_vsx_move): New.
6534 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
6535 (*vsx_le_perm_load_v4si): New.
6536 (*vsx_le_perm_load_v8hi): New.
6537 (*vsx_le_perm_load_v16qi): New.
6538 (*vsx_le_perm_store_v2di): New.
6539 (*vsx_le_perm_store_v4si): New.
6540 (*vsx_le_perm_store_v8hi): New.
6541 (*vsx_le_perm_store_v16qi): New.
6542 (*vsx_xxpermdi2_le_<mode>): New.
6543 (*vsx_xxpermdi4_le_<mode>): New.
6544 (*vsx_xxpermdi8_le_V8HI): New.
6545 (*vsx_xxpermdi16_le_V16QI): New.
6546 (*vsx_lxvd2x2_le_<mode>): New.
6547 (*vsx_lxvd2x4_le_<mode>): New.
6548 (*vsx_lxvd2x8_le_V8HI): New.
6549 (*vsx_lxvd2x16_le_V16QI): New.
6550 (*vsx_stxvd2x2_le_<mode>): New.
6551 (*vsx_stxvd2x4_le_<mode>): New.
6552 (*vsx_stxvd2x8_le_V8HI): New.
6553 (*vsx_stxvd2x16_le_V16QI): New.
6555 2013-10-07 Renlin Li <Renlin.Li@arm.com>
6557 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
6559 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6561 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
6562 loop to work also for 31bit ABI.
6563 Save the stack pointer for frame_size > 0.
6565 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6567 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
6568 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
6569 constraint letters from expanders.
6570 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
6571 retry count to general_operand.
6572 ("tabort"): Give operand 0 a mode.
6573 ("tabort_1"): Add mode and constraint letter for operand 0.
6574 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
6576 2013-10-04 Jeff Law <law@redhat.com>
6578 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
6580 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
6582 (thread_across_edge): Here. Call it.
6584 2013-10-04 Cary Coutant <ccoutant@google.com>
6586 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
6587 discarding a location list expression (or a piece of one).
6589 2013-10-03 Jan Hubicka <jh@suse.cz>
6591 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
6592 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
6593 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6595 2013-10-03 Jan Hubicka <jh@suse.cz>
6597 * config/i386/i386.c (ix86_option_override_internal): Do not enable
6598 accumulate-outgoing-args when producing unwind info.
6600 2013-10-03 Wei Mi <wmi@google.com>
6602 * lra-constraints.c (insert_move_for_subreg): New function
6603 extracted from simplify_operand_subreg.
6604 (simplify_operand_subreg): Add reload for paradoxical subreg.
6606 2013-10-03 Rong Xu <xur@google.com>
6608 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
6609 the candidate of builtin_expect such that we should fix the
6610 size/time estimation.
6611 (estimate_function_body_sizes): Do the acutally size/time fix-up
6614 2013-10-03 Rong Xu <xur@google.com>
6616 * predict.c (tree_predict_by_opcode): Get the probability
6617 for builtin_expect from param builtin_expect_probability.
6618 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
6619 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
6620 * doc/invoke.texi: Add documentation for builtin-expect-probability.
6622 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
6625 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
6626 * calls.c (alloca_call_p): Use get_callee_fndecl.
6627 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
6628 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
6630 (vrp_visit_stmt): Remove duplicated code.
6632 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
6634 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
6635 ceildf2, btruncdf2, instead of vsx_* name.
6637 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
6638 iterators to only do V2DF and V4SF here. Move the DF code to
6639 rs6000.md where it is combined with SF mode. Replace <VSv> with
6640 just 'v' since only vector operations are handled with these insns
6641 after moving the DF support to rs6000.md.
6642 (vsx_sub<mode>3): Likewise.
6643 (vsx_mul<mode>3): Likewise.
6644 (vsx_div<mode>3): Likewise.
6645 (vsx_fre<mode>2): Likewise.
6646 (vsx_neg<mode>2): Likewise.
6647 (vsx_abs<mode>2): Likewise.
6648 (vsx_nabs<mode>2): Likewise.
6649 (vsx_smax<mode>3): Likewise.
6650 (vsx_smin<mode>3): Likewise.
6651 (vsx_sqrt<mode>2): Likewise.
6652 (vsx_rsqrte<mode>2): Likewise.
6653 (vsx_fms<mode>4): Likewise.
6654 (vsx_nfma<mode>4): Likewise.
6655 (vsx_copysign<mode>3): Likewise.
6656 (vsx_btrunc<mode>2): Likewise.
6657 (vsx_floor<mode>2): Likewise.
6658 (vsx_ceil<mode>2): Likewise.
6659 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
6660 (vsx_sminsf3): Likewise.
6661 (vsx_fmadf4): Likewise.
6662 (vsx_fmsdf4): Likewise.
6663 (vsx_nfmadf4): Likewise.
6664 (vsx_nfmsdf4): Likewise.
6665 (vsx_cmpdf_internal1): Likewise.
6667 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
6668 simpler to select whether a target has SPE or traditional floating
6669 point support in iterators.
6670 (TARGET_DF_SPE): Likewise.
6671 (TARGET_SF_FPR): Likewise.
6672 (TARGET_DF_FPR): Likewise.
6673 (TARGET_SF_INSN): Macros to say whether floating point support
6674 exists for a given operation for expanders.
6675 (TARGET_DF_INSN): Likewise.
6677 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
6678 combining of SF/DF mode operations, using both traditional and VSX
6686 (abs<mode>2): Combine SF/DF modes using traditional floating point
6687 instructions. Add support for using the upper DF registers with
6688 VSX support, and SF registers with power8-vector support. Update
6689 expanders for operations supported by both the SPE and traditional
6690 floating point units.
6691 (abs<mode>2_fpr): Likewise.
6692 (nabs<mode>2): Likewise.
6693 (nabs<mode>2_fpr): Likewise.
6694 (neg<mode>2): Likewise.
6695 (neg<mode>2_fpr): Likewise.
6696 (add<mode>3): Likewise.
6697 (add<mode>3_fpr): Likewise.
6698 (sub<mode>3): Likewise.
6699 (sub<mode>3_fpr): Likewise.
6700 (mul<mode>3): Likewise.
6701 (mul<mode>3_fpr): Likewise.
6702 (div<mode>3): Likewise.
6703 (div<mode>3_fpr): Likewise.
6704 (sqrt<mode>3): Likewise.
6705 (sqrt<mode>3_fpr): Likewise.
6706 (fre<Fs>): Likewise.
6707 (rsqrt<mode>2): Likewise.
6708 (cmp<mode>_fpr): Likewise.
6709 (smax<mode>3): Likewise.
6710 (smin<mode>3): Likewise.
6711 (smax<mode>3_vsx): Likewise.
6712 (smin<mode>3_vsx): Likewise.
6713 (negsf2): Delete SF operations that are merged with DF.
6720 (fmasf4_fpr): Likewise.
6721 (fmssf4_fpr): Likewise.
6722 (nfmasf4_fpr): Likewise.
6723 (nfmssf4_fpr): Likewise.
6724 (sqrtsf2): Likewise.
6725 (rsqrtsf_internal1): Likewise.
6726 (smaxsf3): Likewise.
6727 (sminsf3): Likewise.
6728 (cmpsf_internal1): Likewise.
6729 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6730 (negdf2): Delete DF operations that are merged with SF.
6732 (nabsdf2): Likewise.
6738 (rsqrtdf_internal1): Likewise.
6739 (fmadf4_fpr): Likewise.
6740 (fmsdf4_fpr): Likewise.
6741 (nfmadf4_fpr): Likewise.
6742 (nfmsdf4_fpr): Likewise.
6743 (sqrtdf2): Likewise.
6744 (smaxdf3): Likewise.
6745 (smindf3): Likewise.
6746 (cmpdf_internal1): Likewise.
6747 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6748 (btrunc<mode>2): Delete separate expander, and combine with the
6749 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6750 (btrunc<mode>2_fpr): Likewise.
6751 (ceil<mode>2): Likewise.
6752 (ceil<mode>2_fpr): Likewise.
6753 (floor<mode>2): Likewise.
6754 (floor<mode>2_fpr): Likewise.
6755 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6756 Add support for using the upper registers with VSX and
6757 power8-vector. Move insns to be closer to the define_expands. On
6758 VSX systems, prefer the traditional form of FMA over the VSX
6759 version, since the traditional form allows the target not to
6760 overlap with the inputs.
6761 (fms<mode>4_fpr): Likewise.
6762 (nfma<mode>4_fpr): Likewise.
6763 (nfms<mode>4_fpr): Likewise.
6765 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6766 Richard Earnshaw <richard.earnshaw@arm.com>
6768 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6769 (struct mult_cost_table): Likewise.
6770 (struct mem_cost_table): Likewise.
6771 (struct fp_cost_table): Likewise.
6772 (struct vector_cost_table): Likewise.
6773 (cpu_cost_table): Likewise.
6774 * config/arm/arm.opt (mold-rts-costs): New option.
6775 (mnew-generic-costs): Likewise.
6776 * config/arm/arm.c (generic_extra_costs): New table.
6777 (cortexa15_extra_costs): Likewise.
6778 (arm_slowmul_tune): Use NULL as new costs.
6779 (arm_fastmul_tune): Likewise.
6780 (arm_strongarm_tune): Likewise.
6781 (arm_xscale_tune): Likewise.
6782 (arm_9e_tune): Likewise.
6783 (arm_v6t2_tune): Likewise.
6784 (arm_cortex_a5_tune): Likewise.
6785 (arm_cortex_a9_tune): Likewise.
6786 (arm_v6m_tune): Likewise.
6787 (arm_fa726te_tune): Likewise.
6788 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6789 (arm_cortex_tune): Use generict_extra_costs.
6790 (shifter_op_p): New function.
6791 (arm_unspec_cost): Likewise.
6792 (LIBCALL_COST): Define.
6793 (arm_new_rtx_costs): New function.
6794 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6795 table is available. Use old costs otherwise unless mnew-generic-costs
6797 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6798 (cpu_cost_table): Declare.
6800 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6803 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6804 (*subs_mul_imm_<mode>)
6805 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6806 (*sub_<shift>_<mode>)
6807 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6808 Remove k constraint.
6810 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6812 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6814 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6816 2013-10-02 Teresa Johnson <tejohnson@google.com>
6818 * predict.c (probably_never_executed): New function.
6819 (probably_never_executed_bb_p): Invoke probably_never_executed.
6820 (probably_never_executed_edge_p): Ditto.
6821 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6822 Treat profile insanities conservatively.
6824 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6826 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6828 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6830 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6831 use it. Use smaller increase for scratch. Don't increase reject
6832 for early clobber scratch.
6833 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6834 setting eliminated regs except setting fp from hfp.
6835 (lra_eliminate): Check lra_insn_recog_data on NULL.
6837 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6840 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6841 setting -mvsx-timode by default until the underlying problem is fixed.
6842 (RS6000_CPU, power7 defaults): Likewise.
6844 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6846 * config/x-linux (host-linux.o): Remove header dependencies.
6847 Use $(COMPILE) and $(POSTCOMPILE).
6848 * config/t-linux-android (linux-android.o): Ditto.
6850 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6852 * Makefile.in (expmed.o-warn): Remove.
6854 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6856 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6857 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6859 2013-10-02 Teresa Johnson <tejohnson@google.com>
6861 * dojump.c (do_jump_1): Divide probability between
6862 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6864 2013-10-02 Tom Tromey <tromey@redhat.com>
6866 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6868 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6870 * tree-flow.h: Remove some prototypes.
6871 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6872 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6873 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6874 mark_virtual_phi_result_for_renaming): Relocate here.
6875 * tree-into-ssa.h: Add prototypes.
6876 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6877 single_pred_before_succ_order.
6878 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6879 (nonfreeing_call_p) Move to gimple.c.
6880 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6882 * basic-block.h (single_pred_before_succ_order): Add prototype.
6883 * gimple.c (nonfreeing_call_p): Relocate here.
6884 * gimple.h: Add prototype.
6885 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6886 * tree-ssa-dom.h: New file. Relocate prototypes here.
6887 * tree-ssa.h: Include tree-ssa-dom.h.
6889 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6891 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6892 Use $(COMPILE) and $(POSTCOMPILE).
6894 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6896 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6898 * tree-flow.h: Remove some prototypes.
6899 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6900 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6901 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6902 propagate_tree_value*): Move from here to...
6903 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6904 propagate_tree_value*): Relocate here.
6905 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6906 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6907 * gimple-fold.c: Remove gimple-fold.h from include list.
6908 * tree-vrp.c: Remove gimple-fold.h from include list.
6909 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6910 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6911 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6912 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6913 * sese.c: Add tree-ssa-propagate.h to include list.
6915 2013-10-02 Richard Biener <rguenther@suse.de>
6917 * tree-loop-distribution.c: Include tree-vectorizer.h for
6919 (enum partition_kind): Remove PKIND_REDUCTION.
6920 (struct partition_s): Remove has_writes member, add reduction_p member.
6921 (partition_alloc): Adjust.
6922 (partition_builtin_p): Likewise.
6923 (partition_has_writes): Remove.
6924 (partition_reduction_p): New function.
6925 (partition_merge_into): Likewise.
6926 (generate_code_for_partition): Commonize builtin partition
6928 (rdg_cannot_recompute_vertex_p): Remove.
6929 (already_processed_vertex_p): Likewise.
6930 (rdg_flag_vertex): Do not set has_writes.
6931 (classify_partition): Adjust.
6932 (rdg_build_partitions): Do not set has_writes, treat all
6933 partitions as useful.
6934 (distribute_loop): Record number of library calls generated. Adjust.
6935 (tree_loop_distribution): Report number of loops and library
6936 calls generated as opt-info.
6938 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6940 * tree-flow.h: Include new .h files. Move prototypes.
6941 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6942 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6943 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6944 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6945 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6946 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6947 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6948 * tree.h (get_ref_base_and_extent): Move prototype out.
6949 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6951 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6952 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6953 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6954 * tree-scalar-evolution.c: Include tree.h.
6955 * sese.c: Include tree.h.
6956 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6957 * dwarf2out.c: Include tree-dfa.h.
6958 * tree-chrec.c: Include tree.h.
6959 * tree-data-ref.c: Include tree.h.
6961 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6963 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6966 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6968 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6969 Use $(COMPILE) and $(POSTCOMPILE).
6970 (sol2-cxx.o): Likewise.
6971 (sol2-stubs.o): Likewise.
6973 * config/x-solaris (host-solaris.o): Likewise.
6975 * config/sparc/t-sparc (sparc.o): Remove.
6976 (sparc-c.o): Remove header dependencies.
6977 Use $(COMPILE) and $(POSTCOMPILE).
6978 * config/sparc/x-sparc: Likewise.
6980 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6982 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6983 * config/arc/arc700.md: Likewise.
6984 * config/arc/arc-modes.def: Likewise.
6985 * config/arc/arc-simd.h: Likewise.
6986 * config/arc/t-arc-uClibc: Likewise.
6987 * config/arc/t-arc-newlib: Likewise.
6989 2013-10-02 Renlin Li <renlin.li@arm.com>
6991 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6993 (aarch64_expand_epilogue): Likewise.
6995 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6996 Yufeng Zhang <yufeng.zhang@arm.com>
6998 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
7000 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
7001 'base_in' represent a conversion and legal_cast_p_1 holds; set
7002 'base_in' with the returned value from get_unwidened.
7004 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7006 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
7007 plus_constant is not used.
7009 2013-10-01 Wei Mi <wmi@google.com>
7011 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
7012 * config/i386/i386.md: Add define_peephole2 to
7013 break partial reg stall for cvtss2sd/cvtsd2ss.
7015 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7017 * config/arc/arc.c (pass_arc_ifcvt::clone):
7018 Update for ctxt_ -> m_ctxt change.
7020 2013-10-01 Jeff Law <law@redhat.com>
7022 * tree-ssa-threadupdate.c (struct redirection_data): Delete
7023 outgoing_edge and intermediate_edge fields. Instead store the path.
7024 (redirection_data::hash): Hash on the last edge's destination index.
7025 (redirection_data::equal): Check the entire thread path.
7026 (lookup_redirectio_data): Corresponding changes.
7027 (create_edge_and_update_destination_phis): Likewise.
7028 (thread_single_edge): Likewise.
7030 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7031 Diego Novillo <dnovillo@google.com>
7033 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
7034 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
7035 (vld32wh_insn, vld32wl_insn): Delete commented-out old
7036 versions of these patterns.
7038 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
7039 (__builtin_arc_aligned): Likewise.
7041 * config/arc/arc.md: Expand adc_0 comment stating the intended
7042 purpose and why it isn't ready.
7043 Replace commented out call_value_via_label_mixed with a
7044 plain comment about bl_s.
7046 * config/arc/arc.c (stdio.h): Don't include directly.
7047 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
7048 Remove if (1) condition.
7049 (arc_encode_section_info): Fix comment.
7051 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7053 * config/arc/arc.c (arc_conditional_register_usage):
7054 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
7055 Also set reg_alloc_order for DMA config regs.
7057 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7058 Jeremy Bennett <jeremy.bennett@embecosm.com>
7060 * doc/install.texi (--with-cpu): Mention ARC.
7061 (arc-*-elf32): New paragraph.
7062 (arc-linux-uclibc): Likewise.
7063 * doc/md.texi (Machine Constraints): Add ARC part.
7064 * doc/invoke.texi: (menu): Add ARC Options.
7065 (Machine Dependent Options) <ARC Options>: Add synopsis.
7066 (node ARC Options): Add.
7067 * doc/extend.texi (long_call / short_call attribute): Add ARC.
7068 (ARC Built-in Functions): New section defining
7069 generic ARC built-in functions.
7070 (ARC SIMD Built-in Functions): New section defining SIMD specific
7072 (Declaring Attributes of Functions): Extended
7073 description of short_call and long_call attributes for ARC and
7074 added index entries.
7076 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
7077 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
7078 Joern Rennecke <joern.rennecke@embecosm.com>
7079 Muhammad Khurram Riaz <khurram.riaz@arc.com>
7080 Brendan Kehoe <brendan@zen.org>
7081 Michael Eager <eager@eagercon.com>
7082 Simon Cook <simon.cook@embecosm.com>
7083 Jeremy Bennett <jeremy.bennett@embecosm.com>
7085 * config/arc, common/config/arc: New directories.
7087 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
7088 Brendan Kehoe <brendan@zen.org>
7089 Simon Cook <simon.cook@embecosm.com>
7091 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
7093 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
7095 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
7096 * tree-ssa-coalesce.h: New. Move prototype to here.
7097 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
7098 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
7099 (gimple_can_coalesce_p): Move to...
7100 * gimple.c (gimple_can_coalesce_p): Here.
7102 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
7104 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
7105 (dump_decl_set): Move to gimple.c.
7106 * gimple.h: Don't include tree-ssa-operands.h.
7107 (dump_decl_set): Add prototype.
7108 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
7109 Move to gimple-ssa.h.
7110 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
7111 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
7112 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
7113 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
7114 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
7116 (dump_decl_set): Relocate here.
7117 * gimple-ssa.h: New file.
7118 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
7119 Relocate from gimple.h.
7120 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
7121 * tree-ssa-operands.c (swap_ssa_operands): Rename from
7122 swap_tree_operands and remove non-ssa path.
7123 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
7124 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
7126 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
7127 vect_is_simple_reduction_1): Use swap_ssa_operands.
7128 * tree-flow.h: Move various prototypes to tree-phinodes.h.
7129 (enum need_phi_state): Move to tree-into-ssa.c.
7130 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
7131 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
7132 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
7133 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
7134 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
7135 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
7136 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
7137 num_imm_uses): Move to ssa-iterators.h.
7138 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
7139 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
7141 (op_iter_done, op_iter_next_def, op_iter_next_tree,
7142 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
7143 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
7144 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
7145 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
7146 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
7147 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
7148 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
7149 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
7150 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
7151 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
7152 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
7153 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
7154 (set_phi_nodes): Move to tree-phinodes.h.
7155 * tree-ssa-operands.h (enum ssa_op_iter_type,
7156 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
7157 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
7158 (dump_decl_set): Remove prototype.
7159 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
7160 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
7161 (set_phi_nodes): Relocate from tree-flow-inline.h.
7162 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
7164 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
7165 include list. Temporarily add gimple.h to include list.
7166 * ssa-iterators.h: New file.
7167 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
7168 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
7169 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
7170 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
7171 Relocate from tree-ssa-operands.h.
7172 (delink_imm_use, link_imm_use_to_list, link_imm_use,
7173 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
7174 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
7175 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
7176 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
7177 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
7178 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
7179 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
7180 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
7181 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
7182 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
7183 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
7184 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
7185 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
7186 Relocate from tree-flow-inline.h.
7187 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
7189 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
7192 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
7193 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
7194 Insert '\t' to output template.
7195 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
7196 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
7197 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
7198 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
7200 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
7202 * doc/install.texi (Host/target specific installation notes for GCC):
7203 Put @anchor before @heading.
7204 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
7205 Use @email for email addresses.
7207 2013-10-01 Jeff Law <law@redhat.com>
7209 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
7210 a vec. Only delete the path if we create one without successfully
7211 registering a jump thread.
7212 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
7214 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
7215 (paths): New vector of jump threading paths.
7216 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
7217 (THREAD_PATH): New accessor macro for the entire thread path.
7218 (lookup_redirection_data): Get intermediate and final outgoing edge
7219 from the thread path.
7220 (create_edge_and_update_destination_phis): Copy the threading path.
7221 (ssa_fix_duplicate_block_edges): Get edges and block types from the
7222 jump threading path.
7223 (ssa_redirect_edges): Get edges and block types from the jump threading
7224 path. Free the path vector.
7225 (thread_block): Get edges from the jump threading path. Look at the
7226 entire path to see if we thread to a loop exit. If we cancel a jump
7227 thread request, then free the path vector.
7228 (thread_single_edge): Get edges and block types from the jump threading
7229 path. Free the path vector.
7230 (thread_through_loop_header): Get edges and block types from the jump
7231 threading path. Free the path vector.
7232 (mark_threaded_blocks): Iterate over the vector of paths and store
7233 the path on the appropriate edge. Get edges and block types from the
7234 jump threading path.
7235 (mark_threaded_blocks): Get edges and block types from the jump
7236 threading path. Free the path vector.
7237 (thread_through_all_blocks): Use the vector of paths rather than
7238 a vector of 3-edge sets.
7239 (register_jump_thread): Accept pointer to a path vector rather
7240 than the path vector itself. Store the path vector for later use.
7243 2013-10-01 Jakub Jelinek <jakub@redhat.com>
7244 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7247 * config/s390/s390.c (s390_split_branches): Modify check for table
7249 (s390_chunkify_start): Rearrange table jump insn check in order to
7250 deal with compare and branch insns correctly.
7252 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
7256 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
7257 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
7258 define_insn_and_split.
7259 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
7260 (shiftsi3_compare): New pattern.
7262 * config/arm/unspecs.md (UNSPEC_RRX): New.
7264 2013-10-01 Alan Modra <amodra@gmail.com>
7266 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
7267 casing inout operands.
7269 2013-10-01 Richard Biener <rguenther@suse.de>
7271 PR tree-optimization/58553
7272 * tree-loop-distribution.c (struct partition_s): Add niter member.
7273 (classify_partition): Populate niter member for the partition
7274 and properly identify whether the relevant store happens before
7275 or after the loop exit.
7276 (generate_memset_builtin): Use niter member from the partition.
7277 (generate_memcpy_builtin): Likewise.
7279 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7281 * vec.h (vec_prefix, vec): Prefix member names with "m_".
7282 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
7284 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7286 * basic-block.h (edge_list): Prefix member names with "m_".
7287 * context.h (context): Likewise.
7288 * domwalk.h (dom_walker): Likewise.
7289 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
7290 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
7291 * hash-table.h (hash_table): Likewise.
7292 * machmode.h (bit_field_mode_iterator): Likewise.
7293 * pass_manager.h (pass_list): Likewise.
7294 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
7295 * tree-pass.h (pass_data): Likewise.
7296 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
7297 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
7298 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
7299 * asan.c (pass_data_asan): Update accordingly.
7300 * cfganal.c (control_dependences::find_control_dependence): Likewise.
7301 (control_dependences::control_dependences): Likewise.
7302 (control_dependences::~control_dependences): Likewise.
7303 (control_dependences::~control_dependences): Likewise.
7304 (control_dependences::get_edges_dependent_on): Likewise.
7305 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
7306 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
7307 * context.c (gcc::context::context): Likewise.
7308 * cprop.c (pass_rtl_cprop::clone): Likewise.
7309 * domwalk.c (dom_walker::walk): Likewise.
7310 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
7311 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
7312 * mode-switching.c (pass_mode_switching::clone): Likewise.
7313 * passes.c (opt_pass::opt_pass): Likewise.
7314 (pass_manager::pass_manager): Likewise.
7315 * predict.c (pass_strip_predict_hints::clone): Likewise.
7316 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
7317 (pass_split_all_insns::clone): Likewise.
7318 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
7320 (bit_field_mode_iterator::next_mode): Likewise.
7321 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
7322 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
7323 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
7324 * tree-complex.c (pass_lower_complex::clone): Likewise.
7325 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
7326 * tree-object-size.c (pass_object_sizes::clone): Likewise.
7327 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
7328 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
7329 (pass_fold_builtins::clone): Likewise.
7330 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
7331 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
7332 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
7333 (pass_cd_dce::clone): Likewise.
7334 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
7335 (pass_phi_only_cprop::clone): Likewise.
7336 * tree-ssa-dse.c (pass_dse::clone): Likewise.
7337 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
7338 * tree-ssa-loop.c (pass_lim::clone): Likewise.
7339 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
7340 * tree-ssa-pre.c (pass_fre::clone): Likewise.
7341 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
7342 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
7343 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
7344 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
7345 * tree-vrp.c (pass_vrp::clone): Likewise.
7346 * tsan.c (pass_tsan::clone): Likewise.
7348 2013-09-30 Jakub Jelinek <jakub@redhat.com>
7351 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
7352 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
7355 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
7356 optimization, punt if sign_bit_p looked through any zero extension.
7358 2013-09-30 Teresa Johnson <tejohnson@google.com>
7360 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
7361 Update redirected out edge count in joiner case.
7362 (ssa_redirect_edges): Common the joiner and non-joiner cases
7363 so that joiner case gets profile updates.
7365 2013-09-30 Richard Biener <rguenther@suse.de>
7367 PR tree-optimization/58554
7368 * tree-loop-distribution.c (classify_partition): Require
7369 unconditionally executed stores for memcpy and memset recognition.
7370 (tree_loop_distribution): Calculate dominance info.
7372 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
7374 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
7375 (NO_PROFILE_COUNTERS): Likewise.
7376 (PROFILE_HOOK): Likewise.
7377 (FUNCTION_PROFILER): Likewise.
7378 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
7380 2013-09-30 Iain Sandoe <iain@codesourcery.com>
7382 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
7383 calls and defines in TARGET_MACHO conditional.
7384 (load_macho_picbase_di): Likewise.
7385 (reload_macho_picbase): Likewise.
7386 (reload_macho_picbase_si): Likewise.
7387 (reload_macho_picbase_di): Likewise.
7388 (nonlocal_goto_receiver): Likewise.
7390 2013-09-30 Nick Clifton <nickc@redhat.com>
7392 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
7393 that use the MSP430X ISA.
7394 (msp430_option_override): Scan -mmcu command line option for any
7395 MCU name that supports the MSP430X ISA.
7396 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
7397 -mmcu options which enable the MSP430X ISA.
7399 2013-09-30 Richard Biener <rguenther@suse.de>
7402 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
7403 before looking for setjmp-like calls.
7405 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7408 * config/darwin-protos.h (machopic_get_function_picbase): New.
7409 * config/darwin.c (machopic_get_function_picbase): New.
7410 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
7411 label for a new func. (load_macho_picbase_di): Likewise.
7412 (reload_macho_picbase): New expand.
7413 (reload_macho_picbase_si): New insn.
7414 (reload_macho_picbase_di): New insn.
7415 (nonlocal_goto_receiver): New define and split.
7416 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
7417 (unspecv enum): Add UNSPECV_NLGR.
7419 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7421 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
7422 that altivec registers are correctly sized on Darwin.
7424 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7426 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
7427 darwin-driver.o): Use COMPILE and POSTCOMPILE.
7428 * config/x-darwin (host-darwin.o): Likewise.
7429 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
7430 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
7431 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
7433 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
7435 * doc/invoke.texi: Fix usage of @tie{} command.
7437 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
7439 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
7441 2013-09-28 Jan Hubicka <jh@suse.cz>
7443 * config/i386/x86-tune.def: Add documentation for each of the options;
7446 2013-09-28 Jan Hubicka <jh@suse.cz>
7448 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
7450 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
7451 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
7452 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
7453 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
7454 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
7456 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
7458 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
7459 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
7460 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
7461 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
7462 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
7463 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
7464 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
7465 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
7466 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
7467 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
7468 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
7469 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
7470 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
7471 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
7472 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
7473 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
7474 gimple.h, godump.c, graphite-clast-to-gimple.c,
7475 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
7476 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
7477 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
7478 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
7479 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
7480 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
7481 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
7482 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
7483 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
7484 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
7485 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
7486 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
7487 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
7488 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
7489 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
7490 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
7491 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
7492 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
7493 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
7494 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
7495 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
7496 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
7497 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
7498 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
7499 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7500 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7501 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
7502 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
7503 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
7504 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7505 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7506 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
7507 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
7508 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
7509 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7510 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
7511 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
7512 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
7513 whitespace before "(".
7515 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
7517 * expr.h (extract_bit_field): Remove packedp parameter.
7518 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
7519 from forward declaration.
7520 (store_split_bit_field): Remove packedp arg from calls to
7521 extract_fixed_bit_field.
7522 (extract_bit_field_1): Remove packedp parameter and packedp
7523 argument from recursive calls and calls to extract_fixed_bit_field.
7524 (extract_bit_field): Remove packedp parameter and corresponding
7525 arg to extract_bit_field_1.
7526 (extract_fixed_bit_field): Remove packedp parameter. Remove code
7528 (extract_split_bit_field): Remove packedp arg from call to
7529 extract_fixed_bit_field.
7530 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
7531 (copy_blkmode_from_reg): Likewise.
7532 (copy_blkmode_to_reg): Likewise.
7533 (read_complex_part): Likewise.
7534 (store_field): Likewise.
7535 (expand_expr_real_1): Likewise.
7536 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
7537 to extract_bit_field.
7538 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
7539 call to extract_bit_field.
7540 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
7541 call to extract_bit_field.
7542 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
7543 and special packedp behavior from -fstrict-volatile-bitfields
7546 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7548 * lra-eliminations.c (init_elim_table): Guard value_p.
7550 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7552 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
7553 DFmode, DImode, and SFmode in the upper VSX registers based on the
7554 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
7555 if -mpower8-vector. Combine -mvsx-timode handling with the rest
7556 of the VSX register handling.
7558 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
7560 (zero_extendsidi2_lfiwzx): Add support for loading into the
7561 Altivec registers with -mpower8-vector. Use wu/wv constraints to
7562 only do VSX memory options on Altivec registers.
7563 (extendsidi2_lfiwax): Likewise.
7564 (extendsfdf2_fpr): Likewise.
7565 (mov<mode>_hardfloat, SF/SD modes): Likewise.
7566 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
7567 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
7568 (movdi_internal64): Likewise.
7570 2013-09-27 Xinliang David Li <davidxl@google.com>
7572 * opts.c (finish_options): Adjust parameters
7573 according to vect cost model.
7574 (common_handle_option): Set dynamic vect cost
7576 targhooks.c (default_add_stmt_cost): Compute stmt cost
7578 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
7579 Use helper function.
7580 * tree-vectorizer.h (unlimited_cost_model): New function.
7581 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
7582 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
7584 (vect_enhance_data_refs_alignment): Ditto.
7585 * flag-types.h: New enum.
7586 * common/config/i386/i386-common.c (ix86_option_init_struct):
7587 No need to initialize vect_cost_model flag.
7588 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
7591 2013-09-27 Diego Novillo <dnovillo@google.com>
7593 * gimple.h (enum ssa_mode): Remove.
7595 2013-09-27 Paulo Matos <pmatos@broadcom.com>
7597 * cfgloop.h (number_of_loops): Fix typo in check for null.
7599 2013-09-27 Jakub Jelinek <jakub@redhat.com>
7602 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
7603 are children of outermost saved_cfun's loop, and set it up to
7604 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
7605 if loop != loop0 and SESE region contains bbs that belong to loop0.
7607 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7609 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
7610 (binary_scale_code_p, get_base_term, get_index_term): New functions.
7611 (set_address_segment, set_address_base, set_address_index)
7612 (set_address_disp): Accept the argument unconditionally.
7613 (baseness): Remove must_be_base_p and must_be_index_p checks.
7614 (decompose_normal_address): Classify as much as possible in the
7617 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7619 * cse.c (count_reg_usage): Handle INT_LIST.
7620 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
7621 * reginfo.c (reg_scan_mark_refs): Likewise.
7622 * reload1.c (eliminate_regs_1): Likewise.
7624 2013-09-27 Iain Sandoe <iain@codesourcery.com>
7627 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
7630 2013-09-27 Richard Biener <rguenther@suse.de>
7632 PR tree-optimization/58459
7633 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
7634 restriction not propagating into loops.
7636 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
7638 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
7639 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
7640 * doc/tree-ssa.texi (Walking use-def chains): Delete.
7642 2013-09-26 Richard Biener <rguenther@suse.de>
7644 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
7646 2013-09-26 Richard Biener <rguenther@suse.de>
7648 * alias.h (component_uses_parent_alias_set): Rename to ...
7649 (component_uses_parent_alias_set_from): ... this.
7650 * alias.c (component_uses_parent_alias_set): Rename to ...
7651 (component_uses_parent_alias_set_from): ... this and return
7653 (reference_alias_ptr_type_1): Use the result from
7654 component_uses_parent_alias_set_from instead of stripping
7655 components one at a time.
7656 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
7658 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7660 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
7661 Move prototypes to...
7662 * tree-ssa-ter.h: New File. Move prototypes here.
7663 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
7664 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
7665 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7666 from is_replaceable_p.
7667 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
7668 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
7669 (process_replaceable): Use ter_is_replaceable_p.
7670 (find_replaceable_in_bb): Use ter_is_replaceable_p.
7671 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
7672 newly refactored ssa_is_replaceable_p.
7673 * cfgexpand.c: Include tree-outof-ssa.h.
7674 * ssaexpand.h: Delete.
7676 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7678 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
7679 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
7681 (create_gimple_tmp): Delete.
7682 (get_expr_type, build_assign, build_type_cast): Move to...
7683 * gimple-builder.c: New File.
7684 (get_expr_type): Relocate from gimple.c.
7685 (build_assign, build_type_cast): Change to only create ssanames.
7686 * gimple.h: Move prototypes to...
7687 * gimple-builder.h: New File. Here.
7688 * tree-ssa.h: And here.
7689 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
7690 count_uses_and_derefs): Relocate from gimple.c.
7691 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
7692 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
7693 * tree-ssa-math-opts (execute_cse_reciprocals): Use
7694 gimple_replace_ssa_lhs.
7695 * asan.c: Include gimple-builder.h.
7696 * Makefile.in: Add gimple-builder.o.
7698 2013-09-26 Richard Biener <rguenther@suse.de>
7700 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
7701 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
7702 (loe_visit_block): Use gcc_checking_assert.
7703 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
7704 gimple_assign_ssa_name_copy_p.
7705 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
7708 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
7710 * config/rs6000/t-rs6000 (rs6000.o): Remove.
7711 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
7713 2013-09-26 Richard Biener <rguenther@suse.de>
7715 PR tree-optimization/58539
7716 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
7717 the fact that debug statements are not taking part in loop-closed
7720 2013-09-26 Nick Clifton <nickc@redhat.com>
7722 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
7723 time warning message.
7724 (msp430_print_operand_raw): Delete unused letter parameter.
7725 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7726 (msp430_print_operand_address): New function.
7727 (msp430_print_operand): Move address printing code from here to
7729 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7731 (zero_extendpsisi2): Likewise.
7732 (extendpsisi2): New pattern.
7733 (andneghi3): New pattern.
7735 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7737 * config/aarch64/aarch64.opt (mlra): New option.
7738 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7739 (TARGET_LRA_P): Define.
7741 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7743 * expr.c (expand_assignment): Remove obsolete comment.
7745 2013-09-25 Jeff Law <law@redhat.com>
7747 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7748 tree-ssa-threadupdate.h.
7749 (register_jump_thread): Similarly.
7750 * tree-ssa-threadupdate.h: New header file.
7751 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7752 * tree-vrp.c: Likewise.
7753 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7754 (thread_around_empty_blocks): Change type of path vector argument to
7755 an edge,type pair from just an edge. Initialize both elements when
7756 appending to a jump threading path. Tweak references to elements
7758 (thread_across_edge): Similarly. Release memory for the elements
7760 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7761 (dump_jump_thread_path): New function broken out from
7762 register_jump_thread.
7763 (register_jump_thread): Use dump_jump_thread_path. Change type of
7764 path vector entries. Search the path for NULL edges and dump
7765 the path if one is found. Tweak the conversion of path to 3-edge
7766 form to use the block copy type information embedded in the path.
7768 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7770 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7772 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7773 Vladimir Makarov <vmakarov@redhat.com>
7775 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7776 from the least significant bit.
7777 (strip_address_mutations): Add bitfield operations handling.
7778 (must_be_index_p): Add shifting and rotate operations handling.
7779 (set_address_base): Use must_be_base_p predicate.
7780 (set_address_index): Use must_be_index_p predicate.
7782 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7783 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7784 Sergey Lega <sergey.s.lega@intel.com>
7785 Anna Tikhonova <anna.tikhonova@intel.com>
7786 Ilya Tocar <ilya.tocar@intel.com>
7787 Andrey Turetskiy <andrey.turetskiy@intel.com>
7788 Ilya Verbin <ilya.verbin@intel.com>
7789 Kirill Yukhin <kirill.yukhin@intel.com>
7790 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7792 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7794 (ix86_expand_vector_move_misalign): Support new unaligned load and
7795 stores and use new names.
7796 (CODE_FOR_sse2_storedqu): Rename to ...
7797 (CODE_FOR_sse2_storedquv16qi): ... this.
7798 (CODE_FOR_sse2_loaddqu): Rename to ...
7799 (CODE_FOR_sse2_loaddquv16qi): ... this.
7800 (CODE_FOR_avx_loaddqu256): Rename to ...
7801 (CODE_FOR_avx_loaddquv32qi): ... this.
7802 (CODE_FOR_avx_storedqu256): Rename to ...
7803 (CODE_FOR_avx_storedquv32qi): ... this.
7804 * config/i386/i386.md (fpint_logic): New.
7805 * config/i386/sse.md (VMOVE): Extend for AVX512.
7809 (VI_UNALIGNED_LOADSTORE): Ditto.
7810 (sse2_avx_avx512f): Ditto.
7811 (sse2_avx2): Extend for AVX512.
7812 (sse4_1_avx2): Ditto.
7813 (avx2_avx512f): New.
7814 (sse): Extend for AVX512.
7817 (avxsizesuffix): Ditto.
7818 (sseintvecmode): Ditto.
7820 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7821 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7822 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7823 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7824 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7825 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7826 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7827 (STORENT_MODE): Extend for AVX512.
7828 (*absneg<mode>2): Replace constraint "x" with "v".
7829 (*mul<mode>3): Ditto.
7830 (*ieee_smin<mode>3): Ditto.
7831 (*ieee_smax<mode>3): Ditto.
7832 (avx_cmp<mode>3): Replace VF with VF_128_256.
7833 (*<sse>_maskcmp<mode>3_comm): Ditto.
7834 (<sse>_maskcmp<mode>3): Ditto.
7835 (<sse>_andnot<mode>3): Extend for AVX512.
7836 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7837 (<code><mode>3, fpint_logic): New.
7838 (*<code><mode>3): Extend for AVX512.
7839 (avx512flogicsuff): New.
7840 (avx512f_<logic><mode>): Ditto.
7841 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7843 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7844 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7845 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7846 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7847 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7848 (xop_vpermil2<mode>3): Ditto.
7849 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7850 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7851 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7852 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7854 2013-09-25 Tom Tromey <tromey@redhat.com>
7856 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7857 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7858 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7859 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7860 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7861 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7862 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7863 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7864 (GRAPHITE_HTAB_H): Remove.
7866 2013-09-25 Tom Tromey <tromey@redhat.com>
7868 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7870 2013-09-25 Tom Tromey <tromey@redhat.com>
7872 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7874 2013-09-25 Tom Tromey <tromey@redhat.com>
7876 * config/i386/t-i386 (i386.o): Remove.
7877 (i386-c.o): Use COMPILE and POSTCOMPILE.
7879 2013-09-25 Tom Tromey <tromey@redhat.com>
7881 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7883 2013-09-25 Tom Tromey <tromey@redhat.com>
7885 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7886 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7887 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7888 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7889 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7890 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7891 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7892 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7893 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7894 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7895 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7896 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7897 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7898 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7899 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7900 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7901 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7902 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7903 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7904 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7905 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7906 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7907 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7908 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7909 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7910 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7911 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7912 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7913 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7914 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7915 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7916 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7917 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7918 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7919 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7920 (graphite-clast-to-gimple.o, graphite-dependences.o)
7921 (graphite-interchange.o, graphite-poly.o)
7922 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7923 (graphite-optimize-isl.o, tree-vect-loop.o)
7924 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7925 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7926 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7927 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7928 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7929 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7930 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7931 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7932 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7933 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7934 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7935 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7936 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7937 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7938 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7939 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7940 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7941 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7942 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7943 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7944 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7945 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7946 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7947 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7948 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7949 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7950 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7951 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7952 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7953 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7954 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7955 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7956 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7957 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7958 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7959 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7960 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7961 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7962 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7963 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7964 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7965 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7966 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7967 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7968 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7969 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7970 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7971 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7972 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7973 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7974 (gcov-dump.o): Remove.
7975 (default-c.o): Use COMPILE and POSTCOMPILE.
7976 (CFLAGS-gcc.o): New variable.
7977 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7979 2013-09-25 Tom Tromey <tromey@redhat.com>
7981 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7982 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7983 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7984 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7985 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7986 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7987 (c-family/array-notation-common.o, c-family/stub-objc.o)
7988 (c-family/c-ubsan.o): Remove.
7990 2013-09-25 Tom Tromey <tromey@redhat.com>
7992 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7994 2013-09-25 Tom Tromey <tromey@redhat.com>
7996 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7997 to add -DENABLE_SHARED_LIBGCC.
7998 (gcc.o): Don't use subshell.
8000 2013-09-25 Tom Tromey <tromey@redhat.com>
8002 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
8003 * configure.ac: Don't invoke AM_PROG_CC_C_O.
8004 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
8005 * configure, config.in: Rebuild.
8007 2013-09-25 Tom Tromey <tromey@redhat.com>
8009 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
8010 (COMPILE, POSTCOMPILE): New variables.
8011 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
8012 (DEPFILES): New variable.
8013 Include ".Po" files.
8014 * configure.ac: Add checks for dependency checking.
8015 * configure, aclocal.m4: Regenerate.
8017 2013-09-25 Tom Tromey <tromey@redhat.com>
8019 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
8020 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
8022 2013-09-25 Tom Tromey <tromey@redhat.com>
8024 * Makefile.in (generated_files): Add options.h,
8025 target-hooks-def.h, insn-opinit.h,
8026 common/common-target-hooks-def.h, pass-instances.def,
8027 c-family/c-target-hooks-def.h.
8029 2013-09-25 Jeff Law <law@redhat.com>
8031 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
8032 than foo[foo.length () - 1] to access last member in a vec.
8033 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
8035 2013-09-25 Richard Biener <rguenther@suse.de>
8038 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
8040 2013-09-25 Jan Hubicka <jh@suse.cz>
8042 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
8045 2013-09-25 Marek Polacek <polacek@redhat.com>
8048 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
8049 when determining the type name.
8051 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
8053 * config/sh/sh.md: Fix formatting.
8055 2013-09-24 Xinliang David Li <davidxl@google.com>
8057 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
8058 max peel iterations parameter.
8059 * param.def: New parameter.
8060 * doc/invoke.texi: Document New parameter.
8062 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
8064 * gimple-pretty-print.c: Various whitespace tweaks.
8065 * tree-core.h: Likewise.
8066 * tree-pretty-print.c: Likewise.
8067 * tree-ssa-alias.c: Likewise.
8068 * tree-ssa-copy.c: Likewise.
8069 * tree-ssanames.c: Likewise.
8070 * tree-ssanames.h: Likewise.
8071 * tree-vrp.c: Likewise.
8073 2013-09-24 Alan Modra <amodra@gmail.com>
8077 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
8078 for output operands that disallow regs. Don't use EXPAND_WRITE on
8081 2013-09-24 Richard Biener <rguenther@suse.de>
8084 * tree.c (reference_alias_ptr_type): Move ...
8085 * alias.c (reference_alias_ptr_type): ... here and implement
8086 in terms of the new reference_alias_ptr_type_1.
8087 (ref_all_alias_ptr_type_p): New helper.
8088 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
8089 use ref_all_alias_ptr_type_p.
8090 (get_deref_alias_set): Add flag_strict_aliasing check here.
8091 (reference_alias_ptr_type_1): New function, split out from ...
8092 (get_alias_set): ... here.
8093 (alias_ptr_types_compatible_p): New function.
8094 * alias.h (reference_alias_ptr_type): Declare.
8095 (alias_ptr_types_compatible_p): Likewise.
8096 * tree.h (reference_alias_ptr_type): Remove.
8097 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
8098 to compare MEM_REF alias types.
8100 2013-09-24 Richard Biener <rguenther@suse.de>
8102 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
8104 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8106 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
8107 reload helper function arrays into a single array reg_addr.
8108 (reload_fpr_gpr): Likewise.
8109 (reload_gpr_vsx): Likewise.
8110 (reload_vsx_gpr): Likewise.
8111 (struct rs6000_reg_addr): Likewise.
8112 (reg_addr): Likewise.
8113 (rs6000_debug_reg_global): Change rs6000_vector_reload,
8114 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
8115 (rs6000_init_hard_regno_mode_ok): Likewise.
8116 (rs6000_secondary_reload_direct_move): Likewise.
8117 (rs6000_secondary_reload): Likewise.
8119 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
8120 constraints: wu, ww, and wy. Repurpose wv constraint added during
8121 power8 changes. Put wg constraint in alphabetical order.
8123 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
8124 for future work to add ISA 2.07 VSX single precision support.
8125 (-mvsx-scalar-double): Change default from -1 to 1, update
8126 documentation comment.
8127 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
8128 (-mupper-regs-df): New debug switch to control whether DF values
8129 can go in the traditional Altivec registers.
8130 (-mupper-regs-sf): New debug switch to control whether SF values
8131 can go in the traditional Altivec registers.
8133 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
8135 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
8136 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
8137 Add new constraints, wu/ww/wy. Repurpose wv constraint.
8138 (rs6000_debug_legitimate_address_p): Print if we are running
8139 before, during, or after reload.
8140 (rs6000_secondary_reload): Add a comment.
8141 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
8143 * config/rs6000/constraints.md (wa constraint): Sort w<x>
8144 constraints. Update documentation string.
8145 (wd constraint): Likewise.
8146 (wf constraint): Likewise.
8147 (wg constraint): Likewise.
8148 (wn constraint): Likewise.
8149 (ws constraint): Likewise.
8150 (wt constraint): Likewise.
8151 (wx constraint): Likewise.
8152 (wz constraint): Likewise.
8153 (wu constraint): New constraint for ISA 2.07 SFmode scalar
8155 (ww constraint): Likewise.
8156 (wy constraint): Likewise.
8157 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
8158 in the previous submissions.
8159 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
8161 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
8163 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
8164 (REG_BR_PROB): Say that the probability is stored in an int_list.
8165 * reg-notes.def: Update commentary to mention INT_LIST.
8166 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
8167 (INT_LIST): New rtx.
8168 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
8169 * rtlanal.c (int_reg_note_p): New function.
8170 (alloc_reg_note): Assert that the note does not have an int argument.
8171 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
8172 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
8173 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
8174 rather than an INSN_LIST. Handle INT_LIST.
8175 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
8176 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
8177 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
8178 Manipulate them as ints rather than rtxes.
8179 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
8180 * regmove.c (copy_src_to_dest): Likewise.
8181 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
8183 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
8184 into the cases that need it.
8185 * config/arm/arm.c (arm_unwind_emit): Likewise.
8187 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
8188 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
8189 * loop-doloop.c (add_test, doloop_modify): Likewise.
8190 * loop-unswitch.c (compare_and_jump_seq): Likewise.
8191 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
8192 * predict.c (combine_predictions_for_insn): Likewise.
8193 * print-rtl.c (print_rtx): Handle INT_LIST.
8194 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
8195 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
8196 * config/arm/arm.c (emit_unlikely_jump): Likewise.
8197 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
8198 (ix86_split_fp_branch, predict_jump): Likewise.
8199 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
8200 * config/sh/sh.c (expand_cbranchsi4): Likewise.
8201 * config/spu/spu.c (ea_load_store_inline): Likewise.
8203 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
8204 value of a REG_BR_PROB note.
8205 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
8206 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
8207 * emit-rtl.c (try_split): Likewise.
8208 * predict.c (br_prob_note_reliable_p): Likewise.
8209 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
8210 * reorg.c (mostly_true_jump): Likewise.
8211 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
8212 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
8213 * config/i386/i386.c (ix86_print_operand): Likewise.
8214 * config/ia64/ia64.c (ia64_print_operand): Likewise.
8215 * config/mmix/mmix.c (mmix_print_operand): Likewise.
8216 * config/rs6000/rs6000.c (output_cbranch): Likewise.
8217 * config/s390/s390.c (s390_expand_tbegin): Likewise.
8218 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
8219 * config/sparc/sparc.c (output_cbranch): Likewise.
8220 * config/spu/spu.c (get_branch_target): Likewise.
8221 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
8222 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
8224 2013-09-23 Jan Hubicka <jh@suse.cz>
8226 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
8228 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
8229 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
8231 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
8232 (gimple_fold_call): Dump inconsistent devirtualizations; add
8233 sanity check for type based devirtualizations.
8234 * ipa-prop.c: Include ipa-utils.h
8235 (ipa_intraprocedural_devirtualization): Add sanity check.
8236 (try_make_edge_direct_virtual_call): Likewise.
8238 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
8240 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
8241 assignment statements.
8243 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
8245 * gimple-pretty-print.c (dump_ssaname_info): New function.
8246 (dump_gimple_phi): Call it.
8247 (pp_gimple_stmt_1): Likewise.
8248 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
8249 (range_info_def): Declare.
8250 * tree-pretty-print.c (pp_double_int): New function.
8251 (dump_generic_node): Call it.
8252 * tree-pretty-print.h (pp_double_int): Declare.
8253 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
8254 * tree-ssanames.h (range_info_def): New structure.
8255 (value_range_type): Move definition here.
8256 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
8258 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
8260 (set_range_info): New function.
8261 (get_range_info): Likewise.
8262 (duplicate_ssa_name_range_info): Likewise.
8263 (duplicate_ssa_name_fn): Check pointer type and call
8264 duplicate_ssa_name_range_info.
8265 * tree-ssa-copy.c (fini_copy_prop): Likewise.
8266 * tree-vrp.c (value_range_type): Remove definition, now in
8268 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
8269 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
8270 (SSA_NAME_RANGE_INFO): New macro.
8272 2013-09-23 Richard Biener <rguenther@suse.de>
8274 PR tree-optimization/58464
8275 * tree-ssa-pre.c (phi_trans_lookup): Remove.
8276 (phi_trans_add): Change to add conditionally on being not present.
8277 (phi_translate_1): Remove recursion detection here.
8278 (phi_translate): Pre-seed the cache with NULL to catch
8279 recursion here in a more generic way.
8280 (bitmap_find_leader): Adjust comment.
8281 (get_representative_for): Dump value-numbers.
8282 (create_expression_by_pieces): Likewise.
8283 (insert_into_preds_of_block): Likewise.
8285 2013-09-23 Christian Bruel <christian.bruel@st.com>
8288 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
8289 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
8291 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
8294 2013-09-20 Renlin Li <renlin.li@arm.com>
8296 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
8298 (aarch64_expand_epilogue): Likewise.
8299 (aarch64_legitimize_reload_address): Likewise.
8301 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
8303 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
8304 NULL_TREE before pushing them onto the vector. Likewise for labels.
8306 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
8308 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
8309 comparison operators when -fno-trapping-math is in effect.
8310 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
8311 comparison operators in TFmode and assert that unsupported operators
8313 (ia64_print_operand): Likewise.
8315 2013-09-21 Jan Hubicka <jh@suse.cz>
8317 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
8318 (sse_typeless_stores): Enable for core
8319 (sse_load0_by_pxor): Likewise.
8320 (four_jump_limit): Disable for core.
8321 (pad_returns): Likewise.
8322 (avoid_vector_decode): Likewise.
8323 (fuse_cmp_and_branch): Enable for cores.
8324 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
8326 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8329 * config/pa/pa.c (pa_option_override): Disable auto increment and
8330 decrement instructions until reload is completed.
8332 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
8333 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
8334 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
8336 2013-09-20 DJ Delorie <dj@redhat.com>
8337 Nick Clifton <nickc@redhat.com>
8339 * config/rl78/rl78.c: Various whitespace and comment tweaks.
8340 (need_to_save): Save bank 0 on interrupts.
8341 (characterize_address): Strip far address wrappers.
8342 (rl78_as_legitimate_address): Likewise.
8343 (transcode_memory_rtx): Likewise.
8344 (rl78_peep_movhi_p): Disable this peephole after devirt.
8345 (rl78_propogate_register_origins): Forget all origins when a
8347 * config/rl78/rl78-virt.md: Various whitespace tweaks.
8348 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
8350 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
8351 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
8352 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
8354 * config/rl78/constraints.md: Various whitespace and paren tweaks.
8356 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8358 * config/pa/pa.md: In "scc" insn patterns, change output template to
8359 handle const0_rtx in reg_or_0_operand operands.
8361 2013-09-20 Martin Husemann <martin@NetBSD.org>
8364 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
8365 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
8367 2013-09-20 Richard Biener <rguenther@suse.de>
8370 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
8371 remembering SSA name version and block index.
8372 (new_scev_info_str): Adjust.
8373 (hash_scev_info): Likewise. Also hash the block index.
8374 (eq_scev_info): Adjust.
8375 (find_var_scev_info): Likewise.
8376 (struct instantiate_cache_entry): Remove.
8377 (struct instantiate_cache_type): Use a htab to map name, block
8379 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8380 (get_instantiated_value_entry): Likewise.
8381 (hash_idx_scev_info, eq_idx_scev_info): New functions.
8382 (instantiate_scev_name): Adjust.
8384 2013-09-20 Jeff Law <law@redhat.com>
8386 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
8388 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
8390 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
8391 Call aarch64_simd_expand_args to update op[argc].
8393 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8395 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
8398 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8400 * gengtype.c (file_rules): Added rule for *.cc files.
8401 (get_output_file_with_visibility): Give fatal message when no
8404 2013-09-20 Renlin Li <renlin.li@arm.com>
8406 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
8407 (aarch64_expand_epilogue): Likewise.
8408 (aarch64_legitimize_reload_address): Likewise.
8410 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
8413 * expr.c (expand_assignment): Remove misalignp code path.
8415 2013-09-20 Marek Polacek <polacek@redhat.com>
8418 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
8419 TYPE_PRECISION. Add asserts.
8421 2013-09-20 Richard Biener <rguenther@suse.de>
8423 PR tree-optimization/58453
8424 * tree-loop-distribution.c (distribute_loop): Apply the cost
8425 model for -ftree-loop-distribute-patterns, too.
8427 2013-09-20 Richard Biener <rguenther@suse.de>
8430 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
8431 make type comparison less strict.
8433 2013-09-20 Alan Modra <amodra@gmail.com>
8435 * configure: Regenerate.
8436 * aclocal.m4: Regenerate.
8438 2013-09-20 Marek Polacek <polacek@redhat.com>
8441 * doc/extend.texi: Document that attribute used is meant to be used
8442 on variables with static storage duration.
8444 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8446 PR tree-optimization/58472
8447 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
8448 simd_lane_access set inv_p = false.
8449 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
8450 the simduid magic VAR_DECL.
8452 2013-09-19 Jan Hubicka <jh@suse.cz>
8454 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
8456 2013-09-17 Jeff Law <law@redhat.com>
8458 * tree-ssa-dom.c (record_temporary_equivalences): New function
8459 split out of dom_opt_dom_walker::after_dom_children.
8460 (dom_opt_dom_walker::thread_across_edge): Move common code
8461 in here from dom_opt_dom_walker::after_dom_children.
8462 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
8464 2013-09-19 Jan Hubicka <jh@suse.cz>
8466 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
8467 (TARGET_GENERIC): Use PROCESOR_GENERIC
8468 (enum processor_type): Unify generic32 and 64.
8469 * i386.md (cpu): Likewise.
8470 * x86-tune.def (use_leave): Enable for generic32.
8471 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
8472 * athlon.md: Change generic64 to generic in all occurences.
8473 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
8474 (ix86_target_macros_internal): Likewise.
8475 * driver-i386.c (host_detect_local_cpu): Likewise.
8476 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
8478 (generic_memcpy, generic_memset, generic_cost): This one.
8479 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
8480 (m_GENERIC32, m_GENERIC64): Remove.
8481 (m_GENERIC): Turn into one flag.
8482 (processor_target): Unify generic tunnings.
8483 (ix86_option_override_internal): Replace generic32/64 by generic.
8484 (ix86_issue_rate): Likewise.
8485 (ix86_adjust_cost): Likewise.
8487 2013-09-19 Jan Hubicka <jh@suse.cz>
8489 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
8490 of speculative flag.
8492 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8494 * omp-low.c (expand_omp_sections): Always pass len - 1 to
8495 GOMP_sections_start, even if !exit_reachable.
8497 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8499 * lra-constraints.c (need_for_all_save_p): Use macro
8500 HARD_REGNO_CALL_PART_CLOBBERED.
8501 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
8502 set up pseudo conflict hard regs.
8504 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8507 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
8510 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8512 PR rtl-optimization/58438
8513 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
8514 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
8515 from previous subpasses.
8517 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
8519 * arm.c (arm_get_frame_offsets): Validate architecture supports
8520 LDRD/STRD before accepting the tuning preference.
8521 (arm_expand_prologue): Likewise.
8522 (arm_expand_epilogue): Likewise.
8524 2013-09-18 Richard Biener <rguenther@suse.de>
8526 PR tree-optimization/58417
8527 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
8528 have chrecs with symbols defined in the loop as operands.
8529 (chrec_fold_multiply): Likewise.
8530 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
8531 parameters before folding binary operations.
8532 (struct instantiate_cache_entry_hasher): Remove.
8533 (struct instantiate_cache_type): Use a pointer-map.
8534 (instantiate_cache_type::instantiate_cache_type): New function.
8535 (instantiate_cache_type::get): Likewise.
8536 (instantiate_cache_type::set): Likewise.
8537 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8538 (get_instantiated_value_entry): Likewise.
8539 (global_cache): New global.
8540 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
8541 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
8542 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
8543 (instantiate_scev_name): Adjust.
8544 (instantiate_scev): Construct global instead of local cache.
8545 (resolve_mixers): Likewise.
8547 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
8548 Paolo Carlini <paolo.carlini@oracle.com>
8551 * doc/implement-cxx.texi: Fix references to the C++ standards.
8553 2013-09-18 Jakub Jelinek <jakub@redhat.com>
8555 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
8556 * tree-vect-data-refs.c (vect_analyze_data_refs): For
8557 simd_lane_access drs, update also DR_ALIGNED_TO.
8559 2013-09-18 Marek Polacek <polacek@redhat.com>
8562 * doc/extend.texi: Document no_sanitize_undefined attribute.
8563 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
8564 no_sanitize_undefined attribute.
8566 2013-09-18 Nick Clifton <nickc@redhat.com>
8568 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
8569 (ASM_DECLARE_FUNCTION_NAME): Define.
8571 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
8573 * compare-elim.c (find_comparison_dom_walker): New class
8574 (find_comparisons_in_bb): Rename to
8575 find_comparison_dom_walker::before_dom_children
8576 (find_comparisons): Adjust
8577 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
8579 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
8580 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
8581 (init_walk_dominator_tree): Remove declaration.
8582 (fini_walk_dominator_tree): Remove declaration.
8583 * fwprop.c (single_def_use_dom_walker): New class
8584 (single_def_use_enter_block): Convert to
8585 single_def_use_dom_walker::before_dom_children.
8586 (single_def_use_leave_block): Convert to
8587 single_def_use_dom_walker::after_dom_children.
8588 (build_single_def_use_links): Adjust.
8589 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
8591 (find_candidates_in_block): Convert to
8592 find_candidates_dom_walker::before_dom_children.
8593 (execute_strength_reduction): Adjust.
8594 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
8595 (sese_dom_walker): New class.
8596 (sese_dom_walker::sese_dom_walker): New constructor.
8597 (sese_dom_walker::~sese_dom_walker): New destructor.
8598 (build_sese_conditions_before): Convert to
8599 sese_dom_walker::before_dom_children.
8600 (build_sese_conditions_after): Convert to
8601 sese_dom_walker::after_dom_children.
8602 (build_poly_scop): Adjust
8603 * tree-into-ssa.c (rewrite_dom_walker): New class
8604 (rewrite_enter_block): Convert to
8605 rewrite_dom_walker::before_dom_children.
8606 (rewrite_leave_block): Convert to
8607 rewrite_dom_walker::after_dom_children.
8608 (rewrite_update_dom_walker): New class.
8609 (rewrite_update_enter_block): Convert to
8610 rewrite_update_dom_walker::before_dom_children.
8611 (rewrite_update_leave_block): Convert to
8612 rewrite_update_dom_walker::after_dom_children.
8613 (rewrite_blocks, rewrite_into_ssa): Adjust.
8614 (mark_def_dom_walker): New class.
8615 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
8616 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
8617 (mark_def_sites_blocks): Convert to
8618 mark_def_dom_walker::before_dom_children.
8619 (mark_def_site_blocks): Remove.
8620 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
8621 (tree_ssa_dominator_optimize): Adjust.
8622 (dom_thread_across_edge): Convert to method
8623 dom_opt_dom_walker::thread_across_edge.
8624 (dom_opt_enter_block): Convert to member function
8625 dom_opt_dom_walker::before_dom_children.
8626 (dom_opt_leave_block): Convert to member function
8627 dom_opt_dom_walker::after_dom_children.
8628 * tree-ssa-dse.c (dse_dom_walker): New class.
8629 (dse_enter_block): Convert to member function
8630 dse_dom_walker::before_dom_children.
8631 (tree_ssa_dse): Adjust.
8632 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
8633 (determine_invariantness_stmt): Convert to method
8634 invariantness_dom_walker::before_dom_children.
8635 (determine_invariantness): Remove
8636 (move_computations_dom_walker): New class.
8637 (move_computations_stmt): Convert to method
8638 move_computations_dom_walker::before_dom_children.
8639 (move_computations, tree_ssa_lim): Adjust.
8640 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
8641 (nt_init_block): Convert to method
8642 notrappping_dom_walker::before_dom_children.
8643 (nt_fini_block): Convert to method
8644 method nontrapping_dom_walker::after_dom_children.
8645 (get_non_trapping): Adjust.
8646 * tree-ssa-pre.c (eliminate_dom_walker): New class.
8647 (eliminate_bb): Convert to method
8648 eliminate_dom_walker::before_dom_children.
8649 (eliminate_leave_block): Convert to method
8650 eliminate_dom_walker::after_dom_children.
8651 (eliminate): Adjust.
8652 * tree-ssa-strlen.c (strlen_dom_walker): New class.
8653 (strlen_enter_block): Convert to method
8654 strlen_dom_walker::before_dom_children.
8655 (strlen_leave_block): Convert to method
8656 method strlen_dom_walker::after_dom_children.
8657 (tree_ssa_strlen): Adjust.
8658 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
8659 (tree_ssa_uncprop): Adjust.
8660 (uncprop_leave_block): Convert to method
8661 uncprop_dom_walker::after_dom_children.
8662 (uncprop_leave_block): Convert to method
8663 uncprop_dom_walker::before_dom_children.
8665 2013-09-18 Bin Cheng <bin.cheng@arm.com>
8667 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
8668 branch in same basic block. Check both src and dest of the move insn.
8670 2013-09-17 Nick Clifton <nickc@redhat.com>
8672 * config/rl78/rl78-real.md (bf): New pattern.
8674 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
8675 (rl78_print_operand): Do not put a # before a %B.
8676 * config/rl78/rl78.opt: Tweak doc strings.
8678 2013-09-17 DJ Delorie <dj@redhat.com>
8680 * config/rl78/constraints.md (Wcv): Allow up to $r31.
8681 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
8682 (rl78_option_override): Likewise, if -mallregs.
8683 (is_virtual_register): Likewise.
8684 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
8685 (REGNO_OK_FOR_BASE_P): Likewise.
8686 * config/rl78/rl78.opt (-mallregs): New.
8688 2013-09-17 Nick Clifton <nickc@redhat.com>
8690 * config/rl78/rl78.c (need_to_save): Change return type to bool.
8691 For interrupt functions: save all call clobbered registers if the
8692 interrupt handler is not a leaf function.
8693 (rl78_expand_prologue): Always recompute the frame information.
8694 For interrupt functions: only select bank 0 if one of the bank 0
8695 registers is going to be psuhed.
8697 2013-09-17 DJ Delorie <dj@redhat.com>
8699 * config/rl78/constraints.md: For each W* constraint, rename to C*
8700 and create a W* constraint that checks for an optional ES: prefix
8702 * config/rl78/rl78.md (UNS_ES_ADDR): New.
8703 (es_addr): New. Used to wrap far addresses.
8704 * config/rl78/rl78-protos.h (rl78_es_addr): New.
8705 (rl78_es_base): New.
8706 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
8707 wrapped far addresses.
8708 (rl78_print_operand_1): Unwrap far addresses before processing.
8709 (rl78_lo16): Wrap far addresses in unspecs.
8710 (rl78_es_addr): New.
8711 (rl78_es_base): New.
8712 (insn_ok_now): Check for not-yet-wrapped far addresses.
8713 (transcode_memory_rtx): Properly re-wrap far addresses.
8715 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
8717 * config/sparc/t-rtems: Add leon3 multilibs.
8719 2013-09-17 Cong Hou <congh@google.com>
8721 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
8722 when checking the dot production pattern. The type of rhs operand
8723 of multiply is now checked correctly.
8725 2013-09-17 Jeff Law <law@redhat.com>
8727 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8728 edge implied equivalences into successor phis.
8729 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8730 here from tree-ssa-threadedge.c.
8731 (mark_threaded_blocks): When threading through a joiner, if both
8732 successors of the joiner's clone reach the same block, verify the
8733 PHI arguments are equal. If not, cancel the jump threading request.
8734 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8735 tree-ssa-threadupdate.c
8736 (thread_across_edge): Don't check PHI argument equality when
8737 threading through joiner block here.
8739 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8741 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8742 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8743 (warn_uninit, warn_uninitialized_vars,
8744 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8745 Move to tree-ssa-uninit.c.
8746 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8747 (has_undefined_value_p): New. Pass dependant parts of
8748 ssa_undefined_value_p.
8749 (uninit_undefined_value_p): Use has_undefined_value_p.
8750 (warn_uninit, warn_uninitialized_vars,
8751 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8752 Move from tree-ssa.c.
8753 * tree-ssa.h: Adjust prototypes.
8755 2013-09-17 Jan Hubicka <jh@suse.cz>
8758 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8759 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8760 FUNCTION_NOT_OPTIMIZED.
8761 * ipa-inline-analysis.c (compute_inline_parameters): Function
8762 not optimized is not inlinable unless it is alwaysinline.
8763 (inline_analyze_function): Force calls in not optimized
8764 function not inlinable.
8766 2013-09-17 Jan Hubicka <jh@suse.cz>
8769 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8771 * ipa.c (function_and_variable_visibility): Likewise.
8772 * ipa-profile.c (ipa_profile): Likewise.
8774 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8777 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8778 attribute, and don't inline the resolver in this case.
8780 2013-09-17 Teresa Johnson <tejohnson@google.com>
8782 * coverage.c (get_coverage_counts): Add missing newline.
8784 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8786 PR tree-optimization/58088
8787 * fold-const.c (mask_with_trailing_zeros): New function.
8788 (fold_binary_loc): Make sure we don't recurse infinitely
8789 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8790 Use mask_with_trailing_zeros where appropriate.
8792 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8794 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8795 of 'prev' var to get better distance estimation.
8797 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8799 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8800 * tree-inline.c (is_parameter_of): New predicate.
8801 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8802 a parameter has been remapped.
8803 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8804 (optimize_inline_calls): Initialize transform_parameter.
8805 (copy_gimple_seq_and_replace_locals): Likewise.
8806 (tree_function_versioning): Likewise.
8807 (maybe_inline_call_in_expr): Likewise.
8809 2013-09-17 Nick Clifton <nickc@redhat.com>
8811 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8812 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8814 (is_attr_func): New function.
8815 (msp430_is_interrupt_func): New function.
8816 (is_naked_func): New function.
8817 (is_reentrant_func): New function.
8818 (is_critical_func): New function.
8819 (msp430_start_function): Add annotations for function attributes.
8820 (msp430_attr): New function.
8821 (msp430_attribute_table): New.
8822 (msp430_function_section): New function.
8823 (TARGET_ASM_FUNCTION_SECTION): Define.
8824 (msp430_builtin): New enum.
8825 (msp430_init_builtins): New function.
8826 (msp430_builtin_devl): New function.
8827 (msp430_expand_builtin): New function.
8828 (TARGET_INIT_BUILTINS): Define.
8829 (TARGET_EXPAND_BUILTINS): Define.
8830 (TARGET_BUILTIN_DECL): Define.
8831 (msp430_expand_prologue): Add support for naked, interrupt,
8832 critical and reentrant functions.
8833 (msp430_expand_epilogue): Likewise.
8834 (msp430_print_operand): Handle 'O' character.
8835 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8837 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8838 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8839 (pushm): Use a 'n' rather than an 'i' constraint.
8840 (msp_return): Add generation of the interrupt return instruction.
8841 (disable_interrupts): New pattern.
8842 (enable_interrupts): New pattern.
8843 (push_intr_state): New pattern.
8844 (pop_intr_state): New pattern.
8845 (bic_SR): New pattern.
8846 (bis_SR): New pattern.
8847 * doc/extend.texi: Document MSP430 function attributes and builtin
8850 2013-09-17 Richard Biener <rguenther@suse.de>
8852 PR tree-optimization/58432
8853 * tree-loop-distribution.c (tree_loop_distribution): Also
8854 scan PHIs for outside loop uses and seed a partition from them.
8856 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8858 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8859 (restructure_reference): Call backtrace_base_for_ref.
8861 2013-09-17 Alan Modra <amodra@gmail.com>
8864 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8867 2013-09-16 DJ Delorie <dj@redhat.com>
8869 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8870 vregs location for RL78/G10.
8871 (rl78_expand_prologue): Avoid SEL on G10.
8872 (rl78_expand_epilogue): Likewise.
8873 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8874 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8875 __RL78_G10__ when appropriate.
8876 (ASM_SPEC): Pass -mg10 along to the assembler.
8877 * config/rl78/rl78.md (sel_rb): Disable for G10.
8878 * config/rl78/rl78.opt: Add -mg10 option.
8879 * config/rl78/t-rl78: Add -mg10 multilib.
8881 2013-09-16 Xinliang David Li <davidxl@google.com>
8883 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8884 * omp-low.c (omp_max_vf): Ditto.
8885 (expand_omp_simd): Ditto.
8886 * tree-vectorizer.c (vectorize_loops): Ditto.
8887 (gate_vect_slp): Ditto.
8888 (gate_increase_alignment): Ditto.
8889 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8890 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8891 (gate_tree_vectorize): Name change.
8892 (tree_vectorize): Ditto.
8893 (pass_vectorize::gate): Call new function.
8894 (pass_vectorize::execute): Ditto.
8895 * opts.c: O3 default setting change.
8896 (finish_options): Check new flag.
8897 * doc/invoke.texi: Document new flags.
8898 * common.opt: New flags.
8900 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8902 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8903 (Condition Code): Don't reference it.
8904 * doc/tm.texi: Regenerate.
8906 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8909 * lra-constraints.c (undo_optional_reloads): Consider all optional
8910 reload even if it did not get a hard reg.
8912 2013-09-16 Teresa Johnson <tejohnson@google.com>
8914 * dumpfile.c (dump_loc): Remove newline emission.
8915 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8916 emission to dump_printf_loc calls where missing.
8917 (vect_mark_for_runtime_alias_test): Ditto.
8918 (vect_analyze_data_ref_dependence): Ditto.
8919 (vect_analyze_data_ref_dependences): Ditto.
8920 (vect_slp_analyze_data_ref_dependence): Ditto.
8921 (vect_slp_analyze_data_ref_dependences): Ditto.
8922 (vect_compute_data_ref_alignment): Ditto.
8923 (vect_update_misalignment_for_peel): Ditto.
8924 (vect_verify_datarefs_alignment): Ditto.
8925 (vector_alignment_reachable_p): Ditto.
8926 (vect_get_data_access_cost): Ditto.
8927 (vect_enhance_data_refs_alignment): Ditto.
8928 (vect_find_same_alignment_drs): Ditto.
8929 (vect_analyze_data_refs_alignment): Ditto.
8930 (vect_analyze_group_access): Ditto.
8931 (vect_analyze_data_ref_access): Ditto.
8932 (vect_analyze_data_ref_accesses): Ditto.
8933 (vect_prune_runtime_alias_test_list): Ditto.
8934 (vect_analyze_data_refs): Ditto.
8935 (vect_create_addr_base_for_vector_ref): Ditto.
8936 (vect_create_data_ref_ptr): Ditto.
8937 (vect_grouped_store_supported): Ditto.
8938 (vect_grouped_load_supported): Ditto.
8939 * value-prof.c (check_counter): Ditto.
8940 (check_ic_target): Ditto.
8941 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8942 (vect_recog_widen_mult_pattern): Ditto.
8943 (vect_recog_widen_sum_pattern): Ditto.
8944 (vect_recog_over_widening_pattern): Ditto.
8945 (vect_recog_widen_shift_pattern): Ditto.
8946 (vect_recog_rotate_pattern): Ditto.
8947 (vect_recog_vector_vector_shift_pattern): Ditto.
8948 (vect_recog_divmod_pattern): Ditto.
8949 (vect_recog_mixed_size_cond_pattern): Ditto.
8950 (vect_recog_bool_pattern): Ditto.
8951 (vect_pattern_recog_1): Ditto.
8952 (vect_pattern_recog): Ditto.
8953 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8954 (vect_is_simple_iv_evolution): Ditto.
8955 (vect_analyze_scalar_cycles_1): Ditto.
8956 (vect_get_loop_niters): Ditto.
8957 (vect_analyze_loop_1): Ditto.
8958 (vect_analyze_loop_form): Ditto.
8959 (vect_analyze_loop_operations): Ditto.
8960 (vect_analyze_loop_2): Ditto.
8961 (vect_analyze_loop): Ditto.
8962 (report_vect_op): Ditto.
8963 (vect_is_slp_reduction): Ditto.
8964 (vect_is_simple_reduction_1): Ditto.
8965 (vect_get_known_peeling_cost): Ditto.
8966 (vect_estimate_min_profitable_iters): Ditto.
8967 (vect_model_reduction_cost): Ditto.
8968 (vect_model_induction_cost): Ditto.
8969 (get_initial_def_for_induction): Ditto.
8970 (vect_create_epilog_for_reduction): Ditto.
8971 (vectorizable_reduction): Ditto.
8972 (vectorizable_induction): Ditto.
8973 (vectorizable_live_operation): Ditto.
8974 (vect_loop_kill_debug_uses): Ditto.
8975 (vect_transform_loop): Ditto.
8976 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8977 (vect_stmt_relevant_p): Ditto.
8978 (process_use): Ditto.
8979 (vect_mark_stmts_to_be_vectorized): Ditto.
8980 (vect_model_simple_cost): Ditto.
8981 (vect_model_promotion_demotion_cost): Ditto.
8982 (vect_model_store_cost): Ditto.
8983 (vect_get_store_cost): Ditto.
8984 (vect_model_load_cost): Ditto.
8985 (vect_get_load_cost): Ditto.
8986 (vect_init_vector_1): Ditto.
8987 (vect_get_vec_def_for_operand): Ditto.
8988 (vect_finish_stmt_generation): Ditto.
8989 (vectorizable_call): Ditto.
8990 (vectorizable_conversion): Ditto.
8991 (vectorizable_assignment): Ditto.
8992 (vectorizable_shift): Ditto.
8993 (vectorizable_operation): Ditto.
8994 (vectorizable_store): Ditto.
8995 (vectorizable_load): Ditto.
8996 (vectorizable_condition): Ditto.
8997 (vect_analyze_stmt): Ditto.
8998 (vect_transform_stmt): Ditto.
8999 (vect_is_simple_use): Ditto.
9000 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
9001 (vect_can_advance_ivs_p): Ditto.
9002 (vect_update_ivs_after_vectorizer): Ditto.
9003 (vect_do_peeling_for_loop_bound): Ditto.
9004 (vect_gen_niters_for_prolog_loop): Ditto.
9005 (vect_update_inits_of_drs): Ditto.
9006 (vect_create_cond_for_alias_checks): Ditto.
9007 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
9008 (vect_build_slp_tree_1): Ditto.
9009 (vect_supported_load_permutation_p): Ditto.
9010 (vect_analyze_slp_instance): Ditto.
9011 (vect_analyze_slp): Ditto.
9012 (vect_make_slp_decision): Ditto.
9013 (vect_detect_hybrid_slp): Ditto.
9014 (vect_bb_vectorization_profitable_p): Ditto.
9015 (vect_slp_analyze_bb_1): Ditto.
9016 (vect_update_slp_costs_according_to_vf): Ditto.
9017 (vect_get_mask_element): Ditto.
9018 (vect_transform_slp_perm_load): Ditto.
9019 (vect_schedule_slp_instance): Ditto.
9020 (vect_schedule_slp): Ditto.
9021 (vect_slp_transform_bb): Ditto.
9022 * profile.c (read_profile_edge_counts): Ditto.
9023 (compute_branch_probabilities): Ditto.
9024 * coverage.c (get_coverage_counts): Ditto.
9026 2013-09-16 Diego Novillo <dnovillo@google.com>
9028 * tree-core.h: Add missing comment lines from refactoring of tree.h.
9030 2013-09-16 Jan Hubicka <jh@suse.cz>
9032 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
9033 abstract functions; for static functions check the presence of body.
9035 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
9037 * config/aarch64/aarch64-simd-builtins.def (fma): New.
9038 * config/aarch64/aarch64-simd.md
9039 (aarch64_mla_elt<mode>): New.
9040 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
9041 (aarch64_mls_elt<mode>): Likewise.
9042 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
9043 (aarch64_fma4_elt<mode>): Likewise.
9044 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
9045 (aarch64_fma4_elt_to_128v2df): Likewise.
9046 (aarch64_fma4_elt_to_64df): Likewise.
9047 (fnma<mode>4): Likewise.
9048 (aarch64_fnma4_elt<mode>): Likewise.
9049 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
9050 (aarch64_fnma4_elt_to_128v2df): Likewise.
9051 (aarch64_fnma4_elt_to_64df): Likewise.
9052 * config/aarch64/iterators.md (VDQSF): New.
9053 * config/aarch64/arm_neon.h
9054 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
9055 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
9057 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
9059 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
9060 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
9061 (aarch64_mul3_elt_to_128df): Likewise.
9062 (aarch64_mul3_elt_to_64v2df): Likewise.
9063 * config/aarch64/iterators.md (VEL): Also handle DFmode.
9065 (VMUL_CHANGE_NLANES) Likewise.
9068 * config/aarch64/arm_neon.h
9069 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
9071 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
9073 * config/aarch64/arm_neon.h
9074 (vcvtx_high_f32_f64): Fix parameters.
9076 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9077 Uros Bizjak <ubizjak@gmail.com>
9079 * config/alpha.c: Include tree-ssanames.h.
9081 2013-09-16 Richard Biener <rguenther@suse.de>
9083 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
9084 (dot_rdg_1): Handle control_dd.
9085 (create_edge_for_control_dependence): New function.
9086 (create_rdg_edges): Add control dependences if asked for.
9087 (build_rdg): Likewise.
9088 (generate_loops_for_partition): If there are not necessary
9089 control stmts remove all their dependencies.
9090 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
9091 (distribute_loop): Pass on control dependences.
9092 (tree_loop_distribution): Compute control dependences and remove
9093 restriction on number of loop nodes.
9095 2013-09-16 Jakub Jelinek <jakub@redhat.com>
9097 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
9100 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
9102 * cse.c (try_const_anchors): Punt on CC modes.
9104 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9106 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
9108 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
9111 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
9114 2013-09-14 Iain Sandoe <iain@codesourcery.com>
9117 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
9119 (darwin_objc1_section): Likewise.
9120 (darwin_file_end): Emit Image Info section when required.
9122 2013-09-14 Jan Hubicka <jh@suse.cz>
9124 * tree-into-ssa.c (gate_into_ssa): New.
9125 (pass_data_build_ssa): Use it.
9126 * cgraph.h (expand_thunk): Update prototype.
9127 * cgraphunit.c (analyze_function): Expand thunks early.
9128 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
9129 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
9130 set lowered flag; do not add new function.
9131 (assemble_thunks_and_aliases): Update.
9132 * tree-ssa.c (gate_init_datastructures): New gate.
9133 (pass_data_init_datastructures): Use it.
9135 2013-09-14 Iain Sandoe <iain@codesourcery.com>
9138 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
9139 xmm register set described in the psABI.
9141 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
9143 * dwarf2out.c (should_emit_struct_debug): Add check
9144 for type_decl variable is not NULL.
9146 2013-09-13 Jacek Caban <jacek@codeweavers.com>
9148 * config.gcc: Use new winnt-c.c target hooks
9149 * config/t-winnt: New file
9150 * config/winnt-c.c: New file
9151 * doc/tm.texi.in: Document new hook
9152 * doc/tm.texi: Regenerated
9154 2013-09-13 Jan Hubicka <jh@suse.cz>
9157 * ipa-inline.c (check_callers): New function.
9158 (check_caller_edge): Remove.
9159 (want_inline_function_to_all_callers_p): Also permit alises that are
9161 (inline_to_all_callers): Terminate the walk when devirtualization
9162 introduce new calls.
9164 2013-09-13 Jan Hubicka <jh@suse.cz>
9166 * ipa-inline-analysis.c (struct growth_data): Add node.
9167 (do_estimate_growth_1): Fix detection of recursion.
9169 2013-09-13 Jakub Jelinek <jakub@redhat.com>
9171 PR tree-optimization/58392
9172 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
9173 to avoid shadowing of outer loop variable. If
9174 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
9175 replace_by_duplicate_decl simduid of loops that have it set and
9176 set dest_cfun->has_simduid_loops and/or
9177 dest_cfun->has_force_vect_loops.
9178 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
9179 instead of maybe_lookup_decl.
9180 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
9181 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
9182 call. Copy over force_vect and copy and remap simduid. Set
9183 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
9184 (copy_cfg_body): Remove blocks_to_copy argument. Use
9185 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
9186 caller. Don't set cfun->has_simduid_loops and/or
9187 cfun->has_force_vect_loops here.
9188 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
9190 (expand_call_inline, tree_function_versioning): Adjust copy_body
9193 2013-09-13 Martin Jambor <mjambor@suse.cz>
9196 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
9197 the assert if the edge was a speculative one.
9199 2013-09-13 Richard Biener <rguenther@suse.de>
9201 * tree-data-ref.h (known_dependences_p): Move here ...
9202 * tree-loop-distribution.c (known_dependences_p): ... from here.
9203 (dump_rdg_component, debug_rdg_component): Remove.
9205 (generate_loops_for_partition): Use gimple_uid instead of
9206 relying on matching stmt visit order.
9207 (rdg_build_partitions): Take starting stmt vector.
9208 (ldist_gen): Merge into ...
9209 (distribute_loop): ... this function. Do not compute starting
9211 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
9213 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9215 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
9216 Set type attribute correctly. Set predicable_short_it attribute.
9217 (cmpsi_shiftsi): Remove %? from output template.
9219 2013-09-13 Richard Biener <rguenther@suse.de>
9221 * tree-loop-distribution.c (struct rdg_component,
9222 rdg_defs_used_in_other_loops_p, free_rdg_components,
9223 rdg_build_components): Remove.
9224 (stmts_from_loop): Do not record virtual PHIs.
9225 (generate_loops_for_partition): Skip virtual PHIs.
9226 (build_rdg_partition_for_component): Rename to ...
9227 (build_rdg_partition_for_vertex): ... this and adjust.
9228 (rdg_build_partitions): Take a vector of starting vertices
9229 instead of components. Remove unnecessary leftover handling.
9230 (ldist_gen): Do not build components or record other stores.
9231 (distribute_loop): Do not distribute loops containing stmts
9234 2013-09-13 Christian Bruel <christian.bruel@st.com>
9237 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
9239 2013-09-13 Kai Tietz <ktietz@redhat.com>
9241 * config.gcc: Separate cases for mingw and cygwin targets,
9242 and add 64-bit cygwin target case.
9244 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
9245 dll-export inline-functions.
9246 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
9248 2013-09-13 Jeff Law <law@redhat.com>
9252 2013-09-06 Jeff Law <law@redhat.com>
9254 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9255 edge implied equivalences into successor phis.
9257 2013-09-12 DJ Delorie <dj@redhat.com>
9259 * config/rl78/rl78-virt.md: Change from | to \; for asm line
9262 2013-09-12 Brooks Moses <bmoses@google.com>
9265 * Makefile.in: Do not install driver binaries in $(target)/bin.
9267 2013-09-12 DJ Delorie <dj@redhat.com>
9269 * config/rl78/rl78.opt (mrelax): New.
9270 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
9271 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
9273 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
9275 (rl78_expand_epilogue): Likewise.
9277 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
9280 * lra-eliminations.c (remove_reg_equal_offset_note): New.
9281 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
9282 pointer elimination with using remove_reg_equal_offset_note.
9284 2013-09-12 DJ Delorie <dj@redhat.com>
9286 * config/msp430/: New port.
9287 * config.gcc (msp430): Added.
9288 * doc/invoke.texi: Document MSP430 options.
9289 * doc/install.texi: Document msp430-elf
9290 * doc/md.texi: Document msp430-elf
9291 * doc/contrib.texi: Document msp430-elf
9293 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
9296 2013-09-12 Martin Jambor <mjambor@suse.cz>
9299 * ipa-prop.c (remove_described_reference): Give up if the edge in the
9300 reference descriptor is NULL.
9301 (ipa_edge_removal_hook): If owning a reference descriptor, set its
9304 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9306 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
9307 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
9308 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
9309 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
9310 make_temp_ssa_name): move to tree-ssanames.h
9311 * tree-ssa-alias.h: Move prototype.
9312 * tree-ssa.h: Include tree-ssanames.h.
9313 * tree-ssanames.c (FREE_SSANAMES): Move to here.
9314 * tree-ssanames.h: New. Move items from tree-flow*.h
9315 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
9317 2013-09-12 Richard Biener <rguenther@suse.de>
9319 PR tree-optimization/58404
9320 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
9321 propagate non-invariant addresses into dereferences wrapped
9322 in component references.
9324 2013-09-12 Richard Biener <rguenther@suse.de>
9326 PR tree-optimization/58402
9327 * passes.def: Move pass_late_warn_uninitialized later.
9329 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9331 * tree-ssa.h: New. Move content from tree-flow.h and
9333 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
9334 Move prototypes belonging to tree-ssa.c.
9335 * tree-flow-inline.h (redirect_edge_var_map_def,
9336 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
9338 * gimple.h: Adjust prototypes.
9339 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
9341 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
9342 * tree.h: Move prototype to tree-ssa.h.
9343 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
9344 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
9345 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
9346 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
9347 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
9348 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
9349 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
9350 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
9351 graphite-blocking.c, graphite-clast-to-gimple.c,
9352 graphite-dependences.c, graphite-interchange.c,
9353 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
9354 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
9355 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
9356 ipa-reference.c, ipa-split.c, ipa-utils.c,
9357 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
9358 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
9359 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
9360 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
9361 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
9362 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
9363 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
9364 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
9365 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
9366 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
9367 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
9368 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
9369 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
9370 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
9371 value-prof.c, var-tracking.c,
9372 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
9374 2013-09-12 Richard Biener <rguenther@suse.de>
9376 PR tree-optimization/58396
9377 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
9378 (build_rdg): Take a loop-nest parameter, fix memleaks.
9379 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
9381 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9383 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
9386 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
9388 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
9389 The FMA instruction names should have a 'v' prefix.
9391 2013-09-12 Richard Biener <rguenther@suse.de>
9393 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
9394 (dot_rdg): Use popen instead of system in optional code.
9395 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
9396 (already_processed_vertex_p): Adjust.
9397 (has_anti_or_output_dependence, predecessor_has_mem_write,
9398 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
9399 rdg_flag_uses): Remove.
9400 (rdg_flag_vertex): Simplify.
9401 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
9403 (build_rdg_partition_for_component): Process the first vertex
9404 of a component only.
9405 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
9407 2013-09-12 Alan Modra <amodra@gmail.com>
9409 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
9411 2013-09-11 DJ Delorie <dj@redhat.com>
9412 Nick Clifton <nickc@redhat.com>
9414 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
9415 (rl78_stack_based_mem): New.
9416 * config/rl78/constraints.md (Iv08): New.
9425 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
9427 (movsi): Change from expand to insn-and-split.
9428 (ashrsi3): Clobber AX.
9432 * config/rl78/rl78.md (CC_REG): Fix.
9433 (addsi3): Allow memory and immediate operands.
9434 (addsi3_internal): Split into...
9435 (addsi3_internal_virt): ...new, and ...
9436 (addsi3_internal_real): ...new.
9438 (subsi3_internal_virt): New.
9439 (subsi3_internal_real): New.
9440 (mulsi3): Add memory operand.
9441 (mulsi3_rl78): Likewise.
9442 (mulsi3_g13): Likewise.
9443 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
9444 (cbranchqi4_real): Add more constraint options.
9445 (cbranchhi4_real): Expand pattern.
9446 (cbranchhi4_real_signed): New.
9447 (cbranchhi4_real_inverted): New.
9448 (cbranchsi4_real_lt): New.
9449 (cbranchsi4_real_ge): New.
9450 (cbranchsi4_real_signed): New.
9451 (cbranchsi4_real): New.
9453 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
9455 (lshrsi3_virt): Likewise.
9456 (ashlsi3_virt): Likewise.
9457 (cbranchqi4_virt_signed): New.
9458 (cbranchhi4_virt_signed): New.
9459 (cbranchsi4_virt): New.
9460 * config/rl78/rl78.c: Whitespace fixes throughout.
9461 (move_elim_pass): New.
9462 (pass_data_rl78_move_elim): New.
9463 (pass_rl78_move_elim): New.
9464 (make_pass_rl78_move_elim): New.
9465 (rl78_devirt_info): Run devirt earlier.
9466 (rl78_move_elim_info): New.
9467 (rl78_asm_file_start): Register it.
9468 (rl78_split_movsi): New.
9469 (rl78_as_legitimate_address): Allow virtual base registers when
9471 (rl78_addr_space_convert): Remove spurious debug stuff.
9472 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
9473 (rl78_print_operand): More cases for not printing '#'.
9474 (rl78_expand_compare): Remove most of the logic.
9475 (content_memory): New.
9476 (clear_content_memory): New.
9477 (get_content_index): New.
9478 (get_content_name): New.
9479 (display_content_memory): New.
9480 (update_content): New.
9481 (record_content): New.
9482 (already_contains): New.
9483 (insn_ok_now): Re-recog insns with virtual registers.
9484 (add_postponed_content_update): New.
9485 (process_postponed_content_update): New.
9486 (gen_and_emit_move): New.
9487 (transcode_memory_rtx): Record new location content.
9488 Use gen_and_emit_move.
9489 (force_into_acc): New.
9490 (move_to_acc): Use gen_and_emit_move.
9491 (move_from_acc): Likewise.
9492 (move_acc_to_reg): Likewise.
9493 (move_to_x): Likewise.
9494 (move_to_hl): Likewise.
9495 (move_to_de): Likewise.
9496 (rl78_alloc_physical_registers_op1): Record location content.
9497 (has_constraint): New.
9498 (rl78_alloc_physical_registers_op2): Record location content.
9500 (rl78_alloc_physical_registers_ro1): Likewise.
9501 (rl78_alloc_physical_registers_cmp): Likewise.
9502 (rl78_alloc_physical_registers_umul): Likewise.
9503 (rl78_alloc_address_registers_macax): New.
9504 (rl78_alloc_physical_registers): Initialize and set location
9505 content memory as needed.
9506 (rl78_reorg): Make sure split2 is called.
9507 (rl78_rtx_costs): New.
9509 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
9511 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
9512 for (not (neg ...)) and (neg (not ...)) cases.
9514 2013-09-11 Richard Biener <rguenther@suse.de>
9517 * passes.def: Split critical edges before late uninit warning passes.
9518 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
9520 2013-09-11 Jakub Jelinek <jakub@redhat.com>
9522 PR tree-optimization/58385
9523 * fold-const.c (build_range_check): If both low and high are NULL,
9524 use omit_one_operand_loc to preserve exp side-effects.
9526 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9528 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
9530 2013-09-11 Richard Biener <rguenther@suse.de>
9532 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
9533 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
9534 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
9535 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
9536 stmts_from_loop, known_dependences_p, build_empty_rdg,
9537 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
9538 * tree-loop-distribution.c: ... here.
9539 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
9540 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
9541 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
9542 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
9543 * tree-loop-distribution.c: ... here.
9544 * tree-loop-distribution.c: Include gimple-pretty-print.h.
9545 (struct partition_s): Add loops member.
9546 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
9547 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
9548 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
9550 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9551 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9552 Sergey Lega <sergey.s.lega@intel.com>
9553 Anna Tikhonova <anna.tikhonova@intel.com>
9554 Ilya Tocar <ilya.tocar@intel.com>
9555 Andrey Turetskiy <andrey.turetskiy@intel.com>
9556 Ilya Verbin <ilya.verbin@intel.com>
9557 Kirill Yukhin <kirill.yukhin@intel.com>
9558 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9560 * config/i386/constraints.md (k): New.
9562 * config/i386/i386.c (const regclass_map): Add new mask registers.
9563 (dbx_register_map): Ditto.
9564 (dbx64_register_map): Ditto.
9565 (svr4_dbx_register_map): Ditto.
9566 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
9568 (ix86_preferred_reload_class): Disable constants for mask registers.
9569 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
9570 (ix86_hard_regno_mode_ok): Support new mask registers.
9571 (x86_order_regs_for_local_alloc): Ditto.
9572 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
9573 (FIXED_REGISTERS): Add new mask registers.
9574 (CALL_USED_REGISTERS): Ditto.
9575 (REG_ALLOC_ORDER): Ditto.
9576 (VALID_MASK_REG_MODE): New.
9577 (FIRST_MASK_REG): Ditto.
9578 (LAST_MASK_REG): Ditto.
9579 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
9580 (MAYBE_MASK_CLASS_P): New.
9581 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
9582 (REG_CLASS_CONTENTS): Ditto.
9583 (MASK_REGNO_P): New.
9584 (ANY_MASK_REG_P): Ditto.
9585 (HI_REGISTER_NAMES): Add new mask registers.
9586 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
9587 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
9589 (attribute "type"): Add mskmov, msklog.
9590 (attribute "length_immediate"): Support them.
9591 (attribute "memory"): Ditto.
9592 (attribute "prefix_0f"): Ditto.
9593 (*movhi_internal): Support new mask registers.
9594 (*movqi_internal): Ditto.
9595 (define_split): Split out clobber pattern is a logic
9596 insn on mask registers.
9597 (*k<logic><mode>): New.
9598 (*andhi_1): Extend to support mask regs.
9599 (*andqi_1): Extend to support mask regs.
9601 (define_split): Split and-not to and and not if operands
9603 (*<code><mode>_1): Separate HI mode to new pattern...
9604 (*<code>hi_1): This.
9605 (*<code>qi_1): Extend to support mask regs.
9607 (kortestzhi): Ditto.
9608 (kortestchi): Ditto.
9610 (*one_cmpl<mode>2_1): Remove HImode and handle it...
9611 (*one_cmplhi2_1): ...Here, now with mask registers support.
9612 (*one_cmplqi2_1): Support new mask registers.
9613 (HI/QImode arithmetics splitter): Don't split if mask registers
9615 (HI/QImode not splitter): Ditto.
9616 * config/i386/predicated.md (mask_reg_operand): New.
9617 (general_reg_operand): Ditto.
9619 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9621 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
9622 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
9624 2013-09-10 Jeff Law <law@redhat.com>
9626 PR tree-optimization/58380
9627 * tree-ssa-threadupdate.c (thread_block): Recognize another case
9628 of threading through a buried loop header.
9630 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
9631 return value for single successor case.
9633 2013-09-10 Jan Hubicka <jh@suse.cz>
9635 * ipa-devirt.c (ipa_devirt): Enable with LTO.
9637 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
9640 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
9641 support conditional execution.
9642 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
9644 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
9646 * lra.c (lra): Clear lra_optional_reload_pseudos before every
9648 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
9649 Check destination too to check move insn.
9650 (undo_optional_reloads): Add check that the original peudo did not
9651 changed its allocation and the optional reload was inherited on last
9652 inheritance pass. Break loop after deciding to keep optional reload.
9653 (lra_undo_inheritance): Add check that inherited pseudo still in
9656 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
9658 * config/aarch64/aarch64.md (generic_sched): New.
9659 * config/aarch64/aarch64-generic.md (load): Make conditional
9660 on generic_sched attribute.
9661 (nonload): Likewise.
9663 2013-09-10 Jan Hubicka <jh@suse.cz>
9665 * lto-cgraph.c: Include ipa-utils.h.
9666 (compute_ltrans_boundary): Also add possible targets into the boundary.
9668 2013-09-10 Jan Hubicka <jh@suse.cz>
9670 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
9671 VAR_DECL of vtable rather than full expression.
9673 2013-09-10 Jan Hubicka <jh@suse.cz>
9674 Paolo Carlini <paolo.carlini@oracle.com>
9676 * cgraphunit.c (analyze_functions): Save input_location, set it
9677 to UNKNOWN_LOCATION and restore it at the end.
9679 2013-09-10 Martin Jambor <mjambor@suse.cz>
9681 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
9682 represented by a thunk.
9684 2013-09-10 Jeff Law <law@redhat.com>
9686 PR tree-optimization/58343
9687 * tree-ssa-threadupdate.c (thread_block): Identify and disable
9688 jump threading requests through loop headers buried in the middle
9689 of a jump threading path.
9691 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
9692 in return value/type.
9694 2013-09-10 Jakub Jelinek <jakub@redhat.com>
9696 PR rtl-optimization/58365
9697 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
9698 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
9701 2013-09-10 Richard Biener <rguenther@suse.de>
9703 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
9704 * tree-data-ref.c (create_rdg_vertices): Collect all data
9705 references, signal failure to the caller, use data-ref API.
9706 (build_rdg): Compute data references only once. Maintain lifetime
9707 of data references and data dependences from within RDG.
9708 (free_rdg): Free dependence relations.
9709 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
9710 inventing extra dependences.
9711 (distribute_loop): Update for RDG API changes.
9713 2013-09-10 Kai Tietz <ktietz@redhat.com>
9715 * doc/invoke.texi (fms-extensions): Document changed
9716 behavior for ms-abi targets.
