]> git.ipfire.org Git - thirdparty/gcc.git/blob - gcc/ChangeLog
pretty-print.h (pp_newline_and_flush): Declare.
[thirdparty/gcc.git] / gcc / ChangeLog
1 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
2
3 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4 definition.
5 (pp_newline_and_indent): Likewise.
6 (pp_separate_with): Likewise.
7 * pretty-print.c (pp_newline_and_flush): Define.
8 (pp_newline_and_indent): Likewise.
9 (pp_separate_with): Likewise.
10
11 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12
13 PR target/58218
14 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
15 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
16
17 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
18
19 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
20 * gcc/config/i386/i386.md (*movti_internal): Use
21 predicate to determine if EVEX is needed.
22 (*movsi_internal): Ditto.
23 (*movdf_internal): Ditto.
24 (*movsf_internal): Ditto.
25 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
26
27 2013-08-23 Jakub Jelinek <jakub@redhat.com>
28
29 PR tree-optimization/58209
30 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
31 (find_tail_calls): Give up for pointer result types if m is non-NULL.
32 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
33 emit POINTER_PLUS_EXPR.
34 (create_tailcall_accumulator): For pointer result type accumulate in
35 sizetype type.
36
37 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
38
39 * configure.ac: Add backslashes missing from the last change.
40 * configure: Regenerate.
41
42 2013-08-22 Jan Hubicka <jh@susue.cz>
43
44 * ipa.c (function_and_variable_visibility): First remember function
45 was global and then make it local.
46
47 2013-08-22 Julian Brown <julian@codesourcery.com>
48
49 * configure.ac: Add aarch64 to list of arches which use "nop" in
50 debug_line test.
51 * configure: Regenerate.
52
53 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
54
55 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
56 gnu_libc_has_function.
57 * config/s390/tpf.h: Likewise.
58
59 2013-08-22 Jan Hubicka <jh@susue.cz>
60
61 * timevar.c (validate_phases): Add cast.
62
63 2013-08-22 Jan Hubicka <jh@susue.cz>
64
65 * timevar.c (validate_phases): Use size_t for memory.
66 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
67
68 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
69
70 * pretty-print.h (output_buffer::output_buffer): Declare.
71 (pretty_printer::pretty_printer): Likewise.
72 (pp_construct): Remove.
73 * pretty-print.c (output_buffer::output_buffer): Define.
74 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
75 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
76 (print_gimple_expr): Likewise.
77 (print_gimple_seq): Likewise.
78 (gimple_dump_bb): Likewise.
79 * sched-vis.c (dump_value_slim): Likewise.
80 (dump_insn_slim): Likewise.
81 (dump_rtl_slim): Likewise.
82 (str_pattern_slim): Likewise.
83 * tree-mudflap.c (mf_varname_tree): Likewise.
84 * graph.c (print_graph_cfg): Likewise.
85 (start_graph_dump): Likewise.
86 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
87 placement-new.
88 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
89 pretty printer initialization.
90 * coretypes.h (diagnostic_context): Remove superflous type alias
91 declaration.
92 (pretty_printer): Likewise. Declare directly as a class.
93 (pretty_print_info): Remove declaration as class.
94 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
95 and pp_clear_output_area.
96 (asan_add_global): Likewise.
97
98 2013-08-22 Jan Hubicka <jh@suse.cz>
99
100 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
101 * ipa-utils.h (update_type_inheritance_graph): Declare.
102 (possible_polymorphic_call_target_p): Declare.
103 (possible_polymorphic_call_target_p): New.
104 * ipa-devirt.c: Update toplevel comments.
105 (cached_polymorphic_call_targets): Move up.
106 (odr_type_d): Move ID down.
107 (polymorphic_type_binfo_p): Update comment.
108 (odr_hasher::remove): Likewise;
109 (get_odr_type): Set anonymous_namespace.
110 (dump_odr_type): Dump it.
111 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
112 (maybe_record_node): Record node in cached_polymorphic_call_targets.
113 (record_binfo): Add comment.
114 (free_polymorphic_call_targets_hash): Do not ICE when cache is not built.
115 (devirt_node_removal_hook): Do not iCE when cache is freed.
116 (possible_polymorphic_call_target_p): New predicate.
117 (update_type_inheritance_graph): New function.
118
119 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
120 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
121 Sergey Lega <sergey.s.lega@intel.com>
122 Anna Tikhonova <anna.tikhonova@intel.com>
123 Ilya Tocar <ilya.tocar@intel.com>
124 Andrey Turetskiy <andrey.turetskiy@intel.com>
125 Ilya Verbin <ilya.verbin@intel.com>
126 Kirill Yukhin <kirill.yukhin@intel.com>
127 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
128
129 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
130 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
131 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
132 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
133 (OPTION_MASK_ISA_AVX2_UNSET): Update.
134 (OPTION_MASK_ISA_AVX512F_UNSET): New.
135 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
136 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
137 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
138 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
139 OPT_mavx512pf, OPT_mavx512er cases.
140 * config/i386/constraints.md (v): New constraint.
141 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
142 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
143 (bit_AVX512CD): New.
144 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
145 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
146 * config/i386/i386-c.c (ix86_target_macros_internal):
147 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
148 __AVX512PF__.
149 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
150 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
151 * config/i386/i386.c (regclass_map, dbx_register_map)
152 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
153 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
154 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
155 -mavx512pf options.
156 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
157 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
158 -mavx512pf options. Fix formatting.
159 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
160 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
161 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
162 -mavx512cd, -mavx512pf options.
163 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
164 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
165 (ix86_preferred_output_reload_class): Replace SSE_REGS with
166 ALL_SSE_REGS.
167 (ix86_hard_regno_mode_ok): Support 512-bit registers.
168 (ix86_set_reg_reg_cost): Ditto.
169 (x86_order_regs_for_local_alloc): Ditto.
170 (MAX_VECT_LEN): Extend to 64-byte.
171 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
172 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
173 (TARGET_AVX512ER, TARGET_AVX512CD): New.
174 (BIGGEST_ALIGNMENT): Extend to 512-bits.
175 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
176 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
177 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
178 (SSE_REG_MODE_P): Support new modes.
179 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
180 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
181 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
182 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
183 (REG_CLASS_CONTENTS): Add new registers.
184 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
185 (EXT_REX_SSE_REGNO_P): New.
186 (HI_REGISTER_NAMES): Add new registers.
187 * config/i386/i386.md: Define constants for new registers.
188 (mode): Add new 512-bit modes.
189 (prefix): Support evex prefix.
190 (isa): Support avx512f, noavx512f, fma_avx512f.
191 (ssemodesuffix): Add new 512-bit modes.
192 (movxi): New.
193 (*movxi_internal_avx512f): Ditto.
194 (*movdi_internal): Replace constraint "x" with the new constraint "v".
195 Support MODE_XI.
196 (*movsi_internal): Likewise.
197 (*movdf_internal): Likewise.
198 (*movsf_internal): Likewise.
199 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
200 (<code><mode>3): Likewise.
201 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd): New.
202 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
203 with the new constraint "v".
204 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
205 modes.
206 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
207 with the new constraint "v".
208 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
209 (<sse2>_storedqu<avxsizesuffix>): Likewise.
210 (*<plusminus_insn><mode>3): Likewise.
211 (<sse>_vm<plusminus_insn><mode>3): Likewise.
212 (*mul<mode>3): Likewise.
213 (<sse>_vmmul<mode>3): Likewise.
214 (<sse>_div<mode>3): Likewise.
215 (<sse>_vmdiv<mode>3): Likewise.
216 (<sse>_sqrt<mode>2): Likewise.
217 (<sse>_vmsqrt<mode>2): Likewise.
218 (*<code><mode>3_finite): Likewise.
219 (*<code><mode>3) <smaxmin>: Likewise.
220 (<sse>_vm<code><mode>3): Likewise.
221 (*<code><mode>3) <any_logic>: Likewise.
222 (*fma_fmadd_<mode>): Likewise.