9717 * config/i386/i386.c (ix86_option_override_internal):
9718 Set default value of option -fms-extension for ms-abi targets.
9720 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9722 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
9724 2013-09-10 Alan Modra <amodra@gmail.com>
9727 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
9729 2013-09-10 Alan Modra <amodra@gmail.com>
9731 * config/rs6000/predicates.md (add_cint_operand): New.
9732 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9733 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9734 using add_cint_operand.
9735 (largetoc_high_plus_aix): Likewise.
9737 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9739 PR tree-optimization/58364
9740 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9741 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9742 the current range can't be an unconditional true or false.
9744 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9746 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9748 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9750 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9752 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9755 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9757 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9759 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9760 NO_REGS for immediate that can't be moved directly into FP_REGS.
9762 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9764 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9765 comparison with negated operand.
9766 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9769 2013-09-09 Richard Biener <rguenther@suse.de>
9772 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9773 of a subloop record all its block as affecting loop-closed SSA form.
9775 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9777 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9778 of an rtx/bitpos pair.
9779 (store_fixed_bit_field): Update accordingly.
9781 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9783 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9785 * builtins.c (expand_errno_check): Likewise.
9786 * dwarf2cfi.c (init_return_column_size): Likewise.
9787 * except.c (sjlj_mark_call_sites): Likewise.
9788 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9789 * lra-constraints.c (emit_inc): Likewise.
9790 * ree.c (combine_set_extension): Likewise.
9791 * regmove.c (fixup_match_2): Likewise.
9792 * reload1.c (inc_for_reload): Likewise.
9794 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9796 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9797 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9798 Use gen_int_mode with the mode of the associated simplify_* call.
9799 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9800 * expmed.c (expand_shift_1): Likewise.
9801 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9802 * loop-iv.c (iv_number_of_iterations): Likewise.
9803 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9804 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9806 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9808 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9809 of the associated expand_* call.
9810 (asan_emit_stack_protection): Likewise.
9811 * builtins.c (round_trampoline_addr): Likewise.
9812 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9813 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9814 (emit_store_flag): Likewise.
9815 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9816 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9818 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9819 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9820 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9821 * modulo-sched.c (generate_prolog_epilog): Likewise.
9822 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9823 (expand_ctz, expand_ffs, expand_unop): Likewise.
9825 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9827 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9828 of the associated gen_rtx_* call.
9829 * caller-save.c (init_caller_save): Likewise.
9830 * combine.c (find_split_point, make_extraction): Likewise.
9831 (make_compound_operation): Likewise.
9832 * dwarf2out.c (mem_loc_descriptor): Likewise.
9833 * explow.c (plus_constant, probe_stack_range): Likewise.
9834 * expmed.c (expand_mult_const): Likewise.
9835 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9836 * reload1.c (init_reload): Likewise.
9837 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9838 * var-tracking.c (adjust_mems): Likewise.
9839 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9840 rather than gen_rtx_fmt_ee.
9842 2013-09-09 Jan Hubicka <jh@suse.cz>
9845 * value-prof.c (gimple_ic): Copy also abnormal edges.
9847 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9849 * asan.c (asan_shadow_cst): Use gen_int_mode.
9851 2013-09-08 Jan Hubicka <jh@suse.cz>
9853 * ipa-profile.c: Add toplevel comment.
9854 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9855 (contains_hot_call_p): New function.
9856 (ipa_propagate_frequency): Set frequencies based on counts when
9858 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9859 profile; do not tamper with profile after inlining if it is read.
9861 2013-09-08 Jan Hubicka <jh@suse.cz>
9863 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9866 2013-09-08 Jan Hubicka <jh@suse.cz>
9868 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9871 2013-09-08 Jeff Law <law@redhat.com>
9874 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9877 2013-09-08 Andi Kleen <ak@linux.intel.com>
9879 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9881 2013-09-08 Jan Hubicka <jh@suse.cz>
9883 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9885 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9887 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9888 for non-debug insns.
9889 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9890 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9891 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9894 2013-09-08 Jan Hubicka <jh@suse.cz>
9896 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9897 targets and devirtualize to BUILT_IN_UNREACHABLE.
9898 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9899 * ipa.c (walk_polymorphic_call_targets): New function.
9900 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9901 functions; use the new timevar.
9902 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9903 was removed from the program.
9904 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9905 not consider it in the walk when its vtable is dead.
9906 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9908 (devirt_variable_node_removal_hook): New function.
9909 (possible_polymorphic_call_targets): Also register
9910 devirt_variable_node_removal_hook.
9911 (ipa_devirt): Do not do non-speculative devirtualization.
9912 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9914 2013-09-08 Jan Hubicka <jh@suse.cz>
9916 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9917 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9918 varpool_remove_variable_insertion_hook): Declare.
9919 * varpool.c (varpool_node_hook_list): New structure.
9920 (first_varpool_node_removal_hook,
9921 first_varpool_variable_insertion_hook): New variables.
9922 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9923 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9924 varpool_remove_variable_insertion_hook,
9925 varpool_call_variable_insertion_hooks): New functions.
9926 (varpool_remove_node): Use it.
9928 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9931 * diagnostic.c (diagnostic_build_prefix): When s.file is
9932 "<built-in>" don't output line and column numbers.
9934 2013-09-06 Jan Hubicka <jh@suse.cz>
9936 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9937 * lto-streamer-out.c: Stream thunks, too.
9938 * lto-streamer-in.c (input_function): Pop cfun here
9939 (lto_read_body): Instead of here.
9941 2013-09-06 Caroline Tice <cmtice@google.com>
9943 * doc/install.texi: Add documentation for the --enable-vtable-verify
9944 and the --disable-libvtv configure options.
9946 2013-09-06 Jeff Law <law@redhat.com>
9948 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9949 edge implied equivalences into successor phis.
9951 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9953 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9955 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9957 * resource.c (mark_target_live_regs): Compute resources taking
9958 into account if a call is predicated or not.
9960 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9962 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9963 the compiler in the function names.
9965 2013-09-06 Jan Hubicka <jh@suse.cz>
9968 * ipa-inline.c (has_caller_p): New function.
9969 (want_inline_function_to_all_callers_p): Use it.
9970 (sum_callers, inline_to_all_callers): Break out from ...
9971 (ipa_inline): ... here.
9973 2013-09-06 Jan Hubicka <jh@suse.cz>
9975 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9976 only when AVX is enabled.
9978 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9980 * config/aarch64/aarch64.md
9981 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9982 is fpsimd_<load/store>2.
9983 (load_pair<mode>): Likewise.
9984 (store_pair<mode>): Likewise.
9986 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9988 * config/arm/types.md (type): Add "mrs" type.
9989 * config/aarch64/aarch64.md
9990 (aarch64_load_tp_hard): Make type "mrs".
9992 (load_tp_hard): Make type "mrs".
9993 * config/arm/cortex-a15.md: Update with new attributes.
9994 * config/arm/cortex-a5.md: Update with new attributes.
9995 * config/arm/cortex-a53.md: Update with new attributes.
9996 * config/arm/cortex-a7.md: Update with new attributes.
9997 * config/arm/cortex-a8.md: Update with new attributes.
9998 * config/arm/cortex-a9.md: Update with new attributes.
9999 * config/arm/cortex-m4.md: Update with new attributes.
10000 * config/arm/cortex-r4.md: Update with new attributes.
10001 * config/arm/fa526.md: Update with new attributes.
10002 * config/arm/fa606te.md: Update with new attributes.
10003 * config/arm/fa626te.md: Update with new attributes.
10004 * config/arm/fa726te.md: Update with new attributes.
10006 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10008 * config/aarch64/aarch64.md
10009 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
10010 (*movtf_aarch64): Likewise.
10011 * config/arm/arm.md
10012 (thumb1_movdi_insn): Use "multiple" for type where more than one
10013 instruction is used for a move.
10014 (*arm32_movhf): Likewise.
10015 (*thumb_movdf_insn): Likewise.
10017 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10019 * config/arm/types.md (type): Rename fcpys to fmov.
10020 * config/arm/vfp.md
10021 (*arm_movsi_vfp): Rename type fcpys as fmov.
10022 (*thumb2_movsi_vfp): Likewise
10023 (*movhf_vfp_neon): Likewise
10024 (*movhf_vfp): Likewise
10025 (*movsf_vfp): Likewise
10026 (*thumb2_movsf_vfp): Likewise
10027 (*movsfcc_vfp): Likewise
10028 (*thumb2_movsfcc_vfp): Likewise
10029 * config/aarch64/aarch64-simd.md
10030 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
10031 * config/aarch64/aarch64.md
10032 (*movsi_aarch64): Replace type mov_reg with fmovs.
10033 (*movdi_aarch64): Likewise
10034 (*movsf_aarch64): Likewise
10035 (*movdf_aarch64): Likewise
10037 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
10038 * config/arm/iwmmxt.md
10039 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
10040 * config/arm/arm1020e.md: Update with new attributes.
10041 * config/arm/cortex-a15-neon.md: Update with new attributes.
10042 * config/arm/cortex-a5.md: Update with new attributes.
10043 * config/arm/cortex-a53.md: Update with new attributes.
10044 * config/arm/cortex-a7.md: Update with new attributes.
10045 * config/arm/cortex-a8-neon.md: Update with new attributes.
10046 * config/arm/cortex-a9.md: Update with new attributes.
10047 * config/arm/cortex-m4-fpu.md: Update with new attributes.
10048 * config/arm/cortex-r4f.md: Update with new attributes.
10049 * config/arm/marvell-pj4.md: Update with new attributes.
10050 * config/arm/vfp11.md: Update with new attributes.
10052 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10054 * config/aarch64/aarch64.md
10055 (*madd<mode>): Fix type attribute.
10056 (*maddsi_uxtw): Likewise.
10057 (*msub<mode>): Likewise.
10058 (*msubsi_uxtw): Likewise.
10059 (<su_optab>maddsidi4): Likewise.
10060 (<su_optab>msubsidi4): Likewise.
10062 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10064 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
10065 * config/arm/arm.md (core_cycles): Remove fdiv.
10066 * config/arm/vfp.md:
10067 (*sqrtsf2_vfp): Update for attribute changes.
10068 (*sqrtdf2_vfp): Likewise.
10069 * config/aarch64/aarch64.md:
10070 (sqrt<mode>2): Update for attribute changes.
10071 * config/arm/arm1020e.md: Update with new attributes.
10072 * config/arm/cortex-a15-neon.md: Update with new attributes.
10073 * config/arm/cortex-a5.md: Update with new attributes.
10074 * config/arm/cortex-a53.md: Update with new attributes.
10075 * config/arm/cortex-a7.md: Update with new attributes.
10076 * config/arm/cortex-a8-neon.md: Update with new attributes.
10077 * config/arm/cortex-a9.md: Update with new attributes.
10078 * config/arm/cortex-m4-fpu.md: Update with new attributes.
10079 * config/arm/cortex-r4f.md: Update with new attributes.
10080 * config/arm/marvell-pj4.md: Update with new attributes.
10081 * config/arm/vfp11.md: Update with new attributes.
10083 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10085 * config/arm/types.md
10086 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
10087 * config/aarch64/aarch64.md
10088 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
10090 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
10091 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
10092 (float<GPI:mode><GPF:mode>2): Likewise.
10093 * config/arm/vfp.md
10094 (*truncsisf2_vfp): Update with new attributes.
10095 (*truncsidf2_vfp): Likewise.
10096 (fixuns_truncsfsi2): Likewise.
10097 (fixuns_truncdfsi2): Likewise.
10098 (*floatsisf2_vfp): Likewise.
10099 (*floatsidf2_vfp): Likewise.
10100 (floatunssisf2): Likewise.
10101 (floatunssidf2): Likewise.
10102 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
10103 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
10104 * config/arm/arm1020e.md: Update with new attributes.
10105 * config/arm/cortex-a15-neon.md: Update with new attributes.
10106 * config/arm/cortex-a5.md: Update with new attributes.
10107 * config/arm/cortex-a53.md: Update with new attributes.
10108 * config/arm/cortex-a7.md: Update with new attributes.
10109 * config/arm/cortex-a8-neon.md: Update with new attributes.
10110 * config/arm/cortex-a9.md: Update with new attributes.
10111 * config/arm/cortex-m4-fpu.md: Update with new attributes.
10112 * config/arm/cortex-r4f.md: Update with new attributes.
10113 * config/arm/marvell-pj4.md: Update with new attributes.
10114 * config/arm/vfp11.md: Update with new attributes.
10116 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10118 * config/aarch64/arm_neon.h
10119 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
10120 (vqtbx<1,2,3,4><q>_s8): Likewise.
10122 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10124 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
10125 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
10126 (add<mode>3): Add type attribute.
10127 (add<mode>3): Likewise.
10128 (usadd<mode>3): Likewise.
10129 (ssadd<mode>3): Likewise.
10130 (sub<mode>3): Likewise.
10131 (sub<mode>3): Likewise.
10132 (ussub<mode>3): Likewise.
10133 (sssub<mode>3): Likewise.
10134 (ssmulsa3): Likewise.
10135 (usmulusa3): Likewise.
10136 (arm_usatsihi): Likewise.
10137 * config/arm/vfp.md
10138 (*movdi_vfp): Add types for all instructions.
10139 (*movdi_vfp_cortexa8): Likewise.
10140 (*movhf_vfp_neon): Likewise.
10141 (*movhf_vfp): Likewise.
10142 (*movdf_vfp): Likewise.
10143 (*thumb2_movdf_vfp): Likewise.
10144 (*thumb2_movdfcc_vfp): Likewise.
10145 * config/arm/arm.md: Add type attribute to all insn patterns.
10146 (*thumb1_adddi3): Add type attribute.
10147 (*arm_adddi3): Likewise.
10148 (*adddi_sesidi_di): Likewise.
10149 (*adddi_zesidi_di): Likewise.
10150 (*thumb1_addsi3): Likewise.
10151 (addsi3_compare0): Likewise.
10152 (*addsi3_compare0_scratch): Likewise.
10153 (*compare_negsi_si): Likewise.
10154 (cmpsi2_addneg): Likewise.
10155 (*addsi3_carryin_<optab>): Likewise.
10156 (*addsi3_carryin_alt2_<optab>): Likewise.
10157 (*addsi3_carryin_clobercc_<optab>): Likewise.
10158 (*subsi3_carryin): Likewise.
10159 (*subsi3_carryin_const): Likewise.
10160 (*subsi3_carryin_compare): Likewise.
10161 (*subsi3_carryin_compare_const): Likewise.
10162 (*arm_subdi3): Likewise.
10163 (*thumb_subdi3): Likewise.
10164 (*subdi_di_zesidi): Likewise.
10165 (*subdi_di_sesidi): Likewise.
10166 (*subdi_zesidi_di): Likewise.
10167 (*subdi_sesidi_di): Likewise.
10168 (*subdi_zesidi_ze): Likewise.
10169 (thumb1_subsi3_insn): Likewise.
10170 (*arm_subsi3_insn): Likewise.
10171 (*anddi3_insn): Likewise.
10172 (*anddi_zesidi_di): Likewise.
10173 (*anddi_sesdi_di): Likewise.
10174 (*ne_zeroextracts): Likewise.
10175 (*ne_zeroextracts): Likewise.
10176 (*ite_ne_zeroextr): Likewise.
10177 (*ite_ne_zeroextr): Likewise.
10178 (*anddi_notdi_di): Likewise.
10179 (*anddi_notzesidi): Likewise.
10180 (*anddi_notsesidi): Likewise.
10181 (andsi_notsi_si): Likewise.
10182 (thumb1_bicsi3): Likewise.
10183 (*iordi3_insn): Likewise.
10184 (*iordi_zesidi_di): Likewise.
10185 (*iordi_sesidi_di): Likewise.
10186 (*thumb1_iorsi3_insn): Likewise.
10187 (*xordi3_insn): Likewise.
10188 (*xordi_zesidi_di): Likewise.
10189 (*xordi_sesidi_di): Likewise.
10190 (*arm_xorsi3): Likewise.
10191 (*andsi_iorsi3_no): Likewise.
10192 (*smax_0): Likewise.
10193 (*smax_m1): Likewise.
10194 (*arm_smax_insn): Likewise.
10195 (*smin_0): Likewise.
10196 (*arm_smin_insn): Likewise.
10197 (*arm_umaxsi3): Likewise.
10198 (*arm_uminsi3): Likewise.
10199 (*minmax_arithsi): Likewise.
10200 (*minmax_arithsi_): Likewise.
10201 (*satsi_<SAT:code>): Likewise.
10202 (arm_ashldi3_1bit): Likewise.
10203 (arm_ashrdi3_1bit): Likewise.
10204 (arm_lshrdi3_1bit): Likewise.
10205 (*arm_negdi2): Likewise.
10206 (*thumb1_negdi2): Likewise.
10207 (*arm_negsi2): Likewise.
10208 (*thumb1_negsi2): Likewise.
10209 (*negdi_extendsid): Likewise.
10210 (*negdi_zero_extend): Likewise.
10211 (*arm_abssi2): Likewise.
10212 (*thumb1_abssi2): Likewise.
10213 (*arm_neg_abssi2): Likewise.
10214 (*thumb1_neg_abss): Likewise.
10215 (one_cmpldi2): Likewise.
10216 (extend<mode>di2): Likewise.
10217 (*compareqi_eq0): Likewise.
10218 (*arm_extendhisi2addsi): Likewise.
10219 (*arm_movdi): Likewise.
10220 (*thumb1_movdi_insn): Likewise.
10221 (*arm_movt): Likewise.
10222 (*thumb1_movsi_insn): Likewise.
10223 (pic_add_dot_plus_four): Likewise.
10224 (pic_add_dot_plus_eight): Likewise.
10225 (tls_load_dot_plus_eight): Likewise.
10226 (*thumb1_movhi_insn): Likewise.
10227 (*thumb1_movsf_insn): Likewise.
10228 (*movdf_soft_insn): Likewise.
10229 (*thumb_movdf_insn): Likewise.
10230 (cbranchsi4_insn): Likewise.
10231 (cbranchsi4_scratch): Likewise.
10232 (*negated_cbranchsi4): Likewise.
10233 (*tbit_cbranch): Likewise.
10234 (*tlobits_cbranch): Likewise.
10235 (*tstsi3_cbranch): Likewise.
10236 (*cbranchne_decr1): Likewise.
10237 (*addsi3_cbranch): Likewise.
10238 (*addsi3_cbranch_scratch): Likewise.
10239 (*arm_cmpdi_insn): Likewise.
10240 (*arm_cmpdi_unsig): Likewise.
10241 (*arm_cmpdi_zero): Likewise.
10242 (*thumb_cmpdi_zero): Likewise.
10243 (*deleted_compare): Likewise.
10244 (*mov_scc): Likewise.
10245 (*mov_negscc): Likewise.
10246 (*mov_notscc): Likewise.
10247 (*cstoresi_eq0_thumb1_insn): Likewise.
10248 (cstoresi_nltu_thumb1): Likewise.
10249 (cstoresi_ltu_thu): Likewise.
10250 (thumb1_addsi3_addgeu): Likewise.
10251 (*arm_jump): Likewise.
10252 (*thumb_jump): Likewise.
10253 (*check_arch2): Likewise.
10254 (arm_casesi_internal): Likewise.
10255 (thumb1_casesi_dispatch): Likewise.
10256 (*arm_indirect_jump): Likewise.
10257 (*thumb1_indirect_jump): Likewise.
10259 (*and_scc): Likewise.
10260 (*ior_scc): Likewise.
10261 (*compare_scc): Likewise.
10262 (*cond_move): Likewise.
10263 (*cond_arith): Likewise.
10264 (*cond_sub): Likewise.
10265 (*cmp_ite0): Likewise.
10266 (*cmp_ite1): Likewise.
10267 (*cmp_and): Likewise.
10268 (*cmp_ior): Likewise.
10269 (*ior_scc_scc): Likewise.
10270 (*ior_scc_scc_cmp): Likewise.
10271 (*and_scc_scc): Likewise.
10272 (*and_scc_scc_cmp): Likewise.
10273 (*and_scc_scc_nod): Likewise.
10274 (*negscc): Likewise.
10275 (movcond_addsi): Likewise.
10276 (movcond): Likewise.
10277 (*ifcompare_plus_move): Likewise.
10278 (*if_plus_move): Likewise.
10279 (*ifcompare_move_plus): Likewise.
10280 (*if_move_plus): Likewise.
10281 (*ifcompare_arith_arith): Likewise.
10282 (*if_arith_arith): Likewise.
10283 (*ifcompare_arith_move): Likewise.
10284 (*if_arith_move): Likewise.
10285 (*ifcompare_move_arith): Likewise.
10286 (*if_move_arith): Likewise.
10287 (*ifcompare_move_not): Likewise.
10288 (*if_move_not): Likewise.
10289 (*ifcompare_not_move): Likewise.
10290 (*if_not_move): Likewise.
10291 (*ifcompare_shift_move): Likewise.
10292 (*if_shift_move): Likewise.
10293 (*ifcompare_move_shift): Likewise.
10294 (*if_move_shift): Likewise.
10295 (*ifcompare_shift_shift): Likewise.
10296 (*ifcompare_not_arith): Likewise.
10297 (*ifcompare_arith_not): Likewise.
10298 (*if_arith_not): Likewise.
10299 (*ifcompare_neg_move): Likewise.
10300 (*if_neg_move): Likewise.
10301 (*ifcompare_move_neg): Likewise.
10302 (*if_move_neg): Likewise.
10303 (prologue_thumb1_interwork): Likewise.
10304 (*cond_move_not): Likewise.
10305 (*sign_extract_onebit): Likewise.
10306 (*not_signextract_onebit): Likewise.
10307 (stack_tie): Likewise.
10308 (align_4): Likewise.
10309 (align_8): Likewise.
10310 (consttable_end): Likewise.
10311 (consttable_1): Likewise.
10312 (consttable_2): Likewise.
10313 (consttable_4): Likewise.
10314 (consttable_8): Likewise.
10315 (consttable_16): Likewise.
10316 (*thumb1_tablejump): Likewise.
10317 (prefetch): Likewise.
10318 (force_register_use): Likewise.
10319 (thumb_eh_return): Likewise.
10320 (load_tp_hard): Likewise.
10321 (load_tp_soft): Likewise.
10322 (tlscall): Likewise.
10323 (*arm_movtas_ze): Likewise.
10324 (*arm_rev): Likewise.
10325 (*arm_revsh): Likewise.
10326 (*arm_rev16): Likewise.
10327 * config/arm/thumb2.md
10328 (*thumb2_smaxsi3): Likewise.
10329 (*thumb2_sminsi3): Likewise.
10330 (*thumb32_umaxsi3): Likewise.
10331 (*thumb2_uminsi3): Likewise.
10332 (*thumb2_negdi2): Likewise.
10333 (*thumb2_abssi2): Likewise.
10334 (*thumb2_neg_abss): Likewise.
10335 (*thumb2_movsi_insn): Likewise.
10336 (tls_load_dot_plus_four): Likewise.
10337 (*thumb2_movhi_insn): Likewise.
10338 (*thumb2_mov_scc): Likewise.
10339 (*thumb2_mov_negs): Likewise.
10340 (*thumb2_mov_negs): Likewise.
10341 (*thumb2_mov_nots): Likewise.
10342 (*thumb2_mov_nots): Likewise.
10343 (*thumb2_movsicc_): Likewise.
10344 (*thumb2_movsfcc_soft_insn): Likewise.
10345 (*thumb2_indirect_jump): Likewise.
10346 (*thumb2_and_scc): Likewise.
10347 (*thumb2_ior_scc): Likewise.
10348 (*thumb2_ior_scc_strict_it): Likewise.
10349 (*thumb2_cond_move): Likewise.
10350 (*thumb2_cond_arith): Likewise.
10351 (*thumb2_cond_ari): Likewise.
10352 (*thumb2_cond_sub): Likewise.
10353 (*thumb2_negscc): Likewise.
10354 (*thumb2_movcond): Likewise.
10355 (thumb2_casesi_internal): Likewise.
10356 (thumb2_casesi_internal_pic): Likewise.
10357 (*thumb2_alusi3_short): Likewise.
10358 (*thumb2_mov<mode>_shortim): Likewise.
10359 (*thumb2_addsi_short): Likewise.
10360 (*thumb2_subsi_short): Likewise.
10361 (thumb2_addsi3_compare0): Likewise.
10362 (*thumb2_cbz): Likewise.
10363 (*thumb2_cbnz): Likewise.
10364 (*thumb2_one_cmplsi2_short): Likewise.
10365 (*thumb2_negsi2_short): Likewise.
10366 (*orsi_notsi_si): Likewise.
10367 * config/arm/arm1020e.md: Update with new attributes.
10368 * config/arm/arm1026ejs.md: Update with new attributes.
10369 * config/arm/arm1136jfs.md: Update with new attributes.
10370 * config/arm/arm926ejs.md: Update with new attributes.
10371 * config/arm/cortex-a15.md: Update with new attributes.
10372 * config/arm/cortex-a5.md: Update with new attributes.
10373 * config/arm/cortex-a53.md: Update with new attributes.
10374 * config/arm/cortex-a7.md: Update with new attributes.
10375 * config/arm/cortex-a8.md: Update with new attributes.
10376 * config/arm/cortex-a9.md: Update with new attributes.
10377 * config/arm/cortex-m4.md: Update with new attributes.
10378 * config/arm/cortex-r4.md: Update with new attributes.
10379 * config/arm/fa526.md: Update with new attributes.
10380 * config/arm/fa606te.md: Update with new attributes.
10381 * config/arm/fa626te.md: Update with new attributes.
10382 * config/arm/fa726te.md: Update with new attributes.
10384 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10386 * config/aarch64/aarch64-simd.md
10387 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
10388 <vwx> iterator to ensure correct register choice.
10389 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
10390 (aarch64_sqdmull_n<mode>): Likewise.
10391 (aarch64_sqdmull2_n<mode>_internal): Likewise.
10392 * config/aarch64/arm_neon.h
10393 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
10394 (vml<as><q>_n_<su>16): Likewise.
10395 (vml<as>l_high_lane<q>_<su>16): Likewise.
10396 (vml<as>l_high_n_<su>16): Likewise.
10397 (vml<as>l_lane<q>_<su>16): Likewise.
10398 (vml<as>l_n_<su>16): Likewise.
10399 (vmul<q>_lane<q>_<su>16): Likewise.
10400 (vmul<q>_n_<su>16): Likewise.
10401 (vmull_lane<q>_<su>16): Likewise.
10402 (vmull_n_<su>16): Likewise.
10403 (vmull_high_lane<q>_<su>16): Likewise.
10404 (vmull_high_n_<su>16): Likewise.
10405 (vqrdmulh<q>_n_s16): Likewise.
10407 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
10409 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
10410 have the correct lane parameter.
10412 2013-09-06 Richard Biener <rguenther@suse.de>
10414 * cfganal.c (control_dependences::~control_dependences):
10415 Properly free all of the vector.
10417 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
10420 * config/i386/i386.c (ix86_conditional_register_usage):
10421 Proper initialize extended SSE registers.
10423 2013-09-06 Jan Hubicka <jh@suse.cz>
10425 PR tree-optimization/58311
10426 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
10428 2013-09-06 Jan Hubicka <jh@suse.cz>
10430 * Makefile.in (tree-sra.o): Update dependencies.
10431 * tree-sra.c: Include ipa-utils.h
10432 (scan_function): Use recursive_call_p.
10433 (has_caller_p): New function.
10434 (cgraph_for_node_and_aliases): Count also callers of aliases.
10436 2013-09-06 Jan Hubicka <jh@suse.cz>
10438 PR middle-end/58094
10439 * cgraph.h (symtab_semantically_equivalent_p): Declare.
10440 * tree-tailcall.c: Include ipa-utils.h.
10441 (find_tail_calls): Use it.
10442 * ipa-pure-const.c (check_call): Likewise.
10443 * ipa-utils.c (recursive_call_p): New function.
10444 * ipa-utils.h (recursive_call_p): Dclare.
10445 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
10446 (symtab_semantically_equivalent_p): New function.
10447 * Makefile.in (tree-tailcall.o): Update dependencies.
10449 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
10451 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
10452 non-inlinable part.
10454 2013-09-06 Richard Biener <rguenther@suse.de>
10456 * lto-streamer.h (lto_global_var_decls): Remove.
10457 * Makefile.in (OBJS): Remove lto-symtab.o.
10458 (lto-symtab.o): Remove.
10459 (GTFILES): Remove lto-symtab.c
10460 * lto-symtab.c: Move to lto/
10462 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10464 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
10465 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
10466 (UNSPEC_FPINT_RINT): New constant definitions.
10467 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
10468 definition with 2 attributes.
10469 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
10470 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
10473 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10475 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
10476 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
10477 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
10478 attribute to "z196_alone".
10479 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
10482 2013-09-06 Richard Biener <rguenther@suse.de>
10484 * basic-block.h (class control_dependences): New.
10485 * tree-ssa-dce.c (control_dependence_map): Remove.
10487 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
10488 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
10489 find_pdom, find_control_dependence, find_all_control_dependences):
10491 (mark_control_dependent_edges_necessary,
10492 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
10493 tree_dce_done, perform_tree_ssa_dce): Adjust.
10494 * cfganal.c (set_control_dependence_map_bit,
10495 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
10496 find_all_control_dependences): Move from tree-ssa-dce.c and
10497 implement as methods of control_dependences class.
10498 (control_dependences::control_dependences): New.
10499 (control_dependences::~control_dependences): Likewise.
10500 (control_dependences::get_edges_dependent_on): Likewise.
10501 (control_dependences::get_edge): Likewise.
10503 2013-09-04 Jan Hubicka <jh@suse.cz>
10505 * tree.c (types_same_for_odr): Drop overactive check.
10506 * ipa-devirt.c (hash_type_name): Likewise.
10508 2013-09-04 Jan Hubicka <jh@suse.cz>
10510 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
10511 (analyze_functions): ... here.
10513 2013-09-04 Jan Hubicka <jh@suse.cz>
10515 PR middle-end/58201
10516 * cgraphunit.c (analyze_functions): Clear AUX fields
10517 after processing; initialize assembler name has.
10519 2013-09-05 Jeff Law <law@redhat.com>
10521 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
10522 from thread_around_empty_block. Record threading path into PATH.
10523 Recurse if threading through the initial block is successful.
10524 (thread_across_edge): Corresponding changes to slightly simplify.
10526 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10528 * config/aarch64/aarch64.md
10529 (type): Remove frecpe, frecps, frecpx.
10530 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
10531 fix to be a TARGET_SIMD instruction.
10532 (aarch64_frecps): Remove.
10533 * config/aarch64/aarch64-simd.md
10534 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
10535 (aarch64_frecps<mode>): Handle all float/vector of float modes.
10537 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10538 Sofiane Naci <sofiane.naci@arm.com>
10540 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
10541 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
10542 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
10543 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
10544 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
10545 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
10546 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
10547 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
10548 into "clz, "rbit". Rename "shift" to "shift_imm".
10549 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
10550 changes. Update for attribute changes all occurrences of arlo_* and
10552 * config/arm/arm-fixed.md: Update for attribute changes
10553 all occurrences of arlo_* types.
10554 * config/arm/thumb2.md: Update for attribute changes all occurrences
10556 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
10557 (cortexa7_older_only): Likewise.
10558 (cortexa7_younger): Likewise.
10559 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10560 (1020alu_shift_op): Likewise.
10561 (1020alu_shift_reg_op): Likewise.
10562 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10563 (alu_shift_op): Likewise.
10564 (alu_shift_reg_op): Likewise.
10565 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10566 (11_alu_shift_op): Likewise.
10567 (11_alu_shift_reg_op): Likewise.
10568 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10569 (9_alu_shift_reg_op): Likewise.
10570 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
10572 (cortex_a15_alu_shift): Likewise.
10573 (cortex_a15_alu_shift_reg): Likewise.
10574 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
10576 (cortex_a5_alu_shift): Likewise.
10577 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
10579 (cortex_a53_alu_shift): Likewise.
10580 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
10582 (cortex_a7_alu_reg): Likewise.
10583 (cortex_a7_alu_shift): Likewise.
10584 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
10586 (cortex_a8_alu_shift): Likewise.
10587 (cortex_a8_alu_shift_reg): Likewise.
10588 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10589 (cortex_a9_dp_shift): Likewise.
10590 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
10592 * config/arm/cortex-r4.md
10593 (cortex_r4_alu): Update for attribute changes.
10594 (cortex_r4_mov): Likewise.
10595 (cortex_r4_alu_shift_reg): Likewise.
10596 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10597 (526_alu_shift_op): Likewise.
10598 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10599 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10600 (626te_alu_shift_op): Likewise.
10601 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
10602 (726te_alu_shift_op): Likewise.
10603 (726te_alu_shift_reg_op): Likewise.
10604 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10605 (mp626_alu_shift_op): Likewise.
10606 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
10607 (pj4_alu_conds): Likewise.
10608 (pj4_shift): Likewise.
10609 (pj4_shift_conds): Likewise.
10610 (pj4_alu_shift): Likewise.
10611 (pj4_alu_shift_conds): Likewise.
10612 * config/aarch64/aarch64.md: Update for attribute change
10613 all occurrences of arlo_* and shift* types.
10615 2013-09-05 Mike Stump <mikestump@comcast.net>
10617 * tree.h: Move documentation for tree_function_decl to tree-core.h
10618 with the declaration.
10620 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
10623 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
10624 looking for widest mode.
10626 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
10628 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
10630 2013-09-05 Richard Biener <rguenther@suse.de>
10632 PR tree-optimization/58137
10633 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
10634 Do not create vectors of pointers.
10635 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
10636 types for the components of the vector initializer.
10637 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
10638 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
10640 2013-09-05 Martin Jambor <mjambor@suse.cz>
10642 * ipa-prop.c (remove_described_reference): Accept missing references,
10643 return false if that hppens, otherwise return true.
10644 (cgraph_node_for_jfunc): New function.
10645 (try_decrement_rdesc_refcount): Likewise.
10646 (try_make_edge_direct_simple_call): Use them.
10647 (ipa_edge_removal_hook): Remove references from rdescs.
10648 (ipa_edge_duplication_hook): Clone rdescs and their references
10649 when the new edge has the same caller as the old one.
10650 * cgraph.c (cgraph_resolve_speculation): Remove speculative
10651 reference before removing any edges.
10653 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
10655 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
10657 * thumb2.md (thumb2_storewb_parisi): New pattern.
10659 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
10661 * config/aarch64/aarch64-option-extensions.def: Add
10662 AARCH64_OPT_EXTENSION of 'crc'.
10663 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
10664 (AARCH64_ISA_CRC): Ditto.
10665 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
10666 description of the CRC extension.
10668 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
10670 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
10671 * config/rs6000/linux.h: Ditto.
10672 * alpha/linux.h: Ditto.
10673 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
10674 no_c99_libc_has_function.
10675 * config/c6x/uclinux-elf.h: Ditto.
10676 * config/lm32/uclinux-elf.h: Ditto.
10677 * config/m68k/uclinux.h: Ditto.
10678 * config/moxie/uclinux.h: Ditto.
10679 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
10680 (crisv32-*-linux*, cris-*-linux*): Ditto.
10681 * config/bfin/bfin.c: Include "tm_p.h".
10683 2013-09-05 Richard Biener <rguenther@suse.de>
10685 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
10686 check for a definition without a basic-block.
10688 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10689 Sofiane Naci <sofiane.naci@arm.com>
10691 * config/aarch64/aarch64.md
10692 (*movti_aarch64): Rename r_2_f and f_2_r.
10693 (*movsf_aarch64): Likewise.
10694 (*movdf_aarch64): Likewise.
10695 (*movtf_aarch64): Likewise.
10696 (aarch64_movdi_<mode>low): Likewise.
10697 (aarch64_movdi_<mode>high): Likewise.
10698 (aarch64_mov<mode>high_di): Likewise.
10699 (aarch64_mov<mode>low_di): Likewise.
10700 (aarch64_movtilow_tilow): Likewise.
10701 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
10702 values to config/arm/types.md
10703 (attribute "conds"): Update for attribute change.
10704 (anddi3_insn): Likewise.
10705 (iordi3_insn): Likewise.
10706 (xordi3_insn): Likewise.
10707 (one_cmpldi2): Likewise.
10708 * config/arm/types.md (type): Add Neon types.
10709 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
10710 use "type" attribute.
10711 (movmisalign<mode>_neon_store): Likewise.
10712 (movmisalign<mode>_neon_load): Likewise.
10713 (vec_set<mode>_internal): Likewise.
10714 (vec_setv2di_internal): Likewise.
10715 (vec_extract<mode>): Likewise.
10716 (vec_extractv2di): Likewise.
10717 (add<mode>3_neon): Likewise.
10718 (adddi3_neon): Likewise.
10719 (sub<mode>3_neon): Likewise.
10720 (subdi3_neon): Likewise.
10721 (mul<mode>3_neon): Likewise.
10722 (mul<mode>3add<mode>_neon): Likewise.
10723 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
10724 (fma<VCVTF:mode>4)): Likewise.
10725 (fma<VCVTF:mode>4_intrinsic): Likewise.
10726 (fmsub<VCVTF:mode>4)): Likewise.
10727 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
10728 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
10729 (ior<mode>3): Likewise.
10730 (and<mode>3): Likewise.
10731 (anddi3_neon): Likewise.
10732 (orn<mode>3_neon): Likewise.
10733 (orndi3_neon): Likewise.
10734 (bic<mode>3_neon): Likewise.
10735 (bicdi3_neon): Likewise.
10736 (xor<mode>3): Likewise.
10737 (one_cmpl<mode>2): Likewise.
10738 (abs<mode>2): Likewise.
10739 (neg<mode>2): Likewise.
10740 (umin<mode>3_neon): Likewise.
10741 (umax<mode>3_neon): Likewise.
10742 (smin<mode>3_neon): Likewise.
10743 (smax<mode>3_neon): Likewise.
10744 (vashl<mode>3): Likewise.
10745 (vashr<mode>3_imm): Likewise.
10746 (vlshr<mode>3_imm): Likewise.
10747 (ashl<mode>3_signed): Likewise.
10748 (ashl<mode>3_unsigned): Likewise.
10749 (neon_load_count): Likewise.
10750 (ashldi3_neon_noclobber): Likewise.
10751 (signed_shift_di3_neon): Likewise.
10752 (unsigned_shift_di3_neon): Likewise.
10753 (ashrdi3_neon_imm_noclobber): Likewise.
10754 (lshrdi3_neon_imm_noclobber): Likewise.
10755 (widen_ssum<mode>3): Likewise.
10756 (widen_usum<mode>3): Likewise.
10757 (quad_halves_<code>v4si): Likewise.
10758 (quad_halves_<code>v4sf): Likewise.
10759 (quad_halves_<code>v8hi): Likewise.
10760 (quad_halves_<code>v16qi): Likewise.
10761 (reduc_splus_v2di): Likewise.
10762 (neon_vpadd_internal<mode>): Likewise.
10763 (neon_vpsmin<mode>): Likewise.
10764 (neon_vpsmax<mode>): Likewise.
10765 (neon_vpumin<mode>): Likewise.
10766 (neon_vpumax<mode>): Likewise.
10767 (ss_add<mode>_neon): Likewise.
10768 (us_add<mode>_neon): Likewise.
10769 (ss_sub<mode>_neon): Likewise.
10770 (us_sub<mode>_neon): Likewise.
10771 (neon_vadd<mode>_unspec): Likewise.
10772 (neon_vaddl<mode>): Likewise.
10773 (neon_vaddw<mode>): Likewise.
10774 (neon_vhadd<mode>): Likewise.
10775 (neon_vqadd<mode>): Likewise.
10776 (neon_vaddhn<mode>): Likewise.
10777 (neon_vmul<mode>): Likewise.
10778 (neon_vmla<mode>): Likewise.
10779 (neon_vmlal<mode>): Likewise.
10780 (neon_vmls<mode>): Likewise.
10781 (neon_vmlsl<mode>): Likewise.
10782 (neon_vqdmulh<mode>): Likewise.
10783 (neon_vqdmlal<mode>): Likewise.
10784 (neon_vqdmlsl<mode>): Likewise.
10785 (neon_vmull<mode>): Likewise.
10786 (neon_vqdmull<mode>): Likewise.
10787 (neon_vsub<mode>_unspec): Likewise.
10788 (neon_vsubl<mode>): Likewise.
10789 (neon_vsubw<mode>): Likewise.
10790 (neon_vqsub<mode>): Likewise.
10791 (neon_vhsub<mode>): Likewise.
10792 (neon_vsubhn<mode>): Likewise.
10793 (neon_vceq<mode>): Likewise.
10794 (neon_vcge<mode>): Likewise.
10795 (neon_vcgeu<mode>): Likewise.
10796 (neon_vcgt<mode>): Likewise.
10797 (neon_vcgtu<mode>): Likewise.
10798 (neon_vcle<mode>): Likewise.
10799 (neon_vclt<mode>): Likewise.
10800 (neon_vcage<mode>): Likewise.
10801 (neon_vcagt<mode>): Likewise.
10802 (neon_vtst<mode>): Likewise.
10803 (neon_vabd<mode>): Likewise.
10804 (neon_vabdl<mode>): Likewise.
10805 (neon_vaba<mode>): Likewise.
10806 (neon_vabal<mode>): Likewise.
10807 (neon_vmax<mode>): Likewise.
10808 (neon_vmin<mode>): Likewise.
10809 (neon_vpaddl<mode>): Likewise.
10810 (neon_vpadal<mode>): Likewise.
10811 (neon_vpmax<mode>): Likewise.
10812 (neon_vpmin<mode>): Likewise.
10813 (neon_vrecps<mode>): Likewise.
10814 (neon_vrsqrts<mode>): Likewise.
10815 (neon_vqabs<mode>): Likewise.
10816 (neon_vqneg<mode>): Likewise.
10817 (neon_vcls<mode>): Likewise.
10818 (clz<mode>2): Likewise.
10819 (popcount<mode>2): Likewise.
10820 (neon_vrecpe): Likewise.
10821 (neon_vrsqrte): Likewise.
10822 (neon_vget_lane<mode>_sext_internal): Likewise.
10823 (neon_vget_lane<mode>_zext_internal): Likewise.
10824 (neon_vdup_n<mode>): Likewise.
10825 (neon_vdup_nv2di): Likewise.
10826 (neon_vdpu_lane<mode>_internal): Likewise.
10827 (neon_vswp<mode>): Likewise.
10828 (float<mode><V_cvtto>2): Likewise.
10829 (floatuns<mode><V_cvtto>2): Likewise.
10830 (fix_trunc<mode><V_cvtto>)2): Likewise
10831 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10832 (neon_vcvt<mode>): Likewise.
10833 (neon_vcvtv4sfv4hf): Likewise.
10834 (neon_vcvtv4hfv4sf): Likewise.
10835 (neon_vcvt_n<mode>): Likewise.
10836 (neon_vmovn<mode>): Likewise.
10837 (neon_vqmovn<mode>): Likewise.
10838 (neon_vqmovun<mode>): Likewise.
10839 (neon_vmovl<mode>): Likewise.
10840 (neon_vmul_lane<mode>): Likewise.
10841 (neon_vmull_lane<mode>): Likewise.
10842 (neon_vqdmull_lane<mode>): Likewise.
10843 (neon_vqdmulh_lane<mode>): Likewise.
10844 (neon_vmla_lane<mode>): Likewise.
10845 (neon_vmlal_lane<mode>): Likewise.
10846 (neon_vqdmlal_lane<mode>): Likewise.
10847 (neon_vmls_lane<mode>): Likewise.
10848 (neon_vmlsl_lane<mode>): Likewise.
10849 (neon_vqdmlsl_lane<mode>): Likewise.
10850 (neon_vext<mode>): Likewise.
10851 (neon_vrev64<mode>): Likewise.
10852 (neon_vrev32<mode>): Likewise.
10853 (neon_vrev16<mode>): Likewise.
10854 (neon_vbsl<mode>_internal): Likewise.
10855 (neon_vshl<mode>): Likewise.
10856 (neon_vqshl<mode>): Likewise.
10857 (neon_vshr_n<mode>): Likewise.
10858 (neon_vshrn_n<mode>): Likewise.
10859 (neon_vqshrn_n<mode>): Likewise.
10860 (neon_vqshrun_n<mode>): Likewise.
10861 (neon_vshl_n<mode>): Likewise.
10862 (neon_vqshl_n<mode>): Likewise.
10863 (neon_vqshlu_n<mode>): Likewise.
10864 (neon_vshll_n<mode>): Likewise.
10865 (neon_vsra_n<mode>): Likewise.
10866 (neon_vsri_n<mode>): Likewise.
10867 (neon_vsli_n<mode>): Likewise.
10868 (neon_vtbl1v8qi): Likewise.
10869 (neon_vtbl2v8qi): Likewise.
10870 (neon_vtbl3v8qi): Likewise.
10871 (neon_vtbl4v8qi): Likewise.
10872 (neon_vtbx1v8qi): Likewise.
10873 (neon_vtbx2v8qi): Likewise.
10874 (neon_vtbx3v8qi): Likewise.
10875 (neon_vtbx4v8qi): Likewise.
10876 (neon_vtrn<mode>_internal): Likewise.
10877 (neon_vzip<mode>_internal): Likewise.
10878 (neon_vuzp<mode>_internal): Likewise.
10879 (neon_vld1<mode>): Likewise.
10880 (neon_vld1_lane<mode>): Likewise.
10881 (neon_vld1_dup<mode>): Likewise.
10882 (neon_vld1_dupv2di): Likewise.
10883 (neon_vst1<mode>): Likewise.
10884 (neon_vst1_lane<mode>): Likewise.
10885 (neon_vld2<mode>): Likewise.
10886 (neon_vld2_lane<mode>): Likewise.
10887 (neon_vld2_dup<mode>): Likewise.
10888 (neon_vst2<mode>): Likewise.
10889 (neon_vst2_lane<mode>): Likewise.
10890 (neon_vld3<mode>): Likewise.
10891 (neon_vld3qa<mode>): Likewise.
10892 (neon_vld3qb<mode>): Likewise.
10893 (neon_vld3_lane<mode>): Likewise.
10894 (neon_vld3_dup<mode>): Likewise.
10895 (neon_vst3<mode>): Likewise.
10896 (neon_vst3qa<mode>): Likewise.
10897 (neon_vst3qb<mode>): Likewise.
10898 (neon_vst3_lane<mode>): Likewise.
10899 (neon_vld4<mode>): Likewise.
10900 (neon_vld4qa<mode>): Likewise.
10901 (neon_vld4qb<mode>): Likewise.
10902 (neon_vld4_lane<mode>): Likewise.
10903 (neon_vld4_dup<mode>): Likewise.
10904 (neon_vst4<mode>): Likewise.
10905 (neon_vst4qa<mode>): Likewise.
10906 (neon_vst4qb<mode>): Likewise.
10907 (neon_vst4_lane<mode>): Likewise.
10908 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10909 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10910 (neon_vec_<US>mult_lo_<mode>): Likewise.
10911 (neon_vec_<US>mult_hi_<mode>): Likewise.
10912 (neon_vec_<US>shiftl_<mode>): Likewise.
10913 (neon_unpack<US>_<mode>): Likewise.
10914 (neon_vec_<US>mult_<mode>): Likewise.
10915 (vec_pack_trunc_<mode>): Likewise.
10916 (neon_vec_pack_trunk_<mode>): Likewise.
10917 (neon_vabd<mode>_2): Likewise.
10918 (neon_vabd<mode>_3): Likewise.
10919 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10920 (thumb2_movsi_vfp): Likewise.
10921 (movdi_vfp): Likewise.
10922 (movdi_vfp_cortexa8): Likewise.
10923 (movhf_vfp_neon): Likewise.
10924 (movhf_vfp): Likewiwse.
10925 (movsf_vfp): Likewiwse.
10926 (thumb2_movsf_vfp): Likewiwse.
10927 (movdf_vfp): Likewise.
10928 (thumb2_movdf_vfp): Likewise.
10929 (movsfcc_vfp): Likewise.
10930 (thumb2_movsfcc_vfp): Likewise.
10931 (movdfcc_vfp): Likewise.
10932 (thumb2_movdfcc_vfp): Likewise.
10933 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10934 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10935 (v10_v2c): Likewise.
10936 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10938 (cortex_a15_neon_int_2): Likewise.
10939 (cortex_a15_neon_int_3): Likewise.
10940 (cortex_a15_neon_int_4): Likewise.
10941 (cortex_a15_neon_int_5): Likewise.
10942 (cortex_a15_neon_vqneg_vqabs): Likewise.
10943 (cortex_a15_neon_vmov): Likewise.
10944 (cortex_a15_neon_vaba): Likewise.
10945 (cortex_a15_neon_vaba_qqq): Likewise.
10946 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10947 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10948 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10949 scalar_64_32_long_scalar): Likewise.
10950 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10951 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10952 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10953 lotype_qdd_64_32_long): Likewise.
10954 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10955 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10956 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10957 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10958 (cortex_a15_neon_shift_1): Likewise.
10959 (cortex_a15_neon_shift_2): Likewise.
10960 (cortex_a15_neon_shift_3): Likewise.
10961 (cortex_a15_neon_vshl_ddd): Likewise.
10962 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10963 (cortex_a15_neon_vsra_vrsra): Likewise.
10964 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10965 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10966 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10967 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10968 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10969 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10970 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10971 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10972 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10973 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10974 (cortex_a15_neon_bp_simple): Likewise.
10975 (cortex_a15_neon_bp_2cycle): Likewise.
10976 (cortex_a15_neon_bp_3cycle): Likewise.
10977 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10978 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10979 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10980 (cortex_a15_neon_vld2_4_regs): Likewise.
10981 (cortex_a15_neon_vld3_vld4): Likewise.
10982 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10983 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10984 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10985 (cortex_a15_neon_vst3_vst4): Likewise.
10986 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10987 (cortex_a15_neon_vld3_vld4_lane" 10
10988 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10989 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10990 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10991 (cortex_a15_neon_ldm_2): Likewise.0
10992 (cortex_a15_neon_stm_2): Likewise.
10993 (cortex_a15_neon_mcr): Likewise.
10994 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10995 (cortex_a15_neon_mrc): Likewise.
10996 (cortex_a15_neon_mrrc): Likewise.
10997 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10999 (cortex_a15_alu_shift): Likewise.
11000 (cortex_a15_alu_shift_reg): Likewise.
11001 (cortex_a15_mult32): Likewise.
11002 (cortex_a15_mult64): Likewise.
11003 (cortex_a15_block): Likewise.
11004 (cortex_a15_branch): Likewise.
11005 (cortex_a15_load1): Likewise.
11006 (cortex_a15_load3): Likewise.
11007 (cortex_a15_store1): Likewise.
11008 (cortex_a15_store3): Likewise.
11009 (cortex_a15_call): Likewise.
11010 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
11011 (cortex_a5_f2r): Likewise.
11012 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
11014 (cortex_a53_f2r): Likewise.
11015 * config/arm/cortex-a7.md
11016 (cortex_a7_branch): Update for attribute change.
11017 (cortex_a7_call): Likewise.
11018 (cortex_a7_alu_imm): Likewise.
11019 (cortex_a7_alu_reg): Likewise.
11020 (cortex_a7_alu_shift): Likewise.
11021 (cortex_a7_mul): Likewise.
11022 (cortex_a7_load1): Likewise.
11023 (cortex_a7_store1): Likewise.
11024 (cortex_a7_load2): Likewise.
11025 (cortex_a7_store2): Likewise.
11026 (cortex_a7_load3): Likewise.
11027 (cortex_a7_store3): Likewise.
11028 (cortex_a7_load4): Likewise.
11029 (cortex_a7_store4): Likewise.
11030 (cortex_a7_fpalu): Likewise.
11031 (cortex_a7_fconst): Likewise.
11032 (cortex_a7_fpmuls): Likewise.
11033 (cortex_a7_neon_mul): Likewise.
11034 (cortex_a7_fpmacs): Likewise.
11035 (cortex_a7_neon_mla: Likewise.
11036 (cortex_a7_fpmuld: Likewise.
11037 (cortex_a7_fpmacd: Likewise.
11038 (cortex_a7_fpfmad: Likewise.
11039 (cortex_a7_fdivs: Likewise.
11040 (cortex_a7_fdivd: Likewise.
11041 (cortex_a7_r2f: Likewise.
11042 (cortex_a7_f2r: Likewise.
11043 (cortex_a7_f_flags: Likewise.
11044 (cortex_a7_f_loads: Likewise.
11045 (cortex_a7_f_loadd: Likewise.
11046 (cortex_a7_f_stores: Likewise.
11047 (cortex_a7_f_stored: Likewise.
11048 (cortex_a7_neon): Likewise.
11049 * config/arm/cortex-a8-neon.md
11050 (cortex_a8_neon_mrc): Update for attribute change.
11051 (cortex_a8_neon_mrrc): Likewise.
11052 (cortex_a8_neon_int_1): Likewise.
11053 (cortex_a8_neon_int_2): Likewise.
11054 (cortex_a8_neon_int_3): Likewise.
11055 (cortex_a8_neon_int_4): Likewise.
11056 (cortex_a8_neon_int_5): Likewise.
11057 (cortex_a8_neon_vqneg_vqabs): Likewise.
11058 (cortex_a8_neon_vmov): Likewise.
11059 (cortex_a8_neon_vaba): Likewise.
11060 (cortex_a8_neon_vaba_qqq): Likewise.
11061 (cortex_a8_neon_vsma): Likewise.
11062 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11063 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
11064 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
11065 long_scalar): Likewise.
11066 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11067 (cortex_a8_neon_mla_qqq_8_16): Likewise.
11068 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
11069 long_scalar_qdd_64_32_long): Likewise.
11070 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
11071 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
11072 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
11073 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
11074 (cortex_a8_neon_shift_1): Likewise.
11075 (cortex_a8_neon_shift_2): Likewise.
11076 (cortex_a8_neon_shift_3): Likewise.
11077 (cortex_a8_neon_vshl_ddd): Likewise.
11078 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
11079 (cortex_a8_neon_vsra_vrsra): Likewise.
11080 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
11081 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
11082 (cortex_a8_neon_fp_vsum): Likewise.
11083 (cortex_a8_neon_fp_vmul_ddd): Likewise.
11084 (cortex_a8_neon_fp_vmul_qqd): Likewise.
11085 (cortex_a8_neon_fp_vmla_ddd): Likewise.
11086 (cortex_a8_neon_fp_vmla_qqq): Likewise.
11087 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
11088 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
11089 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
11090 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
11091 (cortex_a8_neon_bp_simple): Likewise.
11092 (cortex_a8_neon_bp_2cycle): Likewise.
11093 (cortex_a8_neon_bp_3cycle): Likewise.
11094 (cortex_a8_neon_ldr): Likewise.
11095 (cortex_a8_neon_str): Likewise.
11096 (cortex_a8_neon_vld1_1_2_regs): Likewise.
11097 (cortex_a8_neon_vld1_3_4_regs): Likewise.
11098 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
11099 (cortex_a8_neon_vld2_4_regs): Likewise.
11100 (cortex_a8_neon_vld3_vld4): Likewise.
11101 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
11102 (cortex_a8_neon_vst1_3_4_regs): Likewise.
11103 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
11104 (cortex_a8_neon_vst3_vst4): Likewise.
11105 (cortex_a8_neon_vld1_vld2_lane): Likewise.
11106 (cortex_a8_neon_vld3_vld4_lane): Likewise.
11107 (cortex_a8_neon_vst1_vst2_lane): Likewise.
11108 (cortex_a8_neon_vst3_vst4_lane): Likewise.
11109 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
11110 (cortex_a8_neon_mcr): Likewise.
11111 (cortex_a8_neon_mcr_2_mcrr): Likewise.
11112 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11113 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
11115 (ca9_neon_mrrc): Likewise.
11116 (cortex_a9_neon_int_1): Likewise.
11117 (cortex_a9_neon_int_2): Likewise.
11118 (cortex_a9_neon_int_3): Likewise.
11119 (cortex_a9_neon_int_4): Likewise.
11120 (cortex_a9_neon_int_5): Likewise.
11121 (cortex_a9_neon_vqneg_vqabs): Likewise.
11122 (cortex_a9_neon_vmov): Likewise.
11123 (cortex_a9_neon_vaba): Likewise.
11124 (cortex_a9_neon_vaba_qqq): Likewise.
11125 (cortex_a9_neon_vsma): Likewise.
11126 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11127 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
11128 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
11129 long_scalar): Likewise.
11130 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
11131 (cortex_a9_neon_mla_qqq_8_16): Likewise.
11132 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
11133 long_scalar_qdd_64_32_long): Likewise.
11134 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
11135 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
11136 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
11137 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
11138 (cortex_a9_neon_shift_1): Likewise.
11139 (cortex_a9_neon_shift_2): Likewise.
11140 (cortex_a9_neon_shift_3): Likewise.
11141 (cortex_a9_neon_vshl_ddd): Likewise.
11142 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
11143 (cortex_a9_neon_vsra_vrsra): Likewise.
11144 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
11145 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
11146 (cortex_a9_neon_fp_vsum): Likewise.
11147 (cortex_a9_neon_fp_vmul_ddd): Likewise.
11148 (cortex_a9_neon_fp_vmul_qqd): Likewise.
11149 (cortex_a9_neon_fp_vmla_ddd): Likewise.
11150 (cortex_a9_neon_fp_vmla_qqq): Likewise.
11151 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
11152 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
11153 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
11154 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
11155 (cortex_a9_neon_bp_simple): Likewise.
11156 (cortex_a9_neon_bp_2cycle): Likewise.
11157 (cortex_a9_neon_bp_3cycle): Likewise.
11158 (cortex_a9_neon_ldr): Likewise.
11159 (cortex_a9_neon_str): Likewise.
11160 (cortex_a9_neon_vld1_1_2_regs): Likewise.
11161 (cortex_a9_neon_vld1_3_4_regs): Likewise.
11162 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
11163 (cortex_a9_neon_vld2_4_regs): Likewise.
11164 (cortex_a9_neon_vld3_vld4): Likewise.
11165 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
11166 (cortex_a9_neon_vst1_3_4_regs): Likewise.
11167 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
11168 (cortex_a9_neon_vst3_vst4): Likewise.
11169 (cortex_a9_neon_vld1_vld2_lane): Likewise.
11170 (cortex_a9_neon_vld3_vld4_lane): Likewise.
11171 (cortex_a9_neon_vst1_vst2_lane): Likewise.
11172 (cortex_a9_neon_vst3_vst4_lane): Likewise.
11173 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
11174 (cortex_a9_neon_mcr): Likewise.
11175 (cortex_a9_neon_mcr_2_mcrr): Likewise.
11176 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11177 (cortex_a9_fps): Likewise.
11178 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
11180 (cortex_m4_fmuls): Likewise.
11181 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
11183 (cortex_r4_mrc): Likewise.
11184 * config/arm/iterators.md: Update comment referring to neon_type.
11185 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
11186 (iwmmxt_movsi_insn): Likewise.
11187 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
11189 (pj4_core_to_vfp): Likewise.
11190 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
11192 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
11193 (vfp_fstore): Likewise.
11194 * doc/md.texi: Change references to neon_type to refer to type.
11196 2013-09-04 Dodji Seketeli <dodji@redhat.com>
11198 * tree.h (DECL_BUILT_IN): Fix typo in comment.
11200 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
11202 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
11203 lglobl if not weak.
11205 2013-09-04 Easwaran Raman <eraman@google.com>
11207 PR middle-end/57370
11208 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
11209 (build_and_add_sum): Use it.
11210 (appears_later_in_bb): Simplify code.
11212 2013-09-04 Teresa Johnson <tejohnson@google.com>
11214 * dumpfile.c (dump_finish): Don't close stderr/stdout.
11216 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
11218 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
11220 2013-09-04 Jan Hubicka <jh@suse.cz>
11222 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
11223 * ipa-devirt.c: Include diganostic.h
11224 (odr_type_d): Add types and types_set.
11225 (hash_type_name): Work for types with vtables during LTO.
11226 (odr_hasher::remove): Fix comment; destroy types_set.
11227 (add_type_duplicate): New function,
11228 (get_odr_type): Use it.
11229 (dump_type_inheritance_graph): Dump type duplicates.
11230 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
11232 * tree.c (types_same_for_odr): Give exact answers on types with
11235 2013-09-04 Dodji Seketeli <dodji@redhat.com>
11237 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
11238 explaining their differences.
11240 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
11242 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
11244 2013-09-03 Jeff Law <law@redhat.com>
11246 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
11247 when not threading through a joiner block. Pass joiner/no joiner
11248 state to register_jump_thread.
11249 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
11250 state from argument rather than implying on path length.
11251 Dump the entire jump thread path into debugging dump.
11252 * tree-flow.h (register_jump_thread): Update prototype.
11254 2013-08-29 Xinliang David Li <davidxl@google.com>
11256 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11257 Remove a trivial gcc_assert.
11259 2013-08-29 Xinliang David Li <davidxl@google.com>
11261 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
11262 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
11263 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11264 Delay base decl alignment adjustment.
11265 * tree-vectorizer.c (vect_destroy_datarefs): New function.
11266 * tree-vectorizer.h: New data structure.
11267 (set_dr_misalignment): New function.
11268 (dr_misalignment): Ditto.
11269 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
11270 (vectorizable_load): Ditto.
11271 (ensure_base_align): New function.
11272 (vectorize_loops): Add dbg_cnt support.
11273 (execute_vect_slp): Ditto.
11274 * dbgcnt.def: New debug counter.
11275 * Makefile: New dependency.
11277 2013-09-03 Meador Inge <meadori@codesourcery.com>
11281 2013-08-30 Meador Inge <meadori@codesourcery.com>
11283 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11285 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
11287 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
11288 function descriptor.
11290 2013-09-03 Richard Biener <rguenther@suse.de>
11292 * tree-affine.c (add_elt_to_tree): Fix association issue,
11293 avoid useless converts and make sure to always return a
11294 properly typed result.
11296 2013-09-03 Richard Biener <rguenther@suse.de>
11298 PR middle-end/57656
11299 * fold-const.c (negate_expr_p): Fix division case.
11300 (negate_expr): Likewise.
11302 2013-09-03 Richard Biener <rguenther@suse.de>
11305 * tree-streamer-out.c: Include tm.h.
11306 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
11308 2013-09-03 Jan Hubicka <jh@suse.cz>
11310 * tree-profile.c (tree_profiling): Cleanup CFG when done.
11312 2013-09-03 Alan Modra <amodra@gmail.com>
11314 * config.gcc (powerpc*-*-linux*): Add support for little-endian
11315 multilibs to big-endian target and vice versa.
11316 * config/rs6000/t-linux64: Use := assignment on all vars.
11317 (MULTILIB_EXTRA_OPTS): Remove fPIC.
11318 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
11319 * config/rs6000/t-linux64le: New file.
11320 * config/rs6000/t-linux64bele: New file.
11321 * config/rs6000/t-linux64lebe: New file.
11323 2013-09-02 Jan Hubicka <jh@suse.cz>
11325 * ipa-inline-transform.c (inline_transform): Do not
11326 optimize_inline_calls when not optimizing.
11328 2013-09-02 Jan Hubicka <jh@suse.cz>
11330 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
11331 duplicated nodes for assembler names.
11332 * symtab.c (symtab_unregister_node): Do not attempt to unlink
11333 hard registers from assembler name hash.
11335 2013-09-02 Jan Hubicka <jh@suse.cz>
11337 * ipa-split.c (execute_split_functions): Split externally visible
11338 functions called once.
11340 2013-09-02 Martin Jambor <mjambor@suse.cz>
11343 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
11344 linked list. When finding the correct duplicate, also consider also
11345 the caller in additon to its inlined_to node.
11347 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
11349 * config/aarch64/aarch64-simd-builtins.def
11350 (dup_lane_scalar): Remove.
11351 * config/aarch64/aarch64-simd.md
11352 (aarch64_simd_dup): Add 'w->w' alternative.
11353 (aarch64_dup_lane<mode>): Allow for VALL.
11354 (aarch64_dup_lane_scalar<mode>): Remove.
11355 (aarch64_dup_lane_<vswap_width_name><mode>): New.
11356 (aarch64_get_lane_signed<mode>): Add w->w altenative.
11357 (aarch64_get_lane_unsigned<mode>): Likewise.
11358 (aarch64_get_lane<mode>): Likewise.
11359 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
11360 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
11361 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
11362 (VCON): Change container of V2SF.
11363 (vswap_width_name): Likewise.
11364 * config/aarch64/arm_neon.h
11365 (__aarch64_vdup_lane_any): New.
11366 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11367 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
11368 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11370 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
11372 PR middle-end/56382
11373 * expr.c (emit_move_complex): Do not move complex FP values as parts if
11374 the source or the destination is a single hard register.
11376 2013-09-02 Richard Biener <rguenther@suse.de>
11378 PR middle-end/57511
11379 * tree-scalar-evolution.c (instantiate_scev_name): Allow
11382 2013-09-02 Richard Biener <rguenther@suse.de>
11384 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
11385 arithmetic to sizetype.
11387 2013-09-02 Bin Cheng <bin.cheng@arm.com>
11389 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
11390 Find auto-increment use both before and after candidate.
11392 2013-09-02 Marek Polacek <polacek@redhat.com>
11394 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
11396 2013-09-01 Jan Hubicka <jh@suse.cz>
11398 * Makefile.in: Add ipa-profile.o
11399 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
11400 * cgraph.c (struct cgraph_propagate_frequency_data,
11401 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
11402 ipa-profile.c; replace cgraph_ by ipa_ prefix.
11403 * cgraph.h (cgraph_propagate_frequency): Remove.
11404 * ipa-inline-analysis.c: Include ipa-utils.h;
11405 drop duplicated cfgloop.h.
11406 (inline_update_callee_summaries): Update.
11407 * ipa-profile.c: New file.
11408 * ipa-utils.h (ipa_propagate_frequency): Declare.
11409 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
11410 data-streamer.h, value-prof.h.
11411 (symtab_remove_unreachable_nodes): Update profile.
11412 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
11413 account_time_size, cmp_counts, dump_histogram,
11414 ipa_profile_generate_summary, ipa_profile_write_summary,
11415 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
11416 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
11417 Move to ipa-profile.c.
11419 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
11421 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
11423 2013-09-01 Jan Hubicka <jh@suse.cz>
11425 * common.opt (fdevirtualize-speculatively): New function.
11426 * invoke.texi (fdevirtualize-speculatively): Document.
11427 * ipa-devirt.c: Include ipa-inline.h
11428 (likely_target_p): New function.
11429 (ipa_devirt): New function.
11430 (gate_ipa_devirt): New function.
11431 (pass_data_ipa_devirt): New static var.
11432 (pass_ipa_devirt): Likewise.
11433 (make_pass_ipa_devirt): New function.
11434 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
11435 (common_handle_option): Disable devirtualization when
11436 value range profiling is available.
11437 * passes.def (pass_ipa_devirt): Add.
11438 * timever.def (TV_IPA_DEVIRT): New timevar.
11439 * tree-pass.h (make_pass_ipa_devirt):
11441 2013-09-01 Iain Sandoe <iain@codesourcery.com>
11443 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
11444 include sanitize(undefined).
11446 2013-08-31 Diego Novillo <dnovillo@google.com>
11448 * Makefile.in (TREE_CORE_H): Define.
11450 (GTFILES): Add tree-core.h.
11451 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
11453 * tree-core.h: New file.
11454 Move all data structures, enum, typedefs, global
11455 declarations and constants from ...
11456 * tree.h: ... here.
11458 2013-08-31 Jan Hubicka <jh@suse.cz>
11460 * bulitins.c (expand_builtin): Do not early exit for gcov
11461 instrumented functions.
11463 2013-08-31 Marek Polacek <polacek@redhat.com>
11465 * ubsan.c: Include tm_p.h.
11467 2013-08-31 Jan Hubicka <jh@suse.cz>
11469 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
11472 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
11473 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
11474 * tree-cfg.c (verify_gimple_label): ... here.
11475 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
11476 (ipa_merge_profiles): New function.
11477 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
11478 (lto_input_function_body): Likewise.
11479 * ipa-utils.h (ipa_merge_profiles): Declare.
11480 * lto-streamer.h (lto_input_function_body): Update prototype.
11481 (emit_label_in_global_context_p): Remove.
11482 * lto-symtab.c: Include ipa-utils.h
11483 (lto_cgraph_replace_node): Use ipa_merge_profiles.
11485 2013-08-31 Jan Hubicka <jh@suse.cz>
11487 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
11489 2013-08-31 Jan Hubicka <jh@suse.cz>
11491 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
11493 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
11495 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
11496 "cmp" RTX before signed_comparison_operator check to account
11497 for "code" changes.
11499 2013-08-30 Jan Hubicka <jh@suse.cz>
11501 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
11502 (detect_type_change_1): Rename to ...
11503 (detect_type_change): ... this one; early return on non-polymorphic
11505 (detect_type_change_ssa): Add comp_type parameter; update
11506 use of detect_type_change.
11507 (compute_complex_assign_jump_func): Add param_type parameter;
11508 update use of detect_type_change_ssa.
11509 (compute_complex_ancestor_jump_func): Likewise.
11510 (ipa_get_callee_param_type): New function.
11511 (ipa_compute_jump_functions_for_edge): Compute parameter type;
11512 update calls to the jump function computation functions.
11514 2013-08-30 Teresa Johnson <tejohnson@google.com>
11515 Steven Bosscher <steven@gcc.gnu.org>
11517 * cfgrtl.c (fixup_new_cold_bb): New routine.
11518 (commit_edge_insertions): Invoke fixup_partitions.
11519 (find_partition_fixes): New routine.
11520 (fixup_partitions): Ditto.
11521 (verify_hot_cold_block_grouping): Update comments.
11522 (rtl_verify_edges): Invoke find_partition_fixes.
11523 (rtl_verify_bb_pointers): Update comments.
11524 (rtl_verify_bb_layout): Ditto.
11525 * basic-block.h (probably_never_executed_edge_p): Declare.
11526 (fixup_partitions): Ditto.
11527 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
11528 * bb-reorder.c (sanitize_hot_paths): New function.
11529 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
11530 sanitize_hot_paths.
11531 * predict.c (probably_never_executed_edge_p): New routine.
11532 * cfg.c (check_bb_profile): Add partition insanity warnings.
11534 2013-08-30 Meador Inge <meadori@codesourcery.com>
11536 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11538 2013-08-30 Marek Polacek <polacek@redhat.com>
11540 * Makefile.in (ubsan.o): Add.
11541 (c-family/c-ubsan.o): Add.
11542 (builtins.o): Add ubsan.h dependency.
11543 * ubsan.h: New file.
11544 * ubsan.c: New file.
11545 * common.opt: Add -fsanitize=undefined option.
11546 (flag_sanitize): Add variable.
11547 (fsanitize=): Add option. Add Driver.
11548 (fsanitize=thread): Remove option.
11549 (fsanitize=address): Likewise.
11550 (static-libubsan): New option.
11551 * doc/invoke.texi: Document the new flag and -static-libubsan.
11552 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
11553 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
11554 * builtin-attrs.def (ATTR_COLD): Define.
11555 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11556 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
11557 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
11558 * flag-types.h (sanitize_code): New enum.
11559 * opts.c (common_handle_option): Parse command line arguments
11560 of -fsanitize=. Add -fsanitize=unreachable option.
11561 * varasm.c (get_variable_section): Adjust.
11562 (assemble_noswitch_variable): Likewise.
11563 (assemble_variable): Likewise.
11564 (output_constant_def_contents): Likewise.
11565 (categorize_decl_for_section): Likewise.
11566 (place_block_symbol): Likewise.
11567 (output_object_block): Likewise.
11568 * builtins.def: Likewise.
11569 * toplev.c (compile_file): Likewise.
11570 (process_options): Likewise.
11571 * cppbuiltin.c: Likewise.
11572 * tsan.c (tsan_pass): Likewise.
11573 (tsan_gate): Likewise.
11574 (tsan_gate_O0): Likewise.
11575 * cfgexpand.c (partition_stack_vars): Likewise.
11576 (expand_stack_vars): Likewise.
11577 (defer_stack_allocation): Likewise.
11578 (expand_used_vars): Likewise.
11579 * cfgcleanup.c (old_insns_match_p): Likewise.
11580 * asan.c (asan_finish_file): Likewise.
11581 (asan_instrument): Likewise.
11582 (gate_asan): Likewise.
11583 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
11584 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11585 (asan_global_struct): Use pointer_sized_int_node instead
11586 calling build_nonstandard_integer_type.
11587 (initialize_sanitizer_builtins): Likewise.
11588 (asan_finish_file): Likewise.
11589 * gcc.c: Document %{%:function(args):X}.
11590 (static_spec_functions): Add sanitize.
11591 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
11592 store funcval != NULL there.
11593 (do_spec_1): Adjust handle_spec_function caller.
11594 (handle_braces): Allow %:function(args) as condition.
11595 (sanitize_spec_function): New function.
11596 (ADD_STATIC_LIBUBSAN_LIBS): Define.
11597 (LIBUBSAN_SPEC): Likewise.
11598 (LIBUBSAN_EARLY_SPEC): Likewise.
11599 (SANITIZER_SPEC): Handle libubsan.
11600 (SANITIZER_EARLY_SPEC): Likewise.
11601 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
11602 instead of fsanitize=address.
11603 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
11604 instead of fsanitize=address*.
11605 * builtins.c: Include ubsan.h.
11606 (fold_builtin_0): Instrument __builtin_unreachable.
11607 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
11608 instead of flag_asan.
11609 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
11610 (pointer_sized_int_node): Define.
11611 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
11613 2013-08-30 Mike Stump <mikestump@comcast.net>
11615 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
11618 2013-08-29 Jan Hubicka <jh@suse.cz>
11620 * cgraph.c (cgraph_function_body_availability): Handle weakref
11622 * passes.def: Remove pass_fixup_cfg.
11623 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
11624 track when we need to remove functions.
11625 (gate_ipa_inline): Execute inlining always; add comment why.
11626 (pass_data_ipa_inline): Remove TODO_remove_functions.
11627 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
11628 do not produce summaries.
11629 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
11630 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
11631 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
11632 constant pool and vtable.
11634 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
11636 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
11637 New arm_neon.h's internal macros to specify 64-bit constants.
11638 Avoid using stdint.h's macros.
11640 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
11642 * recog.c (verify_changes): Verify that changes[i].old is non-zero
11643 before applying REG_P.
11645 2013-08-30 Jakub Jelinek <jakub@redhat.com>
11647 PR tree-optimization/58277
11648 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
11649 after seeing too many stmts with vdef in between dombb and current
11650 bb, invalidate everything.
11652 2013-08-30 Richard Biener <rguenther@suse.de>
11654 * fold-const.c (fold_single_bit_test): Fix overflow test.
11656 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
11658 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
11659 and which can live in a register, always retrieve the value on entry.
11660 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
11661 passed by invisible reference specially.
11662 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
11663 (vt_add_function_parameter): Correctly deal with a parameter passed by
11664 invisible reference.
11666 2013-08-30 Jan Hubicka <jh@suse.cz>
11668 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
11670 2013-08-30 Richard Biener <rguenther@suse.de>
11672 PR tree-optimization/58228
11673 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
11674 allow invariant loads in nested loop vectorization.
11676 2013-08-30 Richard Biener <rguenther@suse.de>
11678 PR tree-optimization/58223
11679 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
11680 (has_anti_or_output_dependence): ... this and adjust to also
11681 look for output dependences.
11682 (mark_nodes_having_upstream_mem_writes): Adjust.
11683 (rdg_flag_uses): Likewise.
11685 2013-08-30 Richard Biener <rguenther@suse.de>
11687 PR tree-optimization/58010
11688 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
11689 assert that we have a loop-closed PHI.
11691 2013-08-29 Jan Hubicka <jh@suse.cz>
11693 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
11694 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
11695 * lto-section-in.c (lto_free_function_in_decl_state): New function.
11696 (lto_free_function_in_decl_state_for_node): New function.
11698 2013-08-29 Xinliang David Li <davidxl@google.com>
11700 * loop-unroll.c (report_unroll_peel): Minor message change.
11701 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
11702 Emit alignment peeling message with default -fopt-info.
11703 (vect_loop_versioning): Emit loop version info message.
11704 * tree-vectorizer.c (vectorize_loops): Minor message change.
11705 (execute_vect_slp): Ditto.
11707 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
11709 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
11710 of the clone from the DECL_NAME of the original function.
11712 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
11714 * passes.c (register_pass): Add overload.
11715 * tree-pass.h (register_pass): Forward declare it. Add comment.
11717 2013-08-29 Jan Hubicka <jh@suse.cz>
11719 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
11720 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
11721 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
11722 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11723 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
11724 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
11725 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11726 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
11728 2013-08-29 Teresa Johnson <tejohnson@google.com>
11730 * dumpfile.c (dump_loc): Output column number.
11731 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11732 * doc/invoke.texi: Document optall -fopt-info flag.
11733 * profile.c (read_profile_edge_counts): Use new dump framework.
11734 (compute_branch_probabilities): Ditto.
11735 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11736 when pass not in any opt group.
11737 * pass_manager.h (pass_manager::get_pass_profile): New method.
11738 * value-prof.c (check_counter): Use new dump framework.
11739 (check_ic_target): Ditto.
11740 * coverage.c (get_coverage_counts): Ditto.
11741 (coverage_init): Setup new dump framework.
11743 2013-08-29 Richard Biener <rguenther@suse.de>
11745 PR tree-optimization/58246
11746 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11747 handle the dominance check inside a basic-block.
11749 2013-08-29 Richard Biener <rguenther@suse.de>
11751 PR middle-end/57287
11752 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11753 of default defs that appear in abnormal PHI nodes.
11755 2013-08-29 Richard Biener <rguenther@suse.de>
11757 PR tree-optimization/57685
11758 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11759 single-use operands to avoid exponential complexity.
11761 2013-08-28 Dehao Chen <dehao@google.com>
11763 * ipa-inline.c (edge_badness): Fix integer underflow.
11765 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11767 * gtm-builtins.def (_ITM_free): Declare leaf.
11769 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11772 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11773 (*tls_local_dynamic_base_64_largepic): Likewise.
11774 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11775 Remove predicate from call operand.
11776 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11777 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11779 2013-08-28 Jeff Law <law@redhat.com>
11781 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11782 checks for the number of predecessors and successors allowed.
11783 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11784 which require copying a joiner block if there is a request which
11785 is a subpath that requires no joiner block copying.
11787 2013-08-28 Jan Hubicka <jh@suse.cz>
11789 * lto-streamer-out.c (DFS_write_tree_body): Drop
11790 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11791 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11792 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11793 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11794 Do not read DECL_ERROR_ISSUED.
11795 (unpack_ts_decl_with_vis_value_fields): Do not read
11797 (lto_input_ts_binfo_tree_pointers): Do not read
11798 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11799 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11800 write DECL_ERROR_ISSUED..
11801 (pack_ts_decl_with_vis_value_fields): Do not write
11803 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11804 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11805 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11806 * tree.h (tree_decl_common): Update comment.
11807 (DECL_ERROR_ISSUED): Remove.
11809 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11811 PR middle-end/58257
11812 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11814 2013-08-28 Jan Hubicka <jh@suse.cz>
11816 * builtins.def (free): Declare leaf.
11818 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11820 * gdbhooks.py: New.
11821 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11822 * configure: Regenerate.
11824 2013-08-27 Martin Jambor <mjambor@suse.cz>
11826 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11827 (ipa_ancestor_jf_data): Likewise.
11828 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11829 (ipa_get_jf_pass_through_type_preserved): New function.
11830 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11831 (ipa_get_jf_ancestor_type_preserved): New function.
11832 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11833 (ipa_get_jf_ancestor_result): Likewise.
11834 (propagate_vals_accross_pass_through): Use
11835 ipa_get_jf_pass_through_result to do all the value mappings.
11836 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11837 type_preserved flag.
11838 (ipa_set_jf_cst_copy): New function.
11839 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11840 (ipa_set_jf_arith_pass_through): Likewise.
11841 (ipa_set_ancestor_jf): Likewise.
11842 (compute_complex_assign_jump_func): Set type_preserved instead of
11844 (ipa_compute_jump_functions_for_edge): Likewise.
11845 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11846 (update_jump_functions_after_inlining): Update type_preserved.
11847 Explicitely create jump functions when combining one with pass_through.
11848 (ipa_write_jump_function): Stream the type_preserved flags.
11849 (ipa_read_jump_function): Likewise.
11851 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11852 Aldy Hernandez <aldyh@redhat.com>
11854 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11855 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11856 * function.h (struct function): Add has_force_vect_loops and
11858 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11859 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11861 * gimple.def: Update CLAUSES comments.
11862 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11863 (gimple_build_omp_for): Add argument to prototype.
11864 (gimple_omp_for_kind): New.
11865 (gimple_omp_for_set_kind): New.
11866 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11867 GOVD_DATA_SHARE_CLASS.
11868 (enum omp_region_type): Add ORT_SIMD.
11869 (gimple_add_tmp_var): Handle ORT_SIMD.
11870 (gimplify_var_or_parm_decl): Same.
11871 (is_gimple_stmt): Same.
11872 (omp_firstprivatize_variable): Same.
11873 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11874 (omp_notice_variable): Handle ORT_SIMD.
11875 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11876 (omp_check_private): Handle ORT_SIMD.
11877 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11878 OMP_CLAUSE_SAFELEN.
11879 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11880 Handle OMP_CLAUSE_LASTPRIVATE.
11881 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11882 OMP_CLAUSE_SAFELEN.
11883 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11884 (gimplify_expr): Handle OMP_SIMD.
11885 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11886 (expand_GOMP_SIMD_VF): New.
11887 (expand_GOMP_SIMD_LAST_LANE): New.
11888 * internal-fn.def (GOMP_SIMD_LANE): New.
11889 (GOMP_SIMD_VF): New.
11890 (GOMP_SIMD_LAST_LANE): New.
11891 * omp-low.c: Include target.h.
11892 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11893 OMP_CLAUSE_SAFELEN.
11894 (check_omp_nesting_restrictions): Same.
11896 (lower_rec_simd_input_clauses): New.
11897 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11899 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11900 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11901 (expand_omp_build_assign): New.
11902 (expand_omp_for_init_counts): New.
11903 (expand_omp_for_init_vars): New.
11904 (extract_omp_for_update_vars): New.
11905 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11906 and rewrite accordingly.
11907 (expand_omp_simd): New.
11908 (expand_omp_for): Use expand_omp_simd.
11909 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11910 (lower_omp_for): Do not lower the body.
11911 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11912 in their own loops.
11913 * tree-flow.h (find_omp_clause): Remove prototype.
11914 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11915 forcing vectorization of the loop, or if flag_tree_vectorize.
11916 (gate_tree_if_conversion): Similarly.
11917 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11918 gimple_build_omp_for.
11919 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11920 * tree-parloops (create_parallel_loop): Pass kind argument to
11921 gimple_build_omp_for.
11922 * tree-pretty-print.c (dump_omp_clause): Add cases for
11923 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11924 OMP_CLAUSE__SIMDUID_.
11925 (dump_generic_node): Handle OMP_SIMD.
11926 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11927 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11928 unroll OMP_SIMD loops here.
11929 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11930 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11932 (vect_analyze_data_refs): Handle simd loops.
11933 * tree-vect-loop.c (vectorizable_live_operation): Handle
11935 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11936 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11937 (vectorizable_load): Same.
11938 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11939 (struct simduid_to_vf): New.
11940 (simduid_to_vf::hash): New.
11941 (simduid_to-vf::equal): New.
11942 (struct simd_array_to_simduid): New.
11943 (simd_array_to_simduid::hash): New.
11944 (simd_array_to_simduid::equal): New.
11945 (adjust_simduid_builtins): New.
11946 (struct note_simd_array_uses_struct): New.
11947 (note_simd_array_uses_cb): New.
11948 (note_simd_array_uses): New.
11949 (vectorize_loops): Handle simd hints and adjust simd builtins
11951 * tree-vectorizer.h (struct _stmt_vec_info): Add
11952 simd_lane_access_p field.
11953 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11954 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11955 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11956 (omp_clause_code_name): Same.
11957 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11958 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11959 * tree.def (OMP_SIMD): New entry.
11960 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11961 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11962 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11963 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11964 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11965 (OMP_CLAUSE_LINEAR_STEP): New.
11966 (OMP_CLAUSE_SAFELEN_EXPR): New.
11967 (OMP_CLAUSE__SIMDUID__DECL): New.
11968 (find_omp_clause): New prototype.
11970 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11972 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11973 Haswell processor detection.
11975 2013-08-27 Christian Widmer <shadow@umbrox.de>
11978 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11979 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11980 AVX2 capable processors.
11982 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11984 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11985 of vget_low_* with implementations in terms of other intrinsics.
11987 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11989 PR middle-end/57219
11990 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11991 values to -1, 0 and 1.
11993 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11995 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11996 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11997 (<optab><mode>3_insn): Remove.
11998 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11999 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
12000 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
12001 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
12002 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
12003 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
12004 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
12005 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
12006 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
12007 (ror<mode>3_insn): Likewise.
12008 * config/aarch64/predicates.md (aarch64_simd_register): New.
12010 2013-08-27 Richard Biener <rguenther@suse.de>
12012 PR tree-optimization/57521
12013 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
12014 one edge is non-critical.
12015 (find_phi_replacement_condition): Make sure to use a non-critical
12016 edge. Cleanup and remove old bug workarounds.
12017 (bb_postdominates_preds): Remove.
12018 (if_convertible_loop_p_1): Do not compute post-dominators.
12019 (combine_blocks): Do not free post-dominators.
12020 (main_tree_if_conversion): Likewise.
12021 (pass_data_if_conversion): Add TODO_verify_ssa.
12023 2013-08-27 DJ Delorie <dj@redhat.com>
12025 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
12027 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
12029 * function.c (assign_parm_find_data_types): Set passed_mode and
12030 nominal_mode to the TYPE_MODE of nominal_type for the built
12031 pointer type in case of the struct-pass-by-reference.
12033 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
12035 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
12036 (UINT16_TYPE): Change default to "unsigned int".
12038 * config/avr/avr.opt (mfract-convert-truncate): New option.
12039 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
12040 is set, round negative fractional integers according to n1169
12041 when converting to integer types.
12043 2013-08-26 Jan Hubicka <jh@suse.cz>
12045 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
12046 methods can not be called indirectly when their address is not taken.
12048 2013-08-26 Jan Hubicka <jh@suse.cz>
12050 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
12053 2013-08-26 Jan Hubicka <jh@suse.cz>
12055 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
12058 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
12060 * reload.c (find_valid_class): Allow classes that do not include
12061 FIRST_PSEUDO_REGISTER - 1.
12063 2013-08-26 Jan Hubicka <jh@suse.cz>
12065 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
12066 fix edge count/frequency when speculation failed; fix type check
12067 for the direct call.
12069 2013-08-26 Jan Hubicka <jh@suse.cz>
12071 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
12073 2013-08-26 Jan Hubicka <jh@suse.cz>
12075 * ipa-inline-transform.c (inline_transform): Be ready for basic block
12076 to be changed by edge redirection.
12078 2013-08-26 Jan Hubicka <jh@suse.cz>
12080 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
12081 formating; add sanity check.
12082 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
12083 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
12084 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
12085 (resolve_noninline_speculation): Update callee keys, too.
12087 2013-08-26 Jan Hubicka <jh@suse.cz>
12089 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
12090 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
12092 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
12094 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
12097 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
12099 * config/i386/i386.c (ix86_debug_options): Remove prototype.
12100 (x86_64_elf_select_section): Ditto.
12101 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
12103 (ix86_pass_by_reference): Ditto.
12104 (output_set_got): Ditto.
12105 (ix86_unary_operator_ok): Ditto.
12106 (ix86_expand_builtin): Ditto.
12108 2013-08-23 Jan Hubicka <jh@suse.cz>
12110 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
12112 2013-08-23 Jan Hubicka <jh@suse.cz>
12114 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
12115 (tree_decl_with_vis): Add FINAL field.
12117 2013-08-23 Jeff Law <law@redhat.com>
12119 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
12120 the debugging dump when the expression is fully redundant.
12122 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
12124 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
12125 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
12126 * pretty-print.c (pp_formatted_text_data): Likewise.
12127 (pp_write_text_to_stream): Likewise.
12128 (pp_write_text_as_dot_label_to_stream): Likewise.
12129 (pp_append_r): Likewise.
12130 (pp_format): Likewise.
12131 (pp_flush): Likewise.
12132 (pp_clear_output_area): Likewise.
12133 (pp_append_text): Likewise.
12134 (pp_formatted_text): Likewise.
12135 (pp_remaining_character_count_for_line): Likewise.
12136 (pp_newline): Likewise.
12137 (pp_character): Likewise.
12138 (output_buffer::~output_buffer): Define.
12139 (pretty_printer::~pretty_printer): Destruct output buffer.
12140 * pretty-print.h (output_buffer::~output_buffer): Declare.
12141 (pretty_printer::~pretty_printer): Declare virtual.
12143 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
12146 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
12147 HOST_WIDE_INT_M1U): New macros.
12148 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
12149 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
12150 unsigned -1 for lshift.
12151 * cse.c (cse_insn): Likewise.
12152 * double-int.c (rshift_double, lshift_double): Likewise.
12153 * builtins.c (fold_builtin_bitop): Likewise.
12154 * combine.c (force_to_mode): Likewise.
12155 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
12156 * simplify-rtx.c (simplify_const_unary_operation,
12157 simplify_const_binary_operation): Likewise.
12158 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
12159 check_va_list_escapes): Likewise.
12160 * rtlanal.c (nonzero_bits1): Likewise.
12161 * expmed.c (expand_smod_pow2): Likewise.
12162 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
12164 2013-08-23 Jan Hubicka <jh@suse.cz>
12166 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
12167 as having address taken.
12169 2013-08-23 Jan Hubicka <jh@suse.cz>
12171 * ipa-utils.h (method_class_type): Declare.
12172 * ipa-devirt.c (method_class_type): Export.
12174 * cgraphunit.c (analyze_functions): Do basic devirtualization;
12175 do not walk base classes of anonymous types.
12177 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
12179 PR rtl-optimization/58220
12180 PR regression/58221
12181 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
12182 handle SEQUENCE insns properly.
12184 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
12186 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
12188 (pp_newline_and_indent): Likewise.
12189 (pp_separate_with): Likewise.
12190 * pretty-print.c (pp_newline_and_flush): Define.
12191 (pp_newline_and_indent): Likewise.
12192 (pp_separate_with): Likewise.
12194 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12197 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
12198 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
12200 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
12202 * config/i386/predicates.md (ext_sse_reg_operand): New.
12203 * config/i386/i386.md (*movti_internal): Use
12204 predicate to determine if EVEX is needed.
12205 (*movsi_internal): Ditto.
12206 (*movdf_internal): Ditto.
12207 (*movsf_internal): Ditto.
12208 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
12210 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12212 PR tree-optimization/58209
12213 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
12214 (find_tail_calls): Give up for pointer result types if m is non-NULL.
12215 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
12216 emit POINTER_PLUS_EXPR.
12217 (create_tailcall_accumulator): For pointer result type accumulate in
12220 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
12222 * configure.ac: Add backslashes missing from the last change.
12223 * configure: Regenerate.
12225 2013-08-22 Jan Hubicka <jh@susue.cz>
12227 * ipa.c (function_and_variable_visibility): First remember function
12228 was global and then make it local.
12230 2013-08-22 Julian Brown <julian@codesourcery.com>
12232 * configure.ac: Add aarch64 to list of arches which use "nop" in
12234 * configure: Regenerate.
12236 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12238 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
12239 gnu_libc_has_function.
12240 * config/s390/tpf.h: Likewise.
12242 2013-08-22 Jan Hubicka <jh@susue.cz>
12244 * timevar.c (validate_phases): Add cast.
12246 2013-08-22 Jan Hubicka <jh@susue.cz>
12248 * timevar.c (validate_phases): Use size_t for memory.
12249 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
12251 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
12253 * pretty-print.h (output_buffer::output_buffer): Declare.
12254 (pretty_printer::pretty_printer): Likewise.
12255 (pp_construct): Remove.
12256 * pretty-print.c (output_buffer::output_buffer): Define.
12257 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
12258 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
12259 (print_gimple_expr): Likewise.
12260 (print_gimple_seq): Likewise.
12261 (gimple_dump_bb): Likewise.
12262 * sched-vis.c (dump_value_slim): Likewise.
12263 (dump_insn_slim): Likewise.
12264 (dump_rtl_slim): Likewise.
12265 (str_pattern_slim): Likewise.
12266 * tree-mudflap.c (mf_varname_tree): Likewise.
12267 * graph.c (print_graph_cfg): Likewise.
12268 (start_graph_dump): Likewise.
12269 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
12271 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
12272 pretty printer initialization.
12273 * coretypes.h (diagnostic_context): Remove superflous type alias
12275 (pretty_printer): Likewise. Declare directly as a class.
12276 (pretty_print_info): Remove declaration as class.
12277 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
12278 and pp_clear_output_area.
12279 (asan_add_global): Likewise.
12281 2013-08-22 Jan Hubicka <jh@suse.cz>
12283 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
12284 * ipa-utils.h (update_type_inheritance_graph): Declare.
12285 (possible_polymorphic_call_target_p): Declare.
12286 (possible_polymorphic_call_target_p): New.
12287 * ipa-devirt.c: Update toplevel comments.
12288 (cached_polymorphic_call_targets): Move up.
12289 (odr_type_d): Move ID down.
12290 (polymorphic_type_binfo_p): Update comment.
12291 (odr_hasher::remove): Likewise;
12292 (get_odr_type): Set anonymous_namespace.
12293 (dump_odr_type): Dump it.
12294 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
12295 (maybe_record_node): Record node in cached_polymorphic_call_targets.
12296 (record_binfo): Add comment.
12297 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
12299 (devirt_node_removal_hook): Do not iCE when cache is freed.
12300 (possible_polymorphic_call_target_p): New predicate.
12301 (update_type_inheritance_graph): New function.
12303 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
12304 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
12305 Sergey Lega <sergey.s.lega@intel.com>
12306 Anna Tikhonova <anna.tikhonova@intel.com>
12307 Ilya Tocar <ilya.tocar@intel.com>
12308 Andrey Turetskiy <andrey.turetskiy@intel.com>
12309 Ilya Verbin <ilya.verbin@intel.com>
12310 Kirill Yukhin <kirill.yukhin@intel.com>
12311 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12313 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
12314 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
12315 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
12316 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
12317 (OPTION_MASK_ISA_AVX2_UNSET): Update.
12318 (OPTION_MASK_ISA_AVX512F_UNSET): New.
12319 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
12320 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
12321 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
12322 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
12323 OPT_mavx512pf, OPT_mavx512er cases.
12324 * config/i386/constraints.md (v): New constraint.
12325 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
12326 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
12327 (bit_AVX512CD): New.
12328 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12329 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
12330 * config/i386/i386-c.c (ix86_target_macros_internal):
12331 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
12333 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
12334 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
12335 * config/i386/i386.c (regclass_map, dbx_register_map)
12336 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
12337 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
12338 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
12339 -mavx512pf options.
12340 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
12341 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
12342 -mavx512pf options. Fix formatting.
12343 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
12344 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
12345 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
12346 -mavx512cd, -mavx512pf options.
12347 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
12348 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
12349 (ix86_preferred_output_reload_class): Replace SSE_REGS with
12351 (ix86_hard_regno_mode_ok): Support 512-bit registers.
12352 (ix86_set_reg_reg_cost): Ditto.
12353 (x86_order_regs_for_local_alloc): Ditto.
12354 (MAX_VECT_LEN): Extend to 64-byte.
12355 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
12356 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
12357 (TARGET_AVX512ER, TARGET_AVX512CD): New.
12358 (BIGGEST_ALIGNMENT): Extend to 512-bits.
12359 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
12360 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
12361 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
12362 (SSE_REG_MODE_P): Support new modes.
12363 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
12364 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
12365 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
12366 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
12367 (REG_CLASS_CONTENTS): Add new registers.
12368 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
12369 (EXT_REX_SSE_REGNO_P): New.
12370 (HI_REGISTER_NAMES): Add new registers.
12371 * config/i386/i386.md: Define constants for new registers.
12372 (mode): Add new 512-bit modes.
12373 (prefix): Support evex prefix.
12374 (isa): Support avx512f, noavx512f, fma_avx512f.
12375 (ssemodesuffix): Add new 512-bit modes.
12377 (*movxi_internal_avx512f): Ditto.
12378 (*movdi_internal): Replace constraint "x" with the new constraint "v".
12380 (*movsi_internal): Likewise.
12381 (*movdf_internal): Likewise.
12382 (*movsf_internal): Likewise.
12383 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
12384 (<code><mode>3): Likewise.
12385 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
12387 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
12388 with the new constraint "v".
12389 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
12391 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
12392 with the new constraint "v".
12393 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
12394 (<sse2>_storedqu<avxsizesuffix>): Likewise.
12395 (*<plusminus_insn><mode>3): Likewise.
12396 (<sse>_vm<plusminus_insn><mode>3): Likewise.
12397 (*mul<mode>3): Likewise.
12398 (<sse>_vmmul<mode>3): Likewise.
12399 (<sse>_div<mode>3): Likewise.
12400 (<sse>_vmdiv<mode>3): Likewise.
12401 (<sse>_sqrt<mode>2): Likewise.
12402 (<sse>_vmsqrt<mode>2): Likewise.
12403 (*<code><mode>3_finite): Likewise.
12404 (*<code><mode>3) <smaxmin>: Likewise.
12405 (<sse>_vm<code><mode>3): Likewise.
12406 (*<code><mode>3) <any_logic>: Likewise.
12407 (*fma_fmadd_<mode>): Likewise.
12408 (*fma_fmsub_<mode>): Likewise.
12409 (*fma_fnmadd_<mode>): Likewise.
12410 (*fma_fnmsub_<mode>): Likewise.
12411 (*fma_fmaddsub_<mode>): Likewise.
12412 (*fma_fmsubadd_<mode>): Likewise.
12413 (*fmai_fmadd_<mode>): Likewise.
12414 (*fmai_fmsub_<mode>): Likewise.
12415 (*fmai_fnmadd_<mode>): Likewise.
12416 (*fmai_fnmsub_<mode>): Likewise.
12417 (sse_cvtsi2ss): Likewise.
12418 (sse_cvtsi2ssq): Likewise.
12419 (sse_cvtss2si): Likewise.
12420 (sse_cvtss2si_2): Likewise.
12421 (sse_cvtss2siq): Likewise.
12422 (sse_cvtss2siq_2): Likewise.
12423 (sse_cvttss2si): Likewise.
12424 (sse_cvtss2siq_2): Likewise.
12425 (float<sseintvecmodelower><mode>2): Likewise.
12426 (sse2_cvtsd2si_2): Likewise.
12427 (sse2_cvtsd2siq_2): Likewise.
12428 (*<plusminus_insn><mode>3): Likewise.
12429 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
12430 (*<sse4_1_avx2>_mul<mode>3): Likewise.
12431 (ashr<mode>3): Likewise.
12432 (<shift_insn><mode>3): Likewise.
12433 (avx2_<code><mode>3): Likewise.
12434 (*avx2_<code><mode>3): Likewise.
12435 (*andnot<mode>3): Likewise.
12436 (*<code><mode>3) <any_logic>: Likewise.
12437 (abs<mode>2): Likewise.
12438 (avx2_permvar<mode>): Likewise.
12439 (avx2_perm<mode>_1): Likewise.
12440 (*avx_vpermilp<mode>): Likewise.
12441 (avx_vpermilvar<mode>3): Likewise.
12442 (avx2_ashrv<mode>): Likewise.
12443 (avx2_<shift_insn>v<mode>): Likewise.
12444 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
12446 * doc/rtl.texi: Document XImode.
12448 2013-08-21 Jeff Law <law@redhat.com>
12450 * tree-flow.h (register_jump_thread): Pass vector of edges
12451 instead of each important edge.
12452 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
12453 thread path into a vector and pass that to register_jump_thread.
12454 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
12455 passed in edge vector to the current 3-edge form.