223 (*fma_fmsub_<mode>): Likewise.
224 (*fma_fnmadd_<mode>): Likewise.
225 (*fma_fnmsub_<mode>): Likewise.
226 (*fma_fmaddsub_<mode>): Likewise.
227 (*fma_fmsubadd_<mode>): Likewise.
228 (*fmai_fmadd_<mode>): Likewise.
229 (*fmai_fmsub_<mode>): Likewise.
230 (*fmai_fnmadd_<mode>): Likewise.
231 (*fmai_fnmsub_<mode>): Likewise.
232 (sse_cvtsi2ss): Likewise.
233 (sse_cvtsi2ssq): Likewise.
234 (sse_cvtss2si): Likewise.
235 (sse_cvtss2si_2): Likewise.
236 (sse_cvtss2siq): Likewise.
237 (sse_cvtss2siq_2): Likewise.
238 (sse_cvttss2si): Likewise.
239 (sse_cvtss2siq_2): Likewise.
240 (float<sseintvecmodelower><mode>2): Likewise.
241 (sse2_cvtsd2si_2): Likewise.
242 (sse2_cvtsd2siq_2): Likewise.
243 (*<plusminus_insn><mode>3): Likewise.
244 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
245 (*<sse4_1_avx2>_mul<mode>3): Likewise.
246 (ashr<mode>3): Likewise.
247 (<shift_insn><mode>3): Likewise.
248 (avx2_<code><mode>3): Likewise.
249 (*avx2_<code><mode>3): Likewise.
250 (*andnot<mode>3): Likewise.
251 (*<code><mode>3) <any_logic>: Likewise.
252 (abs<mode>2): Likewise.
253 (avx2_permvar<mode>): Likewise.
254 (avx2_perm<mode>_1): Likewise.
255 (*avx_vpermilp<mode>): Likewise.
256 (avx_vpermilvar<mode>3): Likewise.
257 (avx2_ashrv<mode>): Likewise.
258 (avx2_<shift_insn>v<mode>): Likewise.
259 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
260 -mavx512cd.
261 * doc/rtl.texi: Document XImode.
262
263 2013-08-21 Jeff Law <law@redhat.com>
264
265 * tree-flow.h (register_jump_thread): Pass vector of edges
266 instead of each important edge.
267 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
268 thread path into a vector and pass that to register_jump_thread.
269 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
270 passed in edge vector to the current 3-edge form.
271
272 Revert:
273 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
274
275 * dce.c (fini_dce): Call df_analyze again just in case
276 delete_unmarked_insns removed anything.
277
278 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
279
280 * reload.h (struct reg_equivs): Rename to ..
281 (struct reg_equivs_s): .. this.
282
283 2013-08-20 Martin Liska <marxin.liska@gmail.com>
284
285 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
286
287 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
288
289 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest
290 comment.
291
292 2013-08-21 Jeff Law <law@redhat.com>
293
294 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
295 simplify assignments too. If the RHS collapses to a singleton
296 range, then return the value for the range.
297
298 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
299
300 * config/i386/sse.md (V16): Rename to...
301 (VMOVE): this.
302 (mov<mode>): Update iterator name.
303 (*mov<mode>_internal): Ditto.
304 (push<mode>1): Ditto.
305 (movmisalign<mode>): Ditto.
306
307 2013-08-20 Jan Hubicka <jh@suse.cz>
308
309 PR bootstrap/58186
310 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
311 entry for direct edges.
312 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
313
314 2013-08-20 David Malcolm <dmalcolm@redhat.com>
315
316 Revert my last two changes, r201865 and r201864:
317
318 Revert r201865:
319 2013-08-20 David Malcolm <dmalcolm@redhat.com>
320
321 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
322 instances can own GC refs.
323
324 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
325 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
326 (gcc::context::gt_pch_nx): Likewise.
327 (gcc::context::gt_pch_nx): Likewise.
328 * ggc.h (gt_ggc_mx <T>): New.
329 (gt_pch_nx_with_op <T>): New.
330 (gt_pch_nx <T>): New.
331 * passes.c (opt_pass::gt_ggc_mx): New.
332 (opt_pass::gt_pch_nx): New.
333 (opt_pass::gt_pch_nx_with_op): New.
334 (pass_manager::gt_ggc_mx): New.
335 (pass_manager::gt_pch_nx): New.
336 (pass_manager::gt_pch_nx_with_op): New.
337 (pass_manager::operator new): Use
338 ggc_internal_cleared_alloc_stat rather than xcalloc.
339 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
340 (pass_manager::gt_ggc_mx): New.
341 (pass_manager::gt_pch_nx): New.
342 (pass_manager::gt_pch_nx_with_op): New.
343 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
344 (opt_pass::operator new): New.
345 (opt_pass::gt_ggc_mx): New.
346 (opt_pass::gt_pch_nx): New.
347 (opt_pass::gt_pch_nx_with_op): New.
348
349 Revert r201864:
350 2013-08-20 David Malcolm <dmalcolm@redhat.com>
351
352 * Makefile.in (GTFILES): Add context.h.
353 * context.c (gcc::context::operator new): New.
354 (gcc::context::gt_ggc_mx): New.
355 (gcc::context::gt_pch_nx): New.
356 (gcc::context::gt_pch_nx): New.
357 * context.h (gcc::context): Add GTY((user)) marking.
358 (gcc::context::operator new): New.
359 (gcc::context::gt_ggc_mx): New.
360 (gcc::context::gt_pch_nx): New.
361 (gcc::context::gt_pch_nx): New.
362 (g): Add GTY marking.
363 (gt_ggc_mx (gcc::context *)): New.
364 (gt_pch_nx (gcc::context *)): New.
365 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
366 void *cookie)): New.
367 * gengtype.c (open_base_files) <ifiles>: Add context.h.
368
369 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
370
371 * dce.c (fini_dce): Call df_analyze again just in case
372 delete_unmarked_insns removed anything.
373
374 2013-08-20 Teresa Johnson <tejohnson@google.com>
375
376 PR rtl-optimizations/57451
377 * final.c (reemit_insn_block_notes): Prevent lexical blocks
378 from crossing split section boundaries.
379
380 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
381
382 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
383 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
384 with MULTLIB_DEFAULTS.
385
386 2013-08-20 Nick Clifton <nickc@redhat.com>
387
388 * target.def (narrow_volatile_bitfield): Note that the default
389 value is false, not !TARGET_STRICT_ALIGN.
390 * doc/tm.texi: Regenerate.
391
392 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
393
394 Fix LIB_SPEC for systems without libpthread.
395
396 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
397 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
398 for Android.
399 * config/i386/linux-common.h: Likewise.
400 * config/mips/linux-common.h: Likewise.
401
402 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
403
404 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
405 checks.
406
407 2013-08-20 David Malcolm <dmalcolm@redhat.com>
408
409 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
410 instances can own GC refs.
411
412 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
413 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
414 (gcc::context::gt_pch_nx): Likewise.
415 (gcc::context::gt_pch_nx): Likewise.
416 * ggc.h (gt_ggc_mx <T>): New.
417 (gt_pch_nx_with_op <T>): New.
418 (gt_pch_nx <T>): New.
419 * passes.c (opt_pass::gt_ggc_mx): New.
420 (opt_pass::gt_pch_nx): New.
421 (opt_pass::gt_pch_nx_with_op): New.
422 (pass_manager::gt_ggc_mx): New.
423 (pass_manager::gt_pch_nx): New.
424 (pass_manager::gt_pch_nx_with_op): New.
425 (pass_manager::operator new): Use
426 ggc_internal_cleared_alloc_stat rather than xcalloc.
427 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
428 (pass_manager::gt_ggc_mx): New.
429 (pass_manager::gt_pch_nx): New.
430 (pass_manager::gt_pch_nx_with_op): New.
431 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
432 (opt_pass::operator new): New.
433 (opt_pass::gt_ggc_mx): New.
434 (opt_pass::gt_pch_nx): New.