12458 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12460 * dce.c (fini_dce): Call df_analyze again just in case
12461 delete_unmarked_insns removed anything.
12463 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
12465 * reload.h (struct reg_equivs): Rename to ..
12466 (struct reg_equivs_s): .. this.
12468 2013-08-20 Martin Liska <marxin.liska@gmail.com>
12470 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
12472 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12474 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
12476 2013-08-21 Jeff Law <law@redhat.com>
12478 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
12479 simplify assignments too. If the RHS collapses to a singleton
12480 range, then return the value for the range.
12482 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
12484 * config/i386/sse.md (V16): Rename to...
12486 (mov<mode>): Update iterator name.
12487 (*mov<mode>_internal): Ditto.
12488 (push<mode>1): Ditto.
12489 (movmisalign<mode>): Ditto.
12491 2013-08-20 Jan Hubicka <jh@suse.cz>
12494 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
12495 entry for direct edges.
12496 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
12498 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12500 Revert my last two changes, r201865 and r201864:
12503 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12505 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12506 instances can own GC refs.
12508 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12509 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12510 (gcc::context::gt_pch_nx): Likewise.
12511 (gcc::context::gt_pch_nx): Likewise.
12512 * ggc.h (gt_ggc_mx <T>): New.
12513 (gt_pch_nx_with_op <T>): New.
12514 (gt_pch_nx <T>): New.
12515 * passes.c (opt_pass::gt_ggc_mx): New.
12516 (opt_pass::gt_pch_nx): New.
12517 (opt_pass::gt_pch_nx_with_op): New.
12518 (pass_manager::gt_ggc_mx): New.
12519 (pass_manager::gt_pch_nx): New.
12520 (pass_manager::gt_pch_nx_with_op): New.
12521 (pass_manager::operator new): Use
12522 ggc_internal_cleared_alloc_stat rather than xcalloc.
12523 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12524 (pass_manager::gt_ggc_mx): New.
12525 (pass_manager::gt_pch_nx): New.
12526 (pass_manager::gt_pch_nx_with_op): New.
12527 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12528 (opt_pass::operator new): New.
12529 (opt_pass::gt_ggc_mx): New.
12530 (opt_pass::gt_pch_nx): New.
12531 (opt_pass::gt_pch_nx_with_op): New.
12534 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12536 * Makefile.in (GTFILES): Add context.h.
12537 * context.c (gcc::context::operator new): New.
12538 (gcc::context::gt_ggc_mx): New.
12539 (gcc::context::gt_pch_nx): New.
12540 (gcc::context::gt_pch_nx): New.
12541 * context.h (gcc::context): Add GTY((user)) marking.
12542 (gcc::context::operator new): New.
12543 (gcc::context::gt_ggc_mx): New.
12544 (gcc::context::gt_pch_nx): New.
12545 (gcc::context::gt_pch_nx): New.
12546 (g): Add GTY marking.
12547 (gt_ggc_mx (gcc::context *)): New.
12548 (gt_pch_nx (gcc::context *)): New.
12549 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12550 void *cookie)): New.
12551 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12553 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12555 * dce.c (fini_dce): Call df_analyze again just in case
12556 delete_unmarked_insns removed anything.
12558 2013-08-20 Teresa Johnson <tejohnson@google.com>
12560 PR rtl-optimizations/57451
12561 * final.c (reemit_insn_block_notes): Prevent lexical blocks
12562 from crossing split section boundaries.
12564 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12566 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
12567 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
12568 with MULTLIB_DEFAULTS.
12570 2013-08-20 Nick Clifton <nickc@redhat.com>
12572 * target.def (narrow_volatile_bitfield): Note that the default
12573 value is false, not !TARGET_STRICT_ALIGN.
12574 * doc/tm.texi: Regenerate.
12576 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
12578 Fix LIB_SPEC for systems without libpthread.
12580 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
12581 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
12583 * config/i386/linux-common.h: Likewise.
12584 * config/mips/linux-common.h: Likewise.
12586 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12588 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
12591 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12593 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12594 instances can own GC refs.
12596 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12597 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12598 (gcc::context::gt_pch_nx): Likewise.
12599 (gcc::context::gt_pch_nx): Likewise.
12600 * ggc.h (gt_ggc_mx <T>): New.
12601 (gt_pch_nx_with_op <T>): New.
12602 (gt_pch_nx <T>): New.
12603 * passes.c (opt_pass::gt_ggc_mx): New.
12604 (opt_pass::gt_pch_nx): New.
12605 (opt_pass::gt_pch_nx_with_op): New.
12606 (pass_manager::gt_ggc_mx): New.
12607 (pass_manager::gt_pch_nx): New.
12608 (pass_manager::gt_pch_nx_with_op): New.
12609 (pass_manager::operator new): Use
12610 ggc_internal_cleared_alloc_stat rather than xcalloc.
12611 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12612 (pass_manager::gt_ggc_mx): New.
12613 (pass_manager::gt_pch_nx): New.
12614 (pass_manager::gt_pch_nx_with_op): New.
12615 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12616 (opt_pass::operator new): New.
12617 (opt_pass::gt_ggc_mx): New.
12618 (opt_pass::gt_pch_nx): New.
12619 (opt_pass::gt_pch_nx_with_op): New.
12621 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12623 * Makefile.in (GTFILES): Add context.h.
12624 * context.c (gcc::context::operator new): New.
12625 (gcc::context::gt_ggc_mx): New.
12626 (gcc::context::gt_pch_nx): New.
12627 (gcc::context::gt_pch_nx): New.
12628 * context.h (gcc::context): Add GTY((user)) marking.
12629 (gcc::context::operator new): New.
12630 (gcc::context::gt_ggc_mx): New.
12631 (gcc::context::gt_pch_nx): New.
12632 (gcc::context::gt_pch_nx): New.
12633 (g): Add GTY marking.
12634 (gt_ggc_mx (gcc::context *)): New.
12635 (gt_pch_nx (gcc::context *)): New.
12636 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12637 void *cookie)): New.
12638 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12640 2013-08-20 Alan Modra <amodra@gmail.com>
12643 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
12644 (rs6000_emit_epilogue): Likewise.
12646 2013-08-19 Dehao Chen <dehao@google.com>
12648 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
12650 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
12651 Jakub Jelinek <jakub@redhat.com>
12653 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
12654 (BUILT_IN_FABSD64): Likewise.
12655 (BUILT_IN_FABSD128): Likewise.
12656 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
12657 (fold_builtin_1): Likewise.
12658 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
12659 destination and source operands.
12660 (*abstd2_fpr): Likewise.
12661 (*nabstd2_fpr): Likewise.
12663 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
12665 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
12668 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
12670 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
12673 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
12675 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
12676 * builtins.c (default_libc_has_function): New.
12677 (gnu_libc_has_function): Ditto.
12678 (no_c99_libc_has_function): Ditto.
12679 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
12680 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
12681 (fold_builtin_sincos): Likewise.
12682 (fold_builtin_cexp): Likewise.
12683 * builtins.def (DEF_C94_BUILTIN): Likewise.
12684 (DEF_C99_BUILTIN): Likewise.
12685 (DEF_C99_C90RES_BUILTIN): Likewise.
12686 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
12687 definitions to using this define.
12688 * config/darwin-protos.h (darwin_libc_has_function): New.
12689 * config/darwin.c (darwin_libc_has_function): Ditto.
12690 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
12691 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
12692 * config/darwin.h: Ditto.
12693 * config/elfos.h: Ditto.
12694 * config/freebsd.h: Ditto.
12695 * config/i386/cygming.h: Ditto.
12696 * config/i386/djgpp.h: Ditto.
12697 * config/i386/i386-interix.h: Ditto.
12698 * config/microblaze/microblaze.h: Ditto.
12699 * config/mmix/mmix.h: Ditto.
12700 * config/gnu-user.h: Ditto.
12701 * config/ia64/hpux.h: Ditto.
12702 * config/pa/pa-hpux.h: Ditto.
12703 * config/pdp11/pdp11.h: Ditto.
12704 * config/picochip/picochip.h: Ditto.
12705 * config/linux.h: Ditto.
12706 * config/netbsd.h: Ditto.
12707 * config/openbsd.h: Ditto.
12708 * config/rs6000/aix43.h: Ditto.
12709 * config/rs6000/aix51.h: Ditto.
12710 * config/rs6000/aix52.h: Ditto.
12711 * config/rs6000/aix53.h: Ditto.
12712 * config/rs6000/aix61.h: Ditto.
12713 * config/rs6000/darwin.h: Ditto.
12714 * config/rs6000/linux.h: Ditto.
12715 * config/rs6000/linux64.h: Ditto.
12716 * config/s390/tpf.h: Ditto.
12717 * config/sol2-10.h: Ditto.
12718 * config/sol2.h: Ditto.
12719 * config/vms/vms.h: Ditto.
12720 * config/vxworks.h: Ditto.
12721 * config/linux-android.c (linux_android_libc_has_function):
12722 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
12723 * config/linux-protos.h (linux_android_libc_has_function):
12725 * config/i386/i386.c (ix86_libc_has_function): New.
12726 * config/i386/i386-protos.h
12727 (ix86_libc_has_function): New declaration.
12728 * config/i386/i386.md
12729 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12730 ("isinf<mode>2): Likewise.
12731 * convert.c (convert_to_integer): Using new target hook
12732 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12733 TARGET_C99_FUNCTIONS.
12734 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12735 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12736 * coretypes.h (function_class): New enum for different
12737 classes of functions.
12738 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12739 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12740 (TARGET_HAS_SINCOS): Likewise.
12741 (TARGET_LIBC_HAS_FUNCTION): New.
12742 * doc/tm.texi: Regenerated.
12743 * targhooks.h (default_libc_has_function): New declaration.
12744 (no_c99_libc_has_function): Ditto.
12745 (gnu_libc_has_function): Ditto.
12746 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12747 and TARGET_HAS_SINCOS.
12749 2013-08-18 Jan Hubicka <jh@suse.cz>
12751 * Makeifle-in (ipa-devirt.o): New.
12752 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12753 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12754 (analyze_functions): Look into possible targets of polymorphic call.
12755 * dumpfile.c (dump_files): Add type-inheritance dump.
12756 * dumpfile.h (TDI_inheritance): New.
12757 * ipa-devirt.c: New file.
12758 * ipa-utils.h (odr_type_d): Forward declare.
12759 (odr_type): New type.
12760 (build_type_inheritance_graph): Declare.
12761 (possible_polymorphic_call_targets): Declare and introduce inline
12762 variant when only edge is pased.
12763 (dump_possible_polymorphic_call_targets): Likewise.
12764 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12765 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12766 (types_same_for_odr): ... here.
12767 * tree.h (type_in_anonymous_namespace_p): Declare.
12769 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12771 PR tree-optimization/58006
12772 * tree-parloops.c (take_address_of): Don't ICE if get_name
12774 (eliminate_local_variables_stmt): Remove clobber stmts.
12776 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12778 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12779 error message is issued for an alias to undefined symbol.
12781 2013-08-18 Jan Hubicka <jh@suse.cz>
12783 * cgraph.c (cgraph_create_indirect_edge): Discover
12784 polymorphic calls and record basic info into indirect_info.
12785 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12786 devirtualization, ignore objc function calls.
12787 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12788 call with no parm index info.
12789 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12790 * tree.c (virtual_method_call_p): New function.
12791 * tree.h (virtual_method_call_p): Declare.
12793 2013-08-16 Jan Hubicka <jh@suse.cz>
12795 PR middle-end/58179
12796 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12798 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12800 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12803 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12805 * gengtype.c (type_for_name): Add special-case support for
12806 locating types within the "gcc::" namespace.
12807 (open_base_files): Emit a "using namespace gcc" directive.
12809 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12812 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12813 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12815 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12816 array instead of each individual operand as a separate argument.
12817 (emit_fusion_gpr_load): Likewise.
12818 (expand_fusion_gpr_load): Add new function declaration.
12820 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12821 signature to have the operands passed as an array, instead of as
12822 separate arguments. Allow ZERO_EXTEND to be in the memory
12823 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12824 depend on the register live/dead flags when peepholes are run.
12825 (expand_fusion_gpr_load): New function to be called from the
12826 peephole2 pass, to change the register that addis sets to be the
12828 (emit_fusion_gpr_load): Change the calling signature to have the
12829 operands passed as an array, instead of as separate arguments.
12830 Allow ZERO_EXTEND to be in the memory address, and also
12831 SIGN_EXTEND if -mpower8-fusion-sign.
12833 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12834 unspec enumeration.
12835 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12836 adjust the register addis loads up in the peephole2 pass. Do not
12837 depend on the register live/dead state when the peephole pass is done.
12839 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12841 * gengtype.c (create_user_defined_type): Ensure that the kind
12842 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12843 declaration is seen before the GTY((user)) marking.
12845 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12848 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12850 2013-08-16 Jan Hubicka <jh@suse.cz>
12852 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12853 arugment expected_type.
12854 (gimple_fold_call): Use it.
12855 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12856 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12857 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12858 (try_make_edge_direct_virtual_call): Likewise.
12859 * tree.c (obj_type_ref_class): New.
12860 * tree.h (obj_type_ref_class): Use it.
12862 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12864 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12865 (rtl_slim_pp): Likewise.
12866 (init_rtl_slim_pretty_print): Likewise.
12867 (dump_value_slim): Don't call it. Use local pretty printer.
12868 (dump_insn_slim): Likewise.
12869 (dump_rtl_slim): Likewise.
12870 (str_pattern_slim): Likewise.
12871 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12874 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12876 PR tree-optimization/58164
12877 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12878 walk gimple_goto_dest of GIMPLE_GOTO.
12880 PR tree-optimization/58165
12881 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12882 bi_call must be the last stmt in a bb, don't split_block, instead
12883 use fallthru edge from it and give up if there is none.
12884 Release conds vector when returning early.
12886 2013-08-14 Xinliang David Li <davidxl@google.com>
12888 * config/i386/i386.c (ix86_option_override_internal):
12889 Remove unused variable and field.
12891 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12894 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12895 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12896 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12897 and Linux, correct BLKmode alignment when 128-bit alignment is
12898 required and compatibility flag is not set.
12899 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12900 for zero-size arguments when compatibility flag is not set.
12902 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12904 PR tree-optimization/58145
12905 * tree-sra.c (build_ref_for_offset): If prev_base has
12906 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12908 2013-08-14 Xinliang David Li <davidxl@google.com>
12910 * config/i386/i386.c (ix86_option_override_internal):
12911 Fix uninitialized variable error.
12913 2013-08-14 Xinliang David Li <davidxl@google.com>
12915 * config/i386/i386.opt: Define two new options.
12916 * config/i386/x86-tune.def: Add arch selector field in macros.
12917 * config/i386/i386.h: Adjust macro definition.
12918 * config/i386/i386.c (ix86_option_override_internal):
12920 (parse_mtune_ctrl_str): New function.
12921 (set_ix86_tune_features): New function.
12922 (ix86_function_specific_restore): Call the new helper function.
12924 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12926 PR rtl-optimization/57662
12927 * sel-sched.c (code_motion_process_successors): When the current insn
12928 is removed after the recursive traversal, break from the loop.
12929 Add comments and debug printouts.
12931 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12932 Alexandre Oliva <aoliva@redhat.com>
12935 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12936 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12937 there also UNSPEC_PLTOFF.
12939 2013-08-14 Marek Polacek <polacek@redhat.com>
12941 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12942 NUM_CONDITIONS bit positions.
12944 2013-08-13 Cary Coutant <ccoutant@google.com>
12946 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12947 (attr_checksum): Hash vector contents instead of pointer.
12948 (attr_checksum_ordered): Likewise.
12950 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12952 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12953 when Pmode != word_mode. Add length_address attribute.
12954 (sse3_monitor_<mode>): Merge from sse3_monitor and
12955 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12956 Pmode != word_mode. Update insn length attribute.
12957 * config/i386/i386.c (ix86_option_override_internal): Update
12958 ix86_gen_monitor selection for merged sse3_monitor insn.
12960 2013-08-13 Julian Brown <julian@codesourcery.com>
12962 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12963 perform invalid legitimization on greater-than-word-size modes for
12964 TARGET_E500_DOUBLE.
12966 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12968 * ira.c (setup_class_translate_array): Use aclass instead of cl
12969 for classes not fully covered by allocno classes.
12971 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12973 PR tree-optimization/57661
12974 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12975 * tree-inline.c (tree_function_versioning): Initialize it.
12976 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12977 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12978 that is not being copied.
12981 * asan.c (instrument_strlen_call): Fix typo in comment.
12982 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12984 2013-08-13 Steve Ellcey <sellcey@mips.com>
12986 * config/mips/mips.md (prefetch): Use lw instead of ld on
12987 loongson in 32bit mode.
12989 2013-08-13 Nick Clifton <nickc@redhat.com>
12991 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12993 2013-08-13 Jan Hubicka <jh@suse.cz>
12995 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12996 introduced edge; fix typo in sanity check.
12997 (cgraph_resolve_speculation): Export; improve diagnostic.
12998 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12999 speculation at type mismatch.
13000 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
13001 (cgraph_resolve_speculation): Declare.
13002 (symtab_can_be_discarded): New function.
13003 * value-prof.c (gimple_ic_transform): Remove actual transform code.
13004 * ipa-inline-transform.c (speculation_removed): New global var.
13005 (clone_inlined_nodes): See if speculation can be removed.
13006 (inline_call): If speculations was removed, we growths may not match.
13007 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
13008 (speculation_useful_p): New function.
13009 (resolve_noninline_speculation): New function.
13010 (inline_small_functions): Resolve useless speculations.
13011 * ipa-inline.h (speculation_useful_p): Declare
13012 * ipa.c (can_replace_by_local_alias): Simplify.
13013 (ipa_profile): Produce speculative calls in non-lto, too;
13014 add simple cost model; produce local aliases.
13016 2013-08-13 David Malcolm <dmalcolm@redhat.com>
13018 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
13021 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
13023 * config/i386/i386.c (ix86_function_versions): Use error + inform.
13025 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
13027 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
13028 iterator instead of X87MODEF.
13030 2013-08-12 Perez Read <netfirewall@gmail.com>
13033 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
13034 operand 0 for intel asm alternative.
13035 (*movabs<mode>_2): Ditto for operand 1.
13037 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
13039 * config/aarch64/arm_none.h
13040 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
13042 2013-08-12 Nick Clifton <nickc@redhat.com>
13044 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
13046 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
13048 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
13049 expand for QI/HImode operand to produce more effictive code for
13050 unsigned char(short) --> float(double) conversion.
13052 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
13054 * doc/invoke.texi: Mention that -ftls-model does not force the final
13057 2013-08-12 Marek Polacek <polacek@redhat.com>
13058 Marc Glisse <marc.glisse@inria.fr>
13060 PR tree-optimization/57980
13061 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
13062 when creating -1 constant.
13064 2013-08-10 Jan Hubicka <jh@suse.cz>
13066 Workaround binutils PR14342.
13067 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
13068 (gimple_init_edge_profiler): Likewise.
13069 (gimple_gen_ic_func_profiler): Likewise.
13071 2013-08-09 Jan Hubicka <jh@suse.cz>
13073 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
13075 2013-08-09 Xinliang David Li <davidxl@google.com>
13077 * config/i386/stringop.def: New file.
13078 * config/i386/stringop.opt: New file.
13079 * config/i386/i386-opts.h: Include stringopt.def.
13080 * config/i386/i386.opt: Include stringopt.opt.
13081 * config/i386/i386.c (ix86_option_override_internal):
13082 Override default size based stringop inline strategies with options.
13083 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
13086 2013-08-09 Jan Hubicka <jh@suse.cz>
13088 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
13090 2013-08-09 Jan Hubicka <jh@suse.cz>
13092 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
13094 (dump_cgraph_node): Dump profile-id.
13095 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
13096 and common_target_probability.
13097 * lto-cgraph.c (lto_output_edge): Stream common targets.
13098 (lto_output_node): Stream profile ids.
13099 (input_node): Stream profile ids.
13100 (input_edge): Stream common targets.
13101 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
13102 * ipa.c: Include value-prof.h
13103 (ipa_profile_generate_summary): Turn indirect call statement histograms
13104 into common targets.
13105 (ipa_profile): Turn common targets into speculative edges.
13107 2013-08-09 Jan Hubicka <jh@suse.cz>
13109 * cgraph.h (cgraph_node): Add profile_id.
13110 * value-prof.c (cgraph_node_map): Turn into pointer_map.
13111 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
13112 (del_node_map): Update.
13113 (find_func_by_funcdef_no): Replace by ...
13114 (find_func_by_profile_id): ... this one.
13115 (gimple_ic_transform): Do not remove useful histograms when
13116 speculation is not done; dump info when indirect call removal
13118 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
13119 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
13120 (__gcov_indirect_call_profiler_v2): .. this one.
13121 * profile.h (init_node_map): Update.
13122 * coverage.c (coverage_compute_profile_id): New function.
13123 * coverage.h (coverage_compute_profile_id): Declare.
13124 * tree-profile.c (init_ic_make_global_vars): Make
13125 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
13126 (gimple_init_edge_profiler): Update prototype of
13127 __gcov_indirect_call_profiler.
13128 (gimple_gen_ic_func_profiler): Simplify.
13129 (tree_profiling): Use init_node_map
13131 2013-08-09 Jan Hubicka <jh@suse.cz>
13133 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
13134 non-speculative refs.
13135 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
13136 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
13137 (cgraph_set_call_stmt): Likewise.
13138 (cgraph_create_edge_1): Fix release checking compilatoin;
13139 clear lto_stmt_uid.
13140 (cgraph_free_edge): Free indirect info.
13141 (cgraph_turn_edge_to_speculative): New function.
13142 (cgraph_speculative_call_info): New function.
13143 (cgraph_make_edge_direct): Return direct edge; handle speculation.
13144 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
13145 (dump_cgraph_node): Dump speculation.
13146 (verify_edge_count_and_frequency): Accept speculative edges.
13147 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
13148 (verify_cgraph_node): Handle speculation.
13149 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
13150 (cgraph_set_call_stmt): Update prototype.
13151 (cgraph_make_edge_direct): Update prototype.
13152 (cgraph_speculative_call_info): Declare.
13153 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
13154 to change; update call of ipa_find_references.
13155 * ipa-ref.c (ipa_record_reference): Fix return value; clear
13156 lto_stmt_uid and speculative flags.
13157 (ipa_dump_references): Dump speculation.
13158 (ipa_clone_references): Clone speculative flag.
13159 (ipa_clone_referring): Likewise.
13160 (ipa_clone_ref): New function.
13161 (ipa_find_reference): Look into lto_stmt_uids
13162 (ipa_clear_stmts_in_references): Do not clear speculative calls.
13163 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
13164 (ipa_find_reference): Update declaration.
13165 (ipa_clone_ref): Declare.
13166 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
13167 stream speculative flag.
13168 (lto_output_ref): Stream statements uids and speculation.
13169 (input_ref): Likewise.
13170 (input_edge): Stream speuclation.
13171 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
13172 (cgraph_set_call_stmt_including_clones): Handle speculation.
13173 * ipa-inline.c (heap_edge_removal_hook): New function.
13174 (inline_small_functions): Register it.
13175 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
13176 also initialize refs.
13177 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
13179 (try_make_edge_direct_simple_call): Likewise.
13180 (try_make_edge_direct_simple_call): Likewise.
13181 (update_indirect_edges_after_inlining): Likewise.
13182 (remove_described_reference): Look proper lto_stmt_uid.
13183 (propagate_controlled_uses): Likewise.
13184 (propagate_controlled_uses): Liekwise.
13185 * tree-inline.c (copy_bb): Copy speculative edges.
13186 (redirect_all_calls): New function.
13187 (copy_cfg_body): Do redirection after loop info is updated.
13188 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
13190 2013-08-09 Jan Hubicka <jh@suse.cz>
13192 * lto-streamer-out.c (output_function): Renumber PHIs.
13193 * lto-streamer-in.c (input_function): Likewise.
13195 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
13197 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
13198 (get_lane_unsigned): Likewise.
13199 (dup_lane_scalar): Likewise.
13200 (get_lane): enable for VALL.
13201 * config/aarch64/aarch64-simd.md
13202 (aarch64_dup_lane_scalar<mode>): Remove.
13203 (aarch64_get_lane_signed<mode>): Likewise.
13204 (aarch64_get_lane_unsigned<mode>): Likewise.
13205 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
13206 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
13207 (aarch64_get_lane<mode>): Enable for all vector modes.
13208 (aarch64_get_lanedi): Remove misleading constraints.
13209 * config/aarch64/arm_neon.h
13210 (__aarch64_vget_lane_any): Define.
13211 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
13212 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
13213 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
13214 * config/aarch64/iterators.md (VDQQH): New.
13215 (VDQQHS): Likewise.
13216 (vwcore): Likewise.
13218 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
13220 * configure.ac: Add GAS check for LEON instructions on SPARC.
13221 * configure: Regenerate.
13222 * config.in: Likewise.
13223 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
13225 * config/sparc/sparc.opt (LEON, LEON3): New masks.
13226 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
13228 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
13229 (AS_LEON_FLAG): New macro.
13230 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
13231 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
13232 Deal with LEON and LEON3 for the memory model.
13233 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
13234 (atomic_compare_and_swap<mode>_1): Likewise.
13235 (*atomic_compare_and_swap<mode>_1): Likewise.
13237 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13239 * config/arm/neon.md (vcond): Fix floating-point vector
13240 comparisons against 0.
13242 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
13244 * lra-constraints.c (emit_spill_move): Remove assert.
13245 (process_alt_operands): Add more debugging
13246 output. Increase reject for spilling into memory. Decrease
13247 reject for reloading scratch.
13248 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
13250 2013-08-08 Steve Ellcey <sellcey@mips.com>
13252 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
13253 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
13254 micromips incompatible. Add nan2008.
13255 (MULTILIB_DIRNAMES): Add nan2008.
13256 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13257 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
13258 and micromips incompatible. Add nan2008.
13259 (MULTILIB_DIRNAMES): Add nan2008.
13260 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13262 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
13264 PR rtl-optimization/58079
13265 * combine.c (combine_simplify_rtx): Avoid using SUBST if
13266 simplify_comparison has widened a comparison with an integer.
13268 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13270 * config/arm/neon.md (movmisalign<mode>): Disable when we
13271 don't allow unaligned accesses.
13272 (*movmisalign<mode>_neon_store): Likewise.
13273 (*movmisalign<mode>_neon_load): Likewise.
13274 (*movmisalign<mode>_neon_store): Likewise.
13275 (*movmisalign<mode>_neon_load): Likewise.
13277 2013-08-08 Jan Hubicka <jh@suse.cz>
13279 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
13280 (make_pass_rebuild_cgraph_edges): Also clear references.
13281 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
13282 * ipa-inline-transform.c (inline_transform): Remove all references
13284 * cgraphunit.c (expand_function): Remove all references after
13286 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
13287 (ipa_find_reference): Rewrite to iterator.
13288 (remove_stmt_references): Likewise.
13289 (ipa_clear_stmts_in_references): New function.
13290 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
13291 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
13293 * ipa-split.c (split_function): Remove references in split function.
13295 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13298 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
13299 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
13301 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13304 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
13305 suggested mode for the assignment isn't compatible with the
13306 registers required.
13308 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13311 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
13313 2013-08-07 Xinliang David Li <davidxl@google.com>
13315 * config/i386/i386.opt: New option -mtune-ctrl=.
13316 * config/i386/x86-tune.def: New file.
13317 * config/i386/i386.h: include x86-tune.def.
13318 * config/i386/i386.c (ix86_option_override_internal):
13319 Parsing -mtune-ctrl= option and set tune features.
13321 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
13324 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
13325 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
13328 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13330 * rtl.h (update_alignments): Declare.
13331 * final.c (grow_label_align): New function extracted from...
13332 (shorten_branches): ...here. Call it.
13333 (update_alignments): New function.
13334 * reorg.c (sibling_labels): New variable.
13335 (get_label_before): Add SIBLING parameter. If it is non-zero, push
13336 the new label along with it onto the sibling_labels vector.
13337 (fill_simple_delay_slots): Adjust call to get_label_before.
13338 (fill_slots_from_thread): Likewise.
13339 (relax_delay_slots): Likewise.
13340 (make_return_insns): Likewise.
13341 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
13343 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13345 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
13346 document its semantics.
13347 (diagnostic_report_diagnostic): Adjust accordingly.
13349 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13351 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
13352 (sparc_option_override): ...and port to new C++ pass API.
13353 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
13355 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
13357 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
13359 2013-08-06 Caroline Tice <cmtice@google.com>
13361 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
13362 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
13363 * tree-pass.h: Add pass_vtable_verify.
13364 * varasm.c (assemble_variable): Add code to properly set the comdat
13365 section and name for the .vtable_map_vars section.
13366 (assemble_vtyv_preinit_initializer): New function.
13367 (default_sectin_type_flags): Make sure .vtable_map_vars section has
13369 * output.h: Add function decl for assemble_vtv_preinit_initializer.
13370 * vtable-verify.c: New file.
13371 * vtable-verify.h: New file.
13372 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
13373 initialiation levels.
13374 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
13375 * passes.def: Insert pass_vtable_verify.
13376 * aclocal.m4: Reorder includes.
13377 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
13378 -fvtv-counts options.
13379 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
13380 as appropriate, if -fvtable-verify=... is used.
13381 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
13382 -fvtable-verify=... is used.
13383 * Makefile.in (OBJS): Add vtable-verify.o to list.
13384 (vtable-verify.o): Add new build rule.
13385 (GTFILES): Add vtable-verify.c to list.
13386 * common.opt (fvtable-verify=): New flag.
13387 (vtv_priority): Values for fvtable-verify= flag.
13388 (fvtv-counts): New flag.
13389 (fvtv-debug): New flag.
13390 * tree.h (save_vtable_map_decl): New extern function decl.
13392 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13394 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
13395 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
13396 (pass_data_rl78_devirt): ...new pass_data instance and...
13397 (make_pass_rl78_devirt): ...new function.
13398 (rl78_asm_file_start): Port pass registration to new C++ API.
13400 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13402 * coretypes.h (rtl_opt_pass): Add.
13403 (gcc::context): Add.
13404 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
13405 (epiphany_init): Port to new C++ pass API.
13406 (epiphany_optimize_mode_switching): Likewise.
13407 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
13408 (pass_manager::get_pass_mode_switching): New.
13409 (pass_manager::get_pass_peephole2): New.
13410 * mode-switching.c (pass_mode_switching): Add clone method.
13411 * recog.c (pass_peephole2): Add clone method.
13412 (pass_split_all_insns): Add clone method.
13414 2013-08-06 David Malcolm <dmalcolm@redhat.com>
13416 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
13417 (mips_option_override): ...here, porting to new C++ API for passes.
13419 2013-08-06 Jan Hubicka <jh@suse.cz>
13421 * cgraph.c (cgraph_get_body): New function based on lto.c
13423 * cgraph.h (cgraph_get_body): Declare.
13424 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
13426 * cgraphunit.c (expand_function): Get body prior expanding.
13427 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
13428 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
13430 * passes.c (do_per_function_toporder): Get body.
13431 * tree-inline.c (expand_call_inline): Get body prior inlining it.
13432 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
13434 2013-08-06 Martin Jambor <mjambor@suse.cz>
13437 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
13438 is not re-finalized. Rename second parameter to no_collect.
13440 2013-08-06 Martin Jambor <mjambor@suse.cz>
13442 PR middle-end/58041
13443 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
13444 MEM_REF has proper alignment information.
13446 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
13449 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
13451 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
13452 rtx (*) (rtx, ...) with insn_gen_fn.
13453 * genoutput.c (output_insn_data): Cast gen_? function pointers to
13454 insn_gen_fn::stored_funcptr. Add initializer braces.
13456 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13458 Rewrite how instances of passes are cloned to remove assumptions
13459 about their sizes (thus allowing pass subclasses to have
13460 additional data fields, albeit non-GC-managed ones at this point).
13462 * passes.c (make_pass_instance): Now that passes have clone
13463 methods, rewrite this function to eliminate XNEW and memcpy
13464 calls that used hardcoded sizes. Since this function no longer
13465 creates pass instances, rename it to...
13466 (add_pass_instance): ...this. Document the old way that passes were
13467 numbered and flagged, and rework this function to continue using it.
13468 (next_pass_1): Add an initial_pass argument for use by
13470 (position_pass): When adding multiple instances of a pass, use
13471 the pass's clone method, rather than relying on the XNEW/memcpy
13472 within the former make_pass_instance (now add_pass_instance).
13473 (pass_manager::pass_manager): When invoking next_pass_1, also supply
13474 the initial instance of the current pass within the pass manager.
13476 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13478 This is the automated part of the conversion of passes from C
13479 structs to C++ classes.
13481 Patch autogenerated by refactor_passes.py from
13482 https://github.com/davidmalcolm/gcc-refactoring-scripts
13483 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
13485 * asan.c (pass_asan): Convert from a global struct to a subclass of
13486 gimple_opt_pass along with...
13487 (pass_data_asan): ...new pass_data instance and...
13488 (make_pass_asan): ...new function.
13489 (pass_asan_O0): Convert from a global struct to a subclass of
13490 gimple_opt_pass along with...
13491 (pass_data_asan_O0): ...new pass_data instance and...
13492 (make_pass_asan_O0): ...new function.
13493 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
13494 subclass of rtl_opt_pass along with...
13495 (pass_data_inc_dec): ...new pass_data instance and...
13496 (make_pass_inc_dec): ...new function.
13497 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
13498 a subclass of rtl_opt_pass along with...
13499 (pass_data_reorder_blocks): ...new pass_data instance and...
13500 (make_pass_reorder_blocks): ...new function.
13501 (pass_duplicate_computed_gotos): Convert from a global struct to a
13502 subclass of rtl_opt_pass along with...
13503 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
13504 (make_pass_duplicate_computed_gotos): ...new function.
13505 (pass_partition_blocks): Convert from a global struct to a subclass of
13506 rtl_opt_pass along with...
13507 (pass_data_partition_blocks): ...new pass_data instance and...
13508 (make_pass_partition_blocks): ...new function.
13509 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
13510 struct to a subclass of rtl_opt_pass along with...
13511 (pass_data_branch_target_load_optimize1): ...new pass_data instance
13513 (make_pass_branch_target_load_optimize1): ...new function.
13514 (pass_branch_target_load_optimize2): Convert from a global struct to a
13515 subclass of rtl_opt_pass along with...
13516 (pass_data_branch_target_load_optimize2): ...new pass_data instance
13518 (make_pass_branch_target_load_optimize2): ...new function.
13519 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
13520 of rtl_opt_pass along with...
13521 (pass_data_jump): ...new pass_data instance and...
13522 (make_pass_jump): ...new function.
13523 (pass_jump2): Convert from a global struct to a subclass of
13524 rtl_opt_pass along with...
13525 (pass_data_jump2): ...new pass_data instance and...
13526 (make_pass_jump2): ...new function.
13527 * cfgexpand.c (pass_expand): Convert from a global struct to a
13528 subclass of rtl_opt_pass along with...
13529 (pass_data_expand): ...new pass_data instance and...
13530 (make_pass_expand): ...new function.
13531 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
13532 of rtl_opt_pass along with...
13533 (pass_data_free_cfg): ...new pass_data instance and...
13534 (make_pass_free_cfg): ...new function.
13535 (pass_into_cfg_layout_mode): Convert from a global struct to a
13536 subclass of rtl_opt_pass along with...
13537 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
13538 (make_pass_into_cfg_layout_mode): ...new function.
13539 (pass_outof_cfg_layout_mode): Convert from a global struct to a
13540 subclass of rtl_opt_pass along with...
13541 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
13542 (make_pass_outof_cfg_layout_mode): ...new function.
13543 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
13544 struct to a subclass of gimple_opt_pass along with...
13545 (pass_data_build_cgraph_edges): ...new pass_data instance and...
13546 (make_pass_build_cgraph_edges): ...new function.
13547 (pass_rebuild_cgraph_edges): Convert from a global struct to a
13548 subclass of gimple_opt_pass along with...
13549 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
13550 (make_pass_rebuild_cgraph_edges): ...new function.
13551 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
13552 subclass of gimple_opt_pass along with...
13553 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
13555 (make_pass_remove_cgraph_callee_edges): ...new function.
13556 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
13557 struct to a subclass of rtl_opt_pass along with...
13558 (pass_data_stack_adjustments): ...new pass_data instance and...
13559 (make_pass_stack_adjustments): ...new function.
13560 * combine.c (pass_combine): Convert from a global struct to a subclass
13561 of rtl_opt_pass along with...
13562 (pass_data_combine): ...new pass_data instance and...
13563 (make_pass_combine): ...new function.
13564 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
13565 global struct to a subclass of rtl_opt_pass along with...
13566 (pass_data_compare_elim_after_reload): ...new pass_data instance
13568 (make_pass_compare_elim_after_reload): ...new function.
13569 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
13570 of rtl_opt_pass along with...
13571 (pass_data_rtl_cprop): ...new pass_data instance and...
13572 (make_pass_rtl_cprop): ...new function.
13573 * cse.c (pass_cse): Convert from a global struct to a subclass of
13574 rtl_opt_pass along with...
13575 (pass_data_cse): ...new pass_data instance and...
13576 (make_pass_cse): ...new function.
13577 (pass_cse2): Convert from a global struct to a subclass of
13578 rtl_opt_pass along with...
13579 (pass_data_cse2): ...new pass_data instance and...
13580 (make_pass_cse2): ...new function.
13581 (pass_cse_after_global_opts): Convert from a global struct to a
13582 subclass of rtl_opt_pass along with...
13583 (pass_data_cse_after_global_opts): ...new pass_data instance and...
13584 (make_pass_cse_after_global_opts): ...new function.
13585 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
13586 of rtl_opt_pass along with...
13587 (pass_data_ud_rtl_dce): ...new pass_data instance and...
13588 (make_pass_ud_rtl_dce): ...new function.
13589 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
13590 rtl_opt_pass along with...
13591 (pass_data_fast_rtl_dce): ...new pass_data instance and...
13592 (make_pass_fast_rtl_dce): ...new function.
13593 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
13594 a subclass of rtl_opt_pass along with...
13595 (pass_data_df_initialize_opt): ...new pass_data instance and...
13596 (make_pass_df_initialize_opt): ...new function.
13597 (pass_df_initialize_no_opt): Convert from a global struct to a
13598 subclass of rtl_opt_pass along with...
13599 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
13600 (make_pass_df_initialize_no_opt): ...new function.
13601 (pass_df_finish): Convert from a global struct to a subclass of
13602 rtl_opt_pass along with...
13603 (pass_data_df_finish): ...new pass_data instance and...
13604 (make_pass_df_finish): ...new function.
13605 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
13606 rtl_opt_pass along with...
13607 (pass_data_rtl_dse1): ...new pass_data instance and...
13608 (make_pass_rtl_dse1): ...new function.
13609 (pass_rtl_dse2): Convert from a global struct to a subclass of
13610 rtl_opt_pass along with...
13611 (pass_data_rtl_dse2): ...new pass_data instance and...
13612 (make_pass_rtl_dse2): ...new function.
13613 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
13614 subclass of rtl_opt_pass along with...
13615 (pass_data_dwarf2_frame): ...new pass_data instance and...
13616 (make_pass_dwarf2_frame): ...new function.
13617 * except.c (pass_set_nothrow_function_flags): Convert from a global
13618 struct to a subclass of rtl_opt_pass along with...
13619 (pass_data_set_nothrow_function_flags): ...new pass_data instance
13621 (make_pass_set_nothrow_function_flags): ...new function.
13622 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
13623 subclass of rtl_opt_pass along with...
13624 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
13626 (make_pass_convert_to_eh_region_ranges): ...new function.
13627 * final.c (pass_compute_alignments): Convert from a global struct to a
13628 subclass of rtl_opt_pass along with...
13629 (pass_data_compute_alignments): ...new pass_data instance and...
13630 (make_pass_compute_alignments): ...new function.
13631 (pass_final): Convert from a global struct to a subclass of
13632 rtl_opt_pass along with...
13633 (pass_data_final): ...new pass_data instance and...
13634 (make_pass_final): ...new function.
13635 (pass_shorten_branches): Convert from a global struct to a subclass of
13636 rtl_opt_pass along with...
13637 (pass_data_shorten_branches): ...new pass_data instance and...
13638 (make_pass_shorten_branches): ...new function.
13639 (pass_clean_state): Convert from a global struct to a subclass of
13640 rtl_opt_pass along with...
13641 (pass_data_clean_state): ...new pass_data instance and...
13642 (make_pass_clean_state): ...new function.
13643 * function.c (pass_instantiate_virtual_regs): Convert from a global
13644 struct to a subclass of rtl_opt_pass along with...
13645 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
13646 (make_pass_instantiate_virtual_regs): ...new function.
13647 (pass_leaf_regs): Convert from a global struct to a subclass of
13648 rtl_opt_pass along with...
13649 (pass_data_leaf_regs): ...new pass_data instance and...
13650 (make_pass_leaf_regs): ...new function.
13651 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
13652 subclass of rtl_opt_pass along with...
13653 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
13655 (make_pass_thread_prologue_and_epilogue): ...new function.
13656 (pass_match_asm_constraints): Convert from a global struct to a
13657 subclass of rtl_opt_pass along with...
13658 (pass_data_match_asm_constraints): ...new pass_data instance and...
13659 (make_pass_match_asm_constraints): ...new function.
13660 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
13661 subclass of rtl_opt_pass along with...
13662 (pass_data_rtl_fwprop): ...new pass_data instance and...
13663 (make_pass_rtl_fwprop): ...new function.
13664 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
13665 rtl_opt_pass along with...
13666 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
13667 (make_pass_rtl_fwprop_addr): ...new function.
13668 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
13669 rtl_opt_pass along with...
13670 (pass_data_rtl_pre): ...new pass_data instance and...
13671 (make_pass_rtl_pre): ...new function.
13672 (pass_rtl_hoist): Convert from a global struct to a subclass of
13673 rtl_opt_pass along with...
13674 (pass_data_rtl_hoist): ...new pass_data instance and...
13675 (make_pass_rtl_hoist): ...new function.
13676 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
13677 subclass of gimple_opt_pass along with...
13678 (pass_data_lower_cf): ...new pass_data instance and...
13679 (make_pass_lower_cf): ...new function.
13680 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
13681 from a global struct to a subclass of gimple_opt_pass along with...
13682 (pass_data_strength_reduction): ...new pass_data instance and...
13683 (make_pass_strength_reduction): ...new function.
13684 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
13685 of rtl_opt_pass along with...
13686 (pass_data_rtl_ifcvt): ...new pass_data instance and...
13687 (make_pass_rtl_ifcvt): ...new function.
13688 (pass_if_after_combine): Convert from a global struct to a subclass of
13689 rtl_opt_pass along with...
13690 (pass_data_if_after_combine): ...new pass_data instance and...
13691 (make_pass_if_after_combine): ...new function.
13692 (pass_if_after_reload): Convert from a global struct to a subclass of
13693 rtl_opt_pass along with...
13694 (pass_data_if_after_reload): ...new pass_data instance and...
13695 (make_pass_if_after_reload): ...new function.
13696 * init-regs.c (pass_initialize_regs): Convert from a global struct to
13697 a subclass of rtl_opt_pass along with...
13698 (pass_data_initialize_regs): ...new pass_data instance and...
13699 (make_pass_initialize_regs): ...new function.
13700 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
13701 of ipa_opt_pass_d along with...
13702 (pass_data_ipa_cp): ...new pass_data instance and...
13703 (make_pass_ipa_cp): ...new function.
13704 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
13705 global struct to a subclass of gimple_opt_pass along with...
13706 (pass_data_inline_parameters): ...new pass_data instance and...
13707 (make_pass_inline_parameters): ...new function.
13708 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
13709 subclass of gimple_opt_pass along with...
13710 (pass_data_early_inline): ...new pass_data instance and...
13711 (make_pass_early_inline): ...new function.
13712 (pass_ipa_inline): Convert from a global struct to a subclass of
13713 ipa_opt_pass_d along with...
13714 (pass_data_ipa_inline): ...new pass_data instance and...
13715 (make_pass_ipa_inline): ...new function.
13716 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
13717 struct to a subclass of gimple_opt_pass along with...
13718 (pass_data_local_pure_const): ...new pass_data instance and...
13719 (make_pass_local_pure_const): ...new function.
13720 (pass_ipa_pure_const): Convert from a global struct to a subclass of
13721 ipa_opt_pass_d along with...
13722 (pass_data_ipa_pure_const): ...new pass_data instance and...
13723 (make_pass_ipa_pure_const): ...new function.
13724 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
13725 to a subclass of ipa_opt_pass_d along with...
13726 (pass_data_ipa_reference): ...new pass_data instance and...
13727 (make_pass_ipa_reference): ...new function.
13728 * ipa-split.c (pass_split_functions): Convert from a global struct to
13729 a subclass of gimple_opt_pass along with...
13730 (pass_data_split_functions): ...new pass_data instance and...
13731 (make_pass_split_functions): ...new function.
13732 (pass_feedback_split_functions): Convert from a global struct to a
13733 subclass of gimple_opt_pass along with...
13734 (pass_data_feedback_split_functions): ...new pass_data instance and...
13735 (make_pass_feedback_split_functions): ...new function.
13736 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13737 global struct to a subclass of simple_ipa_opt_pass along with...
13738 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13740 (make_pass_ipa_function_and_variable_visibility): ...new function.
13741 (pass_ipa_free_inline_summary): Convert from a global struct to a
13742 subclass of simple_ipa_opt_pass along with...
13743 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13744 (make_pass_ipa_free_inline_summary): ...new function.
13745 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13746 subclass of ipa_opt_pass_d along with...
13747 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13749 (make_pass_ipa_whole_program_visibility): ...new function.
13750 (pass_ipa_profile): Convert from a global struct to a subclass of
13751 ipa_opt_pass_d along with...
13752 (pass_data_ipa_profile): ...new pass_data instance and...
13753 (make_pass_ipa_profile): ...new function.
13754 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13755 ipa_opt_pass_d along with...
13756 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13757 (make_pass_ipa_cdtor_merge): ...new function.
13758 * ira.c (pass_ira): Convert from a global struct to a subclass of
13759 rtl_opt_pass along with...
13760 (pass_data_ira): ...new pass_data instance and...
13761 (make_pass_ira): ...new function.
13762 (pass_reload): Convert from a global struct to a subclass of
13763 rtl_opt_pass along with...
13764 (pass_data_reload): ...new pass_data instance and...
13765 (make_pass_reload): ...new function.
13766 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13767 subclass of rtl_opt_pass along with...
13768 (pass_data_cleanup_barriers): ...new pass_data instance and...
13769 (make_pass_cleanup_barriers): ...new function.
13770 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13771 of rtl_opt_pass along with...
13772 (pass_data_loop2): ...new pass_data instance and...
13773 (make_pass_loop2): ...new function.
13774 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13775 rtl_opt_pass along with...
13776 (pass_data_rtl_loop_init): ...new pass_data instance and...
13777 (make_pass_rtl_loop_init): ...new function.
13778 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13779 rtl_opt_pass along with...
13780 (pass_data_rtl_loop_done): ...new pass_data instance and...
13781 (make_pass_rtl_loop_done): ...new function.
13782 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13783 subclass of rtl_opt_pass along with...
13784 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13785 (make_pass_rtl_move_loop_invariants): ...new function.
13786 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13787 rtl_opt_pass along with...
13788 (pass_data_rtl_unswitch): ...new pass_data instance and...
13789 (make_pass_rtl_unswitch): ...new function.
13790 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13791 subclass of rtl_opt_pass along with...
13792 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13794 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13795 (pass_rtl_doloop): Convert from a global struct to a subclass of
13796 rtl_opt_pass along with...
13797 (pass_data_rtl_doloop): ...new pass_data instance and...
13798 (make_pass_rtl_doloop): ...new function.
13799 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13800 a subclass of rtl_opt_pass along with...
13801 (pass_data_lower_subreg): ...new pass_data instance and...
13802 (make_pass_lower_subreg): ...new function.
13803 (pass_lower_subreg2): Convert from a global struct to a subclass of
13804 rtl_opt_pass along with...
13805 (pass_data_lower_subreg2): ...new pass_data instance and...
13806 (make_pass_lower_subreg2): ...new function.
13807 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13808 struct to a subclass of ipa_opt_pass_d along with...
13809 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13810 (make_pass_ipa_lto_gimple_out): ...new function.
13811 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13812 of ipa_opt_pass_d along with...
13813 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13814 (make_pass_ipa_lto_finish_out): ...new function.
13815 * mode-switching.c (pass_mode_switching): Convert from a global struct
13816 to a subclass of rtl_opt_pass along with...
13817 (pass_data_mode_switching): ...new pass_data instance and...
13818 (make_pass_mode_switching): ...new function.
13819 * modulo-sched.c (pass_sms): Convert from a global struct to a
13820 subclass of rtl_opt_pass along with...
13821 (pass_data_sms): ...new pass_data instance and...
13822 (make_pass_sms): ...new function.
13823 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13824 subclass of gimple_opt_pass along with...
13825 (pass_data_expand_omp): ...new pass_data instance and...
13826 (make_pass_expand_omp): ...new function.
13827 (pass_lower_omp): Convert from a global struct to a subclass of
13828 gimple_opt_pass along with...
13829 (pass_data_lower_omp): ...new pass_data instance and...
13830 (make_pass_lower_omp): ...new function.
13831 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13832 of gimple_opt_pass along with...
13833 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13834 (make_pass_diagnose_omp_blocks): ...new function.
13835 * passes.c (pass_early_local_passes): Convert from a global struct to
13836 a subclass of simple_ipa_opt_pass along with...
13837 (pass_data_early_local_passes): ...new pass_data instance and...
13838 (make_pass_early_local_passes): ...new function.
13839 (pass_all_early_optimizations): Convert from a global struct to a
13840 subclass of gimple_opt_pass along with...
13841 (pass_data_all_early_optimizations): ...new pass_data instance and...
13842 (make_pass_all_early_optimizations): ...new function.
13843 (pass_all_optimizations): Convert from a global struct to a subclass
13844 of gimple_opt_pass along with...
13845 (pass_data_all_optimizations): ...new pass_data instance and...
13846 (make_pass_all_optimizations): ...new function.
13847 (pass_all_optimizations_g): Convert from a global struct to a subclass
13848 of gimple_opt_pass along with...
13849 (pass_data_all_optimizations_g): ...new pass_data instance and...
13850 (make_pass_all_optimizations_g): ...new function.
13851 (pass_rest_of_compilation): Convert from a global struct to a subclass
13852 of rtl_opt_pass along with...
13853 (pass_data_rest_of_compilation): ...new pass_data instance and...
13854 (make_pass_rest_of_compilation): ...new function.
13855 (pass_postreload): Convert from a global struct to a subclass of
13856 rtl_opt_pass along with...
13857 (pass_data_postreload): ...new pass_data instance and...
13858 (make_pass_postreload): ...new function.
13859 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13860 subclass of rtl_opt_pass along with...
13861 (pass_data_gcse2): ...new pass_data instance and...
13862 (make_pass_gcse2): ...new function.
13863 * postreload.c (pass_postreload_cse): Convert from a global struct to
13864 a subclass of rtl_opt_pass along with...
13865 (pass_data_postreload_cse): ...new pass_data instance and...
13866 (make_pass_postreload_cse): ...new function.
13867 * predict.c (pass_profile): Convert from a global struct to a subclass
13868 of gimple_opt_pass along with...
13869 (pass_data_profile): ...new pass_data instance and...
13870 (make_pass_profile): ...new function.
13871 (pass_strip_predict_hints): Convert from a global struct to a subclass
13872 of gimple_opt_pass along with...
13873 (pass_data_strip_predict_hints): ...new pass_data instance and...
13874 (make_pass_strip_predict_hints): ...new function.
13875 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13876 of rtl_opt_pass along with...
13877 (pass_data_peephole2): ...new pass_data instance and...
13878 (make_pass_peephole2): ...new function.
13879 (pass_split_all_insns): Convert from a global struct to a subclass of
13880 rtl_opt_pass along with...
13881 (pass_data_split_all_insns): ...new pass_data instance and...
13882 (make_pass_split_all_insns): ...new function.
13883 (pass_split_after_reload): Convert from a global struct to a subclass
13884 of rtl_opt_pass along with...
13885 (pass_data_split_after_reload): ...new pass_data instance and...
13886 (make_pass_split_after_reload): ...new function.
13887 (pass_split_before_regstack): Convert from a global struct to a
13888 subclass of rtl_opt_pass along with...
13889 (pass_data_split_before_regstack): ...new pass_data instance and...
13890 (make_pass_split_before_regstack): ...new function.
13891 (pass_split_before_sched2): Convert from a global struct to a subclass
13892 of rtl_opt_pass along with...
13893 (pass_data_split_before_sched2): ...new pass_data instance and...
13894 (make_pass_split_before_sched2): ...new function.
13895 (pass_split_for_shorten_branches): Convert from a global struct to a
13896 subclass of rtl_opt_pass along with...
13897 (pass_data_split_for_shorten_branches): ...new pass_data instance
13899 (make_pass_split_for_shorten_branches): ...new function.
13900 * ree.c (pass_ree): Convert from a global struct to a subclass of
13901 rtl_opt_pass along with...
13902 (pass_data_ree): ...new pass_data instance and...
13903 (make_pass_ree): ...new function.
13904 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13905 subclass of rtl_opt_pass along with...
13906 (pass_data_stack_regs): ...new pass_data instance and...
13907 (make_pass_stack_regs): ...new function.
13908 (pass_stack_regs_run): Convert from a global struct to a subclass of
13909 rtl_opt_pass along with...
13910 (pass_data_stack_regs_run): ...new pass_data instance and...
13911 (make_pass_stack_regs_run): ...new function.
13912 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13913 subclass of rtl_opt_pass along with...
13914 (pass_data_cprop_hardreg): ...new pass_data instance and...
13915 (make_pass_cprop_hardreg): ...new function.
13916 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13917 subclass of rtl_opt_pass along with...
13918 (pass_data_reginfo_init): ...new pass_data instance and...
13919 (make_pass_reginfo_init): ...new function.
13920 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13921 of rtl_opt_pass along with...
13922 (pass_data_regmove): ...new pass_data instance and...
13923 (make_pass_regmove): ...new function.
13924 * regrename.c (pass_regrename): Convert from a global struct to a
13925 subclass of rtl_opt_pass along with...
13926 (pass_data_regrename): ...new pass_data instance and...
13927 (make_pass_regrename): ...new function.
13928 * reorg.c (pass_delay_slots): Convert from a global struct to a
13929 subclass of rtl_opt_pass along with...
13930 (pass_data_delay_slots): ...new pass_data instance and...
13931 (make_pass_delay_slots): ...new function.
13932 (pass_machine_reorg): Convert from a global struct to a subclass of
13933 rtl_opt_pass along with...
13934 (pass_data_machine_reorg): ...new pass_data instance and...
13935 (make_pass_machine_reorg): ...new function.
13936 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13937 of rtl_opt_pass along with...
13938 (pass_data_sched): ...new pass_data instance and...
13939 (make_pass_sched): ...new function.
13940 (pass_sched2): Convert from a global struct to a subclass of
13941 rtl_opt_pass along with...
13942 (pass_data_sched2): ...new pass_data instance and...
13943 (make_pass_sched2): ...new function.
13944 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13945 to a subclass of rtl_opt_pass along with...
13946 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13947 (make_pass_stack_ptr_mod): ...new function.
13948 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13949 to a subclass of rtl_opt_pass along with...
13950 (pass_data_rtl_store_motion): ...new pass_data instance and...
13951 (make_pass_rtl_store_motion): ...new function.
13952 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13953 of gimple_opt_pass along with...
13954 (pass_data_tracer): ...new pass_data instance and...
13955 (make_pass_tracer): ...new function.
13956 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13957 to a subclass of gimple_opt_pass along with...
13958 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13959 (make_pass_diagnose_tm_blocks): ...new function.
13960 (pass_lower_tm): Convert from a global struct to a subclass of
13961 gimple_opt_pass along with...
13962 (pass_data_lower_tm): ...new pass_data instance and...
13963 (make_pass_lower_tm): ...new function.
13964 (pass_tm_init): Convert from a global struct to a subclass of
13965 gimple_opt_pass along with...
13966 (pass_data_tm_init): ...new pass_data instance and...
13967 (make_pass_tm_init): ...new function.
13968 (pass_tm_mark): Convert from a global struct to a subclass of
13969 gimple_opt_pass along with...
13970 (pass_data_tm_mark): ...new pass_data instance and...
13971 (make_pass_tm_mark): ...new function.
13972 (pass_tm_edges): Convert from a global struct to a subclass of
13973 gimple_opt_pass along with...
13974 (pass_data_tm_edges): ...new pass_data instance and...
13975 (make_pass_tm_edges): ...new function.
13976 (pass_tm_memopt): Convert from a global struct to a subclass of
13977 gimple_opt_pass along with...
13978 (pass_data_tm_memopt): ...new pass_data instance and...
13979 (make_pass_tm_memopt): ...new function.
13980 (pass_ipa_tm): Convert from a global struct to a subclass of
13981 simple_ipa_opt_pass along with...
13982 (pass_data_ipa_tm): ...new pass_data instance and...
13983 (make_pass_ipa_tm): ...new function.
13984 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13985 subclass of gimple_opt_pass along with...
13986 (pass_data_call_cdce): ...new pass_data instance and...
13987 (make_pass_call_cdce): ...new function.
13988 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13989 subclass of gimple_opt_pass along with...
13990 (pass_data_build_cfg): ...new pass_data instance and...
13991 (make_pass_build_cfg): ...new function.
13992 (pass_split_crit_edges): Convert from a global struct to a subclass of
13993 gimple_opt_pass along with...
13994 (pass_data_split_crit_edges): ...new pass_data instance and...
13995 (make_pass_split_crit_edges): ...new function.
13996 (pass_warn_function_return): Convert from a global struct to a
13997 subclass of gimple_opt_pass along with...
13998 (pass_data_warn_function_return): ...new pass_data instance and...
13999 (make_pass_warn_function_return): ...new function.
14000 (pass_warn_function_noreturn): Convert from a global struct to a
14001 subclass of gimple_opt_pass along with...
14002 (pass_data_warn_function_noreturn): ...new pass_data instance and...
14003 (make_pass_warn_function_noreturn): ...new function.
14004 (pass_warn_unused_result): Convert from a global struct to a subclass
14005 of gimple_opt_pass along with...
14006 (pass_data_warn_unused_result): ...new pass_data instance and...
14007 (make_pass_warn_unused_result): ...new function.
14008 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
14009 a subclass of gimple_opt_pass along with...
14010 (pass_data_merge_phi): ...new pass_data instance and...
14011 (make_pass_merge_phi): ...new function.
14012 * tree-complex.c (pass_lower_complex): Convert from a global struct to
14013 a subclass of gimple_opt_pass along with...
14014 (pass_data_lower_complex): ...new pass_data instance and...
14015 (make_pass_lower_complex): ...new function.
14016 (pass_lower_complex_O0): Convert from a global struct to a subclass of
14017 gimple_opt_pass along with...
14018 (pass_data_lower_complex_O0): ...new pass_data instance and...
14019 (make_pass_lower_complex_O0): ...new function.
14020 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
14021 subclass of gimple_opt_pass along with...
14022 (pass_data_lower_eh): ...new pass_data instance and...
14023 (make_pass_lower_eh): ...new function.
14024 (pass_refactor_eh): Convert from a global struct to a subclass of
14025 gimple_opt_pass along with...
14026 (pass_data_refactor_eh): ...new pass_data instance and...
14027 (make_pass_refactor_eh): ...new function.
14028 (pass_lower_resx): Convert from a global struct to a subclass of
14029 gimple_opt_pass along with...
14030 (pass_data_lower_resx): ...new pass_data instance and...
14031 (make_pass_lower_resx): ...new function.
14032 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
14033 of gimple_opt_pass along with...
14034 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
14035 (make_pass_lower_eh_dispatch): ...new function.
14036 (pass_cleanup_eh): Convert from a global struct to a subclass of
14037 gimple_opt_pass along with...
14038 (pass_data_cleanup_eh): ...new pass_data instance and...
14039 (make_pass_cleanup_eh): ...new function.
14040 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
14041 to a subclass of simple_ipa_opt_pass along with...
14042 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
14043 (make_pass_ipa_lower_emutls): ...new function.
14044 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
14045 a subclass of gimple_opt_pass along with...
14046 (pass_data_if_conversion): ...new pass_data instance and...
14047 (make_pass_if_conversion): ...new function.
14048 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
14049 subclass of gimple_opt_pass along with...
14050 (pass_data_build_ssa): ...new pass_data instance and...
14051 (make_pass_build_ssa): ...new function.
14052 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
14053 global struct to a subclass of gimple_opt_pass along with...
14054 (pass_data_loop_distribution): ...new pass_data instance and...
14055 (make_pass_loop_distribution): ...new function.
14056 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
14057 subclass of gimple_opt_pass along with...
14058 (pass_data_mudflap_1): ...new pass_data instance and...
14059 (make_pass_mudflap_1): ...new function.
14060 (pass_mudflap_2): Convert from a global struct to a subclass of
14061 gimple_opt_pass along with...
14062 (pass_data_mudflap_2): ...new pass_data instance and...
14063 (make_pass_mudflap_2): ...new function.
14064 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
14065 subclass of gimple_opt_pass along with...
14066 (pass_data_mudflap_1): ...new pass_data instance and...
14067 (make_pass_mudflap_1): ...new function.
14068 (pass_mudflap_2): Convert from a global struct to a subclass of
14069 gimple_opt_pass along with...
14070 (pass_data_mudflap_2): ...new pass_data instance and...
14071 (make_pass_mudflap_2): ...new function.
14072 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
14073 gimple_opt_pass along with...
14074 (pass_data_nrv): ...new pass_data instance and...
14075 (make_pass_nrv): ...new function.
14076 (pass_return_slot): Convert from a global struct to a subclass of
14077 gimple_opt_pass along with...
14078 (pass_data_return_slot): ...new pass_data instance and...
14079 (make_pass_return_slot): ...new function.
14080 * tree-object-size.c (pass_object_sizes): Convert from a global struct
14081 to a subclass of gimple_opt_pass along with...
14082 (pass_data_object_sizes): ...new pass_data instance and...
14083 (make_pass_object_sizes): ...new function.
14084 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
14085 global struct to a subclass of gimple_opt_pass along with...
14086 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
14088 (make_pass_cleanup_cfg_post_optimizing): ...new function.
14089 (pass_fixup_cfg): Convert from a global struct to a subclass of
14090 gimple_opt_pass along with...
14091 (pass_data_fixup_cfg): ...new pass_data instance and...
14092 (make_pass_fixup_cfg): ...new function.
14093 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
14094 (make_pass_mudflap_1): ...new function.
14095 (pass_mudflap_2): Replace declaration with that of...
14096 (make_pass_mudflap_2): ...new function.
14097 (pass_asan): Replace declaration with that of...
14098 (make_pass_asan): ...new function.
14099 (pass_asan_O0): Replace declaration with that of...
14100 (make_pass_asan_O0): ...new function.
14101 (pass_tsan): Replace declaration with that of...
14102 (make_pass_tsan): ...new function.
14103 (pass_tsan_O0): Replace declaration with that of...
14104 (make_pass_tsan_O0): ...new function.
14105 (pass_lower_cf): Replace declaration with that of...
14106 (make_pass_lower_cf): ...new function.
14107 (pass_refactor_eh): Replace declaration with that of...
14108 (make_pass_refactor_eh): ...new function.
14109 (pass_lower_eh): Replace declaration with that of...
14110 (make_pass_lower_eh): ...new function.
14111 (pass_lower_eh_dispatch): Replace declaration with that of...
14112 (make_pass_lower_eh_dispatch): ...new function.
14113 (pass_lower_resx): Replace declaration with that of...
14114 (make_pass_lower_resx): ...new function.
14115 (pass_build_cfg): Replace declaration with that of...
14116 (make_pass_build_cfg): ...new function.
14117 (pass_early_tree_profile): Replace declaration with that of...
14118 (make_pass_early_tree_profile): ...new function.
14119 (pass_cleanup_eh): Replace declaration with that of...
14120 (make_pass_cleanup_eh): ...new function.
14121 (pass_sra): Replace declaration with that of...
14122 (make_pass_sra): ...new function.
14123 (pass_sra_early): Replace declaration with that of...
14124 (make_pass_sra_early): ...new function.
14125 (pass_early_ipa_sra): Replace declaration with that of...
14126 (make_pass_early_ipa_sra): ...new function.
14127 (pass_tail_recursion): Replace declaration with that of...
14128 (make_pass_tail_recursion): ...new function.
14129 (pass_tail_calls): Replace declaration with that of...
14130 (make_pass_tail_calls): ...new function.
14131 (pass_tree_loop): Replace declaration with that of...
14132 (make_pass_tree_loop): ...new function.
14133 (pass_tree_loop_init): Replace declaration with that of...
14134 (make_pass_tree_loop_init): ...new function.
14135 (pass_lim): Replace declaration with that of...
14136 (make_pass_lim): ...new function.
14137 (pass_tree_unswitch): Replace declaration with that of...
14138 (make_pass_tree_unswitch): ...new function.
14139 (pass_predcom): Replace declaration with that of...
14140 (make_pass_predcom): ...new function.
14141 (pass_iv_canon): Replace declaration with that of...
14142 (make_pass_iv_canon): ...new function.
14143 (pass_scev_cprop): Replace declaration with that of...
14144 (make_pass_scev_cprop): ...new function.
14145 (pass_empty_loop): Replace declaration with that of...
14146 (make_pass_empty_loop): ...new function.
14147 (pass_record_bounds): Replace declaration with that of...
14148 (make_pass_record_bounds): ...new function.
14149 (pass_graphite): Replace declaration with that of...
14150 (make_pass_graphite): ...new function.
14151 (pass_graphite_transforms): Replace declaration with that of...
14152 (make_pass_graphite_transforms): ...new function.
14153 (pass_if_conversion): Replace declaration with that of...
14154 (make_pass_if_conversion): ...new function.
14155 (pass_loop_distribution): Replace declaration with that of...
14156 (make_pass_loop_distribution): ...new function.
14157 (pass_vectorize): Replace declaration with that of...
14158 (make_pass_vectorize): ...new function.
14159 (pass_slp_vectorize): Replace declaration with that of...
14160 (make_pass_slp_vectorize): ...new function.
14161 (pass_complete_unroll): Replace declaration with that of...
14162 (make_pass_complete_unroll): ...new function.
14163 (pass_complete_unrolli): Replace declaration with that of...
14164 (make_pass_complete_unrolli): ...new function.
14165 (pass_parallelize_loops): Replace declaration with that of...
14166 (make_pass_parallelize_loops): ...new function.
14167 (pass_loop_prefetch): Replace declaration with that of...
14168 (make_pass_loop_prefetch): ...new function.
14169 (pass_iv_optimize): Replace declaration with that of...
14170 (make_pass_iv_optimize): ...new function.
14171 (pass_tree_loop_done): Replace declaration with that of...
14172 (make_pass_tree_loop_done): ...new function.
14173 (pass_ch): Replace declaration with that of...
14174 (make_pass_ch): ...new function.
14175 (pass_ccp): Replace declaration with that of...
14176 (make_pass_ccp): ...new function.
14177 (pass_phi_only_cprop): Replace declaration with that of...
14178 (make_pass_phi_only_cprop): ...new function.
14179 (pass_build_ssa): Replace declaration with that of...
14180 (make_pass_build_ssa): ...new function.
14181 (pass_build_alias): Replace declaration with that of...
14182 (make_pass_build_alias): ...new function.
14183 (pass_build_ealias): Replace declaration with that of...
14184 (make_pass_build_ealias): ...new function.
14185 (pass_dominator): Replace declaration with that of...
14186 (make_pass_dominator): ...new function.
14187 (pass_dce): Replace declaration with that of...
14188 (make_pass_dce): ...new function.
14189 (pass_dce_loop): Replace declaration with that of...
14190 (make_pass_dce_loop): ...new function.
14191 (pass_cd_dce): Replace declaration with that of...
14192 (make_pass_cd_dce): ...new function.
14193 (pass_call_cdce): Replace declaration with that of...
14194 (make_pass_call_cdce): ...new function.
14195 (pass_merge_phi): Replace declaration with that of...
14196 (make_pass_merge_phi): ...new function.
14197 (pass_split_crit_edges): Replace declaration with that of...
14198 (make_pass_split_crit_edges): ...new function.
14199 (pass_pre): Replace declaration with that of...
14200 (make_pass_pre): ...new function.
14201 (pass_profile): Replace declaration with that of...
14202 (make_pass_profile): ...new function.
14203 (pass_strip_predict_hints): Replace declaration with that of...
14204 (make_pass_strip_predict_hints): ...new function.
14205 (pass_lower_complex_O0): Replace declaration with that of...
14206 (make_pass_lower_complex_O0): ...new function.
14207 (pass_lower_complex): Replace declaration with that of...
14208 (make_pass_lower_complex): ...new function.
14209 (pass_lower_vector): Replace declaration with that of...
14210 (make_pass_lower_vector): ...new function.
14211 (pass_lower_vector_ssa): Replace declaration with that of...
14212 (make_pass_lower_vector_ssa): ...new function.
14213 (pass_lower_omp): Replace declaration with that of...
14214 (make_pass_lower_omp): ...new function.
14215 (pass_diagnose_omp_blocks): Replace declaration with that of...
14216 (make_pass_diagnose_omp_blocks): ...new function.
14217 (pass_expand_omp): Replace declaration with that of...
14218 (make_pass_expand_omp): ...new function.
14219 (pass_expand_omp_ssa): Replace declaration with that of...
14220 (make_pass_expand_omp_ssa): ...new function.
14221 (pass_object_sizes): Replace declaration with that of...
14222 (make_pass_object_sizes): ...new function.
14223 (pass_strlen): Replace declaration with that of...
14224 (make_pass_strlen): ...new function.
14225 (pass_fold_builtins): Replace declaration with that of...
14226 (make_pass_fold_builtins): ...new function.
14227 (pass_stdarg): Replace declaration with that of...
14228 (make_pass_stdarg): ...new function.
14229 (pass_early_warn_uninitialized): Replace declaration with that of...
14230 (make_pass_early_warn_uninitialized): ...new function.
14231 (pass_late_warn_uninitialized): Replace declaration with that of...
14232 (make_pass_late_warn_uninitialized): ...new function.
14233 (pass_cse_reciprocals): Replace declaration with that of...
14234 (make_pass_cse_reciprocals): ...new function.
14235 (pass_cse_sincos): Replace declaration with that of...
14236 (make_pass_cse_sincos): ...new function.
14237 (pass_optimize_bswap): Replace declaration with that of...
14238 (make_pass_optimize_bswap): ...new function.
14239 (pass_optimize_widening_mul): Replace declaration with that of...
14240 (make_pass_optimize_widening_mul): ...new function.
14241 (pass_warn_function_return): Replace declaration with that of...
14242 (make_pass_warn_function_return): ...new function.
14243 (pass_warn_function_noreturn): Replace declaration with that of...
14244 (make_pass_warn_function_noreturn): ...new function.
14245 (pass_cselim): Replace declaration with that of...
14246 (make_pass_cselim): ...new function.
14247 (pass_phiopt): Replace declaration with that of...
14248 (make_pass_phiopt): ...new function.
14249 (pass_forwprop): Replace declaration with that of...
14250 (make_pass_forwprop): ...new function.
14251 (pass_phiprop): Replace declaration with that of...
14252 (make_pass_phiprop): ...new function.
14253 (pass_tree_ifcombine): Replace declaration with that of...
14254 (make_pass_tree_ifcombine): ...new function.
14255 (pass_dse): Replace declaration with that of...
14256 (make_pass_dse): ...new function.
14257 (pass_nrv): Replace declaration with that of...
14258 (make_pass_nrv): ...new function.
14259 (pass_rename_ssa_copies): Replace declaration with that of...
14260 (make_pass_rename_ssa_copies): ...new function.
14261 (pass_sink_code): Replace declaration with that of...
14262 (make_pass_sink_code): ...new function.
14263 (pass_fre): Replace declaration with that of...
14264 (make_pass_fre): ...new function.
14265 (pass_check_data_deps): Replace declaration with that of...
14266 (make_pass_check_data_deps): ...new function.
14267 (pass_copy_prop): Replace declaration with that of...
14268 (make_pass_copy_prop): ...new function.
14269 (pass_vrp): Replace declaration with that of...
14270 (make_pass_vrp): ...new function.
14271 (pass_uncprop): Replace declaration with that of...
14272 (make_pass_uncprop): ...new function.
14273 (pass_return_slot): Replace declaration with that of...
14274 (make_pass_return_slot): ...new function.
14275 (pass_reassoc): Replace declaration with that of...
14276 (make_pass_reassoc): ...new function.
14277 (pass_rebuild_cgraph_edges): Replace declaration with that of...
14278 (make_pass_rebuild_cgraph_edges): ...new function.
14279 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
14280 (make_pass_remove_cgraph_callee_edges): ...new function.
14281 (pass_build_cgraph_edges): Replace declaration with that of...
14282 (make_pass_build_cgraph_edges): ...new function.
14283 (pass_local_pure_const): Replace declaration with that of...
14284 (make_pass_local_pure_const): ...new function.
14285 (pass_tracer): Replace declaration with that of...
14286 (make_pass_tracer): ...new function.
14287 (pass_warn_unused_result): Replace declaration with that of...
14288 (make_pass_warn_unused_result): ...new function.
14289 (pass_diagnose_tm_blocks): Replace declaration with that of...
14290 (make_pass_diagnose_tm_blocks): ...new function.
14291 (pass_lower_tm): Replace declaration with that of...
14292 (make_pass_lower_tm): ...new function.
14293 (pass_tm_init): Replace declaration with that of...
14294 (make_pass_tm_init): ...new function.
14295 (pass_tm_mark): Replace declaration with that of...
14296 (make_pass_tm_mark): ...new function.
14297 (pass_tm_memopt): Replace declaration with that of...
14298 (make_pass_tm_memopt): ...new function.
14299 (pass_tm_edges): Replace declaration with that of...
14300 (make_pass_tm_edges): ...new function.
14301 (pass_split_functions): Replace declaration with that of...
14302 (make_pass_split_functions): ...new function.
14303 (pass_feedback_split_functions): Replace declaration with that of...
14304 (make_pass_feedback_split_functions): ...new function.
14305 (pass_strength_reduction): Replace declaration with that of...
14306 (make_pass_strength_reduction): ...new function.
14307 (pass_ipa_lower_emutls): Replace declaration with that of...
14308 (make_pass_ipa_lower_emutls): ...new function.
14309 (pass_ipa_function_and_variable_visibility): Replace declaration with
14311 (make_pass_ipa_function_and_variable_visibility): ...new function.
14312 (pass_ipa_tree_profile): Replace declaration with that of...
14313 (make_pass_ipa_tree_profile): ...new function.
14314 (pass_early_local_passes): Replace declaration with that of...
14315 (make_pass_early_local_passes): ...new function.
14316 (pass_ipa_whole_program_visibility): Replace declaration with that
14318 (make_pass_ipa_whole_program_visibility): ...new function.
14319 (pass_ipa_lto_gimple_out): Replace declaration with that of...
14320 (make_pass_ipa_lto_gimple_out): ...new function.
14321 (pass_ipa_increase_alignment): Replace declaration with that of...
14322 (make_pass_ipa_increase_alignment): ...new function.
14323 (pass_ipa_inline): Replace declaration with that of...
14324 (make_pass_ipa_inline): ...new function.
14325 (pass_ipa_free_lang_data): Replace declaration with that of...
14326 (make_pass_ipa_free_lang_data): ...new function.
14327 (pass_ipa_free_inline_summary): Replace declaration with that of...
14328 (make_pass_ipa_free_inline_summary): ...new function.
14329 (pass_ipa_cp): Replace declaration with that of...
14330 (make_pass_ipa_cp): ...new function.
14331 (pass_ipa_reference): Replace declaration with that of...
14332 (make_pass_ipa_reference): ...new function.
14333 (pass_ipa_pure_const): Replace declaration with that of...
14334 (make_pass_ipa_pure_const): ...new function.
14335 (pass_ipa_pta): Replace declaration with that of...
14336 (make_pass_ipa_pta): ...new function.
14337 (pass_ipa_lto_finish_out): Replace declaration with that of...
14338 (make_pass_ipa_lto_finish_out): ...new function.
14339 (pass_ipa_tm): Replace declaration with that of...
14340 (make_pass_ipa_tm): ...new function.
14341 (pass_ipa_profile): Replace declaration with that of...
14342 (make_pass_ipa_profile): ...new function.
14343 (pass_ipa_cdtor_merge): Replace declaration with that of...
14344 (make_pass_ipa_cdtor_merge): ...new function.
14345 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
14347 (make_pass_cleanup_cfg_post_optimizing): ...new function.
14348 (pass_init_datastructures): Replace declaration with that of...
14349 (make_pass_init_datastructures): ...new function.
14350 (pass_fixup_cfg): Replace declaration with that of...
14351 (make_pass_fixup_cfg): ...new function.
14352 (pass_expand): Replace declaration with that of...
14353 (make_pass_expand): ...new function.
14354 (pass_instantiate_virtual_regs): Replace declaration with that of...
14355 (make_pass_instantiate_virtual_regs): ...new function.
14356 (pass_rtl_fwprop): Replace declaration with that of...
14357 (make_pass_rtl_fwprop): ...new function.
14358 (pass_rtl_fwprop_addr): Replace declaration with that of...
14359 (make_pass_rtl_fwprop_addr): ...new function.
14360 (pass_jump): Replace declaration with that of...
14361 (make_pass_jump): ...new function.
14362 (pass_jump2): Replace declaration with that of...
14363 (make_pass_jump2): ...new function.
14364 (pass_lower_subreg): Replace declaration with that of...
14365 (make_pass_lower_subreg): ...new function.
14366 (pass_cse): Replace declaration with that of...
14367 (make_pass_cse): ...new function.
14368 (pass_fast_rtl_dce): Replace declaration with that of...
14369 (make_pass_fast_rtl_dce): ...new function.
14370 (pass_ud_rtl_dce): Replace declaration with that of...
14371 (make_pass_ud_rtl_dce): ...new function.
14372 (pass_rtl_dce): Replace declaration with that of...
14373 (make_pass_rtl_dce): ...new function.
14374 (pass_rtl_dse1): Replace declaration with that of...
14375 (make_pass_rtl_dse1): ...new function.
14376 (pass_rtl_dse2): Replace declaration with that of...
14377 (make_pass_rtl_dse2): ...new function.
14378 (pass_rtl_dse3): Replace declaration with that of...
14379 (make_pass_rtl_dse3): ...new function.
14380 (pass_rtl_cprop): Replace declaration with that of...
14381 (make_pass_rtl_cprop): ...new function.
14382 (pass_rtl_pre): Replace declaration with that of...
14383 (make_pass_rtl_pre): ...new function.
14384 (pass_rtl_hoist): Replace declaration with that of...
14385 (make_pass_rtl_hoist): ...new function.
14386 (pass_rtl_store_motion): Replace declaration with that of...
14387 (make_pass_rtl_store_motion): ...new function.
14388 (pass_cse_after_global_opts): Replace declaration with that of...
14389 (make_pass_cse_after_global_opts): ...new function.
14390 (pass_rtl_ifcvt): Replace declaration with that of...
14391 (make_pass_rtl_ifcvt): ...new function.
14392 (pass_into_cfg_layout_mode): Replace declaration with that of...
14393 (make_pass_into_cfg_layout_mode): ...new function.
14394 (pass_outof_cfg_layout_mode): Replace declaration with that of...
14395 (make_pass_outof_cfg_layout_mode): ...new function.
14396 (pass_loop2): Replace declaration with that of...
14397 (make_pass_loop2): ...new function.
14398 (pass_rtl_loop_init): Replace declaration with that of...
14399 (make_pass_rtl_loop_init): ...new function.
14400 (pass_rtl_move_loop_invariants): Replace declaration with that of...
14401 (make_pass_rtl_move_loop_invariants): ...new function.
14402 (pass_rtl_unswitch): Replace declaration with that of...
14403 (make_pass_rtl_unswitch): ...new function.
14404 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
14405 (make_pass_rtl_unroll_and_peel_loops): ...new function.
14406 (pass_rtl_doloop): Replace declaration with that of...
14407 (make_pass_rtl_doloop): ...new function.
14408 (pass_rtl_loop_done): Replace declaration with that of...
14409 (make_pass_rtl_loop_done): ...new function.
14410 (pass_web): Replace declaration with that of...
14411 (make_pass_web): ...new function.
14412 (pass_cse2): Replace declaration with that of...
14413 (make_pass_cse2): ...new function.
14414 (pass_df_initialize_opt): Replace declaration with that of...
14415 (make_pass_df_initialize_opt): ...new function.
14416 (pass_df_initialize_no_opt): Replace declaration with that of...
14417 (make_pass_df_initialize_no_opt): ...new function.
14418 (pass_reginfo_init): Replace declaration with that of...
14419 (make_pass_reginfo_init): ...new function.
14420 (pass_inc_dec): Replace declaration with that of...
14421 (make_pass_inc_dec): ...new function.
14422 (pass_stack_ptr_mod): Replace declaration with that of...
14423 (make_pass_stack_ptr_mod): ...new function.
14424 (pass_initialize_regs): Replace declaration with that of...
14425 (make_pass_initialize_regs): ...new function.
14426 (pass_combine): Replace declaration with that of...
14427 (make_pass_combine): ...new function.
14428 (pass_if_after_combine): Replace declaration with that of...
14429 (make_pass_if_after_combine): ...new function.
14430 (pass_ree): Replace declaration with that of...
14431 (make_pass_ree): ...new function.
14432 (pass_partition_blocks): Replace declaration with that of...
14433 (make_pass_partition_blocks): ...new function.
14434 (pass_match_asm_constraints): Replace declaration with that of...
14435 (make_pass_match_asm_constraints): ...new function.
14436 (pass_regmove): Replace declaration with that of...
14437 (make_pass_regmove): ...new function.
14438 (pass_split_all_insns): Replace declaration with that of...
14439 (make_pass_split_all_insns): ...new function.
14440 (pass_fast_rtl_byte_dce): Replace declaration with that of...
14441 (make_pass_fast_rtl_byte_dce): ...new function.
14442 (pass_lower_subreg2): Replace declaration with that of...
14443 (make_pass_lower_subreg2): ...new function.
14444 (pass_mode_switching): Replace declaration with that of...
14445 (make_pass_mode_switching): ...new function.
14446 (pass_sms): Replace declaration with that of...
14447 (make_pass_sms): ...new function.
14448 (pass_sched): Replace declaration with that of...
14449 (make_pass_sched): ...new function.
14450 (pass_ira): Replace declaration with that of...
14451 (make_pass_ira): ...new function.
14452 (pass_reload): Replace declaration with that of...
14453 (make_pass_reload): ...new function.
14454 (pass_clean_state): Replace declaration with that of...
14455 (make_pass_clean_state): ...new function.
14456 (pass_branch_prob): Replace declaration with that of...
14457 (make_pass_branch_prob): ...new function.
14458 (pass_value_profile_transformations): Replace declaration with that
14460 (make_pass_value_profile_transformations): ...new function.
14461 (pass_postreload_cse): Replace declaration with that of...
14462 (make_pass_postreload_cse): ...new function.
14463 (pass_gcse2): Replace declaration with that of...
14464 (make_pass_gcse2): ...new function.
14465 (pass_split_after_reload): Replace declaration with that of...
14466 (make_pass_split_after_reload): ...new function.
14467 (pass_branch_target_load_optimize1): Replace declaration with that
14469 (make_pass_branch_target_load_optimize1): ...new function.
14470 (pass_thread_prologue_and_epilogue): Replace declaration with that
14472 (make_pass_thread_prologue_and_epilogue): ...new function.
14473 (pass_stack_adjustments): Replace declaration with that of...
14474 (make_pass_stack_adjustments): ...new function.
14475 (pass_peephole2): Replace declaration with that of...
14476 (make_pass_peephole2): ...new function.
14477 (pass_if_after_reload): Replace declaration with that of...
14478 (make_pass_if_after_reload): ...new function.
14479 (pass_regrename): Replace declaration with that of...
14480 (make_pass_regrename): ...new function.
14481 (pass_cprop_hardreg): Replace declaration with that of...
14482 (make_pass_cprop_hardreg): ...new function.
14483 (pass_reorder_blocks): Replace declaration with that of...
14484 (make_pass_reorder_blocks): ...new function.
14485 (pass_branch_target_load_optimize2): Replace declaration with that
14487 (make_pass_branch_target_load_optimize2): ...new function.
14488 (pass_leaf_regs): Replace declaration with that of...
14489 (make_pass_leaf_regs): ...new function.
14490 (pass_split_before_sched2): Replace declaration with that of...
14491 (make_pass_split_before_sched2): ...new function.
14492 (pass_compare_elim_after_reload): Replace declaration with that of...
14493 (make_pass_compare_elim_after_reload): ...new function.
14494 (pass_sched2): Replace declaration with that of...
14495 (make_pass_sched2): ...new function.
14496 (pass_stack_regs): Replace declaration with that of...
14497 (make_pass_stack_regs): ...new function.
14498 (pass_stack_regs_run): Replace declaration with that of...
14499 (make_pass_stack_regs_run): ...new function.
14500 (pass_df_finish): Replace declaration with that of...
14501 (make_pass_df_finish): ...new function.
14502 (pass_compute_alignments): Replace declaration with that of...
14503 (make_pass_compute_alignments): ...new function.
14504 (pass_duplicate_computed_gotos): Replace declaration with that of...
14505 (make_pass_duplicate_computed_gotos): ...new function.
14506 (pass_variable_tracking): Replace declaration with that of...
14507 (make_pass_variable_tracking): ...new function.
14508 (pass_free_cfg): Replace declaration with that of...
14509 (make_pass_free_cfg): ...new function.
14510 (pass_machine_reorg): Replace declaration with that of...
14511 (make_pass_machine_reorg): ...new function.
14512 (pass_cleanup_barriers): Replace declaration with that of...
14513 (make_pass_cleanup_barriers): ...new function.
14514 (pass_delay_slots): Replace declaration with that of...
14515 (make_pass_delay_slots): ...new function.
14516 (pass_split_for_shorten_branches): Replace declaration with that of...
14517 (make_pass_split_for_shorten_branches): ...new function.
14518 (pass_split_before_regstack): Replace declaration with that of...
14519 (make_pass_split_before_regstack): ...new function.
14520 (pass_convert_to_eh_region_ranges): Replace declaration with that
14522 (make_pass_convert_to_eh_region_ranges): ...new function.
14523 (pass_shorten_branches): Replace declaration with that of...
14524 (make_pass_shorten_branches): ...new function.
14525 (pass_set_nothrow_function_flags): Replace declaration with that of...
14526 (make_pass_set_nothrow_function_flags): ...new function.
14527 (pass_dwarf2_frame): Replace declaration with that of...
14528 (make_pass_dwarf2_frame): ...new function.
14529 (pass_final): Replace declaration with that of...
14530 (make_pass_final): ...new function.
14531 (pass_rtl_seqabstr): Replace declaration with that of...
14532 (make_pass_rtl_seqabstr): ...new function.
14533 (pass_release_ssa_names): Replace declaration with that of...
14534 (make_pass_release_ssa_names): ...new function.
14535 (pass_early_inline): Replace declaration with that of...
14536 (make_pass_early_inline): ...new function.
14537 (pass_inline_parameters): Replace declaration with that of...
14538 (make_pass_inline_parameters): ...new function.
14539 (pass_update_address_taken): Replace declaration with that of...
14540 (make_pass_update_address_taken): ...new function.
14541 (pass_convert_switch): Replace declaration with that of...
14542 (make_pass_convert_switch): ...new function.
14543 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
14544 to a subclass of simple_ipa_opt_pass along with...
14545 (pass_data_ipa_tree_profile): ...new pass_data instance and...
14546 (make_pass_ipa_tree_profile): ...new function.
14547 * tree-sra.c (pass_sra_early): Convert from a global struct to a
14548 subclass of gimple_opt_pass along with...
14549 (pass_data_sra_early): ...new pass_data instance and...
14550 (make_pass_sra_early): ...new function.
14551 (pass_sra): Convert from a global struct to a subclass of
14552 gimple_opt_pass along with...
14553 (pass_data_sra): ...new pass_data instance and...
14554 (make_pass_sra): ...new function.
14555 (pass_early_ipa_sra): Convert from a global struct to a subclass of
14556 gimple_opt_pass along with...
14557 (pass_data_early_ipa_sra): ...new pass_data instance and...
14558 (make_pass_early_ipa_sra): ...new function.
14559 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
14560 subclass of gimple_opt_pass along with...
14561 (pass_data_ccp): ...new pass_data instance and...
14562 (make_pass_ccp): ...new function.
14563 (pass_fold_builtins): Convert from a global struct to a subclass of
14564 gimple_opt_pass along with...
14565 (pass_data_fold_builtins): ...new pass_data instance and...
14566 (make_pass_fold_builtins): ...new function.
14567 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
14568 subclass of gimple_opt_pass along with...
14569 (pass_data_copy_prop): ...new pass_data instance and...
14570 (make_pass_copy_prop): ...new function.
14571 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
14572 global struct to a subclass of gimple_opt_pass along with...
14573 (pass_data_rename_ssa_copies): ...new pass_data instance and...
14574 (make_pass_rename_ssa_copies): ...new function.
14575 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
14576 subclass of gimple_opt_pass along with...
14577 (pass_data_dce): ...new pass_data instance and...
14578 (make_pass_dce): ...new function.
14579 (pass_dce_loop): Convert from a global struct to a subclass of
14580 gimple_opt_pass along with...
14581 (pass_data_dce_loop): ...new pass_data instance and...
14582 (make_pass_dce_loop): ...new function.
14583 (pass_cd_dce): Convert from a global struct to a subclass of
14584 gimple_opt_pass along with...
14585 (pass_data_cd_dce): ...new pass_data instance and...
14586 (make_pass_cd_dce): ...new function.
14587 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
14588 subclass of gimple_opt_pass along with...
14589 (pass_data_dominator): ...new pass_data instance and...
14590 (make_pass_dominator): ...new function.
14591 (pass_phi_only_cprop): Convert from a global struct to a subclass of
14592 gimple_opt_pass along with...
14593 (pass_data_phi_only_cprop): ...new pass_data instance and...
14594 (make_pass_phi_only_cprop): ...new function.
14595 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
14596 subclass of gimple_opt_pass along with...
14597 (pass_data_dse): ...new pass_data instance and...
14598 (make_pass_dse): ...new function.
14599 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
14600 a subclass of gimple_opt_pass along with...
14601 (pass_data_forwprop): ...new pass_data instance and...
14602 (make_pass_forwprop): ...new function.
14603 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
14604 struct to a subclass of gimple_opt_pass along with...
14605 (pass_data_tree_ifcombine): ...new pass_data instance and...
14606 (make_pass_tree_ifcombine): ...new function.
14607 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
14608 subclass of gimple_opt_pass along with...
14609 (pass_data_ch): ...new pass_data instance and...
14610 (make_pass_ch): ...new function.
14611 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
14612 subclass of gimple_opt_pass along with...
14613 (pass_data_tree_loop): ...new pass_data instance and...
14614 (make_pass_tree_loop): ...new function.
14615 (pass_tree_loop_init): Convert from a global struct to a subclass of
14616 gimple_opt_pass along with...
14617 (pass_data_tree_loop_init): ...new pass_data instance and...
14618 (make_pass_tree_loop_init): ...new function.
14619 (pass_lim): Convert from a global struct to a subclass of
14620 gimple_opt_pass along with...
14621 (pass_data_lim): ...new pass_data instance and...
14622 (make_pass_lim): ...new function.
14623 (pass_tree_unswitch): Convert from a global struct to a subclass of
14624 gimple_opt_pass along with...
14625 (pass_data_tree_unswitch): ...new pass_data instance and...
14626 (make_pass_tree_unswitch): ...new function.
14627 (pass_predcom): Convert from a global struct to a subclass of
14628 gimple_opt_pass along with...
14629 (pass_data_predcom): ...new pass_data instance and...
14630 (make_pass_predcom): ...new function.
14631 (pass_vectorize): Convert from a global struct to a subclass of
14632 gimple_opt_pass along with...
14633 (pass_data_vectorize): ...new pass_data instance and...
14634 (make_pass_vectorize): ...new function.
14635 (pass_graphite): Convert from a global struct to a subclass of
14636 gimple_opt_pass along with...
14637 (pass_data_graphite): ...new pass_data instance and...
14638 (make_pass_graphite): ...new function.
14639 (pass_graphite_transforms): Convert from a global struct to a subclass
14640 of gimple_opt_pass along with...
14641 (pass_data_graphite_transforms): ...new pass_data instance and...
14642 (make_pass_graphite_transforms): ...new function.
14643 (pass_check_data_deps): Convert from a global struct to a subclass of
14644 gimple_opt_pass along with...
14645 (pass_data_check_data_deps): ...new pass_data instance and...
14646 (make_pass_check_data_deps): ...new function.
14647 (pass_iv_canon): Convert from a global struct to a subclass of
14648 gimple_opt_pass along with...
14649 (pass_data_iv_canon): ...new pass_data instance and...
14650 (make_pass_iv_canon): ...new function.
14651 (pass_scev_cprop): Convert from a global struct to a subclass of
14652 gimple_opt_pass along with...
14653 (pass_data_scev_cprop): ...new pass_data instance and...
14654 (make_pass_scev_cprop): ...new function.
14655 (pass_record_bounds): Convert from a global struct to a subclass of
14656 gimple_opt_pass along with...
14657 (pass_data_record_bounds): ...new pass_data instance and...
14658 (make_pass_record_bounds): ...new function.
14659 (pass_complete_unroll): Convert from a global struct to a subclass of
14660 gimple_opt_pass along with...
14661 (pass_data_complete_unroll): ...new pass_data instance and...
14662 (make_pass_complete_unroll): ...new function.
14663 (pass_complete_unrolli): Convert from a global struct to a subclass of
14664 gimple_opt_pass along with...
14665 (pass_data_complete_unrolli): ...new pass_data instance and...
14666 (make_pass_complete_unrolli): ...new function.
14667 (pass_parallelize_loops): Convert from a global struct to a subclass
14668 of gimple_opt_pass along with...
14669 (pass_data_parallelize_loops): ...new pass_data instance and...
14670 (make_pass_parallelize_loops): ...new function.
14671 (pass_loop_prefetch): Convert from a global struct to a subclass of
14672 gimple_opt_pass along with...
14673 (pass_data_loop_prefetch): ...new pass_data instance and...
14674 (make_pass_loop_prefetch): ...new function.
14675 (pass_iv_optimize): Convert from a global struct to a subclass of
14676 gimple_opt_pass along with...
14677 (pass_data_iv_optimize): ...new pass_data instance and...
14678 (make_pass_iv_optimize): ...new function.
14679 (pass_tree_loop_done): Convert from a global struct to a subclass of
14680 gimple_opt_pass along with...
14681 (pass_data_tree_loop_done): ...new pass_data instance and...
14682 (make_pass_tree_loop_done): ...new function.
14683 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
14684 struct to a subclass of gimple_opt_pass along with...
14685 (pass_data_cse_reciprocals): ...new pass_data instance and...
14686 (make_pass_cse_reciprocals): ...new function.
14687 (pass_cse_sincos): Convert from a global struct to a subclass of
14688 gimple_opt_pass along with...
14689 (pass_data_cse_sincos): ...new pass_data instance and...
14690 (make_pass_cse_sincos): ...new function.
14691 (pass_optimize_bswap): Convert from a global struct to a subclass of
14692 gimple_opt_pass along with...
14693 (pass_data_optimize_bswap): ...new pass_data instance and...
14694 (make_pass_optimize_bswap): ...new function.
14695 (pass_optimize_widening_mul): Convert from a global struct to a
14696 subclass of gimple_opt_pass along with...
14697 (pass_data_optimize_widening_mul): ...new pass_data instance and...
14698 (make_pass_optimize_widening_mul): ...new function.
14699 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
14700 subclass of gimple_opt_pass along with...
14701 (pass_data_phiopt): ...new pass_data instance and...
14702 (make_pass_phiopt): ...new function.
14703 (pass_cselim): Convert from a global struct to a subclass of
14704 gimple_opt_pass along with...
14705 (pass_data_cselim): ...new pass_data instance and...
14706 (make_pass_cselim): ...new function.
14707 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
14708 subclass of gimple_opt_pass along with...
14709 (pass_data_phiprop): ...new pass_data instance and...
14710 (make_pass_phiprop): ...new function.
14711 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
14712 subclass of gimple_opt_pass along with...
14713 (pass_data_pre): ...new pass_data instance and...
14714 (make_pass_pre): ...new function.
14715 (pass_fre): Convert from a global struct to a subclass of
14716 gimple_opt_pass along with...
14717 (pass_data_fre): ...new pass_data instance and...
14718 (make_pass_fre): ...new function.
14719 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
14720 subclass of gimple_opt_pass along with...
14721 (pass_data_reassoc): ...new pass_data instance and...
14722 (make_pass_reassoc): ...new function.
14723 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
14724 subclass of gimple_opt_pass along with...
14725 (pass_data_sink_code): ...new pass_data instance and...
14726 (make_pass_sink_code): ...new function.
14727 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
14728 subclass of gimple_opt_pass along with...
14729 (pass_data_strlen): ...new pass_data instance and...
14730 (make_pass_strlen): ...new function.
14731 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14732 struct to a subclass of gimple_opt_pass along with...
14733 (pass_data_build_alias): ...new pass_data instance and...
14734 (make_pass_build_alias): ...new function.
14735 (pass_build_ealias): Convert from a global struct to a subclass of
14736 gimple_opt_pass along with...
14737 (pass_data_build_ealias): ...new pass_data instance and...
14738 (make_pass_build_ealias): ...new function.
14739 (pass_ipa_pta): Convert from a global struct to a subclass of
14740 simple_ipa_opt_pass along with...
14741 (pass_data_ipa_pta): ...new pass_data instance and...
14742 (make_pass_ipa_pta): ...new function.
14743 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14744 subclass of gimple_opt_pass along with...
14745 (pass_data_uncprop): ...new pass_data instance and...
14746 (make_pass_uncprop): ...new function.
14747 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14748 global struct to a subclass of gimple_opt_pass along with...
14749 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14750 (make_pass_late_warn_uninitialized): ...new function.
14751 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14752 to a subclass of gimple_opt_pass along with...
14753 (pass_data_init_datastructures): ...new pass_data instance and...
14754 (make_pass_init_datastructures): ...new function.
14755 (pass_early_warn_uninitialized): Convert from a global struct to a
14756 subclass of gimple_opt_pass along with...
14757 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14758 (make_pass_early_warn_uninitialized): ...new function.
14759 (pass_update_address_taken): Convert from a global struct to a
14760 subclass of gimple_opt_pass along with...
14761 (pass_data_update_address_taken): ...new pass_data instance and...
14762 (make_pass_update_address_taken): ...new function.
14763 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14764 struct to a subclass of gimple_opt_pass along with...
14765 (pass_data_release_ssa_names): ...new pass_data instance and...
14766 (make_pass_release_ssa_names): ...new function.
14767 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14768 subclass of gimple_opt_pass along with...
14769 (pass_data_stdarg): ...new pass_data instance and...
14770 (make_pass_stdarg): ...new function.
14771 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14772 global struct to a subclass of gimple_opt_pass along with...
14773 (pass_data_convert_switch): ...new pass_data instance and...
14774 (make_pass_convert_switch): ...new function.
14775 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14776 to a subclass of gimple_opt_pass along with...
14777 (pass_data_tail_recursion): ...new pass_data instance and...
14778 (make_pass_tail_recursion): ...new function.
14779 (pass_tail_calls): Convert from a global struct to a subclass of
14780 gimple_opt_pass along with...
14781 (pass_data_tail_calls): ...new pass_data instance and...