435 (opt_pass::gt_pch_nx_with_op): New.
436
437 2013-08-20 David Malcolm <dmalcolm@redhat.com>
438
439 * Makefile.in (GTFILES): Add context.h.
440 * context.c (gcc::context::operator new): New.
441 (gcc::context::gt_ggc_mx): New.
442 (gcc::context::gt_pch_nx): New.
443 (gcc::context::gt_pch_nx): New.
444 * context.h (gcc::context): Add GTY((user)) marking.
445 (gcc::context::operator new): New.
446 (gcc::context::gt_ggc_mx): New.
447 (gcc::context::gt_pch_nx): New.
448 (gcc::context::gt_pch_nx): New.
449 (g): Add GTY marking.
450 (gt_ggc_mx (gcc::context *)): New.
451 (gt_pch_nx (gcc::context *)): New.
452 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
453 void *cookie)): New.
454 * gengtype.c (open_base_files) <ifiles>: Add context.h.
455
456 2013-08-20 Alan Modra <amodra@gmail.com>
457
458 PR target/57865
459 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
460 (rs6000_emit_epilogue): Likewise.
461
462 2013-08-19 Dehao Chen (dehao@google.com)
463
464 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
465
466 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
467 Jakub Jelinek <jakub@redhat.com>
468
469 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
470 (BUILT_IN_FABSD64): Likewise.
471 (BUILT_IN_FABSD128): Likewise.
472 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
473 (fold_builtin_1): Likewise.
474 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping destination
475 and source operands.
476 (*abstd2_fpr): Likewise.
477 (*nabstd2_fpr): Likewise.
478
479 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
480
481 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
482 JUMP_P and INSN_P.
483
484 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
485
486 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
487 incomplete.
488
489 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
490
491 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
492 * builtins.c (default_libc_has_function): New.
493 (gnu_libc_has_function): Ditto.
494 (no_c99_libc_has_function): Ditto.
495 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
496 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
497 (fold_builtin_sincos): Likewise.
498 (fold_builtin_cexp): Likewise.
499 * builtins.def (DEF_C94_BUILTIN): Likewise.
500 (DEF_C99_BUILTIN): Likewise.
501 (DEF_C99_C90RES_BUILTIN): Likewise.
502 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
503 definitions to using this define.
504 * config/darwin-protos.h (darwin_libc_has_function): New.
505 * config/darwin.c: (darwin_libc_has_function: Ditto.
506 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
507 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
508 * config/darwin.h: Ditto.
509 * config/elfos.h: Ditto.
510 * config/freebsd.h: Ditto.
511 * config/i386/cygming.h: Ditto.
512 * config/i386/djgpp.h: Ditto.
513 * config/i386/i386-interix.h: Ditto.
514 * config/microblaze/microblaze.h: Ditto.
515 * config/mmix/mmix.h: Ditto.
516 * config/gnu-user.h: Ditto.
517 * config/ia64/hpux.h: Ditto.
518 * config/pa/pa-hpux.h: Ditto.
519 * config/pdp11/pdp11.h: Ditto.
520 * config/picochip/picochip.h: Ditto.
521 * config/linux.h: Ditto.
522 * config/netbsd.h: Ditto.
523 * config/openbsd.h: Ditto.
524 * config/rs6000/aix43.h: Ditto.
525 * config/rs6000/aix51.h: Ditto.
526 * config/rs6000/aix52.h: Ditto.
527 * config/rs6000/aix53.h: Ditto.
528 * config/rs6000/aix61.h: Ditto.
529 * config/rs6000/darwin.h: Ditto.
530 * config/rs6000/linux.h: Ditto.
531 * config/rs6000/linux64.h: Ditto.
532 * config/s390/tpf.h: Ditto.
533 * config/sol2-10.h: Ditto.
534 * config/sol2.h: Ditto.
535 * config/vms/vms.h: Ditto.
536 * config/vxworks.h: Ditto.
537 * config/linux-android.c (linux_android_libc_has_function):
538 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
539 * config/linux-protos.h (linux_android_libc_has_function):
540 New declaration.
541 * config/i386/i386.c (ix86_libc_has_function): New.