14782 (make_pass_tail_calls): ...new function.
14783 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14784 struct to a subclass of gimple_opt_pass along with...
14785 (pass_data_lower_vector): ...new pass_data instance and...
14786 (make_pass_lower_vector): ...new function.
14787 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14788 gimple_opt_pass along with...
14789 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14790 (make_pass_lower_vector_ssa): ...new function.
14791 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14792 to a subclass of gimple_opt_pass along with...
14793 (pass_data_slp_vectorize): ...new pass_data instance and...
14794 (make_pass_slp_vectorize): ...new function.
14795 (pass_ipa_increase_alignment): Convert from a global struct to a
14796 subclass of simple_ipa_opt_pass along with...
14797 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14798 (make_pass_ipa_increase_alignment): ...new function.
14799 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14800 gimple_opt_pass along with...
14801 (pass_data_vrp): ...new pass_data instance and...
14802 (make_pass_vrp): ...new function.
14803 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14804 subclass of simple_ipa_opt_pass along with...
14805 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14806 (make_pass_ipa_free_lang_data): ...new function.
14807 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14808 gimple_opt_pass along with...
14809 (pass_data_tsan): ...new pass_data instance and...
14810 (make_pass_tsan): ...new function.
14811 (pass_tsan_O0): Convert from a global struct to a subclass of
14812 gimple_opt_pass along with...
14813 (pass_data_tsan_O0): ...new pass_data instance and...
14814 (make_pass_tsan_O0): ...new function.
14815 * var-tracking.c (pass_variable_tracking): Convert from a global
14816 struct to a subclass of rtl_opt_pass along with...
14817 (pass_data_variable_tracking): ...new pass_data instance and...
14818 (make_pass_variable_tracking): ...new function.
14819 * web.c (pass_web): Convert from a global struct to a subclass of
14820 rtl_opt_pass along with...
14821 (pass_data_web): ...new pass_data instance and...
14822 (make_pass_web): ...new function.
14823 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14824 declaration with that of...
14825 (make_pass_mode_switch_use): ...new function.
14826 (pass_resolve_sw_modes): Replace declaration with that of...
14827 (make_pass_resolve_sw_modes): ...new function.
14828 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14829 from a global struct to a subclass of rtl_opt_pass along with...
14830 (pass_data_mode_switch_use): ...new pass_data instance and...
14831 (make_pass_mode_switch_use): ...new function.
14832 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14833 from a global struct to a subclass of rtl_opt_pass along with...
14834 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14835 (make_pass_resolve_sw_modes): ...new function.
14836 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14837 struct to a subclass of rtl_opt_pass along with...
14838 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14839 (make_pass_insert_vzeroupper): ...new function.
14840 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14841 global struct to a subclass of rtl_opt_pass along with...
14842 (pass_data_work_around_errata): ...new pass_data instance and...
14843 (make_pass_work_around_errata): ...new function.
14844 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14845 struct to a subclass of rtl_opt_pass along with...
14846 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14847 (make_pass_mips_machine_reorg2): ...new function.
14849 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14851 * passes.c (pass_manager::operator new): New.
14853 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14855 Handwritten part of conversion of passes to C++ classes.
14857 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14858 (toplev.o): Add dep on PASS_MANAGER_H.
14859 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14860 of early local pases to reflect this moving from a global to a
14861 member of gcc::pass_manager.
14862 (cgraph_add_new_function): Likewise.
14863 * lto-cgraph.c (lto_output_node): Update for conversion of
14864 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14865 * passes.c (opt_pass::clone): New.
14866 (opt_pass::gate): New.
14867 (opt_pass::execute): New.
14868 (opt_pass::opt_pass): New.
14869 (pass_manager::execute_early_local_passes): New.
14870 (pass_manager::execute_pass_mode_switching): new.
14871 (finish_optimization_passes): Convert to...
14872 (pass_manager::finish_optimization_passes): ...this.
14873 (finish_optimization_passes): Update for conversion of passes to
14875 (register_dump_files_1): Use has_gate since we cannot portably
14876 check a vtable entry against NULL.
14877 (dump_one_pass): Likewise.
14878 (ipa_write_summaries_2): Likewise.
14879 (ipa_write_optimization_summaries_1): Likewise.
14880 (ipa_read_summaries_1): Likewise.
14881 (ipa_read_optimization_summaries_1): Likewise.
14882 (execute_ipa_stmt_fixups): Likewise.
14883 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14884 pass-creation functions rather than wiring up globals, and
14885 storing the results in fields of pass_manager generated using
14886 pass-instances.def.
14887 (pass_manager::dump_profile_report): Update for conversion of
14888 passes to C++ classes.
14889 (pass_manager::execute_ipa_summary_passes): Likewise.
14890 (execute_one_ipa_transform_pass): Likewise.
14891 (execute_one_pass): Use has_gate and has_execute since we cannot
14892 portably check a vtable entry against NULL.
14893 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14894 (pass_manager): Use pass-instances.def to add fields for the
14895 various pass instances.
14896 * toplev.c (finalize): Update for move of
14897 finish_optimization_passes to a method of gcc::pass_manager.
14898 * toplev.h (finish_optimization_passes): Move to method of class
14900 * tree-pass.h (struct pass_data): New.
14901 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14902 (opt_pass::gate): Convert to virtual function.
14903 (opt_pass::~opt_pass): New.
14904 (opt_pass::clone): New.
14905 (opt_pass::execute): Convert to virtual function.
14906 (opt_pass::opt_pass): New.
14907 (opt_pass::ctxt_): new.
14908 (gimple_opt_pass): Convert to subclass of opt_pass.
14909 (gimple_opt_pass::gimple_opt_pass): New.
14910 (rtl_opt_pass): Convert to subclass of opt_pass.
14911 (rtl_opt_pass::rtl_opt_pass): New.
14912 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14913 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14914 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14915 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14916 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14917 invocation of pass_mode_switching to reflect this moving from a
14918 global to a member of gcc::pass_manager.
14919 (ix86_option_override): Rework how pass_insert_vzeroupper is
14920 added to the pass_manager to reflect autogenerated changes.
14921 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14923 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14925 PR rtl-optimization/57708
14926 * recog.c (peep2_find_free_register): Validate all regs in a
14929 2013-08-05 Jan Hubicka <jh@suse.cz>
14932 * cgraph.c (verify_cgraph_node): Accept local flags from other
14934 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14935 (function_and_variable_visibility): Likewise.
14936 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14938 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14940 * graph.c (init_graph_slim_pretty_print): Remove.
14941 (print_graph_cfg): Do not call it. Use local pretty printer.
14942 (start_graph_dump): Likewise.
14944 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14946 * gimple-pretty-print.c (buffer): Remove.
14947 (initialized): Likewise.
14948 (maybe_init_pretty_print): Likewise.
14949 (print_gimple_stmt): Do not call it. Use non-static local
14950 pretty_printer variable.
14951 (print_gimple_expr): Likewise.
14952 (print_gimple_seq): Likewise.
14953 (gimple_dump_bb): Likewise.
14955 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14957 * asan.c (asan_pp): Remove.
14958 (asan_pp_initialized): Likewise.
14959 (asan_pp_initialize): Likewise.
14960 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14961 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14962 (asan_add_global): Likewise.
14964 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14966 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14967 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14968 * pretty-print.c (pp_formatted_text_data): Likewise.
14969 (pp_indent): Rename from pp_base_indent.
14970 (pp_format): Rename from pp_base_format.
14971 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14972 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14973 (pp_flush): Rename from pp_base_flush.
14974 (pp_set_line_maximum_length): Rename from
14975 pp_base_set_line_maximum_length.
14976 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14977 (pp_set_prefix): Rename from pp_base_set_prefix.
14978 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14979 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14980 (pp_append_text): Rename from pp_base_append_text.
14981 (pp_formatted_text): Rename from pp_base_formatted_text.
14982 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14983 (pp_remaining_character_count_for_line): Rename from
14984 pp_base_remaining_character_count_for_line.
14985 (pp_newline): Rename from pp_base_newline.
14986 (pp_character): Rename from pp_base_character.
14987 (pp_string): Rename from pp_base_string.
14988 (pp_maybe_space): Rename from pp_base_maybe_space.
14989 * asan.c (asan_pp_string): Adjust.
14990 (asan_emit_stack_protection): Likewise.
14991 (asan_add_global): Likewise.
14992 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14993 * tree-mudflap.c (mf_varname_tree): Likewise.
14994 * tree-pretty-print.c (pp_tree_identifier): Rename from
14995 pp_base_tree_identifier.
14996 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14997 Declare as function.
14999 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
15001 * pretty-print.h (pp_bar_bar): New.
15002 (pp_ampersand_ampersand): Likewise.
15003 (pp_less_equal): Likewise.
15004 (pp_greater_equal): Likewise.
15005 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
15006 printer functions instead of pp_string or operators and punctuators.
15007 (dump_gimple_call): Likewise.
15008 (dump_gimple_omp_for): Likewise.
15009 (dump_gimple_transaction): Likewise.
15010 (dump_gimple_phi): Likewise.
15011 (pp_gimple_stmt_1): Likewise.
15012 * sched-vis.c (print_insn): Likewise.
15013 * tree-mudflap.c (mf_varname_tree): Likewise.
15014 * tree-pretty-print.c (dump_block_node): Likewise.
15015 (dump_generic_node): Likewise.
15017 2013-08-02 Jan Hubicka <jh@suse.cz>
15019 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
15021 * lto-streamer-out.c (tree_is_indexable): Results decls and
15022 parm decls are not indexable.
15023 (DFS_write_tree_body): Do not follow args and results.
15024 (hash_tree): Likewise.
15025 (output_functions): Rearrange so struct function is needed
15026 only when real body is output; be able to also ouptut abstract
15027 functions; output DECL_ARGUMENTS and DECL_RESULT.
15028 (lto_output): When not in WPA, ale store abstract functions.
15029 (write_symbol): Do not care about RESULT_DECL.
15030 (output_symbol_p): Handle correctly sbtract decls.
15031 * lto-streamer-in.c (input_function): Rearrange so struct
15032 function can be NULL at entry; allow streaming of
15033 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
15034 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
15035 sanity check during LTO.
15036 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
15037 RESULT_DECl and DECL_ARGUMENTS.
15038 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
15041 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
15043 * pretty-print.h (pp_underscore): New.
15045 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
15046 printer functions instead of pp_character.
15047 (dump_binary_rhs): Likewise.
15048 (dump_ternary_rhs): Likewise.
15049 (dump_gimple_call_args): Likewise.
15050 (pp_points_to_solution): Likewise.
15051 (dump_gimple_call): Likewise.
15052 (dump_gimple_switch): Likewise.
15053 (dump_gimple_cond): Likewise.
15054 (dump_gimple_bind): Likewise.
15055 (dump_gimple_try): Likewise.
15056 (dump_gimple_omp_for): Likewise.
15057 (dump_gimple_omp_continue): Likewise.
15058 (dump_gimple_omp_single): Likewise.
15059 (dump_gimple_omp_sections): Likewise.
15060 (dump_gimple_omp_block): Likewise.
15061 (dump_gimple_omp_critical): Likewise.
15062 (dump_gimple_transaction): Likewise.
15063 (dump_gimple_asm): Likewise.
15064 (dump_gimple_phi): Likewise.
15065 (dump_gimple_omp_parallel): Likewise.
15066 (dump_gimple_omp_task): Likewise.
15067 (dump_gimple_omp_atomic_load): Likewise.
15068 (dump_gimple_omp_atomic_store): Likewise.
15069 (dump_gimple_mem_ops): Likewise.
15070 (pp_gimple_stmt_1): Likewise.
15071 (pp_cfg_jump): Likewise.
15072 (dump_implicit_edges): Likewise.
15073 (gimple_dump_bb_for_graph): Likewise.
15074 * graph.c (draw_cfg_node): Likewise.
15075 * langhooks.c (lhd_print_error_function): Likewise.
15076 * sched-vis.c (print_exp): Likewise.
15077 (print_value): Likewise.
15078 (print_pattern): Likewise.
15079 (print_insn): Likewise.
15080 (rtl_dump_bb_for_graph): Likewise.
15081 * tree-pretty-print.c (dump_function_declaration): Likewise.
15082 (dump_array_domain): Likewise.
15083 (dump_omp_clause): Likewise.
15084 (dump_location): Likewise.
15085 (dump_generic_node): Likewise.
15086 (print_struct_decl): Likewise.
15087 * diagnostic.c (diagnostic_show_locus): Use pp_space.
15089 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15091 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
15092 candidate table when replacing a candidate statement.
15093 (replace_rhs_if_not_dup): Likewise.
15094 (replace_one_candidate): Likewise.
15096 2013-08-02 Jan Hubicka <jh@suse.cz>
15097 Martin Liska <marxin.liska@gmail.com>
15099 * cgraphunit.c (add_new_function): Fix logic when adding from
15101 (assemble_thunk): Rename to ...
15102 (expand_thunk); .. this one; export; get it working with
15103 general functions; make produced gimple valid.
15104 * cgraph.h (expand_thunk): Declare.
15106 2013-08-02 Jan Hubicka <jh@suse.cz>
15108 * ipa-cp.c (gather_context_independent_values): Use
15109 ipa_get_param_move_cost.
15110 (get_replacement_map): Remove PARAM; move parameter folding
15112 (create_specialized_node): Update.
15113 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
15114 assert that we have gimple body; update move_cost.
15115 (count_formal_params): Assert that we have gimple body.
15116 (ipa_dump_param): New function.
15117 (ipa_alloc_node_params): Break out from ...
15118 (ipa_initialize_node_params): ... here.
15119 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
15120 (ipa_write_node_info): Stream move costs.
15121 (ipa_read_node_info): Read move costs.
15122 (ipa_update_after_lto_read): Do not recompute node params.
15123 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
15124 (ipa_get_param): Check we are not in WPA.
15125 (ipa_get_param_move_cost): New.
15126 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
15127 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
15128 parm numbers to be present.
15130 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
15132 PR rtl-optimization/58048
15133 * lra-constraints.c (process_alt_operands): Don't check asm
15134 operand on register.
15136 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
15138 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
15139 the implied StoreLoad barrier for atomic operations if before.
15141 2013-08-02 Jan Hubicka <jh@suse.cz>
15142 Martin Liska <marxin.liska@gmail.com>
15144 * cgraph.c (cgraph_function_body_availability): Do not check
15146 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
15147 symtab_node_availability): Declare.
15148 * ipa.c (can_replace_by_local_alias): New.
15149 (function_and_variable_visibility): Use it.
15150 * symtab.c (symtab_for_node_and_aliases,
15151 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
15153 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
15155 PR rtl-optimization/57963
15156 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
15157 (lra_constraints): Use them.
15159 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
15161 * config/arm/types.md (define_attr "type"): Add "load_acq"
15163 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
15165 (cortex_a53_store1): Likewise.
15167 2013-08-01 Jan Hubicka <jh@suse.cz>
15169 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
15170 partitions are not needed.
15172 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
15174 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
15175 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
15176 MAYBE_NON_Q_CLASS_P where appropriate.
15178 2013-08-01 Jan Hubicka <jh@suse.cz>
15180 * cgraph.h (release_function_body): Declare.
15181 * tree.c (free_lang_data_in_decl): Free, parameters and return values
15182 of unused delcarations.
15184 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15186 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
15187 RTL form when subtracting a constant.
15189 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15191 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
15192 Generate canonical plus rtx with negated immediate instead of minus
15194 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
15196 2013-08-01 Jan Hubicka <jh@suse.cz>
15198 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
15199 (cgraph_release_function_body): Likewise.
15200 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
15201 * cgraph.h (cgrpah_node): Rename abstract_and_needed
15202 to used_as_abstract_origin.
15203 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
15204 symbols used as abstract origins.
15205 * cgraphunit.c (analyze_functions): Update.
15206 * ipa.c (symtab_remove_unreachable_nodes): Recompute
15207 used_as_abstract_origin.
15208 * tree-inline.c (tree_function_versioning): Update
15209 used_as_abstract_origin; be ready for DECL_RESULT and
15210 DECL_ARGUMENTS to be NULL.
15212 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
15213 for abstract functions.
15214 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
15217 2013-08-01 Jan Hubicka <jh@suse.cz>
15219 * profile.c (compute_value_histograms): Fix thinko.
15221 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15223 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
15224 aarch-common-protos.h to extra_headers.
15225 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15226 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
15227 * config/aarch64/t-aarch64 (aarch-common.o): Define.
15229 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15231 * config/aarch64/aarch64.md (define_attr "type"): Delete.
15232 Include "../arm/types.md". Define "type" attribute for all patterns.
15233 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
15236 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
15238 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
15239 to support power8 load fusion.
15240 (fusion_gpr_mem_load): Likewise.
15242 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
15244 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
15245 declarations for power8 load fusion.
15246 (emit_fusion_gpr_load): Likewise.
15248 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
15249 tuning for power8, turn on fusion mode by default. Turn on sign
15250 extending fusion mode if normal fusion mode is on, and we are at
15252 (fusion_gpr_load_p): New function, return true if we can fuse an
15253 addis instruction with a dependent load to a GPR.
15254 (emit_fusion_gpr_load): Emit the instructions for power8 load
15257 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
15258 (VSX load fusion peepholes): New peepholes to fuse together an
15259 addi instruction with a VSX load instruction.
15261 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
15262 peepholes to fuse an addis instruction with a load to a GPR base
15263 register. If we are supporting sign extending fusions, convert
15264 sign extending loads to zero extending loads and add an explicit
15267 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15269 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
15270 aarch-common-protos.h to extra_headers.
15271 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15272 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
15273 (arm_early_store_addr_dep): Likewise.
15274 (arm_no_early_alu_shift_dep): Likewise.
15275 (arm_no_early_alu_shift_value_dep): Likewise.
15276 (arm_no_early_mul_dep): Likewise.
15277 (arm_no_early_store_addr_dep): Likewise.
15278 (arm_mac_accumulator_is_mul_result): Likewise.
15279 (arm_mac_accumulator_is_result): Likewise.
15280 * config/arm/aarch-common.c: ... here. New file.
15281 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
15283 (arm_early_store_addr_dep): Likewise.
15284 (arm_no_early_alu_shift_dep): Likewise.
15285 (arm_no_early_alu_shift_value_dep): Likewise.
15286 (arm_no_early_mul_dep): Likewise.
15287 (arm_no_early_store_addr_dep): Likewise.
15288 (arm_mac_accumulator_is_mul_result): Likewise.
15289 (arm_mac_accumulator_is_result): Likewise.
15290 * config/arm/aarch-common-protos.h: ... here. New file.
15291 * config/arm/t-arm (aarch-common.o): Define.
15293 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15295 * config/arm/arm.md: Include new file "types.md".
15296 (define_attr "type"): Move from here to ...
15297 (define_attr "mul32"): Likewise.
15298 (define_attr "mul64"): Likewise.
15299 * config/arm/types.md: ... here. New file.
15301 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
15303 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
15304 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
15306 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15308 * gen-pass-instances.awk: Fix offset of substr().
15310 2013-07-31 David Malcolm <dmalcolm@redhat.com>
15312 * Makefile.in (pass-instances.def): New.
15313 (passes.o): Replace dependency on passes.def with one on
15316 * gen-pass-instances.awk: New.
15318 * passes.c (pass_manager::pass_manager): Use pass-instances.def
15319 rather than passes.def, updating local definition of NEXT_PASS
15320 macro to add an extra NUM parameter (currently unused).
15322 2013-07-30 David Malcolm <dmalcolm@redhat.com>
15324 * Makefile.in (PASS_MANAGER_H): New.
15325 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
15326 (passes.o): Likewise.
15327 (statistics.o): Likewise.
15328 (cgraphunit.o): Likewise.
15329 (context.o): Depend on PASS_MANAGER_H.
15331 * pass_manager.h: New.
15333 * cgraphunit.c (cgraph_add_new_function): Update for moves
15334 of globals to fields of pass_manager.
15335 (analyze_function): Likewise.
15336 (expand_function): Likewise.
15337 (ipa_passes): Likewise.
15338 (compile): Likewise.
15340 * context.c (context::context): New.
15341 * context.h (context::context): New.
15342 (context::get_passes): New.
15343 (context::passes_): New.
15345 * lto-cgraph.c (input_node): Update for moves of globals to
15346 fields of pass_manager.
15348 * passes.c (all_passes): Remove, in favor of a field of the
15349 same name within the new class pass_manager.
15350 (all_small_ipa_passes): Likewise.
15351 (all_lowering_passes): Likewise.
15352 (all_regular_ipa_passes): Likewise.
15353 (all_late_ipa_passes): Likewise.
15354 (all_lto_gen_passes): Likewise.
15355 (passes_by_id): Likewise.
15356 (passes_by_id_size): Likewise.
15357 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15358 the new class pass_manager.
15359 (set_pass_for_id): Convert to...
15360 (pass_manager::set_pass_for_id): ...method.
15361 (get_pass_for_id): Convert to...
15362 (pass_manager::get_pass_for_id): ...method.
15363 (register_one_dump_file): Move body of implementation into...
15364 (pass_manager::register_one_dump_file): ...here.
15365 (register_dump_files_1): Convert to...
15366 (pass_manager::register_dump_files_1): ...method.
15367 (register_dump_files): Convert to...
15368 (pass_manager::register_dump_files): ...method.
15369 (create_pass_tab): Update for moves of globals to fields of
15371 (dump_passes): Move body of implementation into...
15372 (pass_manager::dump_passes): ...here.
15373 (register_pass): Move body of implementation into...
15374 (pass_manager::register_pass): ...here.
15375 (init_optimization_passes): Convert into...
15376 (pass_manager::pass_manager): ...constructor for new
15377 pass_manager class, and initialize the pass_lists array.
15378 (check_profile_consistency): Update for moves of globals to
15379 fields of pass_manager.
15380 (dump_profile_report): Move body of implementation into...
15381 (pass_manager::dump_profile_report): ...here.
15382 (ipa_write_summaries_1): Update for moves of pass lists from
15383 being globals to fields of pass_manager.
15384 (ipa_write_optimization_summaries): Likewise.
15385 (ipa_read_summaries): Likewise.
15386 (ipa_read_optimization_summaries): Likewise.
15387 (execute_all_ipa_stmt_fixups): Likewise.
15389 * statistics.c (statistics_fini): Update for moves of globals to
15390 fields of pass_manager.
15392 * toplev.c (general_init): Replace call to
15393 init_optimization_passes with construction of the pass_manager
15396 * tree-pass.h (all_passes): Remove, in favor of a field of the
15397 same name within the new class pass_manager.
15398 (all_small_ipa_passes): Likewise.
15399 (all_lowering_passes): Likewise.
15400 (all_regular_ipa_passes): Likewise.
15401 (all_lto_gen_passes): Likewise.
15402 (all_late_ipa_passes): Likewise.
15403 (passes_by_id): Likewise.
15404 (passes_by_id_size): Likewise.
15405 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15406 the new class pass_manager.
15407 (get_pass_for_id): Remove.
15409 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15411 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
15414 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15416 * arm.md (mulhi3): New expand pattern.
15418 2013-07-30 Jan Hubicka <jh@suse.cz>
15419 Martin Liska <marxin.liska@gmail.com>
15421 * profile.c (compute_value_histograms): Do not ICE when
15422 there is mismatch only on some counters.
15424 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15426 PR rtl-optimization/57637
15427 * function.c (move_insn_for_shrink_wrap): Also check the
15428 GEN set of the LIVE problem for the liveness analysis
15429 if it exists, otherwise give up.
15431 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15433 PR tree-optimization/57993
15434 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
15435 replaced statement in the candidate table.
15436 (phi_add_costs): Return infinite cost when the hidden basis does
15437 not dominate all phis on which the candidate is dependent.
15438 (replace_one_candidate): Record replaced statement in the
15441 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
15443 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
15444 (ashlv2si3): New expander.
15445 (*ashlv2si3_i): New define_insn_and_split.
15446 * predicates.md (float_operation): Allow patterns with three
15447 basic sub-patterns.
15449 PR rtl-optimization/58021
15450 * mode-switching.c (create_pre_exit): Always split off preceding
15451 insns if we are not at the basic block head.
15453 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
15455 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
15456 (UCLIBC_DYNAMIC_LINKER): New macro.
15457 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
15459 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
15460 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
15462 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
15463 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
15464 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
15465 for SF and DF modes. Use ieee_quad_format for TF mode.
15466 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
15467 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
15468 (mips_option_override): Handle `-mnan=legacy'.
15469 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
15470 `-mabs=2008' and `-mnan=2008'.
15471 (OPTION_DEFAULT_SPECS): Add "nan" default.
15472 (ASM_SPEC): Handle `-mnan='.
15473 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
15474 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
15475 comment accordingly.
15476 (neg<mode>2): Likewise.
15477 * config/mips/mips.opt (mabs, mnan): New options.
15478 * doc/install.texi (Configuration): Document `--with-nan=' option.
15479 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
15481 (MIPS Options): Document them.
15482 * config.gcc <mips*-*-*>: Handle `--with-nan='.
15483 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
15484 * configure: Regenerate.
15485 * config.in: Regenerate.
15487 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15489 * config/i386/i386.md (float post-reload splitters): Do not check
15490 for subregs of SSE registers.
15492 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15493 H.J. Lu <hongjiu.lu@intel.com>
15497 * config/i386/i386.md (post-reload splitter
15498 to avoid partial SSE reg dependency stalls): New pattern.
15500 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
15502 * config/s390/s390.md ("movcc"): Swap load and store instructions.
15504 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
15506 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15507 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
15509 2013-07-26 Cary Coutant <ccoutant@google.com>
15511 * dwarf2out.c (die_checksum_ordered): Don't include template
15512 instantiations in signature.
15513 (is_template_parameter): New function.
15514 (is_template_instantiation): New function.
15515 (generate_skeleton_bottom_up): Don't include template instantiations
15517 (generate_skeleton): Likewise.
15518 (break_out_comdat_types): Move recursive call to break out nested
15520 (prune_unused_types_mark_generic_parms_dies): Call
15521 is_template_parameter.
15523 2013-07-26 Ian Bolton <ian.bolton@arm.com>
15525 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
15526 uses vector registers.
15527 * config/aarch64/iterators.md: Add attributes rtn and vas.
15529 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15530 Richard Earnshaw <richard.earnshaw@arm.com>
15532 * combine.c (simplify_comparison): Re-canonicalize operands
15534 * config/arm/arm.md (movcond_addsi): New splitter.
15536 2013-07-25 Sterling Augustine <saugustine@google.com>
15538 * dwarf2out.c (size_of_pubnames): Move code to...
15539 (include_pubname_in_output): ...here. New.
15540 (want_pubnames): Rearrange.
15541 (output_pubnames): Call include_pubname_in_output. Move assertion.
15543 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15545 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
15547 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15550 * doc/extend.texi: Remove obsolete builtins. Fix
15551 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
15553 2013-07-25 Jan Hubicka <jh@suse.cz>
15555 * cgraph.c (release_function_body): Break out from ...
15556 (cgraph_release_function_body): ... this one; also release DECL_RESULT
15557 and DECL_ARGUMENTS.
15558 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
15559 old_tree in the map.
15560 (create_specialized_node): Update.
15561 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
15563 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
15564 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
15565 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
15566 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
15569 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15571 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
15572 addsi3_carryin_alt2_<optab>): Correct output template.
15574 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15576 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
15577 Adjust for arm_restrict_it.
15578 Remove trailing whitespace.
15580 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
15582 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
15583 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
15585 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
15587 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
15589 PR rtl-optimization/57960
15590 * lra-constraints.c (process_alt_operands): Use the right mode
15591 when checking strict_low.
15593 2013-07-25 Jan Hubicka <jh@suse.cz>
15595 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
15596 * cgraph.c (cgraph_remove_node): Do not release function body
15597 when in cgraph streaming.
15598 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
15599 in other partitions are not considered reachable; fix handling of
15602 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15604 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
15606 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15611 * config/arm/arm.md ("*sibcall_insn): Replace use of
15612 Ss with US. Adjust output for v5 and v4t.
15613 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
15615 * config/arm/constraints.md ("Ss"): Rename to US.
15617 2013-07-25 Terry Guo <terry.guo@arm.com>
15619 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
15620 shift_add/shift_sub0/shift_sub1 RTXs.
15622 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
15623 Anton Blanchard <anton@au1.ibm.com>
15625 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
15626 (altivec_vpks<VI_char>ss): Likewise.
15627 (altivec_vpks<VI_char>us): Likewise.
15628 (altivec_vpku<VI_char>us): Likewise.
15629 (altivec_vpku<VI_char>um): Likewise.
15631 2013-07-24 David Malcolm <dmalcolm@redhat.com>
15633 Introduce context class.
15635 * Makefile.in (CONTEXT_H): New.
15636 (OBJS): Add context.o.
15637 (toplev.o): Add CONTEXT_H to dependencies.
15640 * toplev.c (general_init): Create the singleton gcc::context instance.
15646 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
15648 PR rtl-optimization/57968
15649 * mode-switching.c (create_pre_exit): Allow instructions that
15650 don't set a return register to need a non-exit mode.
15652 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15653 Anton Blanchard <anton@au1.ibm.com>
15655 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
15656 operands to vperm for little endian.
15657 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
15658 of lvsl to create the control mask for a vperm for little endian.
15660 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15661 Anton Blanchard <anton@au1.ibm.com>
15663 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
15664 two operands for little-endian.
15666 2013-07-23 Steve Ellcey <sellcey@mips.com>
15668 * config/mips/mips.c (mips_case_values_threshold): New.
15669 (TARGET_CASE_VALUES_THRESHOLD): Define.
15671 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15672 Anton Blanchard <anton@au1.ibm.com>
15674 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
15675 selection of field for vector splat in little endian mode.
15677 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
15679 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
15680 expanders to rs6000.md.
15681 (ior<mode>3): Likewise.
15682 (and<mode>3): Likewise.
15683 (one_cmpl<mode>2): Likewise.
15684 (nor<mode>3): Likewise.
15685 (andc<mode>3): Likewise.
15686 (eqv<mode>3): Likewise.
15687 (nand<mode>3): Likewise.
15688 (orc<mode>3): Likewise.
15690 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
15693 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
15694 to split multi-word logical operations.
15695 (rs6000_split_logical_di): Likewise.
15696 (rs6000_split_logical): Likewise.
15698 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
15699 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
15700 and allow TImode operations in 32-bit.
15701 (vsx_and<mode>3_64bit): Likewise.
15702 (vsx_ior<mode>3_32bit): Likewise.
15703 (vsx_ior<mode>3_64bit): Likewise.
15704 (vsx_xor<mode>3_32bit): Likewise.
15705 (vsx_xor<mode>3_64bit): Likewise.
15706 (vsx_one_cmpl<mode>2_32bit): Likewise.
15707 (vsx_one_cmpl<mode>2_64bit): Likewise.
15708 (vsx_nor<mode>3_32bit): Likewise.
15709 (vsx_nor<mode>3_64bit): Likewise.
15710 (vsx_andc<mode>3_32bit): Likewise.
15711 (vsx_andc<mode>3_64bit): Likewise.
15712 (vsx_eqv<mode>3_32bit): Likewise.
15713 (vsx_eqv<mode>3_64bit): Likewise.
15714 (vsx_nand<mode>3_32bit): Likewise.
15715 (vsx_nand<mode>3_64bit): Likewise.
15716 (vsx_orc<mode>3_32bit): Likewise.
15717 (vsx_orc<mode>3_64bit): Likewise.
15719 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
15720 logical types in GPRs.
15722 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
15723 logical insns to rs6000.md, and allow TImode operations in
15725 (altivec_ior<mode>3): Likewise.
15726 (altivec_xor<mode>3): Likewise.
15727 (altivec_one_cmpl<mode>2): Likewise.
15728 (altivec_nor<mode>3): Likewise.
15729 (altivec_andc<mode>3): Likewise.
15731 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15732 attributes for moving the 128-bit logical operations into
15734 (BOOL_REGS_OUTPUT): Likewise.
15735 (BOOL_REGS_OP1): Likewise.
15736 (BOOL_REGS_OP2): Likewise.
15737 (BOOL_REGS_UNARY): Likewise.
15738 (BOOL_REGS_AND_CR0): Likewise.
15739 (one_cmpl<mode>2): Add support for DI logical operations on
15740 32-bit, splitting the operations to 32-bit.
15741 (anddi3): Likewise.
15742 (iordi3): Likewise.
15743 (xordi3): Likewise.
15744 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15745 changes to combine the 32/64-bit code, allow logical operations on
15746 TI mode in 32-bit, and to use similar match_operator patterns like
15747 scalar mode uses. Combine the Altivec and VSX code for logical
15748 operations, and move it here.
15749 (ior<mode>3, 128-bit types): Likewise.
15750 (xor<mode>3, 128-bit types): Likewise.
15751 (one_cmpl<mode>3, 128-bit types): Likewise.
15752 (nor<mode>3, 128-bit types): Likewise.
15753 (andc<mode>3, 128-bit types): Likewise.
15754 (eqv<mode>3, 128-bit types): Likewise.
15755 (nand<mode>3, 128-bit types): Likewise.
15756 (orc<mode>3, 128-bit types): Likewise.
15757 (and<mode>3_internal): Likewise.
15758 (bool<mode>3_internal): Likewise.
15759 (boolc<mode>3_internal1): Likewise.
15760 (boolc<mode>3_internal2): Likewise.
15761 (boolcc<mode>3_internal1): Likewise.
15762 (boolcc<mode>3_internal2): Likewise.
15763 (eqv<mode>3_internal1): Likewise.
15764 (eqv<mode>3_internal2): Likewise.
15765 (one_cmpl1<mode>3_internal): Likewise.
15767 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15769 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15770 Rename flag_stack_usage to flag_stack_usage_info.
15772 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15774 * config/microblaze/sync.md: New file.
15775 * config/microblaze/microblaze.md: Include sync.md
15776 * config/microblaze/microblaze.c: Add print_operand 'y'.
15777 * config/microblaze/constraints.md: Add memory_contraint
15778 'Q' which is a single register.
15780 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15782 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15784 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15786 * reload.c (find_reloads): Exit loop once we find this operand
15787 cannot be reloaded somehow for this alternative.
15789 * reload.c (find_reloads): Exit loop once we find a hard register.
15791 * rtlanal.c (computed_jump_p): Exit loop once we find label
15794 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15796 * cfgloopmanip.c (remove_path): Exit loop after setting
15799 * gensupport.c (subst_dup): Avoid loop if code is not
15800 MATCH_DUP nor MATCH_OP_DUP.
15802 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15804 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15806 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15808 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15809 true for SP_REGNUM if mode == ptr_mode.
15810 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15811 with value R0_REGNUM + 31.
15813 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15815 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15816 pad pointer-typed argument downward.
15818 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15820 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15821 and __ILP32__ when the ILP32 model is in use.
15823 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15825 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15826 (aarch64_load_symref_appropriately): In the case of
15827 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15828 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15829 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15830 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15831 if 'mode' doesn't equal to 'ptr_mode'.
15832 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15833 'vcall_offset'; change to call aarch64_emit_move differently depending
15834 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15835 to calculate the upper bound of 'vcall_offset'.
15836 (aarch64_cannot_force_const_mem): Change to also return true if
15838 (aarch64_legitimize_reload_address): In the case of large
15839 displacements, add new local variable 'xmode' and an assertion
15840 based on it; change to use 'xmode' to generate the new rtx and
15842 (aarch64_asm_trampoline_template): Change to generate the template
15843 differently depending on TARGET_ILP32 or not; change to use
15844 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15845 (aarch64_trampoline_size): Removed.
15846 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15847 and replace immediate literals with it. Change to use 'ptr_mode'
15848 instead of 'DImode' and call convert_memory_address if the mode
15849 of 'fnaddr' doesn't equal to 'ptr_mode'.
15850 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15852 (aarch64_elf_asm_destructor): Likewise.
15853 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15854 on TARGET_ILP32 instead of aarch64_trampoline_size.
15855 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15856 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15857 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15858 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15859 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15860 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15861 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15862 depending on the value of 'mode'.
15863 (add_losym_<mode>): New.
15864 (ldr_got_small_<mode>): New, based on ldr_got_small.
15865 (ldr_got_small): Remove.
15866 (ldr_got_small_sidi): New.
15867 * config/aarch64/iterators.md (P): New.
15868 (PTR): Change to 'ptr_mode' in the condition.
15870 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15872 * config.gcc (aarch64*-*-*): Support --with-abi.
15873 (aarch64*-*-elf): Support --with-multilib-list.
15874 (aarch64*-*-linux*): Likewise.
15875 (supported_defaults): Add abi to aarch64*-*-*.
15876 * configure.ac: Mention AArch64 for --with-multilib-list.
15877 * configure: Re-generated.
15878 * config/aarch64/biarchilp32.h: New file.
15879 * config/aarch64/biarchlp64.h: New file.
15880 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15882 (MULTILIB_DEFAULTS): Ditto.
15883 (DRIVER_SELF_SPECS): Ditto.
15884 (ASM_SPEC): Update to also substitute -mabi.
15885 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15886 file whose name depends on -mabi= and -mbig-endian.
15887 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15889 (POINTER_SIZE): New define.
15890 (POINTERS_EXTEND_UNSIGNED): Ditto.
15891 (enum aarch64_abi_type): New enumeration tag.
15892 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15893 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15894 (TARGET_ILP32): New define.
15895 * config/aarch64/aarch64.opt (mabi): New.
15896 (aarch64_abi): New.
15897 (ilp32, lp64): New values for -mabi.
15898 * config/aarch64/t-aarch64 (comma): New define.
15899 (MULTILIB_OPTIONS): Ditto.
15900 (MULTILIB_DIRNAMES): Ditto.
15901 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15902 * doc/invoke.texi: Document -mabi for AArch64.
15904 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15906 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15908 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15909 Anton Blanchard <anton@au1.ibm.com>
15911 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15912 endianness when selecting field to splat.
15914 2013-07-22 Eric Christopher <echristo@gmail.com>
15916 * dwarf2out.c (die_odr_checksum): New function to use
15917 CHECKSUM_ macros and ULEB128 for DIE tag.
15918 (generate_type_signature): Use.
15920 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15922 * config.gcc (sparc*-*-*): Accept leon3 processor.
15923 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15924 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15925 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15926 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15927 (mfix-ut699): Adjust comment.
15928 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15929 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15930 (CPP_CPU_SPEC): Likewise.
15931 (ASM_CPU_SPEC): Likewise.
15932 * config/sparc/sparc.c (leon3_cost): New constant.
15933 (sparc_option_override): Add leon3 support.
15934 (mem_ref): New function.
15935 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15936 (sparc_do_work_around_errata): Look into the instruction in the delay
15937 slot and adjust accordingly. Add fix for the data cache nullify issues
15938 of the UT699. Change insertion position for the NOP.
15939 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15940 (leon3_load): New reservation.
15941 (leon_store): Bump latency to 2.
15942 (grfpu): New automaton.
15943 (grfpu_alu): New unit.
15944 (grfpu_ds): Likewise.
15945 (leon_fp_alu): Adjust.
15946 (leon_fp_mult): Delete.
15947 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15948 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15949 * config/sparc/sparc.md (cpu): Add leon3.
15950 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15951 (swapsi): Likewise.
15952 (atomic_test_and_set): Likewise.
15953 (ldstub): Likewise.
15955 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15957 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15958 default for R5900 targets.
15959 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15960 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15961 * config/mips/mips.c (mips_option_override): Report an error for
15962 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15963 for -march=r5900 -mhard-float.
15965 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15967 * df-problems.c (can_move_insns_across): Exit loop once we
15968 find a non-fixed, non-global register.
15970 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15973 * omega.c (omega_eliminate_red): Break after setting red_found.
15974 (omega_problem_has_red_equations): Similarly after setting found.
15975 (omega_query_variable): Similarly after setting coupled.
15977 2013-07-22 Marek Polacek <polacek@redhat.com>
15979 * gimplify.c: Don't include gimple.h twice.
15981 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15983 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15984 instead of TARGET_THUMB1.
15985 (Pz): New constraint.
15986 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15988 (compare_negsi_si): Likewise.
15989 (compare_addsi2_op0): Likewise.
15990 (compare_addsi2_op1): Likewise.
15991 (addsi3_carryin_<optab>): Likewise.
15992 (addsi3_carryin_alt2_<optab>): Likewise.
15993 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15994 for arm_restrict_it.
15995 (subsi3_carryin): Likewise.
15996 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15997 (minmax_arithsi): Disable for arm_restrict_it.
15998 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15999 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
16000 (satsi_<SAT:code>_shift): Likewise.
16001 (arm_shiftsi3): Add alternative for 16-bit encoding.
16002 (arm32_movhf): Disable for arm_restrict_it.
16003 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
16004 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
16006 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
16008 * config/arm/arm.md (attribute "insn"): Delete.
16009 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
16010 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
16011 (not_shiftsi): Update for attribute change.
16012 (not_shiftsi_compare0): Likewise.
16013 (not_shiftsi_compare0_scratch): Likewise.
16014 (arm_one_cmplsi2): Likewise.
16015 (thumb1_one_cmplsi2): Likewise.
16016 (notsi_compare0): Likewise.
16017 (notsi_compare0_scratch): Likewise.
16018 (thumb1_movdi_insn): Likewise.
16019 (arm_movsi_insn): Likewise.
16020 (movhi_insn_arch4): Likewise.
16021 (movhi_bytes): Likewise.
16022 (arm_movqi_insn): Likewise.
16023 (thumb1_movqi_insn): Likewise.
16024 (arm32_movhf): Likewise.
16025 (thumb1_movhf): Likewise.
16026 (arm_movsf_soft_insn): Likewise.
16027 (thumb1_movsf_insn): Likewise.
16028 (thumb_movdf_insn): Likewise.
16029 (movsicc_insn): Likewise.
16030 (movsfcc_soft_insn): Likewise.
16031 (and_scc): Likewise.
16032 (cond_move): Likewise.
16033 (if_move_not): Likewise.
16034 (if_not_move): Likewise.
16035 (if_shift_move): Likewise.
16036 (if_move_shift): Likewise.
16037 (if_shift_shift): Likewise.
16038 (if_not_arith): Likewise.
16039 (if_arith_not): Likewise.
16040 (cond_move_not): Likewise.
16041 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
16042 (neon_mov<mode>): Likewise.
16043 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
16044 (thumb2_movsi_vfp): Likewise.
16045 (movsf_vfp): Likewise.
16046 (thumb2_movsf_vfp): Likewise.
16047 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16049 (cortexa7_older_only): Likewise.
16050 (cortexa7_younger): Likewise.
16051 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
16052 (1020alu_shift_op): Likewise.
16053 (1020alu_shift_reg_op): Likewise.
16054 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
16055 (alu_shift_op): Likewise.
16056 (alu_shift_reg_op): Likewise.
16057 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
16058 (11_alu_shift_op): Likewise.
16059 (11_alu_shift_reg_op): Likewise.
16060 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
16061 (9_alu_shift_reg_op): Likewise.
16062 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
16064 (cortex_a15_alu_shift): Likewise.
16065 (cortex_a15_alu_shift_reg): Likewise.
16066 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
16067 (cortex_a5_alu_shift): Likewise.
16068 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
16070 (cortex_a53_alu_shift): Likewise.
16071 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
16073 (cortex_a7_alu_reg): Likewise.
16074 (cortex_a7_alu_shift): Likewise.
16075 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
16076 (cortex_a8_alu_shift): Likewise.
16077 (cortex_a8_alu_shift_reg): Likewise.
16078 (cortex_a8_mov): Likewise.
16079 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
16080 (cortex_a9_dp_shift): Likewise.
16081 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
16082 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
16083 (cortex_r4_mov): Likewise.
16084 (cortex_r4_alu_shift): Likewise.
16085 (cortex_r4_alu_shift_reg): Likewise.
16086 * config/arm/fa526.md (526_alu_op): Update for attribute change.
16087 (526_alu_shift_op): Likewise.
16088 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
16089 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
16090 (626te_alu_shift_op): Likewise.
16091 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
16092 (726te_alu_op): Likewise.
16093 (726te_alu_shift_op): Likewise.
16094 (726te_alu_shift_reg_op): Likewise.
16095 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
16096 (mp626_alu_shift_op): Likewise.
16097 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
16098 (pj4_alu_e1_conds): Likewise.
16099 (pj4_alu): Likewise.
16100 (pj4_alu_conds): Likewise.
16101 (pj4_shift): Likewise.
16102 (pj4_shift_conds): Likewise.
16103 (pj4_alu_shift): Likewise.
16104 (pj4_alu_shift_conds): Likewise.
16106 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16108 * config/arm/predicates.md (shiftable_operator_strict_it):
16110 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
16111 Disable cond_exec version for arm_restrict_it.
16112 (thumb2_smaxsi3): Convert to generate cond_exec.
16113 (thumb2_sminsi3): Likewise.
16114 (thumb32_umaxsi3): Likewise.
16115 (thumb2_uminsi3): Likewise.
16116 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
16117 (thumb2_neg_abssi2): Likewise.
16118 (thumb2_mov_scc): Add alternative for 16-bit encoding.
16119 (thumb2_movsicc_insn): Adjust alternatives.
16120 (thumb2_mov_negscc): Disable for arm_restrict_it.
16121 (thumb2_mov_negscc_strict_it): New pattern.
16122 (thumb2_mov_notscc_strict_it): New pattern.
16123 (thumb2_mov_notscc): Disable for arm_restrict_it.
16124 (thumb2_ior_scc): Likewise.
16125 (thumb2_ior_scc_strict_it): New pattern.
16126 (thumb2_cond_move): Adjust for arm_restrict_it.
16127 (thumb2_cond_arith): Disable for arm_restrict_it.
16128 (thumb2_cond_arith_strict_it): New pattern.
16129 (thumb2_cond_sub): Adjust for arm_restrict_it.
16130 (thumb2_movcond): Likewise.
16131 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
16132 (thumb2_zero_extendhisi2_v6): Likewise.
16133 (thumb2_zero_extendqisi2_v6): Likewise.
16134 (orsi_notsi_si): Likewise.
16135 (orsi_not_shiftsi_si): Likewise.
16137 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
16139 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
16140 instruction sequence is 1 byte shorter.
16142 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
16144 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
16145 it is not needed after split.
16147 2013-07-20 Iain Sandoe <iain@codesourcery.com>
16150 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
16151 second label for nonlocal goto receivers. Don't output pic base labels
16152 unless we're producing PIC; mark that action unreachable().
16153 (ix86_save_reg): If the function contains a nonlocal label, save the
16155 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
16156 * config/darwin.c (emitted_pic_label_num): New GTY.
16157 (update_pic_label_number_if_needed): New.
16158 (machopic_output_function_base_name): Adjust for nonlocal receiver
16160 (machopic_should_output_picbase_label): New.
16161 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
16162 (nonlocal_goto_receiver): New insn and split.
16164 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
16166 * config/aarch64/aarch64-builtins.c
16167 (aarch64_fold_builtin): Fold abs in all modes.
16168 * config/aarch64/aarch64-simd-builtins.def
16169 (abs): Enable for all modes.
16170 * config/aarch64/arm_neon.h
16171 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
16172 (vabs_f64): Add missing intrinsic.
16174 2013-07-19 Ian Bolton <ian.bolton@arm.com>
16176 * config/aarch64/arm_neon.h (vabs_s64): New function
16178 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
16181 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
16182 * config/avr/avr.md (adjust_len): Add `round'.
16183 * config/avr/avr-protos.h (avr_out_round): New prototype.
16184 (avr_out_plus): Add `out_label' argument.
16185 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
16186 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
16187 Handle the case where `insn' is just a pattern.
16188 (avr_out_bitop): Handle the case where `insn' is just a pattern.
16189 (avr_out_round): New function.
16190 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
16192 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
16194 * config/microblaze/microblaze.c (microblaze_expand_prologue):
16195 Add check for flag_stack_usage to handle -fstack-usage support
16197 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
16199 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
16200 interaction for new Power8 flags and VSX.
16202 2013-07-18 Sriraman Tallam <tmsriram@google.com>
16204 PR middle-end/57698
16205 * tree-inline.c (expand_call_inline): Emit errors during
16206 early_inlining only if optimization is not turned on.
16208 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16212 * passes.c (init_optimization_passes): Move the construction of
16213 the pass hierarchy into a new passes.def file.
16215 * Makefile.in (passes.o): Add dependency on passes.def.
16217 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16219 * passes.c (init_optimization_passes): Introduce macros for
16220 constructing the tree of passes (INSERT_PASSES_AFTER,
16221 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
16222 TERMINATE_PASS_LIST).
16224 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
16225 Wei Mi <wmi@google.com>
16227 PR rtl-optimization/57878
16228 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
16230 (reload_pseudo_compare_func): Check nregs first for reload
16233 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16235 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
16237 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
16239 * read-rtl.c (validate_const_int): Once an invalid character is
16240 seen, quit the loop.
16242 * gengtype.c (write_roots): Similarly once we find the "deletable"
16243 or "if_marked" option.
16245 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16247 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
16248 "xtab" and "sat". Move value "clz" from here to ...
16249 (attriubte "type"): ... here.
16250 (satsi_<SAT:code>): Delete "insn" attribute.
16251 (satsi_<SAT:code>_shift): Likewise.
16252 (arm_zero_extendqisi2addsi): Likewise.
16253 (arm_extendqisi2addsi): Likewise.
16254 (clzsi2): Update for attribute changes.
16255 (rbitsi2): Likewise.
16256 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
16258 (arm_usatsihi): Likewise.
16259 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
16261 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16263 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
16264 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
16265 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
16266 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
16267 in alphabetical order.
16268 (attribute "core_cycles"): Update for attribute changes.
16269 (arm_addsi3): Likewise.
16270 (addsi3_compare0): Likewise.
16271 (addsi3_compare0_scratch): Likewise.
16272 (addsi3_compare_op1): Likewise.
16273 (addsi3_compare_op2): Likewise.
16274 (compare_addsi2_op0): Likewise.
16275 (compare_addsi2_op1): Likewise.
16276 (addsi3_carryin_shift_<optab>): Likewise.
16277 (subsi3_carryin_shift): Likewise.
16278 (rsbsi3_carryin_shift): Likewise.
16279 (arm_subsi3_insn): Likewise.
16280 (subsi3_compare0): Likewise.
16281 (subsi3_compare): Likewise.
16282 (arm_andsi3_insn): Likewise.
16283 (thumb1_andsi3_insn): Likewise.
16284 (andsi3_compare0): Likewise.
16285 (andsi3_compare0_scratch): Likewise.
16286 (zeroextractsi_compare0_scratch
16287 (andsi_not_shiftsi_si): Likewise.
16288 (iorsi3_insn): Likewise.
16289 (iorsi3_compare0): Likewise.
16290 (iorsi3_compare0_scratch): Likewise.
16291 (arm_xorsi3): Likewise.
16292 (thumb1_xorsi3_insn): Likewise.
16293 (xorsi3_compare0): Likewise.
16294 (xorsi3_compare0_scratch): Likewise.
16295 (satsi_<SAT:code>_shift): Likewise.
16297 (arm_shiftsi3): Likewise.
16298 (shiftsi3_compare0): Likewise.
16299 (not_shiftsi): Likewise.
16300 (not_shiftsi_compare0): Likewise.
16301 (not_shiftsi_compare0_scratch): Likewise.
16302 (arm_one_cmplsi2): Likewise.
16303 (thumb_one_complsi2): Likewise.
16304 (notsi_compare0): Likewise.
16305 (notsi_compare0_scratch): Likewise.
16306 (thumb1_zero_extendhisi2): Likewise.
16307 (arm_zero_extendhisi2): Likewise.
16308 (arm_zero_extendhisi2_v6): Likewise.
16309 (arm_zero_extendhisi2addsi): Likewise.
16310 (thumb1_zero_extendqisi2): Likewise.
16311 (thumb1_zero_extendqisi2_v6): Likewise.
16312 (arm_zero_extendqisi2): Likewise.
16313 (arm_zero_extendqisi2_v6): Likewise.
16314 (arm_zero_extendqisi2addsi): Likewise.
16315 (thumb1_extendhisi2): Likewise.
16316 (arm_extendhisi2): Likewise.
16317 (arm_extendhisi2_v6): Likewise.
16318 (arm_extendqisi): Likewise.
16319 (arm_extendqisi_v6): Likewise.
16320 (arm_extendqisi2addsi): Likewise.
16321 (thumb1_extendqisi2): Likewise.
16322 (thumb1_movdi_insn): Likewise.
16323 (arm_movsi_insn): Likewise.
16324 (movsi_compare0): Likewise.
16325 (movhi_insn_arch4): Likewise.
16326 (movhi_bytes): Likewise.
16327 (arm_movqi_insn): Likewise.
16328 (thumb1_movqi_insn): Likewise.
16329 (arm32_movhf): Likewise.
16330 (thumb1_movhf): Likewise.
16331 (arm_movsf_soft_insn): Likewise.
16332 (thumb1_movsf_insn): Likewise.
16333 (movdf_soft_insn): Likewise.
16334 (thumb_movdf_insn): Likewise.
16335 (arm_cmpsi_insn): Likewise.
16336 (cmpsi_shiftsi): Likewise.
16337 (cmpsi_shiftsi_swp): Likewise.
16338 (arm_cmpsi_negshiftsi_si): Likewise.
16339 (movsicc_insn): Likewise.
16340 (movsfcc_soft_insn): Likewise.
16341 (arith_shiftsi): Likewise.
16342 (arith_shiftsi_compare0
16343 (arith_shiftsi_compare0_scratch
16344 (sub_shiftsi): Likewise.
16345 (sub_shiftsi_compare0
16346 (sub_shiftsi_compare0_scratch
16347 (and_scc): Likewise.
16348 (cond_move): Likewise.
16349 (if_plus_move): Likewise.
16350 (if_move_plus): Likewise.
16351 (if_move_not): Likewise.
16352 (if_not_move): Likewise.
16353 (if_shift_move): Likewise.
16354 (if_move_shift): Likewise.
16355 (if_shift_shift): Likewise.
16356 (if_not_arith): Likewise.
16357 (if_arith_not): Likewise.
16358 (cond_move_not): Likewise.
16359 (thumb1_ashlsi3): Set type attribute.
16360 (thumb1_ashrsi3): Likewise.
16361 (thumb1_lshrsi3): Likewise.
16362 (thumb1_rotrsi3): Likewise.
16363 (shiftsi3_compare0_scratch): Likewise.
16364 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
16365 (neon_mov<mode>): Likewise.
16366 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
16368 (thumb2_movsi_insn): Likewise.
16369 (thumb2_cmpsi_neg_shiftsi): Likewise.
16370 (thumb2_extendqisi_v6): Likewise.
16371 (thumb2_zero_extendhisi2_v6): Likewise.
16372 (thumb2_zero_extendqisi2_v6): Likewise.
16373 (thumb2_shiftsi3_short): Likewise.
16374 (thumb2_addsi3_compare0_scratch): Likewise.
16375 (orsi_not_shiftsi_si): Likewise.
16376 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
16377 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
16379 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
16380 (1020alu_shift_op): Likewise.
16381 (1020alu_shift_reg_op): Likewise.
16382 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
16383 (alu_shift_op): Likewise.
16384 (alu_shift_reg_op): Likewise.
16385 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
16386 (11_alu_shift_op): Likewise.
16387 (11_alu_shift_reg_op): Likewise.
16388 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
16389 (9_alu_shift_reg_op): Likewise.
16390 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
16392 (cortex_a15_alu_shift): Likewise.
16393 (cortex_a15_alu_shift_reg): Likewise.
16394 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
16396 (cortex_a5_alu_shift): Likewise.
16397 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
16399 (cortex_a53_alu_shift): Likewise.
16400 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
16402 (cortex_a7_alu_reg): Likewise.
16403 (cortex_a7_alu_shift): Likewise.
16404 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
16406 (cortex_a8_alu_shift): Likewise.
16407 (cortex_a8_alu_shift_reg): Likewise.
16408 (cortex_a8_mov): Likewise.
16409 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
16410 (cortex_a9_dp_shift): Likewise.
16411 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
16413 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
16415 (cortex_r4_mov): Likewise.
16416 (cortex_r4_alu_shift): Likewise.
16417 (cortex_r4_alu_shift_reg): Likewise.
16418 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
16419 (526_alu_shift_op): Likewise.
16420 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
16421 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
16422 (626te_alu_shift_op): Likewise.
16423 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
16424 (726te_alu_op): Likewise.
16425 (726te_alu_shift_op): Likewise.
16426 (726te_alu_shift_reg_op): Likewise.
16427 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
16428 (mp626_alu_shift_op): Likewise.
16429 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
16430 (pj4_alu_e1_conds): Likewise.
16431 (pj4_alu): Likewise.
16432 (pj4_alu_conds): Likewise.
16433 (pj4_shift): Likewise.
16434 (pj4_shift_conds): Likewise.
16435 (pj4_alu_shift): Likewise.
16436 (pj4_alu_shift_conds): Likewise.
16437 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16439 (cortexa7_older_only): Likewise.
16440 (cortexa7_younger): Likewise.
16442 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16444 * ipa-pure-const.c (generate_summary): Rename to...
16445 (pure_const_generate_summary): ... this.
16447 2013-07-17 Iain Sandoe <iain@codesourcery.com>
16449 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
16451 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
16454 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
16457 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16459 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
16460 enabled without -march=zEC12.
16461 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
16464 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16466 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
16467 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
16468 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
16469 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
16470 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
16471 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
16472 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
16474 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
16475 and ISA_HAS_NMADD3_NMSUB3.
16476 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
16477 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
16478 (nmsub4<mode>, nmsub3<mode>): Likewise.
16479 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
16481 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16483 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
16484 TARGET_MIPS5400 checking.
16486 2013-07-16 Jakub Jelinek <jakub@redhat.com>
16487 Peter Bergner <bergner@vnet.ibm.com>
16489 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
16490 registers in the comment.
16491 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
16492 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
16493 rather than FIRST_PSEUDO_REGISTERS.
16495 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
16497 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
16498 enable extra ISA flags with TARGET_HTM.
16500 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16502 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
16505 2013-07-15 Cong Hou <congh@google.com>
16507 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
16508 in compare function for sorting.
16510 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
16512 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
16513 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
16514 * config/rs6000/rs6000.opt: Add -mhtm option.
16515 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
16516 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
16517 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16518 __HTM__ if the HTM instructions are available.
16519 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
16520 htm_spr_reg_operand): New define_predicates.
16521 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
16522 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
16524 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
16525 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
16526 HTM builtin functions.
16527 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
16528 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
16529 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
16530 (rs6000_builtin_mask_calculate): Likewise.
16531 (rs6000_option_override_internal): Likewise.
16532 (bdesc_htm): Add new HTM builtin support.
16533 (htm_spr_num): New function.
16534 (htm_spr_regno): Likewise.
16535 (rs6000_htm_spr_icode): Likewise.
16536 (htm_expand_builtin): Likewise.
16537 (htm_init_builtins): Likewise.
16538 (rs6000_expand_builtin): Add support for HTM builtin functions.
16539 (rs6000_init_builtins): Likewise.
16540 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
16542 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
16543 (TARGET_HTM, MASK_HTM): Define macros.
16544 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
16545 (FIXED_REGISTERS): Likewise.
16546 (CALL_USED_REGISTERS): Likewise.
16547 (CALL_REALLY_USED_REGISTERS): Likewise.
16548 (REG_ALLOC_ORDER): Likewise.
16549 (enum reg_class): Likewise.
16550 (REG_CLASS_NAMES): Likewise.
16551 (REG_CLASS_CONTENTS): Likewise.
16552 (REGISTER_NAMES): Likewise.
16553 (ADDITIONAL_REGISTER_NAMES): Likewise.
16554 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
16555 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
16556 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
16557 * config/rs6000/htm.md: New file.
16558 * config/rs6000/htmintrin.h: New file.
16559 * config/rs6000/htmxlintrin.h: New file.
16561 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
16563 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16564 Define SYMBOL_TINY_GOT, update comment.
16565 * config/aarch64/aarch64.c
16566 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
16567 (aarch64_expand_mov_immediate): Likewise.
16568 (aarch64_print_operand): Likewise.
16569 (aarch64_classify_symbol): Likewise.
16570 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
16571 (ldr_got_tiny): Define.
16573 2013-07-13 Tobias Grosser <tobias@grosser.es>
16575 PR tree-optimization/54094
16576 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
16577 scheduling dimension for the parallelism check from the polyhedral
16578 information in the AST.
16579 * graphite-dependences.c (carries_deps): Do not assume the schedule is
16582 2013-07-13 Jason Merrill <jason@redhat.com>
16584 * print-tree.c (debug_vec_tree): Use debug_raw.
16585 (debug_raw (vec<tree, va_gc> &)): New.
16586 (debug_raw (vec<tree, va_gc> *)): New.
16587 * tree.h: Declare them.
16589 2013-07-13 Bin Cheng <bin.cheng@arm.com>
16591 * ifcvt.c (ifcvt_after_combine): New static variable.
16592 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
16594 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
16595 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
16596 rest_of_handle_if_after_reload): Pass new argument for if_convert.
16598 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
16600 * config/mips/mips.c (mips_expand_call): Remove empty statement.
16602 2013-07-12 Michael Matz <matz@suse.de>
16604 PR middle-end/55771
16605 * convert.c (convert_to_real): Reject non-float inner types.
16607 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
16609 * config/aarch64/aarch64-protos.h
16610 (aarch64_simd_immediate_valid_for_move): Remove.
16611 * config/aarch64/aarch64.c (simd_immediate_info): New member.
16612 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
16614 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
16616 2013-07-11 Steve Ellcey <sellcey@mips.com>
16618 * config/mips/mips.c (mips_conditional_register_usage): Do not
16619 use t[0-7] registers in MIPS16 mode when optimizing for size.
16621 2013-07-11 Sriraman Tallam <tmsriram@google.com>
16623 * config/i386/i386.c (dispatch_function_versions): Fix array
16624 indexing of function_version_info to match actual_versions.
16626 2013-07-11 Teresa Johnson <tejohnson@google.com>
16628 * vec.h (struct va_gc): Move release out-of-line.
16629 (va_gc::release): Call ggc_free on released vec.
16631 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16633 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
16634 Require GOT register as additional operand in UNSPEC.
16635 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
16636 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16637 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16638 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
16639 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
16640 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
16641 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
16643 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
16646 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
16647 name seen by assembler/linker rather if available.
16649 2013-07-11 Andreas Schwab <schwab@suse.de>
16651 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
16653 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
16655 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
16657 2013-07-10 Joseph Myers <joseph@codesourcery.com>
16659 * doc/tm.texi.in: Move hook documentation to ....
16660 * target.def: ... here.
16662 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
16663 text on @hook line.
16664 * doc/tm.texi: Regenerate.
16666 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
16669 * doc/invoke.texi: Document -Wconditionally-supported.
16671 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16674 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
16677 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16680 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
16681 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
16682 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
16683 Remove duplicate devices.
16684 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
16685 * config/avr/t-multilib: Regenerate.
16686 * config/avr/avr-tables.opt: Regenerate.
16687 * doc/avr-mmcu.texi: Regenerate.
16689 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16692 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
16694 2013-07-10 Graham Stott <graham.stott@btinternet.com>
16696 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
16697 the cost of MULT when optimizing for size.
16699 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16701 * config/cr16/cr16-protos.h: Don't include target.h.
16703 2013-07-09 Joseph Myers <joseph@codesourcery.com>
16705 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
16706 adjust register size for TDmode and TFmode for VSX registers.
16708 2013-07-08 Kai Tietz <ktietz@redhat.com>
16711 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
16712 hook_bool_const_tree_true.
16714 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16716 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
16717 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
16718 * config/s390/s390.md: Define FPR*_REGNUM constants.
16719 Fix FPR2_REGNUM constant (18 -> 17).
16720 ("*trunc<BFP:mode><DFP_ALL:mode>2")
16721 ("*trunc<DFP_ALL:mode><BFP:mode>2")
16722 ("trunc<BFP:mode><DFP_ALL:mode>2")
16723 ("trunc<DFP_ALL:mode><BFP:mode>2")
16724 ("*extend<BFP:mode><DFP_ALL:mode>2")
16725 ("*extend<DFP_ALL:mode><BFP:mode>2")
16726 ("extend<BFP:mode><DFP_ALL:mode>2")
16727 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16730 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16732 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16734 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16736 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16737 and cfun_fpr_bit_p to cfun_fpr_save_p.
16738 (s390_frame_area, s390_register_info, s390_frame_info)
16739 (s390_emit_prologue, s390_emit_epilogue)
16740 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16742 * config/s390/s390.h: Define *_REGNUM macros for floating point
16745 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16747 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16749 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16751 PR rtl-optimization/57786
16752 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16753 and break out of the loop when it is set to false.
16755 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16758 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16759 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16760 (const_int 63)) 0)).
16761 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16762 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16763 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16765 PR rtl-optimization/57829
16766 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16767 mask bits outside of mode are just sign-extension from mode to HWI.
16769 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16771 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16772 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16773 adjust_address instead of change_address to keep info about alignment.
16774 (emit_strmov): Remove.
16775 (emit_memmov): New function.
16776 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16777 (expand_movmem_epilogue): Likewise and return updated rtx for
16779 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16780 destination and source.
16781 (decide_alignment): Refactor, handle vector_loop.
16782 (ix86_expand_movmem): Likewise.
16783 (ix86_expand_setmem): Likewise.
16784 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16786 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16788 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16789 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16791 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16793 * config/i386/sse.md (sse_movlhps): Change alternative 3
16794 of operand 2 to "m".
16796 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16799 * config/i386/sse.md (iptr): New mode attribute.
16800 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16801 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16802 (<sse>_vmmul<mode>3): Ditto.
16803 (<sse>_vmdiv<mode>3): Ditto.
16804 (sse_vmrcpv4sf2): Ditto.
16805 (<sse>_vmsqrt<mode>2): Ditto.
16806 (sse_vmrsqrtv4sf2): Ditto.
16807 (<sse>_vm<code><mode>3): Ditto.
16808 (avx_vmcmp<mode>3): Ditto.
16809 (<sse>_vmmaskcmp<mode>3): Ditto.
16810 (<sse>_comi): Ditto.
16811 (<sse>_ucomi): Ditto.
16812 (*xop_vmfrcz_<mode>): Ditto.
16813 (*fmai_fmadd_<mode>): Ditto.
16814 (*fmai_fmsub_<mode>): Ditto.
16815 (*fmai_fnmadd_<mode>): Ditto.
16816 (*fmai_fnmsub_<mode>): Ditto.
16817 (*fma4i_vmfmadd_<mode>): Ditto.
16818 (*fma4i_vmfmsub_<mode>): Ditto.
16819 (*fma4i_vmfnmadd_<mode>): Ditto.
16820 (*fma4i_vmfnmsub_<mode>): Ditto.
16821 (*xop_vmfrcz_<mode>): Ditto.
16822 (sse_cvtps2pi): Ditto.
16823 (sse_cvttps2pi): Ditto.
16824 (sse_cvtss2si): Ditto.
16825 (sse_cvtss2si_2): Ditto.
16826 (sse_cvtss2siq_2): Ditto.
16827 (sse_cvttss2si): Ditto.
16828 (sse_cvttss2siq): Ditto.
16829 (sse_cvtsd2si): Ditto.
16830 (sse_cvtsd2si_2): Ditto.
16831 (sse_cvtsd2siq_2): Ditto.
16832 (sse_cvttsd2si): Ditto.
16833 (sse_cvttsd2siq): Ditto.
16834 (sse_cvtsd2ss): Ditto.
16835 (sse_cvtss2sd): Ditto.
16836 (avx2_pbroadcast<mode>): Ditto.
16837 (avx2_pbroadcast<mode>_1): Ditto.
16838 (*avx_vperm_broadcast_v4sf): Ditto.
16840 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16841 (sse_movlhps): Ditto.
16842 (sse_storehps): Ditto.
16843 (sse_loadhps): Ditto.
16844 (sse_storelps): Ditto.
16845 (sse_loadlps): Ditto.
16846 (*vec_concatv4sf): Ditto.
16847 (*vec_interleave_highv2df): Ditto.
16848 (*vec_interleave_lowv2df): Ditto.
16849 (*vec_extractv2df_1_sse): Ditto.
16850 (*vec_extractv2df_0_sse): Ditto.
16851 (sse2_storelpd): Ditto.
16852 (sse2_loadlpd): Ditto.
16853 (sse2_movsd): Ditto.
16854 (*vec_concatv4si): Ditto.
16855 (vec_concatv2di): Ditto.
16857 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16858 for Intel asm dialect.
16859 (mmx_punpcklwd): Ditto.
16860 (mmx_punpckldq): Ditto.
16862 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16863 for intel assembler dialect.
16865 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16868 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16869 for BUILT_IN_C{LZ,LRSB}*.
16870 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16871 * tree-vrp.c (extract_range_basic): Handle
16872 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16873 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16874 fall thru to code calling set_value*.
16875 * builtins.c (expand_builtin): Remove *IMAX cases.
16876 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16877 if width is bigger than 2*HWI.
16879 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16881 PR rtl-optimization/55342
16882 * lra-int.h (lra_subreg_reload_pseudos): New.
16883 * lra.c: Add undoing optional reloads to the block diagram.
16884 (lra_subreg_reload_pseudos): New.
16885 (lra_optional_reload_pseudos): Change comments.
16886 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16887 lra_optional_reload_pseudos after undo transformations.
16888 * lra-assigns.c (pseudo_prefix_title): New.
16889 (lra_setup_reg_renumber): Use it.
16890 (spill_for): Ditto. Check subreg reload pseudos too.
16891 (assign_by_spills): Consider subreg reload pseudos too.
16892 * lra-constraints.c (simplify_operand_subreg): Use
16893 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16894 (curr_insn_transform): Recognize and do optional reloads.
16895 (undo_optional_reloads): New.
16896 (lra_undo_inheritance): Call undo_optional_reloads.
16898 2013-07-05 Thomas Quinot <quinot@adacore.com>
16900 * tree-complex.c (expand_complex_operations_1): Fix typo.
16902 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16904 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16905 (tune_params): New member 'const vec_costs'.
16906 * config/aarch64/aarch64.c (generic_vector_cost): New.
16907 (generic_tunings): New member 'generic_vector_cost'.
16908 (aarch64_builtin_vectorization_cost): New.
16909 (aarch64_add_stmt_cost): New.
16910 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16911 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16913 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16916 * config/i386/predicates.md (vsib_address_operand): Disallow
16917 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16919 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16921 PR middle-end/55030
16922 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16923 expand_builtin_setjmp_receiver.
16924 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16925 with NULL for the label parameter.
16926 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16927 the frame-pointer. Adjust comments.
16928 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16929 only if LABEL is non-NULL.
16931 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16933 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16934 (ARM_ABI_AAPCS64): Ditto.
16936 (ARM_DEFAULT_ABI): Ditto.
16938 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16940 * config/aarch64/aarch64-builtins.c
16941 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16942 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16944 * config/aarch64/aarch64-simd.md
16945 (aarch64_ld1<VALL:mode>): New.
16946 (aarch64_st1<VALL:mode>): Likewise.
16947 * config/aarch64/arm_neon.h
16948 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16950 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16952 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16954 (ix86_option_override_internal):Turn on all -mavx target flags by
16955 default as they are dependent on AVX anyway.
16957 2013-07-02 Cary Coutant <ccoutant@google.com>
16959 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16960 deterministic hash.
16961 (loc_checksum_ordered): Likewise.
16962 (hash_loc_operands): Remove inline keyword.
16964 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16966 PR tree-optimization/57741
16967 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16968 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16969 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16970 Allow REAL_CST step_exprs if flag_associative_math.
16971 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16973 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16975 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16977 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16979 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16981 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16983 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16985 (iorsi3_insn): Likewise.
16986 (arm_xorsi3): Likewise.
16988 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16990 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16992 (attribute "type"): ... here, and prefix with "wmmx_".
16993 (attribute "core_cycles"): Update for attribute changes.
16994 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16995 (tbcstv4hi): Likewise.
16996 (tbcstv2si): Likewise.
16997 (iwmmxt_iordi3): Likewise.
16998 (iwmmxt_xordi3): Likewise.
16999 (iwmmxt_anddi3): Likewise.
17000 (iwmmxt_nanddi3): Likewise.
17001 (iwmmxt_arm_movdi): Likewise.
17002 (iwmmxt_movsi_insn): Likewise.
17003 (mov<mode>_internal): Likewise.
17004 (and<mode>3_iwmmxt): Likewise.
17005 (ior<mode>3_iwmmxt): Likewise.
17006 (xor<mode>3_iwmmxt): Likewise.
17007 (add<mode>3_iwmmxt): Likewise.
17008 (ssaddv8qi3): Likewise.
17009 (ssaddv4hi3): Likewise.
17010 (ssaddv2si3): Likewise.
17011 (usaddv8qi3): Likewise.
17012 (usaddv4hi3): Likewise.
17013 (usaddv2si3): Likewise.
17014 (sub<mode>3_iwmmxt): Likewise.
17015 (sssubv8qi3): Likewise.
17016 (sssubv4hi3): Likewise.
17017 (sssubv2si3): Likewise.
17018 (ussubv8qi3): Likewise.
17019 (ussubv4hi3): Likewise.
17020 (ussubv2si3): Likewise.
17021 (mulv4hi3_iwmmxt): Likewise.
17022 (smulv4hi3_highpart): Likewise.
17023 (umulv4hi3_highpart): Likewise.
17024 (iwmmxt_wmacs): Likewise.
17025 (iwmmxt_wmacsz): Likewise.
17026 (iwmmxt_wmacu): Likewise.
17027 (iwmmxt_wmacuz): Likewise.
17028 (iwmmxt_clrdi): Likewise.
17029 (iwmmxt_clrv8qi): Likewise.
17030 (iwmmxt_clr4hi): Likewise.
17031 (iwmmxt_clr2si): Likewise.
17032 (iwmmxt_uavgrndv8qi3): Likewise.
17033 (iwmmxt_uavgrndv4hi3): Likewise.
17034 (iwmmxt_uavgv8qi3): Likewise.
17035 (iwmmxt_uavgv4hi3): Likewise.
17036 (iwmmxt_tinsrb): Likewise.
17037 (iwmmxt_tinsrh): Likewise.
17038 (iwmmxt_tinsrw): Likewise.
17039 (iwmmxt_textrmub): Likewise.
17040 (iwmmxt_textrmsb): Likewise.
17041 (iwmmxt_textrmuh): Likewise.
17042 (iwmmxt_textrmsh): Likewise.
17043 (iwmmxt_textrmw): Likewise.
17044 (iwmxxt_wshufh): Likewise.
17045 (eqv8qi3): Likewise.
17046 (eqv4hi3): Likewise.
17047 (eqv2si3): Likewise.
17048 (gtuv8qi3): Likewise.
17049 (gtuv4hi3): Likewise.
17050 (gtuv2si3): Likewise.
17051 (gtv8qi3): Likewise.
17052 (gtv4hi3): Likewise.
17053 (gtv2si3): Likewise.
17054 (smax<mode>3_iwmmxt): Likewise.
17055 (umax<mode>3_iwmmxt): Likewise.
17056 (smin<mode>3_iwmmxt): Likewise.
17057 (umin<mode>3_iwmmxt): Likewise.
17058 (iwmmxt_wpackhss): Likewise.
17059 (iwmmxt_wpackwss): Likewise.
17060 (iwmmxt_wpackdss): Likewise.
17061 (iwmmxt_wpackhus): Likewise.
17062 (iwmmxt_wpackwus): Likewise.
17063 (iwmmxt_wpackdus): Likewise.
17064 (iwmmxt_wunpckihb): Likewise.
17065 (iwmmxt_wunpckihh): Likewise.
17066 (iwmmxt_wunpckihw): Likewise.
17067 (iwmmxt_wunpckilb): Likewise.
17068 (iwmmxt_wunpckilh): Likewise.
17069 (iwmmxt_wunpckilw): Likewise.
17070 (iwmmxt_wunpckehub): Likewise.
17071 (iwmmxt_wunpckehuh): Likewise.
17072 (iwmmxt_wunpckehuw): Likewise.
17073 (iwmmxt_wunpckehsb): Likewise.
17074 (iwmmxt_wunpckehsh): Likewise.
17075 (iwmmxt_wunpckehsw): Likewise.
17076 (iwmmxt_wunpckelub): Likewise.
17077 (iwmmxt_wunpckeluh): Likewise.
17078 (iwmmxt_wunpckeluw): Likewise.
17079 (iwmmxt_wunpckelsb): Likewise.
17080 (iwmmxt_wunpckelsh): Likewise.
17081 (iwmmxt_wunpckelsw): Likewise.
17082 (ror<mode>3): Likewise.
17083 (ashr<mode>3_iwmmxt): Likewise.
17084 (lshr<mode>3_iwmmxt): Likewise.
17085 (ashl<mode>3_iwmmxt): Likewise.
17086 (ror<mode>3_di): Likewise.
17087 (ashr<mode>3_di): Likewise.
17088 (lshr<mode>3_di): Likewise.
17089 (ashl<mode>3_di): Likewise.
17090 (iwmmxt_wmadds): Likewise.
17091 (iwmmxt_wmaddu): Likewise.
17092 (iwmmxt_tmia): Likewise.
17093 (iwmmxt_tmiaph): Likewise.
17094 (iwmmxt_tmiabb): Likewise.
17095 (iwmmxt_tmiatb): Likewise.
17096 (iwmmxt_tmiabt): Likewise.
17097 (iwmmxt_tmiatt): Likewise.
17098 (iwmmxt_tmovmskb): Likewise.
17099 (iwmmxt_tmovmskh): Likewise.
17100 (iwmmxt_tmovmskw): Likewise.
17101 (iwmmxt_waccb): Likewise.
17102 (iwmmxt_wacch): Likewise.
17103 (iwmmxt_waccw): Likewise.
17104 (iwmmxt_waligni): Likewise.
17105 (iwmmxt_walignr): Likewise.
17106 (iwmmxt_walignr0): Likewise.
17107 (iwmmxt_walignr1): Likewise.
17108 (iwmmxt_walignr2): Likewise.
17109 (iwmmxt_walignr3): Likewise.
17110 (iwmmxt_wsadb): Likewise.
17111 (iwmmxt_wsadh): Likewise.
17112 (iwmmxt_wsadbz): Likewise.
17113 (iwmmxt_wsadhz): Likewise.
17114 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
17115 (iwmmxt_wabsdiffb): Likewise.
17116 (iwmmxt_wabsdiffh): Likewise.
17117 (iwmmxt_wabsdiffw): Likewise.
17118 (iwmmxt_waddsubhx): Likewise
17119 (iwmmxt_wsubaddhx): Likewise.
17120 (addc<mode>3): Likewise.
17121 (iwmmxt_avg4): Likewise.
17122 (iwmmxt_avg4r): Likewise.
17123 (iwmmxt_wmaddsx): Likewise.
17124 (iwmmxt_wmaddux): Likewise.
17125 (iwmmxt_wmaddsn): Likewise.
17126 (iwmmxt_wmaddun): Likewise.
17127 (iwmmxt_wmulwsm): Likewise.
17128 (iwmmxt_wmulwum): Likewise.
17129 (iwmmxt_wmulsmr): Likewise.
17130 (iwmmxt_wmulumr): Likewise.
17131 (iwmmxt_wmulwsmr): Likewise.
17132 (iwmmxt_wmulwumr): Likewise.
17133 (iwmmxt_wmulwl): Likewise.
17134 (iwmmxt_wqmulm): Likewise.
17135 (iwmmxt_wqmulwm): Likewise.
17136 (iwmmxt_wqmulmr): Likewise.
17137 (iwmmxt_wqmulwmr): Likewise.
17138 (iwmmxt_waddbhusm): Likewise.
17139 (iwmmxt_waddbhusl): Likewise.
17140 (iwmmxt_wqmiabb): Likewise.
17141 (iwmmxt_wqmiabt): Likewise.
17142 (iwmmxt_wqmiatb): Likewise.
17143 (iwmmxt_wqmiatt): Likewise.
17144 (iwmmxt_wqmiabbn): Likewise.
17145 (iwmmxt_wqmiabtn): Likewise.
17146 (iwmmxt_wqmiatbn): Likewise.
17147 (iwmmxt_wqmiattn): Likewise.
17148 (iwmmxt_wmiabb): Likewise.
17149 (iwmmxt_wmiabt): Likewise.
17150 (iwmmxt_wmiatb): Likewise.
17151 (iwmmxt_wmiatt): Likewise.
17152 (iwmmxt_wmiabbn): Likewise.
17153 (iwmmxt_wmiabtn): Likewise.
17154 (iwmmxt_wmiatbn): Likewise.
17155 (iwmmxt_wmiattn): Likewise.
17156 (iwmmxt_wmiawbb): Likewise.
17157 (iwmmxt_wmiawbt): Likewise.
17158 (iwmmxt_wmiawtb): Likewise.
17159 (iwmmxt_wmiawtt): Likewise.
17160 (iwmmxt_wmiawbbn): Likewise.
17161 (iwmmxt_wmiawbtn): Likewise.
17162 (iwmmxt_wmiawtbn): Likewise.
17163 (iwmmxt_wmiawttn): Likewise.
17164 (iwmmxt_wmerge): Likewise.
17165 (iwmmxt_tandc<mode>3): Likewise.
17166 (iwmmxt_torc<mode>3): Likewise.
17167 (iwmmxt_torvsc<mode>3): Likewise.
17168 (iwmmxt_textrc<mode>3): Likewise.
17169 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
17170 (wmmxt_pack): Likewise.
17171 (wmmxt_mult_c1): Likewise.
17172 (wmmxt_mult_c2): Likewise.
17173 (wmmxt_alu_c1): Likewise.
17174 (wmmxt_alu_c2): Likewise.
17175 (wmmxt_alu_c3): Likewise.
17176 (wmmxt_transfer_c1): Likewise.
17177 (wmmxt_transfer_c2): Likewise.
17178 (wmmxt_transfer_c3): Likewise.
17179 (marvell_f_iwmmxt_wstr): Likewise.
17180 (marvell_f_iwmmxt_wldr): Likewise.
17182 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
17184 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
17186 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17189 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17190 * lra-constraints.c (need_for_split_p): Check call used hard regs
17191 living through calls.
17193 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
17194 call used regs for call insn.
17196 2013-06-28 Jakub Jelinek <jakub@redhat.com>
17199 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
17200 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
17202 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
17204 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
17206 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17208 * lra-constraints.c (need_for_split_p): Check call used hard regs
17209 living through calls.
17211 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
17214 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
17215 to tie with any other modes. Eliminate Altivec vector mode tests,
17216 since these are a subset of ALTIVEC or VSX vector modes. Simplify
17217 code, to return 0 if testing MODE2 for a condition, if we've
17218 already tested MODE1 for the same condition.
17220 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17222 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
17225 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17227 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17228 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
17230 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17232 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
17234 (aarch64_symbolic_constant_p): Remove.
17235 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
17236 static. Fix line length and white space.
17237 (aarch64_symbolic_constant_p): Remove.
17238 * config/aarch64/predicates.md (aarch64_valid_symref):
17239 Use aarch64_classify_symbol_expression.
17241 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17243 * config/arm/constraints.md (Ts): New constraint.
17244 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
17246 (compare_scc): Use "Ts" constraint for operand 0.
17247 (ior_scc_scc): Likewise.
17248 (and_scc_scc): Likewise.
17249 (and_scc_scc_nodom): Likewise.
17250 (ior_scc_scc_cmp): Likewise for operand 7.
17251 (and_scc_scc_cmp): Likewise.
17252 * config/arm/thumb2.md (thumb2_movsi_insn):
17253 Add alternatives for 16-bit encodings.
17254 (thumb2_movhi_insn): Likewise.
17255 (thumb2_movsicc_insn): Likewise.
17256 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
17257 (thumb2_negscc): Use "Ts" constraint.
17258 Move mvn instruction outside cond_exec block.
17259 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
17260 for 16-bit encodings.
17262 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17264 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
17266 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
17267 (mulsi3subsi): Likewise.
17268 (mulsidi3adddi): Likewise.
17269 (mulsidi3_v6): Likewise.
17270 (umulsidi3_v6): Likewise.
17271 (umulsidi3adddi_v6): Likewise.
17272 (smulsi3_highpart_v6): Likewise.
17273 (umulsi3_highpart_v6): Likewise.
17274 (mulhisi3tb): Likewise.
17275 (mulhisi3bt): Likewise.
17276 (mulhisi3tt): Likewise.
17277 (maddhisi4): Likewise.
17278 (maddhisi4tb): Likewise.
17279 (maddhisi4tt): Likewise.
17280 (maddhidi4): Likewise.
17281 (maddhidi4tb): Likewise.
17282 (maddhidi4tt): Likewise.
17283 (zeroextractsi_compare0_scratch): Likewise.
17284 (insv_zero): Likewise.
17285 (insv_t2): Likewise.
17286 (anddi_notzesidi_di): Likewise.
17287 (anddi_notsesidi_di): Likewise.
17288 (andsi_notsi_si): Likewise.
17289 (iordi_zesidi_di): Likewise.
17290 (xordi_zesidi_di): Likewise.
17291 (andsi_iorsi3_notsi): Likewise.
17292 (smax_0): Likewise.
17293 (smax_m1): Likewise.
17294 (smin_0): Likewise.
17295 (not_shiftsi): Likewise.
17296 (unaligned_loadsi): Likewise.
17297 (unaligned_loadhis): Likewise.
17298 (unaligned_loadhiu): Likewise.
17299 (unaligned_storesi): Likewise.
17300 (unaligned_storehi): Likewise.
17301 (extv_reg): Likewise.
17302 (extzv_t2): Likewise.
17303 (divsi3): Likewise.
17304 (udivsi3): Likewise.
17305 (arm_zero_extendhisi2addsi): Likewise.
17306 (arm_zero_extendqisi2addsi): Likewise.
17307 (compareqi_eq0): Likewise.
17308 (arm_extendhisi2_v6): Likewise.
17309 (arm_extendqisi2addsi): Likewise.
17310 (arm_movt): Likewise.
17311 (thumb2_ldrd): Likewise.
17312 (thumb2_ldrd_base): Likewise.
17313 (thumb2_ldrd_base_neg): Likewise.
17314 (thumb2_strd): Likewise.
17315 (thumb2_strd_base): Likewise.
17316 (thumb2_strd_base_neg): Likewise.
17317 (arm_negsi2): Add alternative for 16-bit encoding.
17318 (arm_one_cmplsi2): Likewise.
17320 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17322 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
17323 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
17324 (movdfcc): Likewise.
17325 * config/arm/vfp.md (*thumb2_movsf_vfp):
17326 Disable predication for arm_restrict_it.
17327 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
17328 (*thumb2_movdfcc_vfp): Likewise.
17329 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
17330 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
17331 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
17332 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
17333 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
17334 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
17335 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
17336 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
17337 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
17338 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
17339 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
17340 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
17341 Disable predication for arm_restrict_it.
17343 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
17345 * config/i386/bmiintrin.h (_bextr_u32): New.
17346 (_bextr_u64): Ditto.
17348 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
17350 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
17351 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
17352 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
17353 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
17354 * config/mips/n32-elf.h: ...this new file.
17356 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
17359 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
17360 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
17362 2013-06-27 Catherine Moore <clm@codesourcery.com>
17364 * config/mips/mips-tables.opt: Regenerate.
17365 * config/mips/mips-cpus.def: Add m14ke and m14kec.
17366 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
17367 * doc/invoke.texi: Add -m14kc.
17369 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17372 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
17373 constraints of operand 1 and 2.
17376 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
17377 to match RTL canonicalization. Swap predicates and
17378 constraints of operand 1 and 2.
17380 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
17382 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
17383 Process OP_INOUT regs for splitting too.
17385 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17387 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
17388 decl before the loop, initialize to NULL.
17389 (vectorizable_load): Initialize ptr_incr to NULL.
17391 2013-06-27 Martin Jambor <mjambor@suse.cz>
17394 * ipa-ref.h (ipa_maybe_record_reference): Declare.
17395 * ipa-ref.c (ipa_maybe_record_reference): New function.
17396 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
17397 * ipa-cp.c (create_specialized_node): Record potential references from
17399 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
17401 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17403 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
17404 parameter 'mode' of type 'enum machine_mode mode'; change to pass
17405 'mode' to force_reg.
17406 (aarch64_add_offset): Update calls to aarch64_force_temporary.
17407 (aarch64_expand_mov_immediate): Likewise.
17409 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17411 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
17412 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
17414 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17416 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
17417 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17418 (struct machine_function): Add tbegin_p.
17419 (s390_canonicalize_comparison): Fold CC mode compares to
17420 conditional jump if possible.
17421 (s390_emit_jump): Return the emitted jump.
17422 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
17423 Handle CCRAWmode compares.
17424 (s390_option_override): Default to -mhtm if available.
17425 (s390_reg_clobbered_rtx): Handle floating point regs as well.
17426 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
17427 FPRs instead of df_regs_ever_live_p.
17428 (s390_optimize_nonescaping_tx): New function.
17429 (s390_init_frame_layout): Extend clobbered_regs array to cover
17431 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
17432 (s390_expand_tbegin): New function.
17433 (enum s390_builtin): New enum definition.
17434 (code_for_builtin): New array definition.
17435 (s390_init_builtins): New function.
17436 (s390_expand_builtin): New function.
17437 (TARGET_INIT_BUILTINS): Define.
17438 (TARGET_EXPAND_BUILTIN): Define.
17439 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
17440 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
17441 (s390_alc_comparison): Likewise.
17442 * config/s390/s390-modes.def: Add CCRAWmode.
17443 * config/s390/s390.h (processor_flags): Add PF_TX.
17444 (TARGET_CPU_HTM): Define macro.
17445 (TARGET_HTM): Define macro.
17446 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
17447 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
17448 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17449 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
17450 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
17451 (TBEGIN_MASK, TBEGINC_MASK): New constants.
17452 ("*cc_to_int"): Move up.
17453 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
17454 constants other than 0.
17455 ("*ccraw_to_int"): New insn and splitter definition.
17456 ("tbegin", "tbegin_nofloat", "tbegin_retry")
17457 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
17458 ("tx_assist"): New expander.
17459 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
17460 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
17461 * config/s390/s390.opt: Add -mhtm option.
17462 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
17463 * config/s390/htmxlintrin.h: New file.
17464 * config/s390/htmintrin.h: New file.
17465 * config/s390/s390intrin.h: New file.
17466 * doc/extend.texi: Document htm builtins.
17467 * config.gcc: Add the new header files to extra_headers.
17469 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
17471 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
17472 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
17474 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
17475 Pat Haugen <pthaugen@us.ibm.com>
17476 Peter Bergner <bergner@vnet.ibm.com>
17478 * config/rs6000/power8.md: New.
17479 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
17480 setting for power8 entry.
17481 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
17482 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
17483 test for Power4/Power5 only.
17484 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
17486 (force_new_group): Adjust comment.
17487 * config/rs6000/rs6000.md: Include power8.md.
17489 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
17491 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
17492 * config/arm/arm-protos.h (arm_max_conditional_execute): New
17494 (tune_params): Update comment.
17495 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
17496 (arm_max_conditional_execute): New function.
17497 (thumb2_final_prescan_insn): Use max_insn_skipped and
17498 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
17500 2013-06-25 Jakub Jelinek <jakub@redhat.com>
17502 PR tree-optimization/57705
17503 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
17504 SSA_NAME step, provided that it is not defined inside the loop.
17505 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
17506 (get_initial_def_for_induction): Handle SSA_NAME IV step.
17508 2013-06-25 Martin Jambor <mjambor@suse.cz>
17510 PR middle-end/57670
17511 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
17512 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
17514 (ipa_note_param_call): Initialize member_ptr flag.
17515 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
17516 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
17517 (ipa_write_indirect_edge_info): Stream member_ptr flag.
17518 (ipa_read_indirect_edge_info): Likewise.
17520 2013-06-25 Richard Biener <rguenther@suse.de>
17522 PR middle-end/56977
17523 * passes.c (init_optimization_passes): Move pass_fold_builtins
17524 and pass_dce earlier with -Og.
17526 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
17528 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
17529 <BIT_FIELD_REF>: Remove trailing TAB.
17530 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
17533 2013-06-24 Martin Jambor <mjambor@suse.cz>
17535 PR tree-optimization/57358
17536 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
17537 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
17538 (ipa_analyze_params_uses): Generate pessimistic info when true.
17540 2013-06-24 Martin Jambor <mjambor@suse.cz>
17542 PR tree-optimization/57539
17543 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
17544 global.inlined_to of the new node to it. All callers changed.
17545 * ipa-inline-transform.c (clone_inlined_nodes): New variable
17546 inlining_into, pass it to cgraph_clone_node.
17547 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
17548 ipa_free_edge_args_substructures.
17549 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
17550 rdesc linked list. Do not assert rdesc edges have inlined caller.
17551 Assert we have found an rdesc in the rdesc list.
17553 2013-06-24 Richard Biener <rguenther@suse.de>
17555 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
17556 (pointer_set_lookup): Declare.
17557 (class pointer_map): New template class implementing a
17558 generic pointer to T map.
17559 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
17560 pointer_map<T>::contains, pointer_map<T>::insert,
17561 pointer_map<T>::traverse): New functions.
17562 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
17563 (pointer_set_lookup): New function.
17564 (pointer_set_contains): Use pointer_set_lookup.
17565 (pointer_set_insert): Likewise.
17566 (insert_aux): Remove.
17567 (struct pointer_map_t): Embed a pointer_set_t.
17568 (pointer_map_create): Adjust.
17569 (pointer_map_destroy): Likewise.
17570 (pointer_map_contains): Likewise.
17571 (pointer_map_insert): Likewise.
17572 (pointer_map_traverse): Likewise.
17573 * tree-streamer.h (struct streamer_tree_cache_d): Use a
17574 pointer_map<unsigned> instead of a pointer_map_t.
17575 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
17576 (streamer_tree_cache_lookup): Likewise.
17577 (streamer_tree_cache_create): Likewise.
17578 (streamer_tree_cache_delete): Likewise.
17579 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
17580 pointer_map<unsigned> instead of a pointer_map_t.
17581 (lto_init_tree_ref_encoder): Adjust.
17582 (lto_destroy_tree_ref_encoder): Likewise.
17583 * lto-section-out.c (lto_output_decl_index): Likewise.
17584 (lto_record_function_out_decl_state): Likewise.
17585 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
17587 2013-06-24 Richard Biener <rguenther@suse.de>
17589 PR tree-optimization/57488
17590 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
17592 2013-06-24 Alan Modra <amodra@gmail.com>
17594 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
17595 (gen_easy_altivec_constant): Likewise.
17596 * config/rs6000/predicates.md (easy_vector_constant_add_self,
17597 easy_vector_constant_msb): Likewise.
17599 2013-06-23 Jakub Jelinek <jakub@redhat.com>
17602 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
17603 add missing return true.
17605 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
17608 * config/sh/predicates.md (general_extend_operand): Invoke
17609 general_movsrc_operand for memory operands.
17610 (general_movsrc_operand): Allow reg+reg addressing, do not use
17611 general_operand for memory operands.
17613 2013-06-23 Sriraman Tallam <tmsriram@google.com>
17615 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
17616 when current target options does not apply.
17617 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
17618 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
17619 * config/i386/bmiintrin.h: Pass appropriate target
17620 attributes to header.
17621 * config/i386/mmintrin.h: Ditto.
17622 * config/i386/nmmintrin.h: Ditto.
17623 * config/i386/avx2intrin.h: Ditto.
17624 * config/i386/fxsrintrin.h: Ditto.
17625 * config/i386/tbmintrin.h: Ditto.
17626 * config/i386/xsaveintrin.h: Ditto.
17627 * config/i386/f16cintrin.h: Ditto.
17628 * config/i386/xtestintrin.h: Ditto.
17629 * config/i386/xsaveoptintrin.h: Ditto.
17630 * config/i386/bmi2intrin.h: Ditto.
17631 * config/i386/lzcntintrin.h: Ditto.
17632 * config/i386/smmintrin.h: Ditto.
17633 * config/i386/wmmintrin.h: Ditto.
17634 * config/i386/x86intrin.h: Remove all header include guards.
17635 * config/i386/prfchwintrin.h: Ditto.
17636 * config/i386/pmmintrin.h: Ditto.
17637 * config/i386/tmmintrin.h: Ditto.
17638 * config/i386/xmmintrin.h: Ditto.
17639 * config/i386/popcntintrin.h: Ditto.
17640 * config/i386/rdseedintrin.h: Ditto.
17641 * config/i386/ammintrin.h: Ditto.
17642 * config/i386/emmintrin.h: Ditto.
17643 * config/i386/immintrin.h: Remove all header include guards.
17644 * config/i386/fma4intrin.h: Ditto.
17645 * config/i386/lwpintrin.h: Ditto.
17646 * config/i386/xopintrin.h: Ditto.
17647 * config/i386/ia32intrin.h: Ditto.
17648 * config/i386/avxintrin.h: Ditto.
17649 * config/i386/rtmintrin.h: Ditto.
17650 * config/i386/fmaintrin.h: Ditto.
17651 * config/i386/mm3dnow.h: Ditto.
17653 2013-06-22 Sriraman Tallam <tmsriram@google.com>
17655 * common/config/i386/i386-common.c: Handle LZCNT.
17657 2013-06-22 Andi Kleen <ak@linux.intel.com>
17659 * doc/extend.texi: Use __atomic_store_n instead of
17660 __atomic_store in HLE example.
17662 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
17664 * config/sh/sh.c: Remove <cstdlib> workaround.
17666 2013-06-21 Andi Kleen <ak@linux.intel.com>
17668 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
17670 2013-06-21 Andi Kleen <ak@linux.intel.com>
17672 * doc/extend.texi: Document that __atomic_clear and
17673 __atomic_test_and_set should only be used with bool.
17675 2013-06-20 Jan Hubicka <jh@suse.cz>
17677 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
17678 types_same_for_odr.
17679 * tree.c (decls_same_for_odr): New function.
17680 (same_for_edr): New function.
17681 (types_same_for_odr): New function.
17682 (get_binfo_at_offset): Use it.
17683 * tree.h (types_same_for_odr): Declare.
17685 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
17686 Jason Merrill <jason@redhat.com>
17688 * system.h: Include <cstdlib> as well as <stdlib.h>.
17690 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
17693 * config/i386/i386.c (construct_container): Report error if
17694 long double is used with disabled x87 float returns.
17696 2013-06-20 Jan Hubicka <jh@suse.cz>
17698 * lto-cgraph.c (input_symtab): Do not set cgraph state.
17700 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
17702 PR rtl-optimization/57425
17703 PR rtl-optimization/57569
17704 * alias.c (write_dependence_p): Remove parameters mem_mode and
17705 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
17706 Changed all callers.
17707 (canon_anti_dependence): Get comments and semantics in sync.
17708 Add parameter mem_canonicalized. Changed all callers.
17709 * rtl.h (canon_anti_dependence): Update prototype.
17711 2013-06-20 Richard Biener <rguenther@suse.de>
17713 * data-streamer-in.c (streamer_read_uhwi): Optimize single
17714 byte case, inline streamer_read_uchar and defer section
17717 2013-06-20 Richard Biener <rguenther@suse.de>
17719 PR tree-optimization/57584
17720 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
17721 SSA names into the expanded expression that take part in
17722 abnormal coalescing.
17724 2013-06-19 Sharad Singhai <singhai@google.com>
17726 * gcov.c (print_usage): Handle new option.
17727 (process_args): Ditto.
17728 (get_gcov_intermediate_filename): New function.
17729 (output_intermediate_file): New function.
17730 (output_gcov_file): New function
17731 (generate_results): Handle new option.
17732 (release_function): Relase demangled name.
17733 (read_graph_file): Handle demangled name.
17734 (output_lines): Ditto.
17735 * doc/gcov.texi: Document gcov intermediate format.
17737 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17740 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17741 (lra_emit_add): Use the functions. Add comment about Y as an
17744 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17747 * collect2.c (collect_atexit): New.
17748 (collect_exit): Delete.