542 * config/i386/i386-protos.h
543 (ix86_libc_has_function): New declaration.
544 * config/i386/i386.md
545 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
546 ("isinf<mode>2): Likewise.
547 * convert.c (convert_to_integer): Using new target hook
548 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
549 TARGET_C99_FUNCTIONS.
550 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
551 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
552 * coretypes.h (function_class): New enum for different
553 classes of functions.
554 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
555 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
556 (TARGET_HAS_SINCOS): Likewise.
557 (TARGET_LIBC_HAS_FUNCTION): New.
558 * doc/tm.texi: Regenerated.
559 * targhooks.h (default_libc_has_function): New declaration.
560 (no_c99_libc_has_function): Ditto.
561 (gnu_libc_has_function): Ditto.
562 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
563 and TARGET_HAS_SINCOS.
564
565 2013-08-18 Jan Hubicka <jh@suse.cz>
566
567 * Makeifle-in (ipa-devirt.o): New.
568 (GTFILES): Add ipa-utils.h and ipa-devirt.c
569 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
570 (analyze_functions): Look into possible targets of polymorphic call.
571 * dumpfile.c (dump_files): Add type-inheritance dump.
572 * dumpfile.h (TDI_inheritance): New.
573 * ipa-devirt.c: New file.
574 * ipa-utils.h (odr_type_d): Forward declare.
575 (odr_type): New type.
576 (build_type_inheritance_graph): Declare.
577 (possible_polymorphic_call_targets): Declare and introduce inline
578 variant when only edge is pased.
579 (dump_possible_polymorphic_call_targets): Likewise.
580 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
581 * tree.c (type_in_anonymous_namespace_p): Break out from ...
582 (types_same_for_odr): ... here.
583 * tree.h (type_in_anonymous_namespace_p): Declare.
584
585 2013-08-18 Jakub Jelinek <jakub@redhat.com>
586
587 PR tree-optimization/58006
588 * tree-parloops.c (take_address_of): Don't ICE if get_name
589 returns NULL.
590 (eliminate_local_variables_stmt): Remove clobber stmts.
591
592 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
593
594 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
595 error message is issued for an alias to undefined symbol.
596
597 2013-08-18 Jan Hubicka <jh@suse.cz>
598
599 * cgraph.c (cgraph_create_indirect_edge): Discover
600 polymorphic calls and record basic info into indirect_info.
601 * gimple-fold.c (gimple_fold_call): When doing BINFO based
602 devirtualization, ignore objc function calls.
603 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
604 call with no parm index info.
605 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
606 * tree.c (virtual_method_call_p): New function.
607 * tree.h (virtual_method_call_p): Declare.
608
609 2013-08-16 Jan Hubicka <jh@suse.cz>
610
611 PR middle-end/58179
612 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
613
614 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
615
616 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
617 attribute.
618
619 2013-08-16 David Malcolm <dmalcolm@redhat.com>
620
621 * gengtype.c (type_for_name): Add special-case support for
622 locating types within the "gcc::" namespace.
623 (open_base_files): Emit a "using namespace gcc" directive.
624
625 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
626
627 PR target/58160
628 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
629 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
630
631 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
632 array instead of each individual operand as a separate argument.
633 (emit_fusion_gpr_load): Likewise.
634 (expand_fusion_gpr_load): Add new function declaration.
635
636 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
637 signature to have the operands passed as an array, instead of as
638 separate arguments. Allow ZERO_EXTEND to be in the memory
639 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
640 depend on the register live/dead flags when peepholes are run.
641 (expand_fusion_gpr_load): New function to be called from the
642 peephole2 pass, to change the register that addis sets to be the
643 target register.
644 (emit_fusion_gpr_load): Change the calling signature to have the
645 operands passed as an array, instead of as separate arguments.
646 Allow ZERO_EXTEND to be in the memory address, and also
647 SIGN_EXTEND if -mpower8-fusion-sign.
648
649 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
650 unspec enumeration.