17749 (main): Register collect_atexit with atexit.
17750 (collect_wait): Change collect_exit to exit.
17752 * collect2.h (collect_exit): Delete.
17753 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17755 2013-06-19 Wei Mi <wmi@google.com>
17757 PR rtl-optimization/57518
17758 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17759 if regno is used in paradoxical subreg.
17760 (update_equiv_regs): Check pdx_subregs[regno] before
17761 set a reg to be equivalent with a mem.
17763 2013-06-19 Matthias Klose <doko@ubuntu.com>
17766 * file-find.h (find_a_file): Add a mode parameter.
17767 * file-find.c (find_a_file): Likewise.
17768 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17769 with X_OK for the executables.
17770 * collect2.c (main): Call find_a_file with X_OK.
17772 2013-06-19 Steve Ellcey <sellcey@mips.com>
17775 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17776 Use NEXT_INSN instead of next_real_insn.
17778 2013-06-19 Jan Hubicka <jh@suse.cz>
17780 * cgraph.h (const_value_known_p): Replace by ...
17781 (ctor_for_folding): .. this one.
17782 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17783 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17784 * expr.c (expand_expr_real_1): Likewise.
17785 (string_constant): Likewise.
17786 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17787 * ipa.c (process_references): Likewise.
17788 (symtab_remove_unreachable_nodes): Likewise.
17789 * ipa-inline-analysis.c (param_change_prob): Likewise.
17790 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17791 (get_base_constructor): Likwise.
17792 * varpool.c (varpool_remove_node): Likewise.
17793 (varpool_remove_initializer): LIkewise.
17794 (dump_varpool_node): LIkwise.
17795 (const_value_known_p): Rewrite to ...
17796 (ctor_for_folding): ... this one.
17798 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17801 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17802 PERSONALITY in $PATH derived prefixes.
17804 2013-06-19 Jeff Law <law@redhat.com>
17806 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17809 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17810 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17813 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17815 * config/arm/vfp.md: Move VFP instruction classification documentation
17817 * config/arm/arm.md: ... here. Update instruction classification
17820 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17822 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17823 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17824 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17826 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17829 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17830 with NEXT_INSN. Use tablejump_p to check for jump table data
17833 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17836 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17837 that now in C++ the value is correct per the C++ standards.
17839 2013-06-19 Richard Biener <rguenther@suse.de>
17841 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17842 for global context.
17844 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17847 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17850 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17851 with next_active_insn.
17853 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17855 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17856 functions are inlined during failures to flag an error.
17857 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17858 in early inline pass.
17860 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17862 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17865 2013-06-18 Julian Brown <julian@codesourcery.com>
17867 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17868 Permit virtual register pre-reload if !strict.
17869 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17871 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17873 * config/arm/neon.md (movmisalign<mode>): Use
17874 neon_perm_struct_or_reg_operand instead of
17875 neon_struct_or_register_operand.
17876 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17877 neon_permissive_struct_operand instead of neon_struct_operand.
17878 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17879 neon_vector_mem_operand.
17880 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17881 neon_vector_mem_operand.
17882 (neon_permissive_struct_operand): New.
17883 (neon_struct_or_register_operand): Rename to...
17884 (neon_perm_struct_or_reg_operand): This. Adjust call to
17885 neon_vector_mem_operand.
17887 2013-06-18 Richard Biener <rguenther@suse.de>
17889 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17890 * lto-streamer.h: Include pointer-set.h.
17891 (struct lto_decl_slot): Remove.
17892 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17893 Remove next_index entry.
17894 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17895 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17896 (lto_init_tree_ref_encoder): Adjust.
17897 (lto_destroy_tree_ref_encoder): Likewise.
17898 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17899 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17900 (lto_output_decl_index): Adjust.
17901 (lto_new_out_decl_state): Likewise.
17902 (lto_record_function_out_decl_state): Likewise.
17903 * lto-streamer-out.c (copy_function): Likewise.
17905 2013-06-18 Richard Biener <rguenther@suse.de>
17907 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17908 * cgraphunit.c: Include cfgloop.h.
17909 (init_lowered_empty_function): Initialize the loop tree.
17910 (assemble_thunk): Insert new BBs into loops.
17912 2013-06-18 Richard Biener <rguenther@suse.de>
17914 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17915 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17916 the map from cache entry to cache index optional.
17917 (streamer_tree_cache_replace_tree): Adjust accordingly.
17918 (streamer_tree_cache_append): Likewise.
17919 (streamer_tree_cache_delete): Likewise.
17920 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17921 streamer cache map from cache entry to cache index.
17922 * lto-streamer-out.c (create_output_block): Adjust.
17924 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17926 * config/arm/arm.md (attribute "insn"): Move multiplication and
17927 division attributes to...
17928 (attribute "type"): ... here. Remove mult.
17929 (attribute "mul32"): New attribute.
17930 (attribute "mul64"): Add umaal.
17931 (*arm_mulsi3): Update attributes.
17932 (*arm_mulsi3_v6): Likewise.
17933 (*thumb_mulsi3): Likewise.
17934 (*thumb_mulsi3_v6): Likewise.
17935 (*mulsi3_compare0): Likewise.
17936 (*mulsi3_compare0_v6): Likewise.
17937 (*mulsi_compare0_scratch): Likewise.
17938 (*mulsi_compare0_scratch_v6): Likewise.
17939 (*mulsi3addsi): Likewise.
17940 (*mulsi3addsi_v6): Likewise.
17941 (*mulsi3addsi_compare0): Likewise.
17942 (*mulsi3addsi_compare0_v6): Likewise.
17943 (*mulsi3addsi_compare0_scratch): Likewise.
17944 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17945 (*mulsi3subsi): Likewise.
17946 (*mulsidi3adddi): Likewise.
17947 (*mulsi3addsi_v6): Likewise.
17948 (*mulsidi3adddi_v6): Likewise.
17949 (*mulsidi3_nov6): Likewise.
17950 (*mulsidi3_v6): Likewise.
17951 (*umulsidi3_nov6): Likewise.
17952 (*umulsidi3_v6): Likewise.
17953 (*umulsidi3adddi): Likewise.
17954 (*umulsidi3adddi_v6): Likewise.
17955 (*smulsi3_highpart_nov6): Likewise.
17956 (*smulsi3_highpart_v6): Likewise.
17957 (*umulsi3_highpart_nov6): Likewise.
17958 (*umulsi3_highpart_v6): Likewise.
17959 (mulhisi3): Likewise.
17960 (*mulhisi3tb): Likewise.
17961 (*mulhisi3bt): Likewise.
17962 (*mulhisi3tt): Likewise.
17963 (maddhisi4): Likewise.
17964 (*maddhisi4tb): Likewise.
17965 (*maddhisi4tt): Likewise.
17966 (maddhidi4): Likewise.
17967 (*maddhidi4tb): Likewise.
17968 (*maddhidi4tt): Likewise.
17969 (divsi3): Likewise.
17970 (udivsi3): Likewise.
17971 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17972 (thumb2_mulsi_short_compare0): Likewise.
17973 (thumb2_mulsi_short_compare0_scratch): Likewise.
17974 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17975 (1020mult2): Likewise.
17976 (1020mult3): Likewise.
17977 (1020mult4): Likewise.
17978 (1020mult5): Likewise.
17979 (1020mult6): Likewise.
17980 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17982 (cortex_a15_mult64): Likewise.
17983 (cortex_a15_sdiv): Likewise.
17984 (cortex_a15_udiv): Likewise.
17985 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17991 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17992 (pj4_ir_div): Likewise.
17993 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17994 (11_mult2): Likewise.
17995 (11_mult3): Likewise.
17996 (11_mult4): Likewise.
17997 (11_mult5): Likewise.
17998 (11_mult6): Likewise.
17999 (11_mult7): Likewise.
18000 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
18001 (cortex_a8_mla): Likewise.
18002 (cortex_a8_mull): Likewise.
18003 (cortex_a8_smulwy): Likewise.
18004 (cortex_a8_smlald): Likewise.
18005 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
18006 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
18007 (cortex_r4_mul_3): Likewise.
18008 (cortex_r4_mla_4): Likewise.
18009 (cortex_r4_mla_3): Likewise.
18010 (cortex_r4_smlald): Likewise.
18011 (cortex_r4_mull): Likewise.
18012 (cortex_r4_sdiv): Likewise.
18013 (cortex_r4_udiv): Likewise.
18014 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
18015 (cortex_a7_idiv): Likewise.
18016 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
18017 (9_mult2): Likewise.
18018 (9_mult3): Likewise.
18019 (9_mult4): Likewise.
18020 (9_mult5): Likewise.
18021 (9_mult6): Likewise.
18022 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
18023 (cortex_a53_sdiv): Likewise.
18024 (cortex_a53_udiv): Likewise.
18025 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
18026 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
18027 (mp626_mult2): Likewise.
18028 (mp626_mult3): Likewise.
18029 (mp626_mult4): Likewise.
18030 * config/arm/fa526.md (526_mult1): Update attribute change.
18031 (526_mult2): Likewise.
18032 * config/arm/arm-generic.md (mult): Update attribute change.
18033 (mult_ldsched_strongarm): Likewise.
18034 (mult_ldsched): Likewise.
18035 (multi_cycle): Likewise.
18036 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
18037 * config/arm/fa606te.md (606te_mult1): Update attribute change.
18038 (606te_mult2): Likewise.
18039 (606te_mult3): Likewise.
18040 (606te_mult4): Likewise.
18041 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
18042 (cortex_a9_mac16): Likewise.
18043 (cortex_a9_multiply): Likewise.
18044 (cortex_a9_mac): Likewise.
18045 (cortex_a9_multiply_long): Likewise.
18046 * config/arm/fa626te.md (626te_mult1): Update attribute change.
18047 (626te_mult2): Likewise.
18048 (626te_mult3): Likewise.
18049 (626te_mult4): Likewise.
18051 2013-06-18 Richard Biener <rguenther@suse.de>
18054 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
18056 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18059 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
18060 with next_active_insn.
18062 2013-06-18 Alan Modra <amodra@gmail.com>
18064 * config/rs6000/rs6000.h (enum data_align): New.
18065 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
18066 (DATA_ABI_ALIGNMENT): Define.
18067 (CONSTANT_ALIGNMENT): Correct comment.
18068 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
18069 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
18071 2013-06-17 David Malcolm <dmalcolm@redhat.com>
18073 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
18074 ATTRIBUTE_UNUSED marking.
18076 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
18078 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
18079 alternative and update.
18080 (aarch64_dup_lanedi): Delete.
18081 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
18082 * config/aarch64/aarch64-simd-builtins.def: Update.
18084 2013-06-17 Richard Biener <rguenther@suse.de>
18086 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
18087 (lto_input_scc): Declare.
18088 (lto_input_tree_1): Likewise.
18089 (struct lto_stats_d): Add num_tree_bodies_output and
18090 num_pickle_refs_output.
18091 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
18092 (lto_read_tree_1): Split out from ...
18093 (lto_read_tree): ... this.
18094 (lto_input_scc): New function.
18095 (lto_input_tree_1): Split out from ...
18096 (lto_input_tree): ... this. Handle LTO_tree_scc.
18097 (lto_data_in_create): Create the streamer cache without hashes.
18098 * lto-streamer-out.c (create_output_block): Create the streamer
18099 cache with hashes when not doing WPA.
18100 (lto_write_tree_1): Split out from ...
18101 (lto_write_tree): ... this.
18102 (get_symbol_initial_value): New function.
18103 (lto_output_tree_1): Split out from ...
18104 (lto_output_tree): ... this. Write trees as series of SCCs
18105 using a DFS walk via DFS_write_tree.
18106 (struct sccs, struct scc_entry): New types.
18107 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
18108 (DFS_write_tree_body): New function.
18109 (DFS_write_tree): Likewise.
18110 (hash_tree): Likewise.
18111 (scc_entry_compare): Likewise.
18112 (hash_scc): Likewise.
18113 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
18114 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
18115 TREE_CHAIN as regular reference.
18116 (streamer_read_integer_cst): Remove.
18117 (streamer_get_pickled_tree): Adjust.
18118 * tree-streamer-out.c (streamer_write_chain): Disable streaming
18119 of DECL_EXTERNALs in BLOCK_VARS for now.
18120 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
18122 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
18123 Add hash value argument and record that if hashes are recorded
18125 (streamer_tree_cache_insert_1): Adjust.
18126 (streamer_tree_cache_insert): Likewise.
18127 (streamer_tree_cache_insert_at): Rename to ...
18128 (streamer_tree_cache_replace_tree): ... this and adjust.
18129 (streamer_tree_cache_append): Adjust.
18130 (record_common_node): Likewise.
18131 (streamer_tree_cache_create): Add argument whether to
18132 record hash values together with trees.
18133 (streamer_tree_cache_delete): Adjust.
18134 * tree-streamer.h (struct streamer_tree_cache_d): Add
18136 (streamer_read_integer_cst): Remove.
18137 (streamer_tree_cache_insert): Adjust.
18138 (streamer_tree_cache_append): Likewise.
18139 (streamer_tree_cache_insert_at): Rename to ...
18140 (streamer_tree_cache_replace_tree): ... this and adjust.
18141 (streamer_tree_cache_create): Add argument whether to record hashes.
18142 (streamer_tree_cache_get): Rename to ...
18143 (streamer_tree_cache_get_tree): ... this.
18144 (streamer_tree_cache_get_hash): New function.
18145 * tree.c (cache_integer_cst): New function.
18146 * tree.h (cache_integer_cst): Declare.
18147 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
18148 * lto-symtab.c (lto_varpool_replace_node): Only release
18149 DECL_INITIAL of non-prevailing decls.
18150 * varpool.c (varpool_remove_initializer): Do not release
18151 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
18153 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
18155 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
18156 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
18157 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
18158 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
18159 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
18160 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
18161 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
18162 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
18163 instead of TARGET_64BIT.
18164 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
18165 Require ISA_HAS_<D>DIV.
18167 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
18169 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
18170 (mips*-*-linux*): Move default with_llsc setting to where other
18172 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
18174 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
18175 Likewise. Remove default with_tune setting. Move default float
18176 setting to its own block. Handle with_llsc in the same block as above.
18178 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
18180 PR rtl-optimization/57425
18181 PR rtl-optimization/57569
18182 * alias.c (write_dependence_p): Add new parameters mem_mode,
18183 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
18184 Changed all callers.
18185 (canon_anti_dependence): New function.
18186 * cse.c (check_dependence): Use canon_anti_dependence.
18187 * cselib.c (cselib_invalidate_mem): Likewise.
18188 * rtl.h (canon_anti_dependence): Declare.
18190 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
18192 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
18193 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
18194 ".set mips3" for 64-bit targets.
18196 2013-06-15 Dehao Chen <dehao@google.com>
18198 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
18199 * gimple-low.c (gimple_check_call_matching_types): Likewise.
18200 (gimple_check_call_args): Likewise.
18201 * value-prof.c (check_ic_target): Likewise.
18202 * ipa-inline.c (early_inliner): Likewise.
18203 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
18204 * cgraph.c (cgraph_create_edge_1): Likewise.
18205 (cgraph_make_edge_direct): Likewise.
18207 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
18210 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
18211 rs6000_output_move_128bit to handle emitting quad memory
18212 operations. Set attribute length to 8 bytes.
18214 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
18216 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
18218 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
18219 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
18220 (aarch64_<su>mlsl<mode>): Likewise.
18222 2013-06-14 Mike Stump <mikestump@comcast.net>
18224 * Makefile.in (TARGET_H): Add insn-codes.h.
18226 2013-06-14 Alan Modra <amodra@gmail.com>
18228 PR middle-end/57134
18229 PR middle-end/57586
18230 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
18231 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
18232 bitfield expansion when EXPAND_MEMORY.
18233 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
18235 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
18237 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
18238 test for clearing quad memory on 32-bit later.
18240 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18242 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
18243 (fold_negate_expr): Likewise.
18244 (fold_real_zero_addition_p): Handle vectors.
18245 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
18247 2013-06-14 Alan Modra <amodra@gmail.com>
18249 * varasm.c (force_const_mem): Revert 2013-06-07 change.
18251 2013-06-13 Jan Hubicka <jh@suse.cz>
18253 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18254 Local comdats are not externally visible.
18255 * symtab.c (dump_symtab_base): Dump externally visible.
18256 (verify_symtab_base): Verify back links in the symtab hash.
18258 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18260 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
18261 CONVERT_EXPR as equal nodes.
18263 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18265 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
18267 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18269 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
18270 Generalize to complex and vector.
18271 * tree.c (build_all_ones_cst): New function.
18272 * tree.h (build_all_ones_cst): Declare it.
18274 2013-06-13 Alan Modra <amodra@gmail.com>
18276 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
18277 * config/rs6000/rs6000.md (signbittf2): New insn.
18278 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
18279 (abstf2_internal, cmptf_internal2): Likewise.
18280 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
18282 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
18283 Pat Haugen <pthaugen@us.ibm.com>
18284 Peter Bergner <bergner@vnet.ibm.com>
18286 * config/rs6000/rs6000.c (emit_load_locked): Add support for
18287 power8 byte, half-word, and quad-word atomic instructions.
18288 (emit_store_conditional): Likewise.
18289 (rs6000_expand_atomic_compare_and_swap): Likewise.
18290 (rs6000_expand_atomic_op): Likewise.
18292 * config/rs6000/sync.md (larx): Add new modes for power8.
18294 (AINT): New mode iterator to include TImode as well as normal
18295 integer modes on power8.
18296 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
18297 that VSX registers are not considered. Use AINT mode iterator
18298 instead of INT1 to allow inclusion of quad word atomic operations
18300 (load_locked<mode>): Likewise.
18301 (store_conditional<mode>): Likewise.
18302 (atomic_compare_and_swap<mode>): Likewise.
18303 (atomic_exchange<mode>): Likewise.
18304 (atomic_nand<mode>): Likewise.
18305 (atomic_fetch_<fetchop_name><mode>): Likewise.
18306 (atomic_nand_fetch<mode>): Likewise.
18307 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
18309 (ATOMIC): On power8, add QImode, HImode modes.
18310 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
18311 modes that promote to SImode.
18312 (load_lockedti): Convert TImode arguments to PTImode, so that we
18313 get a guaranteed even/odd register pair.
18314 (load_lockedpti): Likewise.
18315 (store_conditionalti): Likewise.
18316 (store_conditionalpti): Likewise.
18318 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
18319 atomic load/store instructions.
18322 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
18324 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
18326 (insn_count): New attribute, with most cases extracted from...
18327 (length): ...here. Redefine most cases in terms of insn_count.
18328 (single_insn): Delete.
18329 (can_delay): Use insn_count to check for single instructions.
18330 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
18331 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
18332 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
18333 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
18334 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
18335 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
18336 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
18337 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
18338 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
18339 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
18340 rather than "length".
18341 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
18342 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
18343 Use "insn_count" rather than "length".
18344 * config/mips/mips-dsp.md
18345 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
18346 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
18349 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
18351 PR tree-optimization/57361
18352 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
18354 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
18356 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
18358 (aarch64_simd_combine<mode>): New instruction expansion.
18359 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
18360 function prototype.
18361 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
18362 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
18364 2013-06-12 Jan Hubicka <jh@suse.cz>
18366 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
18367 decl has when in streaming stage.
18368 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
18369 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
18371 2013-06-12 Roland Stigge <stigge@antcom.de>
18374 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
18376 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18378 PR tree-optimization/57537
18379 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
18380 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
18382 2013-06-12 Richard Biener <rguenther@suse.de>
18384 * data-streamer.h (streamer_write_char_stream): CSE
18385 obs->current_pointer.
18386 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
18387 streamer_write_char_stream manually and optimize the resulting loop.
18388 (streamer_write_hwi_stream): Likewise.
18390 2013-06-12 Jan Hubicka <jh@suse.cz>
18392 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
18393 * cgraph.h (varpool_create_empty_node): Declare.
18394 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
18396 * symtab.c (symtab_unregister_node): Be lax about missin entries
18398 (symtab_get_node): Update comment.
18399 * varpool.c (varpool_create_empty_node): Break out from ...
18400 (varpool_node_for_decl): ... here.
18401 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
18403 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
18405 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
18406 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
18407 part. Use straight-line flow at the end.
18408 <COMPONENT_REF>: Remove superfluous else.
18409 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
18411 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18414 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
18415 target hook even for !TREE_PUBLIC decls. If no resolution info
18416 is available, return false for common and external decls.
18418 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18420 * config/rl78/constraints.md (U): New constraint.
18421 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
18424 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18427 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
18428 to allow returning address to AT_PLATFORM name.
18430 2013-06-11 Jan Hubicka <jh@suse.cz>
18432 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
18433 * cgraph.h (symtab_node_base): Add weakref flag.
18434 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
18435 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
18436 (output_weakrefs): Use weakref flag.
18437 * fold-const.c (simple_operand_p): Handle WEAK.
18438 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
18439 * ipa.c (varpool_externally_visible_p): Drop weakref.
18440 (function_and_variable_visibility): Update comment; fix weakref
18441 sanity checks; do not clear DECL_WEAK on them.
18442 * lto-cgraph.c (lto_output_node): update.
18443 (lto_output_varpool_node): Update.
18444 (input_overwrite_node): Update.
18445 (input_node): Update.
18446 (input_varpool_node): Update.
18447 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
18448 (lto_symtab_merge_symbols): Add sanity check.
18449 (lto_symtab_prevailing_decl): Do not special case weakrefs.
18450 * passes.c (rest_of_decl_compilation): Set static flag, too.
18451 * symtab.c (dump_symtab_base): Dump weakref.
18452 (verify_symtab_base): Sanity check weakrefs.
18453 (symtab_make_decl_local): Remove duplicated code.
18454 (symtab_alias_ultimate_target): Simplify.
18455 * varpool.c (varpool_create_variable_alias): Set weakref flag.
18457 2013-06-11 Tom de Vries <tom@codesourcery.com>
18459 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
18460 sequence_vect == NULL.
18462 2013-06-11 DJ Delorie <dj@redhat.com>
18464 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
18465 (rl78_unwind_word_mode): New.
18467 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18469 * final.c (debug_prefix_maps): Make static.
18471 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18473 * function.c (initial_trampoline): Remove stray copy.
18475 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
18477 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
18479 2013-06-11 Martin Jambor <mjambor@suse.cz>
18481 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
18482 within bounds at the beginning of the function.
18484 2013-06-11 Alan Modra <amodra@gmail.com>
18486 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
18488 (get_named_section): Don't NULL !DECL_P decl.
18490 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
18492 * doc/invoke.texi (core-avx2): Document.
18494 (atom): Updated with MOVBE.
18496 2013-06-11 Richard Biener <rguenther@suse.de>
18498 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
18500 2013-06-11 Anton Blanchard <anton@samba.org>
18502 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
18503 correct shift value in little-endian mode.
18505 2013-06-11 Jakub Jelinek <jakub@redhat.com>
18508 * varasm.c (get_variable_align): Move #endif to the right place.
18510 2013-06-10 Cary Coutant <ccoutant@google.com>
18512 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
18513 for hash so that hash table traversal order is deterministic.
18515 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
18516 Pat Haugen <pthaugen@us.ibm.com>
18517 Peter Bergner <bergner@vnet.ibm.com>
18519 * config/rs6000/vector.md (GPR move splitter): Do not split moves
18520 of vectors in GPRS if they are direct moves or quad word load or
18523 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
18525 (direct_move_p): Likewise.
18526 (quad_load_store_p): Likewise.
18528 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
18529 classes into bins based on the physical register type.
18530 (reg_class_to_reg_type): Likewise.
18531 (IS_STD_REG_TYPE): Likewise.
18532 (IS_FP_VECT_REG_TYPE): Likewise.
18533 (reload_fpr_gpr): Arrays to determine what insn to use if we can
18534 use direct move instructions.
18535 (reload_gpr_vsx): Likewise.
18536 (reload_vsx_gpr): Likewise.
18537 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
18538 information that is a simplification of register classes. Also
18539 precalculate direct move reload helpers.
18540 (direct_move_p): New function to return true if the operation can
18541 be done as a direct move instruciton.
18542 (quad_load_store_p): New function to return true if the operation
18543 is a quad memory operation.
18544 (rs6000_legitimize_address): If quad memory, only allow register
18545 indirect for TImode addresses.
18546 (rs6000_legitimate_address_p): Likewise.
18547 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
18548 (rs6000_reload_register_type): Likewise.
18549 (register_to_reg_type): Return register type.
18550 (rs6000_secondary_reload_simple_move): New helper function for
18551 secondary reload and secondary memory needed to identify anything
18552 that is a simple move, and does not need reloading.
18553 (rs6000_secondary_reload_direct_move): New helper function for
18554 secondary reload to identify cases that can be done with several
18555 instructions via the direct move instructions.
18556 (rs6000_secondary_reload_move): New helper function for secondary
18557 reload to identify moves between register types that can be done.
18558 (rs6000_secondary_reload): Add support for quad memory operations
18559 and for direct move.
18560 (rs6000_secondary_memory_needed): Likewise.
18561 (rs6000_debug_secondary_memory_needed): Change argument names.
18562 (rs6000_output_move_128bit): New function to return the move to
18563 use for 128-bit moves, including knowing about the various
18564 limitations of quad memory operations.
18566 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
18567 memory operations. call rs6000_output_move_128bit for the actual
18568 instruciton(s) to generate.
18569 (vsx_movti_64bit): Likewise.
18571 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
18572 (UNSPEC_P8V_MTVSRWZ): Likewise.
18573 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
18574 (UNSPEC_P8V_MTVSRD): Likewise.
18575 (UNSPEC_P8V_XXPERMDI): Likewise.
18576 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
18577 (UNSPEC_FUSION_GPR): Likewise.
18578 (FMOVE128_GPR): New iterator for direct move.
18579 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
18580 (f32_sv): Likewise.
18581 (f32_dm): Likewise.
18582 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
18583 loads and direct move instructions.
18584 (zero_extendsidi2_lfiwzx): Likewise.
18585 (extendsidi2_lfiwax): Likewise.
18586 (extendsidi2_nocell): Likewise.
18587 (floatsi<mode>2_lfiwax): Likewise.
18588 (lfiwax): Likewise.
18589 (floatunssi<mode>2_lfiwzx): Likewise.
18590 (lfiwzx): Likewise.
18591 (fix_trunc<mode>_stfiwx): Likewise.
18592 (fixuns_trunc<mode>_stfiwx): Likewise.
18593 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
18594 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
18595 (parity<mode>2_cmpb): Set length/type attr.
18596 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
18597 for 'mr.' to fast_compare.
18598 (bpermd_<mode>): Change type attr to popcnt.
18599 (p8_fmrgow_<mode>): New insns for power8 direct move support.
18600 (p8_mtvsrwz_1): Likewise.
18601 (p8_mtvsrwz_2): Likewise.
18602 (reload_fpr_from_gpr<mode>): Likewise.
18603 (p8_mtvsrd_1): Likewise.
18604 (p8_mtvsrd_2): Likewise.
18605 (p8_xxpermdi_<mode>): Likewise.
18606 (reload_vsx_from_gpr<mode>): Likewise.
18607 (reload_vsx_from_gprsf): Likewise.
18608 (p8_mfvsrd_3_<mode>): LIkewise.
18609 (reload_gpr_from_vsx<mode>): Likewise.
18610 (reload_gpr_from_vsxsf): Likewise.
18611 (p8_mfvsrd_4_disf): Likewise.
18612 (multi-word GPR splits): Do not split direct moves or quad memory
18615 2013-06-10 David Malcolm <dmalcolm@redhat.com>
18617 * tree-into-ssa.c (interesting_blocks): Make static.
18619 2013-06-10 Jakub Jelinek <jakub@redhat.com>
18622 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
18623 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
18624 Use DATA_ABI_ALIGNMENT for that case instead if defined.
18625 (get_variable_align): New function.
18626 (get_variable_section, emit_bss, emit_common,
18627 assemble_variable_contents, place_block_symbol): Use
18628 get_variable_align instead of DECL_ALIGN.
18629 (assemble_noswitch_variable): Add align argument, use it
18630 instead of DECL_ALIGN.
18631 (assemble_variable): Adjust caller. Use get_variable_align
18632 instead of DECL_ALIGN.
18633 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
18635 (DATA_ABI_ALIGNMENT): Define.
18636 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
18637 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
18638 opt is false, only return the psABI mandated alignment increase.
18639 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
18640 (DATA_ABI_ALIGNMENT): ... this.
18641 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
18642 (DATA_ABI_ALIGNMENT): ... this.
18643 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
18644 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
18645 (DATA_ABI_ALIGNMENT): ... this.
18646 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
18647 * doc/tm.texi: Regenerated.
18649 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
18651 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
18652 cmp_code to construct REG_EQUAL note.
18654 2013-06-09 Jakub Jelinek <jakub@redhat.com>
18657 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
18658 that operands[2] doesn't overlap with operands[0].
18660 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
18661 Jan Hubicka <jh@suse.cz>
18663 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
18664 hack to mark symbols as used.
18666 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
18668 PR rtl-optimization/57559
18669 * lra-constraints.c (process_alt_operands): Don't discourage
18670 memory with known offset for offsetable memory constraint.
18671 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
18673 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
18675 * varasm.c (struct oc_local_state): Reorder fields.
18676 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
18677 and adjust accordingly.
18678 (output_constructor): Reorder initialization code and adjust call to
18679 output_constructor_bitfield.
18681 2013-06-07 Jan Hubicka <jh@suse.cz>
18683 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
18685 2013-06-07 David Malcolm <dmalcolm@redhat.com>
18687 * tree-object-size.c (unknown): Make const.
18689 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18691 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
18692 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
18693 for last alternative in the cpu_facility attribute.
18695 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18698 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
18699 (xordi3): Change operand 2 constraint to arm_xordi_operand.
18700 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
18701 * config/arm/constraints.md (Dg): New constraint.
18702 * config/arm/neon.md (xordi3_neon): Remove.
18703 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
18704 * config/arm/predicates.md (arm_xordi_operand): New predicate.
18706 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18708 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
18709 Clean up alternatives.
18711 2013-06-07 Alan Modra <amodra@gmail.com>
18713 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
18716 2013-06-07 Alan Modra <amodra@gmail.com>
18718 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
18720 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18722 * config/arm/constraints.md (Df): New constraint.
18723 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
18724 Correct length attribute for last two alternatives.
18726 2013-06-07 Alan Modra <amodra@gmail.com>
18728 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
18729 override user -mfp-in-toc.
18730 (offsettable_ok_by_alignment): Consider just the current access
18731 rather than the whole object, unless BLKmode. Handle
18732 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18733 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18734 for -mcmodel=medium.
18735 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18736 override user -mfp-in-toc or -msum-in-toc. Default to
18737 -mno-fp-in-toc for -mcmodel=medium.
18739 2013-06-06 DJ Delorie <dj@redhat.com>
18741 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18742 TARGET_VALID_POINTER_MODE.
18744 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18745 Pat Haugen <pthaugen@us.ibm.com>
18746 Peter Bergner <bergner@vnet.ibm.com>
18748 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18749 Document new power8 builtins.
18751 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18752 condition code register, to allow 128-bit logical operations to be
18753 done in the VSX or GPR registers.
18754 (nor<mode>3): Use the canonical form for nor.
18755 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18756 vclz*, and vpopcnt* vector instructions.
18757 (nand<mode>3): Likewise.
18758 (orc<mode>3): Likewise.
18759 (clz<mode>2): LIkewise.
18760 (popcount<mode>2): Likewise.
18762 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18763 that only the GPRs are recognized.
18765 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18766 support for new power8 builtins.
18768 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18770 (xscvdpspn): Likewise.
18776 (vpopcnt): Likewise.
18777 (vpopcntb): Likewise.
18778 (vpopcnth): Likewise.
18779 (vpopcntw): Likewise.
18780 (vpopcntd): Likewise.
18782 (vmrgew): Likewise.
18783 (vmrgow): Likewise.
18785 (eqv_v16qi3): Likewise.
18786 (eqv_v8hi3): Likewise.
18787 (eqv_v4si3): Likewise.
18788 (eqv_v2di3): Likewise.
18789 (eqv_v4sf3): Likewise.
18790 (eqv_v2df3): Likewise.
18792 (nand_v16qi3): Likewise.
18793 (nand_v8hi3): Likewise.
18794 (nand_v4si3): Likewise.
18795 (nand_v2di3): Likewise.
18796 (nand_v4sf3): Likewise.
18797 (nand_v2df3): Likewise.
18799 (orc_v16qi3): Likewise.
18800 (orc_v8hi3): Likewise.
18801 (orc_v4si3): Likewise.
18802 (orc_v2di3): Likewise.
18803 (orc_v4sf3): Likewise.
18804 (orc_v2df3): Likewise.
18806 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18807 allow power8 quad mode in 64-bit.
18808 (rs6000_builtin_vectorized_function): Add support to vectorize
18809 ISA 2.07 count leading zeros, population count builtins.
18810 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18811 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18812 (builtin_function_type): Add vgbbd builtin function which takes an
18814 (altivec_expand_vec_perm_const): Add support for new power8 merge
18817 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18818 that does not include TImdoe for use with 32-bit.
18819 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18821 (UNSPEC_VSX_CVDPSPN): Likewise.
18822 (vsx_xscvdpspn): Likewise.
18823 (vsx_xscvspdpn): Likewise.
18824 (vsx_xscvdpspn_scalar): Likewise.
18825 (vsx_xscvspdpn_directmove): Likewise.
18826 (vsx_and<mode>3): Split logical operations into 32-bit and
18827 64-bit. Add support to do logical operations on TImode as well as
18828 VSX vector types. Allow logical operations to be done in either
18829 VSX registers or in general purpose registers in 64-bit mode. Add
18830 splitters if GPRs were used. For AND, add clobber of CCmode to
18831 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18833 (vsx_and<mode>3_32bit): Likewise.
18834 (vsx_and<mode>3_64bit): Likewise.
18835 (vsx_ior<mode>3): Likewise.
18836 (vsx_ior<mode>3_32bit): Likewise.
18837 (vsx_ior<mode>3_64bit): Likewise.
18838 (vsx_xor<mode>3): Likewise.
18839 (vsx_xor<mode>3_32bit): Likewise.
18840 (vsx_xor<mode>3_64bit): Likewise.
18841 (vsx_one_cmpl<mode>2): Likewise.
18842 (vsx_one_cmpl<mode>2_32bit): Likewise.
18843 (vsx_one_cmpl<mode>2_64bit): Likewise.
18844 (vsx_nor<mode>3): Likewise.
18845 (vsx_nor<mode>3_32bit): Likewise.
18846 (vsx_nor<mode>3_64bit): Likewise.
18847 (vsx_andc<mode>3): Likewise.
18848 (vsx_andc<mode>3_32bit): Likewise.
18849 (vsx_andc<mode>3_64bit): Likewise.
18850 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18851 and xxlorc instructions.
18852 (vsx_eqv<mode>3_64bit): Likewise.
18853 (vsx_nand<mode>3_32bit): Likewise.
18854 (vsx_nand<mode>3_64bit): Likewise.
18855 (vsx_orc<mode>3_32bit): Likewise.
18856 (vsx_orc<mode>3_64bit): Likewise.
18858 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18860 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18862 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18863 (p8_vmrgow): Likewise.
18864 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18865 GPRs to be split under VSX.
18866 (p8v_clz<mode>2): Add power8 count leading zero support.
18867 (p8v_popcount<mode>2): Add power8 population count support.
18868 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18871 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18874 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18876 (vec_nand): Likewise.
18877 (vec_vclz): Likewise.
18878 (vec_vclzb): Likewise.
18879 (vec_vclzd): Likewise.
18880 (vec_vclzh): Likewise.
18881 (vec_vclzw): Likewise.
18882 (vec_vgbbd): Likewise.
18883 (vec_vmrgew): Likewise.
18884 (vec_vmrgow): Likewise.
18885 (vec_vpopcnt): Likewise.
18886 (vec_vpopcntb): Likewise.
18887 (vec_vpopcntd): Likewise.
18888 (vec_vpopcnth): Likewise.
18889 (vec_vpopcntw): Likewise.
18891 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18893 PR rtl-optimization/57468
18894 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18897 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18899 PR rtl-optimization/57459
18900 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18901 type when setting live regs.
18903 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18905 * config/s390/s390.opt (mlra): New option.
18906 * config/s390/s390.c (s390_decompose_address): Check displacement
18907 for all registers for LRA.
18908 (s390_secondary_reload): Don't used secondary reloads for LRA.
18909 (s390_lra_p): New function.
18910 (TARGET_LRA_P): Define.
18911 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18912 of attribute cpu_facility to zarch for the last alternative.
18913 (*cmpmem_short): Ditto.
18915 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18917 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18918 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18919 (arm_expand_prologue): Likewise.
18921 2013-06-06 Teresa Johnson <tejohnson@google.com>
18924 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18925 as this is now done by redirect_edge_and_branch_force.
18926 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18927 barriers, and fix interaction with splitting.
18928 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18929 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18930 reflect changes made in the routine.
18931 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18932 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18933 since this is called in cfglayout mode, and replace partition fixup
18934 with assert as that is now done by force_nonfallthru_and_redirect.
18935 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18936 already be marked with region crossing note.
18937 (insert_section_boundary_note): Make non-static, gate on flag
18938 has_bb_partition, rewrite to also check for multiple partitions.
18939 (rest_of_handle_reorder_blocks): Remove call to
18940 insert_section_boundary_note, now done later during free_cfg.
18941 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18942 * bb-reorder.h (insert_section_boundary_note): Declare.
18943 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18944 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18945 invoke insert_section_boundary_note.
18946 (try_redirect_by_replacing_jump): Remove unnecessary
18947 check for region crossing note.
18948 (fixup_partition_crossing): New function.
18949 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18950 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18951 in non-cfglayout mode.
18952 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18953 remove old code that tried to do this. Emit barrier correctly
18954 when we are in cfglayout mode.
18955 (last_bb_in_partition): New function.
18956 (rtl_split_edge): Correctly fixup partition boundaries.
18957 (commit_one_edge_insertion): Remove old code that tried to
18958 fixup region crossing edge since this is now handled in
18959 split_block, and set up insertion point correctly since
18960 block may now end in a jump.
18961 (verify_hot_cold_block_grouping): Guard against checking when not in
18962 linearized RTL mode.
18963 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18965 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18966 rtl_verify_flow_info, so not called in cfglayout mode.
18967 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18968 (fixup_reorder_chain): Remove old code that attempted to fixup region
18969 crossing note as this is now handled in force_nonfallthru_and_redirect.
18970 (duplicate_insn_chain): Don't duplicate switch section notes.
18971 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18973 * basic-block.h (emit_barrier_after_bb): Declare.
18975 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18977 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18978 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18979 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18981 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18983 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18985 * config/arm/ldmstm.md: Regenerate.
18987 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18989 * config/arm/sync.md (atomic_loaddi_1):
18990 Disable predication for arm_restrict_it.
18991 (arm_load_exclusive<mode>): Likewise.
18992 (arm_load_exclusivesi): Likewise.
18993 (arm_load_exclusivedi): Likewise.
18994 (arm_load_acquire_exclusive<mode>): Likewise.
18995 (arm_load_acquire_exclusivesi): Likewise.
18996 (arm_load_acquire_exclusivedi): Likewise.
18997 (arm_store_exclusive<mode>): Likewise.
18998 (arm_store_exclusive<mode>): Likewise.
18999 (arm_store_release_exclusivedi): Likewise.
19000 (arm_store_release_exclusive<mode>): Likewise.
19002 2013-06-06 Richard Biener <rguenther@suse.de>
19004 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
19006 (lto_tag_is_tree_code_p): Adjust.
19007 (lto_tag_is_gimple_code_p): Likewise.
19008 (lto_gimple_code_to_tag): Likewise.
19009 (lto_tag_to_gimple_code): Likewise.
19010 (lto_tree_code_to_tag): Likewise.
19011 (lto_tag_to_tree_code): Likewise.
19012 * data-streamer.h (streamer_write_hwi_in_range): Use
19013 uhwi streaming to stream the normalized range.
19014 (streamer_read_hwi_in_range): Likewise.
19016 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19018 * config/arm/arm.md (enabled_for_depr_it): New attribute.
19019 (predicable_short_it): Likewise.
19020 (predicated): Likewise.
19021 (enabled): Handle above.
19022 (define_cond_exec): Set predicated attribute to yes.
19024 2013-06-05 Mike Stump <mikestump@comcast.net>
19026 * gdbinit.in (__FUNCTION__): Add.
19028 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
19030 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
19031 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
19033 2013-06-05 Jan Hubicka <jh@suse.cz>
19035 * varasm.c (mark_decl_referenced): Revert the removal until targets
19038 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
19040 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
19041 instead of mark_decl_referenced.
19043 2013-06-05 Jan Hubicka <jh@suse.cz>
19045 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
19046 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
19047 and symtab_used_from_object_file_p.
19048 (cgraph_make_node_local_1): Clear forced_by_abi.
19049 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
19050 * cgraph.h (symtab_node_base): Add forced_by_abi.
19051 (decide_is_variable_needed): Remove.
19052 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
19053 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
19054 (decide_is_symbol_needed): ... this one; handle symbols in general;
19055 always analyze virtuals; honnor forced_by_abi.
19056 (cgraph_finalize_function): Update.
19057 (varpool_finalize_decl): Update.
19058 (symbol_defined_and_needed): Remove.
19059 (analyze_functions): Update.
19060 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
19061 output_refs, input_overwrite_node): Handle forced_by_abi.
19062 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
19063 (address_taken_from_non_vtable_p): ... this one.
19064 (comdat_can_be_unshared_p_1): New function.
19065 (cgraph_comdat_can_be_unshared_p): Rename to ...
19066 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
19067 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
19068 (function_and_variable_visibility): Clear forced_by_abi as needed.
19069 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
19070 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
19071 * symtab.c (dump_symtab_base): Dump forced_by_abi.
19072 * varpool.c (decide_is_variable_needed): Remove.
19074 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19076 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
19077 (arm_option_override): Override arm_restrict_it where appropriate.
19078 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
19079 * config/arm/arm.opt (mrestrict-it): New command-line option.
19080 * doc/invoke.texi: Document -mrestrict-it.
19082 2013-06-05 David Malcolm <dmalcolm@redhat.com>
19084 * tsan.c (tsan_atomic_table): Make const.
19086 2013-06-05 Richard Biener <rguenther@suse.de>
19088 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
19089 index associated with the tree we are supposed to replace.
19090 * tree-streamer-out.c (pack_ts_base_value_fields): Output
19091 TREE_ASM_WRITTEN as zero for everything but SSA names.
19093 2013-06-05 David Malcolm <dmalcolm@redhat.com>
19095 * tree-ssa-structalias.c (call_stmt_vars): Make static.
19097 2013-06-04 Jan Hubicka <jh@suse.cz>
19099 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
19100 (input_node, input_varpool_node): Handle correctly external same
19102 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
19103 nodes at ltrans stage.
19105 2013-06-04 Jan Hubicka <jh@suse.cz>
19107 * ipa-inline.c (update_caller_keys): Fix availability test.
19108 (update_callee_keys): Likewise.
19109 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
19110 to follow ELF standard.
19112 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
19114 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
19115 (mips64r5900el-*-elf*): New configurations.
19116 * config/mips/mips-cpus.def (r5900): New processor.
19117 * config/mips/mips-tables.opt: Regenerate.
19118 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
19119 (mips_issue_rate): Handle PROCESSOR_R5900.
19120 (mips_reorg_process_insns): Force reorder mode for the R5900.
19121 * config/mips/mips.h (TARGET_MIPS5900): Define.
19122 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
19124 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
19126 * config/mips/mips.md (processor): Add r5900.
19127 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
19129 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19131 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
19132 into function to generate MOVI instruction.
19133 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
19134 (aarch64_preferred_simd_mode): Turn into wrapper.
19135 (aarch64_output_scalar_simd_mov_immediate): New function.
19136 * config/aarch64/aarch64-protos.h: Add prototype for above.
19138 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19140 * config/aarch64/aarch64.c (simd_immediate_info): Remove
19141 element_char member.
19142 (sizetochar): Return signed char.
19143 (aarch64_simd_valid_immediate): Remove elchar and other
19144 unnecessary variables.
19145 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
19146 Calculate element_char as required.
19147 * config/aarch64/aarch64-protos.h: Update and move prototype
19148 for aarch64_output_simd_mov_immediate.
19149 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
19152 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19154 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
19155 information completed by aarch64_simd_valid_immediate.
19156 (aarch64_legitimate_constant_p): Update arguments.
19157 (aarch64_simd_valid_immediate): Work with struct rather than many
19159 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
19160 (aarch64_simd_make_constant): Update arguments.
19161 (aarch64_output_simd_mov_immediate): Work with struct rather than
19162 many pointers. Output immediate directly rather than as operand.
19163 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
19165 * config/aarch64/constraints.md (Dn): Update arguments.
19167 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19169 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
19171 (aarch64_simd_immediate_valid_for_move): Remove.
19172 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
19173 (aarch64_simd_make_constant): Update call.
19174 (aarch64_output_simd_mov_immediate): Update call.
19175 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
19177 * config/aarch64/constraints.md (Dn): Update call.
19179 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19181 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
19182 return type to bool for prototype.
19183 (aarch64_legitimate_constant_p): Check for true instead of not -1.
19184 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
19185 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
19187 2013-06-04 Catherine Moore <clm@codesourcery.com>
19189 * config/mips/mips.opt (meva): New.
19190 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
19191 (ASM_SPEC): Handle -meva.
19192 * doc/invoke.texi (meva): Document.
19194 2013-06-04 Alan Modra <amodra@gmail.com>
19196 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
19199 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19201 * rtl.def: Add extra fourth optional field to define_cond_exec.
19202 * gensupport.c (process_one_cond_exec): Process attributes from
19204 * doc/md.texi: Document fourth field in define_cond_exec.
19206 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
19208 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
19209 out the processing order as in store_bit_field_1.
19211 2013-06-04 Jan Hubicka <jh@suse.cz>
19213 PR middle-end/57500
19214 * cgraphunit.c (cgraph_process_same_body_aliases): Create
19215 non-VAR_DECL node if it does not exist yet.
19217 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
19219 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
19220 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
19221 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
19222 target_cpu_default setting.
19224 2013-06-03 Teresa Johnson <tejohnson@google.com>
19226 * dumpfile.c (opt_info_switch_p): Change -fopt-info
19227 default to -fopt-info=optimized instead of all.
19228 * doc/invoke.texi: Ditto.
19229 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
19230 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
19231 (execute_vect_slp): Emit BB vectorization success under
19232 MSG_OPTIMIZED_LOCATIONS.
19233 * tree-vect-slp.c (vect_slp_transform_bb): Change
19234 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
19235 * tree-vect-loop.c (vect_transform_loop): Ditto.
19237 2013-06-03 Jason Merrill <jason@redhat.com>
19240 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
19241 Use TARGET_EXPR for C++.
19243 2013-06-03 Jakub Jelinek <jakub@redhat.com>
19245 PR rtl-optimization/57268
19246 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
19247 if DEBUG_INSN_P (insn).
19250 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19252 PR rtl-optimization/57268
19253 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19254 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19256 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19258 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
19259 (ix86_avoid_lea_for_addr): Likewise.
19260 (exact_dependency_1): Likewise.
19261 (ix86_adjust_cost): Likewise.
19262 (swap_top_of_ready_list): Fix formatting and !reload_completed check
19264 (do_reorder_for_imul): Fix typo, formatting and
19265 !reload_completed check removed.
19266 (ix86_sched_reorder): Fix typo and formatting.
19267 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
19270 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
19272 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
19274 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
19276 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
19277 <STRING_CST>: Likewise.
19278 <VECTOR_CST>: Likewise.
19280 2013-06-01 Janus Weil <janus@gcc.gnu.org>
19281 Mikael Morin <mikael@gcc.gnu.org>
19283 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
19284 * config.in: Regenerated.
19285 * configure: Regenerated.
19287 2013-06-01 Jan Hubicka <jh@suse.cz>
19289 PR middle-end/57366
19290 * cgraphunit.c (compile): When weakref is not supported,
19291 set up transparent aliases before final output pass.
19292 * varasm.c (assemble_alias): Do not try to do it here.
19294 2013-06-01 Jan Hubicka <jh@suse.cz>
19296 PR middle-end/57467
19297 * passes.c (for_per_function): Skip unanalyzed functions.
19299 2013-06-01 Jan Hubicka <jh@suse.cz>
19301 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
19302 (lto_symtab_merge_symbols_1): ... this one.
19303 (lto_symtab_merge_cgraph_nodes): Rename to ...
19304 (lto_symtab_merge_symbols): ... this one; simplify.
19305 * cgraph.c (same_body_aliases_done): Rename to ...
19306 (cpp_implicit_aliases_done): ... this one.
19307 (cgraph_create_function_alias): Update.
19308 (cgraph_same_body_alias): Update.
19309 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
19310 (verify_edge_corresponds_to_fndecl): Simplify.
19311 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
19312 (cgraph_node): Remove same_body_alias.
19313 (varpool_node): Remove alias_of and extra_name_alias.
19314 (same_body_aliases_done): Rename to ..
19315 (cpp_implicit_aliases_done): ... this one.
19316 (symtab_alias_ultimate_target): Add default parameter.
19317 (symtab_resolve_alias): New function.
19318 (fixup_same_cpp_alias_visibility): Declare.
19319 (cgraph_function_node): Add default parameter.
19320 (cgraph_node_asm_name): Likewise.
19321 (cgraph_function_or_thunk_node): Add default parameter; do
19322 not ICE when it is NULL.
19323 (varpool_variable_node): Likewise.
19324 * tree-emutls.c (create_emultls_var): Update.
19325 (ipa_lower_emutls): Update.
19326 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
19327 (cgraph_reset_node): Reset alias info.
19328 (cgraph_finalize_function): Update.
19329 (fixup_same_cpp_alias_visibility): Move to symtab.c.
19330 (analyze_function): Simplify.
19331 (cgraph_process_same_body_aliases): Simplify.
19332 (analyze_functions): Fixup same body aliases.
19333 (handle_alias_pairs): Simplify.
19334 (assemble_thunk): Update.
19335 (assemble_thunks_and_aliases): Update.
19336 (output_weakrefs): Rewrite.
19337 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
19338 (lto_output_varpool_node): Likewise.
19339 (compute_ltrans_boundary): Remve assert.
19340 (get_alias_symbol): New functoin.
19341 (input_node): Rewrite alias handling.
19342 (input_varpool_node): Likewise.
19343 * ipa-pure-const.c (propagate_pure_const): Fix formating.
19344 * ipa.c (process_references): Handle weakrefs correctly.
19345 (symtab_remove_unreachable_nodes): Likewise.
19346 * trans-mem.c (get_cg_data): Update.
19347 (ipa_tm_create_version_alias): Update.
19348 (ipa_tm_execute): Update.
19349 * symtab.c (dump_symtab_base): Dump aliases.
19350 (verify_symtab_base): Verify aliases.
19351 (symtab_node_availability): New function.
19352 (symtab_alias_ultimate_target): Simplify.
19353 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
19354 handle all the fixup cases.
19355 (symtab_resolve_alias): New function.
19356 * passes.c (ipa_write_summaries): Handle weakrefs.
19357 * varpool.c (varpool_analyze_node): Simplify.
19358 (assemble_aliases): Update.
19359 (varpool_create_variable_alias): Simplify.
19360 (varpool_extra_name_alias): Simplify.
19361 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
19362 (lto_symtab_merge_symbols): ... this one.
19364 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
19367 PR rtl-optimization/57268
19368 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19369 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19371 2013-06-01 Tobias Burnus <burnus@net-b.de>
19373 Partially reverted:
19374 2013-05-31 Tobias Burnus <burnus@net-b.de>
19376 PR middle-end/57073
19377 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
19380 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19382 PR rtl-optimization/57268
19383 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19384 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19386 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19388 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
19389 unordered comparison operators when -fno-trapping-math is in effect
19391 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
19392 and implement unordered comparison operators properly on the e500.
19394 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19396 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
19397 for constant scalar integers.
19398 (simplify_relational_operation_1): Likewise.
19400 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
19402 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
19403 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
19406 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19407 Igor Zamyatin <igor.zamyatin@intel.com>
19409 Silvermont (SLM) architecture performance tuning.
19410 * config/i386/i386.h (enum ix86_tune_indices): Add
19411 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
19412 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
19414 * config/i386/i386.c (initial_ix86_tune_features)
19415 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
19416 (ix86_lea_outperforms): Handle Silvermont tuning.
19417 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
19419 (ix86_use_lea_for_mov): Likewise.
19420 (ix86_avoid_lea_for_addr): Likewise.
19421 (ix86_lea_for_add_ok): Likewise.
19422 (exact_dependency_1): New function.
19423 (exact_store_load_dependency): Likewise.
19424 (ix86_adjust_cost): Handle Silvermont tuning.
19425 (do_reoder_for_imul): Likewise.
19426 (swap_top_of_ready_list): New function.
19427 (ix86_sched_reorder): Changed to handle Silvermont tuning.
19429 * config/i386/i386.md (peepholes that split memory operand in fp
19432 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19434 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
19435 Remove un-necessary braces.
19437 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19439 * config/aarch64/aarch64.c (aarch64_classify_symbol):
19440 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
19442 2013-05-31 Tobias Burnus <burnus@net-b.de>
19444 PR middle-end/57073
19445 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
19447 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19450 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
19451 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
19452 * config/arm/neon.md (iordi3_neon): Remove.
19453 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
19454 * config/arm/predicates.md (imm_for_neon_logic_operand):
19455 Move to earlier in the file.
19456 (neon_logic_op2): Likewise.
19457 (arm_iordi_operand_neon): New predicate.
19459 2013-05-31 Richard Biener <rguenther@suse.de>
19461 PR tree-optimization/57478
19462 PR tree-optimization/57453
19463 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
19466 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19468 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
19469 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
19471 2013-05-30 Tobias Burnus <burnus@net-b.de>
19472 Thomas Koenig <tkoenig@gcc.gnu.org>
19474 PR middle-end/57073
19475 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
19476 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
19478 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
19480 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
19482 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
19484 * target.def (register_usage_leveling_p): New hook.
19485 * targhooks.c (default_register_usage_leveling_p): New.
19486 * targhooks.h (default_register_usage_leveling_p): New prototype.
19487 * lra-assigns.c (register_usage_leveling_p): Use the hook.
19488 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
19489 * doc/tm.texi: Update.
19490 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
19492 2013-05-30 Ian Bolton <ian.bolton@arm.com>
19494 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
19495 (*insv_reg<mode>): New define_insn.
19497 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
19499 PR rtl-optimization/57439
19500 * postreload.c (move2add_valid_value_p): Check that we have
19501 a zero subreg_regno_offset when accessing the register in
19502 the requested mode.
19504 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19505 Igor Zamyatin <igor.zamyatin@intel.com>
19507 Silvermont (SLM) architecture pipeline model, tuning and
19509 * config.gcc: Add slm config options and target.
19511 * config/i386/slm.md: New.
19513 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
19515 * config/i386/i386-c.c (ix86_target_macros_internal): New case
19517 (ix86_target_macros_internal): Likewise.
19519 * config/i386/i386.c (slm_cost): New cost.
19520 (m_SLM): New macro flag.
19521 (initial_ix86_tune_features): Set m_SLM.
19522 (x86_accumulate_outgoing_args): Likewise.
19523 (x86_arch_always_fancy_math_387): Likewise.
19524 (processor_target_table): Add slm cost.
19525 (cpu_names): Add slm cpu name.
19526 (x86_option_override_internal): Set SLM ISA.
19527 (ix86_issue_rate): New case PROCESSOR_SLM.
19528 (ia32_multipass_dfa_lookahead): Likewise.
19529 (fold_builtin_cpu): Add slm.
19531 * config/i386/i386.h (TARGET_SLM): New target macro.
19532 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
19533 (processor_type): Add PROCESSOR_SLM.
19535 * config/i386/i386.md (cpu): Add new value "slm".
19536 (slm.md): Include slm.md.
19538 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
19539 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19541 * config/arm/arm-protos.h: Add and update function protos.
19542 * config/arm/arm.c (use_simple_return_p): New added.
19543 (thumb2_expand_return): Check simple_return flag.
19544 * config/arm/arm.md: Add simple_return and conditional simple_return.
19545 * config/arm/iterators.md: Add iterator for return and simple_return.
19547 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19549 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
19550 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
19551 (arm_emit_vfp_multi_reg_pop): Likewise.
19552 (thumb2_emit_ldrd_pop): Likewise.
19553 (arm_expand_epilogue): Add misc REG_CFA notes.
19554 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
19556 2013-05-29 Lawrence Crowl <crowl@google.com>
19558 * config/arm/t-arm: Update for below.
19560 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
19561 Change type to hash_table. Update dependent calls and types.
19563 * config/i386/t-cygming: Update for below.
19565 * config/i386/t-interix: Update for below.
19567 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
19568 Change type to hash_table. Update dependent calls and types.
19569 (i386_find_on_wrapper_list::wrappers): Likewise.
19571 * config/ia64/t-ia64: Update for below.
19573 * config/ia64/ia64.c (bundle_state_table):
19574 Change type to hash_table. Update dependent calls and types.
19576 * config/mips/mips.c (mips_reorg_process_insns::htab):
19577 Change type to hash_table. Update dependent calls and types.
19579 * config/sol2.c (solaris_comdat_htab):
19580 Change type to hash_table. Update dependent calls and types.
19582 * config/t-sol2: Update for above.
19584 2013-05-29 Teresa Johnson <tejohnson@google.com>
19586 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
19587 functions are not yet marked as defined.
19589 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
19590 Pat Haugen <pthaugen@us.ibm.com>
19591 Peter Bergner <bergner@vnet.ibm.com>
19593 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
19597 (vrotl<mode>3): Likewise.
19598 (vashl<mode>3): Likewise.
19599 (vlshr<mode>3): Likewise.
19600 (vashr<mode>3): Likewise.
19602 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
19603 support for power8 V2DI builtins.
19605 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
19606 power8 V2DI builtins.
19607 (vupkhsw): Likewise.
19608 (vupklsw): Likewise.
19609 (vaddudm): Likewise.
19610 (vminsd): Likewise.
19611 (vmaxsd): Likewise.
19612 (vminud): Likewise.
19613 (vmaxud): Likewise.
19614 (vpkudum): Likewise.
19615 (vpksdss): Likewise.
19616 (vpkudus): Likewise.
19617 (vpksdus): Likewise.
19622 (vsubudm): Likewise.
19623 (vcmpequd): Likewise.
19624 (vcmpgtsd): Likewise.
19625 (vcmpgtud): Likewise.
19626 (vcmpequd_p): Likewise.
19627 (vcmpgtsd_p): Likewise.
19628 (vcmpgtud_p): Likewise.
19629 (vupkhsw): Likewise.
19630 (vupklsw): Likewise.
19631 (vaddudm): Likewise.
19632 (vmaxsd): Likewise.
19633 (vmaxud): Likewise.
19634 (vminsd): Likewise.
19635 (vminud): Likewise.
19636 (vpksdss): Likewise.
19637 (vpksdus): Likewise.
19638 (vpkudum): Likewise.
19639 (vpkudus): Likewise.
19644 (vsubudm): Likewise.
19646 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
19647 support for power8 V2DI instructions.
19649 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
19650 power8 V2DI instructions. Combine pack and unpack insns to use an
19651 iterator for each mode. Check whether a particular mode supports
19652 Altivec instructions instead of just checking TARGET_ALTIVEC.
19653 (UNSPEC_VPKUWUM): Likewise.
19654 (UNSPEC_VPKSHSS): Likewise.
19655 (UNSPEC_VPKSWSS): Likewise.
19656 (UNSPEC_VPKUHUS): Likewise.
19657 (UNSPEC_VPKSHUS): Likewise.
19658 (UNSPEC_VPKUWUS): Likewise.
19659 (UNSPEC_VPKSWUS): Likewise.
19660 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
19661 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
19662 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
19663 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
19664 (UNSPEC_VUPKHSB): Likewise.
19665 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
19666 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
19667 (UNSPEC_VUPKHSH): Likewise.
19668 (UNSPEC_VUPKLSB): Likewise.
19669 (UNSPEC_VUPKLSH): Likewise.
19671 (VI_char): Likewise.
19672 (VI_scalar): Likewise.
19673 (VI_unit): Likewise.
19675 (VP_small): Likewise.
19676 (VP_small_lc): Likewise.
19677 (VU_char): Likewise.
19678 (add<mode>3): Likewise.
19679 (altivec_vaddcuw): Likewise.
19680 (altivec_vaddu<VI_char>s): Likewise.
19681 (altivec_vadds<VI_char>s): Likewise.
19682 (sub<mode>3): Likewise.
19683 (altivec_vsubcuw): Likewise.
19684 (altivec_vsubu<VI_char>s): Likewise.
19685 (altivec_vsubs<VI_char>s): Likewise.
19686 (altivec_vavgs<VI_char>): Likewise.
19687 (altivec_vcmpbfp): Likewise.
19688 (altivec_eq<mode>): Likewise.
19689 (altivec_gt<mode>): Likewise.
19690 (altivec_gtu<mode>): Likewise.
19691 (umax<mode>3): Likewise.
19692 (smax<mode>3): Likewise.
19693 (umin<mode>3): Likewise.
19694 (smin<mode>3): Likewise.
19695 (altivec_vpkuhum): Likewise.
19696 (altivec_vpkuwum): Likewise.
19697 (altivec_vpkshss): Likewise.
19698 (altivec_vpkswss): Likewise.
19699 (altivec_vpkuhus): Likewise.
19700 (altivec_vpkshus): Likewise.
19701 (altivec_vpkuwus): Likewise.
19702 (altivec_vpkswus): Likewise.
19703 (altivec_vpks<VI_char>ss): Likewise.
19704 (altivec_vpks<VI_char>us): Likewise.
19705 (altivec_vpku<VI_char>us): Likewise.
19706 (altivec_vpku<VI_char>um): Likewise.
19707 (altivec_vrl<VI_char>): Likewise.
19708 (altivec_vsl<VI_char>): Likewise.
19709 (altivec_vsr<VI_char>): Likewise.
19710 (altivec_vsra<VI_char>): Likewise.
19711 (altivec_vsldoi_<mode>): Likewise.
19712 (altivec_vupkhsb): Likewise.
19713 (altivec_vupkhs<VU_char>): Likewise.
19714 (altivec_vupkls<VU_char>): Likewise.
19715 (altivec_vupkhsh): Likewise.
19716 (altivec_vupklsb): Likewise.
19717 (altivec_vupklsh): Likewise.
19718 (altivec_vcmpequ<VI_char>_p): Likewise.
19719 (altivec_vcmpgts<VI_char>_p): Likewise.
19720 (altivec_vcmpgtu<VI_char>_p): Likewise.
19721 (abs<mode>2): Likewise.
19722 (vec_unpacks_hi_v16qi): Likewise.
19723 (vec_unpacks_hi_v8hi): Likewise.
19724 (vec_unpacks_lo_v16qi): Likewise.
19725 (vec_unpacks_hi_<VP_small_lc>): Likewise.
19726 (vec_unpacks_lo_v8hi): Likewise.
19727 (vec_unpacks_lo_<VP_small_lc>): Likewise.
19728 (vec_pack_trunc_v8h): Likewise.
19729 (vec_pack_trunc_v4si): Likewise.
19730 (vec_pack_trunc_<mode>): Likewise.
19732 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19734 (vec_vmaxsd): Likewise.
19735 (vec_vmaxud): Likewise.
19736 (vec_vminsd): Likewise.
19737 (vec_vminud): Likewise.
19738 (vec_vpksdss): Likewise.
19739 (vec_vpksdus): Likewise.
19740 (vec_vpkudum): Likewise.
19741 (vec_vpkudus): Likewise.
19742 (vec_vrld): Likewise.
19743 (vec_vsld): Likewise.
19744 (vec_vsrad): Likewise.
19745 (vec_vsrd): Likewise.
19746 (vec_vsubudm): Likewise.
19747 (vec_vupkhsw): Likewise.
19748 (vec_vupklsw): Likewise.
19750 2013-05-29 Jan Hubicka <jh@suse.cz>
19752 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19753 flags; reorder rest of fields in more consistent way.
19754 (varpool_node): Remove analyzed, finalized and alias.
19755 (cgraph_ndoe): Likewise.
19756 (symtab_alias_ultimate_target): New function.
19757 (cgraph_function_node): Move offline.
19758 (cgraph_reset_node): Declare.
19759 (cgraph_comdat_can_be_unshared_p): Remove.
19760 (varpool_remove_initializer): Declare.
19761 (varpool_first_defined_variable, varpool_next_defined_variable
19762 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19763 (cgraph_function_with_gimple_body_p): Update.
19764 (varpool_all_refs_explicit_p): Update.
19765 (symtab_alias_target): New function.
19766 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19767 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19768 (cgraph_function_or_thunk_node): Simplify using
19769 symtab_alias_ultimate_target.
19770 (varpool_variable_node): Likewise.
19771 * cgraph.c (cgraph_create_function_alias): Update.
19772 (cgraph_add_thunk): Update.
19773 (cgraph_remove_node): Update.
19774 (dump_cgraph_node): Do not dump removed flags.
19775 (cgraph_function_body_availability): Update.
19776 (cgraph_propagate_frequency): Update.
19777 (verify_cgraph_node): Check sanity of local flag.
19778 (cgraph_function_node): Move here from cgraph.h; revamp for
19779 cgraph_function_or_thunk_node.
19780 * lto-symtab.c (lto_varpool_replace_node): Update.
19781 (lto_symtab_resolve_can_prevail_p): Update.
19782 (lto_symtab_merge_cgraph_nodes): Update.
19783 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19784 propagate_constants_accross_call, devirtualization_time_bonus,
19785 ipcp_propagate_stage): Update.
19786 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19787 * ipa-inline-transform.c (clone_inlined_nodes,
19788 preserve_function_body_p): Update.
19789 * ipa-reference.c (propagate): Update.
19790 (write_node_summary_p): Update.
19791 * toplev.c (wrapup_global_declaration_2): Update.
19792 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19793 (analyze_function) ... this one.
19794 (cgraph_process_new_functions): Update.
19795 (cgraph_reset_node): Export.
19796 (cgraph_finalize_function): Update.
19797 (cgraph_add_new_function): Update.
19798 (process_function_and_variable_attributes): Update.
19799 (varpool_finalize_decl): Update.
19800 (symbol_finalized): Remove.
19801 (symbol_finalized_and_needed): Rename to ...
19802 (symbol_defined_and_needed): ... update.
19803 (cgraph_analyze_functions): Update.
19804 (handle_alias_pairs): Update.
19805 (mark_functions_to_output): Update.
19806 (assemble_thunk): Update.
19807 (output_in_order): Update.
19808 (output_weakrefs): Update.
19809 (finalize_compilation_unit): Update.
19810 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19811 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19812 input_node, input_varpool_node): Update.
19813 * dbxout.c (dbxout_expand_expr): Update.
19814 * cgraphclones.c (cgraph_clone_node): Update.
19815 (cgraph_copy_node_for_versioning): Update.
19816 (cgraph_materialize_clone): Update.
19817 (cgraph_materialize_all_clones): Update.
19818 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19819 propagate_pure_const, propagate_nothrow): Update.
19820 * lto-streamer-out.c (lto_output, write_symbol): Update.
19821 * ipa-utils.c (ipa_reverse_postorder): Update.
19822 * ipa-inline.c (can_inline_edge_p): Update.
19823 (update_caller_keys, ipa_inline): Update.
19824 * dwarf2out.c (reference_to_unused,
19825 premark_types_used_by_global_vars_helper): Update.
19826 * tree-eh.c (tree_could_trap_p): Update.
19827 * ipa-split.c (consider_split, execute_split_functions): Update.
19828 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19829 has_addr_references_p): Update; move ahead in file for better
19831 (process_references): Simplify.
19832 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19833 bodies are removed.
19834 (cgraph_comdat_can_be_unshared_p): Make static.
19835 (cgraph_externally_visible_p): Update.
19836 (varpool_externally_visible_p): Update.
19837 (function_and_variable_visibility): Update.
19838 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19839 ipa_tm_mark_force_output_node): Update.
19840 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19841 estimate_edge_devirt_benefit, inline_generate_summary,
19842 inline_write_summary): Update.
19843 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19844 * ipa-prop.c (ipa_compute_jump_functions): Update.
19845 (ipa_print_node_params, ipa_prop_read_section,
19846 ipa_update_after_lto_read, read_replacements_section): Update.
19847 * varasm.c (mark_decl_referenced): Update.
19848 (assemble_alias, dump_tm_clone_pairs): Update.
19849 * tree-inline.c (copy_bb): Update.
19850 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19852 * symtab.c (dump_symtab_base): Print new flags.
19853 (verify_symtab_base): Verify new flags.
19854 (symtab_alias_ultimate_target): New function.
19855 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19856 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19858 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19860 * i386.c (ix86_get_function_versions_dispatcher,
19861 ix86_generate_version_dispatcher_body): Update.
19862 (fold_builtin_cpu): Use varpool_add_new_variable.
19863 * varpool.c (varpool_remove_initializer): Break out from ...
19864 (varpool_remove_node): ... this one.
19865 (dump_varpool_node, varpool_node_for_asm,
19866 cgraph_variable_initializer_availability, varpool_analyze_node,
19867 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19868 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19871 2013-05-29 Jan Hubicka <jh@suse.cz>
19873 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19875 2013-05-29 Easwaran Raman <eraman@google.com>
19877 PR tree-optimization/57442
19878 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19879 when control exits the main loop.
19881 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19883 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19885 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19886 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19887 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19890 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19892 PR tree-optimization/57441
19893 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19894 Don't limit size of incr_vec to number of candidates.
19896 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19898 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19899 and mips16 directories.
19900 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19901 (MULTILIB_DIRNAMES): Ditto.
19902 (MULTILIB_EXCEPTIONS): Add new exceptions.
19903 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19904 (MULTILIB_DIRNAMES): Ditto.
19905 (MULTILIB_EXCEPTIONS): Add new exceptions.
19907 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19908 Marcus Shawcroft <marcus.shawcroft@arm.com>
19910 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19911 SYMBOL_TINY_ABSOLUTE.
19912 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19913 SYMBOL_TINY_ABSOLUTE.
19914 (aarch64_expand_mov_immediate): Likewise.
19915 (aarch64_classify_symbol): Likewise.
19916 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19917 Permit SYMBOL_TINY_ABSOLUTE.
19918 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19920 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19921 Marcus Shawcroft <marcus.shawcroft@arm.com>
19923 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19924 Refactor if/switch. Replace gcc_assert with if.
19926 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19928 * config/i386/i386.c (initial_ix86_tune_features): Enable
19929 FP Reassociation for AMD bdver1 and bdver2.
19931 2013-05-29 Martin Jambor <mjambor@suse.cz>
19933 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19934 and IMAGPART_EXPR do not occur within other handled_components.
19936 2013-05-29 Richard Biener <rguenther@suse.de>
19938 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19939 access on whether the use is in the BB we currently try to
19941 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19942 vectorize to vect_bb_slp_scalar_cost.
19944 2013-05-29 Richard Biener <rguenther@suse.de>
19946 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19947 computing scalar cost offsetted by stmts that are kept live
19949 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19950 for computation of scalar cost.
19952 2013-05-28 Steve Ellcey <sellcey@mips.com>
19954 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19956 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19958 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19959 array notation built-in reduction functions.
19960 * doc/passes.texi (Passes): Added documentation about changes done
19962 * doc/invoke.texi (C Dialect Options): Added documentation about
19963 the -fcilkplus flag.
19964 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19965 (BUILTINS_DEF): Depend on cilkplus.def.
19966 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19967 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19968 * cilkplus.def: New file.
19970 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19972 PR rtl-optimization/57439
19973 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19975 2013-05-28 Easwaran Raman <eraman@google.com>
19977 PR tree-optimization/57337
19978 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19979 (find_insert_point): Correctly identify the insertion point
19980 when two statements with the same UID is compared.
19982 2013-05-28 Richard Biener <rguenther@suse.de>
19984 PR tree-optimization/56787
19985 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19986 from the list of data references.
19987 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19989 (vect_analyze_loop_operations): Likewise.
19990 (vect_transform_loop): Remove clobbers.
19992 2013-05-28 Martin Jambor <mjambor@suse.cz>
19994 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19995 and REALPART_EXPRs have scalar type.
19997 2013-05-28 Richard Biener <rguenther@suse.de>
19999 PR tree-optimization/57411
20000 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
20002 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
20003 virtual operand propagation.
20005 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
20007 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
20008 destination register for bmasksi_vis.
20009 (vector_init_bshuffle): Likewise.
20010 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
20012 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
20014 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
20015 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
20016 mode if the instruction isn't available in the original mode.
20017 * config/sparc/sparc.opt (mfix-ut699): New option.
20018 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
20019 (divdf3): Turn into expander.
20020 (divdf3_nofix): New insn.
20021 (divdf3_fix): Likewise.
20022 (divsf3): Disable if -mfix-ut699.
20023 (sqrtdf2): Turn into expander.
20024 (sqrtdf2_nofix): New insn.
20025 (sqrtdf2_fix): Likewise.
20026 (sqrtsf2): Disable if -mfix-ut699.
20028 2013-05-27 Richard Biener <rguenther@suse.de>
20030 PR middle-end/57412
20031 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
20032 block for the new loop.
20034 2013-05-27 Richard Biener <rguenther@suse.de>
20036 PR tree-optimization/57343
20037 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
20038 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
20039 (number_of_iterations_cond): Do not build the folded tree.
20041 2013-05-27 Richard Biener <rguenther@suse.de>
20044 PR middle-end/57381
20045 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
20046 OEP_CONSTANT_ADDRESS_OF retained.
20048 PR tree-optimization/57417
20049 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
20050 for unchanged base.
20051 (set_ssa_val_to): Compare addresses using
20052 get_addr_base_and_unit_offset.
20054 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
20056 PR rtl-optimization/56833
20057 * postreload.c (move2add_record_mode): New function.
20058 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
20059 (move2add_use_add2_insn): Use move2add_record_sym_value.
20060 (move2add_use_add3_insn): Likewise.
20061 (reload_cse_move2add): Use move2add_valid_value_p and
20062 move2add_record_mode. Invalidate call-clobbered and REG_INC
20063 affected regs by setting reg_mode to VOIDmode.
20064 (move2add_note_store): Don't pretend the inside of a SUBREG is
20065 the actual destination. Invalidate single/leading registers by
20066 setting reg_mode to VOIDmode.
20067 Use move2add_record_sym_value, move2add_valid_value_p and
20068 move2add_record_mode.
20070 2013-05-27 Richard Biener <rguenther@suse.de>
20072 PR tree-optimization/57396
20073 * tree-affine.c (double_int_constant_multiple_p): Properly
20074 return false for val == 0 and div != 0.
20076 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
20078 * config/mips/mips.h: Use #elif in preprocessor conditions.
20080 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
20083 * config/mips/constraints.md (kl): New constraint.
20084 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
20085 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
20086 constraint for operand 0. Split after CSE for MIPS16. Emit a move
20087 from LO for MIPS16.
20088 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
20090 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
20093 * config/mips/mips.c (mips_can_inline_p): New function.
20094 (TARGET_CAN_INLINE_P): Define.
20096 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
20098 * sched-int.h (ds_t, dw_t): Make unsigned int.
20099 Fix documentation that describes how all the ds_t bits are used.
20100 Reserve the last bit for delayed-branch scheduling.
20101 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
20102 (BITS_PER_DEP_WEAK): Fix definition and documentation.
20103 (gen_dep_weak_1): Remove prototype.
20104 * sched-deps.c (get_dep_weak_1): Make static.
20105 * target.def (speculate_insn, needs_block_p, gen_spec_check,
20106 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
20107 * doc/tm.texi: Regenerate.
20108 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
20110 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
20113 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
20115 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
20116 Sandra Loosemore <sandra@codesourcery.com>
20118 * config.gcc (powerpc-*): Allow native for with-cpu.
20120 2013-05-24 Jeff Law <law@redhat.com>
20122 PR tree-optimization/57124
20123 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
20124 conversion feeding a condition if the range has an overflow
20125 if -fstrict-overflow. Add warnings for when we do make the
20128 2013-05-24 Dehao Chen <dehao@google.com>
20130 * tree-cfg.c (locus_discrim_map): Fix the typo.
20131 (locus_discrim_hasher): Likewise.
20132 (locus_discrim_hasher::hash): Likewise.
20133 (locus_discrim_hasher::equal): Likewise.
20135 2013-05-24 Martin Jambor <mjambor@suse.cz>
20137 PR tree-optimization/57294
20138 * cgraph.h (ipa_record_stmt_references): Declare.
20139 * cgraphbuild.c (ipa_record_stmt_references): New function.
20140 (build_cgraph_edges): Use ipa_record_stmt_references.
20141 (rebuild_cgraph_edges): Likewise.
20142 (cgraph_rebuild_references): Likewise.
20143 * ipa-prop.c (ipa_modify_call_arguments): Discard references
20144 associated with the old statement and build references from the
20145 newly built statements.
20146 * ipa-ref.c (ipa_remove_stmt_references): New function.
20147 * ipa-ref.h (ipa_remove_stmt_references): Declare.
20149 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
20151 * lra-constraints.c (emit_spill_move): Use smaller mode for
20153 (check_and_process_move): Consider mem-reg moves for secondary
20155 (curr_insn_transform): Don't lose insns emitted before for
20156 secondary memory moves.
20157 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
20158 reg set up in the current insn.
20160 2013-05-24 Dehao Chen <dehao@google.com>
20162 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
20164 (locus_descrim_hasher::equal): Likewise.
20165 (build_gimple_cfg): New discriminator assignment algorithm.
20166 (make_edges): Likewise.
20167 (next_discriminator_for_locus): Likewise.
20168 (same_line_p): Likewise.
20169 (assign_discriminators): Likewise.
20170 (make_cond_expr_edges): Likewise.
20171 (make_gimple_switch_edges): Likewise.
20172 (make_goto_expr_edges): Likewise.
20173 (make_gimple_asm_edges): Likewise.
20175 2013-05-24 Ian Bolton <ian.bolton@arm.com>
20177 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
20178 X format specifier to only display bottom 16 bits.
20179 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
20180 immediate to match for operand 2, since it will be masked.
20182 2013-05-24 Richard Biener <rguenther@suse.de>
20184 PR tree-optimization/57287
20185 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
20186 all SSA names that occur in abnormal PHIs.
20188 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
20191 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
20192 that index is not negative.
20194 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
20196 PR rtl-optimization/55177
20197 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
20198 (simplify_byte_swapping_operation): New.
20199 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
20200 (simplify_relational_operation_1): Deal with BSWAP.
20202 2013-05-23 Richard Henderson <rth@redhat.com>
20205 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
20206 (ix86_reorg): Call it.
20208 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
20211 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
20212 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
20213 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
20215 2013-05-23 Christian Bruel <christian.bruel@st.com>
20218 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
20220 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20221 Marcus Shawcroft <marcus.shawcroft@arm.com>
20223 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
20224 * config/aarch64/constraints.md (Usa): Remove.
20225 * doc/md.texi (AArch64 Usa): Remove.
20227 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20228 Marcus Shawcroft <marcus.shawcroft@arm.com>
20230 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
20231 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
20232 * config/aarch64/predicates.md (aarch64_const_address): Remove.
20233 (aarch64_mov_operand): Use aarch64_mov_operand_p.
20235 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
20237 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
20238 instruction (AdvSIMD).
20239 * config/aarch64/aarch64-builtins.c
20240 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
20241 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
20243 2013-05-23 Martin Jambor <mjambor@suse.cz>
20245 PR middle-end/57347
20246 * tree.h (contains_bitfld_component_ref_p): Declare.
20247 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
20248 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
20250 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
20251 not access a bit-field. Assert all final offsets are byte-aligned.
20253 2013-05-23 Richard Biener <rguenther@suse.de>
20255 PR tree-optimization/57380
20256 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
20257 least one invariant or re-used load.
20258 * passes.c (init_optimization_passes): Move pass_phiprop before
20261 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
20263 * config/aarch64/aarch64-simd.md
20264 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
20266 2013-05-23 Richard Biener <rguenther@suse.de>
20268 PR middle-end/57381
20269 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
20270 OEP_CONSTANT_ADDRESS_OF retained.
20272 2013-05-23 Jakub Jelinek <jakub@redhat.com>
20274 PR middle-end/57344
20275 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
20276 don't lower unit. Handle unit not being always BITS_PER_WORD.
20278 2013-05-23 Richard Biener <rguenther@suse.de>
20280 PR rtl-optimization/57341
20281 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
20282 instead of true_dependence.
20284 2013-05-22 David Malcolm <dmalcolm@redhat.com>
20286 * bb-reorder.c (branch_threshold): Make const.
20287 (exec_threshold): Ditto.
20289 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20290 Pat Haugen <pthaugen@us.ibm.com>
20291 Peter Bergner <bergner@vnet.ibm.com>
20293 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
20294 documentation for the power8 crypto builtins.
20296 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
20298 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
20299 macros for defining power8 builtin functions.
20300 (BU_P8V_AV_2): Likewise.
20301 (BU_P8V_AV_P): Likewise.
20302 (BU_P8V_VSX_1): Likewise.
20303 (BU_P8V_OVERLOAD_1): Likewise.
20304 (BU_P8V_OVERLOAD_2): Likewise.
20305 (BU_CRYPTO_1): Likewise.
20306 (BU_CRYPTO_2): Likewise.
20307 (BU_CRYPTO_3): Likewise.
20308 (BU_CRYPTO_OVERLOAD_1): Likewise.
20309 (BU_CRYPTO_OVERLOAD_2): Likewise.
20310 (XSCVSPDP): Fix typo, point to the correct instruction.
20311 (VCIPHER): Add power8 crypto builtins.
20312 (VCIPHERLAST): Likewise.
20313 (VNCIPHER): Likewise.
20314 (VNCIPHERLAST): Likewise.
20315 (VPMSUMB): Likewise.
20316 (VPMSUMH): Likewise.
20317 (VPMSUMW): Likewise.
20318 (VPERMXOR_V2DI): Likewise.
20319 (VPERMXOR_V4SI: Likewise.
20320 (VPERMXOR_V8HI: Likewise.
20321 (VPERMXOR_V16QI: Likewise.
20322 (VSHASIGMAW): Likewise.
20323 (VSHASIGMAD): Likewise.
20324 (VPMSUM): Likewise.
20325 (VPERMXOR): Likewise.
20326 (VSHASIGMA): Likewise.
20328 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
20329 __CRYPTO__ if the crypto instructions are available.
20330 (altivec_overloaded_builtins): Add support for overloaded power8
20333 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
20334 support for power8 crypto builtins.
20335 (builtin_function_type): Likewise.
20336 (altivec_init_builtins): Add support for builtins that take vector
20337 long long (V2DI) arguments.
20339 * config/rs6000/crypto.md: New file, define power8 crypto
20342 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20343 Pat Haugen <pthaugen@us.ibm.com>
20344 Peter Bergner <bergner@vnet.ibm.com>
20346 * doc/invoke.texi (Option Summary): Add power8 options.
20347 (RS/6000 and PowerPC Options): Likewise.
20349 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
20350 constraints.md instead of rs6000.h. Reorder w* constraints. Add
20351 wm, wn, wr documentation.
20353 * config/rs6000/constraints.md (wm): New constraint for VSX
20354 registers if direct move instructions are enabled.
20355 (wn): New constraint for no registers.
20356 (wq): New constraint for quad word even GPR registers.
20357 (wr): New constraint if 64-bit instructions are enabled.
20358 (wv): New constraint if power8 vector instructions are enabled.
20359 (wQ): New constraint for quad word memory locations.
20361 * config/rs6000/predicates.md (const_0_to_15_operand): New
20362 constraint for 0..15 for crypto instructions.
20363 (gpc_reg_operand): If VSX allow registers in VSX registers as well
20364 as GPR and floating point registers.
20365 (int_reg_operand): New predicate to match only GPR registers.
20366 (base_reg_operand): New predicate to match base registers.
20367 (quad_int_reg_operand): New predicate to match even GPR registers
20368 for quad memory operations.
20369 (vsx_reg_or_cint_operand): New predicate to allow vector logical
20370 operations in both GPR and VSX registers.
20371 (quad_memory_operand): New predicate for quad memory operations.
20372 (reg_or_indexed_operand): New predicate for direct move support.
20374 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
20375 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
20376 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
20377 (POWERPC_MASKS): Add power8 options.
20378 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
20381 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
20382 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
20384 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
20385 (-mpower8-fusion): New power8 options.
20386 (-mpower8-fusion-sign): Likewise.
20387 (-mpower8-vector): Likewise.
20388 (-mcrypto): Likewise.
20389 (-mdirect-move): Likewise.
20390 (-mquad-memory): Likewise.
20392 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
20393 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
20395 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
20396 (rs6000_debug_vector_unit): Add p8_vector.
20397 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
20398 definitions. Also print fusion state.
20399 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
20400 (rs6000_builtin_mask_calculate): Add power8 builtin support.
20401 (rs6000_option_override_internal): Add support for power8.
20402 (rs6000_common_init_builtins): Add debugging for skipped builtins
20403 if -mdebug=builtin.
20404 (rs6000_adjust_cost): Add power8 support.
20405 (rs6000_issue_rate): Likewise.
20406 (insn_must_be_first_in_group): Likewise.
20407 (insn_must_be_last_in_group): Likewise.
20408 (force_new_group): Likewise.
20409 (rs6000_register_move_cost): Likewise.
20410 (rs6000_opt_masks): Likewise.
20412 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
20413 power8 capable assembler, default to power7 options.
20414 (TARGET_DIRECT_MOVE): Likewise.
20415 (TARGET_CRYPTO): Likewise.
20416 (TARGET_P8_VECTOR): Likewise.
20417 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
20418 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
20419 (VECTOR_MEM_P8_VECTOR_P): Likewise.
20420 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
20421 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
20422 (TARGET_XSCVDPSPN): Likewise.
20423 (TARGET_XSCVSPDPN): Likewsie.
20424 (TARGET_SYNC_HI_QI): Likewise.
20425 (TARGET_SYNC_TI): Likewise.
20426 (MASK_CRYPTO): Likewise.
20427 (MASK_DIRECT_MOVE): Likewise.
20428 (MASK_P8_FUSION): Likewise.
20429 (MASK_P8_VECTOR): Likewise.
20430 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
20431 temporary used by some of the direct move instructions to get two FP
20432 temporary registers does not force creation of a stack frame.
20433 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
20434 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
20435 that any VSX registers are tieable, even if they are also an
20436 Altivec vector mode.
20437 (r6000_reg_class_enum): Add wm, wr, wv constraints.
20438 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
20439 (RS6000_BTM_CRYPTO): Likewise.
20440 (RS6000_BTM_COMMON): Likewise.
20442 * config/rs6000/rs6000.md (cpu attribute): Add power8.
20443 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
20444 (enum rs6000_vector): Add power8 vector support.
20446 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20450 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
20451 (any_sibcall_could_use_r3): this and handle indirect calls.
20452 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
20454 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20456 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
20458 2013-05-22 Richard Biener <rguenther@suse.de>
20460 PR middle-end/57349
20461 * profile.c (branch_prob): Do not split blocks that are
20462 abnormally receiving from ECF_RETURNS_TWICE functions.
20464 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20466 * recog.c (offsettable_address_addr_space_p): Fix calculation of
20467 address mode. Move pointer mode initialization to the same place.
20469 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
20471 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
20472 while it has any effect.
20474 2013-05-21 Easwaran Raman <eraman@google.com>
20476 PR tree-optimization/57322
20477 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
20478 UID of the statement added to the BB to be 1.
20480 2013-05-21 Jakub Jelinek <jakub@redhat.com>
20482 PR tree-optimization/57331
20483 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
20484 of conversion from pointer type to integral type with integer.
20486 2013-05-21 Martin Jambor <mjambor@suse.cz>
20489 * ipa-prop.c (ipa_read_node_info): Process param_used and
20490 controlled_uses in the same order as when writing.
20492 2013-05-21 Magnus Granberg <baldrick@free.fr>
20495 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
20497 2013-05-21 Richard Biener <rguenther@suse.de>
20499 PR tree-optimization/57318
20500 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
20501 estimate stmts with side-effects as likely eliminated.
20503 2013-05-21 Richard Biener <rguenther@suse.de>
20505 PR tree-optimization/57330
20506 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
20507 preserve the call stmts fntype.
20509 2013-05-21 Richard Biener <rguenther@suse.de>
20511 PR tree-optimization/57303
20512 * tree-ssa-sink.c (statement_sink_location): Improve killing
20513 stmt detection and properly handle self-assignments.
20515 2013-05-21 Christian Bruel <christian.bruel@st.com>
20517 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
20518 spanning registers. LEAF_REG_REMAP is supported only for contiguous
20519 registers. Set register size out of the PARALLEL loop.
20521 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
20524 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
20525 (*fmasf4, *fmasf4_media): New insns.
20527 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20529 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
20530 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
20531 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
20532 (mips_idiv_insns): Update the comments to say that the returned
20533 instruction counts are in units of BASE_INSN_LENGTH.
20534 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
20535 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
20536 using 2 rather than 4 as the length of indirect MIPS16 and
20537 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
20538 length of a NOP. Don't divide MIPS16 lengths by 2.
20539 (mips16_split_long_branches): Assume a branch is long if the
20540 length is greater than 4 rather than 8.
20541 * config/mips/mips.md (length): Give MIPS16 lengths directly,
20542 rather than multiplying them by 2. Multiply instruction counts
20543 by BASE_INSN_LENGTH rather than 4.
20544 (*jump_mips16, tls_get_tp_mips16_<mode>)
20545 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
20547 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20549 * config/mips/mips.md (extended_mips16): Remove branch case.
20550 (length): Remove duplicated extended_mips16 test.
20552 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20554 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
20556 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
20558 * recog.h (Recog_data): Rename to...
20559 (recog_data_d): ...this.
20560 (recog_data): Update accordingly.
20561 * recog.c (recog_data): Likewise.
20562 * reload.c (save_recog_data): Likewise.
20563 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
20564 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
20566 2013-05-17 Julian Brown <julian@codesourcery.com>
20568 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
20569 found in a REG_EQUAL note, invalidate it.
20571 2013-05-17 Easwaran Raman <eraman@google.com>
20573 * tree-ssa-reassoc.c (find_insert_point): New function.
20574 (insert_stmt_after): Likewise.
20575 (get_def_stmt): Likewise.
20576 (ensure_ops_are_available): Likewise.
20577 (not_dominated_by): Likewise.
20578 (rewrite_expr_tree): Do not move statements beyond what is
20579 necessary. Remove call to swap_ops_for_binary_stmt...
20580 (reassociate_bb): ... and move it here.
20581 (build_and_add_sum): Assign UIDs for new statements.
20582 (linearize_expr): Likewise.
20583 (do_reassoc): Renumber gimple statement UIDs.
20585 2013-05-17 Jan Hubicka <jh@suse.cz>
20587 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
20589 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
20590 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
20592 (output_weakrefs): Update.
20594 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
20595 Martin Jambor <mjambor@suse.cz>
20597 PR middle-end/57276
20598 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
20599 value that corresponds to the given aggval is found in values vector.
20601 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
20603 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
20604 sse, sse2, sse3, ssse3 and sse4a flags to options.
20606 2013-05-17 David Malcolm <dmalcolm@redhat.com>
20608 * gengtype-state.c: (s_expr_writer): New class, to handle
20609 prettifying of output layout of s-expressions.
20610 (state_writer): New class, to write out gtype.state.
20611 (state_written_type_count): Move this variable into member data of
20613 (s_expr_writer::s_expr_writer): New code: constructor for new class
20614 (state_writer::state_writer(): ditto
20615 (s_expr_writer::write_new_line): New function
20616 (s_expr_writer::write_any_indent): ditto
20617 (s_expr_writer::begin_s_expr): ditto
20618 (s_expr_writer::end_s_expr): ditto
20619 (write_state_fileloc): convert to method of state_writer...
20620 (state_writer:: write_state_fileloc): ...and use methods of
20621 s_expr_writer to write indentation into the gtype.state output file
20622 to visually represent the hierarchical structure of the list
20624 (write_state_fields): ditto, renaming to...
20625 (state_writer::write_state_fields)
20626 (write_state_a_string): ditto, renaming to...
20627 (state_writer::write_state_a_string)
20628 (write_state_string_option): ditto, renaming to...
20629 (state_writer::write_state_string_option)
20630 (write_state_type_option): ditto, renaming to...
20631 (state_writer::write_state_type_option)
20632 (write_state_nested_option): ditto, renaming to...
20633 (state_writer::write_state_nested_option)
20634 (write_state_option): ditto, renaming to...
20635 (state_writer::write_state_option)
20636 (write_state_options): ditto, renaming to...
20637 (state_writer::write_state_options)
20638 (write_state_lang_bitmap): ditto, renaming to...
20639 (state_writer::write_state_lang_bitmap)
20640 (write_state_version): ditto, renaming to...
20641 (state_writer::write_state_version)
20642 (write_state_scalar_type): ditto, renaming to...
20643 (state_writer::write_state_scalar_type)
20644 (write_state_string_type): ditto, renaming to...
20645 (state_writer::write_state_string_type)
20646 (write_state_undefined_type): ditto, renaming to...
20647 (state_writer::write_state_undefined_type)
20648 (write_state_struct_union_type): ditto, renaming to...
20649 (state_writer::write_state_struct_union_type)
20650 (write_state_struct_type): ditto, renaming to...
20651 (state_writer::write_state_struct_type)
20652 (write_state_user_struct_type): ditto, renaming to...
20653 (state_writer::write_state_user_struct_type)
20654 (write_state_lang_struct_type): ditto, renaming to...
20655 (state_writer::write_state_lang_struct_type)
20656 (write_state_param_struct_type): ditto, renaming to...
20657 (state_writer::write_state_param_struct_type)
20658 (write_state_pointer_type): ditto, renaming to...
20659 (state_writer::write_state_pointer_type)
20660 (write_state_array_type): ditto, renaming to...
20661 (state_writer::write_state_array_type)
20662 (write_state_gc_used): ditto, renaming to...
20663 (state_writer::write_state_gc_used)
20664 (write_state_common_type_content): ditto, renaming to...
20665 (state_writer::write_state_common_type_content)
20666 (write_state_type): ditto, renaming to...
20667 (state_writer::write_state_type)
20668 (write_state_pair_list): ditto, renaming to...
20669 (state_writer::write_state_pair_list)
20670 (write_state_pair): ditto, renaming to...
20671 (state_writer::write_state_pair)
20672 (write_state_typedefs): ditto, renaming to...
20673 (state_writer::write_state_typedefs)
20674 (write_state_structures): ditto, renaming to...
20675 (state_writer::write_state_structures)
20676 (write_state_param_structs): ditto, renaming to...
20677 (state_writer::write_state_param_structs)
20678 (write_state_variables): ditto, renaming to...
20679 (state_writer::write_state_variables)
20680 (write_state_srcdir): ditto, renaming to...
20681 (state_writer::write_state_srcdir)
20682 (write_state_files_list): ditto, renaming to...
20683 (state_writer::write_state_files_list)
20684 (write_state_languages): ditto, renaming to...
20685 (state_writer::write_state_languages)
20686 (write_state): create a state_writer instance and use it when
20687 writing out the state file
20689 2013-05-17 Mike Stump <mikestump@comcast.net>
20691 PR rtl-optimization/57304
20692 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
20693 accessing DF_REF_REAL_LOC.
20695 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20697 PR rtl-optimization/57281
20698 PR rtl-optimization/57300
20699 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
20700 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
20701 what the other splitter did if the registers are dead.
20703 2013-05-17 Richard Biener <rguenther@suse.de>
20705 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
20708 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20710 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
20713 2013-05-17 Marek Polacek <polacek@redhat.com>
20715 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
20716 length when doing non-zero store of storing '\0' to '\0'.
20718 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20720 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
20721 vect_external_def oprnd1 with loop_vinfo, try to emit
20722 optional cast, negation and and stmts on the loop preheader
20723 edge instead of into the pattern def seq.
20725 PR tree-optimization/57051
20726 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
20727 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
20729 2013-05-16 Nick Clifton <nickc@redhat.com>
20731 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20732 (rl78_is_naked_func): New function.
20733 (rl78_expand_prologue): Skip prologue generation for naked functions.
20734 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20735 * doc/extend.texi (naked): Add RL78 to the list of processors
20736 that supports this attribute.
20738 2013-05-16 Jeff Law <law@redhat.com>
20740 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20742 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20744 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20745 cache parameters using detect_caches_amd also for CYRIX,
20746 NSC and TM2 signatures.
20748 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20749 Dzianis Kahanovich <mahatma@eu.by>
20753 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20754 VIA/Centaur processors and determine their cache parameters
20755 using detect_caches_amd.
20757 2013-05-16 Teresa Johnson <tejohnson@google.com>
20759 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20760 (rtl_verify_edges): New function.
20761 (rtl_verify_bb_insns): Ditto.
20762 (rtl_verify_bb_pointers): Ditto.
20763 (rtl_verify_bb_insn_chain): Ditto.
20764 (rtl_verify_fallthru): Ditto.
20765 (rtl_verify_bb_layout): Ditto.
20766 (rtl_verify_flow_info_1): Outline checks into new functions.
20767 (rtl_verify_flow_info): Ditto.
20769 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20771 * cfghooks.c (copy_bbs): Add update_dominance argument.
20772 * cfghooks.h (copy_bbs): Update prototype.
20773 * tree-cfg.c (gimple_duplicate_sese_region):
20774 Add update_dominance argument.
20775 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20776 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20777 gimple_duplicate_sese_region call.
20778 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20779 Update copy_bbs call.
20780 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20781 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20783 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20785 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20786 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20787 vect_recog_rotate_pattern.
20788 (vect_recog_rotate_pattern): New function.
20790 2013-05-16 Jason Merrill <jason@redhat.com>
20792 * Makefile.in (LLINKER): New variable.
20793 (mostlyclean): Remove link mutex.
20794 * configure.ac: Handle --enable-link-mutex.
20795 * lock-and-run.sh: New script.
20797 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20800 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20803 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20805 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20807 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20809 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20810 * config/arm/arm.c (next_consecutive_mem): New function.
20811 (gen_movmem_ldrd_strd): Likewise.
20812 * config/arm/arm.md (movmemqi): Update condition and code.
20813 (unaligned_loaddi, unaligned_storedi): New patterns.
20815 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20817 * config.gcc: Obsolete *-*-solaris2.9*.
20818 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20820 2013-05-16 Richard Biener <rguenther@suse.de>
20822 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20823 earlier, after GRAPHITE transforms and IV canonicalization.
20825 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20827 * omp-low.c (extract_omp_for_data): For collapsed loops,
20828 if at least one of the loops is known at compile time to
20829 iterate zero times, set count to 0.
20830 (expand_omp_regimplify_p): New function.
20831 (expand_omp_for_generic): For collapsed loops, if at least
20832 one of the loops isn't known to iterate at least once,
20833 add runtime check with setting count to 0.
20834 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20835 For unsigned types if it isn't known at compile time that
20836 the loop will iterate at least once, add runtime check to bypass
20837 the whole loop if initial condition isn't true.
20839 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20841 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20843 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20845 PR middle-end/57286
20846 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20847 transformations to avoid an infinite loop.
20849 2013-05-16 Marek Polacek <polacek@redhat.com>
20851 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20853 2013-05-15 Leif Ekblad <leif@rdos.net>
20855 * config/i386/i386.c (ix86_decompose_address): Use
20856 DEFAULT_TLS_SEG_REG to access TLS segment register.
20857 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20858 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20859 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20861 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20864 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20865 sibling calls to functions that would normally be lazily bound,
20866 unless $gp is call-clobbered.
20868 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20870 * config/i386/i386.c (ix86_option_override_internal): Update
20871 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20872 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20873 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20874 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20876 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20878 2013-05-15 Andreas Schwab <schwab@suse.de>
20880 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20881 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20884 2013-05-15 Teresa Johnson <tejohnson@google.com>
20886 * loop-unroll.c (report_unroll_peel): Check decision before
20887 emitting unroll/peel message.
20889 2013-05-15 Teresa Johnson <tejohnson@google.com>
20891 * function.h (has_bb_partition): New rtl_data flag.
20892 (bb_reorder_complete): Ditto.
20893 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20894 instead of flag_reorder_blocks_and_partition.
20895 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20896 with some enhancements.
20897 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20898 * bb-reorder.c (connect_traces): Check for has_bb_partition
20899 instead of flag_reorder_blocks_and_partition.
20900 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20901 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20902 verify_hot_cold_block_grouping.
20903 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20905 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20908 * config/arm/predicates.md (call_insn_operand): New predicate.
20909 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20910 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20911 if insn is not a tail call.
20912 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20914 * config/arm/arm.h (enum reg_class): New caller save register class.
20915 (REG_CLASS_NAMES): Likewise.
20916 (REG_CLASS_CONTENTS): Likewise.
20917 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20920 2013-05-15 Richard Biener <rguenther@suse.de>
20922 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20923 of MSG_OPTIMIZED_LOCATIONS.
20924 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20925 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20927 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20928 of MSG_OPTIMIZED_LOCATIONS.
20929 (execute_vect_slp): Likewise.
20930 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20931 (vect_create_cond_for_alias_checks): Likewise.
20932 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20933 (vect_recog_widen_mult_pattern): Likewise.
20934 (vect_recog_widen_sum_pattern): Likewise.
20935 (vect_recog_over_widening_pattern): Likewise.
20936 (vect_recog_widen_shift_pattern): Likewise.
20937 (vect_recog_vector_vector_shift_pattern): Likewise.
20938 (vect_recog_divmod_pattern): Likewise.
20939 (vect_recog_mixed_size_cond_pattern): Likewise.
20940 (vect_recog_bool_pattern): Likewise.
20941 (vect_pattern_recog_1): Likewise.
20943 2013-05-15 Martin Jambor <mjambor@suse.cz>
20945 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20946 non-functions to builtin_unreachable.
20947 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20948 correct when new direct edges were discovered.
20950 2013-05-15 Martin Jambor <mjambor@suse.cz>
20952 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20953 header, print symbol order instead of node uid, print more information
20954 about indirect edge targets.
20955 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20957 (ipa_make_edge_direct_to_target): Likewise.
20958 (remove_described_reference): Likewise.
20959 (propagate_controlled_uses): Likewise.
20960 (ipa_print_node_params): Also print symbol order.
20961 (ipcp_transform_function): Print symbol order instead of node uids.
20962 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20963 (cgraph_get_create_real_symbol_node): Likewise.
20964 * ipa-cp.c (print_lattice): Likewise.
20965 (print_all_lattices): Likewise.
20966 (determine_versionability): Likewise.
20967 (initialize_node_lattices): Likewise.
20968 (estimate_local_effects): Likewise.
20969 (update_profiling_info): Likewise.
20970 (create_specialized_node): Likewise.
20971 (perhaps_add_new_callers): Likewise.
20972 (decide_about_value): Likewise.
20973 (decide_whether_version_node): Likewise.
20974 (identify_dead_nodes): Likewise.
20975 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20976 (dump_inline_summary): Likewise.
20977 (estimate_node_size_and_time): Likewise.
20978 (inline_analyze_function): Likewise.
20979 * ipa-inline.c (report_inline_failed_reason): Likewise.
20980 (want_early_inline_function_p): Likewise.
20981 (edge_badness): Likewise.
20982 (update_edge_key): Likewise.
20983 (inline_small_functions): Likewise. Add dumping of order to two other
20985 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20986 instead of node uids.
20987 (propagate_pure_const): Likewise.
20988 (propagate_pure_const): Likewise.
20989 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20990 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20991 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20993 * tree-pretty-print.c (dump_function_header): Likewise.
20994 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20995 Print symbol order instead of node uids.
20997 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20999 * config/s390/s390.c (s390_register_move_cost): Don't impose the
21000 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
21002 2013-05-15 Richard Biener <rguenther@suse.de>
21004 PR tree-optimization/57275
21005 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
21006 return value for fail to do runtime alias checks for gather loads.
21008 2013-05-15 Jan Hubicka <jh@suse.cz>
21012 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
21013 weakrefs are not external.
21014 (lto_symtab_merge_decls): Fix thinko when dealing with
21015 non-lto_symtab decls.
21016 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
21017 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
21018 * varpool.c (dump_varpool_node): Dump more flags.
21020 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21022 * config/i386/i386.c (processor_alias_table): Add instruction
21023 FSGSBASE for AMD bdver3 architecture.
21025 2013-05-14 Jakub Jelinek <jakub@redhat.com>
21027 * tree.c (warn_deprecated_use): Print file:line using locus color.
21028 * diagnostic.c (diagnostic_report_current_module): Print file:line
21029 and file:line:column using locus color.
21031 2013-05-14 Mike Stump <mikestump@comcast.net>
21033 * gdbinit.in: Add __null.
21035 2013-05-14 Mike Stump <mikestump@comcast.net>
21037 * recog.h: Rename struct recog_data to Recog_data.
21038 * recog.c: Likewise.
21039 * reload.c (can_reload_into): Likewise.
21040 * config/picochip/picochip.c: Likewise.
21042 2013-05-14 Mike Stump <mikestump@comcast.net>
21044 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
21046 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
21048 * resource.h (struct resources): Remove unch_memory member.
21049 (CLEAR_RESOURCE): Don't clear unch_memory.
21050 * resource.c (mark_referenced_resources): Don't set it.
21051 (mark_set_resources): Likewise.
21052 (mark_target_live_regs): Don't clear it.
21053 (init_resource_info): Likewise.
21054 * reorg.c (resource_conflicts_p): Don't compare it.
21055 (redundant_insn): Don't set it.
21057 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
21059 * emit-rtl.c (next_label): Remove unused function.
21060 (skip_consecutive_labels, link_cc0_insns): Move to ...
21061 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
21062 only place where these functions are used, and make them static.
21064 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
21066 * fold-const.c (fold_negate_expr): Handle vectors.
21067 (fold_truth_not_expr): Make it static.
21068 (fold_invert_truthvalue): New static function.
21069 (invert_truthvalue_loc): Handle vectors. Do not call
21070 fold_truth_not_expr directly.
21071 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
21072 <TRUTH_NOT_EXPR>: Do not cast to boolean.
21073 (fold_comparison): Handle vector constants.
21074 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
21075 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
21076 * tree.h (fold_truth_not_expr): Remove declaration.
21078 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
21080 * config/aarch64/aarch64-simd.md
21081 (aarch64_vcond_internal<mode>): Rename to...
21082 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
21083 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
21084 float modes. Clarify all iterator modes.
21085 (vcond<mode><mode>): Use new name for vcond expanders.
21086 (vcond<v_cmp_result><mode>): Likewise.
21087 (vcondu<mode><mode>: Likewise.
21088 * config/aarch64/iterators.md (VDQF_COND): New.
21090 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
21093 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
21094 variable for the shift amount. Check that we shift by non-negative
21097 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
21100 * config/arm/arm.h (EPILOGUE_USES): Only return true
21101 for LR_REGNUM after epilogue_completed.
21103 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
21105 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
21106 is error_mark_node.
21108 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21111 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
21112 and Solaris 11+/x86 with gld.
21113 * configure: Regenerate.
21115 2013-05-14 Jakub Jelinek <jakub@redhat.com>
21117 * expmed.c (expand_shift_1): Canonicalize rotates by
21118 constant bitsize / 2 to bitsize - 1.
21119 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
21120 case ROTATERT>: Likewise.
21123 2013-05-10 Jakub Jelinek <jakub@redhat.com>
21125 * config/i386/i386.md (rotateinv): New code attr.
21126 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
21127 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
21128 roll $31, %eax, etc.
21130 2013-05-14 Richard Biener <rguenther@suse.de>
21132 PR middle-end/57235
21133 * tree-eh.c (sink_clobbers): Give up for successors with
21134 multiple predecessors and no virtual uses.
21136 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
21138 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
21139 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
21141 2013-05-14 Jakub Jelinek <jakub@redhat.com>
21143 PR middle-end/57251
21144 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
21145 the case when both op0 and op1 have VOIDmode.
21147 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
21149 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
21150 in multiply-accumulate mode.
21152 2013-05-13 Guozhi Wei <carrot@google.com>
21154 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
21156 2013-05-13 Kai Tietz <ktietz@redhat.com>
21159 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
21160 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
21161 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
21162 * config/i386/i386.c (ix86_option_override_internal): Likewise.
21163 (ix86_expand_prologue): Likewise.
21164 (ix86_expand_split_stack_prologue): Likewise.
21165 (legitimate_pic_address_disp_p): Likewise.
21166 (legitimize_pic_address): Likewise.
21167 (legitimize_tls_address): Likewise.
21168 (legitimize_pe_coff_symbol): Likewise.
21169 (output_pic_addr_const): Likewise.
21170 (construct_plt_address): Likewise.
21171 (ix86_expand_call): Likewise.
21172 (x86_output_mi_thunk): Likewise.
21173 (x86_function_profiler): Likewise.
21175 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
21177 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
21178 similar switch cases.
21179 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
21180 (aarch64_simd_mov_to_<mode>low): Delete.
21181 (aarch64_simd_mov_to_<mode>high): Delete.
21182 (move_lo_quad_<mode>): Add w<-r alternative.
21183 (aarch64_simd_move_hi_quad_<mode>): Likewise.
21184 (aarch64_simd_mov_from_*): Update type attribute.
21185 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
21188 2013-05-13 Jan Hubicka <jh@suse.cz>
21190 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
21191 * config/i386/i386.c (ix86_compute_frame_layout,
21192 ix86_expand_epilogue, emit_i387_cw_initialization,
21193 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
21194 ix86_local_alignment): Fix use of size/speed predicates.
21196 2013-05-13 Jakub Jelinek <jakub@redhat.com>
21198 PR tree-optimization/45216
21199 PR tree-optimization/57157
21200 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
21201 the (-Y) & (B - 1) variant if OP is |.
21202 * expmed.c (expand_shift_1): For rotations by const0_rtx just
21203 return shifted. Use (-op1) & (prec - 1) as other_amount
21204 instead of prec - op1.
21206 2013-05-13 Martin Jambor <mjambor@suse.cz>
21208 PR middle-end/42371
21209 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
21210 (ipa_constant_data): New type.
21211 (ipa_jump_func): Use ipa_constant_data to hold information about
21212 constant jump functions.
21213 (ipa_get_jf_constant): Adjust to jump function type changes.
21214 (ipa_get_jf_constant_rdesc): New function.
21215 (ipa_param_descriptor): New field controlled_uses.
21216 (ipa_get_controlled_uses): New function.
21217 (ipa_set_controlled_uses): Likewise.
21218 * ipa-ref.h (ipa_find_reference): Declare.
21219 * ipa-prop.c (ipa_cst_ref_desc): New type.
21220 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
21222 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
21223 New parameter cs. Adjust all callers.
21224 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
21225 (remove_described_reference): New function.
21226 (jfunc_rdesc_usable): Likewise.
21227 (try_make_edge_direct_simple_call): Decrement controlled use count,
21228 attempt to remove reference if it hits zero.
21229 (combine_controlled_uses_counters): New function.
21230 (propagate_controlled_uses): Likewise.
21231 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
21232 (ipa_edge_duplication_hook): Duplicate reference descriptions.
21233 (ipa_print_node_params): Print described use counter.
21234 (ipa_write_jump_function): Adjust to jump function type changes.
21235 (ipa_read_jump_function): New parameter CS, pass it to
21236 ipa_set_jf_constant. Adjust caller.
21237 (ipa_write_node_info): Stream controlled use count
21238 (ipa_read_node_info): Likewise.
21239 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
21241 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
21242 count. Remove cloning-added reference if it reaches zero.
21243 * ipa-ref.c (ipa_find_reference): New function.
21245 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21247 * config/i386/i386.c (processor_target_table): Modified default
21248 alignment values for AMD BD and BT architectures.
21250 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
21252 * tree-vect-generic.c (uniform_vector_p): Move ...
21253 * tree.c (uniform_vector_p): ... here.
21254 * tree.h (uniform_vector_p): Declare it.
21255 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
21258 2013-05-13 Jakub Jelinek <jakub@redhat.com>
21260 PR tree-optimization/57230
21261 * tree-ssa-strlen.c (handle_char_store): Record length for
21262 array store from STRING_CST.
21264 PR tree-optimization/57230
21265 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
21268 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
21270 * config/epiphany/epiphany.c (epiphany_init): Check size of
21271 NUM_MODES_FOR_MODE_SWITCHING.
21272 (epiphany_expand_prologue):
21273 Remove CONFIG_REGNUM initial value handling code.
21274 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
21275 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
21276 (emit_set_fp_mode, epiphany_mode_after): Likewise.
21277 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
21278 Don't return 1 for FP_MODE_NONE.
21279 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
21280 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
21281 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
21282 * config/epiphany/epiphany.md (save_config): New pattern.
21284 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
21286 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
21288 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
21290 * config/i386/i386.md (memory): Handle sseishft1.
21291 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
21292 (*vec_extractv2di_1): Ditto.
21294 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
21296 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
21299 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
21301 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
21302 Add mthumb/march=armv7-a multilib.
21303 Add mthumb/march=armv7-r multilib.
21304 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
21306 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
21308 * config/v850/t-rtems: Add more multilibs.
21310 2013-05-10 Richard Biener <rguenther@suse.de>
21312 PR tree-optimization/57214
21313 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
21314 not propagate from SSA names that occur in abnormal PHI nodes.
21316 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
21318 * stor-layout.c (element_precision): New function.
21319 * machmode.h (element_precision): Declare it.
21320 * tree.c (build_minus_one_cst): New function.
21321 (element_precision): Likewise.
21322 * tree.h (build_minus_one_cst): Declare new function.
21323 (element_precision): Likewise.
21324 * fold-const.c (operand_equal_p): Use element_precision.
21325 (fold_binary_loc): Handle vector types.
21326 * convert.c (convert_to_integer): Use element_precision.
21327 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
21330 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
21332 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
21333 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
21334 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
21335 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21336 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
21337 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
21338 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
21339 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21340 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
21341 (Uuw8): New constraints.
21342 (Usb4): Move into alphabetical order.
21343 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
21344 (sd8_operand, ub8_operand, uw8_operand): New predicates.
21345 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
21346 previously unnamed patterns.
21347 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
21348 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
21349 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
21350 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
21351 of set_attr_alternative/if_then_else. Use extended_mips16 instead
21352 of specific lengths.
21354 2013-05-10 Jakub Jelinek <jakub@redhat.com>
21356 * config/i386/i386.md (rotateinv): New code attr.
21357 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
21358 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
21359 roll $31, %eax, etc.
21361 PR tree-optimization/45216
21362 PR tree-optimization/57157
21363 * tree-ssa-forwprop.c (simplify_rotate): New function.
21364 (ssa_forward_propagate_and_combine): Call it.
21366 2013-05-10 Richard Biener <rguenther@suse.de>
21368 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
21369 disable peeling when we version for aliasing.
21370 (vector_alignment_reachable_p): Honor explicit user alignment.
21371 (vect_supportable_dr_alignment): Likewise.
21372 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
21373 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
21374 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
21375 then peeling to arrange for the cost-model check to come first.
21377 2013-05-10 Alan Modra <amodra@gmail.com>
21379 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
21380 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
21381 * configure: Regenerate.
21383 2013-05-10 Alan Modra <amodra@gmail.com>
21386 * varasm.c (default_elf_select_section): Move !DECL_P check..
21387 (get_named_section): ..to here before calling get_section_name.
21389 (default_section_type_flags): Add DECL_P check.
21390 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
21391 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
21393 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
21395 * config/epiphany/epiphany.c (epiphany_expand_prologue):
21396 When using gen_stack_adjust_str with a register offset, add a
21397 REG_FRAME_RELATED_EXPR note.
21399 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
21401 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
21402 (*vec_extractv4si_zext_mem): Ditto.
21403 (*vec_extractv2di): Add 0->x and x->x alternatives.
21404 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
21405 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
21407 2013-05-09 Jason Merrill <jason@redhat.com>
21409 N3639 C++1y VLA support
21410 * gimplify.c (gimplify_vla_decl): Don't touch an existing
21413 * tree.c (build_constructor_va): New.
21414 * tree.h: Declare it.
21416 2013-05-09 Martin Jambor <mjambor@suse.cz>
21419 * gimple-fold.c (canonicalize_constructor_val): Call
21420 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
21422 2013-05-09 Jan Hubicka <jh@suse.cz>
21423 Richard Biener <rguenther@suse.de>
21426 * symtab.c (symtab_make_decl_local): Do not add private names.
21428 2013-05-09 Jan Hubicka <jh@suse.cz>
21431 * symtab.c (insert_to_assembler_name_hash): Handle clones.
21432 (unlink_from_assembler_name_hash): Likewise.
21433 (symtab_prevail_in_asm_name_hash, symtab_register_node,
21434 symtab_unregister_node, symtab_initialize_asm_name_hash,
21435 change_decl_assembler_name): Update.
21437 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
21439 * config/aarch64/aarch64.md: New movtf split.
21440 (*movtf_aarch64): Update.
21441 (aarch64_movdi_tilow): Handle TF modes and rename to
21442 aarch64_movdi_<mode>low.
21443 (aarch64_movdi_tihigh): Handle TF modes and rename to
21444 aarch64_movdi_<mode>high
21445 (aarch64_movtihigh_di): Handle TF modes and rename to
21446 aarch64_mov<mode>high_di
21447 (aarch64_movtilow_di): Handle TF modes and rename to
21448 aarch64_mov<mode>low_di
21449 (aarch64_movtilow_tilow): Remove spurious whitespace.
21450 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
21452 (aarch64_print_operand): Update.
21454 2013-05-09 Alan Modra <amodra@gmail.com>
21456 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
21458 * configure: Regenerate.
21460 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21462 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
21463 splitter preparation statements.
21464 * config/i386/sse.md (*vec_extract* splitters): Ditto.
21465 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
21468 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21470 * gimple-ssa-strength-reduction.c (count_candidates): Change
21471 return value to int.
21472 (analyze_candidates_and_replace): Change type of length to int.
21474 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21476 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
21477 (*vec_extract<mode>): Use VI12_128 mode iterator.
21478 (*vec_extract<mode>_mem): Ditto.
21479 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
21482 2013-05-08 Diego Novillo <dnovillo@google.com>
21487 2012-08-17 Diego Novillo <dnovillo@google.com>
21490 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
21491 * config.in: Regenerate.
21492 * configure: Regenerate.
21493 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
21495 2013-05-08 Jan Hubicka <jh@suse.cz>
21498 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
21499 * cgraph.h (symtab_node_base): Add unique_name.
21500 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
21501 input_overwrite_node, input_varpool_node): Stream unique_name.
21502 * cgraphclones.c (cgraph_create_virtual_clone,
21503 cgraph_function_versioning): Set unique_name.
21504 * ipa.c (function_and_variable_visibility): Set unique_name.
21506 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21508 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
21509 (alloc_cand_and_find_basis): Restrict conditional candidate
21510 processing to CAND_MULTs.
21512 2013-05-08 Jan Hubicka <jh@suse.cz>
21515 lto-symtab.c (lto_symtab_symbol_p): New function.
21516 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
21517 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
21518 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
21519 Skip static symbols.
21521 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
21523 PR tree-optimization/57200
21524 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
21525 Only call inform if the preceding warning_at returns true.
21527 2013-05-07 Han Shen <shenhan@google.com>
21529 * cfgexpand.c (record_or_union_type_has_array_p): New function.
21530 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
21531 * common.opt (fstack-protector-strong): New option.
21532 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
21533 * doc/invoke.texi (Optimization Options): Document
21534 "-fstack-protector-strong".
21535 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
21537 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21539 * config/mips/mips.c (mips_machine_reorg2): Return 0.
21541 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
21543 * ira.c (update_equiv_regs): Add insn having equiv memory even if
21544 it is not lhs of the insn.
21545 (setup_reg_equiv): Remove insn having equiv memory which it is not
21547 * lra-constraints.c (process_address): Try to improve generation
21548 code for address base + disp.
21549 (lra_constraints): Make correct the code for checking insn setting
21550 up backward equivalence. Remove insn only if it is in the init
21552 * lra-eliminations.c (update_reg_eliminate): Change return value.
21553 (lra_eliminate): Use the result.
21555 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
21557 * config/i386/sse.md (ssescalarnummask): New mode attribute.
21558 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
21559 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
21560 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
21561 register target operands.
21562 (*vec_extractv8hi_sse2): New pattern.
21563 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
21564 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
21565 (*vec_extract<mode>_mem): New insn and split pattern.
21567 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
21569 * config/arm/arm.c (arm_asan_shadow_offset): New function.
21570 (TARGET_ASAN_SHADOW_OFFSET): Define.
21571 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
21572 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
21574 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21576 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
21577 (incr_vec_index): Return -1 if increment not found.
21578 (create_add_on_incoming_edge): Assert if increment not found.
21579 (record_increment): Limit number of increments recorded.
21580 (all_phi_incrs_profitable): Return false if an increment not found.
21581 (replace_profitable_candidates): Don't process increments that were
21583 (analyze_candidates_and_replace): Limit size of incr_vec.
21585 2013-05-07 Richard Biener <rguenther@suse.de>
21587 * calls.c (special_function_p): setjmp-like functions are leaf.
21588 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
21589 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
21591 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
21593 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
21594 (aarch64_simd_mov<mode>): New expander.
21595 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
21596 (aarch64_simd_mov_to_<mode>high): Likewise.
21597 (aarch64_simd_mov_from_<mode>low): Likewise.
21598 (aarch64_simd_mov_from_<mode>high): Likewise.
21599 (aarch64_dup_lane<mode>): Update.
21600 (aarch64_dup_lanedi): New instruction pattern.
21601 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
21602 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
21604 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21606 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
21607 (replace_mult_candidate): Remove unnecessary argument; remove
21608 unnecessary parameter from call to introduce_cast_before_cand.
21609 (replace_unconditional_candidate): Remove unnecessary parameter
21610 from call to replace_mult_candidate.
21611 (replace_conditional_candidate): Likewise.
21612 (insert_initializers): Use make_temp_ssa_name.
21613 (introduce_cast_before_cand): Remove unnecessary argument; use
21614 make_temp_ssa_name.
21615 (replace_one_candidate): Remove unnecessary argument; remove
21616 unnecessary parameter from calls to introduce_cast_before_cand.
21617 (replace_profitable_candidates): Remove unnecessary parameters
21618 from calls to replace_one_candidate.
21620 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21622 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
21623 phi def as possibly hiding a basis for a CAND_ADD whose operands
21624 have been commuted in the analysis.
21625 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
21627 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21629 * config/aarch64/aarch64.md
21630 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
21631 shift value between 0-4.
21633 2013-05-07 Richard Biener <rguenther@suse.de>
21635 * double-int.h (rshift): New overload.
21636 * double-int.c (rshift): New function.
21637 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
21638 (create_reference_ops_from_ref): Remove.
21639 (vn_reference_insert): Use shared ops for constructing the
21640 reference and copy it.
21642 2013-05-07 Richard Biener <rguenther@suse.de>
21644 PR middle-end/57190
21645 * tree-eh.c (sink_clobbers): Properly propagate
21646 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
21648 2013-05-07 Jakub Jelinek <jakub@redhat.com>
21650 PR tree-optimization/57149
21651 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
21652 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
21653 collect_phi_def_edges, execute_late_warn_uninitialized): Use
21654 uninit_undefined_value_p instead of ssa_undefined_value_p.
21657 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
21658 for modifier == EXPAND_INITIALIZER.
21660 2013-05-07 Anton Blanchard <anton@samba.org>
21662 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
21663 for powerpc64 little endian.
21664 * configure: Regenerate.
21666 2013-05-06 Graham Stott <grahams@btinternet.com>
21668 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
21669 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
21670 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
21671 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
21673 2013-05-06 Graham Stott <grahams@btinternet.com>
21675 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
21676 codes which allow non-lvalues.
21678 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
21680 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
21681 components are all 1s.
21682 (integer_minus_onep): New function.
21683 * tree.h (integer_minus_onep): Declare it.
21684 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
21685 integer_minus_onep instead of integer_all_onesp.
21687 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21690 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
21691 variations of these patterns.
21693 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21695 * config/i386/i386.md (isa): Add x64_sse4 member.
21696 (enabled): Handle x64_sse4.
21697 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
21698 instruction for 64bit SSE4_1 targets. Update insn attributes.
21699 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
21700 instruction for SSE4_1 targets. Update insn attributes.
21701 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21702 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
21703 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
21705 (*vec_extractv4si): Rename from *sse4_1_pextrd.
21706 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
21707 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
21709 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21712 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
21714 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
21716 * final.c (do_assembler_dialects): Don't handle curly braces and
21717 vertical bar escaped by % as dialect delimiters.
21718 (output_asm_insn): Print curly braces and vertical bar if escaped
21719 by % and ASSEMBLER_DIALECT defined.
21720 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
21721 * doc/tm.texi: Regenerated.
21723 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21725 * config/mips/mips.c: Include tree-pass.h.
21726 (mips_reorg): Split in pre- and post-dbr_schedule parts.
21727 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
21728 (pass_mips_machine_reorg2): New machine specific pass.
21729 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21730 (mips_option_override): Register the new pass.
21731 * rtl.h (cleanup_barriers): Remove prototype.
21732 (dbr_schedule): Likewise.
21733 * jump.c (cleanup_barriers): Make static.
21734 * reorg.c (dbr_schedule): Likewise.
21736 2013-05-06 Richard Biener <rguenther@suse.de>
21738 PR tree-optimization/57185
21739 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21740 SSA names properly.
21742 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21745 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21746 between "first_arg" and "insn", not "last" and "insn".
21748 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21750 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21751 (find_candidates_in_block): Re-enable slsr_process_phi.
21752 (create_phi_basis): Fix double counting of candidate adjustment.
21754 2013-05-06 Richard Biener <rguenther@suse.de>
21756 PR middle-end/57147
21757 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21758 the edge is also fallthru, preserve it and just clear the
21760 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21761 also complex, preserve that and just clear the fallthru flag.
21762 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21763 update virtual operands.
21765 2013-05-06 Alan Modra <amodra@gmail.com>
21767 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21768 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21769 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21770 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21771 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21772 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21774 2013-05-06 Alan Modra <amodra@gmail.com>
21776 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21777 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21778 (DEFAULT_ASM_ENDIAN): Define.
21779 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21780 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21781 Update -K PIC clause from sysv4.h.
21782 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21783 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21785 2013-05-06 Alan Modra <amodra@gmail.com>
21787 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21788 twice for little-endian.
21789 (ashrdi3_no_power, ashrdi3): Support little-endian.
21791 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21794 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21795 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21797 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21798 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21799 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21802 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21803 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21805 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21806 * config/pa/pa.opt: Make mbig-switch a no-op.
21807 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21808 (CASE_VECTOR_MODE): Always return SImode.
21809 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21810 for the !TARGET_BIG_SWITCH case.
21811 * config/pa/pa-linux.h: Likewise.
21812 * config/pa/pa-openbsd.h: Likewise.
21813 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21814 * config/pa/pa.md (short_jump): Remove define_insn.
21815 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21816 (casesi0): Remove define_insn.
21817 (type): Remove btable_branch.
21818 (pa_combine_type): Likewise.
21819 (in_nullified_branch_delay): Likewise.
21820 (in_call_delay): Likewise.
21821 (define_delay): Likewise.
21822 (define_insn_reservation "Z3"): Likewise.
21823 (define_insn_reservation "Z4"): Likewise.
21824 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21825 (pa_adjust_insn_length): Remove adjustment for btable branches.
21826 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21829 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21831 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21832 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21833 Add m->r,x alternatives.
21834 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21835 splitters using SWI48x mode iterator.
21836 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21837 TARGET_64BIT. Add m->x alternative.
21838 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21839 Add o->x alternative. Enable for TARGET_SSE.
21840 (sse_storeq): Remove expander.
21841 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21842 with memory input operand.
21843 (*vec_extractv2di_1 splitter): New.
21844 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21845 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21847 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21849 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21850 (INT_LOWPART): Delete.
21851 (extract_MB): Adjust.
21852 (extract_ME): Adjust.
21853 (print_operand): Adjust.
21855 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21857 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21858 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21859 (reg_or_logical_cint_operand, easy_fp_constant,
21860 logical_const_operand): Delete "CONST_DOUBLE" case.
21861 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21862 "HOST_BITS_PER_WIDE_INT == 64" test.
21863 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21864 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21865 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21866 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21868 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21869 CONST_DOUBLE DImode/VOIDmode case.
21870 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21871 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21872 CONST_DOUBLE VOIDmode case.
21873 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21874 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21875 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21876 Delete CONST_DOUBLE case.
21877 (splitters for mov FMOVE64 const_double): Delete
21878 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21879 "HOST_BITS_PER_WIDE_INT >= 64" test.
21880 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21882 (mov DI const_double): Delete.
21884 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21886 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21887 on op shows all bits zero in mode of a lowpart subreg, return zero.
21889 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21892 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21893 to save TFmode registers and DImode to save TImode registers for
21894 caller save operations.
21895 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21896 mark being partially clobbered since they only use the first
21899 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21900 and TDmode only use the upper 64-bits of each VSX register.
21902 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21904 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21905 (find_candidates_in_block): Disable slsr_process_phi.
21907 2013-05-03 Guozhi Wei <carrot@google.com>
21909 * coverage.c (coverage_obj_init): Move the construction of gcov
21911 (build_init_ctor): ... here.
21913 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21915 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21916 (slsr_cand_d): Redefine def_phi.
21917 (stride_status, phi_adjust_status, count_phis_status): New enums.
21918 (find_phi_def): New.
21919 (find_basis_for_base_expr): New.
21920 (find_basis_for_candidate): Handle hidden bases.
21921 (alloc_cand_and_find_basis): Handle phi candidates.
21922 (slsr_process_phi): New.
21923 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21924 (create_mul_imm_cand): Likewise.
21925 (create_add_ssa_cand): Exclude phi base candidates.
21926 (create_add_imm_cand): Likewise.
21927 (slsr_process_cast): Likewise.
21928 (slsr_process_copy): Likewise.
21929 (find_candidates_in_block): Handle phi candidates.
21930 (dump_candidate): Likewise.
21931 (unconditional_cands): Delete.
21932 (unconditional_cands_with_known_stride_p): Delete.
21933 (phi_dependent_cand_p): New.
21934 (cand_increment): Handle phi-dependent candidates.
21935 (replace_dependent): Delete.
21936 (replace_mult_candidate): New.
21937 (replace_unconditional_candidate): New.
21938 (incr_vec_index): Move to avoid forward reference.
21939 (create_add_on_incoming_edge): New.
21940 (create_phi_basis): New.
21941 (replace_dependents): Delete.
21942 (replace_conditional_candidate): New.
21943 (phi_add_costs): New.
21944 (replace_uncond_cands_and_profitable_phis): New.
21945 (record_increment): Handle phi adjustments.
21946 (record_phi_increments): New.
21947 (record_increments): Handle phi adjustments.
21948 (phi_incr_cost): New.
21949 (lowest_cost_path): Handle phis.
21950 (total_savings): Likewise.
21951 (analyze_increments): Likewise.
21952 (ncd_with_phi): New.
21953 (ncd_of_cand_and_phis): New.
21954 (nearest_common_dominator_for_cands): Handle phi increments.
21955 (all_phi_incrs_profitable): New.
21956 (replace_profitable_candidates): Handle phi-dependent candidates.
21957 (analyze_candidates_and_replace): Likewise.
21959 2013-05-03 Teresa Johnson <tejohnson@google.com>
21962 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21963 do not exceed REG_BR_PROB_BASE.
21965 2013-05-03 Jeff Law <law@redhat.com>
21967 PR tree-optimization/57144
21968 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21969 operand of the condition will bit into the new type when eliminating
21970 a cast feeding a condition.
21972 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21974 PR rtl-optimization/57130
21975 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21976 of COMPARE as in_code to the recursive call if needed.
21978 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21980 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21981 (enabled): Handle new members.
21982 * config/i386/sse.md (*vec_concatv2si): Merge from
21983 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21984 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21986 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21988 PR tree-optimization/57027
21989 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21990 for fnms opportunity, check we got the prerequisite kind
21991 of tree / gimple before using accessor functions.
21993 2013-05-03 Richard Biener <rguenther@suse.de>
21995 * double-int.h (lshift): New overload without precision
21996 and arith argument.
21997 (operator *=, operator +=, operator -=): Move ...
21998 * double-int.c (operator *=, operator +=, operator -=): ... here
21999 and implement more efficiently.
22000 (mul_double_with_sign): Remove.
22001 (lshift_double): Adjust to take unsinged shift argument, push
22002 dispatching code to callers.
22003 (mul_double_wide_with_sign): Add early out for callers that
22004 are not interested in high parts or overflow.
22005 (lshift): New function.
22006 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
22007 dispatch code here.
22008 (lrotate, rrotate): Use logical shifts.
22009 * expr.c (get_inner_reference): Use lshift.
22010 * fixed-value.c (do_fixed_divide): Likewise.
22011 * tree-dfa.c (get_ref_base_and_extent): Likewise.
22012 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
22013 (indirect_refs_may_alias_p): Likewise.
22014 (stmt_kills_ref_p_1): Likewise.
22016 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
22018 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
22020 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
22022 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
22023 scalar form of FABD instruction.
22025 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
22027 * lra-constraints.c (process_alt_operands): Add checking alt
22028 number to choose the best alternative.
22030 2013-05-02 Richard Biener <rguenther@suse.de>
22032 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
22033 bitmap and its handling.
22034 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
22036 2013-05-02 Richard Biener <rguenther@suse.de>
22038 PR middle-end/57140
22039 * tree-inline.c (copy_loops): Properly handle removed loops.
22040 (copy_cfg_body): Mark destination loops for fixup if source
22041 loops needed fixup.
22043 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
22046 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
22047 generating simple_return for naked functions.
22049 2013-05-02 Martin Jambor <mjambor@suse.cz>
22051 PR middle-end/56988
22052 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
22053 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
22055 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
22056 ipa_agg_replacement_value structures.
22057 (known_aggs_to_agg_replacement_list): Likewise.
22058 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
22059 (read_agg_replacement_chain): Likewise.
22060 (ipcp_transform_function): Also check that by_ref flags match.
22062 2013-05-02 Richard Biener <rguenther@suse.de>
22064 * graphds.h (struct graph): Add obstack member.
22065 * graphds.c (new_graph): Initialize obstack and allocate
22067 (add_edge): Allocate edge from the obstack.
22068 (free_graph): Free the obstack instead of all edges and vertices.
22070 2013-05-02 Teresa Johnson <tejohnson@google.com>
22072 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
22074 * cfg.c (update_bb_profile_for_threading): Ditto.
22075 * tree-inline.c (copy_bb): Ditto.
22076 (copy_edges_for_bb): Ditto.
22077 (initialize_cfun): Ditto.
22078 (copy_cfg_body): Ditto.
22079 (expand_call_inline): Ditto.
22080 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
22081 (estimate_node_size_and_time): Ditto.
22082 (inline_merge_summary): Ditto.
22083 * cgraphclones.c (cgraph_clone_edge): Ditto.
22084 (cgraph_clone_node): Ditto.
22085 * sched-rgn.c (compute_dom_prob_ps): Ditto.
22086 (compute_trg_info): Ditto.
22088 2013-05-02 Ian Bolton <ian.bolton@arm.com>
22090 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
22091 S reg when fp attribute set.
22092 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
22094 2013-05-02 Ian Bolton <ian.bolton@arm.com>
22096 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
22098 (*and_one_cmplsi3_compare0_uxtw): Likewise.
22099 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
22100 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
22102 2013-05-02 Richard Biener <rguenther@suse.de>
22104 * tree-scalar-evolution.c (scev_info_hasher): Remove.
22105 (struct instantiate_cache_entry): New type.
22106 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
22107 (struct instantiate_cache_type): New type.
22108 (set_instantiated_value, get_instantiated_value): Remove.
22109 (get_instantiated_value_entry): New function.
22110 (instantiate_scev_name): Use the new cache and adjust.
22111 (instantiate_scev_poly): Adjust.
22112 (instantiate_scev_binary): Likewise.
22113 (instantiate_array_ref): Likewise.
22114 (instantiate_scev_convert): Likewise.
22115 (instantiate_scev_not): Likewise.
22116 (instantiate_scev_3): Likewise.
22117 (instantiate_scev_2): Likewise.
22118 (instantiate_scev_r): Likewise.
22119 (instantiate_scev): Likewise.
22120 (resolve_mixers): Likewise.
22122 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
22125 * lra-constraints.c (best_small_class_operands_num): Remove.
22126 (process_alt_operands): Remove small_class_operands_num. Take
22127 small classes operands into losers and only if the operand is not
22128 matched. Modify debugging output.
22129 (curr_insn_transform): Remove best_small_class_operands_num.
22132 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22134 * config/aarch64/aarch64-builtins.c
22135 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
22136 * config/aarch64/aarch64-simd-builtins.def
22137 (reduc_splus_): Add new modes.
22138 (reduc_uplus_): New.
22139 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
22140 (reduc_uplus_v4sf): Likewise.
22141 (reduc_splus_v4sf): Likewise.
22142 (aarch64_addv<mode>): Likewise.
22143 (reduc_uplus_<mode>): Likewise.
22144 (reduc_splus_<mode>): Likewise.
22145 (aarch64_addvv2di): Likewise.
22146 (reduc_uplus_v2di): Likewise.
22147 (reduc_splus_v2di): Likewise.
22148 (aarch64_addvv2si): Likewise.
22149 (reduc_uplus_v2si): Likewise.
22150 (reduc_splus_v2si): Likewise.
22151 (reduc_<sur>plus_<mode>): New.
22152 (reduc_<sur>plus_v2di): Likewise.
22153 (reduc_<sur>plus_v2si): Likewise.
22154 (reduc_<sur>plus_v4sf): Likewise.
22155 (aarch64_addpv4sf): Likewise.
22156 * config/aarch64/arm_neon.h
22157 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
22158 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
22159 add UNSPEC_SADDV, UNSPEC_UADDV.
22161 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
22163 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22165 * config/aarch64/arm_neon.h
22166 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
22168 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22170 * config/aarch64/aarch64-builtins
22171 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
22173 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22175 * config/aarch64/aarch64-simd-builtins.def
22176 (reduc_smax_): New.
22177 (reduc_smin_): Likewise.
22178 (reduc_umax_): Likewise.
22179 (reduc_umin_): Likewise.
22180 (reduc_smax_nan_): Likewise.
22181 (reduc_smin_nan_): Likewise.
22184 (smax): Update for V2SF, V4SF and V2DF modes.
22187 (smin_nan): Likewise.
22188 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
22189 (<su><maxmin><mode>3): ...This, refactor.
22190 (s<maxmin><mode>3): New.
22191 (<maxmin_uns><mode>3): Likewise.
22192 (reduc_<maxmin_uns>_<mode>): Refactor.
22193 (reduc_<maxmin_uns>_v4sf): Likewise.
22194 (reduc_<maxmin_uns>_v2si): Likewise.
22195 (aarch64_<fmaxmin><mode>: Remove.
22196 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
22198 (vmin<q>_f<32,64>): Likewise.
22199 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
22201 (su): Add mappings for smax, smin, umax, umin.
22203 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
22204 (FMAXMIN): Rename as...
22205 (FMAXMIN_UNS): ...This.
22207 (fmaxminv): Likewise.
22208 (fmaxmin): Likewise.
22210 (maxmin_uns_op): Likewise.
22212 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22214 * config/aarch64/arm_neon.h
22215 (vac<ge, gt><sd>_f<32, 64>): Rename to...
22216 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
22217 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
22219 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22221 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
22222 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
22224 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22226 * config/aarch64/aarch64-simd.md
22227 (vcond<mode>_internal): Handle special cases for constant masks.
22228 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
22229 (vcondu<mode><mode>): Likewise.
22230 (vcond<v_cmp_result><mode>): New.
22232 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22234 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
22235 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
22236 * config/aarch64/aarch64-simd-builtins.def
22237 (cmeq): Update to BUILTIN_VALLDI.
22242 * config/aarch64/arm_neon.h
22243 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
22244 to builtins or C as appropriate.
22246 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22248 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
22250 (cmhi): Rename to...
22252 * config/aarch64/aarch64-simd.md
22253 (simd_mode): Add SF.
22254 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
22255 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
22256 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
22257 (cstore<mode>_neg): ...This.
22258 * config/aarch64/iterators.md
22260 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
22261 (COMPARISONS): New.
22262 (UCOMPARISONS): Likewise.
22263 (optab): Add missing comparisons.
22269 (VCMP_S): Likewise.
22270 (VCMP_U): Likewise.
22271 (V_cmp_result): Add DF, SF modes.
22272 (v_cmp_result): Likewise.
22274 (vmtype): Likewise.
22275 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
22277 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
22279 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
22280 define_insn to define_insn_and_split.
22281 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
22282 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
22283 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
22284 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
22285 (thumb2_negscc): Likewise.
22287 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22289 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
22291 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22293 * config/arm/thumb2.md: Remove trailing whitespaces.
22295 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
22297 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
22298 Use gen_int_mode rather than GEN_INT.
22300 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
22302 * value-prof.c (stream_in_histogram_value): Remove the strayed
22305 2013-04-30 Richard Biener <rguenther@suse.de>
22307 PR middle-end/57122
22308 * cfghooks.c (split_edge): Properly check for the loop latch edge.
22310 2013-04-30 Richard Biener <rguenther@suse.de>
22312 PR middle-end/57107
22313 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
22315 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
22317 PR rtl-optimization/56957
22318 PR rtl-optimization/57105
22319 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
22320 variable. Use just INSN_UID for determining whether an insn
22321 should be only disconnected from the insn stream.
22322 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
22324 2013-04-30 Jakub Jelinek <jakub@redhat.com>
22326 PR tree-optimization/57104
22327 * tsan.c (instrument_expr): Don't instrument accesses to
22328 DECL_HARD_REGISTER VAR_DECLs.
22330 2013-04-30 Richard Biener <rguenther@suse.de>
22332 * function.h (loops_for_fn): New inline function.
22333 (set_loops_for_fn): Likewise.
22334 * cfgloop.h (place_new_loop): Add struct function parameter.
22335 (get_loop): Likewise.
22336 (get_loops): Likewise.
22337 (number_of_loops): Likewise.
22338 (fel_next): Adjust.
22339 (fel_init): Likewise.
22340 * cfg.c (get_loop_copy): Adjust.
22341 * cfgloop.c (flow_loops_dump): Likewise.
22342 (record_loop_exits): Likewise.
22343 (verify_loop_structure): Likewise.
22344 * cfgloopanal.c (mark_irreducible_loops): Likewise.
22345 (estimate_reg_pressure_cost): Likewise.
22346 (mark_loop_exit_edges): Likewise.
22347 * cfgloopmanip.c (place_new_loop): Likewise.
22348 (add_loop): Likewise.
22349 (duplicate_loop): Likewise.
22350 * graph.c (draw_cfg_nodes): Likewise.
22351 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
22352 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
22353 (extract_affine_chrec): Likewise.
22354 (build_scop_iteration_domain): Likewise.
22355 * graphite.c (graphite_initialize): Likewise.
22356 * ira-build.c (create_loop_tree_nodes): Likewise.
22357 (more_one_region_p): Likewise.
22358 (rebuild_regno_allocno_maps): Likewise.
22359 (mark_loops_for_removal): Likewise.
22360 (mark_all_loops_for_removal): Likewise.
22361 (remove_unnecessary_regions): Likewise.
22362 (ira_build): Likewise.
22363 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22364 * loop-init.c (fix_loop_structure): Likewise.
22365 (gate_rtl_move_loop_invariants): Likewise.
22366 (gate_rtl_unswitch): Likewise.
22367 (gate_rtl_unroll_and_peel_loops): Likewise.
22368 (rtl_doloop): Likewise.
22369 * lto-streamer-in.c (input_cfg): Likewise.
22370 * lto-streamer-out.c (output_cfg): Likewise.
22371 * modulo-sched.c (sms_schedule): Likewise.
22372 * predict.c (tree_estimate_probability): Likewise.
22373 (tree_estimate_probability_driver): Likewise.
22374 (estimate_loops): Likewise.
22375 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
22376 (move_sese_region_to_fn): Likewise.
22377 (debug_loop_num): Likewise.
22378 * tree-chrec.c (chrec_evaluate): Likewise.
22379 (hide_evolution_in_other_loops_than_loop): Likewise.
22380 (chrec_component_in_loop_num): Likewise.
22381 (reset_evolution_in_loop): Likewise.
22382 (evolution_function_is_invariant_rec_p): Likewise.
22383 * tree-if-conv.c (main_tree_if_conversion): Likewise.
22384 * tree-inline.c (copy_loops): Likewise.
22385 (copy_cfg_body): Likewise.
22386 (tree_function_versioning): Likewise.
22387 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
22388 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
22390 (add_to_evolution_1): Likewise.
22391 (scev_const_prop): Likewise.
22392 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
22393 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
22394 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
22395 (tree_ssa_lim_initialize): Likewise.
22396 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
22397 (verify_loop_closed_ssa): Likewise.
22398 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
22399 (tree_ssa_loop_im): Likewise.
22400 (tree_ssa_loop_unswitch): Likewise.
22401 (tree_vectorize): Likewise.
22402 (check_data_deps): Likewise.
22403 (tree_ssa_loop_ivcanon): Likewise.
22404 (tree_ssa_loop_bounds): Likewise.
22405 (tree_complete_unroll): Likewise.
22406 (tree_complete_unroll_inner): Likewise.
22407 (tree_parallelize_loops): Likewise.
22408 (tree_ssa_loop_prefetch): Likewise.
22409 (tree_ssa_loop_ivopts): Likewise.
22410 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
22411 * tree-vectorizer.c (vectorize_loops): Likewise.
22413 2013-04-29 Mike Frysinger <vapier@gentoo.org>
22415 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
22416 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
22417 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
22418 with EABI_LINK_SPEC.
22420 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22423 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
22426 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
22429 * lra-constraints.c (process_alt_operands): Discourage a bit more
22430 using memory for pseudos. Print cost dump for alternatives.
22431 Modify cost values for conflicts with early clobbers.
22432 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
22434 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22437 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
22439 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22441 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
22442 from/to S register.
22443 (movdi_aarch64): Support LDR/STR from/to D register.
22445 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22447 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
22448 or higher by default.
22450 2013-04-29 Richard Biener <rguenther@suse.de>
22452 PR middle-end/57075
22453 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
22454 even if not adding abnormal edges for calls that can make
22457 2013-04-29 Richard Biener <rguenther@suse.de>
22459 PR middle-end/57103
22460 * tree-cfg.c (move_stmt_op): Fix condition under which to update
22462 (move_stmt_r): Remove redundant checking.
22464 2013-04-29 Teresa Johnson <tejohnson@google.com>
22467 * basic-block.h (apply_scale): New function.
22468 (apply_probability): Use apply_scale.
22469 * gimple-streamer-in.c (input_bb): Ditto.
22470 * lto-streamer-in.c (input_cfg): Ditto.
22471 * lto-cgraph.c (merge_profile_summaries): Ditto.
22472 * tree-optimize.c (execute_fixup_cfg): Ditto.
22473 * tree-inline.c (copy_bb): Update comment to use apply_scale.
22474 (copy_edges_for_bb): Ditto.
22475 (copy_cfg_body): Ditto.
22477 2013-04-29 Tom de Vries <tom@codesourcery.com>
22479 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
22480 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
22481 (tail_merge_optimize): Handle current_loops == NULL.
22483 2013-04-26 Jeff Law <law@redhat.com>
22485 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
22486 (simplify_cond_using_ranges): Generalize code to simplify
22487 COND_EXPRs where one argument is a constant and the other
22488 is an SSA_NAME created by an integral type conversion.
22490 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22492 * config/arm/arm.md (store_minmaxsi): Use only when
22493 optimize_insn_for_size_p.
22495 2013-04-29 Christian Bruel <christian.bruel@st.com>
22498 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
22500 2013-04-29 Richard Biener <rguenther@suse.de>
22502 PR middle-end/57089
22503 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
22504 loop tree make sure to schedule a fixup for the child as well.
22505 (expand_omp_for_generic): Properly add loops.
22506 (expand_omp_for_static_nochunk): Likewise.
22507 (expand_omp_for_static_chunk): Likewise.
22508 (expand_omp_for): For the degenerate case fixup loops.
22509 (expand_omp_sections): Fix default bb placement in loops.
22510 (expand_omp_atomic_pipeline): Properly add loops.
22512 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22514 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
22516 2013-04-29 Tom de Vries <tom@codesourcery.com>
22518 * tree-ssa-tail-merge.c: Update header comment.
22520 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22522 * config/aarch64/arm_neon.h
22523 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
22524 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
22525 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
22526 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
22527 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
22528 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
22529 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
22530 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
22532 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22534 * config/aarch64/aarch64-simd.md
22535 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
22536 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
22537 fix_trunc, fixuns_trunc.
22538 (ftrunc<VDQF:mode>2): New.
22539 * config/aarch64/iterators.md (optab): Add fix, fixuns.
22540 (fix_trunc_optab): New.
22542 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22544 * config/aarch64/aarch64-builtins.c
22545 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
22546 iceilf, lround, iroundf.
22548 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22551 * config/i386/i386.h (enum ix86_tune_indices)
22552 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
22553 New, split from X86_TUNE_INTER_UNIT_MOVES.
22554 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
22555 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
22556 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
22557 (TARGET_INTER_UNIT_MOVES): Remove.
22558 * config/i386/i386.c (initial_ix86_tune_features): Update.
22559 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
22560 (ix86_expand_convert_uns_didf_sse): Use
22561 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22562 (ix86_expand_vector_init_one_nonzero): Ditto.
22563 (ix86_expand_vector_init_interleave): Ditto.
22564 (inline_secondary_memory_needed): Return true for moves from SSE class
22565 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
22566 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
22567 * config/i386/constraints.md (Yi, Ym): Depend on
22568 TARGET_INTER_UNIT_MOVES_TO_VEC.
22569 (Yj, Yn): New constraints.
22570 * config/i386/i386.md (*movdi_internal): Change constraints of
22571 operand 1 from Yi to Yj and from Ym to Yn.
22572 (*movsi_internal): Ditto.
22573 (*movdf_internal): Ditto.
22574 (*movsf_internal): Ditto.
22575 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
22576 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22577 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
22578 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
22579 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
22580 * config/i386/sse.md (movdi_to_sse): Ditto.
22581 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
22582 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
22583 TARGET_INTER_UNIT_MOVES.
22584 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
22585 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
22586 instead of TARGET_INTER_UNIT_MOVES.
22587 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
22588 operand 1 from Yi to Yj and from Ym to Yn.
22590 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22592 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
22593 (float_truncate_hi_): Likewise.
22594 (float_extend_lo_): Likewise.
22595 (float_truncate_lo_): Likewise.
22596 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
22597 (aarch64_float_extend_lo_v2df): Likewise.
22598 (vec_unpacks_hi_v4sf): Likewise.
22599 (aarch64_float_truncate_lo_v2sf): Likewise.
22600 (aarch64_float_truncate_hi_v4sf): Likewise.
22601 (vec_pack_trunc_v2df): Likewise.
22602 (vec_pack_trunc_df): Likewise.
22604 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22606 * config/aarch64/aarch64-builtins.c
22607 (aarch64_fold_builtin): Fold float conversions.
22608 * config/aarch64/aarch64-simd-builtins.def
22609 (floatv2si, floatv4si, floatv2di): New.
22610 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
22611 * config/aarch64/aarch64-simd.md
22612 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
22613 * config/aarch64/iterators.md (FLOATUORS): New.
22614 (optab): Add float, floatuns.
22615 (su_optab): Likewise.
22617 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22619 * config/aarch64/aarch64-builtins.c
22620 (aarch64_builtin_vectorized_function): Use new names for
22622 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
22623 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
22624 (fcvtzu): Split as...
22625 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
22626 (fcvtas): Split as...
22627 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
22628 (fcvtau): Split as...
22629 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
22630 (fcvtps): Split as...
22631 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
22632 (fcvtpu): Split as...
22633 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
22634 (fcvtms): Split as...
22635 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
22636 (fcvtmu): Split as...
22637 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
22638 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
22639 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
22640 (lfrintnusf, lfrintnudf): Likewise.
22641 * config/aarch64/aarch64-simd.md
22642 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
22644 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
22645 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
22646 (fcvt_pattern): Likewise.
22648 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22650 * config/aarch64/aarch64-simd.md
22651 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
22652 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
22654 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22656 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
22657 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
22658 (vrnd<a,m,n,p>_f32): Implement using builtins.
22659 (vrnd<i,x><q>_f<32, 64>): New.
22661 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22663 * config/aarch64/aarch64-builtins.c
22664 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
22665 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
22666 (frintz): Rename to...
22668 (frintp): Rename to...
22670 (frintm): Rename to...
22672 (frinti): Rename to...
22673 (nearbyint): ...this.
22674 (frintx): Rename to...
22676 (frinta): Rename to...
22678 * config/aarch64/aarch64-simd.md
22679 (aarch64_frint<frint_suffix><mode>): Delete.
22680 (<frint_pattern><mode>2): Convert to insn.
22681 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
22682 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
22683 (frint_pattern): Likewise.
22684 (frint_suffix): Likewise.
22686 2013-04-29 Richard Biener <rguenther@suse.de>
22688 PR tree-optimization/57081
22689 * loop-init.c: Include tree-flow.h.
22690 (loop_optimizer_finalize): Free number of iteration estimates.
22691 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
22693 2013-04-29 Jakub Jelinek <jakub@redhat.com>
22695 PR tree-optimization/57083
22696 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
22697 non-singleton shift count range, zero extend low_bound for uns case.
22699 * config/i386/predicates.md (general_vector_operand): New predicate.
22700 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
22701 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
22702 if they aren't nonimmediate operands. If their original values
22703 satisfy const_vector_equal_evenodd_p, don't shift them.
22704 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
22705 predicates. For the SSE4.1 case force operands[{1,2}] into registers
22706 if not nonimmediate_operand.
22707 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
22708 instead of register_operand.
22709 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
22711 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22713 * stor-layout.c (finalize_size_functions): Allocate a structure and
22714 reset cfun before dumping the functions.
22716 2013-04-27 Jakub Jelinek <jakub@redhat.com>
22718 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
22721 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
22722 use xop_pmacsdqh if uns_p.
22723 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
22724 the immediate rotate count.
22726 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
22728 * rtl.h (struct rtx_def): Add comment for field jump.
22729 (LRA_SUBREG_P): New macro.
22730 * recog.c (register_operand): Check LRA_SUBREG_P.
22731 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22733 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22734 (lra_final_code_change): Skip subreg change for operators.
22735 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22736 if there are no operand changes.
22737 * lra-constraints.c (curr_insn_set): New.
22738 (match_reload): Set LRA_SUBREG_P.
22739 (emit_spill_move): Ditto.
22740 (check_and_process_move): Use curr_insn_set. Process only single
22741 set insns. Don't initialize sec_mem_p and change_p.
22742 (simplify_operand_subreg): Use LRA_SUBREG_P.
22743 (reg_in_class_p): New function.
22744 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22745 of #ifdef. Add code to remove cycling.
22746 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22747 non-null disp. Reload inner instead of disp when base and index
22748 are null. Try to put lo_sum into register.
22749 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22750 (check_and_process_move): Move code for move cost check to
22751 simple_move_p. Remove equiv_substitution.
22752 (simple_move_p): New function.
22753 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22754 curr_insn_set. Call check_and_process_move only for single set
22755 insns. Use the new function. Move call of check_and_process_move
22756 after operand equiv substitution and address process.
22758 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22761 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22762 with nonlocal goto receivers or returns twice calls, ignore
22763 unininitialized values from abnormal edges to nl goto receiver
22764 or returns twice call.
22766 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22768 PR tree-optimization/57051
22769 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22770 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22773 2013-04-26 Richard Biener <rguenther@suse.de>
22775 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22776 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22777 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22778 (fixup_loop_arrays_after_move): New function.
22779 (move_sese_region_to_fn): Properly outline the loop tree parts
22780 of the SESE region.
22782 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22784 * config/i386/i386.md (type, unit): Fix long lines.
22786 2013-04-26 Richard Biener <rguenther@suse.de>
22788 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22789 (lto-streamer-out.o): Likewise.
22790 * cfgloop.c (init_loops_structure): Export, add struct function
22791 argument and adjust.
22792 (flow_loops_find): Adjust.
22793 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22794 (init_loops_structure): Declare.
22795 * lto-streamer-in.c: Include cfgloop.h.
22796 (input_cfg): Input the loop tree.
22797 * lto-streamer-out.c: Include cfgloop.h.
22798 (output_cfg): Output the loop tree.
22799 (output_struct_function_base): Do not drop PROP_loops.
22801 2013-03-26 Richard Biener <rguenther@suse.de>
22803 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22804 (pass_build_cfg): Provide PROP_loops.
22805 (move_sese_region_to_fn): Remove loops that are outlined into fn
22807 * tree-inline.c: Include cfgloop.h.
22808 (initialize_cfun): Do not drop PROP_loops.
22809 (copy_loops): New function.
22810 (copy_cfg_body): Copy loop structure.
22811 (tree_function_versioning): Initialize destination loop tree.
22812 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22813 (pass_parallelize_loops): Do IL verification.
22814 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22815 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22816 the CFG make sure we fixup loops as well.
22817 * tree-ssa-tail-merge.c: Include cfgloop.h.
22818 (replace_block_by): When merging loop latches mark loops for fixup.
22819 * lto-streamer-out.c (output_struct_function_base): Drop
22820 PROP_loops for now.
22821 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22822 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22823 * ipa-split.c: Include cfgloop.h.
22824 (split_function): Add the new return block to the loop tree root.
22825 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22826 whether we have removed the forwarder block.
22827 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22828 * cfgloop.h (place_new_loop): Declare.
22829 * cfgloopmanip.c (place_new_loop): Export.
22830 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22831 (tree-switch-conversion.o): Likewise.
22832 (tree-complex.o): Likewise.
22833 (tree-inline.o): Likewise.
22834 (tree-ssa-tailmerge.o): Likewise.
22835 (ipa-split.o): Likewise.
22836 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22837 (tree-ssa-copy.o): Likewise.
22838 * tree-switch-conversion.c: Include cfgloop.h
22839 (process_switch): If we emit a bit-test cascade, schedule loops
22841 * tree-complex.c: Include cfgloop.h.
22842 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22843 * asan.c: Include cfgloop.h.
22844 (create_cond_insert_point): Properly add new basic-blocks to
22845 loops, schedule loop fixup.
22846 * cfgloop.c (verify_loop_structure): Check that looks are not
22848 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22850 (expand_omp_for_generic): Likewise.
22851 (expand_omp_sections): Likewise.
22852 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22853 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22854 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22855 is initialized, not when loops are present.
22856 * tree-parloops.c (parallelize_loops): Remove checking here.
22857 * passes.c (init_optimization_passes): Schedule a copy-propagation
22858 pass before complete unrolling of inner loops.
22860 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22862 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22863 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22864 (colorize_init): Add argument to _WIN32 version.
22865 * toplev.c: Include diagnostic-color.h.
22866 (process_options): Default to -fdiagnostics-color=auto if
22867 GCC_COLORS env var is in the environment.
22868 * common.opt (fdiagnostics-color=): Add Var and Init.
22869 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22870 env var is in the environment, the default is auto rather than never.
22872 * diagnostic.h (file_name_as_prefix): Add context argument.
22873 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22874 the string as locus.
22875 * langhooks.c (lhd_print_error_function): Adjust caller.
22877 2013-04-25 Lawrence Crowl <crowl@google.com>
22879 * var-tracking.c (shared_hash_def::htab):
22880 Change type to hash_table. Update dependent calls and types.
22882 2013-04-25 Lawrence Crowl <crowl@google.com>
22884 * Makefile.in: Update as needed below.
22886 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22887 Move declaration to after the type's method definitons.
22889 * attribs.c (htab_t scoped_attributes::attribute_hash):
22890 Change type to hash_table. Update dependent calls and types.
22892 * bitmap.c (htab_t bitmap_desc_hash):
22893 Change type to hash_table. Update dependent calls and types.
22895 * cselib.c (htab_t cselib_hash_table):
22896 Change type to hash_table. Update dependent calls and types.
22898 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22899 (hash_string_slot_node): Move implementation into lto-streamer.h
22900 struct string_slot_hasher.
22901 (eq_string_slot_node): Likewise.
22903 * data-streamer-out.c: Update output_block::string_hash_table
22904 dependent calls and types.
22906 * dwarf2cfi.c (htab_t trace_index):
22907 Change type to hash_table. Update dependent calls and types.
22909 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22910 Change type to hash_table. Update dependent calls and types.
22911 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22912 (htab_t optimize_external_refs::map): Likewise.
22913 (htab_t output_comp_unit::extern_map): Likewise.
22914 (htab_t output_comdat_type_unit::extern_map): Likewise.
22915 (htab_t output_macinfo::macinfo_htab): Likewise.
22916 (htab_t optimize_location_lists::htab): Likewise.
22917 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22919 * except.c (htab_t ehspec_hash_type):
22920 Change type to hash_table. Update dependent calls and types.
22921 (assign_filter_values::ttypes): Likewise.
22922 (assign_filter_values::ehspec): Likewise.
22923 (sjlj_assign_call_site_values::ar_hash): Likewise.
22924 (convert_to_eh_region_ranges::ar_hash): Likewise.
22926 * gcse.c (htab_t pre_ldst_table):
22927 Change type to hash_table. Update dependent calls and types.
22929 * ggc-common.c (htab_t saving_htab):
22930 Change type to hash_table. Update dependent calls and types.
22931 (htab_t loc_hash): Likewise.
22932 (htab_t ptr_hash): Likewise.
22933 (call_count): Rename ggc_call_count.
22934 (call_alloc): Rename ggc_call_alloc.
22935 (loc_descriptor): Rename make_loc_descriptor.
22936 (add_statistics): Rename ggc_add_statistics.
22938 * ggc-common.c (saving_htab):
22939 Change type to hash_table. Update dependent calls and types.
22941 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22942 (push_gimplify_context): Likewise.
22943 (pop_gimplify_context): Likewise.
22944 (struct gimple_temp_hash_elt): Added.
22945 (struct gimplify_hasher): Likewise.
22946 (struct gimplify_ctx.temp_htab):
22947 Change type to hash_table. Update dependent calls and types.
22949 * gimple-fold.c: Include gimplify-ctx.h.
22951 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22952 Change type to hash_table. Update dependent calls and types.
22953 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22954 avoid potential global name collision.
22956 * gimplify.c: Include gimplify-ctx.h.
22957 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22958 (htab_t gimplify_ctx::temp_htab):
22959 Update dependent calls and types for new type hash_table.
22960 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22961 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22963 * gimplify-ctx.h: New.
22964 (struct gimple_temp_hash_elt): Move from gimplify.c.
22965 (class gimplify_hasher): New.
22966 (struct gimplify_ctx): Move from gimple.h.
22967 (htab_t gimplify_ctx::temp_htab):
22968 Change type to hash_table. Update dependent calls and types.
22970 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22971 (htab_t ivs_params::newivs_index):
22972 Change type to hash_table. Update dependent calls and types.
22973 (htab_t ivs_params::params_index): Likewise.
22974 (htab_t print_generated_program::params_index): Likewise.
22975 (htab_t gloog::newivs_index): Likewise.
22976 (htab_t gloog::params_index): Likewise.
22978 * graphite.c: Include graphite-htab.h.
22979 4htab_t graphite_transform_loops::bb_pbb_mapping):
22980 Change type to hash_table. Update dependent calls and types.
22982 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22983 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22984 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22986 * graphite-dependences.c: Include graphite-htab.h.
22987 (loop_is_parallel_p): Change hash table type of parameter.
22989 * graphite-htab.h: New.
22990 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22991 (extern find_pbb_via_hash): Move from graphite-poly.h.
22992 (extern loop_is_parallel_p): Move from graphite-poly.h.
22993 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22995 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22996 (extern loop_is_parallel_p): Move to graphite-htab.h.
22997 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22999 * haifa-sched.c (htab_t delay_htab):
23000 Change type to hash_table. Update dependent calls and types.
23001 (htab_t delay_htab_i2): Likewise.
23003 * ira-color.c (htab_t allocno_hard_regs_htab):
23004 Change type to hash_table. Update dependent calls and types.
23006 * ira-costs.c (htab_t cost_classes_htab):
23007 Change type to hash_table. Update dependent calls and types.
23009 * loop-invariant.c (htab_t merge_identical_invariants::eq):
23010 Change type to hash_table. Update dependent calls and types.
23012 * loop-iv.c (htab_t bivs):
23013 Change type to hash_table. Update dependent calls and types.
23015 * loop-unroll.c (htab_t opt_info::insns_to_split):
23016 Change type to hash_table. Update dependent calls and types.
23017 (htab_t opt_info::insns_with_var_to_expand): Likewise.
23019 * lto-streamer.h (struct string_slot): Move from data-streamer.h
23020 (struct string_slot_hasher): New.
23021 (htab_t output_block::string_hash_table):
23022 Change type to hash_table. Update dependent calls and types.
23024 * lto-streamer-in.c (freeing_string_slot_hasher): New.
23025 (htab_t file_name_hash_table):
23026 Change type to hash_table. Update dependent calls and types.
23028 * lto-streamer-out.c: Update output_block::string_hash_table dependent
23031 * lto-streamer.c (htab_t tree_htab):
23032 Change type to hash_table. Update dependent calls and types.
23034 * omp-low.c: Include gimplify-ctx.h.
23036 * passes.c (htab_t name_to_pass_map):
23037 Change type to hash_table. Update dependent calls and types.
23038 (pass_traverse): Rename to passes_pass_traverse.
23040 * plugin.c (htab_t event_tab):
23041 Change type to hash_table. Update dependent calls and types.
23043 * postreload-gcse.c (htab_t expr_table):
23044 Change type to hash_table. Update dependent calls and types.
23045 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
23047 * sese.c (debug_rename_map_1): Make extern.
23048 (htab_t copy_bb_and_scalar_dependences::rename_map):
23049 Change type to hash_table. Update dependent calls and types.
23051 * sese.h (extern debug_rename_map): Move to .c file.
23053 * store-motion.c (htab_t store_motion_mems_table):
23054 Change type to hash_table. Update dependent calls and types.
23056 * trans-mem.c (htab_t tm_new_mem_hash):
23057 Change type to hash_table. Update dependent calls and types.
23059 * tree-browser.c (htab_t TB_up_ht):
23060 Change type to hash_table. Update dependent calls and types.
23062 * tree-cfg.c (htab_t discriminator_per_locus):
23063 Change type to hash_table. Update dependent calls and types.
23065 * tree-complex.c: Include tree-hasher.h
23066 (htab_t complex_variable_components):
23067 Change type to hash_table. Update dependent calls and types.
23069 * tree-eh.c (htab_t finally_tree):
23070 Change type to hash_table. Update dependent calls and types.
23072 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
23073 struct int_tree_hasher.
23074 (extern int_tree_map_eq): Likewise.
23075 (uid_decl_map_hash): Removed.
23076 (extern decl_tree_map_eq): Likewise.
23078 * tree-hasher.h: New.
23079 (struct int_tree_hasher): New.
23080 (typedef int_tree_htab_type): New.
23082 * tree-inline.c: Include gimplify-ctx.h.
23084 * tree-mudflap.c: Include gimplify-ctx.h.
23086 * tree-parloops.c: Include tree-hasher.h.
23087 (htab_t eliminate_local_variables_stmt::decl_address):
23088 Change type to hash_table. Update dependent calls and types.
23089 (htab_t separate_decls_in_region::decl_copies): Likewise.
23091 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
23092 Change type to hash_table. Update dependent calls and types.
23094 * tree-sra.c (candidates):
23095 Change type to hash_table. Update dependent calls and types.
23097 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
23099 (int_tree_map_hash): Likewise.
23101 * tree-ssa-dom.c (htab_t avail_exprs):
23102 Change type to hash_table. Update dependent calls and types.
23104 * tree-ssa-live.c (var_map_base_init::tree_to_index):
23105 Change type to hash_table. Update dependent calls and types.
23107 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
23108 Change type to hash_table. Update dependent calls and types.
23110 * tree-ssa-phiopt.c (seen_ssa_names):
23111 Change type to hash_table. Update dependent calls and types.
23113 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
23114 Change type to hash_table. Update dependent calls and types.
23116 * tree-ssa-uncprop.c (equiv):
23117 Change type to hash_table. Update dependent calls and types.
23119 2013-04-25 Jakub Jelinek <jakub@redhat.com>
23121 PR rtl-optimization/57003
23122 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
23123 call note_stores with kill_clobbered_value callback again after
23124 killing regs_invalidated_by_call.
23126 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23128 * config/aarch64/aarch64-simd.md
23129 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
23130 (aarch64_simd_bsl<mode>): Likewise.
23131 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
23133 2013-04-25 Marek Polacek <polacek@redhat.com>
23135 PR tree-optimization/57066
23136 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
23138 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23140 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
23142 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23144 * config/aarch64/aarch64-builtins.c
23145 (aarch64_fold_builtin): New.
23146 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
23147 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
23148 * config/aarch64/aarch64-simd-builtins.def (abs): New.
23149 * config/aarch64/arm_neon.h
23150 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
23152 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23153 Tejas Belagod <tejas.belagod@arm.com>
23155 * config/aarch64/aarch64-builtins.c
23156 (aarch64_gimple_fold_builtin): New.
23157 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
23158 * config/aarch64/aarch64-simd-builtins.def (addv): New.
23159 * config/aarch64/aarch64-simd.md (addpv4sf): New.
23160 (addvv4sf): Update.
23161 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
23163 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23165 * config/aarch64/aarch64.md
23166 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
23168 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23170 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
23171 (*ngcsi_uxtw): New pattern.
23173 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23174 Julian Brown <julian@codesourcery.com>
23176 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
23177 (TB_DREG): Add T_V4HF.
23178 (v4hf_UP): New macro.
23179 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23180 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23181 Handle initialisation of V4HF. Adjust initialisation of reinterpret
23183 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23184 (arm_vector_mode_supported_p): Handle V4HF.
23185 (arm_mangle_map): Handle V4HFmode.
23186 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
23187 * config/arm/arm_neon_builtins.def: Add entries for
23188 vcvtv4hfv4sf, vcvtv4sfv4hf.
23189 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
23190 (neon_vcvtv4hfv4sf): Likewise.
23191 * config/arm/neon-gen.ml: Handle half-precision floating point
23193 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
23194 * config/arm/arm_neon.h: Regenerate.
23195 * config/arm/neon.ml (type elts): Add F16.
23196 (type vectype): Add T_float16x4, T_floatHF.
23197 (type vecmode): Add V4HF.
23198 (type features): Add Requires_FP_bit feature.
23199 (elt_width): Handle F16.
23200 (elt_class): Likewise.
23201 (elt_of_class_width): Likewise.
23202 (mode_of_elt): Refactor.
23203 (type_for_elt): Handle F16, fix error messages.
23204 (vectype_size): Handle T_float16x4.
23205 (vcvt_sh): New function.
23206 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
23207 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
23208 (string_of_mode): Handle V4HF.
23209 * doc/arm-neon-intrinsics.texi: Regenerate.
23211 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23213 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
23214 format specifier in 'X' case.
23216 2013-04-25 Alan Modra <amodra@gmail.com>
23219 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
23220 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
23221 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
23222 Repeat for many other rotate/shift and mask patterns using subregs.
23223 Name lshiftrt insns.
23224 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
23225 on WORDS_BIG_ENDIAN.
23227 2013-04-25 Alan Modra <amodra@gmail.com>
23229 * config.gcc: Support little-endian powerpc-linux targets.
23230 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
23231 (LINK_OS_LINUX_SPEC): Define.
23232 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
23233 Preserve MASK_LITTLE_ENDIAN.
23234 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
23235 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
23236 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
23237 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
23238 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
23239 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
23240 Correct fp word order for little-endian. Don't shift toc entries
23241 smaller than a word for little-endian.
23242 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
23243 (bswapdi2 splits): Correct low-part subreg for little-endian.
23244 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
23245 low/high where such is correct only for be.
23246 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
23247 little-endian for -mcall-aixdesc.
23249 2013-04-25 Alan Modra <amodra@gmail.com>
23251 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
23252 replace_equiv_address_nv.
23254 2013-04-25 Alan Modra <amodra@gmail.com>
23256 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
23258 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23261 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23262 * rtl.h (struct rtx_def): ...
23264 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23266 PR rtl-optimizations/57046
23267 * lra-constraints (split_reg): Set up lra_risky_transformations_p
23268 for multi-reg splits.
23270 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
23272 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
23274 2013-04-24 Sterling Augustine <saugustine@google.com>
23276 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
23277 (comp_dir_string, debug_str_dwo_section): New.
23278 (DEBUG_STR_DWO_SECTION): Rename to ...
23279 (DEBUG_DWO_STR_SECTION): ... this.
23280 (DEBUG_NORM_STR_SECTION): Delete.
23281 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
23282 (DEBUG_STR_DWO_SECTION_FLAGS): New.
23283 (find_AT_string): Move most logic to ...
23284 (find_AT_string_in_table): ... here. New.
23285 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
23286 add_skeleton_AT_string. Delete logic.
23287 (output_skeleton_debug_sections): Remove call to
23288 add_top_level_skeleton_die_attrs.
23289 (add_comp_dir_attribute): Move logic to comp_dir_string.
23290 (dwarf2out_init): Initialize debug_str_dwo_section.
23291 (output_indirect_string): Call find_string_form.
23292 (output_indirect_strings): Rewrite.
23293 (prune_unused_types): Empty skeleton_debug_str_hash.
23294 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
23295 (dwarf2out_finish): Call output_indirect_strings.
23297 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23299 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
23301 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23303 * rtl.h (struct rtx_def): Add comment for field jump.
23304 (LRA_SUBREG_P): New macro.
23305 * recog.c (register_operand): Check LRA_SUBREG_P.
23306 * lra.c (lra): Add note at the end of RTL code. Align non-empty
23308 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
23309 (lra_final_code_change): Skip subreg change for operators.
23310 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
23311 if there are no operand changes.
23312 * lra-constraints.c (curr_insn_set): New.
23313 (match_reload): Set LRA_SUBREG_P.
23314 (emit_spill_move): Ditto.
23315 (check_and_process_move): Use curr_insn_set. Process only single
23316 set insns. Don't initialize sec_mem_p and change_p.
23317 (simplify_operand_subreg): Use LRA_SUBREG_P.
23318 (reg_in_class_p): New function.
23319 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
23320 of #ifdef. Add code to remove cycling.
23321 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
23322 non-null disp. Reload inner instead of disp when base and index
23323 are null. Try to put lo_sum into register.
23324 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
23325 (check_and_process_move): Move code for move cost check to
23326 simple_move_p. Remove equiv_substitution.
23327 (simple_move_p): New function.
23328 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
23329 curr_insn_set. Call check_and_process_move only for single set
23330 insns. Use the new function. Move call of check_and_process_move
23331 after operand equiv substitution and address process.
23333 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
23335 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
23336 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
23337 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
23339 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23341 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
23343 2013-04-24 Marek Polacek <polacek@redhat.com>
23345 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
23346 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
23347 (select_loops_exit_conditions): Likewise.
23348 (number_of_iterations_for_all_loops): Likewise.
23349 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
23350 (scev_analysis): Likewise.
23352 2013-04-02 Catherine Moore <clm@codesourcery.com>
23353 Chao-ying Fu <fu@mips.com>
23355 * config/mips/micromips.md (jraddiusp): New pattern.
23356 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
23357 instruction if possible.
23359 2013-04-24 Alan Modra <amodra@gmail.com>
23361 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
23363 2013-04-24 Julian Brown <julian@codesourcery.com>
23364 Chung-Lin Tang <cltang@codesourcery.com>
23366 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
23367 dependency behavior in enumeration type DIE generation. Add TODO note
23368 to comments about future DW_FORM_sdata/udata re-work of related code.
23370 2013-04-23 Lawrence Crowl <crowl@google.com>
23372 * Makefile.in: Update as needed below.
23374 * hash-table.h (class hash_table):
23375 Correct many methods with parameter types compare_type to the correct
23376 value_type. (Correct code was unlikely to notice the change.)
23377 (hash_table::elements_with_deleted) New.
23378 (class hashtable::iterator): New.
23379 (hashtable::begin()): New.
23380 (hashtable::end()): New.
23381 (FOR_EACH_HASH_TABLE_ELEMENT): New.
23383 * statistics.c (statistics_hashes):
23384 Change type to hash_table. Update dependent calls and types.
23386 * tree-into-ssa.c (var_infos):
23387 Change type to hash_table. Update dependent calls and types.
23389 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
23390 Change type to hash_table. Update dependent calls and types.
23392 * tree-ssa-loop-im.c (struct mem_ref.refs):
23393 Change type to hash_table. Update dependent calls and types.
23395 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
23396 Change type to hash_table. Update dependent calls and types.
23398 * tree-ssa-sccvn.c (vn_tables_s::nary):
23399 Change type to hash_table. Update dependent calls and types.
23400 (vn_tables_s::phis): Likewise.
23401 (vn_tables_s::references): Likewise.
23403 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
23404 (vn_reference_eq): Update parameter and return types.
23406 * tree-ssa-structalias.c (pointer_equiv_class_table):
23407 Change type to hash_table. Update dependent calls and types.
23408 (location_equiv_class_table): Likewise.
23410 * tree-vect-data-refs.c: Consequential changes for making
23411 peeling a hash_table.
23413 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
23414 (destroy_loop_vec_info): Dependent hash_table update.
23416 * tree-vectorizer.h (peeling_htab):
23417 Change type to hash_table. Update dependent calls and types.
23419 2013-04-23 Shiva Chen <shiva0217@gmail.com>
23421 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
23422 to check the register content is equal or not.
23423 * lra-constraints.c (match_reload): Use lra_assign_reg_val
23424 to assign register content record.
23425 * lra-eliminations.c (update_reg_eliminate): Use
23426 lra_update_reg_val_offset to update register content offset.
23427 * lra-int.h (struct lra_reg): Add offset member.
23428 (lra_reg_val_equal_p): New static inline function.
23429 (lra_update_reg_val_offset): New static inline function.
23430 (lra_assign_reg_val): New static inline function.
23431 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
23432 to assign register content record.
23433 (initialize_lra_reg_info_element): Initial offset to zero.
23435 2013-04-23 Catherine Moore <clm@codesourcery.com>
23437 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
23438 operands. Record compression.
23440 2013-04-23 Xinliang David Li <davidxl@google.com>
23442 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
23444 2013-04-23 Richard Biener <rguenther@suse.de>
23446 PR middle-end/57036
23447 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
23448 parameter, only add abnormal goto edges from the copied body
23449 if the call could perform abnormal gotos.
23450 (copy_cfg_body): Adjust.
23452 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
23454 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
23456 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
23458 * coretypes.h (gimple_stmt_iterator): Add struct to make
23461 2013-04-23 Richard Biener <rguenther@suse.de>
23463 PR tree-optimization/57026
23464 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
23465 from SSA names occuring in abnormal PHI nodes.
23467 2013-04-22 Andi Kleen <ak@linux.intel.com>
23469 * lto/lto.c (print_lto_report_1): Fix LTO report names.
23471 2013-04-22 Andi Kleen <ak@linux.intel.com>
23473 * lto/lto.c (print_lto_report_1): Declare early.
23474 (read_cgraph_and_symbols): Call print_lto_report_1 early.
23476 2013-04-22 Andi Kleen <ak@linux.intel.com>
23478 * common.opt (-flto-report-wpa): Add.
23479 * doc/invoke.texi (-flto-report-wpa): Add.
23480 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
23483 2013-04-22 Xinliang David Li <davidxl@google.com>
23485 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
23486 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
23487 * Makefile.in: New dependency
23489 David Daney <ddaney.cavm@gmail.com>
23491 * configure.ac (gcc_cv_as_micromips_support): Use the
23492 --fatal-warnings option.
23493 * configure: Regenerate.
23495 2013-04-22 Marek Polacek <polacek@redhat.com>
23498 * tsan.c (instrument_expr): Don't instrument expression
23499 in case its size is zero.
23501 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
23505 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23507 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23509 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23511 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
23512 (gimple_stmt_iterator): New typedef.
23513 * gimple.h (gimple_stmt_iterator): Rename to...
23514 (gimple_stmt_iterator_d): ... This.
23515 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
23516 trees be valid for GIMPLE and GENERIC.
23517 (TARGET_GIMPLE_FOLD_BUILTIN): New.
23518 * gimple-fold.c (gimple_fold_call): Call target hook
23519 gimple_fold_builtin.
23520 * hooks.c (hook_bool_gsiptr_false): New.
23521 * hooks.h (hook_bool_gsiptr_false): New.
23522 * target.def (fold_stmt): New.
23523 * doc/tm.texi: Regenerate.
23525 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
23528 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
23529 a set sp if no stack realignment.
23531 2013-04-22 Nick Clifton <nickc@redhat.com>
23533 * config.gcc (tilegx-linux): Extend extra_objs rather than
23535 (tilepro-linux): Likewise.
23537 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23539 * config/aarch64/aarch64-builtins.c
23541 (CF0, CF1, CF2, CF3, CF4, CF10): New.
23542 (VAR<1-12>): Add MAP parameter.
23543 (BUILTIN_*): Likewise.
23544 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
23545 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
23546 (aarch64_ushl_n<mode>): Likewise.
23547 (aarch64_sshr_n<mode>): Likewise.
23548 (aarch64_ushr_n<mode>): Likewise.
23549 (aarch64_<maxmin><mode>): Likewise.
23550 (aarch64_sqrt<mode>): Likewise.
23551 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
23552 (vshr<q>_n_*): Likewise.
23554 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23556 * config/aarch64/aarch64-builtins.c
23557 (aarch64_simd_builtin_type_mode): Handle SF types.
23559 (BUILTIN_GPF): Define.
23560 (aarch64_init_simd_builtins): Handle SF types.
23561 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
23562 (frecps): Likewise.
23563 (frecpx): Likewise.
23564 * config/aarch64/aarch64-simd.md
23565 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
23566 (aarch64_frecpe<mode>): New.
23567 (aarch64_frecps<mode>): Likewise.
23568 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
23569 (v8type): Add frecp<esx>.
23570 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
23571 (aarch64_frecps<mode>): Likewise.
23572 * config/aarch64/iterators.md (FRECP): New.
23573 (frecp_suffix): Likewise.
23574 * config/aarch64/arm_neon.h
23575 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
23577 2013-04-22 Christian Bruel <christian.bruel@st.com>
23580 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
23581 (REG_CLASS_NAMES): Idem.
23582 (REG_CLASS_CONTENTS): Idem.
23583 (REGCLASS_HAS_FP_REG): Idem.
23584 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
23585 (sh_conditional_register_usage): Idem.
23587 2013-04-21 Jeff Law <law@redhat.com>
23589 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
23590 (ssa_forward_propagate_and_combine): Use it.
23592 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23594 * lra.c: Update the flow chart diagram.
23596 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23598 PR rtl-optimization/56847
23599 * lra-constraints.c (process_alt_operands): Discourage alternative
23600 with non-matche doffsettable memory constraint fro memory with
23603 2013-04-19 Richard Biener <rguenther@suse.de>
23605 PR tree-optimization/56982
23606 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
23608 * gimplify.c (gimplify_call_expr): Notice special calls.
23609 (gimplify_modify_expr): Likewise.
23610 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
23611 abnormal control flow receivers.
23612 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
23613 in the same way as cfun->has_nonlocal_labels.
23614 (gimple_purge_dead_abnormal_call_edges): Likewise.
23615 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
23616 receivers start a basic-block.
23618 2013-04-19 Richard Biener <rguenther@suse.de>
23620 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
23622 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
23623 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
23624 (SLP_TREE_LOAD_PERMUTATION): Add.
23625 (vect_transform_slp_perm_load): Adjust prototype.
23626 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
23627 (vect_free_slp_instance): Likewise.
23628 (vect_create_new_slp_node): Likewise.
23629 (vect_supported_slp_permutation_p): Remove.
23630 (vect_slp_rearrange_stmts): Adjust.
23631 (vect_supported_load_permutation_p): Likewise. Inline
23632 vect_supported_slp_permutation_p here.
23633 (vect_analyze_slp_instance): Compute load permutations per
23634 slp node instead of per instance.
23635 (vect_get_slp_defs): Adjust.
23636 (vect_transform_slp_perm_load): Likewise.
23637 (vect_schedule_slp_instance): Remove redundant code.
23638 (vect_schedule_slp): Remove hack for PR56270, add it ...
23639 * tree-vect-stmts.c (vectorizable_load): ... here, do not
23640 CSE loads for SLP. Adjust.
23642 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23644 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
23645 spelling in two comments.
23647 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23650 * config/arm/arm.c (load_multiple_sequence): Require SP
23651 as base register for loads if SP is in the register list.
23653 2013-04-19 Martin Jambor <mjambor@suse.cz>
23655 PR tree-optimization/56718
23656 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
23657 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
23658 and made public. Adjusted all callers.
23659 (ipa_intraprocedural_devirtualization): New function.
23660 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
23661 (ipa_intraprocedural_devirtualization): Likewise.
23662 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
23664 2013-04-19 Richard Biener <rguenther@suse.de>
23666 PR tree-optimization/57000
23667 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
23669 2013-04-19 Terry Guo <terry.guo@arm.com>
23671 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
23673 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
23674 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
23675 (cortex_m4_fmacs): Use new reservations.
23676 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
23678 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
23680 PR rtl-optimization/56999
23681 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
23683 (lra_coalesce): Remove split_origin_bitmap and related code.
23684 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
23685 ranges if necessary.
23687 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
23689 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
23691 (ix86_expand_call): Remove clobbered_registers array and use
23692 x86_64_ms_sysv_extra_clobbered_registers instead.
23693 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
23695 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
23697 * config/i386/i386.md (*call_rex64_ms_sysv): Use
23698 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
23699 (*call_value_rex64_ms_sysv): Ditto.
23701 2013-04-18 Cary Coutant <ccoutant@google.com>
23703 * dwarf2out.c (output_pubnames): Check die_perennial_p of
23704 parent instead of die_mark.
23706 2013-04-18 Diego Novillo <dnovillo@google.com>
23708 * gimple.c (create_gimple_tmp): New.
23709 (get_expr_type): New.
23710 (build_assign): New.
23711 (build_type_cast): New.
23712 * gimple.h (enum ssa_mode): Define.
23713 (gimple_seq_set_location): New.
23714 * asan.c (build_check_stmt): Change some gimple_build_* calls
23715 to use build_assign and build_type_cast.
23717 2013-04-18 Richard Biener <rguenther@suse.de>
23719 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
23720 handle negative step. Remove redundant checks.
23721 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
23722 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
23723 for negative step and grouped loads fail to vectorize.
23725 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23727 * emit-rtl.c (reset_insn_used_flags): New function.
23728 (reset_all_used_flags): Use it.
23729 (verify_insn_sharing): New function.
23730 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23732 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23734 PR tree-optimization/56984
23735 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23736 and (x >> M) >= N don't register any assertion if N << M is the
23739 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23741 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23742 do not delete_insn insns that have not yet been emitted, only
23743 unlink them with remove_insn.
23744 * df-scan.c (df_insn_delete): Revert r197492.
23746 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23748 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23749 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23751 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23753 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23754 define_insn_and_split.
23755 (and_scc,ior_scc,negscc): Likewise.
23756 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23758 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23760 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23761 can benefit from using a sequence of LDRD instructions in epilogue
23762 instead of a single LDM instruction.
23764 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23767 * doc/extend.texi: Fix typo.
23769 2013-04-17 Richard Biener <rguenther@suse.de>
23771 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23772 (vect_build_slp_tree): ... here.
23773 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23774 match. Remove special-casing of mismatched complex loads.
23775 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23776 re-try the match with swapped commutative operands.
23777 (vect_supported_load_permutation_p): Remove special-casing of
23778 mismatched complex loads.
23779 (vect_analyze_slp_instance): Adjust.
23781 2013-04-17 Richard Biener <rguenther@suse.de>
23783 PR rtl-optimization/56921
23784 * cfgloop.h (struct loop): Add simple_loop_desc member.
23785 (struct niter_desc): Mark with GTY(()).
23786 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23787 * loop-iv.c (get_simple_loop_desc): Likewise.
23788 (free_simple_loop_desc): Likewise.
23791 2013-04-16 Richard Biener <rguenther@suse.de>
23793 PR rtl-optimization/56921
23794 * loop-init.c (pass_rtl_move_loop_invariants): Add
23795 TODO_do_not_ggc_collect to todo_flags_finish.
23796 (pass_rtl_unswitch): Same.
23797 (pass_rtl_unroll_and_peel_loops): Same.
23798 (pass_rtl_doloop): Same.
23800 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23802 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23803 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23804 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23806 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23807 * tree-streamer.c (record_common_node): Adjust reference in comment.
23809 2013-04-17 Terry Guo <terry.guo@arm.com>
23811 * config/arm/cortex-m4.md: Add a new bypass.
23813 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23815 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23817 (*subs_<optab><mode>_multp2): New pattern.
23818 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23819 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23821 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23823 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23824 (*subs_mul_imm_<mode>): New pattern.
23826 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23829 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23830 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23831 (vsx_movti_32bit): Change j->wa to O->wa.
23833 2013-04-16 Richard Biener <rguenther@suse.de>
23835 PR rtl-optimization/56921
23836 * loop-init.c (pass_rtl_move_loop_invariants): Add
23837 TODO_do_not_ggc_collect to todo_flags_finish.
23838 (pass_rtl_unswitch): Same.
23839 (pass_rtl_unroll_and_peel_loops): Same.
23840 (pass_rtl_doloop): Same.
23842 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23844 * config/arm/arm.c (emit_multi_reg_push): New declaration
23845 for an existing function.
23846 (arm_emit_strd_push): New function.
23847 (arm_expand_prologue): Used here.
23848 (arm_emit_ldrd_pop): New function.
23849 (arm_expand_epilogue): Used here.
23850 (arm_get_frame_offsets): Update condition.
23851 (arm_emit_multi_reg_pop): Add a special case for load of a single
23852 register with writeback.
23854 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23856 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23859 2013-04-16 Richard Biener <rguenther@suse.de>
23861 PR tree-optimization/56756
23862 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23863 (first_mem_ref_loc): New.
23864 (execute_sm): Place the load temporarily before a previous
23865 access instead of in the latch edge to ensure its SSA dependencies
23866 are defined at points dominating the load.
23868 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23870 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23871 correct fix by moving header and footer insn to the footer of
23872 the merged basic block. Clear BB_END of the merged-away block.
23874 PR middle-end/43631
23875 * emit-rtl.c (make_note_raw): New function.
23876 (link_insn_into_chain): New static inline function.
23877 (add_insn): Use it.
23878 (add_insn_before, add_insn_after): Factor insn chain linking code...
23879 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23880 using link_insn_into_chain.
23881 (note_outside_basic_block_p): New helper function for emit_note_after
23882 and emit_note_before.
23883 (emit_note_after): Use nobb variant of add_insn_after if the note
23884 should not be contained in a basic block.
23885 (emit_note_before): Use nobb variant of add_insn_before if the note
23886 should not be contained in a basic block.
23887 (emit_note_copy): Use make_note_raw.
23888 (emit_note): Likewise.
23889 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23890 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23891 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23892 the moved barrier the tail of the basic block it follows.
23893 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23895 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23897 PR tree-optimization/56962
23898 * gimple-ssa-strength-reduction.c (record_increment): Only set
23899 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23900 either rhs1 or rhs2 is equal to c->base_expr.
23902 2013-04-15 Richard Biener <rguenther@suse.de>
23904 PR tree-optimization/56933
23905 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23907 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23908 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23909 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23910 dependence check ...
23911 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23913 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23914 GROUP_READ_WRITE_DEPENDENCE.
23916 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23918 * emit-rtl.c (reset_all_used_flags): New function.
23919 (verify_rtl_sharing): Call reset_all_used_flags before and after
23920 performing the checks.
23922 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23924 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23925 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23926 * config/arm/constraints.md (De): New constraint.
23927 * config/arm/neon.md (anddi3_neon): Delete.
23928 (neon_vand<mode>): Expand to standard anddi3 pattern.
23929 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23930 Move earlier in the file.
23931 (neon_inv_logic_op2): Likewise.
23932 (arm_anddi_operand_neon): New predicate.
23934 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23936 * configure.ac (gcc_cv_ld_as_needed): Set
23937 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23938 Use -z ignore, -z record on *-*-solaris2*.
23939 (HAVE_LD_AS_NEEDED): Update comment.
23940 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23941 * configure: Regenerate.
23942 * config.in: Regenerate.
23943 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23944 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23945 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23946 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23947 equivalents. Fix markup.
23948 * doc/tm.texi: Regenerate.
23950 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23952 * config/i386/i386.opt: New option mstack-protector-guard=.
23953 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23954 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23955 TARGET_SSP_TLS_GUARD.
23956 * config/i386/i386.c (ix86_option_override_internal): Set
23957 ix86_stack_protector_guard.
23958 * config/i386/i386.md (stack_protect_set): Enable for
23959 TARGET_SSP_TLS_GUARD only.
23960 (stack_protect_set_<mode>): Ditto.
23961 (stack_protect_test): Ditto.
23962 (stack_protect_test_<mode>): Ditto.
23963 * doc/invoke.texi (i386 Option): Document.
23965 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23968 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23969 (S_MODES): Set H_MODE bit.
23970 (SF_MODES): Set only S_MODE and SF_MODE bits.
23971 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23972 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23973 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23974 <MODE_FLOAT>: Likewise.
23976 2013-04-15 Joey Ye <joey.ye@arm.com>
23978 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23980 2013-04-15 Joey Ye <joey.ye@arm.com>
23982 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23984 (thumb_far_jump_used_p): Count instruction size and set
23987 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23989 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23990 * resource.c (mark_target_live_regs): Reformat conditional block.
23992 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23994 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23995 notes, they are emitted only just before final.
23996 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23998 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
24000 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
24001 * cfgrtl.c (delete_insn): Call it here instead.
24002 * lra-spills.c (lra_final_code_change): Use delete_insn.
24003 * haifa-sched.c (sched_remove_insn): Likewise.
24004 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
24005 returning to the nop pool.
24006 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
24007 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
24009 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
24011 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
24012 * doc/tm.texi: Regenerated.
24014 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
24016 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
24019 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
24021 * df-core.c (df_find_def): Compare register numbers.
24022 (df_find_use): Likewise.
24024 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
24027 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
24028 lra_in_progress for return.
24030 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
24032 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
24033 define_insn into define_insn_and_split and emit movsicc patterns.
24035 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
24037 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
24039 2013-04-12 Richard Biener <rguenther@suse.de>
24041 * tree-pass.h (TODO_do_not_ggc_collect): New.
24042 * passes.c (execute_one_ipa_transform_pass): Honor
24043 TODO_do_not_ggc_collect.
24044 (execute_one_pass): Likewise.
24047 2013-04-10 Richard Biener <rguenther@suse.de>
24049 * passes.c (init_optimization_passes): Remove reload pass.
24050 * ira.c (do_reload): Merge into ...
24052 (rest_of_handle_reload): Remove.
24053 (pass_reload): Likewise.
24054 * config/i386/i386.c (ix86_option_override): Refer to ira instead
24055 of reload for vzeroupper pass placement.
24057 2013-04-12 Jakub Jelinek <jakub@redhat.com>
24059 PR tree-optimization/56918
24060 PR tree-optimization/56920
24061 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
24062 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
24063 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
24064 use wide_mul_with_sign method.
24066 2013-04-12 Richard Biener <rguenther@suse.de>
24068 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
24069 not be considered a gimple constant.
24071 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
24073 * fold-const.c (const_binop): Handle vector shifts by a scalar.
24074 (fold_binary_loc): Call const_binop also for mixed vector-scalar
24077 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
24078 Jakub Jelinek <jakub@redhat.com>
24080 * opts.c: Include diagnostic-color.h.
24081 (common_handle_option): Handle OPT_fdiagnostics_color_.
24082 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
24083 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
24084 (diagnostic-color.o): New.
24085 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
24086 (diagnostic_color_rule): New enum.
24087 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
24088 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
24089 the location string.
24090 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
24091 either NULL, or color kind.
24092 * diagnostic-color.c: New file.
24093 * diagnostic-color.h: New file.
24094 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
24096 * doc/invoke.texi (-fdiagnostics-color): Document.
24097 * pretty-print.h (pp_show_color): Define.
24098 (struct pretty_print_info): Add show_color field.
24099 * diagnostic.c: Include diagnostic-color.h.
24100 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
24101 macros. Colorize error:, warning: etc. strings and also the location
24103 (diagnostic_show_locus): Colorize the caret line.
24104 * pretty-print.c: Include diagnostic-color.h.
24105 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
24106 inside of %< %> quotes or quoted through q format modifier.
24108 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24110 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
24112 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24114 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
24115 code in CC_NZ mode.
24116 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
24119 2013-04-11 Marek Polacek <polacek@redhat.com>
24121 PR tree-optimization/48184
24122 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
24124 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
24126 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
24127 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
24128 (skip_simple_arithmetic): Tidy up.
24129 * tree.h (skip_simple_constant_arithmetic): Declare.
24131 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24133 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
24135 2013-04-11 Richard Biener <rguenther@suse.de>
24137 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24138 generate vector constants.
24140 2013-04-11 Richard Biener <rguenther@suse.de>
24142 PR tree-optimization/56878
24143 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
24144 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
24146 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
24147 Prefer to align the DR with the most invariant base address.
24149 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
24151 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
24154 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
24156 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
24157 floating-point vector comparisons against 0.
24159 2013-04-11 Jakub Jelinek <jakub@redhat.com>
24161 PR tree-optimization/56899
24162 * fold-const.c (extract_muldiv_1): Apply distributive law
24163 only if TYPE_OVERFLOW_WRAPS (ctype).
24165 2013-04-11 Bin Cheng <bin.cheng@arm.com>
24168 * ira-costs.c (scan_one_insn): Check whether the source rtx of
24169 loading has side effect.
24171 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
24173 * config/sparc/sparc.c: Include tree-pass.h.
24174 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
24175 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
24176 head of file. Change return type. Split off gate function.
24177 (sparc_gate_work_around_errata): New function.
24178 (pass_work_around_errata): New pass definition.
24179 (insert_pass_work_around_errata) New pass insert definition to
24180 insert pass_work_around_errata just after delayed-branch scheduling.
24181 (sparc_option_override): Insert the pass.
24182 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
24184 2013-04-10 David S. Miller <davem@davemloft.net>
24186 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
24187 or -mcpu=hypersparc.
24189 * target.def (cstore_mode): New hook.
24190 * target.h: Include insn-codes.h
24191 * targhooks.c: Likewise.
24192 (default_cstore_mode): New function.
24193 * targhooks.h: Declare it.
24194 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
24195 * doc/tm.texi: Rebuild.
24196 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
24197 target hook, rather than inspecting the insn_data.
24198 * config/sparc/sparc.c (sparc_cstore_mode): New function.
24199 (TARGET_CSTORE_MODE): Redefine.
24200 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
24202 * config/sparc/predicates.md (cstore_result_operand): New special
24204 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
24205 Use it for operand 0.
24206 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
24207 (*snesi_special): Likewise.
24208 (*snesi_zero): Likewise.
24209 (*seqsi_zero): Likewise.
24210 (*sltu_insn): Likewise.
24211 (*sgeu_insn): Likewise.
24212 (*seqdi_special): Make operand 0 and comparison operation be of
24214 (*snedi_special): Likewise.
24215 (*snedi_special_vis3): Likewise.
24216 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
24217 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
24218 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
24219 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
24220 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
24221 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
24222 (*sltu_extend_sp64): Likewise.
24223 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
24224 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
24225 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
24226 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
24227 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
24229 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
24231 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
24232 (aarch64_start_file): Use the new function.
24234 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
24235 Jason Merrill <jason@redhat.com>
24237 * common.opt: Add -gdwarf.
24238 * opts.c (common_handle_option): Handle it.
24239 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
24241 2013-04-10 Richard Biener <rguenther@suse.de>
24243 * passes.c (execute_todo): Do not call ggc_collect conditional here.
24244 (execute_one_ipa_transform_pass): But unconditionally here.
24245 (execute_one_pass): And here.
24246 (init_optimization_passes): Remove reload pass.
24247 * tree-pass.h (TODO_ggc_collect): Remove.
24248 (pass_reload): Likewise.
24249 * ira.c (do_reload): Merge into ...
24251 (rest_of_handle_reload): Remove.
24252 (pass_reload): Likewise.
24253 * config/i386/i386.c (ix86_option_override): Refer to ira instead
24254 of reload for vzeroupper pass placement.
24255 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
24256 and todo_flags_finish of all passes.
24258 2013-04-10 Richard Biener <rguenther@suse.de>
24260 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
24261 first_const_oprnd field, rename first_def_type to first_op_type.
24262 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
24263 (vect_get_and_check_slp_defs): Always use the type of the
24264 operand. Allow mixed vect_external_def, vect_constant_def types.
24265 (vect_get_constant_vectors): Handle mixed vect_external_def,
24266 vect_constant_def types.
24268 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
24270 PR tree-optimization/55524
24271 * tree-ssa-math-opts.c
24272 (convert_mult_to_fma): Don't use an fms construct
24273 when we don't have an fms operation, but fnma, and it looks
24274 likely that we'll be able to use the latter.
24276 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
24278 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
24280 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
24281 inline fail caused by overwritable functions.
24283 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
24285 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
24286 unnecessary bits in the constant power of two case.
24288 2013-04-10 Richard Biener <rguenther@suse.de>
24290 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
24291 broken code swapping operands.
24292 (vect_build_slp_tree): Do not compute load permutations here.
24293 (vect_analyze_slp_instance): Compute load permutations here,
24294 after building the SLP tree.
24296 2013-04-09 Christian Bruel <christian.bruel@st.com>
24298 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
24299 of next/prev_real_insn.
24301 2013-04-09 Jan Hubicka <jh@suse.cz>
24303 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
24304 Drop aliased parameter.
24305 (function_and_variable_visibility): Do not handle alias pairs.
24306 * cgraph.c (varpool_externally_visible_p): Update prototype.
24307 * varpool.c (varpool_add_new_variable): Update.
24309 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24311 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
24313 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
24315 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
24317 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
24319 2013-04-09 Marek Polacek <polacek@redhat.com>
24321 PR tree-optimization/48762
24322 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
24324 2013-04-09 Richard Biener <rguenther@suse.de>
24326 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
24328 (vect_build_slp_tree): Likewise.
24329 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
24330 calculating the cost of a SLP instance.
24331 (vect_analyze_slp_instance): Use it from here, after building
24334 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24336 PR middle-end/56883
24337 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
24338 expand_omp_for_static_chunk): Use simple_p = true in
24339 force_gimple_operand_gsi calls when assigning to addressable decls.
24341 2013-04-09 Jeff Law <law@redhat.com>
24343 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
24344 when the boolean was created by converting a wider object which
24345 had a boolean range.
24347 2013-04-09 Richard Biener <rguenther@suse.de>
24349 * tree-vectorizer.h (slp_void_p): Remove.
24350 (slp_tree): Typedef before _slp_tree declaration.
24351 (struct _slp_tree): Use a vector of slp_tree as children.
24352 (vect_get_place_in_interleaving_chain): Remove.
24353 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
24355 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
24357 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
24358 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
24359 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
24360 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
24361 Use slp_node instead of slp_void_p and adjust.
24363 2013-04-09 Richard Biener <rguenther@suse.de>
24365 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
24366 work that is not necessary.
24368 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24370 PR tree-optimization/56854
24371 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
24372 forward into clobber stmts if it would change MEM_REF lhs into
24375 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
24377 * tree.c (type_hash_lookup, type_hash_add): Make static.
24378 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
24380 2013-04-09 Richard Biener <rguenther@suse.de>
24382 * tree.h (unsave_expr_now): Remove.
24383 * tree-inline.c (mark_local_for_remap_r): Remove.
24384 (unsave_expr_1): Likewise.
24385 (unsave_r): Likewise.
24386 (unsave_expr_now): Likewise.
24387 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
24388 (propagate_tree_value): Likewise.
24390 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
24392 * doc/rtl.texi (sequence): Rewrite documentation to match the
24393 current use of SEQUENCE rtl objects.
24394 * rtl.def (SEQUENCE): Likewise.
24396 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
24397 Update documentation.
24398 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
24399 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
24401 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
24403 2013-04-08 Teresa Johnson <tejohnson@google.com>
24405 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
24406 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
24408 (estimate_edge_size_and_time): Add comment to suggest using rounding
24410 (estimate_node_size_and_time): Ditto.
24411 (remap_edge_change_prob): Use helper rounding divide methods.
24412 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
24413 (gimple_mod_pow2_value_transform): Ditto.
24414 (gimple_mod_subtract_transform): Ditto.
24415 (gimple_ic_transform): Ditto.
24416 (gimple_stringops_transform): Ditto.
24417 * stmt.c (conditional_probability): Ditto.
24418 (emit_case_dispatch_table): Ditto.
24419 * lto-cgraph.c (merge_profile_summaries): Ditto.
24420 * tree-optimize.c (execute_fixup_cfg): Ditto.
24421 * cfgcleanup.c (try_forward_edges): Ditto.
24422 * cfgloopmanip.c (scale_loop_profile): Ditto.
24424 (duplicate_loop_to_header_edge): Ditto.
24425 (lv_adjust_loop_entry_edge): Ditto.
24426 * tree-vect-loop.c (vect_transform_loop): Ditto.
24427 * profile.c (compute_branch_probabilities): Ditto.
24428 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
24429 * lto-streamer-in.c (input_cfg): Ditto.
24430 * gimple-streamer-in.c (input_bb): Ditto.
24431 * ipa-cp.c (update_profiling_info): Ditto.
24432 (update_specialized_profile): Ditto.
24433 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
24434 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
24436 * sched-rgn.c (compute_dom_prob_ps): Ditto.
24437 (compute_trg_info): Ditto.
24438 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
24439 (purge_dead_edges): Ditto.
24440 * loop-unswitch.c (unswitch_loop): Ditto.
24441 * cgraphclones.c (cgraph_clone_edge): Ditto.
24442 (cgraph_clone_node): Ditto.
24443 * tree-inline.c (copy_bb): Ditto.
24444 (copy_edges_for_bb): Ditto.
24445 (initialize_cfun): Ditto.
24446 (copy_cfg_body): Ditto.
24447 (expand_call_inline): Ditto.
24449 2013-04-08 Kai Tietz <ktietz@redhat.com>
24451 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
24452 TARGET_CYGWIN64 by TARGET_64BIT.
24454 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24456 * config/epiphany/epiphany.md (GPR_1): New constant.
24457 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
24458 * config/epiphany/epiphany.c (gen_compare_reg):
24459 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
24460 is already in place.
24461 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
24462 Don't require being called during rtl expansion; If y operlaps r0,
24464 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
24465 (epiphany_expand_epilogue): Likewise.
24467 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
24468 Don't use CC_FPmode for ORDERED / UNORDERED.
24469 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
24471 * config/epiphany/constraints.md (CnL): New constraint.
24472 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
24473 * config/epiphany/predicates.md (add_operand): Allow 1024.
24475 * config/epiphany/epiphany.md (logical_op): New code iterator.
24476 (op_mnc): New code attribute.
24477 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
24478 (mov_f+1, mov_f+2): New peephole2 patterns.
24480 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
24481 (cstoresi4): Also allow re-use of zero result when doing a NE
24482 comparison to a non-zero operand.
24483 Use (clobber (scratch)) for first insn if the gpr output is not needed.
24485 * config/epiphany/epiphany.md (<insn_opname>v2si3):
24486 Use gen_addsi3_i / gen_subsi3_i.
24488 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24492 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
24493 contain anything but clobbers, at most one __builtin_stack_restore,
24494 optionally debug stmts and final resx, and if it has at least one
24495 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
24496 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
24497 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
24498 which isn't defaut definition, remove them.
24499 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
24500 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
24501 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
24502 with MEM_REF LHS with SSA_NAME address.
24504 2013-04-08 Jeff Law <law@redhat.com>
24506 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
24508 2013-04-08 Richard Biener <rguenther@suse.de>
24510 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
24512 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
24513 determined vector type.
24514 (vect_analyze_data_refs): Likewise.
24515 (vect_get_new_vect_var): Adjust.
24516 (vect_create_destination_var): Preserve SSA name versions.
24517 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
24518 not dump anything here.
24520 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24522 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
24523 Add member lr_slot_known.
24524 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
24526 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
24527 Remove code that sets lr_slot_offset according to what a previous
24528 version of epiphany_emit_save_restore used to do.
24529 (epiphany_emit_save_restore): When doing an lr save or restore,
24530 set/verify lr_slot_known and lr_slot_offset.
24532 2013-04-08 Xinyu Qi <xyqi@marvell.com>
24535 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
24538 2013-04-08 Richard Biener <rguenther@suse.de>
24540 * alias.c (find_base_term): Fix thinko in previous change.
24542 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24544 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
24545 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
24546 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
24547 if possible to compute val.
24548 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
24549 For QImode integers don't require anything about precision. Use
24550 const_with_all_bytes_same to find out if the constant doesn't have
24551 repeated bytes in it.
24553 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24555 * config/s390/s390.c (s390_expand_insv): Only accept insertions
24558 2013-04-08 Marek Polacek <polacek@redhat.com>
24560 PR rtl-optimization/48182
24561 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
24564 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24567 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
24568 nuses, make sure we have a label.
24570 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24573 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
24574 (rs6000_emit_swdiv_low_precision): Remove.
24575 (rs6000_emit_swdiv): Rewrite to handle between one and four
24576 iterations of Newton-Raphson generally; modify required number of
24577 iterations for some cases.
24578 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
24580 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
24582 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
24583 set-but-unused variable.
24585 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
24586 basic blocks of released function bodies garbage-collectable.
24588 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
24589 (struct rtl_opt_pass): Add TODO_df_finish.
24591 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
24593 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24595 * config/arm/constraints.md (q): New constraint.
24596 * config/arm/ldrdstrd.md: New file.
24597 * config/arm/arm.md (ldrdstrd.md) New include.
24598 (arm_movdi): Use "q" instead of "r" constraint
24599 for double-word memory access.
24600 (movdf_soft_insn): Likewise.
24601 * config/arm/vfp.md (movdi_vfp): Likewise.
24602 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
24603 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
24604 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
24605 (mem_ok_for_ldrd_strd): Likewise.
24606 (output_move_double): Update assertion.
24608 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24610 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
24612 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24614 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
24615 define_insn_and_split.
24616 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
24618 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24620 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
24621 define_insn_and_split.
24622 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
24623 (shiftsi3_compare): New pattern.
24624 (rrx): New pattern.
24625 * config/arm/unspecs.md (UNSPEC_RRX): New.
24627 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24629 * config/arm/arm.md (negdi_extendsidi): New pattern.
24630 (negdi_zero_extendsidi): Likewise.
24632 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24634 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
24635 define_insn_and_split.
24636 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
24637 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
24639 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24641 * config/arm/arm.md (arm_subdi3): Convert define_insn into
24642 define_insn_and_split.
24643 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
24644 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
24646 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24648 * config/arm/arm.md (subsi3_carryin): New pattern.
24649 (subsi3_carryin_const): Likewise.
24650 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
24651 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
24653 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24655 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
24657 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24659 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
24660 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
24662 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24664 * config/arm/arm.c (arm_expand_builtin): Change fcode
24665 type to unsigned int.
24667 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24669 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
24671 2013-04-04 Ian Lance Taylor <iant@google.com>
24673 * doc/standards.texi (Standards): The Go frontend supports the Go 1
24676 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
24678 PR middle-end/56729
24679 * df-scan.c (df_insn_delete): Disable failing assert.
24681 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24683 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
24684 New function prototype.
24685 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
24686 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
24687 (arm_builtin_vectorized_function): New function.
24689 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24691 * config/arm/arm_neon_builtins.def: New file.
24692 * config/arm/arm.c (neon_builtin_data): Move contents to
24693 arm_neon_builtins.def.
24694 (enum arm_builtins): Include neon builtin definitions.
24695 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
24696 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
24698 2013-04-04 Marek Polacek <polacek@redhat.com>
24700 PR tree-optimization/48186
24701 * predict.c (maybe_hot_frequency_p): Return false if
24702 HOT_BB_FREQUENCY_FRACTION is 0.
24703 (cgraph_maybe_hot_edge_p): Likewise.
24705 2013-04-04 Richard Biener <rguenther@suse.de>
24707 PR tree-optimization/56826
24708 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
24711 2013-04-04 Richard Biener <rguenther@suse.de>
24713 PR tree-optimization/56213
24714 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
24715 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
24717 2013-04-04 Richard Biener <rguenther@suse.de>
24719 PR tree-optimization/56837
24720 * tree-loop-distribution.c (classify_partition): For non-zero
24721 values require that the value has the same precision as its
24722 mode to be useful as memset value.
24724 2013-04-03 Nick Clifton <nickc@redhat.com>
24726 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
24727 (fmssf4): Use fmsf.s on E3V5 architectures.
24728 (fnmasf4): Use fnmaf.s on E3V5 architectures.
24729 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24731 2013-04-03 Jeff Law <law@redhat.com>
24733 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24734 (lra-eliminations.o): Likewise.
24736 2013-04-03 Teresa Johnson <tejohnson@google.com>
24738 * gcov-io.c (compute_working_sets): Moved most of body of old
24739 compute_working_sets here from profile.c.
24740 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24741 (gcov_working_set_t): Moved typedef here from basic-block.h
24742 (compute_working_set): Declare.
24743 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24744 (get_working_sets): Renamed from compute_working_set,
24745 replace most of body with call to new compute_working_sets.
24746 (get_exec_counts): Replace call to compute_working_sets
24747 to get_working_sets.
24748 * profile.h (get_working_sets): Renamed from compute_working_set.
24749 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24750 to get_working_sets.
24751 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24752 * gcov-dump.c (dump_working_sets): New function.
24754 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24756 * hwint.c (sext_hwi, zext_hwi): New functions.
24757 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24758 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24759 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24760 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24761 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24762 (sext_hwi, zext_hwi): New functions.
24764 2013-04-03 Jeff Law <law@redhat.com>
24766 PR tree-optimization/56799
24767 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24768 back test for widening conversion erroneously dropped in prior change.
24770 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24773 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24774 instead of next_real_insn.
24776 2013-04-03 Marek Polacek <polacek@redhat.com>
24779 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24781 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24784 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24786 (thumb1_output_casesi): Likewise.
24787 (thumb2_output_casesi): Likewise.
24789 2013-04-03 Richard Biener <rguenther@suse.de>
24791 PR tree-optimization/56817
24792 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24794 (tree_unroll_loops_completely_1): ... new function to manually
24795 walk the loop tree, properly defering outer loops of unrolled
24796 loops to later iterations.
24798 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24800 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24801 (vectorizable_load): Likewise.
24802 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24803 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24805 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24807 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24810 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24812 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24814 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24816 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24818 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24820 PR tree-optimization/56790
24821 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24824 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24826 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24828 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24829 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24832 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24835 * tree.h (force_folding_builtin_constant_p): New decl.
24836 * builtins.c (force_folding_builtin_constant_p): New variable.
24837 (fold_builtin_constant_p): Fold immediately also if
24838 force_folding_builtin_constant_p.
24840 2013-04-03 Richard Biener <rguenther@suse.de>
24842 PR tree-optimization/56812
24843 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24844 DRs of the same interleaving chain are independent.
24846 2013-04-02 Jason Merrill <jason@redhat.com>
24848 * gdbinit.in (pbb): Use debug fn.
24850 2013-04-02 Lawrence Crowl <crowl@google.com>
24852 * sese.h (struct ivtype_map_elt_s): Remove unused.
24853 (extern debug_ivtype_map): Remove unused.
24854 (extern eq_ivtype_map_elts): Remove unused.
24855 * sese.c (debug_ivtype_map): Removed unused.
24856 (debug_ivtype_map_1): Removed unused.
24857 (debug_ivtype_elt): Remove unused.
24858 (eq_ivtype_map_elts): Remove unused.
24860 2013-04-02 Kai Tietz <ktietz@redhat.com>
24863 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24864 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24865 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24867 (legitimize_pe_coff_symbol): Likewise.
24868 (is_imported_p): New helper-function.
24869 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24870 for Windows x64 targets.
24871 (ix86_expand_prologue): Optimize for pe-coff targets.
24872 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24873 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24874 medium/large code-model.
24875 (legitimize_pic_address): Likewise.
24876 (legitimize_tls_address): Likewise.
24877 (ix86_expand_call): Likewise.
24878 (x86_output_mi_thunk): Likewise.
24879 (get_dllimport_decl): Add new beimport argument.
24880 (construct_plt_address): Don't assert for x64 pe-coff targets.
24881 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24883 (SYMBOL_FLAG_STUBVAR): New macro.
24884 (SYMBOL_REF_STUBVAR_P): Likewise.
24885 * config/i386/winnt.c (stub_list): New structure.
24886 (stub_head): New local variable.
24887 (i386_pe_record_stub): New function.
24888 (i386_pe_file_end): Emit refptr-stubs.
24890 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24892 PR rtl-optimization/56745
24893 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24894 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24897 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24898 and both of them are MEM_REFs, just compare first argument for
24899 equality and attempt to deal even with differing offsets.
24902 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24903 of gimple_clobber_p to be MEM_REF.
24904 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24905 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24906 after gimplification.
24907 * asan.c (get_mem_ref_of_assignment): Don't instrument
24908 gimple_clobber_p stmts.
24909 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24910 gimple_clobber_p stmt if they have MEM_REF lhs and
24911 are dead because of another gimple_clobber_p stmt.
24912 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24913 gimple_clobber_p stmts like debug stmts.
24914 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24915 that refer to unused VAR_DECLs or uninitialized values.
24916 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24917 gimple_clobber_p stmts if they refer to removed parameters.
24918 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24921 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24923 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24924 using SWI48 mode attribute.
24926 2013-04-02 Wei Mi <wmi@google.com>
24928 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24929 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24930 *<rotate_insn><mode>3_mask in i386.md.
24932 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24934 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24936 2013-04-02 Richard Biener <rguenther@suse.de>
24938 PR tree-optimization/56778
24939 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24940 Runtime alias tests are not supported for gather loads.
24941 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24942 stmts referenced from SSA operands before updating SSA form.
24944 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24945 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24947 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24948 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24949 * config/arm/cortex-a53.md: New file.
24950 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24951 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24952 * config/arm/arm.c (arm_issue_rate): Likewise.
24953 * config/arm/arm-tune.md: Regenerate
24954 * config/arm/arm-tables.opt: Regenerate.
24955 * config/arm/arm-cores.def: Add cortex-a53.
24957 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24959 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24962 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24964 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24965 scalar load/store operations using B/H registers.
24966 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24968 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24970 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24972 * config/aarch64/aarch64.c
24973 (aarch64_simd_scalar_immediate_valid_for_move): New.
24974 * config/aarch64/aarch64-protos.h
24975 (aarch64_simd_scalar_immediate_valid_for_move): New.
24976 * config/aarch64/constraints.md (Dh, Dq): New.
24977 * config/aarch64/iterators.md (hq): New.
24979 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24981 * reorg.c (get_branch_condition): Deal with conditional returns.
24982 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24984 2013-04-01 Wei Mi <wmi@google.com>
24986 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24987 Truncate operand 2 using %b asm operand modifier.
24988 (*<shift_insn><mode>3_mask): Ditto.
24989 (*<rotate_insn><mode>3_mask): Ditto.
24991 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24993 PR middle-end/56798
24994 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24996 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24998 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
25000 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
25002 2013-03-30 Lawrence Crowl <crowl@google.com>
25004 * dse.c (clear_alias_sets): Remove never set.
25005 (disqualified_clear_alias_sets): Remove never set.
25006 (clear_alias_mode_pool): Remove never set.
25007 (dse_step0): Remove condition that is never true.
25008 (canon_address): Remove condition that is never true.
25009 (dse_step7): Remove condition that is never true.
25010 (rest_of_handle_dse): Remove condition that is never true.
25011 (rest_of_handle_dse::did_global): Remove never read from above.
25012 (dse_step2_spill): Remove never called from above.
25013 (dse_step5_spill): Remove never called from above.
25015 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
25017 * doc/md.texi (Standard Names) <casesi>: Update documentation for
25018 JUMP_TABLE_DATA changes.
25019 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
25020 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
25021 (Insns) <jump_table_data>: New entry.
25022 * doc/tm.texi: Regenerate.
25024 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
25026 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
25027 for table jump at the end of a basic block using tablejump_p.
25028 * targhooks.c (default_invalid_within_doloop): Likewise.
25029 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
25030 target hook implementation that is identical to the default hook.
25031 (rs6000_invalid_within_doloop): Remove.
25033 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
25034 unused variable from tablejump_p call.
25036 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
25037 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
25038 (INSN_DELETED_P): Likewise.
25039 (emit_jump_table_data): New prototype.
25040 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
25041 after 4th as unused.
25042 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
25043 * sched-vis.c (print_insn): Likewise.
25044 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
25045 insn for compatibility with back ends that use next_active_insn to
25046 identify jump table data.
25047 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
25048 (remove_insn): Likewise.
25049 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
25051 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
25052 (emit_jump_table_data): New function.
25054 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
25055 basic block, a JUMP_TABLE_DATA never is.
25056 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
25057 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
25058 off from code handling real insns.
25059 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
25060 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
25061 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
25062 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
25063 is not a NONDEBUG_INSN_P.
25064 * ira-costs.c (scan_one_insn): Likewise.
25065 * jump.c (mark_all_labels): Likewise.
25066 (mark_jump_label_1): Likewise.
25067 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
25068 * lra.c (get_insn_freq): Expect all insns reaching here to be in
25070 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
25071 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
25072 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
25073 JUMP_TABLE_DATA_P insns.
25074 (calculate_elim_costs_all_insns): Likewise.
25075 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
25076 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
25077 (delete_output_reload): Code style fixups.
25078 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
25079 insn flags on this non-insn.
25080 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
25081 as scheduling barriers, for pre-change compatibility.
25082 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
25083 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
25085 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
25086 redundant JUMP_TABLE_DATA_P test.
25087 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
25088 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
25089 (frv_for_each_packet): Likewise.
25090 * config/i386/i386.c (min_insn_size): Likewise.
25091 (ix86_avoid_jump_mispredicts): Likewise.
25092 * config/m32r/m32r.c (m32r_is_insn): Likewise.
25093 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
25094 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
25095 (mips16_insn_length): Robustify.
25096 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
25097 (mips16_split_long_branches): Likewise.
25098 * config/pa/pa.c (pa_combine_instructions): Likewise.
25099 * config/rs6000/rs6000.c (get_next_active_insn): Treat
25100 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
25101 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
25102 as contributing to pool range lengths.
25103 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
25104 Remove redundant JUMP_TABLE_DATA_P test.
25105 (sh_loop_align): Likewise.
25106 (split_branches): Likewise.
25107 (sh_insn_length_adjustment): Likewise.
25108 * config/spu/spu.c (get_branch_target): Likewise.
25110 2013-03-29 Jan Hubicka <jh@suse.cz>
25112 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
25113 gcov streaming; stream hot bb threshold to ltrans.
25114 * predict.c (get_hot_bb_threshold): Break out from ....
25115 (maybe_hot_count_p): ... here.
25116 (set_hot_bb_threshold): New function.
25117 * lto-section-in.c (lto_section_name): Add profile.
25118 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
25119 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
25120 and data-streamer.h
25121 (histogram_entry): New structure.
25122 (histogram, histogram_pool): New global vars.
25123 (histogram_hash): New structure.
25124 (histogram_hash::hash): New method.
25125 (histogram_hash::equal): Likewise.
25126 (account_time_size): New function.
25127 (cmp_counts): New function.
25128 (dump_histogram): New function.
25129 (ipa_profile_generate_summary): New function.
25130 (ipa_profile_write_summary): New function.
25131 (ipa_profile_read_summary): New function.
25132 (ipa_profile): Decide on threshold.
25133 (pass_ipa_profile): Add ipa_profile_write_summary and
25134 ipa_profile_read_summary.
25135 * Makefile.in (ipa.o): Update dependencies.
25136 * lto-streamer.h (LTO_section_ipa_profile): New section.
25138 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
25140 * tree.h (VAR_P): New.
25142 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
25145 * doc/invoke.texi ([-fwhole-program]): Fix typo.
25147 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
25149 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
25150 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
25151 (control_flow_insn_p): Likewise.
25152 * cfgrtl.c (duplicate_insn_chain): Likewise.
25153 * final.c (get_attr_length_1): Likewise.
25154 (shorten_branches): Likewise.
25155 (final_scan_insn): Likewise.
25156 * function.c (instantiate_virtual_regs): Likewise.
25157 * gcse.c (insert_insn_end_basic_block): Likewise.
25158 * ira-costs.c (scan_one_insn): Likewise.
25159 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
25160 * lra.c (check_rtl): Likewise.
25161 * reload1.c (elimination_costs_in_insn): Likewise.
25162 * reorg.c (follow_jumps): Likewise.
25164 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
25165 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
25166 (thumb_far_jump_used_p): Likewise.
25167 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
25168 (workaround_speculation): Likewise.
25169 (add_sched_insns_for_speculation): Likewise.
25170 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
25171 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
25172 (frv_for_each_packet): Likewise.
25173 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
25174 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
25175 (final_emit_insn_group_barriers): Likewise.
25176 * config/m32r/m32r.c (m32r_is_insn): Likewise.
25177 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
25178 (mips16_insn_length): Likewise.
25179 * config/pa/pa.c (pa_reorg): Likewise.
25180 (pa_combine_instructions): Likewise.
25181 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
25182 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
25183 (sh_reorg): Likewise.
25184 (split_branches): Likewise.
25185 * config/spu/spu.c (get_branch_target): Likewise.
25187 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
25190 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
25192 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
25193 Fix declaration name.
25195 2013-03-28 Lawrence Crowl <crowl@google.com>
25197 * graphds.h (struct graph.indicies): Remove unused.
25198 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
25199 (SCOP_ORIGINAL_PDDRS): Remove unused.
25200 * sese.h (extern insert_loop_close_phis): Removed unused.
25201 (extern insert_guard_phis): Removed unused.
25202 (extern ivtype_map_elt_info): Removed unused.
25203 (new_ivtype_map_elt): Removed unused.
25204 * sese.c (ivtype_map_elt_info): Removed unused.
25206 2013-03-28 Lawrence Crowl <crowl@google.com>
25208 * Makefile.in: Add several missing include dependences.
25210 (test-dump.o): New. This object is not added to any executable,
25211 but is present for ad-hoc testing.
25213 (debug (const bitmap_head_def &)): New.
25214 (debug (const bitmap_head_def *)): New.
25216 (extern debug (const bitmap_head_def &)): New.
25217 (extern debug (const bitmap_head_def *)): New.
25219 (debug (edge_def &)): New.
25220 (debug (edge_def *)): New.
25222 (debug (basic_block_def &)): New.
25223 (debug (basic_block_def *)): New.
25225 (dump_node (const_tree, int, FILE *)): Correct source file.
25227 (debug (die_struct &)): New.
25228 (debug (die_struct *)): New.
25230 (extern debug (die_struct &)): New.
25231 (extern debug (die_struct *)): New.
25232 * gimple-pretty-print.c
25233 (debug (gimple_statement_d &)): New.
25234 (debug (gimple_statement_d *)): New.
25235 * gimple-pretty-print.h
25236 (extern debug (gimple_statement_d &)): New.
25237 (extern debug (gimple_statement_d *)): New.
25239 (debug (ira_allocno_copy &)): New.
25240 (debug (ira_allocno_copy *)): New.
25241 (debug (ira_allocno &)): New.
25242 (debug (ira_allocno *)): New.
25244 (extern debug (ira_allocno_copy &)): New.
25245 (extern debug (ira_allocno_copy *)): New.
25246 (extern debug (ira_allocno &)): New.
25247 (extern debug (ira_allocno *)): New.
25249 (debug (live_range &)): New.
25250 (debug (live_range *)): New.
25252 (debug (lra_live_range &)): New.
25253 (debug (lra_live_range *)): New.
25255 (debug (lra_live_range &)): New.
25256 (debug (lra_live_range *)): New.
25258 (debug (omega_pb_d &)): New.
25259 (debug (omega_pb_d *)): New.
25261 (extern debug (omega_pb_d &)): New.
25262 (extern debug (omega_pb_d *)): New.
25264 (debug (const rtx_def &)): New.
25265 (debug (const rtx_def *)): New.
25267 (debug_tree (tree): Move within file.
25268 (debug_raw (const tree_node &)): New.
25269 (debug_raw (const tree_node *)): New.
25270 (dump_tree_via_hooks (const tree_node *, int)): New.
25271 (debug (const tree_node &)): New.
25272 (debug (const tree_node *)): New.
25273 (debug_verbose (const tree_node &)): New.
25274 (debug_verbose (const tree_node *)): New.
25275 (debug_head (const tree_node &)): New.
25276 (debug_head (const tree_node *)): New.
25277 (debug_body (const tree_node &)): New.
25278 (debug_body (const tree_node *)): New.
25279 (debug_vec_tree (tree): Move and reimplement in terms of dump.
25280 (debug (vec<tree, va_gc> &)): New.
25281 (debug (vec<tree, va_gc> *)): New.
25283 (extern debug (const rtx_def &)): New.
25284 (extern debug (const rtx_def *)): New.
25286 (debug_raw (simple_bitmap_def &)): New.
25287 (debug_raw (simple_bitmap_def *)): New.
25288 (debug (simple_bitmap_def &)): New.
25289 (debug (simple_bitmap_def *)): New.
25291 (extern debug (simple_bitmap_def &)): New.
25292 (extern debug (simple_bitmap_def *)): New.
25293 (extern debug_raw (simple_bitmap_def &)): New.
25294 (extern debug_raw (simple_bitmap_def *)): New.
25296 (debug (vinsn_def &)): New.
25297 (debug (vinsn_def *)): New.
25298 (debug_verbose (vinsn_def &)): New.
25299 (debug_verbose (vinsn_def *)): New.
25300 (debug (expr_def &)): New.
25301 (debug (expr_def *)): New.
25302 (debug_verbose (expr_def &)): New.
25303 (debug_verbose (expr_def *)): New.
25304 (debug (vec<rtx> &)): New.
25305 (debug (vec<rtx> *)): New.
25307 (extern debug (vinsn_def &)): New.
25308 (extern debug (vinsn_def *)): New.
25309 (extern debug_verbose (vinsn_def &)): New.
25310 (extern debug_verbose (vinsn_def *)): New.
25311 (extern debug (expr_def &)): New.
25312 (extern debug (expr_def *)): New.
25313 (extern debug_verbose (expr_def &)): New.
25314 (extern debug_verbose (expr_def *)): New.
25315 (extern debug (vec<rtx> &)): New.
25316 (extern debug (vec<rtx> *)): New.
25318 (_list_iter_cond_expr): Make inline instead of static.
25320 (debug (sreal &)): New.
25321 (debug (sreal *)): New.
25323 (extern debug (sreal &)): New.
25324 (extern debug (sreal *)): New.
25326 (extern debug_raw (const tree_node &)): New.
25327 (extern debug_raw (const tree_node *)): New.
25328 (extern debug (const tree_node &)): New.
25329 (extern debug (const tree_node *)): New.
25330 (extern debug_verbose (const tree_node &)): New.
25331 (extern debug_verbose (const tree_node *)): New.
25332 (extern debug_head (const tree_node &)): New.
25333 (extern debug_head (const tree_node *)): New.
25334 (extern debug_body (const tree_node &)): New.
25335 (extern debug_body (const tree_node *)): New.
25336 (extern debug (vec<tree, va_gc> &)): New.
25337 (extern debug (vec<tree, va_gc> *)): New.
25339 (debug (struct loop &)): New.
25340 (debug (struct loop *)): New.
25341 (debug_verbose (struct loop &)): New.
25342 (debug_verbose (struct loop *)): New.
25343 * tree-dump.c: Add header dependence.
25345 (extern debug (struct loop &)): New.
25346 (extern debug (struct loop *)): New.
25347 (extern debug_verbose (struct loop &)): New.
25348 (extern debug_verbose (struct loop *)): New.
25350 (debug (data_reference &)): New.
25351 (debug (data_reference *)): New.
25352 (debug (vec<data_reference_p> &)): New.
25353 (debug (vec<data_reference_p> *)): New.
25354 (debug (vec<ddr_p> &)): New.
25355 (debug (vec<ddr_p> *)): New.
25357 (extern debug (data_reference &)): New.
25358 (extern debug (data_reference *)): New.
25359 (extern debug (vec<data_reference_p> &)): New.
25360 (extern debug (vec<data_reference_p> *)): New.
25361 (extern debug (vec<ddr_p> &)): New.
25362 (extern debug (vec<ddr_p> *)): New.
25364 (debug (pt_solution &)): New.
25365 (debug (pt_solution *)): New.
25367 (extern debug (pt_solution &)): New.
25368 (extern debug (pt_solution *)): New.
25370 (debug (_var_map &)): New.
25371 (debug (_var_map *)): New.
25372 (debug (tree_live_info_d &)): New.
25373 (debug (tree_live_info_d *)): New.
25375 (extern debug (_var_map &)): New.
25376 (extern debug (_var_map *)): New.
25377 (extern debug (tree_live_info_d &)): New.
25378 (extern debug (tree_live_info_d *)): New.
25380 2013-03-28 Jan Hubicka <jh@suse.cz>
25382 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
25384 2013-03-28 Ian Bolton <ian.bolton@arm.com>
25386 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
25387 record only when desired or required.
25389 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
25391 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
25392 *vec_extractv2di_1_rex64. Use x64 isa attribute.
25394 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25396 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
25397 (*andsi3_compare0_uxtw): New pattern.
25398 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
25399 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
25401 2013-03-28 Jan Hubicka <jh@suse.cz>
25403 * data-streamer-in.c (streamer_read_gcov_count): New function.
25404 * gimple-streamer-out.c: Include value-prof.h.
25405 (output_gimple_stmt): Output histogram.
25406 (output_bb): Use streamer_write_gcov_count.
25407 * value-prof.c: Include data-streamer.h
25408 (dump_histogram_value): Add HIST_TYPE_MAX.
25409 (stream_out_histogram_value): New function.
25410 (stream_in_histogram_value): New function.
25411 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
25412 (stream_out_histogram_value, stream_in_histogram_value): Declare.
25413 * data-streamer-out.c (streamer_write_gcov_count): New function.
25414 (streamer_write_gcov_count_stream): New function.
25415 * lto-cgraph.c (lto_output_edge): Update counter streaming.
25416 (lto_output_node): Likewise.
25417 (input_node, input_edge): Likewise.
25418 * lto-streamer-out.c (output_cfg): Update streaming.
25419 * lto-streamer-in.c (input_cfg): Likewise.
25420 * data-streamer.h (streamer_write_gcov_count,
25421 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
25422 * gimple-streamer-in.c: Include value-prof.h
25423 (input_gimple_stmt): Input histograms.
25424 (input_bb): Update profile streaming.
25426 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25428 * genmodes.c (emit_max_int): New function.
25429 (emit_insn_modes_h): Added call to emit_max_function.
25430 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
25432 * machmode.def: Fixed comment.
25434 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25436 * combine.c (try_combine): Removed useless assert.
25437 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
25439 2013-03-28 Marek Polacek <polacek@redhat.com>
25440 Richard Biener <rguenther@suse.de>
25442 PR tree-optimization/56695
25443 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
25444 build signed result of a vector comparison.
25445 * tree-cfg.c (verify_gimple_comparison): Check that a result
25446 of a vector comparison has signed type.
25448 2013-03-28 Richard Biener <rguenther@suse.de>
25450 PR tree-optimization/37021
25451 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
25452 do not restrict gaps between groups.
25453 * tree-vect-stmts.c (vectorizable_load): Properly account for
25454 a gap between groups.
25456 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
25458 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
25459 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
25462 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
25464 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
25465 * doc/extend.texi (Named Address Spaces): Ditto.
25466 (Variable Attributes): Ditto.
25468 2013-03-27 Kai Tietz <ktietz@redhat.com>
25470 * config.build: Add support for cygwin x64 target.
25471 * config.gcc: Likewise.
25472 * config.host: Likewise.
25473 * configure.ac: Likewise
25474 * configure: Regenerated.
25476 2013-03-27 Kai Tietz <ktietz@redhat.com>
25478 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
25479 * config/i386/t-cygwin-w64: New file.
25480 * config/i386/cygwin-w64.h: New file.
25481 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
25482 and add support for x64-cygwin target.
25483 (CPP_SPEC): Likewise.
25484 (CXX_WRAP_SPEC_LIST): Undefine before define.
25485 (LIBGCJ_SONAME): Use 15 as version.
25487 2013-03-27 Richard Biener <rguenther@suse.de>
25489 PR tree-optimization/56716
25490 * tree-ssa-structalias.c (perform_var_substitution): Adjust
25491 dumping for ref nodes.
25493 2013-03-27 Martin Jambor <mjambor@suse.cz>
25495 PR tree-optimization/55334
25496 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
25497 restricted pointers to arrays.
25499 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
25501 * Makefile.in (.SUFFIXES): Add .cc.
25502 (.c.o): Apply same recipe for implicit rule .cc.o.
25504 2013-03-27 Richard Biener <rguenther@suse.de>
25506 PR tree-optimization/37021
25507 * tree-vect-data-refs.c (vect_check_strided_load): Allow
25508 REALPART/IMAGPART_EXPRs around the supported refs.
25509 * tree-ssa-structalias.c (find_func_aliases): Assume that
25510 floating-point values are not used to transfer pointers.
25512 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
25514 * target.def (TARGET_HAS_IFUNC_P): New target hook.
25515 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
25516 * doc/tm.texi: Regenerate.
25517 * targhooks.h (default_has_ifunc_p): New.
25518 * targhooks.c (default_has_ifunc_p): Ditto.
25519 * config/linux-protos.h: New file.
25520 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
25521 hook for linux which disables support of indirect functions in android.
25522 * config/linux-android.c: New file.
25523 * config/t-linux-android.c: Ditto.
25524 * config.gcc: Added new object file linux-android.o.
25525 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
25526 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
25527 * varasm.c (do_assemble_alias): Likewise.
25528 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
25529 doesn't support indirect functions.
25530 * configure: Regenerate.
25532 2013-03-27 Bin Cheng <bin.cheng@arm.com>
25535 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
25536 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
25539 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25541 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
25543 2013-03-27 Terry Guo <terry.guo@arm.com>
25545 * config/arm/arm-cores.def: Added core cortex-r7.
25546 * config/arm/arm-tune.md: Regenerated.
25547 * config/arm/arm-tables.opt: Regenerated.
25548 * doc/invoke.texi: Added entry for core cortex-r7.
25550 2013-03-27 Walter Lee <walt@tilera.com>
25552 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
25553 double-decrement of next_scratch_regno.
25555 2013-03-27 Walter Lee <walt@tilera.com>
25557 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
25559 (insn_v1mulus): Ditto.
25560 (insn_v2muls): Ditto.
25562 2013-03-27 Walter Lee <walt@tilera.com>
25564 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
25565 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
25567 2013-03-27 Walter Lee <walt@tilera.com>
25569 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
25570 (*sibcall_value): Ditto.
25572 2013-03-27 Walter Lee <walt@tilera.com>
25574 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
25575 (insn_mnz_v8qi): ... this ...
25576 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
25578 (insn_v<n>mnz): Replaced by ...
25579 (insn_v1mnz): ... this ...
25580 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
25582 (insn_mz_<mode>): Replaced by ...
25583 (insn_mz_v8qi): ... this ...
25584 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
25586 (insn_v<n>mz): Replaced by ...
25587 (insn_v1mz): ... this ...
25588 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
25591 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
25593 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
25595 2013-03-26 Roland McGrath <mcgrathr@google.com>
25597 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
25598 than fprintf with a non-constant, non-format string.
25600 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25602 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
25603 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
25604 operand 0 predicate.
25605 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
25606 attribute. Use general_x64nomem_operand as operand 1 predicate.
25607 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
25608 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25609 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
25610 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25611 (mov<mode>_insv_1): Remove expander. Merge insn with
25612 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
25613 Use general_x64nomem_operand as operand 1 predicate.
25614 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
25615 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
25616 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25617 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
25618 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
25619 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
25620 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25621 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
25622 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
25623 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
25624 (general_x64nomem_operand): Ditto.
25626 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
25628 * config/rtems.opt: Add -pthread option.
25630 2013-03-26 Richard Biener <rguenther@suse.de>
25632 * alias.c (find_base_term): Avoid redundant and not used recursion.
25633 (base_alias_check): Get the initial base term from the caller.
25634 (true_dependence_1): Compute and pass base terms to base_alias_check.
25635 (write_dependence_p): Likewise.
25636 (may_alias_p): Likewise.
25638 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
25640 * config/aarch64/aarch64.c (aarch64_classify_address): Support
25641 PC-relative load in SI modes and above only.
25643 2013-03-26 Xinyu Qi <xyqi@marvell.com>
25645 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
25646 * config/arm/iwmmxt.md (WCGR0): Update.
25647 (WCGR1, WCGR2, WCGR3): Likewise.
25649 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25651 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
25652 Use x64 and nox64 isa attributes.
25654 2013-03-26 Richard Biener <rguenther@suse.de>
25656 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
25657 alignment computations and rely on get_object_alignment_1
25658 for the !TYPE_P case.
25659 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
25661 2013-03-26 Walter Lee <walt@tilera.com>
25663 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
25664 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
25666 2013-03-25 Jeff Law <law@redhat.com>
25668 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
25669 check for INTEGRAL_TYPE_P that was missing due to checking in
25670 wrong version of prior patch.
25672 2013-03-25 Walter Lee <walt@tilera.com>
25674 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
25675 TILEGX_INSN_SHUFFLEBYTES1.
25676 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
25678 (tilegx_builtins): Ditto.
25679 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
25681 2013-03-25 Walter Lee <walt@tilera.com>
25683 * config/tilegx/tilegx.md (floatsisf2): New pattern.
25684 (floatunssisf2): New pattern.
25685 (floatsidf2): New pattern.
25686 (floatunssidf2): New pattern.
25688 2013-03-25 Walter Lee <walt@tilera.com>
25690 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
25691 tests for constraint J, K, N, P.
25693 2013-03-25 Walter Lee <walt@tilera.com>
25695 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
25696 Use indirect/pcrel encoding.
25697 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
25700 2013-03-25 Steve Ellcey <sellcey@mips.com>
25702 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
25703 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
25704 * config/mips/mips.c (mips_option_override): Set IMADD default.
25705 * config/mips/mips.h (PTF_AVOID_IMADD): New.
25706 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
25707 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
25708 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
25709 * doc/invoke.texi (-mimadd/-mno-imadd): New.
25711 2013-03-25 Jeff Law <law@redhat.com>
25713 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
25714 slightly to avoid creating and folding useless trees. Simplify
25715 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
25717 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
25719 * config/i386/i386.md (*zero_extendsidi2): Merge with
25720 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
25721 * config/i386/predicates.md (x86_64_zext_operand): Rename from
25722 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
25723 targets. Clarify comment.
25725 2013-03-25 Martin Jambor <mjambor@suse.cz>
25727 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
25728 pass-through jump functions differently.
25729 (ipa_read_jump_function): Likewise. Also use setter functions to set
25732 2013-03-25 Martin Jambor <mjambor@suse.cz>
25734 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25735 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25737 (ipa_get_indirect_edge_target): New function.
25738 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25739 ipa_get_indirect_edge_target. Update all callers.
25740 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25741 ipa_get_indirect_edge_target_1 instead of calling
25742 ipa_get_indirect_edge_target.
25743 (create_specialized_node): Pass aggvlas to
25744 ipcp_discover_new_direct_edges.
25746 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25748 * config/arm/arm.md (f_sels, f_seld): New types.
25749 (*cmov<mode>): New pattern.
25750 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25753 2013-03-25 Kai Tietz <ktietz@redhat.com>
25755 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25756 POSIX-printf for mingw-hosted builds.
25758 2013-03-25 Richard Biener <rguenther@suse.de>
25760 PR middle-end/56694
25761 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25762 must-not-throw stmt location.
25764 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25766 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25767 Emit load-acquire versions when acq is true.
25768 (arm_emit_store_exclusive): Add rel parameter.
25769 Emit store-release versions when rel is true.
25770 (arm_split_compare_and_swap): Use acquire-release instructions
25772 of barriers when appropriate.
25773 (arm_split_atomic_op): Likewise.
25774 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25775 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25776 (VUNSPEC_SLX): Likewise.
25777 (VUNSPEC_LDA): Likewise.
25778 (VUNSPEC_STL): Likewise.
25779 * config/arm/sync.md (atomic_load<mode>): New pattern.
25780 (atomic_store<mode>): Likewise.
25781 (arm_load_acquire_exclusive<mode>): Likewise.
25782 (arm_load_acquire_exclusivesi): Likewise.
25783 (arm_load_acquire_exclusivedi): Likewise.
25784 (arm_store_release_exclusive<mode>): Likewise.
25786 2013-03-25 Catherine Moore <clm@codesourcery.com>
25788 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25789 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25790 * config/mip/predicates.md (lwsp_swsp_operand,
25791 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25792 sb16_operand, db4_operand, db7_operand, ib3_operand,
25793 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25794 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25795 andi16_operand): New predicates.
25796 * config/mips/mips.md (compression): New attribute.
25797 (enabled): New attribute.
25798 (length): Consider compression in computing length.
25799 (shift_compression): New code attribute.
25800 (*add<mode>3): New operands. Record compression.
25801 (sub<mode>3): Likewise.
25802 (one_cmpl<mode>2): Likewise.
25803 (*and<mode>3): Likewise.
25804 (*ior<mode>3): Likewise.
25805 (unnamed pattern for xor): Likewise.
25806 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25807 (*<optab><mode>3): Likewise.
25808 (*mov<mode>_internal: Likewise.
25809 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25810 (mips_unsigned_immediate_p): New.
25811 (umips_lwsp_swsp_address_p): New.
25812 (m16_based_address_p): New.
25813 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25814 (mips_unsigned_immediate_p): New prototype.
25815 (lwsp_swsp_address_p): New prototype.
25816 (m16_based_address_p): New prototype.
25817 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25818 (mips_signed_immediate_p): New function.
25819 (m16_based_address_p): New function.
25820 (lwsp_swsp_address_p): New function.
25821 (mips_print_operand_punctuation): Recognize short delay slot insns
25822 for microMIPS.add<mode>3"
25824 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25827 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25828 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25830 2013-03-25 Richard Biener <rguenther@suse.de>
25832 PR tree-optimization/56689
25833 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25836 2013-03-25 Richard Biener <rguenther@suse.de>
25838 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25840 (memory_references): Likewise.
25841 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25842 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25843 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25844 (gather_mem_refs_in_loops): Fold into ...
25845 (analyze_memory_references): ... this. Move initialization
25846 to tree_ssa_lim_initialize.
25847 (fill_always_executed_in): Rename to ...
25848 (fill_always_executed_in_1): ... this.
25849 (fill_always_executed_in): Move contains_call computation to
25850 this new function from ...
25851 (tree_ssa_lim_initialize): ... here.
25852 (tree_ssa_lim): Call fill_always_executed_in.
25854 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25856 * postreload.c (reload_combine): Fix code detecting returns.
25858 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25860 * function.c (emit_use_return_register_into_block): On cc0 targets,
25861 do not emit the sequence between cc0 setter and user.
25863 2013-03-25 Kai Tietz <ktietz@redhat.com>
25865 * config/i386/predicates.md (local_symbolic_operand): Interpret
25866 dll-imported symbols as none-local.
25868 2013-03-25 Richard Biener <rguenther@suse.de>
25870 * tree-ssa-loop-im.c (struct depend): Remove.
25871 (struct lim_aux_data): Make depends a vec of gimples.
25872 (free_lim_aux_data): Adjust.
25873 (add_dependency): Likewise.
25874 (set_level): Likewise.
25876 2013-03-25 Richard Biener <rguenther@suse.de>
25878 PR middle-end/56434
25879 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25880 the pointer returned by calls with ECF_MALLOC set.
25882 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25884 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25886 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25888 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25889 using MMXMODE mode iterator.
25890 (*move<mode>_internal): Merge with *movv2sf_internal and
25891 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25893 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25895 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25896 (record_last_mem_set_info): Likewise.
25898 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25899 of XNEWVEC followed by memset.
25900 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25902 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25904 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25905 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25906 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25907 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25908 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25909 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25910 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25911 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25912 BARRIER_P instead of GET_CODE.
25914 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25916 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25917 inaccuracy in the probing code.
25919 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25920 (ctrapdi4): Likewise.
25922 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25924 * calls.c (expand_call): Add missing guard to code handling return
25925 of non-BLKmode structures in MSB.
25926 * function.c (expand_function_end): Likewise.
25928 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25930 * combine.c (try_combine): Adjust comment. Do not add the set of
25931 insn #0 if the destination indirectly is set or dies in insn #2.
25932 Tidy up code to distribute a new note.
25934 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25936 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25937 also for alternatives 16 and 17.
25939 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25941 * config/i386/sse.md (*mov<mode>_internal): Merge with
25942 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25943 Emit insn template depending on type attribute. Use
25944 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25945 movd instead of movq mnemonic for interunit moves. Rewrite mode
25946 attribute calculation. Remove unit attribute calculation.
25947 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25948 Set prefix_data16 attribute for DImode ssemov types.
25949 Use Ym instead of y for SSE-MMX conversion alternatives.
25950 Reorder operand constraints.
25952 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25954 * df.h (df_insn_delete): Adjust prototype.
25955 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25956 and let it decide whether mark the basic block dirty.
25957 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25958 * df-scan.c (df_insn_info_delete): New helper function, split
25959 off from df_insn_delete.
25960 (df_scan_free_bb_info): Use it.
25961 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25963 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25964 that the insn is actually an insn and it has a non-NULL basic block.
25965 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25967 2013-03-22 Richard Biener <rguenther@suse.de>
25969 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25971 (mem_ref_alloc): Do not allocate them.
25972 (refs_independent_p): Do not query or maintain a cache.
25974 2013-03-22 Richard Biener <rguenther@suse.de>
25976 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25977 (gather_mem_refs_in_loops): Do not compute it.
25978 (analyze_memory_references): Do not allocate it.
25979 (tree_ssa_lim_finalize): Do not free it.
25980 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25982 2013-03-22 Richard Biener <rguenther@suse.de>
25984 * is-a.h (as_a): Use gcc_checking_assert.
25986 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25988 * config/aarch64/aarch64.c (aarch64_print_operand): New
25989 format specifier for printing a constant in hex.
25990 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25991 format specifier for printing second operand.
25993 2013-03-22 Richard Biener <rguenther@suse.de>
25995 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25997 (gather_mem_refs_in_loops): Perform store accumulation here.
25998 (create_vop_ref_mapping_loop): Remove.
25999 (create_vop_ref_mapping): Likewise.
26000 (analyze_memory_references): Initialize refs_stored_in_loop.
26001 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
26002 (record_indep_loop): Remove.
26003 (record_dep_loop): New function.
26004 (ref_indep_loop_p_1): Adjust to only walk over references
26005 in the loop, not its subloops.
26006 (ref_indep_loop_p): Rename to ...
26007 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
26008 maintaining a more fine-grained cache.
26009 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
26010 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
26012 2013-03-22 Richard Biener <rguenther@suse.de>
26014 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
26015 (struct mem_ref): Make accesses_in_loop a vec of a vec of
26016 aggregate mem_ref_loc.
26017 (free_mem_ref_locs): Inline into ...
26018 (memref_free): ... this and adjust.
26019 (mem_ref_alloc): Adjust.
26020 (mem_ref_locs_alloc): Remove.
26021 (record_mem_ref_loc): Adjust.
26022 (get_all_locs_in_loop): Rewrite into ...
26023 (for_all_locs_in_loop): ... this iterator.
26024 (rewrite_mem_ref_loc): New functor.
26025 (rewrite_mem_refs): Use for_all_locs_in_loop.
26026 (sm_set_flag_if_changed): New functor.
26027 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
26028 (ref_always_accessed): New functor.
26029 (ref_always_accessed_p): Use for_all_locs_in_loop.
26031 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
26033 * tree-pass.h (PROP_gimple_lvec): New.
26034 * passes.c (dump_properties): Handle PROP_gimple_lvec.
26035 (init_optimization_passes): Move pass_lower_vector.
26036 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
26038 (pass_lower_vector): Provide PROP_gimple_lvec.
26039 (pass_lower_vector_ssa): Likewise.
26040 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
26042 2013-03-21 Mark Wielaard <mjw@redhat.com>
26044 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
26046 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
26048 * config/i386/i386.md (*movdi_internal): Disparage slightly
26049 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
26050 conversion alternatives.
26052 2013-03-21 Jakub Jelinek <jakub@redhat.com>
26054 PR middle-end/48087
26055 * diagnostic.def (DK_WERROR): New kind.
26056 * diagnostic.h (werrorcount): Define.
26057 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
26058 promoted to DK_ERROR, increment DK_WERROR counter instead of
26060 * toplev.c (toplev_main): Call print_ignored_options even if
26061 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
26062 even if just werrorcount is non-zero.
26065 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
26067 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
26068 (string_cst_pool_decl): New function.
26069 (optimize_one_addr_into_implicit_ptr): New function.
26070 (resolve_addr_in_expr): Optimize DWARF location expression
26071 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
26072 which doesn't live in memory, but has DW_AT_location or
26073 DW_AT_const_value, or refers to a string literal, into
26074 DW_OP_GNU_implicit_pointer.
26075 (optimize_location_into_implicit_ptr): New function.
26076 (resolve_addr): If removing DW_AT_location of a variable because
26077 it was DW_OP_addr of address of the variable, but the variable doesn't
26078 live in memory, try to emit const value attribute for the initializer.
26080 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
26082 * tree.h (VECTOR_TYPE_P): New macro.
26083 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
26084 TYPE_MODE): Use it.
26085 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
26086 VEC_COND_EXPR cannot be lvalues.
26087 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
26089 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
26091 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
26092 Restrict the transformation to equal modes.
26094 2013-03-21 Richard Biener <rguenther@suse.de>
26096 PR tree-optimization/39326
26097 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
26098 (MEM_ANALYZABLE): Adjust.
26099 (record_mem_ref_loc): Move bitmap ops ...
26100 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
26101 unanalyzable refs, do not record locations for it.
26102 (analyze_memory_references): Allocate ref zero as shared
26104 (refs_independent_p): Do not test for unanalyzed mems here.
26105 (ref_indep_loop_p_1): Special-case disambiguation against
26106 the unanalyzed ref.
26107 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
26109 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
26111 * config/arm/arm-protos.h (tune_params): Add
26112 prefer_neon_for_64bits field.
26113 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
26114 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
26115 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
26116 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
26117 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
26118 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
26119 (arm_option_override): Handle -mneon-for-64bits new option.
26120 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
26121 (prefer_neon_for_64bits): Declare new variable.
26122 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
26123 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
26124 (arch_enabled): Handle new arch types. Remove support for onlya8
26126 (one_cmpldi2): Use new arch names.
26127 (zero_extend<mode>di2, extend<mode>di2): Ditto.
26128 * config/arm/arm.opt (mneon-for-64bits): Add option.
26129 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
26130 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
26131 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
26133 * doc/invoke.texi (-mneon-for-64bits): Document.
26135 2013-03-21 Richard Biener <rguenther@suse.de>
26137 PR tree-optimization/39326
26138 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
26139 (sort_bbs_in_loop_postorder_cmp): New function.
26140 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
26142 2013-03-21 Richard Biener <rguenther@suse.de>
26144 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
26145 (vect_insert_into_interleaving_chain): Likewise.
26146 (vect_drs_dependent_in_basic_block): Inline ...
26147 (vect_slp_analyze_data_ref_dependence): ... here. New function,
26149 (vect_analyze_data_ref_dependence): ... here. Simplify.
26150 (vect_check_interleaving): Simplify.
26151 (vect_analyze_data_ref_dependences): Likewise. Split out ...
26152 (vect_slp_analyze_data_ref_dependences): ... this new function.
26153 (dr_group_sort_cmp): New function.
26154 (vect_analyze_data_ref_accesses): Compute data-reference groups
26155 here instead of in vect_analyze_data_ref_dependence. Use
26156 a more efficient algorithm.
26157 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
26158 vect_slp_analyze_data_ref_dependences. Call
26159 vect_analyze_data_ref_accesses earlier.
26160 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
26161 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
26162 (vect_slp_analyze_data_ref_dependences): New prototype.
26164 2013-03-21 Richard Biener <rguenther@suse.de>
26166 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
26167 ref is stored in the loop.
26168 (find_refs_for_sm): Walk only over all stores.
26169 (store_motion_loop): Allocate from lim_bitmap_obstack.
26170 (store_motion): Likewise.
26172 2013-03-21 Richard Biener <rguenther@suse.de>
26174 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
26175 Update virtual SSA form.
26177 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26179 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
26180 * configure: Regenerate.
26181 * config.in: Regenerate.
26182 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
26183 if !HAVE_LD_EH_FRAME_CIEV3.
26185 2013-03-21 Richard Biener <rguenther@suse.de>
26187 * tree-cfg.c (verify_expr_no_block): New function.
26188 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
26189 nor DECL_VALUE_EXPR have locations with associated blocks.
26190 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
26191 (clear_unused_block_pointer): Remove code dealing with
26192 blocks in DECL_DEBUG_EXPR locations.
26194 2013-03-21 Richard Biener <rguenther@suse.de>
26196 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
26197 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
26198 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
26199 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
26200 instead of DECL_DEBUG_EXPR_IS_FROM.
26201 * gimplify.c (gimplify_modify_expr): Likewise.
26202 * tree-cfg.c (verify_expr_location_1): Likewise.
26203 * tree-complex.c (create_one_component_var): Likewise.
26204 * tree-sra.c (create_access_replacement): Likewise.
26205 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
26206 (clear_unused_block_pointer): Likewise.
26207 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
26208 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
26209 * var-tracking.c (var_debug_decl): Likewise.
26210 (track_expr_p): Likewise.
26211 * tree-inline.c (add_local_variables): Likewise. Set
26212 DECL_HAS_DEBUG_EXPR_P after copying it.
26213 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
26214 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
26216 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
26219 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
26220 * configure: Regenerate.
26221 * config.in: Regenerate.
26222 * config/i386/i386.md (*movdf_internal): Use
26223 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
26224 movd instead of movq mnemonic for interunit moves.
26225 (*movdi_internal): Ditto.
26227 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26229 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
26230 (abd<mode>_3): New pattern.
26231 (aba<mode>_3): New pattern.
26232 (fabd<mode>_3): New pattern.
26234 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26236 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
26237 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
26238 occurrence of REGISTER_PREFIX as its empty string.
26240 2013-03-20 Jeff Law <law@redhat.com>
26242 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
26243 addititional equivalences for equality comparisons between an SSA_NAME
26244 and a constant where the SSA_NAME was set from a widening conversion.
26246 2013-03-20 Walter Lee <walt@tilera.com>
26248 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
26250 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26252 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
26253 depending on type attribute.
26254 (*movti_internal): Ditto.
26255 (*movtf_internal): Ditto.
26256 (*movxf_internal): Ditto.
26257 (*movdf_internal): Ditto.
26258 (*movsf_internal): Ditto.
26260 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26262 * config/i386/i386.md (*movti_internal): Set prefix attribute to
26263 maybe_vex for sselog1 and ssemov types.
26264 (*movdi_internal): Reorder operand constraints.
26265 (*movsi_internal): Ditto. Set prefix attribute to
26266 maybe_vex for sselog1 and ssemov types.
26267 (*movtf_internal): Set prefix attribute to maybe_vex
26268 for sselog1 and ssemov types.
26269 (*movdf_internal): Ditto. Set prefix_data16 attribute for
26270 DImode ssemov types. Reorder operand constraints.
26271 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
26272 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
26273 attribute for SImode ssemov types. Reorder operand constraints.
26275 2013-03-20 Martin Jambor <mjambor@suse.cz>
26277 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
26278 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
26280 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
26282 * config/rs6000/predicates.md (indexed_address, update_address_mem
26283 update_indexed_address_mem): New predicates.
26284 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
26285 attribute for load/store instructions.
26286 * config/rs6000/dfp.md (movsd_store): Likewise.
26287 (movsd_load): Likewise.
26288 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
26289 (unnamed HI->DI extend define_insn): Likewise.
26290 (unnamed SI->DI extend define_insn): Likewise.
26291 (unnamed QI->SI extend define_insn): Likewise.
26292 (unnamed QI->HI extend define_insn): Likewise.
26293 (unnamed HI->SI extend define_insn): Likewise.
26294 (unnamed HI->SI extend define_insn): Likewise.
26295 (extendsfdf2_fpr): Likewise.
26296 (movsi_internal1): Likewise.
26297 (movsi_internal1_single): Likewise.
26298 (movhi_internal): Likewise.
26299 (movqi_internal): Likewise.
26300 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
26301 attribute for load/store instructions.
26302 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
26304 (mov<mode>_softfloat): Likewise.
26305 (mov<mode>_hardfloat32): Likewise.
26306 (mov<mode>_hardfloat64): Likewise.
26307 (mov<mode>_softfloat64): Likewise.
26308 (movdi_internal32): Likewise.
26309 (movdi_internal64): Likewise.
26310 (probe_stack_<mode>): Likewise.
26312 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
26314 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
26315 floating point, and decimal floating point to reload iterator.
26317 * config/rs6000/constraints.md (wl constraint): New constraints to
26318 return FLOAT_REGS if certain options are used to reduce the number
26319 of separate patterns that exist in the file.
26320 (wx constraint): Likewise.
26321 (wz constraint): Likewise.
26323 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26324 -mdebug=reg, print wg, wl, wx, and wz constraints.
26325 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26326 Initialize the reload functions for 64-bit binary/decimal floating
26328 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26329 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26330 create the buffer on the stack to overcome not having a 32-bit
26332 (rs6000_emit_move): Likewise.
26333 (rs6000_secondary_memory_needed_rtx): Likewise.
26334 (rs6000_alloc_sdmode_stack_slot): Likewise.
26335 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26336 via xxlxor, just like DFmode 0.0.
26338 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
26339 define as 1 if we are running on a power7 or newer.
26340 (enum r6000_reg_class_enum): Add new constraints.
26342 * config/rs6000/dfp.md (movsd): Delete, combine with binary
26343 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26344 with other moves by using conditional constraits (wg). Use LFIWZX
26345 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
26346 (movsd splitter): Likewise.
26347 (movsd_hardfloat): Likewise.
26348 (movsd_softfloat): Likewise.
26350 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
26351 binary and decimal floating point moves.
26352 (fmove_ok): New attributes to combine binary and decimal floating
26353 point moves, and to combine power6x (mfpgpr) moves along normal
26355 (real_value_to_target): Likewise.
26356 (f32_lr): Likewise.
26357 (f32_lm): Likewise.
26358 (f32_li): Likewise.
26359 (f32_sr): Likewise.
26360 (f32_sm): Likewise.
26361 (f32_si): Likewise.
26362 (movsf): Combine binary and decimal floating point moves. Combine
26363 power6x (mfpgpr) moves with other moves by using conditional
26364 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
26365 (mov<mode> for SFmode/SDmode); Likewise.
26366 (SFmode/SDmode splitters): Likewise.
26367 (movsf_hardfloat): Likewise.
26368 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
26369 (movsf_softfloat): Likewise.
26370 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
26372 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
26373 wx and wz constraints.
26375 * config/rs6000/constraints.md (wg constraint): New constraint to
26376 return FLOAT_REGS if -mmfpgpr (power6x) was used.
26378 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
26381 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26382 -mdebug=reg, print wg, wl, wx, and wz constraints.
26383 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26384 Initialize the reload functions for 64-bit binary/decimal floating
26386 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26387 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26388 create the buffer on the stack to overcome not having a 32-bit
26390 (rs6000_emit_move): Likewise.
26391 (rs6000_secondary_memory_needed_rtx): Likewise.
26392 (rs6000_alloc_sdmode_stack_slot): Likewise.
26393 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26394 via xxlxor, just like DFmode 0.0.
26396 * config/rs6000/dfp.md (movdd): Delete, combine with binary
26397 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26398 with other moves by using conditional constraits (wg). Use LFIWZX
26399 and STFIWX for loading SDmode on power7.
26400 (movdd splitters): Likewise.
26401 (movdd_hardfloat32): Likewise.
26402 (movdd_softfloat32): Likewise.
26403 (movdd_hardfloat64_mfpgpr): Likewise.
26404 (movdd_hardfloat64): Likewise.
26405 (movdd_softfloat64): Likewise.
26407 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
26408 64-bit binary and decimal floating point moves.
26409 (FMOVE64X): Likewise.
26410 (movdf): Combine 64-bit binary and decimal floating point moves.
26411 Combine power6x (mfpgpr) moves with other moves by using
26412 conditional constraits (wg).
26413 (mov<mode> for DFmode/DDmode): Likewise.
26414 (DFmode/DDmode splitters): Likewise.
26415 (movdf_hardfloat32): Likewise.
26416 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
26417 (movdf_softfloat32): Likewise.
26418 (movdf_hardfloat64_mfpgpr): Likewise.
26419 (movdf_hardfloat64): Likewise.
26420 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
26421 (movdf_softfloat64): Likewise.
26422 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
26423 (reload_<mode>_load): Move to later in the file so they aren't in
26424 the middle of the floating point move insns.
26425 (reload_<mode>_store): Likewise.
26427 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
26430 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
26431 constraint if -mdebug=reg.
26432 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
26433 Enable using dd reload support if needed.
26435 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
26436 binary and decimal floating point moves in rs6000.md.
26437 (movtd_internal): Likewise.
26439 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
26440 decimal floating point moves.
26442 (movtf_internal): Likewise.
26443 (mov<mode>_internal, TDmode/TFmode): Likewise.
26444 (movtf_softfloat): Likewise.
26445 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
26447 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
26448 movdi_internal64, using wg constraint for move direct operations.
26449 (movdi_internal64): Likewise.
26451 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
26452 MODES_TIEABLE_P for selected modes. Print the numerical value of
26453 the various virtual registers. Use GPR/FPR first/last values,
26454 instead of hard coding the register numbers. Print which modes
26455 have reload functions registered.
26456 (rs6000_option_override_internal): If -mdebug=reg, trace the options
26457 settings before/after setting cpu, target and subtarget settings.
26458 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
26459 and for secondary reload failures in rs6000_secondary_reload_inner.
26460 (rs6000_secondary_reload_fail): Likewise.
26461 (rs6000_secondary_reload_inner): Likewise.
26463 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
26464 macros for first/last GPR and FPR registers.
26465 (LAST_GPR_REGNO): Likewise.
26466 (FIRST_FPR_REGNO): Likewise.
26467 (LAST_FPR_REGNO): Likewise.
26469 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
26470 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
26471 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
26472 (vcond<mode><mode>): Likewise.
26473 (vcondu<mode><mode>): Likewise.
26474 (vector_gtu<mode>): Likewise.
26475 (vector_gte<mode>): Likewise.
26476 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
26477 to prevent the compiler from converting DImode operations to TImode.
26478 (ior<mode>3): Likewise.
26479 (and<mode>3): Likewise.
26480 (one_cmpl<mode>2): Likewise.
26481 (nor<mode>3): Likewise.
26482 (andc<mode>3): Likewise.
26484 * config/rs6000/constraints.md (wt constraint): New constraint
26485 that returns VSX_REGS if TImode is allowed in VSX registers.
26487 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
26488 constant under VSX.
26490 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
26491 similar to TImode, but it is restricted to being in the GPRs.
26493 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
26494 TImode to occupy a single VSX register.
26496 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
26497 -mvsx-timode for power7/power8.
26498 (power7 cpu): Likewise.
26499 (power8 cpu): Likewise.
26501 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
26502 sure that TFmode/TDmode take up two registers if they are ever
26503 allowed in the upper VSX registers.
26504 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
26506 (rs6000_init_hard_regno_mode_ok): Likewise.
26507 (rs6000_debug_reg_global): Add debugging for PTImode and wt
26508 constraint. Print if LRA is turned on.
26509 (rs6000_option_override_internal): Give an error if -mvsx-timode
26510 and VSX is not enabled.
26511 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
26512 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
26513 to reg+offset addressing. Use PTImode when checking offset
26514 addresses for validity.
26515 (reg_offset_addressing_ok_p): Likewise.
26516 (rs6000_legitimate_offset_address_p): Likewise.
26517 (rs6000_legitimize_address): Likewise.
26518 (rs6000_legitimize_reload_address): Likewise.
26519 (rs6000_legitimate_address_p): Likewise.
26520 (rs6000_eliminate_indexed_memrefs): Likewise.
26521 (rs6000_emit_move): Likewise.
26522 (rs6000_secondary_reload): Likewise.
26523 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
26524 reloads to fpr registers to continue to use reg+offset addressing,
26525 but 64-bit reloads to altivec registers need reg+reg addressing.
26526 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
26527 it. Treat LO_SUM like a PLUS operation.
26528 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
26529 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
26530 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
26531 registers to share a register with a smaller sized type, since VSX
26532 puts scalars in the upper 64-bits.
26533 (print_operand): Add support for PTImode.
26534 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
26535 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
26536 registers, but don't have arithmetic support.
26537 (rs6000_memory_move_cost): Add test for VSX.
26538 (rs6000_opt_masks): Add -mvsx-timode.
26540 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
26543 (VSr): Use wt constraint for TImode.
26544 (VSv): Drop TImode support.
26545 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
26546 (vsx_movti_64bit): Likewise.
26547 (vsx_movti_32bit): Likewise.
26548 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
26549 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
26550 one '?' on the appropriate output constraint. Do not allow TImode
26551 logical operations on 32-bit systems.
26552 (vsx_ior<mode>3): Likewise.
26553 (vsx_xor<mode>3): Likewise.
26554 (vsx_one_cmpl<mode>2): Likewise.
26555 (vsx_nor<mode>3): Likewise.
26556 (vsx_andc<mode>3): Likewise.
26557 (vsx_concat_<mode>): Likewise.
26558 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
26560 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
26561 OPTION_MASK_VSX_TIMODE.
26562 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
26563 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
26565 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
26566 (TI2 iterator): New iterator for TImode, PTImode.
26567 (wd mode attribute): Add values for vector types.
26568 (movti_string): Replace TI move operations with operations for TImode
26569 and PTImode. Add support for TImode being allowed in VSX registers.
26570 (mov<mode>_string, TImode/PTImode): Likewise.
26571 (movti_ppc64): Likewise.
26572 (mov<mode>_ppc64, TImode/PTImode): Likewise.
26573 (TI mode splitters): Likewise.
26575 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
26578 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
26580 PR tree-optimization/56355
26581 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
26582 Also handle integers with undefined overflow.
26584 2013-03-20 Catherine Moore <clm@codesourcery.com>
26585 Maciej W. Rozycki <macro@codesourcery.com>
26586 Tom de Vries <tom@codesourcery.com>
26587 Nathan Sidwell <nathan@codesourcery.com>
26588 Iain Sandoe <iain@codesourcery.com>
26589 Nathan Froyd <froydnj@codesourcery.com>
26590 Chao-ying Fu <fu@mips.com>
26592 * doc/extend.texi: (micromips, nomicromips, nocompression):
26593 Document new function attributes.
26594 * doc/invoke.texi (minterlink-compressed, mmicromips,
26595 m14k, m14ke, m14kec): Document new options.
26596 (minterlink-mips16): Update documentation.
26597 * doc/md.texi (ZC, ZD): Document new constraints.
26598 * configure.ac (gcc_cv_as_micromips): Check if linker
26599 supports the .set micromips directive.
26600 * configure: Regenerate.
26601 * config.in: Regenerate.
26602 * config/mips/mips-tables.opt: Regenerate.
26603 * config/mips/micromips.md: New file.
26604 * constraints.md (ZC, ZD): New constraints.
26605 * config/mips/predicates.md (movep_src_register): New predicate.
26606 (movep_src_operand): New predicate.
26607 (non_volatile_mem_operand): New predicate.
26608 * config/mips/mips.md (multimem): New type.
26609 (length): Differentiate between 17-bit and 18-bit branch offsets.
26610 (MOVEP1, MOVEP2): New mode iterator.
26611 (mov_<load>l): Use ZC constraint.
26612 (mov_<load>r): Likewise.
26613 (mov_<store>l): Likewise.
26614 (mov_<store>r): Likewise.
26615 (*branch_equality<mode>_inverted): Add microMIPS support.
26616 (*branch_equality<mode>): Likewise.
26617 (*jump_absolute): Likewise.
26618 (indirect_jump_<mode>): Likewise.
26619 (tablejump_<mode>): Likewise.
26620 (<optab>_internal): Likewise.
26621 (sibcall_internal): Likewise.
26622 (sibcall_value_internal): Likewise.
26623 (prefetch): Use constraint ZD.
26624 * config/mips/mips.opt (minterlink-compressed): New option.
26625 (minterlink-mips16): Now an alias for minterlink-compressed.
26626 (mmicromips): New option.
26627 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26628 (compare_and_swap_12): Likewise.
26629 (sync_add<mode>): Likewise.
26630 (sync_<optab>_12): Likewise.
26631 (sync_old_<optab>_12): Likewise.
26632 (sync_new_<optab>_12): Likewise.
26633 (sync_nand_12): Likewise.
26634 (sync_old_nand_12): Likewise.
26635 (sync_new_nand_12): Likewise.
26636 (sync_sub<mode>): Likewise.
26637 (sync_old_add<mode>): Likewise.
26638 (sync_old_sub<mode>): Likewise.
26639 (sync_new_add<mode>): Likewise.
26640 (sync_new_sub<mode>): Likewise.
26641 (sync_<optab><mode>): Likewise.
26642 (sync_old_<optab><mode>): Likewise.
26643 (sync_new_<optab><mode>): Likewise.
26644 (sync_nand<mode>): Likewise.
26645 (sync_old_nand<mode>): Likewise.
26646 (sync_new_nand<mode>): Likewise.
26647 (sync_lock_test_and_set<mode>): Likewise.
26648 (test_and_set_12): Likewise.
26649 (atomic_compare_and_swap<mode>): Likewise.
26650 (atomic_exchange<mode>_llsc): Likewise.
26651 (atomic_fetch_add<mode>_llsc): Likewise.
26652 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26653 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26654 (umips_save_restore_pattern_p): Likewise.
26655 (umips_load_store_pair_p): Likewise.
26656 (umips_output_load_store_pair): Likewise.
26657 (umips_movep_target_p): Likewise.
26658 (umips_12bit_offset_address_p): Likewise.
26659 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26660 (mips_base_mips16): Rename this...
26661 (mips_base_compression_flags): ...to this. Update all uses.
26662 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26663 (mips_mips16_decl_p): Delete.
26664 (mips_nomips16_decl_p): Delete.
26665 (mips_get_compress_on_flags): New function.
26666 (mips_get_compress_off_flags): New function.
26667 (mips_get_compress_mode): New function.
26668 (mips_get_compress_on_name): New function.
26669 (mips_get_compress_off_name): New function.
26670 (mips_insert_attributes): Support multiple compression types.
26671 (mips_merge_decl_attributes): Likewise.
26672 (umips_12bit_offset_address_p): New function.
26673 (mips_start_function_definition): Emit .set micromips directive.
26674 (mips_call_may_need_jalx_p): New function.
26675 (mips_function_ok_for_sibcall): Add microMIPS support.
26676 (mips_print_operand_punctuation): Support short delay slots and
26678 (umips_swm_mask, umips_swm_encoding): New.
26679 (umips_build_save_restore): New function.
26680 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26681 (was_mips16_p): Remove.
26682 (old_compression_mode): New.
26683 (mips_set_compression_mode): New function.
26684 (mips_set_current_function): Add microMIPS support.
26685 (mips_option_override): Likewise.
26686 (umips_save_restore_pattern_p): New function.
26687 (umips_output_save_restore): New function.
26688 (umips_load_store_pair_p_1): New function.
26689 (umips_load_store_pair_p): New function.
26690 (umips_output_load_store_pair_1): New function.
26691 (umips_output_load_store_pair): New function.
26692 (umips_movep_target_p) New function.
26693 (mips_prepare_pch_save): Add microMIPS support.
26694 * config/mips/mips.h (TARGET_COMPRESSION): New.
26695 (TARGET_CPU_CPP_BUILTINS): Update macro
26696 to use new compression flags and to support microMIPS.
26697 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26698 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26699 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26700 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26701 (ASM_SPEC): Support mmicromips and mno-micromips.
26702 (M16STORE_REG_P): New macro.
26703 (MIPS_CALL): Support TARGET_MICROMIPS.
26704 (MICROMIPS_J): New macro.
26705 (mips_base_mips16): Rename this...
26706 (mips_base_compression_flags): ...to this.
26707 (UMIPS_12BIT_OFFSET_P): New macro.
26708 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26709 (MULTILIB_DIRNAMES): Likewise.
26710 2013-03-20 Richard Biener <rguenther@suse.de>
26712 PR tree-optimization/56661
26713 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
26714 the result does not have to be distinct.
26716 2013-03-20 Richard Biener <rguenther@suse.de>
26718 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
26721 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26722 Steven Bosscher <steven@gcc.gnu.org>
26724 PR rtl-optimization/56605
26725 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
26727 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26730 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26731 that require movd instead of movq.
26733 2013-03-20 Richard Biener <rguenther@suse.de>
26735 * tree-ssa-structalias.c (struct variable_info): Add pointer
26736 to the first field of an aggregate with sub-vars. Make
26737 this and the pointer to the next subfield its ID.
26738 (vi_next): New function.
26739 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26740 storedanything_id, integer_id): Increment by one.
26741 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26742 get_call_clobber_vi): Adjust.
26743 (solution_set_expand): Simplify and speedup.
26744 (solution_set_add): Inline into ...
26745 (set_union_with_increment): ... this. Adjust accordingly.
26746 (do_sd_constraint): Likewise.
26747 (do_ds_constraint): Likewise.
26748 (do_complex_constraint): Simplify.
26749 (build_pred_graph): Adjust.
26750 (solve_graph): Likewise. Simplify and speedup.
26751 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26752 get_constraint_for_component_ref, get_constraint_for_1,
26753 first_vi_for_offset, first_or_preceding_vi_for_offset,
26754 create_function_info_for, create_variable_info_for_1,
26755 create_variable_info_for, intra_create_variable_infos): Adjust.
26756 (init_base_vars): Push NULL for ID zero.
26757 (compute_points_to_sets): Adjust.
26759 2013-03-20 Richard Biener <rguenther@suse.de>
26761 * cfgloop.c (verify_loop_structure): Streamline and avoid
26762 ICEing on corrupt loop tree.
26763 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26766 2013-03-20 Richard Biener <rguenther@suse.de>
26768 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26769 check whether an SSA update is needed.
26771 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26773 * config/mips/constraints.md (T): Rename to...
26777 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26778 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26780 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26782 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26783 (*subsi3_carryin_uxtw): Likewise.
26785 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26787 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26788 (*rorsi3_insn_uxtw): Likewise.
26790 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26792 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26793 (*extrsi5_insn_uxtw): Likewise.
26795 2013-03-19 Richard Biener <rguenther@suse.de>
26797 PR tree-optimization/56273
26798 * passes.c (init_optimization_passes): Move second VRP after DOM.
26800 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26802 * config/i386/i386.md (*movti_internal): Merge from
26803 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26804 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26805 nox64 isa attributes.
26807 2013-03-18 Richard Biener <rguenther@suse.de>
26809 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26811 (merge_node_constraints): Likewise.
26812 (build_succ_graph): Likewise.
26813 (valid_graph_edge): Inline into single caller.
26814 (unify_nodes): Likewise. Use bitmap_set_bit return value
26816 (scc_visit): Fix formatting and variable use.
26817 (do_sd_constraint): Use gcc_checking_assert.
26818 (do_ds_constraint): Likewise.
26819 (do_complex_constraint): Likewise.
26820 (condense_visit): Likewise. Cleanup.
26821 (dump_pred_graph): New function.
26822 (perform_var_substitution): Dump the pred-graph before
26823 variable substitution.
26824 (find_equivalent_node): Use gcc_checking_assert.
26825 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26827 2013-03-18 Richard Biener <rguenther@suse.de>
26829 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26830 Remove cond_expr_stmt_list argument and do not gimplify the
26832 (vect_loop_versioning): Adjust.
26833 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26834 Cleanup to use less temporaries.
26835 (vect_create_data_ref_ptr): Cleanup.
26837 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26839 PR tree-optimization/56635
26840 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26841 require types_compatible_p types.
26843 2013-03-18 Nick Clifton <nickc@redhat.com>
26845 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26846 spurious backslash.
26848 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26849 Add missing line to comment describing function.
26851 2013-03-18 Richard Biener <rguenther@suse.de>
26853 PR tree-optimization/56210
26854 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26855 Handle string / character search functions.
26856 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26858 2013-03-18 Richard Biener <rguenther@suse.de>
26860 PR middle-end/56483
26861 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26862 and implement properly.
26863 * gimple.h (gimple_cond_single_var_p): Remove.
26865 2013-03-18 Richard Biener <rguenther@suse.de>
26867 * tree-data-ref.h (find_data_references_in_loop): Declare.
26868 * tree-data-ref.c (get_references_in_stmt): Use a stack
26869 vector pre-allocated in the callers.
26870 (find_data_references_in_stmt): Adjust.
26871 (graphite_find_data_references_in_stmt): Likewise.
26872 (create_rdg_vertices): Likewise.
26873 (find_data_references_in_loop): Export.
26874 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26875 Compute dependences here...
26876 (vect_analyze_data_refs): ...not here. When we encounter
26877 a non-vectorizable data reference in basic-block vectorization
26878 truncate the data reference vector. Do not bother to
26879 fixup data-dependence information for gather loads.
26880 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26881 of data references, as reported.
26883 2013-03-18 Richard Biener <rguenther@suse.de>
26885 PR tree-optimization/3713
26886 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26887 has_constants and expr.
26888 (stmt_has_constants): Properly valueize SSA names when deciding
26889 whether the stmt has constants.
26891 2013-03-18 Richard Biener <rguenther@suse.de>
26893 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26894 whole function when there is nothing to do.
26895 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26896 * tree-vectorizer.c (vectorize_loops): Update virtual and
26897 loop-closed SSA once.
26898 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26900 2013-03-18 Richard Biener <rguenther@suse.de>
26902 PR middle-end/56113
26903 * domwalk.c (bb_postorder): New global static.
26904 (cmp_bb_postorder): New function.
26905 (walk_dominator_tree): Replace scheme imposing an order for
26906 visiting dominator sons by one sorting them at the time they
26907 are pushed on the stack.
26909 2013-03-18 Richard Biener <rguenther@suse.de>
26911 PR tree-optimization/39326
26912 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26913 (struct mem_ref): Replace mem member with ao_ref typed member.
26914 (MEM_ANALYZABLE): Adjust.
26915 (memref_eq): Likewise.
26916 (mem_ref_alloc): Likewise.
26917 (gather_mem_refs_stmt): Likewise.
26918 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26919 (execute_sm_if_changed_flag_set): Adjust.
26920 (execute_sm): Likewise.
26921 (ref_always_accessed_p): Likewise.
26922 (refs_independent_p): Likewise.
26923 (can_sm_ref_p): Likewise.
26925 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26928 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26929 return 1 even for !unsignedp.
26931 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26933 * config/i386/i386.md (isa): Add x64 and nox64.
26934 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26935 (*pushtf): Enable *roF alternative for x64 isa only.
26936 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26937 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26938 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26939 (*movtf_internal): Merge from *movtf_internal_rex64 and
26940 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26941 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26942 nox64 isa attributes.
26943 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26944 nox64 isa attributes.
26945 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26947 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26949 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26951 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26954 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26955 class names. Remove trailing comma after "ALL_REGS".
26957 2013-03-16 Jan Hubicka <jh@suse.cz>
26959 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26960 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26961 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26962 of cgraph_get_create_node.
26963 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26965 2013-03-16 Jason Merrill <jason@redhat.com>
26968 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26969 with DW_AT_default_value.
26971 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26973 * BASE-VER: Set to 4.9.0.
26975 2013-03-14 Andi Kleen <ak@linux.intel.com>
26978 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26979 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26980 Document _x* TSX intrinsics.
26982 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26983 David Holsgrove <david.holsgrove@xilinx.com>
26985 * configure.ac: Add MicroBlaze TLS support detection.
26986 * configure: Regenerate.
26987 * config/microblaze/microblaze-protos.h
26988 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26989 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26990 * config/microblaze/microblaze.c (microblaze_address_type): Add
26991 ADDRESS_TLS and tls_reloc address types.
26992 (microblaze_address_info): Add tls_reloc.
26993 (TARGET_HAVE_TLS): Define.
26994 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26995 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26996 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26997 load_tls_operand, microblaze_call_tls_get_addr,
26998 microblaze_legitimize_tls_address): New functions.
26999 (microblaze_classify_unspec): Handle UNSPEC_TLS.
27000 (get_base_reg): Use microblaze_tls_symbol_p.
27001 (microblaze_classify_address): Handle TLS.
27002 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
27003 label_mentioned_p and microblaze_tls_referenced_p.
27004 (microblaze_legitimize_address): Handle TLS.
27005 (microblaze_address_insns): Handle ADDRESS_TLS.
27006 (pic_address_needs_scratch): Handle TLS.
27007 (print_operand_address): Handle TLS.
27008 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
27009 (microblaze_expand_move): Handle TLS.
27010 (microblaze_legitimate_constant_p): Check
27011 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
27012 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
27013 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
27014 (PIC_OFFSET_TABLE_REGNUM): Set.
27015 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
27016 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
27017 (addsi3, movsi_internal2, movdf_internal): Update constraints
27018 * config/microblaze/predicates.md (arith_plus_operand): Define
27019 (move_operand): Redefine as move_src_operand,
27020 check microblaze_tls_referenced_p.
27022 2013-03-14 Ian Bolton <ian.bolton@arm.com>
27024 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
27025 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
27027 2013-03-14 Ian Bolton <ian.bolton@arm.com>
27029 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
27032 2013-03-14 Jakub Jelinek <jakub@redhat.com>
27034 PR tree-optimization/53265
27035 * common.opt (Waggressive-loop-optimizations): New option.
27036 * tree-ssa-loop-niter.c: Include tree-pass.h.
27037 (do_warn_aggressive_loop_optimizations): New function.
27038 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
27039 if number_of_latch_executions returned constant.
27040 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
27041 early. If number_of_latch_executions returned constant, set
27042 nb_iterations_upper_bound back to it.
27043 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
27045 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
27046 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
27048 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
27049 (MULTILIB_OSDIRNAMES): Set.
27050 * genmultilib: If defaultosdirname doesn't start with :: , set
27051 defaultosdirname2 instead, clear it and emit two . multilib_raw
27052 entries instead of just one.
27054 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
27056 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
27057 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
27058 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
27059 (SUBTARGET_OVERRIDE_OPTIONS): New.
27061 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
27064 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
27065 (musermode): Convert to Var(TARGET_USERMODE).
27066 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
27067 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
27068 * config/sh/sh.c (sh_option_override): Use
27069 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
27070 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
27072 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
27074 (udivsi3_i4_single, divsi3_i4_single): Use
27075 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
27077 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
27079 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
27082 2013-03-13 Richard Biener <rguenther@suse.de>
27084 PR tree-optimization/56608
27085 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
27086 calls when vectorizing basic-blocks.
27088 2013-03-13 Jakub Jelinek <jakub@redhat.com>
27091 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
27094 2013-03-12 Jakub Jelinek <jakub@redhat.com>
27096 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
27098 2013-03-11 Jan Hubicka <jh@suse.cz>
27101 * lto-streamer-out.c (output_symbol_p): Skip references from
27102 constructors of external variables.
27104 2013-03-11 Jan Hubicka <jh@suse.cz>
27106 PR middle-end/56571
27107 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
27109 * emit-rtl.c (verify_rtx_sharing): Likewise.
27110 (copy_insn_1): Likewise.
27111 * rtl.c (copy_rtx): Likewise.
27113 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
27116 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
27117 output_operand_lossage message.
27119 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
27122 * arm.c (shift_op): Validate RTL pattern on the fly.
27123 (arm_print_operand, case 'S'): Don't use shift_operator to validate
27126 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27129 * config/pa/pa.md (call_value): Check for calls to powf and direct to
27130 new call patterns that clobber %fr12.
27131 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
27132 split and postreload patterns.
27133 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
27134 registers %fr12 and %fr12R as call used.
27136 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
27138 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
27139 (canon_address, record_store, replace_read, check_mem_read_rtx,
27140 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
27141 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
27142 rest_of_handle_dse): Likewise.
27144 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
27146 PR middle-end/56524
27147 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
27149 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
27150 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
27151 (save_optabs_if_changed): Replace with...
27152 (init_tree_optimization_optabs): ...this.
27153 * optabs.c (save_optabs_if_changed): Rename to...
27154 (init_tree_optimization_optabs): ...this. Take the optimization node
27155 as argument. Do nothing if the base optabs are already correct.
27156 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
27157 to recompute optabs.
27158 * function.h (function): Remove optabs field.
27159 * function.c (invoke_set_current_function_hook): Call
27160 init_tree_optimization_optabs. Use the result to initialize
27163 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
27165 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
27166 if GTMA_HAS_NO_INSTRUMENTATION.
27167 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
27168 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
27169 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
27170 * gimple-pretty-print.c (dump_gimple_transaction): Handle
27171 GTMA_HAS_NO_INSTRUMENTATION.
27173 2013-03-08 Jakub Jelinek <jakub@redhat.com>
27175 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
27178 2013-03-08 Marek Polacek <polacek@redhat.com>
27179 Jakub Jelinek <jakub@redhat.com>
27181 PR tree-optimization/56478
27182 * predict.c (is_comparison_with_loop_invariant_p): Change the
27183 type of loop_step to tree.
27184 (predict_loops): Adjust.
27185 (predict_iv_comparison): Perform the computations on double_ints.
27187 2013-03-08 Richard Biener <rguenther@suse.de>
27189 PR tree-optimization/56570
27190 * tree-cfg.c (verify_expr_location_1): Verify locations for
27192 * tree-sra.c (create_access_replacement): Strip locations
27193 from DECL_DEBUG_EXPRs.
27195 2013-03-08 Richard Biener <rguenther@suse.de>
27197 * tree-inline.c (expand_call_inline): Do not associate
27198 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
27199 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
27201 2013-03-08 Richard Biener <rguenther@suse.de>
27203 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
27204 or block changes with -Og. Fix for location / block encoding
27205 changes and PHI arguments with locations.
27207 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
27209 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
27211 (struct output_info): Likewise.
27212 (register_overhead): Remove bad gcc_assert.
27213 (bitmap_find_bit): If there is only a single bitmap element, do not
27214 count a miss as a search.
27215 (print_statistics): Update for counter type changes.
27216 (dump_bitmap_statistics): Likewise. Print headers such that they
27217 are properly lined up with the printed counters.
27219 2013-03-07 Jakub Jelinek <jakub@redhat.com>
27221 PR tree-optimization/56559
27222 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
27223 check that it has only a single use.
27225 2013-03-07 Richard Biener <rguenther@suse.de>
27227 * doc/invoke.texi (fwhole-program): Discourage use in combination
27230 2013-03-06 Jakub Jelinek <jakub@redhat.com>
27232 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
27234 PR tree-optimization/56539
27235 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
27236 instead of GSI_CONTINUE_LINKING as last argument to
27237 force_gimple_operand_gsi. Adjust function comment.
27239 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
27242 PR middle-end/56548
27243 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
27244 promoted mode, convert the result back to the original mode.
27246 2013-03-06 Richard Biener <rguenther@suse.de>
27248 PR middle-end/56294
27249 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
27250 (insert_updated_phi_nodes_compare_uids): New function.
27251 (update_ssa): Sort symbols_to_rename after UID before
27252 traversing it to insert PHI nodes.
27254 2013-03-06 Richard Biener <rguenther@suse.de>
27256 PR middle-end/50494
27257 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
27258 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
27261 2013-02-13 Richard Biener <rguenther@suse.de>
27264 * varasm.c (output_constant_def_1): Get the decl representing
27265 the constant as argument.
27266 (output_constant_def): Wrap output_constant_def_1.
27267 (make_decl_rtl): Use output_constant_def_1 with the decl
27268 representing the constant.
27269 (build_constant_desc): Optionally re-use a decl already
27270 representing the constant.
27271 (tree_output_constant_def): Adjust.
27273 2013-03-06 Joey Ye <joey.ye@arm.com>
27276 * gcc.c (convert_white_space): New function.
27277 (main): Handles white space in function name.
27279 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
27282 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
27283 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
27284 to SH_DIV_CALL_TABLE for TARGET_SH2.
27285 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
27287 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
27288 call-table options.
27290 2013-03-05 Sterling Augustine <saugustine@google.com>
27291 Cary Coutant <ccoutant@google.com>
27294 * dwarf2out.c (resolve_addr): Don't call
27295 remove_loc_list_addr_table_entries a second time for the same
27298 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27301 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
27302 (avoid_complex_debug_insns): New function.
27303 (expand_debug_locations): Call it.
27305 PR rtl-optimization/56484
27306 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
27307 lifetimes of hard registers on small register class machines.
27309 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27311 * config/microblaze/microblaze-protos.h: Rename
27312 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
27313 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
27315 (microblaze_fast_interrupt_function_p): New function.
27316 (microblaze_is_interrupt_handler): Rename to
27317 microblaze_is_interrupt_variant and add fast_interrupt check.
27318 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
27319 (save_restore_insns): Likewise.
27320 (compute_frame_size): Likewise.
27321 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
27322 (microblaze_globalize_label): Likewise.
27323 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
27324 * config/microblaze/microblaze.md: Use wrapper
27325 microblaze_is_interrupt_variant.
27327 2013-03-05 Kai Tietz <ktietz@redhat.com>
27329 * sdbout.c (sdbout_one_type): Switch to current function's section
27330 supporting cold/hot.
27332 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27334 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
27337 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27339 PR middle-end/56461
27340 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
27341 if VALGRIND_GET_VBITS is defined, temporarily make object
27342 memory all defined, and restore previous valgrind addressability
27343 and definability afterwards. Free this_object at the end.
27345 PR middle-end/56461
27346 * lra.c (lra): Call lra_clear_live_ranges if live_p,
27347 right before calling lra_create_live_ranges, also call it
27348 when clearing live_p. Only call lra_clear_live_ranges
27349 at the end if live_p.
27351 PR middle-end/56461
27352 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
27354 2013-03-05 Richard Biener <rguenther@suse.de>
27356 PR tree-optimization/56521
27357 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
27360 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
27363 * except.h (remove_unreachable_eh_regions): New prototype.
27364 * except.c (remove_eh_handler_splicer): New function, split out
27365 of remove_eh_handler.
27366 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
27367 warning about running it on many EH regions one at a time.
27368 (remove_unreachable_eh_regions_worker): New function, walk the
27369 EH tree in depth-first order and remove non-marked regions.
27370 (remove_unreachable_eh_regions): New function.
27371 * tree-eh.c (mark_reachable_handlers): New function, split out
27372 from remove_unreachable_handlers.
27373 (remove_unreachable_handlers): Use mark_reachable_handlers and
27374 remove_unreachable_eh_regions.
27375 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
27376 and remove_unreachable_eh_regions.
27378 2013-03-05 Richard Biener <rguenther@suse.de>
27380 PR middle-end/56525
27381 * loop-init.c (fix_loop_structure): Remove loops in two stages,
27382 not freeing them until the end.
27384 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27386 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
27388 2013-03-05 Richard Biener <rguenther@suse.de>
27390 PR tree-optimization/56270
27391 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
27392 of loads after scheduling an SLP instance.
27394 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27396 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
27398 (check_gcc_parallelize): Run guality.exp as a separate job from
27399 vect.exp with unsorted.exp and $(dg_target_exps) separately from
27400 struct-layout-1.exp with stackalign.exp.
27402 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
27404 PR middle-end/56461
27405 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
27406 load_index sbitmap even if some bit in it isn't set.
27408 PR middle-end/56461
27409 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
27410 (discover_iteration_bound_by_body_walk): Change queues to
27411 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
27412 spelling in comment. Call safe_push on queues[bound_index] directly.
27413 Release queues[queue_index] in every iteration unconditionally.
27414 Release bounds vector.
27416 PR middle-end/56461
27417 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
27418 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
27419 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
27422 2013-03-05 Richard Biener <rguenther@suse.de>
27425 * tree-inline.c (remap_blocks_to_null): New function.
27426 (expand_call_inline): When expanding a call stmt without
27427 an associated block inline remap all callee blocks to NULL.
27429 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27431 PR rtl-optimization/56494
27432 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
27433 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
27434 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
27436 PR middle-end/56461
27437 * sel-sched-ir.c (free_sched_pools): Release
27438 succs_info_pool.stack[succs_info_pool.max_top] vectors too
27439 if succs_info_pool.max_top isn't -1.
27442 * opts.c (opts_obstack, opts_concat): Moved to...
27443 * opts-common.c (opts_obstack, opts_concat): ... here.
27445 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27447 PR middle-end/56461
27448 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
27450 2013-03-04 Martin Jambor <mjambor@suse.cz>
27452 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
27453 all appropriate places.
27455 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
27457 PR tree-optimization/56424
27458 * ipa-split.c (split_function): Do not set the RSO flag if result is
27459 not by reference and its type is a register type.
27461 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
27463 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
27464 (microblaze_legitimate_pic_operand): Likewise
27465 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
27466 new function microblaze_legitimate_pic_operand
27467 * config/microblaze/microblaze-protos.h
27468 (microblaze_legitimate_pic_operand): Declare.
27470 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27472 * config/microblaze/predicates.md (call_insn_simple_operand):
27473 New predicate for supported rtx code types.
27474 * config/microblaze/microblaze.md (call_internal1): Use
27475 call_insn_simple_operand predicate.
27477 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27479 PR middle-end/56461
27480 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
27481 partitions.ordered_remove.
27483 PR middle-end/56461
27484 * tree-vect-stmts.c (vectorizable_conversion): Don't call
27485 vec_oprnds0.create (1) for modifier == NONE.
27487 PR middle-end/56461
27488 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
27489 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
27490 vec_oprnds1 right before pushing anything to it for
27493 PR middle-end/56461
27494 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
27495 set nbbs to 0 instead of having separate code path.
27496 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
27497 instead of false as last argument if returning NULL.
27499 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
27501 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
27502 the attribute is now called "target" instead of "option".
27503 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
27504 * doc/tm.texi.in (Target Attributes): Likewise document the correct
27505 attribute/pragma name for TARGET_OPTION_VALID_P and
27506 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
27507 * doc/tm.texi: Regenerated.
27509 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
27511 * config/microblaze/microblaze.c:
27512 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
27513 * config/microblaze/microblaze.h: Add -mxl-reorder to
27515 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
27516 instructions emitted if TARGET_REORDER.
27517 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
27518 or 0 for -m/-mno case, but initialises as 2 to detect default use case
27521 2013-03-01 Xinliang David Li <davidxl@google.com>
27523 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
27526 2013-03-01 Jakub Jelinek <jakub@redhat.com>
27528 PR middle-end/56461
27529 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
27530 vector even when returning true. Fix up function comment formatting.
27532 PR middle-end/56461
27533 * ira-build.c (ira_loop_nodes_count): New variable.
27534 (create_loop_tree_nodes): Initialize it.
27535 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
27537 PR middle-end/56461
27538 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
27539 method on dr_chain and result_chain.
27540 * tree-vect-stmts.c (vectorizable_store): Only call
27541 result_chain.create if j == 0.
27543 PR middle-end/56461
27544 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
27545 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
27546 before overwriting it.
27548 2013-03-01 Tobias Burnus <burnus@net-b.de>
27550 * doc/extended.texi (C Extensions): Change order in @menu
27552 (Other MIPS Built-in Functions): Move last MIPS entry before
27553 "picoChip Built-in Functions".
27554 (SH Built-in Functions): Move after RX Built-in Functions.
27555 * doc/gcc.texi (Introduction): Change order in @menu
27557 * doc/md.texi (Constraints): Ditto.
27558 * gty.texi (Type Information): Ditto.
27559 (User-provided marking routines for template types): Make
27561 * doc/invoke.texi (AArch64 Options): Move before
27562 "Adapteva Epiphany Options".
27564 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
27565 Jakub Jelinek <jakub@redhat.com>
27568 * asan.c (gate_asan): Lookup no_sanitize_address instead of
27569 no_address_safety_analysis attribute.
27570 * doc/extend.texi (no_address_safety_attribute): Rename to
27571 no_sanitize_address attribute, mention no_address_safety_analysis
27572 attribute as deprecated alias.
27574 2013-02-28 Jakub Jelinek <jakub@redhat.com>
27576 PR middle-end/56461
27577 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
27578 type to vec<vec<tree> > *.
27579 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
27580 to be vec<tree> instead of vec<tree> *, set vec_defs
27581 to vNULL and call vec_defs.create (number_of_vects), adjust other
27583 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
27584 vectorizable_condition): Adjust vect_get_slp_defs callers.
27586 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27588 * config/aarch64/aarch64.c
27589 (aarch64_float_const_representable): Remove unused variable.
27591 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27593 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
27595 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27597 * config/aarch64/aarch64-builtins.c
27598 (aarch64_init_simd_builtins): Make static.
27600 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27602 * config/aarch64/aarch64.c
27603 (aarch64_simd_make_constant): Make static.
27605 2013-02-28 Martin Jambor <mjambor@suse.cz>
27607 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
27608 with no initialization to the RHS of debug statements.
27610 2013-02-28 Martin Jambor <mjambor@suse.cz>
27612 PR tree-optimization/56294
27613 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
27615 (get_access_replacement): Do not call create_access_replacement.
27616 Assert a replacement exists.
27617 (get_repl_default_def_ssa_name): Create the replacement declaration
27620 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27622 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
27623 final_end_function.
27625 2013-02-28 Marek Polacek <polacek@redhat.com>
27627 PR rtl-optimization/56466
27628 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
27629 if we're changing a loop.
27630 (peel_loops_completely): Likewise.
27632 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
27635 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
27637 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27640 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
27641 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
27642 INTX_FTYPE_FX, FX_FTYPE_INTX.
27643 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
27645 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27647 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
27648 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
27649 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
27650 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
27651 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
27652 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
27653 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
27654 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
27655 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
27656 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
27657 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
27658 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
27659 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
27660 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
27661 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
27662 (avrxmega6): Increase max flash segments from 5 to 6.
27663 * config/avr/t-multilib: Regenerate.
27664 * config/avr/avr-tables.opt: Regenerate.
27665 * doc/avr-mmcu.texi: Regenerate.
27667 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27669 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
27670 (avr_device_to_arch): Rename to avr_device_to_ld.
27671 (avr_device_to_as): New prototype.
27672 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
27673 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
27674 * config/avr/driver-avr.c (avr_device_to_as): New.
27675 (avr_device_to_arch): Rename to avr_device_to_ld.
27677 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27679 PR middle-end/56461
27680 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
27681 method on dr_chain and result_chain.
27683 PR middle-end/56461
27684 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
27685 pointer_set_destroy on not_executed_last_iteration.
27687 PR middle-end/56461
27688 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
27690 PR middle-end/56461
27691 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
27692 FOR_EACH_DEFINED_FUNCTION when freeing state.
27694 PR middle-end/56461
27695 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
27697 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
27700 PR middle-end/56461
27701 * ipa-cp.c (decide_whether_version_node): Call vec_free on
27702 known_aggs[i].items and release known_aggs vector.
27704 PR middle-end/56461
27705 * ipa-reference.c (propagate): Free node_info even for alias nodes.
27707 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27709 * config/microblaze/microblaze.c (microblaze_emit_compare):
27710 Use xor for EQ/NE comparisions.
27711 * config/microblaze/microblaze.md (cstoresf4): Add constraints
27712 (cbranchsf4): Adjust operator to comparison_operator.
27714 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27716 PR middle-end/56461
27717 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
27719 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
27720 vec_safe_push, always update *slot.
27721 (redirect_edge_var_map_clear): Use vec_free.
27722 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
27723 (free_var_map_entry): Use vec_free.
27724 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
27725 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
27727 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
27729 PR middle-end/45472
27730 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27731 when the may_trap_p bit of the exprs being merged differs.
27732 Reorder tests for speculativeness in the logical and operator.
27734 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27736 * incpath.c (add_standard_paths): Use reconcat instead of concat
27737 where appropriate and avoid leaking memory.
27739 * opts.h: Include obstack.h.
27740 (opts_concat): New prototype.
27741 (opts_obstack): New declaration.
27742 * opts.c (opts_concat): New function.
27743 (opts_obstack): New variable.
27744 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27745 (finish_options): Use opts_concat instead of concat
27746 and XOBNEWVEC instead of XNEWVEC.
27747 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27748 generate_option): Likewise.
27749 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27750 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27753 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27754 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27756 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27758 PR middle-end/56461
27759 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27761 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27763 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27764 (arm_block_move_unaligned_straight): Likewise.
27765 (arm_adjust_block_mem): Likewise.
27767 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27770 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27771 temp, cond and label.
27772 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27775 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27776 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27779 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27780 of prologue/epilogue functions.
27783 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27784 Remove unused variable cfa_offset.
27785 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27788 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27789 type promotion to unsigned.
27792 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27793 for HOST_WIDE_INT of 32 bit / same size as int.
27794 (arm_block_move_unaligned_straight): Likewise.
27795 (arm_adjust_block_mem): Likewise.
27798 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27801 2013-02-26 Marek Polacek <polacek@redhat.com>
27803 PR tree-optimization/56426
27804 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27806 2013-02-26 Richard Biener <rguenther@suse.de>
27809 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27810 unused variable loops.
27812 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27814 PR tree-optimization/56448
27815 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27816 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27817 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27818 later operands of the references, or even first operand for
27819 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27821 PR tree-optimization/56443
27822 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27823 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27824 to type_for_mode langhook.
27826 2013-02-25 Matt Turner <mattst88@gmail.com>
27828 * doc/invoke.texi: Document r4700.
27830 2013-02-25 Richard Biener <rguenther@suse.de>
27832 PR tree-optimization/56175
27833 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27835 (simplify_bitwise_binary): ... here. Also guard the conversion
27836 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27838 2013-02-25 Catherine Moore <clm@codesourcery.com>
27841 2013-02-24 Catherine Moore <clm@codesourcery.com>
27842 Maciej W. Rozycki <macro@codesourcery.com>
27843 Tom de Vries <tom@codesourcery.com>
27844 Nathan Sidwell <nathan@codesourcery.com>
27845 Iain Sandoe <iain@codesourcery.com>
27846 Nathan Froyd <froydnj@codesourcery.com>
27847 Chao-ying Fu <fu@mips.com>
27849 * doc/extend.texi: (micromips, nomicromips, nocompression):
27850 Document new function attributes.
27851 * doc/invoke.texi (minterlink-compressed, mmicromips,
27852 m14k, m14ke, m14kec): Document new options.
27853 (minterlink-mips16): Update documentation.
27854 * doc/md.texi (ZC, ZD): Document new constraints.
27855 * configure.ac (gcc_cv_as_micromips): Check if linker
27856 supports the .set micromips directive.
27857 * configure: Regenerate.
27858 * config.in: Regenerate.
27859 * config/mips/mips-tables.opt: Regenerate.
27860 * config/mips/micromips.md: New file.
27861 * constraints.md (ZC, AD): New constraints.
27862 * config/mips/predicates.md (movep_src_register): New predicate.
27863 (movep_src_operand): New predicate.
27864 (non_volatile_mem_operand): New predicate.
27865 * config/mips/mips.md (multimem): New type.
27866 (length): Differentiate between 17-bit and 18-bit branch offsets.
27867 (MOVEP1, MOVEP2): New mode iterator.
27868 (mov_<load>l): Use ZC constraint.
27869 (mov_<load>r): Likewise.
27870 (mov_<store>l): Likewise.
27871 (mov_<store>r): Likewise.
27872 (*branch_equality<mode>_inverted): Add microMIPS support.
27873 (*branch_equality<mode>): Likewise.
27874 (*jump_absolute): Likewise.
27875 (indirect_jump_<mode>): Likewise.
27876 (tablejump_<mode>): Likewise.
27877 (<optab>_internal): Likewise.
27878 (sibcall_internal): Likewise.
27879 (sibcall_value_internal): Likewise.
27880 (prefetch): Use constraint ZD.
27881 * config/mips/mips.opt (minterlink-compressed): New option.
27882 (minterlink-mips16): Now an alias for minterlink-compressed.
27883 (mmicromips): New option.
27884 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27885 (compare_and_swap_12): Likewise.
27886 (sync_add<mode>): Likewise.
27887 (sync_<optab>_12): Likewise.
27888 (sync_old_<optab>_12): Likewise.
27889 (sync_new_<optab>_12): Likewise.
27890 (sync_nand_12): Likewise.
27891 (sync_old_nand_12): Likewise.
27892 (sync_new_nand_12): Likewise.
27893 (sync_sub<mode>): Likewise.
27894 (sync_old_add<mode>): Likewise.
27895 (sync_old_sub<mode>): Likewise.
27896 (sync_new_add<mode>): Likewise.
27897 (sync_new_sub<mode>): Likewise.
27898 (sync_<optab><mode>): Likewise.
27899 (sync_old_<optab><mode>): Likewise.
27900 (sync_new_<optab><mode>): Likewise.
27901 (sync_nand<mode>): Likewise.
27902 (sync_old_nand<mode>): Likewise.
27903 (sync_new_nand<mode>): Likewise.
27904 (sync_lock_test_and_set<mode>): Likewise.
27905 (test_and_set_12): Likewise.
27906 (atomic_compare_and_swap<mode>): Likewise.
27907 (atomic_exchange<mode>_llsc): Likewise.
27908 (atomic_fetch_add<mode>_llsc): Likewise.
27909 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27910 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27911 (umips_save_restore_pattern_p): Likewise.
27912 (umips_load_store_pair_p): Likewise.
27913 (umips_output_load_store_pair): Likewise.
27914 (umips_movep_target_p): Likewise.
27915 (umips_12bit_offset_address_p): Likewise.
27916 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27917 (mips_base_mips16): Rename this...
27918 (mips_base_compression_flags): ...to this. Update all uses.
27919 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27920 (mips_mips16_decl_p): Delete.
27921 (mips_nomips16_decl_p): Delete.
27922 (mips_get_compress_on_flags): New function.
27923 (mips_get_compress_off_flags): New function.
27924 (mips_get_compress_mode): New function.
27925 (mips_get_compress_on_name): New function.
27926 (mips_get_compress_off_name): New function.
27927 (mips_insert_attributes): Support multiple compression types.
27928 (mips_merge_decl_attributes): Likewise.
27929 (umips_12bit_offset_address_p): New function.
27930 (mips_start_function_definition): Emit .set micromips directive.
27931 (mips_call_may_need_jalx_p): New function.
27932 (mips_function_ok_for_sibcall): Add microMIPS support.
27933 (mips_print_operand_punctuation): Support short delay slots and
27935 (umips_swm_mask, umips_swm_encoding): New.
27936 (umips_build_save_restore): New function.
27937 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27938 (was_mips16_p): Remove.
27939 (old_compression_mode): New.
27940 (mips_set_compression_mode): New function.
27941 (mips_set_current_function): Add microMIPS support.
27942 (mips_option_override): Likewise.
27943 (umips_save_restore_pattern_p): New function.
27944 (umips_output_save_restore): New function.
27945 (umips_load_store_pair_p_1): New function.
27946 (umips_load_store_pair_p): New function.
27947 (umips_output_load_store_pair_1): New function.
27948 (umips_output_load_store_pair): New function.
27949 (umips_movep_target_p) New function.
27950 (mips_prepare_pch_save): Add microMIPS support.
27951 * config/mips/mips.h (TARGET_COMPRESSION): New.
27952 (TARGET_CPU_CPP_BUILTINS): Update macro
27953 to use new compression flags and to support microMIPS.
27954 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27955 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27956 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27957 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27958 (ASM_SPEC): Support mmicromips and mno-micromips.
27959 (M16STORE_REG_P): New macro.
27960 (MIPS_CALL): Support TARGET_MICROMIPS.
27961 (MICROMIPS_J): New macro.
27962 (mips_base_mips16): Rename this...
27963 (mips_base_compression_flags): ...to this.
27964 (UMIPS_12BIT_OFFSET_P): New macro.
27965 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27966 (MULTILIB_DIRNAMES): Likewise.
27968 2013-02-25 Tom de Vries <tom@codesourcery.com>
27970 PR rtl-optimization/56131
27971 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27972 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27973 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27975 2013-02-25 Tobias Burnus <burnus@net-b.de>
27977 * doc/invoke.texi (-fsanitize=): Move from optimization
27978 to debugging options.
27980 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27982 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27984 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27985 Alexander Monakov <amonakov@ispras.ru>
27987 PR middle-end/56077
27988 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27989 flush pending lists also on non-jumps. Adjust comment.
27991 2013-02-24 Catherine Moore <clm@codesourcery.com>
27992 Maciej W. Rozycki <macro@codesourcery.com>
27993 Tom de Vries <tom@codesourcery.com>
27994 Nathan Sidwell <nathan@codesourcery.com>
27995 Iain Sandoe <iain@codesourcery.com>
27996 Nathan Froyd <froydnj@codesourcery.com>
27997 Chao-ying Fu <fu@mips.com>
27999 * doc/extend.texi: (micromips, nomicromips, nocompression):
28000 Document new function attributes.
28001 * doc/invoke.texi (minterlink-compressed, mmicromips,
28002 m14k, m14ke, m14kec): Document new options.
28003 (minterlink-mips16): Update documentation.
28004 * doc/md.texi (ZC, ZD): Document new constraints.
28005 * configure.ac (gcc_cv_as_micromips): Check if linker
28006 supports the .set micromips directive.
28007 * configure: Regenerate.
28008 * config.in: Regenerate.
28009 * config/mips/mips-tables.opt: Regenerate.
28010 * config/mips/micromips.md: New file.
28011 * constraints.md (ZC, AD): New constraints.
28012 * config/mips/predicates.md (movep_src_register): New predicate.
28013 (movep_src_operand): New predicate.
28014 (non_volatile_mem_operand): New predicate.
28015 * config/mips/mips.md (multimem): New type.
28016 (length): Differentiate between 17-bit and 18-bit branch offsets.
28017 (MOVEP1, MOVEP2): New mode iterator.
28018 (mov_<load>l): Use ZC constraint.
28019 (mov_<load>r): Likewise.
28020 (mov_<store>l): Likewise.
28021 (mov_<store>r): Likewise.
28022 (*branch_equality<mode>_inverted): Add microMIPS support.
28023 (*branch_equality<mode>): Likewise.
28024 (*jump_absolute): Likewise.
28025 (indirect_jump_<mode>): Likewise.
28026 (tablejump_<mode>): Likewise.
28027 (<optab>_internal): Likewise.
28028 (sibcall_internal): Likewise.
28029 (sibcall_value_internal): Likewise.
28030 (prefetch): Use constraint ZD.
28031 * config/mips/mips.opt (minterlink-compressed): New option.
28032 (minterlink-mips16): Now an alias for minterlink-compressed.
28033 (mmicromips): New option.
28034 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
28035 (compare_and_swap_12): Likewise.
28036 (sync_add<mode>): Likewise.
28037 (sync_<optab>_12): Likewise.
28038 (sync_old_<optab>_12): Likewise.
28039 (sync_new_<optab>_12): Likewise.
28040 (sync_nand_12): Likewise.
28041 (sync_old_nand_12): Likewise.
28042 (sync_new_nand_12): Likewise.
28043 (sync_sub<mode>): Likewise.
28044 (sync_old_add<mode>): Likewise.
28045 (sync_old_sub<mode>): Likewise.
28046 (sync_new_add<mode>): Likewise.
28047 (sync_new_sub<mode>): Likewise.
28048 (sync_<optab><mode>): Likewise.
28049 (sync_old_<optab><mode>): Likewise.
28050 (sync_new_<optab><mode>): Likewise.
28051 (sync_nand<mode>): Likewise.
28052 (sync_old_nand<mode>): Likewise.
28053 (sync_new_nand<mode>): Likewise.
28054 (sync_lock_test_and_set<mode>): Likewise.
28055 (test_and_set_12): Likewise.
28056 (atomic_compare_and_swap<mode>): Likewise.
28057 (atomic_exchange<mode>_llsc): Likewise.
28058 (atomic_fetch_add<mode>_llsc): Likewise.
28059 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
28060 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
28061 (umips_save_restore_pattern_p): Likewise.
28062 (umips_load_store_pair_p): Likewise.
28063 (umips_output_load_store_pair): Likewise.
28064 (umips_movep_target_p): Likewise.
28065 (umips_12bit_offset_address_p): Likewise.
28066 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
28067 (mips_base_mips16): Rename this...
28068 (mips_base_compression_flags): ...to this. Update all uses.
28069 (mips_attribute_table): Add micromips, nomicromips and nocompression.
28070 (mips_mips16_decl_p): Delete.
28071 (mips_nomips16_decl_p): Delete.
28072 (mips_get_compress_on_flags): New function.
28073 (mips_get_compress_off_flags): New function.
28074 (mips_get_compress_mode): New function.
28075 (mips_get_compress_on_name): New function.
28076 (mips_get_compress_off_name): New function.
28077 (mips_insert_attributes): Support multiple compression types.
28078 (mips_merge_decl_attributes): Likewise.
28079 (umips_12bit_offset_address_p): New function.
28080 (mips_start_function_definition): Emit .set micromips directive.
28081 (mips_call_may_need_jalx_p): New function.
28082 (mips_function_ok_for_sibcall): Add microMIPS support.
28083 (mips_print_operand_punctuation): Support short delay slots and
28085 (umips_swm_mask, umips_swm_encoding): New.
28086 (umips_build_save_restore): New function.
28087 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
28088 (was_mips16_p): Remove.
28089 (old_compression_mode): New.
28090 (mips_set_compression_mode): New function.
28091 (mips_set_current_function): Add microMIPS support.
28092 (mips_option_override): Likewise.
28093 (umips_save_restore_pattern_p): New function.
28094 (umips_output_save_restore): New function.
28095 (umips_load_store_pair_p_1): New function.
28096 (umips_load_store_pair_p): New function.
28097 (umips_output_load_store_pair_1): New function.
28098 (umips_output_load_store_pair): New function.
28099 (umips_movep_target_p) New function.
28100 (mips_prepare_pch_save): Add microMIPS support.
28101 * config/mips/mips.h (TARGET_COMPRESSION): New.
28102 (TARGET_CPU_CPP_BUILTINS): Update macro
28103 to use new compression flags and to support microMIPS.
28104 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
28105 (MIPS_ARCH_FLOAT_SPEC): Likewise.
28106 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
28107 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
28108 (ASM_SPEC): Support mmicromips and mno-micromips.
28109 (M16STORE_REG_P): New macro.
28110 (MIPS_CALL): Support TARGET_MICROMIPS.
28111 (MICROMIPS_J): New macro.
28112 (mips_base_mips16): Rename this...
28113 (mips_base_compression_flags): ...to this.
28114 (UMIPS_12BIT_OFFSET_P): New macro.
28115 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
28116 (MULTILIB_DIRNAMES): Likewise.
28118 2013-02-24 Jakub Jelinek <jakub@redhat.com>
28121 * target-globals.c (save_target_globals): For init_reg_sets and
28122 target_reinit remporarily set this_fn_optabs to this_target_optabs.
28124 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
28126 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
28127 * config/aarch64/t-aarch64
28128 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
28130 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
28132 PR inline-asm/56148
28133 * lra-constraints.c (process_alt_operands): Reload operand
28134 conflicting with earlier clobber only if no more other conflicting
28137 2013-02-22 Jakub Jelinek <jakub@redhat.com>
28140 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
28141 if not linking a shared library.
28143 2013-02-22 Seth LaForge <sethml@google.com>
28145 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
28147 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
28149 * config/arm/arm.md (split for extendsidi): Update condition.
28150 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
28151 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
28152 (qhs_zextenddi_cstr): Likewise.
28154 2013-02-21 Jakub Jelinek <jakub@redhat.com>
28156 PR middle-end/56420
28157 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
28158 avoid signed wrapping.
28159 (expand_mult): Handle properly multiplication by
28160 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
28161 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
28162 in the compiler if coeff is HOST_WIDE_INT_MIN.
28163 (expand_divmod): Don't make ext_op1 static, change it's type to
28164 uhwi. Avoid undefined behavior in -INTVAL (op1).
28166 PR rtl-optimization/50339
28167 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
28169 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
28170 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
28171 into splitting_ashiftrt field.
28172 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
28174 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
28177 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
28179 PR middle-end/56108
28180 * trans-mem.c (execute_tm_mark): Do not expand transactions that
28181 are sure to go irrevocable.
28183 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
28185 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
28186 scalars are valid operands.
28188 2013-02-21 Martin Jambor <mjambor@suse.cz>
28190 PR tree-optimization/56310
28191 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
28192 only matching indices and non-negative final offsets.
28193 (intersect_aggregates_with_edge): Pass src_idx to
28194 agg_replacements_to_vector. Pass src_idx insstead of index to
28195 intersect_with_agg_replacements.
28197 2013-02-21 Martin Jambor <mjambor@suse.cz>
28199 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
28200 instead of hard-wired defaults.
28202 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
28204 * doc/invoke.texi (MIPS Options): Update documentation of the
28205 floating-point multiply-accumulate instruction restrictions.
28207 2013-02-21 Kostya Serebryany <kcc@google.com>
28209 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
28210 asan_shadow_offset on x86_64 linux.
28212 2013-02-21 Richard Biener <rguenther@suse.de>
28214 PR tree-optimization/56415
28216 2013-02-11 Richard Biener <rguenther@suse.de>
28218 PR tree-optimization/56273
28219 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28222 2013-02-21 Jakub Jelinek <jakub@redhat.com>
28225 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
28228 PR inline-asm/56405
28229 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
28230 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
28232 2013-02-20 Jan Hubicka <jh@suse.cz>
28234 PR tree-optimization/56265
28235 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
28236 when target is referenced for first time.
28238 2013-02-20 Richard Biener <rguenther@suse.de>
28240 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
28241 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
28242 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
28243 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
28244 not return anything.
28245 (rename_ssa_copies): Do not remove unused locals.
28246 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
28247 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
28248 * passes.c (execute_function_todo): Do not schedule unused locals
28249 removal if cleanup_tree_cfg did something.
28250 * tree-ssa-live.c (remove_unused_locals): Dump statistics
28251 about the number of removed locals.
28253 2013-02-20 Richard Biener <rguenther@suse.de>
28255 PR tree-optimization/56398
28256 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
28258 2013-02-20 Martin Jambor <mjambor@suse.cz>
28260 PR tree-optimization/55334
28261 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
28262 restricted pointers to arrays.
28264 2013-02-20 Richard Biener <rguenther@suse.de>
28265 Jakub Jelinek <jakub@redhat.com>
28267 PR tree-optimization/56396
28268 * tree-ssa-ccp.c (n_const_val): New static variable.
28269 (get_value): Return NULL for SSA names we don't have a lattice
28271 (ccp_initialize): Initialize n_const_val.
28272 * tree-ssa-copy.c (n_copy_of): New static variable.
28273 (init_copy_prop): Initialize n_copy_of.
28274 (get_value): Return NULL_TREE for SSA names we don't have a
28277 2013-02-20 Martin Jambor <mjambor@suse.cz>
28279 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
28281 2013-02-20 Richard Biener <rguenther@suse.de>
28283 * genpreds.c (write_lookup_constraint): Do not compare first
28284 letter of the constraint again.
28286 2013-02-20 Richard Biener <rguenther@suse.de>
28288 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
28290 (get_use_iv_cost): Terminate hashtable walk when coming across
28293 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
28295 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
28296 reassociation for avx2 targets.
28298 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28300 * config/microblaze/microblaze.c: microblaze_has_clz = 0
28301 Add version check for v8.10.a to enable microblaze_has_clz
28302 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
28303 version and TARGET_PATTERN_COMPARE check
28304 * config/microblaze/microblaze.md: New clzsi2 instruction
28306 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28308 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
28309 function before branching.
28311 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28313 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
28315 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
28317 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28319 PR middle-end/55889
28320 * sel-sched.c: Include ira.h.
28321 (implicit_clobber_conflict_p): New function.
28322 (moveup_expr): Use it.
28323 * Makefile.in (sel-sched.o): Depend on ira.h.
28325 2013-02-19 Richard Biener <rguenther@suse.de>
28327 PR tree-optimization/56384
28328 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
28329 (vn_hash_type): Split out from ...
28330 (vn_hash_constant_with_type): ... here.
28331 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
28332 (vn_phi_eq): Compare types from vn_phi_s structure.
28333 (vn_phi_lookup): Populate vn_phi_s type.
28334 (vn_phi_insert): Likewise.
28336 2013-02-19 Jakub Jelinek <jakub@redhat.com>
28338 PR tree-optimization/56350
28339 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
28340 if haven't found reduction or nested cycle operand, rather than
28341 asserting we must find it.
28343 PR tree-optimization/56381
28344 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
28347 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
28348 Jakub Jelinek <jakub@redhat.com>
28351 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
28352 (swap_optab_enable): Same.
28353 (init_all_optabs): Use argument instead of global.
28354 * tree.h (struct tree_optimization_option): New field target_optabs.
28355 * expr.h (init_all_optabs): Add argument to prototype.
28356 (TREE_OPTIMIZATION_OPTABS): New.
28357 (save_optabs_if_changed): Protoize.
28358 * optabs.h: Declare this_fn_optabs.
28359 * optabs.c (save_optabs_if_changed): New.
28360 Declare this_fn_optabs.
28361 (init_optabs): Add argument to init_all_optabs() call.
28362 * function.c (invoke_set_current_function_hook): Handle per
28364 * function.h (struct function): New field optabs.
28365 * config/mips/mips.c (mips_set_mips16_mode): Handle when
28366 optimization_current_node has changed.
28367 * target-globals.h (save_target_globals_default_opts): Protoize.
28368 * target-globals.c (save_target_globals_default_opts): New.
28370 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28373 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
28374 registers %fr12 and %fr12R as call used.
28377 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
28378 and HImode, require all displacements to be an integer multiple of
28380 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
28381 only allow QImode and HImode when reload is in progress and strict is
28382 true. Likewise for symbolic addresses. Use base14_operand to check
28383 displacements in REG+BASE addresses.
28385 2013-02-18 Richard Biener <rguenther@suse.de>
28387 PR tree-optimization/56366
28388 * tree-vect-loop.c (get_initial_def_for_induction): Properly
28389 handle sign-conversion of outer-loop initial induction value.
28391 2013-02-18 Richard Biener <rguenther@suse.de>
28393 PR middle-end/56349
28394 * cfghooks.c (merge_blocks): If we merge a latch into another
28395 block adjust references to it.
28396 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
28397 (verify_loop_structure): Verify that a recorded latch is in fact
28400 2013-02-18 Richard Biener <rguenther@suse.de>
28402 PR tree-optimization/56321
28403 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
28404 order SSA name release and virtual operand unlinking.
28406 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28408 * config/microblaze/microblaze.md (save_stack_block): Define.
28409 (restore_stack_block): Likewise.
28411 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28413 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
28414 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
28415 * config/microblaze/microblaze.c (microblaze_option_override):
28416 Bail out early for PIC modes when target does not support PIC.
28418 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28420 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
28421 Replace with a microblaze version.
28422 (microblaze_trampoline_init): Adapt for microblaze.
28423 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
28426 2013-02-16 Jakub Jelinek <jakub@redhat.com>
28427 Dodji Seketeli <dodji@redhat.com>
28430 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
28431 (instrument_mem_region_access): Do not forget to always put
28432 instrumentation of the of 'base' and 'base + len' in a "if (len !=
28433 0) statement, even for cases where either 'base' or 'base + len'
28434 are not instrumented -- because they have been previously
28435 instrumented. Simplify the logic by putting all the statements
28436 instrument 'base + len' inside a sequence, and then insert that
28437 sequence right before the current insertion point. Then, to
28438 instrument 'base + len', just get an iterator on that statement.
28439 And do not forget to update the pointer to iterator the function
28440 received as argument.
28442 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
28444 PR rtl-optimization/56348
28445 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
28447 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
28449 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
28450 (clean_graph_dump_file): Pass base to start_graph_dump.
28452 2013-02-14 Richard Henderson <rth@redhat.com>
28455 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
28457 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
28459 * collect2-aix.h: Define F_LOADONLY.
28461 2013-02-14 Richard Biener <rguenther@suse.de>
28464 * varasm.c (output_constant_def_1): Get the decl representing
28465 the constant as argument.
28466 (output_constant_def): Wrap output_constant_def_1.
28467 (make_decl_rtl): Use output_constant_def_1 with the decl
28468 representing the constant.
28469 (build_constant_desc): Optionally re-use a decl already
28470 representing the constant.
28471 (tree_output_constant_def): Adjust.
28473 2013-02-14 Dodji Seketeli <dodji@redhat.com>
28476 * asan.c (instrument_builtin_call): Really put the length of the
28477 second source argument into src1_len.
28479 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28481 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
28482 argument. If it is false, don't create edge from then_bb to
28484 (insert_if_then_before_iter): Pass true to it.
28485 (build_check_stmt): Pass false to it.
28486 (transform_statements): Flush hash table only on extended basic
28487 block boundaries, rather than at the beginning of every bb.
28488 Don't flush hash table on nonfreeing_call_p calls.
28489 * tree-flow.h (nonfreeing_call_p): New prototype.
28490 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
28492 2013-02-13 David S. Miller <davem@davemloft.net>
28494 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
28496 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
28499 * ira.c (max_regno_before_ira): Move from ...
28501 (fix_reg_equiv_init): Use max_regno_before_ira instead of
28504 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28506 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
28508 2013-02-13 Richard Biener <rguenther@suse.de>
28511 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
28512 globals in MEM_REFs.
28514 2013-02-13 Richard Biener <rguenther@suse.de>
28516 * loop-init.c (loop_optimizer_init): Clear loop state when
28517 re-initializing preserved loops.
28518 * loop-unswitch.c (unswitch_single_loop): Return whether
28519 we unswitched the loop. Do not verify loop state here.
28520 (unswitch_loops): When we unswitched a loop discover new loops.
28522 2013-02-13 Kostya Serebryany <kcc@google.com>
28524 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
28526 * sanitizer.def: Rename __asan_init to __asan_init_v1.
28528 2013-02-12 Dodji Seketeli <dodji@redhat.com>
28530 Avoid instrumenting duplicated memory access in the same basic block
28531 * Makefile.in (asan.o): Add new dependency on hash-table.h
28532 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
28533 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
28534 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
28535 (free_mem_ref_resources, has_mem_ref_been_instrumented)
28536 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
28537 (get_mem_ref_of_assignment): New functions.
28538 (get_mem_refs_of_builtin_call): Extract from
28539 instrument_builtin_call and tweak a little bit to make it fit with
28541 (instrument_builtin_call): Use the new
28542 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
28543 of is_gimple_builtin_call.
28544 (instrument_derefs, instrument_mem_region_access): Insert the
28545 instrumented memory reference into the hash table.
28546 (maybe_instrument_assignment): Renamed instrument_assignment into
28547 this, and change it to advance the iterator when instrumentation
28548 actually happened and return true in that case. This makes it
28549 homogeneous with maybe_instrument_assignment, and thus give a
28550 chance to callers to be more 'regular'.
28551 (transform_statements): Clear the memory reference hash table
28552 whenever we enter a new BB, when we cross a function call, or when
28553 we are done transforming statements. Use
28554 maybe_instrument_assignment instead of instrumentation. No more
28555 need to special case maybe_instrument_assignment and advance the
28556 iterator after calling it; it's now handled just like
28557 maybe_instrument_call. Update comment.
28559 2013-02-13 Richard Biener <rguenther@suse.de>
28561 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
28562 Fix loop discovery code.
28564 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
28566 PR inline-asm/56148
28567 * lra-constraints.c (process_alt_operands): Match early clobber
28568 operand with itself. Check conflicts with earlyclobber only if
28569 the operand is not reloaded. Prefer to reload conflicting operand
28570 if earlyclobber and matching operands are the same.
28572 2013-02-12 Richard Biener <rguenther@suse.de>
28575 * lto-streamer-out.c (write_symbol): Do not output symbols
28576 for hard register variables.
28578 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
28581 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
28582 (umulsidi3_insn, mulsidi3_insn): New insns.
28584 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
28586 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
28587 (struct tune_params): Add vec_costs field.
28588 * config/arm/arm.c (arm_builtin_vectorization_cost)
28589 (arm_add_stmt_cost): New functions.
28590 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
28591 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
28592 (arm_default_vec_cost): New struct of type cpu_vec_costs.
28593 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
28594 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
28595 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
28596 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
28598 2013-02-12 Richard Biener <rguenther@suse.de>
28601 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
28602 decls again if possible.
28604 2013-02-12 Richard Biener <rguenther@suse.de>
28606 PR middle-end/56288
28607 * tree-ssa.c (verify_ssa_name): Fix check, move
28608 SSA_NAME_IN_FREE_LIST check up.
28610 2013-02-12 Jakub Jelinek <jakub@redhat.com>
28611 Steven Bosscher <steven@gcc.gnu.org>
28613 PR rtl-optimization/56151
28614 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
28615 equal to op0 or op1, and last_insn pattern is CODE operation
28616 with MEM dest and one of the operands matches that MEM.
28618 2013-02-11 Sriraman Tallam <tmsriram@google.com>
28620 * doc/extend.texi: Document Function Multiversioning and "default"
28621 parameter string to target attribute.
28622 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
28623 target attribute parameter is "default".
28624 (ix86_compare_version_priority): Remove checks for target attribute.
28625 (ix86_mangle_function_version_assembler_name): Change error to sorry.
28626 Remove check for target attribute equal to NULL. Add assert.
28627 (ix86_generate_version_dispatcher_body): Change error to sorry.
28629 2013-02-11 Iain Sandoe <iain@codesourcery.com>
28630 Jack Howarth <howarth@bromo.med.uc.edu>
28631 Patrick Marlier <patrick.marlier@gmail.com>
28634 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
28635 define ENDFILE_SPEC as TM_DESTRUCTOR.
28636 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
28638 2013-02-11 Alexander Potapenko <glider@google.com>
28639 Jack Howarth <howarth@bromo.med.uc.edu>
28640 Jakub Jelinek <jakub@redhat.com>
28643 * config/darwin.c (cdtor_record): Rename ctor_record.
28644 (sort_cdtor_records): Rename sort_ctor_records.
28645 (finalize_dtors): New routine to sort destructors by
28646 priority before use in assemble_integer.
28647 (machopic_asm_out_destructor): Use finalize_dtors if needed.
28649 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
28651 PR rtl-optimization/56275
28652 * simplify-rtx.c (avoid_constant_pool_reference): Check that
28653 offset is non-negative and less than cmode size before
28654 calling simplify_subreg.
28656 2013-02-11 Richard Biener <rguenther@suse.de>
28658 PR tree-optimization/56264
28659 * cfgloop.h (fix_loop_structure): Adjust prototype.
28660 * loop-init.c (fix_loop_structure): Return the number of
28661 newly discovered loops.
28662 * tree-cfgcleanup.c (repair_loop_structures): When new loops
28663 are discovered, do a full loop-closed SSA rewrite.
28665 2013-02-11 Richard Biener <rguenther@suse.de>
28667 PR tree-optimization/56273
28668 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28670 (check_array_ref): Fix missing newline in dumps.
28671 (search_for_addr_array): Likewise.
28673 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
28675 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
28677 2013-02-09 Jakub Jelinek <jakub@redhat.com>
28680 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
28682 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
28684 PR rtl-optimization/56246
28685 * lra-constraints.c (simplify_operand_subreg): Try to reuse
28687 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
28688 constraints are satisfied.
28690 2013-02-08 Jeff Law <law@redhat.com>
28693 * emit-rtl.c (reg_is_parm_p): New function.
28694 * regs.h (reg_is_parm_p): New prototype.
28695 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
28696 callee-clobbered registers.
28698 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
28701 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
28702 If there is no implicit builtin declaration, just return NULL.
28704 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28706 * config/i386/sse.md (FMAMODEM): New mode iterator.
28707 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
28708 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
28710 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28712 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
28713 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
28714 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
28716 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28718 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
28719 (microblaze*-*-elf): Likewise.
28720 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
28722 * config/microblaze/microblaze-c.c: Add builtin defines for
28723 _LITTLE_ENDIAN and _BIG_ENDIAN.
28724 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
28725 add to TARGET_DEFAULT flags.
28726 Expand ASM_SPEC and LINK_SPEC.
28727 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
28728 * config/microblaze/microblaze.md: Update extendsidi2 and
28729 movdi_internal instructions to use low-order / high-order reg
28731 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28732 options and inversemask / mask of LITTLE_ENDIAN.
28733 * config/microblaze/t-microblaze: Expand multilib options to
28734 include mlittle-endian (le) and update exceptions patterns.
28736 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28738 PR rtl-optimization/56195
28739 * lra-constraints.c (get_reload_reg): Don't reuse regs
28740 if they have smaller mode than requested, if they have
28741 wider mode than requested, try to return a SUBREG.
28743 PR tree-optimization/56250
28744 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28745 if type is unsigned and code isn't MULT_EXPR.
28747 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28749 PR tree-optimization/56064
28750 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28751 bits according to mode.
28752 * fixed-value.h (fixed_from_double_int)
28753 (const_fixed_from_double_int): Adjust comments.
28755 2013-02-08 Richard Biener <rguenther@suse.de>
28758 * lto-streamer.h (struct data_in): Remove current_file, current_line
28759 and current_col members.
28760 * lto-streamer-out.c (lto_output_location): Stream changed bits
28761 en-block for efficiency.
28762 * lto-streamer-in.c (clear_line_info): Remove.
28763 (lto_input_location): Cache current file, line and column
28764 globally via local statics. Read changed bits en-block.
28765 (input_function): Do not call clear_line_info.
28766 (lto_read_body): Likewise.
28767 (lto_input_toplevel_asms): Likewise.
28769 2013-02-08 Michael Matz <matz@suse.de>
28771 PR tree-optimization/52448
28772 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28773 (nt_call_phase): New static.
28774 (add_or_mark_expr): Only mark accesses with newer phase than any
28776 (nonfreeing_call_p): New.
28777 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28778 (nt_fini_block): Keep blocks marked as visited.
28779 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28781 2013-02-08 Richard Biener <rguenther@suse.de>
28783 * ira.c (ira): Free broken dominator information.
28785 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28787 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28789 2013-02-08 Marek Polacek <polacek@redhat.com>
28791 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28793 2013-02-08 Richard Biener <rguenther@suse.de>
28795 PR middle-end/56181
28796 * cfgloop.h (flow_loops_find): Adjust.
28797 (bb_loop_header_p): Declare.
28798 * cfgloop.c (bb_loop_header_p): New function split out from ...
28799 (flow_loops_find): ... here. Adjust function signature,
28800 support incremental loop structure update.
28801 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28802 * cfgloopmanip.c (fix_loop_structure): Move ...
28803 * loop-init.c (fix_loop_structure): ... here.
28804 (apply_loop_flags): Split out from ...
28805 (loop_optimizer_init): ... here.
28806 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28807 in incremental mode, only remove dead loops here.
28809 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28812 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28813 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28814 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28815 (*round<mode>3.libgcc): New insns for fixed-modes.
28816 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28817 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28818 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28819 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28820 implementations. Define to __builtin_avr_absFX,
28821 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28822 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28823 __builtin_avr_countlsFX, respectively.
28824 * config/avr/avr-c.c (target.h): Include it.
28825 (enum avr_builtin_id): New enum.
28826 (avr_resolve_overloaded_builtin): New static function.
28827 (avr_register_target_pragmas): Use it to set
28828 targetm.resolve_overloaded_builtin.
28829 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28830 tree nodes used by DEF_BUILTIN.
28831 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28832 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28833 <AVR_BUILTIN_xxBITS>: Same.
28835 2013-02-08 Richard Biener <rguenther@suse.de>
28837 * cfgloop.c (verify_loop_structure): Properly handle
28838 a loop exiting to another loop header.
28839 * ira-int.h (ira_loops): Remove.
28840 * ira.c (ira_loops): Remove.
28841 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28842 (do_reload): Use loop_optimizer_finalize.
28843 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28844 number_of_loops to access the loop tree.
28845 (more_one_region_p): Likewise.
28846 (finish_loop_tree_nodes): Likewise.
28847 (rebuild_regno_allocno_maps): Likewise.
28848 (mark_loops_for_removal): Likewise.
28849 (mark_all_loops_for_removal): Likewise.
28850 (remove_unnecessary_regions): Likewise.
28851 (ira_build): Likewise.
28852 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28854 2013-02-08 Richard Biener <rguenther@suse.de>
28856 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28857 * ipa-pure-const.c (analyze_function): Avoid calling
28858 mark_irreducible_loops twice.
28859 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28861 2013-02-07 David S. Miller <davem@davemloft.net>
28863 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28865 * var-tracking.c (vt_add_function_parameter): Test the presence of
28866 HAVE_window_save properly and do not remap argument registers when
28867 we have a leaf function.
28869 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28872 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28874 * config/i386/i386.c (ix86_print_operand): Ditto.
28876 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28878 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28880 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28882 PR rtl-optimization/56225
28883 * lra-constraints.c (process_alt_operands): Check that reload hard
28884 reg can hold value for strict_low_part.
28886 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28889 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28890 dwarf2out_end_function.
28891 (in_first_function_p, maybe_at_text_label_p,
28892 first_loclabel_num_not_at_text_label): New variables.
28893 (dwarf2out_var_location): In the first function find out
28894 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28895 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28898 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28900 PR rtl-optimization/56178
28901 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28902 SUBREG of a register. Tidy up related block of code.
28903 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28904 note if the source is a register or a SUBREG of a register.
28906 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28909 * config/rs6000/rs6000.md (ptrm): New mode attr.
28910 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28911 call_value_indirect_aix<pttrsize>,
28912 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28915 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28917 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28918 if -bnortl. Convert to strcmp and strncmp.
28920 2013-02-07 Alan Modra <amodra@gmail.com>
28923 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28924 addresses won't wrap when offsetting.
28925 (rs6000_secondary_reload): Provide secondary reloads needed for
28926 wrapping LO_SUM addresses.
28928 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28930 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28931 MACH, just __MACH__.
28933 2013-02-06 Richard Biener <rguenther@suse.de>
28935 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28936 instead of calling fix_loop_structure.
28938 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28940 PR middle-end/56217
28941 * omp-low.c (use_pointer_for_field): Return false if
28942 lower_send_shared_vars doesn't generate any copy-out code.
28944 2013-02-06 Tom de Vries <tom@codesourcery.com>
28946 PR rtl-optimization/56131
28947 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28948 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28949 of the label is NULL. Add comment.
28951 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28953 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28956 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28957 (STATIC_LIBTSAN_LIBS): Likewise.
28958 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28959 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28960 is defined, don't add anything else beyond that.
28961 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28962 (LINK_COMMAND_SPEC): Use them.
28964 PR tree-optimization/56205
28965 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28966 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28967 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28969 2013-02-05 Richard Biener <rguenther@suse.de>
28971 PR tree-optimization/53342
28972 PR tree-optimization/53185
28973 * tree-vectorizer.h (vect_check_strided_load): Remove.
28974 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28975 not disallow peeling for vectorized strided loads.
28976 (vect_check_strided_load): Make static and simplify.
28977 (vect_analyze_data_refs): Adjust.
28978 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28979 correctly when vectorizing strided loads.
28981 2013-02-05 Richard Biener <rguenther@suse.de>
28983 * doc/install.texi: Refer to ISL, not PPL.
28985 2013-02-05 Jan Hubicka <jh@suse.cz>
28987 PR tree-optimization/55789
28988 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28990 2013-02-05 Jan Hubicka <jh@suse.cz>
28992 PR tree-optimization/55789
28993 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28994 the dead call anyway.
28996 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28999 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
29001 2013-02-04 Alexander Potapenko <glider@google.com>
29002 Jack Howarth <howarth@bromo.med.uc.edu>
29003 Jakub Jelinek <jakub@redhat.com>
29006 * config/darwin.c (sort_ctor_records): Stabilized qsort
29007 on constructor priority by using original position.
29008 (finalize_ctors): New routine to sort constructors by
29009 priority before use in assemble_integer.
29010 (machopic_asm_out_constructor): Use finalize_ctors if needed.
29012 2013-02-04 Jakub Jelinek <jakub@redhat.com>
29015 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
29016 about visibility on artificial decls.
29017 * config/sol2.c (solaris_assemble_visibility): Likewise.
29019 2013-02-04 Kai Tietz <ktietz@redhat.com>
29022 * config/i386/i386.c (function_value_ms_64): Add additional valtype
29023 argument and improve checking of return-argument types for 16-byte
29025 (ix86_function_value_1): Add additional valtype argument on call
29026 of function_value_64.
29027 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
29028 handling infunction_value_64 function.
29030 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
29032 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
29034 2013-02-04 Richard Biener <rguenther@suse.de>
29036 PR tree-optimization/56188
29037 * tree-ssa-structalias.c (label_visit): Consider case with
29038 initially non-empty points-to set.
29039 (perform_var_substitution): Dump node mapping and clean up.
29041 2013-02-04 Richard Guenther <rguenther@suse.de>
29044 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
29045 node prevail as last resort.
29046 (lto_symtab_merge_decls): Remove guard on LTRANS here.
29047 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
29049 2013-02-04 Richard Biener <rguenther@suse.de>
29051 PR tree-optimization/56113
29052 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
29054 (equiv_class_lookup_or_add): ... this.
29055 (label_visit): Adjust and fix error in previous patch.
29056 (perform_var_substitution): Adjust.
29058 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
29060 * config/sh/divtab.c: Fix formatting and comments throughout the file.
29061 * config/sh/sh4-300.md: Likewise.
29062 * config/sh/sh4a.md: Likewise.
29063 * config/sh/constraints.md: Likewise.
29064 * config/sh/sh.md: Likewise.
29065 * config/sh/netbsd-elf.h: Likewise.
29066 * config/sh/predicates.md: Likewise.
29067 * config/sh/sh-protos.h: Likewise.
29068 * config/sh/ushmedia.h: Likewise.
29069 * config/sh/linux.h: Likewise.
29070 * config/sh/sh.c: Likewise.
29071 * config/sh/superh.h: Likewise.
29072 * config/sh/elf.h: Likewise.
29073 * config/sh/sh4.md: Likewise.
29074 * config/sh/sh.h: Likewise.
29076 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29078 * config/pa/constraints.md: Adjust unused letters. Change "T"
29079 constraint to match_test floating_point_store_memory_operand().
29080 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
29081 (base14_operand): New.
29082 (floating_point_store_memory_operand): New.
29083 (integer_store_memory_operand): Revise to use base14_operand and
29084 reg_plus_base_memory_operand.
29085 (move_dest_operand): Allow symbolic_memory_operands.
29086 (symbolic_memory_operand): Check for LO_SOM.
29087 (symbolic_operand): Change default case to break.
29088 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
29089 CONST_DOUBLE values to be reloaded by putting them into memory when
29090 the destination is a floating point register.
29091 (movdf): Remove code to handle CONST_DOUBLE.
29093 (reload_indf_r1): New.
29094 (reload_insf_r1): New.
29095 Consistently use "Q" and "T" constraints with integer and floating
29096 point move instructions, respectively.
29097 (movdi): Remove FAIL.
29098 Change predicate for source operand unamed DImode move from
29099 general_operand to move_src_operand.
29100 (umulsidi3): Change predicate for destination operand to
29102 Likewise for similar unamed patterns.
29103 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
29104 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
29105 (hppa_legitimize_address): Simplify mask calculation.
29106 (pa_emit_move_sequence): Revised handling of secondary reloads from
29107 REG+D addresses for floating point loads and stores. Directly handle
29108 loading CONST0_RTX (mode) to a floating point register.
29109 (pa_secondary_reload): Handle reloading DF and SFmode constant values
29110 to floating point registers. Don't restrict secondary reloads to
29111 floating point registers to integer modes. Revise some comments and
29113 (TARGET_LEGITIMATE_ADDRESS_P): Define.
29114 (pa_legitimate_address_p): New.
29115 (pa_legitimize_reload_address): New.
29116 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
29117 (STRICT_REG_OK_FOR_BASE_P): New.
29118 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
29119 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
29121 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
29122 Andrew Dixie <andrewd@gentrack.com>
29124 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
29127 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
29129 * expmed.c (extract_bit_field_1): Pass the full width of the
29130 structure to get_best_reg_extraction_insn.
29132 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
29135 * configure.ac (use_cxa_atexit): Add AIX.
29136 * configure: Regenerate.
29138 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
29140 2013-02-01 Jakub Jelinek <jakub@redhat.com>
29143 * final.c (need_profile_function): New variable.
29144 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
29145 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
29146 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
29147 notes, targetm.asm_out.function_prologue doesn't emit anything,
29148 HAVE_prologue and profiler should be emitted before prologue,
29149 set need_profile_function instead of emitting it.
29150 (final_scan_insn): If need_profile_function, emit
29151 profile_function on the first NOTE_INSN_BASIC_BLOCK or
29152 NOTE_INSN_FUNCTION_BEG note.
29154 2013-02-01 Richard Henderson <rth@redhat.com>
29156 * config/rs6000/rs6000.md (smulditi3): New.
29159 * config/alpha/alpha.md (umulditi3): New.
29161 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
29163 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
29164 (ASM_OUTPUT_ALIGNED_LOCAL): New.
29166 2013-02-01 Richard Biener <rguenther@suse.de>
29168 PR tree-optimization/56113
29169 * tree-ssa-structalias.c (label_visit): Reduce work for
29170 single-predecessor nodes.
29172 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
29174 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
29175 range isn't testing for zero.
29177 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
29179 PR middle-end/56113
29180 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
29182 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
29183 Nick Clifton <nickc@redhat.com>
29185 * config/v850/constraints.md (Q): Define as a memory constraint.
29186 * config/v850/predicates.md (label_ref_operand): New predicate.
29187 (e3v5_shift_operand): New predicate.
29188 (ior_operator): New predicate.
29189 * config/v850/t-v850: Add e3v5 multilib.
29190 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
29191 (v850_gen_movdi): Prototype.
29192 * config/v850/v850.c: Add support for e3v5 architecture.
29193 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
29195 (construct_save_jarl): Add e3v5 long JARL support.
29196 (v850_adjust_insn_length): New function. Adjust length of call
29197 insns when using e3v5 instructions.
29198 (v850_gen_movdi): New function: Generate instructions to move a
29200 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
29201 (CPP_SPEC): Define __v850e3v5__ as appropriate.
29202 (TARGET_USE_FPU): Enable for e3v5.
29203 (CONST_OK_FOR_W): New macro.
29204 (ADJUST_INSN_LENGTH): Define.
29205 * config/v850/v850.md (UNSPEC_LOOP): Define.
29206 (attr cpu): Add v850e3v5.
29207 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
29208 (movdi): New pattern.
29209 (movdi_internal): New pattern.
29210 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
29211 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
29212 (cstoresf4): Likewise.
29213 (cstoredf4): Likewise.
29214 (insv): New pattern.
29215 (rotlso3_a): New pattern.
29216 (rotlsi3_b): New pattern
29217 (rotlsi3_v850e3v5): New pattern.
29218 (doloop_begin): New pattern.
29219 (fix_loop_counter): New pattern.
29220 (doloop_end): New pattern.
29221 (branch_normal): Add e3v5 long branch support.
29222 (branch_invert): Likewise.
29223 (branch_z_normal): Likewise.
29224 (branch_z_invert): Likewise.
29225 (branch_nz_normal): Likewise.
29226 (branch_nz_invert): Likewise.
29227 (call_internal_short): Add e3v5 register-indirect JARL support.
29228 (call_internal_long): Likewise.
29229 (call_value_internal_short): Likewise.
29230 (call_value_internal_long): Likewise.
29231 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
29232 (mloop): New option.
29233 * config.gcc: Add support for configuring v840e3v5 target.
29234 * doc/invoke.texi: Document new v850 specific command line options.
29236 2013-01-31 Paul Koning <ni1d@arrl.net>
29240 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
29241 children if parent is a class.
29242 (prune_unused_types_prune): Don't add DW_AT_declaration.
29244 2013-01-31 Richard Biener <rguenther@suse.de>
29246 PR tree-optimization/56157
29247 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
29248 match up operand with SLP child.
29250 2013-01-31 Jason Merrill <jason@redhat.com>
29253 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
29254 parameters the first time.
29255 (gen_scheduled_generic_parms_dies): Check completeness here.
29257 2013-01-31 Richard Biener <rguenther@suse.de>
29259 PR middle-end/53073
29260 * common.opt (faggressive-loop-optimizations): New flag,
29261 enabled by default.
29262 * doc/invoke.texi (faggressive-loop-optimizations): Document.
29263 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
29264 infer_loop_bounds_from_undefined by it.
29266 2013-01-31 Richard Biener <rguenther@suse.de>
29268 PR tree-optimization/56150
29269 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
29270 visit virtual operands.
29271 (find_uses_to_rename_bb): Likewise.
29273 2013-01-31 Richard Biener <rguenther@suse.de>
29275 PR tree-optimization/56150
29276 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
29277 mixed store non-store stmts.
29279 2013-01-30 Jakub Jelinek <jakub@redhat.com>
29282 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
29283 LIBASAN_EARLY_SPEC is defined.
29284 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
29285 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
29287 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
29290 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
29291 invalid args instead of ICEing on it.
29292 (ix86_valid_target_attribute_tree): Return error_mark_node if
29293 ix86_valid_target_attribute_inner_p failed.
29294 (ix86_valid_target_attribute_p): Return false only if
29295 ix86_valid_target_attribute_tree returned error_mark_node. Allow
29296 target("default") attribute.
29297 (sorted_attr_string): Change argument from const char * to tree,
29298 merge in all target attribute arguments rather than just one.
29299 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
29300 instead of free. Avoid using strcat.
29301 (ix86_mangle_function_version_assembler_name): Mangle
29302 target("default") as if no target attribute is present. Adjust
29303 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
29304 instead of xmalloc and XDELETEVEC instead of free.
29305 (ix86_function_versions): Don't return true if one of the decls
29306 doesn't have target attribute. If they don't and one of the decls
29307 is DECL_FUNCTION_VERSIONED, report an error. Adjust
29308 sorted_attr_string caller. Use XDELETEVEC instead of free.
29309 (ix86_supports_function_versions): Remove.
29310 (make_name): Fix up formatting.
29311 (make_dispatcher_decl): Remove resolver_name and its initialization.
29312 Avoid leaking memory.
29313 (is_function_default_version): Return true if there is
29314 target("default") attribute rather than no target attribute at all.
29315 (make_resolver_func): Avoid leaking memory.
29316 (ix86_generate_version_dispatcher_body): Likewise.
29317 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
29318 * target.def (supports_function_versions): Remove.
29319 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
29320 * doc/tm.texi: Regenerated.
29322 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
29324 PR rtl-optimization/56144
29325 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
29326 for values with side effects.
29328 2013-01-30 Richard Biener <rguenther@suse.de>
29330 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
29331 (sparseset_pop): Likewise.
29332 * cfganal.c (compute_idf): Likewise. Increase work-stack size
29333 to be able to use quick_push in the worker loop.
29335 2013-01-30 Marek Polacek <polacek@redhat.com>
29337 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
29339 2013-01-30 Richard Biener <rguenther@suse.de>
29342 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
29344 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
29346 PR tree-optimization/56064
29347 * fixed-value.c (fixed_from_double_int): New function.
29348 * fixed-value.h (fixed_from_double_int): New prototype.
29349 (const_fixed_from_double_int): New static inline function.
29350 * fold-const.c (native_interpret_fixed): New static function.
29351 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
29352 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
29353 (native_encode_fixed): New static function.
29354 (native_encode_expr) <FIXED_CST>: Use it.
29355 (native_interpret_int): Move double_int worker code to...
29356 * double-int.c (double_int::from_buffer): ...this new static method.
29357 * double-int.h (double_int::from_buffer): Prototype it.
29359 2013-01-30 Richard Biener <rguenther@suse.de>
29361 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
29362 New pointer-map and obstack.
29363 (init_alias_vars): Allocate pointer-map and obstack.
29364 (delete_points_to_sets): Free them.
29365 (find_what_var_points_to): Cache result.
29366 (find_what_p_points_to): Adjust for changed interface of
29367 find_what_var_points_to.
29368 (compute_points_to_sets): Likewise.
29369 (ipa_pta_execute): Likewise.
29371 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
29373 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
29374 * configure: Regenerate.
29375 * config.in: Regenerate.
29376 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
29377 #nobits/#progbits if supported.
29379 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
29382 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
29383 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
29384 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
29386 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29388 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
29389 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
29391 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29393 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
29394 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
29396 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29398 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
29400 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
29401 * config/arm/cortex-a7.md: New bypasses using
29402 arm_mac_accumulator_is_result.
29404 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29406 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
29407 (cortex_a7_neon_mla): Likewise.
29408 (cortex_a7_fpfmad): New reservation.
29409 (cortex_a7_fpmacs): Use ffmas and update required units.
29410 (cortex_a7_fpmuld): Update required units and latency.
29411 (cortex_a7_fpmacd): Likewise.
29412 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
29413 (cortex_a7_neon). Likewise.
29414 (bypass) Update participating units.
29416 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29418 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
29419 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
29421 * config/arm/vfp11.md (vfp_farith): Use ffmas.
29422 (vfp_fmul): Use ffmad.
29423 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
29424 (cortex_r4_fmacd): Use ffmad.
29425 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
29426 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
29427 (cortex_a9_fmacd): Use ffmad.
29428 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
29429 (cortex_a8_vfp_macd): Use ffmad.
29430 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
29431 (cortex_a5_fpmacd): Use ffmad.
29432 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
29433 (cortex_a15_vfp_macd): Use ffmad.
29434 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29436 2013-01-29 Jason Merrill <jason@redhat.com>
29439 * varasm.c (default_assemble_visibility): Don't warn about
29440 visibility on artificial decls.
29442 2013-01-29 Richard Biener <rguenther@suse.de>
29444 PR tree-optimization/56113
29445 * tree-ssa-structalias.c (equiv_class_lookup): Also return
29447 (label_visit): Free duplicate bitmaps and record the leader instead.
29448 (perform_var_substitution): Adjust.
29450 2013-01-29 Richard Biener <rguenther@suse.de>
29452 PR tree-optimization/55270
29453 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
29454 the CFG, schedule loops for fixup.
29456 2013-01-29 Nick Clifton <nickc@redhat.com>
29458 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
29461 2013-01-28 Leif Ekblad <leif@rdos.net>
29463 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
29464 * config/i386/i386.h (TARGET_RDOS): New macro.
29465 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
29466 * config/i386/i386.c (ix86_option_override_internal): For 64bit
29467 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
29468 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
29469 DEFAULT_LARGE_SECTION_THRESHOLD.
29470 * config/i386/i386.md (R14_REG, R15_REG): New constants.
29471 * config/i386/rdos.h: New file.
29472 * config/i386/rdos64.h: New file.
29474 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
29477 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
29480 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29482 PR rtl-optimization/56117
29483 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
29484 call cselib_lookup_from_insn on the MEM before calling
29485 add_insn_mem_dependence.
29487 2013-01-28 Richard Biener <rguenther@suse.de>
29489 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
29490 to a stmt that didn't have one.
29491 (copy_phis_for_bb): Likewise for PHI arguments.
29492 (copy_debug_stmt): Likewise for debug stmts.
29494 2013-01-28 Richard Biener <rguenther@suse.de>
29496 PR tree-optimization/56034
29497 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
29498 (partition_builtin_p): Adjust.
29499 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
29500 it is the last partition.
29501 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
29502 up the vertex for the definition.
29503 (classify_partition): Classify whether a partition is a
29504 PKIND_REDUCTION, thus has uses outside of the loop.
29505 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
29506 Merge all PKIND_REDUCTION partitions into the last partition.
29507 (tree_loop_distribution): Seed partitions from reductions as well.
29509 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29511 PR tree-optimization/56125
29512 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
29513 pow(x,c) into sqrt(x) * powi(x, n/2) or
29514 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
29515 optimizing for size.
29516 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
29517 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
29520 PR tree-optimization/56094
29521 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
29522 to UNKNOWN_LOCATION while gimplifying expr.
29524 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
29527 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
29528 operand 0 in movabs insn template for -masm=intel asm alternative.
29529 (*movabs<mode>_2): Ditto for operand 1.
29531 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
29534 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
29537 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29539 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
29542 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29544 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
29545 Undef to avoid warning.
29547 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
29549 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
29550 * configure: Regenerate.
29552 2013-01-25 Jakub Jelinek <jakub@redhat.com>
29554 PR tree-optimization/56098
29555 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
29556 for stmts with volatile ops.
29557 (cond_store_replacement): Don't optimize if assign has volatile ops.
29558 (cond_if_else_store_replacement_1): Don't optimize if either
29559 then_assign or else_assign have volatile ops.
29560 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
29563 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29565 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
29567 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29569 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
29570 missing ':' in asm example.
29572 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
29574 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
29575 entries into lane and laneq entries.
29576 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
29577 Remove AdvSIMD scalar modes.
29578 (aarch64_sq<r>dmulh_laneq<mode>): New.
29579 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
29581 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
29582 builtin implementations to relfect changes in RTL in aarch64-simd.md.
29583 * config/aarch64/iterators.md (VCOND): New.
29586 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29589 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
29590 Add NULL LIBNAME argument to existing definitions.
29591 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
29592 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
29593 * config/avr/avr.c (DEF_BUILTIN): Same.
29594 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
29595 (avr_expand_builtin): Expand to a vanilla call if a libgcc
29596 implementation is available (DECL_ASSEMBLER_NAME is set).
29597 (avr_fold_absfx): New static function.
29598 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
29599 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
29600 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
29601 AVR_BUILTIN_ABSLLK.
29602 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
29603 (abshk, absk, abslk, absllk): Provide as static inline functions.
29605 2013-01-25 Marek Polacek <polacek@redhat.com>
29607 PR tree-optimization/56035
29608 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
29610 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
29612 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
29613 (*movtf_internal_rex64): Add (!o,C) alternative
29614 (*movxf_internal_rex64): Ditto.
29615 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
29617 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
29619 * doc/invoke.texi: fix typo.
29620 * doc/objc.texi: fix typo.
29622 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
29624 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
29625 for the first two alternatives.
29627 2013-01-24 Diego Novillo <dnovillo@google.com>
29629 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
29630 (ggc-zone.o): Remove.
29631 * configure.ac: Remove option --with-gc.
29632 * configure: Re-generate.
29633 * doc/install.texi: Remove documentation for --with-gc.
29634 * gengtype.c (write_enum_defn): Remove. Update all users.
29635 (write_Types_process_field): Remove generation of gt_e_* argument.
29636 (output_type_enum): Remove. Update all users.
29637 (write_enum_defn): Remove. Update all users.
29638 (enum alloc_zone): Remove. Update all users.
29639 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
29640 * ggc-common.c (ggc_splay_alloc): Remove first argument.
29641 Update all callers.
29642 (struct ptr_data): Remove field TYPE. Update all users.
29643 (gt_pch_note_object): Remove argument TYPE. Update all users.
29644 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
29646 * ggc-none.c (ggc_alloc_typed_stat): Remove.
29647 (struct alloc_zone): Remove.
29648 (ggc_internal_alloc_zone_stat): Remove.
29649 (ggc_internal_cleared_alloc_zone_stat): Remove.
29650 * ggc-page.c (ggc_alloc_typed_stat): Remove.
29651 (ggc_pch_count_object): Remove last argument. Update all users.
29652 (ggc_pch_alloc_object): Remove last argument. Update all users.
29653 (struct alloc_zone): Remove.
29654 * ggc-zone.c: Remove.
29655 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
29656 (struct alloc_zone): Remove.
29657 (ggc_alloc_typed_stat): Remove.
29658 (ggc_alloc_typed): Remove.
29659 (ggc_splay_alloc): Remove first argument.
29660 (rtl_zone): Remove. Update all users.
29661 (tree_zone): Remove. Update all users.
29662 (tree_id_zone): Remove. Update all users.
29663 (ggc_internal_zone_alloc_stat): Remove. Update all users.
29664 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
29665 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
29666 * tree-ssanames.c: Remove references to zone allocator in comments.
29668 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
29670 * config/avr/avr.c (avr_out_fract): Make register numbers that
29671 might be outside of source operand signed.
29673 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
29675 * config/i386/constraints.md (Yf): New constraint.
29676 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
29677 of f constraint to conditionaly disable x87 register preferences.
29678 (*movdf_internal): Ditto.
29679 (*movsf_internal): Ditto.
29681 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
29683 PR inline-asm/55934
29684 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
29685 that have operands with impossible constraints.
29686 Add a FIXME for a speed-up opportunity.
29687 * lra-constraints.c (process_alt_operands): Verify that a class
29688 selected from constraints on asms is valid for the operand mode.
29689 (curr_insn_transform): Remove incorrect comment.
29691 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
29693 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
29694 TOC operand is a valid symbol ref in the constant pool.
29696 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29698 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
29700 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
29703 * config/avr/stdfix.h: New file.
29704 * t-avr (stdfix-gcc.h): New rule to build it.
29705 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
29707 2013-01-23 Kostya Serebryany <kcc@google.com>
29709 * config/darwin.h: remove dependency on
29710 CoreFoundation (asan on Mac OS).
29712 2013-01-23 Jakub Jelinek <jakub@redhat.com>
29715 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
29716 instead of cmpdi_operand for first comparison operand.
29717 Don't assert that comparison operands aren't both constants.
29719 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29721 * doc/install.texi (Downloading the Source): Update references to
29722 downloading separate components.
29724 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29726 * doc/extend.texi (__int128): Improve grammar.
29728 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29731 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29732 alternative to (o,r).
29733 (*movdi_internal_rex64): Remove (!o,n) alternative.
29734 (DImode immediate->memory splitter): Remove.
29735 (DImode immediate->memory peephole2): Remove.
29736 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29737 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29738 alternative to (!o,*r).
29739 (*movtf_internal_sse): New pattern.
29740 (*movxf_internal_rex64): New pattern.
29741 (*movxf_internal): Disable for TARGET_64BIT.
29742 (*movdf_internal_rex64): Remove (!o,F) alternative.
29744 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29746 PR middle-end/56074
29747 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29748 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29749 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29750 stmt locations where LOCATION_LOCUS of the stmt location is
29751 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29754 * config/i386/i386.md (UNSPEC_STOS): New.
29755 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29756 *strsetqi_1): Add UNSPEC_STOS.
29758 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29761 * doc/invoke.texi: Remove left over -Wsynth example.
29763 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29765 PR tree-optimization/56051
29766 * fold-const.c (fold_binary_loc): Don't fold
29767 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29768 a narrowing conversion, or widening conversion from signed
29771 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29773 PR rtl-optimization/56023
29774 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29775 dependent on debug instruction.
29777 2013-01-21 Martin Jambor <mjambor@suse.cz>
29779 PR middle-end/56022
29780 * function.c (allocate_struct_function): Call
29781 invoke_set_current_function_hook earlier.
29783 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29785 * reload1.c (init_reload): Only initialize reload_obstack
29786 during the first call.
29788 2013-01-21 Marek Polacek <polacek@redhat.com>
29790 * cfgloop.c (verify_loop_structure): Fix up grammar.
29792 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29794 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29795 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29797 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29800 * config/arm/marvell-pj4.md: Update copyright year.
29801 Fix up use of alu to alu_reg and simple_alu_imm.
29803 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29805 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29807 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29810 * lra-constraints.c (curr_insn_transform): Don't reuse original
29811 insn for secondary memory move when memory mode should be different.
29813 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29815 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29816 atomic_storedi_1): New patterns.
29818 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29820 btver2 pipeline descriptions.
29821 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29823 * config/i386/i386.md (btver2_decode): New type attributes.
29824 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29826 * config/i386/btver2.md: New file describing btver2 pipelines.
29828 2013-01-19 Andrew Pinski <apinski@cavium.com>
29830 PR tree-optimization/52631
29831 * tree-ssa-sccvn (visit_use): Before looking up the original
29832 statement, try looking up the simplified expression.
29834 2013-01-19 Anthony Green <green@moxielogic.com>
29836 * config/moxie/moxie.c (moxie_expand_prologue): Set
29837 current_function_static_stack_size.
29839 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29841 PR tree-optimization/56029
29842 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29843 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29845 2013-01-18 Sharad Singhai <singhai@google.com>
29847 PR tree-optimization/55995
29848 * dumpfile.c (dump_loc): Print location only if available.
29849 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29851 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29854 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29855 secondary memory move.
29856 (inherit_reload_reg): Use rclass instead of cl for
29857 check_secondary_memory_needed_p.
29859 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29861 PR middle-end/56015
29862 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29863 the case where writing real complex part of target modifies op1.
29865 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29867 * config/aarch64/aarch64-simd.md
29868 (aarch64_vcond_internal<mode>): Handle unordered cases.
29869 * config/aarch64/iterators.md (v_cmp_result): New.
29871 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29872 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29874 * config/arm/marvell-pj4.md: New file.
29875 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29876 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29877 (generic_vfp): Likewise.
29878 * config/arm/arm-cores.def: Add marvell-pj4.
29879 * config/arm/arm-tune.md: Regenerate.
29880 * config/arm/arm-tables.opt: Regenerate.
29881 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29882 * doc/invoke.texi: Document marvell-pj4.
29884 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29886 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29888 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29893 * var-tracking.c (negative_power_of_two_p): New.
29894 (global_get_addr_cache, local_get_addr_cache): New.
29895 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29896 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29898 (vt_stack_offset_p): Remove.
29899 (vt_canon_true_dep): Always canonicalize loc's address.
29900 (clobber_overlapping_mems): Make sure we have a MEM.
29901 (local_get_addr_clear_given_value): New.
29902 (val_reset): Clear local cached entries.
29903 (compute_bb_dataflow): Create and release the local cache.
29904 Disable duplicate MEMs clobbering.
29905 (emit_notes_in_bb): Clobber MEMs likewise.
29906 (vt_emit_notes): Create and release the local cache.
29907 (vt_initialize, vt_finalize): Create and release the global
29908 cache, respectively.
29909 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29911 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29913 PR libmudflap/53359
29914 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29915 not found in the symtab.
29917 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29920 PR rtl-optimization/55547
29921 PR rtl-optimization/53827
29924 * alias.c (offset_overlap_p): New, factored out of...
29925 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29926 the conservative special case for symbolic constants. Don't
29927 adjust zero sizes on alignment.
29929 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29931 PR rtl-optimization/52573
29932 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29933 REG_UNUSED for the same register.
29935 2013-01-17 Richard Biener <rguenther@suse.de>
29936 Marek Polacek <polacek@redhat.com>
29938 PR rtl-optimization/55833
29939 * loop-unswitch.c (unswitch_loops): Move loop verification...
29940 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29941 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29942 Set it to true when we're removing a loop from hierarchy tree in
29943 an irreducible region.
29944 (fix_bb_placements): Adjust caller.
29945 (fix_loop_placements): Likewise.
29947 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29949 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29950 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29951 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29952 Remove ID. Adjust comments.
29953 * config/avr/avr-c.c (avr_builtin_name): Remove.
29954 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29955 * config/avr/avr.c (avr_tolower): New static function.
29956 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29957 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29958 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29961 2013-01-17 Jan Hubicka <jh@suse.cz>
29963 PR tree-optimization/55273
29964 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29966 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29969 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29970 store through atomic_store<mode>_1.
29971 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29973 2013-01-17 Martin Jambor <mjambor@suse.cz>
29975 PR tree-optimizations/55264
29976 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29977 for virtual methods.
29978 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29979 virtual methods before inlining is over.
29980 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29982 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29985 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29987 PR rtl-optimization/56005
29988 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29989 pending reads for prefetch.
29991 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29993 * config/aarch64/aarch64.md
29994 (*cstoresi_neg_uxtw): New pattern.
29995 (*cmovsi_insn_uxtw): New pattern.
29996 (*<optab>si3_uxtw): New pattern.
29997 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29998 (*<optab>si3_insn_uxtw): New pattern.
29999 (*bswapsi2_uxtw): New pattern.
30001 2013-01-16 Richard Biener <rguenther@suse.de>
30003 * tree-inline.c (tree_function_versioning): Remove set but
30004 never used variable.
30006 2013-01-16 Richard Biener <rguenther@suse.de>
30008 PR tree-optimization/55964
30009 * tree-flow.h (rename_variables_in_loop): Remove.
30010 (rename_variables_in_bb): Likewise.
30011 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
30012 (copy_loop_before): Adjust and delete update-ssa status.
30013 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
30014 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
30015 (rename_variables_in_loop): Remove.
30016 (slpeel_update_phis_for_duplicate_loop): Likewise.
30017 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
30018 use available cfg machinery instead of duplicating it.
30019 Update PHI nodes and perform poor-mans SSA update here.
30020 (slpeel_tree_peel_loop_to_edge): Adjust.
30022 2013-01-16 Richard Biener <rguenther@suse.de>
30024 PR tree-optimization/54767
30025 PR tree-optimization/53465
30026 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
30027 (vrp_visit_phi_node): For PHI arguments coming via backedges
30028 drop all symbolical range information.
30029 (execute_vrp): Compute backedges.
30031 2013-01-16 Richard Biener <rguenther@suse.de>
30033 * doc/install.texi: Update CLooG and ISL requirements to
30036 2013-01-16 Christian Bruel <christian.bruel@st.com>
30039 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
30040 (broken_move): Handle UNSPECV_SP_SWITCH_B.
30041 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
30043 2013-01-16 DJ Delorie <dj@redhat.com>
30045 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
30046 (UNSPECV_SP_SWITCH_E): New.
30047 (sp_switch_1): Change to an unspec.
30048 (sp_switch_2): Change to an unspec. Don't use post-inc when we
30051 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
30053 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
30054 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
30055 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
30056 (expand_mem_thread_fence): Ditto.
30057 (expand_mem_signal_fence): Ditto.
30058 (expand_atomic_load): Ditto.
30059 (expand_atomic_store): Ditto.
30061 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
30063 PR rtl-optimization/55547
30064 PR rtl-optimization/53827
30067 * alias.c (memrefs_conflict_p): Set sizes to negative after
30070 2013-01-15 Jakub Jelinek <jakub@redhat.com>
30073 * function.c (thread_prologue_and_epilogue_insns): Always
30074 add crtl->drap_reg to set_up_by_prologue.set, even if
30075 stack_realign_drap is false.
30077 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
30079 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
30080 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
30081 *call): Fix indention.
30083 2013-01-15 Tom de Vries <tom@codesourcery.com>
30086 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
30089 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
30091 PR rtl-optimization/55153
30092 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
30094 2013-01-15 Martin Jambor <mjambor@suse.cz>
30096 PR tree-optimization/55920
30097 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
30098 accesses as grp_to_be_debug_replaced.
30100 2013-01-15 Jakub Jelinek <jakub@redhat.com>
30102 PR tree-optimization/55920
30103 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
30104 there is non-useless type conversion needed from debug rhs to lhs,
30105 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
30107 2013-01-15 Joseph Myers <joseph@codesourcery.com>
30108 Mikael Pettersson <mikpe@it.uu.se>
30111 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
30113 (ASM_OUTPUT_CASE_LABEL): Remove.
30114 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
30115 * final.c (shorten_branches): Update alignment of labels before
30116 jump tables if CASE_VECTOR_SHORTEN_MODE.
30118 2013-01-15 Richard Biener <rguenther@suse.de>
30121 * system.h: Do not include gmp.h for building host tools.
30123 2013-01-15 Richard Biener <rguenther@suse.de>
30125 PR middle-end/55882
30126 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
30127 account for bitpos when computing alignment.
30129 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
30131 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
30132 (ix86_target_macros_internal): Likewise.
30134 * config/i386/i386.c (m_CORE2I7): Removed.
30135 (m_CORE_HASWELL): New macro.
30136 (m_CORE_ALL): Likewise.
30137 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
30138 (initial_ix86_arch_features): Likewise.
30139 (processor_target_table): Initializations for Core avx2.
30140 (cpu_names): New names "core-avx2".
30141 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
30142 PROCESSOR_CORE_HASWELL.
30143 (ix86_issue_rate): New case.
30144 (ia32_multipass_dfa_lookahead): Likewise.
30145 (ix86_sched_init_global): Likewise.
30147 * config/i386/i386.h (TARGET_HASWELL): New macro.
30148 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
30149 (processor_type): New PROCESSOR_HASWELL.
30151 2013-01-15 Jakub Jelinek <jakub@redhat.com>
30153 PR tree-optimization/55955
30154 * tree-vect-loop.c (vectorizable_reduction): Give up early on
30155 *SHIFT_EXPR and *ROTATE_EXPR codes.
30157 PR tree-optimization/48766
30158 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
30159 -ftrapv disable -fwrapv.
30161 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30164 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
30165 etc. to 1 and not to __flash.
30166 Use LL suffix for __INT24_MAX__ with -mint8.
30167 Use ULL suffix for __UINT24_MAX__ with -mint8.
30169 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30171 * config/avr/avr-arch.h
30172 (struct base_arch_s): Use typedef avr_arch_t instead.
30173 (struct arch_info_s): Use typedef avr_arch_info_t instead.
30174 (struct mcu_type_s): Use typedef avr_mcu_t instead.
30175 * config/avr/avr.c: Same.
30176 * config/avr/avr-devices.c: Same.
30177 * config/avr/driver-avr.c: Same.
30178 * config/avr/gen-avr-mmcu-texi.c: Same.
30179 * config/avr/avr-mcus.def: Adjust comment.
30181 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
30183 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
30184 * config/aarch64/iterators.md (VALLDI): New.
30186 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
30187 Andi Kleen <ak@linux.intel.com>
30190 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
30191 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
30194 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30196 * config/avr/avr-stdint.h: Remove trailing blanks.
30197 * config/avr/avr-log.h: Same.
30198 * config/avr/avr-arch.h: Same.
30199 * config/avr/avr-devices.c: Same.
30200 * config/avr/avr-dimode.md: Same.
30201 * config/avr/predicates.md: Same.
30202 * config/avr/avr-c.c: Same. And fix typo.
30204 * config/avr/avr-protos.h: Same. And:
30205 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
30206 (init_cumulative_args): Rename to avr_init_cumulative_args.
30207 (expand_prologue): Rename to avr_expand_prologue.
30208 (expand_epilogue): Rename to avr_expand_epilogue.
30209 (adjust_insn_length): Rename to avr_adjust_insn_length.
30210 (notice_update_cc): Rename to avr_notice_update_cc.
30211 (final_prescan_insn): Rename to avr_final_prescan_insn.
30212 * config/avr/avr.c: Same.
30213 * config/avr/avr.h: Same.
30214 * config/avr/avr.md: Remove trailing blanks.
30215 (prologue): Use avr_expand_prologue.
30216 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
30218 2013-01-14 Richard Biener <rguenther@suse.de>
30220 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
30221 verify_location, collect_subblocks): New functions.
30222 (verify_gimple_in_cfg): Verify that locations only reference
30223 BLOCKs in the functions BLOCK tree.
30225 2013-01-14 Richard Biener <rguenther@suse.de>
30227 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
30229 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
30231 (insert_out_of_ssa_copy_on_edge): Likewise.
30232 (rewrite_close_phi_out_of_ssa): Likewise.
30233 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
30235 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
30236 propagated constants.
30237 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
30240 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30242 * config/avr/avr-modes.def: Add GPL copyright notice.
30244 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
30246 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
30247 MEMMODEL_MASK to determine memory model.
30248 (atomic_store<mode>): Ditto from operands[2].
30249 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
30251 2013-01-13 Jakub Jelinek <jakub@redhat.com>
30254 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
30255 (fold_gimple_assign): Don't call unshare_expr here.
30256 (fold_ctor_reference): Call unshare_expr.
30258 2013-01-13 Terry Guo <terry.guo@arm.com>
30260 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
30261 * doc/fragments.texi: Document MULTILIB_REUSE.
30262 * gcc.c (multilib_reuse): New internal spec.
30263 (set_multilib_dir): Also search multilib from multilib_reuse.
30264 * genmultilib (tmpmultilib3): Refactor code.
30265 (tmpmultilib4): Ditto.
30266 (multilib_reuse): New multilib argument.
30268 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
30270 * Makefile.in: Update copyright.
30272 2013-01-12 Tom de Vries <tom@codesourcery.com>
30274 PR middle-end/55890
30275 * calls.c (expand_call): Check if arg_nr is valid.
30277 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
30279 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
30280 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
30281 documentation. Add missing '__' in front of
30282 __builtin_ia32_packssdw256.
30284 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30287 * config/s390/s390.c (s390_preferred_reload_class): Do not return
30288 NO_REGS for larl operands.
30289 (s390_reload_larl_operand): Use s390_load_address instead of
30292 2013-01-11 Richard Biener <rguenther@suse.de>
30294 * tree-cfg.c (verify_node_sharing_1): Split out from ...
30295 (verify_node_sharing): ... here.
30296 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
30298 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
30300 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
30301 Substitute TREECHECKING.
30302 * configure: Regenerate.
30303 * Makefile.in (TREECHECKING): New.
30305 2013-01-11 Richard Guenther <rguenther@suse.de>
30307 PR tree-optimization/44061
30308 * tree-vrp.c (extract_range_basic): Compute zero as
30309 value-range for __builtin_constant_p of function parameters.
30311 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
30313 Update copyright years.
30315 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
30317 PR rtl-optimization/55672
30318 * lra-eliminations.c (mark_not_eliminable): Permit addition with
30319 const to be eliminable.
30321 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
30323 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
30324 * configure: Regenerate.
30326 2013-01-10 Richard Biener <rguenther@suse.de>
30328 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
30330 2013-01-10 Richard Biener <rguenther@suse.de>
30333 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
30334 locations for virtual PHI arguments.
30335 (rewrite_update_phi_arguments): Likewise.
30337 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
30339 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
30342 2013-01-10 Jakub Jelinek <jakub@redhat.com>
30344 PR tree-optimization/55921
30345 * tree-complex.c (expand_complex_asm): New function.
30346 (expand_complex_operations_1): Call it for GIMPLE_ASM.
30348 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30351 * config/s390/s390.c (s390_symref_operand_p)
30352 (s390_loadrelative_operand_p): Merge the two functions.
30353 (s390_check_qrst_address, print_operand_address): Add parameters
30354 to s390_loadrelative_operand_p invokation.
30355 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
30356 (s390_reload_larl_operand, s390_secondary_reload): Use
30357 s390_loadrelative_operand_p instead of s390_symref_operand_p.
30358 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
30360 2013-01-09 Mike Stump <mikestump@comcast.net>
30362 * dse.c (record_store): Remove unnecessary assert.
30364 2013-01-09 Jan Hubicka <jh@suse.cz>
30366 PR tree-optimization/55569
30367 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
30368 * cfgloop.h (scale_loop_profile): Likewise.
30370 2013-01-09 Jan Hubicka <jh@suse.cz>
30373 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
30375 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
30377 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
30379 PR middle-end/55114
30380 * expr.h (maybe_emit_group_store): Declare.
30381 * expr.c (maybe_emit_group_store): New function.
30382 * builtins.c (expand_builtin_int_roundingfn): Call it.
30383 (expand_builtin_int_roundingfn_2): Likewise.
30385 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
30387 PR rtl-optimization/55829
30388 * lra-constraints.c (match_reload): Add code for absent output.
30389 (curr_insn_transform): Add code for reloads of matched inputs
30392 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
30394 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
30395 attribute of movddup insn to DF.
30396 (*vec_interleave_lowv2df): Ditto.
30397 (vec_dupv2df): Ditto.
30399 2013-01-09 Jan Hubicka <jh@suse.cz>
30401 PR tree-optimiation/55875
30402 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
30403 EVERY_ITERATION parameter.
30404 (number_of_iterations_exit): Check if exit is executed every iteration.
30405 (idx_infer_loop_bounds): Similarly here.
30406 (n_of_executions_at_most): Simplify
30407 to only test for cases where statement is dominated by the
30408 particular bound; handle correctly the "postdominance" test.
30409 (scev_probably_wraps_p): Use max loop iterations info
30410 as a global bound first.
30412 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
30413 Nick Clifton <nickc@redhat.com>
30415 * config/v850/v850.md (cbranchsf4): New pattern.
30416 (cstoresf4): New pattern.
30417 (cbranchdf4): New pattern.
30418 (cstoredf4): New pattern.
30419 (movsicc): Disallow floating point comparisons.
30420 (cmpsf_le_insn): Fix order of operators.
30421 (cmpsf_lt_insn): Likewise.
30422 (cmpsf_eq_insn): Likewise.
30423 (cmpdf_le_insn): Likewise.
30424 (cmpdf_lt_insn): Likewise.
30425 (cmpdf_eq_insn): Likewise.
30426 (cmpsf_ge_insn): Use LE comparison.
30427 (cmpdf_ge_insn): Likewise.
30428 (cmpsf_gt_insn): Use LT comparison.
30429 (cmpdf_gt_insn): Likewise.
30430 (cmpsf_ne_insn): Delete pattern.
30431 (cmpdf_ne_insn): Delete pattern.
30432 * config/v850/v850.c (v850_gen_float_compare): Use
30433 gen_cmpdf_eq_insn for NE comparison.
30434 (v850_float_z_comparison_operator)
30435 (v850_float_nz_comparison_operator): Move from here ...
30436 * config/v850/predicates.md: ... to here. Move GT and GE
30437 comparisons into v850_float_z_comparison_operator.
30438 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
30440 (v850_float_nz_comparison_operator): Likewise.
30442 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30444 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
30445 with calls to gen_insvsi/gen_insvdi.
30447 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
30449 * config/i386/i386.c (initial_ix86_tune_features): Set up
30450 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
30452 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
30453 Jakub Jelinek <jakub@redhat.com>
30455 PR tree-optimization/48189
30456 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
30457 If nitercst is 0, don't predict the exit edge.
30459 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30461 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
30462 in asm_fprintf with reg_names.
30463 (aarch64_print_operand_address): Likewise.
30464 (aarch64_return_addr): Likewise.
30465 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
30467 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30469 * config/pa/pa.h (VAL_U6_BITS_P): Define.
30470 (INT_U6_BITS): Likewise.
30471 * config/pa/predicates.md (uint6_operand): New predicate.
30472 (shift5_operand, shift6_operand): Likewise.
30473 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
30475 (lshrdi3): Use shift6_operand.
30476 (shrpsi4, shrpdi4): New insn patterns.
30477 (extzv): Delete expander.
30478 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
30479 predicates in unamed zero extract patterns. Tighten common constraint.
30480 (extv): Delete expander.
30481 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
30482 predicates in unamed sign extract patterns. Tighten common constraint.
30483 (insv): Delete expander.
30484 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
30485 predicates in unamed insert patterns. Tighten common constraint.
30486 Change uint32_operand predicate to uint6_operand predicate in unamed
30487 DImode pattern to insert constant values of type 1...1xxxx.
30489 2013-01-04 Jan Hubicka <jh@suse.cz>
30491 PR tree-optimization/55823
30492 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
30495 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30496 Uros Bizjak <ubizjak@gmail.com>
30498 PR rtl-optimization/55845
30499 * df-problems.c (can_move_insns_across): Stop scanning at
30500 volatile_insn_p source instruction or give up if
30501 across_from .. across_to range contains any volatile_insn_p
30504 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30506 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
30507 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
30509 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
30510 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
30512 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30515 * asan.c (asan_clear_shadow): New function.
30516 (asan_emit_stack_protection): Use it.
30518 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30520 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
30521 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
30522 with tab instead of space.
30524 2013-01-08 Nick Clifton <nickc@redhat.com>
30526 * config/rl78/rl78.c (rl78_expand_prologue): Always select
30527 register bank 0 at the start of an interrupt handler.
30528 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
30531 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30533 * config/aarch64/aarch64-simd.md
30534 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
30535 (aarch64_simd_bsl): Likewise.
30536 (aarch64_vcond_internal<mode>): Likewise.
30537 (vcond<mode><mode>): Likewise.
30538 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
30539 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
30541 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30543 * config/aarch64/aarch64-builtins.c
30544 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
30546 2013-01-08 Martin Jambor <mjambor@suse.cz>
30549 * tree-sra.c (analyze_access_subtree): Return true also after
30550 potentially creating a debug-only replacement.
30552 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30554 PR middle-end/55890
30555 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
30557 PR tree-optimization/54120
30558 * tree-vrp.c (range_fits_type_p): Don't allow
30559 src_precision < precision from signed vr to unsigned_p
30560 if vr->min or vr->max is negative.
30561 (simplify_float_conversion_using_ranges): Test can_float_p
30562 against CODE_FOR_nothing.
30564 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30565 Richard Biener <rguenther@suse.de>
30567 PR middle-end/55851
30568 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
30569 types instead of just INTEGER_TYPE types.
30571 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
30573 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
30576 2013-01-07 Steve Ellcey <sellcey@mips.com>
30579 * config/mips/mips.opt: Change mad to mmad to match documentation.
30581 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30584 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
30585 .progmemx.data now.
30587 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30590 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
30591 (avr_addrspace_t): Add .section_name field.
30592 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
30594 (avr_addrspace): Same. Initialize .section_name. Remove last
30595 NULL entry. Put __memx into .progmemx.data.
30596 (progmem_section_prefix): Remove.
30597 (avr_asm_init_sections): No need to initialize progmem_section.
30598 (avr_asm_named_section): Use avr_addrspace[].section_name to get
30599 section name prefix.
30600 (avr_asm_select_section): Ditto. And use get_unnamed_section to
30601 retrieve the progmem section.
30602 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
30603 boundary to run over avr_addrspace[].
30604 (avr_register_target_pragmas): Ditto.
30606 2013-01-06 Jakub Jelinek <jakub@redhat.com>
30608 * varasm.c (output_constant_def_contents): For asan_protect_global
30609 protected strings, adjust DECL_ALIGN if needed, before testing for
30611 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
30612 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
30614 (output_object_block): For asan protected decls, emit asan padding
30615 after their contents.
30616 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
30617 (asan_finish_file): Test it here instead.
30619 2013-01-07 Nick Clifton <nickc@redhat.com>
30620 Matthias Klose <doko@debian.org>
30621 Doug Kwan <dougkwan@google.com>
30622 H.J. Lu <hongjiu.lu@intel.com>
30625 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
30627 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
30629 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
30631 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
30633 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
30635 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30638 * doc/install.texi (Cross-Compiler-Specific Options): Document
30641 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
30643 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
30644 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
30645 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
30646 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
30647 vqmovun_high_s64): Fix source operand number and update copyright.
30649 2013-01-07 Richard Biener <rguenther@suse.de>
30651 PR middle-end/55890
30652 * gimple.h (gimple_call_builtin_p): New overload.
30653 * gimple.c (validate_call): New function.
30654 (gimple_call_builtin_p): Likewise.
30655 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
30656 Use gimple_call_builtin_p.
30657 (find_func_clobbers): Likewise.
30658 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
30659 (strlen_optimize_stmt): Likewise.
30661 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30663 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
30664 (vld1q_dup_*): Likewise.
30665 (vld1_*): Likewise.
30666 (vld1q_*): Likewise.
30667 (vld1_lane_*): Likewise.
30668 (vld1q_lane_*): Likewise.
30670 2013-01-07 Richard Biener <rguenther@suse.de>
30672 * lto-streamer.h (LTO_minor_version): Bump to 2.
30674 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30676 * config/aarch64/aarch64-protos.h
30677 (aarch64_const_double_zero_rtx_p): Rename to...
30678 (aarch64_float_const_zero_rtx_p): ...this.
30679 (aarch64_float_const_representable_p): New.
30680 (aarch64_output_simd_mov_immediate): Likewise.
30681 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
30682 move immediate case.
30683 * config/aarch64/aarch64.c
30684 (aarch64_const_double_zero_rtx_p): Rename to...
30685 (aarch64_float_const_zero_rtx_p): ...this.
30686 (aarch64_print_operand): Allow printing of new constants.
30687 (aarch64_valid_floating_const): New.
30688 (aarch64_legitimate_constant_p): Check for valid floating-point
30690 (aarch64_simd_valid_immediate): Likewise.
30691 (aarch64_vect_float_const_representable_p): New.
30692 (aarch64_float_const_representable_p): Likewise.
30693 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
30694 (aarch64_output_simd_mov_immediate): New.
30695 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
30696 (*movdf_aarch64): Likewise.
30697 * config/aarch64/constraints.md (Ufc): New.
30698 (Y): call aarch64_float_const_zero_rtx.
30699 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
30701 2013-01-07 Richard Biener <rguenther@suse.de>
30703 PR tree-optimization/55888
30704 PR tree-optimization/55862
30705 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
30706 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
30707 not if it is contained therein.
30709 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30711 * config/avr/t-avr: Typo.
30713 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30716 * config/avr/t-avr: Don't automatically rebuild
30717 $(srcdir)/config/avr/t-multilib
30718 $(srcdir)/config/avr/avr-tables.opt
30719 $(srcdir)/doc/avr-mmcu.texi
30720 (avr-mcus): New phony target to build them on request.
30721 (s-avr-mlib, s-avr-mmcu-texi): Remove.
30722 * avr/avr-mcus.def: Adjust comments.
30724 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
30726 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
30728 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30730 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30732 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30734 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30736 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30738 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30739 to generate profiling.
30740 * config/rs6000/aix64.h (LIB_SPEC): Same.
30742 2013-01-04 Andrew Pinski <apinski@cavium.com>
30744 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30746 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30748 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30750 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30752 (ix86_expand_move): Ditto.
30753 (ix86_zero_extend_to_Pmode): Ditto.
30754 (ix86_expand_call): Ditto.
30755 (ix86_expand_special_args_builtin): Ditto.
30756 (ix86_expand_builtin): Ditto.
30758 2013-01-04 Richard Biener <rguenther@suse.de>
30760 PR tree-optimization/55862
30761 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30762 translating them through PHI nodes.
30764 2013-01-04 Martin Jambor <mjambor@suse.cz>
30766 PR tree-optimization/55755
30767 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30768 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30770 2013-01-04 Marek Polacek <polacek@redhat.com>
30772 PR middle-end/55859
30773 * opts.c (default_options_optimization): Clarify error message.
30775 2013-01-04 Richard Biener <rguenther@suse.de>
30777 PR middle-end/55863
30778 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30781 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30784 * config/pa/pa.md (movsi): Revert previous change.
30785 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30788 2013-01-03 Richard Henderson <rth@redhat.com>
30790 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30791 after eliminating TLS symbols.
30793 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30796 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30797 * graphite-poly.c (debug_gmp_value): Likewise.
30799 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30802 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30803 selected code model, define __code_mode_small__, __code_model_medium__,
30804 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30805 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30806 xchg temporary register with %k. Declare temporary register as
30808 [__x86_64__]: For medium and large code models, preserve %rbx register.
30810 2013-01-03 Richard Biener <rguenther@suse.de>
30812 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30813 (dump_subscript): Adjust.
30814 (finalize_ddr_dependent): Do not dump redundant info.
30815 (analyze_siv_subscript): Adjust.
30816 (subscript_dependence_tester): Likewise.
30817 (compute_affine_dependence): Likewise.
30819 2013-01-03 Richard Biener <rguenther@suse.de>
30822 2013-01-03 Richard Biener <rguenther@suse.de>
30824 PR tree-optimization/55857
30825 * tree-vect-stmts.c (vectorizable_load): Do not setup
30826 re-alignment for invariant loads.
30828 2013-01-02 Richard Biener <rguenther@suse.de>
30830 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30831 invariant load do not generate a vector load from the scalar location.
30833 2013-01-03 Richard Biener <rguenther@suse.de>
30835 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30836 for not vectorizing.
30837 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30838 not build INDIRECT_REFs, call get_name once only.
30839 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30840 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30842 2013-01-03 Richard Biener <rguenther@suse.de>
30844 PR tree-optimization/55857
30845 * tree-vect-stmts.c (vectorizable_load): Do not setup
30846 re-alignment for invariant loads.
30848 2013-01-03 Richard Biener <rguenther@suse.de>
30851 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30852 prefer a built-in decl.
30854 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30856 * gcc.c (process_command): Update copyright notice dates.
30857 * gcov.c (print_version): Likewise.
30858 * gcov-dump.c (print_version): Likewise.
30860 PR rtl-optimization/55838
30861 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30862 iv0.step, iv1.step and step.
30864 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30865 Marc Glisse <marc.glisse@inria.fr>
30867 PR tree-optimization/55832
30868 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30869 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30870 integer_{one,zero}_node.
30872 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30875 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30876 * var-tracking.c (reverse_op): Don't add reverse ops to
30877 VALUEs that have already
30878 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30880 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30882 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30884 2013-01-02 Teresa Johnson <tejohnson@google.com>
30886 * dumpfile.c (dump_loc): Print filename with location.
30887 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30888 new location_t parameter to emit complete unroll message with
30889 new dump framework.
30890 (canonicalize_loop_induction_variables): Compute loops location
30891 and pass to try_unroll_loop_completely.
30892 * loop-unroll.c (report_unroll_peel): New function.
30893 (peel_loops_completely): Use new dump format with location
30894 for main dumpfile message, and invoke report_unroll_peel on success.
30895 (decide_unrolling_and_peeling): Ditto.
30896 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30897 by report_unroll_peel.
30898 (decide_peel_completely): Ditto.
30899 (decide_unroll_constant_iterations): Ditto.
30900 (decide_unroll_runtime_iterations): Ditto.
30901 (decide_peel_simple): Ditto.
30902 (decide_unroll_stupid): Ditto.
30903 * cfgloop.c (get_loop_location): New function.
30904 * cfgloop.h (get_loop_location): Declare.
30906 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30908 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30911 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30913 PR middle-end/55198
30914 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30915 BLKmode objects when EXPAND_MEMORY is specified.
30917 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30919 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30921 (fold_builtin_cpu): Do not share cpu model decls across statements.
30923 2013-01-02 Jason Merrill <jason@redhat.com>
30926 * tree.c (build_array_type_1): Revert earlier change.
30928 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30930 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30932 * config/aarch64/aarch64-tune.md: Re-generate.
30934 2013-01-02 Richard Biener <rguenther@suse.de>
30936 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30937 invariant load do not generate a vector load from the scalar location.
30939 2013-01-02 Richard Biener <rguenther@suse.de>
30942 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30943 * configure: Regenerate.
30945 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30947 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30948 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30949 (expand_builtin_int_roundingfn_2): Keep the original target around
30950 for the fallback case.
30952 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30954 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30955 to be clear for sign changes.
30957 2013-01-01 Jan Hubicka <jh@suse.cz>
30959 * ipa-inline-analysis.c: Fix formatting.
30961 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30963 PR tree-optimization/55831
30964 * tree-vect-loop.c (get_initial_def_for_induction): Use
30965 gsi_after_labels instead of gsi_start_bb.
30967 Copyright (C) 2013 Free Software Foundation, Inc.
30969 Copying and distribution of this file, with or without modification,
30970 are permitted in any medium without royalty provided the copyright
30971 notice and this notice are preserved.