651 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
652 adjust the register addis loads up in the peephole2 pass. Do not
653 depend on the register live/dead state when the peephole pass is
654 done.
655
656 2013-08-16 David Malcolm <dmalcolm@redhat.com>
657
658 * gengtype.c (create_user_defined_type): Ensure that the kind
659 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
660 declaration is seen before the GTY((user)) marking.
661
662 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
663
664 PR target/58105
665 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
666
667 2013-08-16 Jan Hubicka <jh@suse.cz>
668
669 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
670 arugment expected_type.
671 (gimple_fold_call): Use it.
672 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
673 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
674 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
675 (try_make_edge_direct_virtual_call): Likewise.
676 * tree.c (obj_type_ref_class): New.
677 * tree.h (obj_type_ref_class): Use it.
678
679 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
680
681 * sched-vis.c (rtl_slim_pp_initialized): Remove.
682 (rtl_slim_pp): Likewise.
683 (init_rtl_slim_pretty_print): Likewise.
684 (dump_value_slim): Don't call it. Use local pretty printer.
685 (dump_insn_slim): Likewise.
686 (dump_rtl_slim): Likewise.
687 (str_pattern_slim): Likewise.
688 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
689 Simplify.
690
691 2013-08-16 Jakub Jelinek <jakub@redhat.com>
692
693 PR tree-optimization/58164
694 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
695 walk gimple_goto_dest of GIMPLE_GOTO.
696
697 PR tree-optimization/58165
698 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
699 bi_call must be the last stmt in a bb, don't split_block, instead
700 use fallthru edge from it and give up if there is none.
701 Release conds vector when returning early.
702
703 2013-08-14 Xinliang David Li <davidxl@google.com>
704
705 * config/i386/i386.c (ix86_option_override_internal):
706 Remove unused variable and field.
707
708 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
709
710 PR target/57949
711 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
712 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
713 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
714 and Linux, correct BLKmode alignment when 128-bit alignment is
715 required and compatibility flag is not set.
716 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
717 for zero-size arguments when compatibility flag is not set.
718
719 2013-08-14 Jakub Jelinek <jakub@redhat.com>
720
721 PR tree-optimization/58145
722 * tree-sra.c (build_ref_for_offset): If prev_base has
723 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
724
725 2013-08-14 Xinliang David Li <davidxl@google.com>
726
727 * config/i386/i386.c (ix86_option_override_internal):
728 Fix uninitialized variable error.
729
730 2013-08-14 Xinliang David Li <davidxl@google.com>
731
732 * config/i386/i386.opt: Define two new options.
733 * config/i386/x86-tune.def: Add arch selector field in macros.
734 * config/i386/i386.h: Adjust macro definition.
735 * config/i386/i386.c (ix86_option_override_internal):
736 Refactor the code.
737 (parse_mtune_ctrl_str): New function.
738 (set_ix86_tune_features): New function.
739 (ix86_function_specific_restore): Call the new helper function.
740
741 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
742
743 PR rtl-optimization/57662
744 * sel-sched.c (code_motion_process_successors): When the current insn
745 is removed after the recursive traversal, break from the loop.
746 Add comments and debug printouts.
747
748 2013-08-14 Jakub Jelinek <jakub@redhat.com>
749 Alexandre Oliva <aoliva@redhat.com>
750
751 PR target/58067
752 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
753 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
754 there also UNSPEC_PLTOFF.
755
756 2013-08-14 Marek Polacek <polacek@redhat.com>
757
758 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
759 NUM_CONDITIONS bit positions.
760
761 2013-08-13 Cary Coutant <ccoutant@google.com>
762
763 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
764 (attr_checksum): Hash vector contents instead of pointer.
765 (attr_checksum_ordered): Likewise.
766
767 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
768
769 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
770 when Pmode != word_mode. Add length_address attribute.
771 (sse3_monitor_<mode>): Merge from sse3_monitor and
772 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
773 Pmode != word_mode. Update insn length attribute.
774 * config/i386/i386.c (ix86_option_override_internal): Update
775 ix86_gen_monitor selection for merged sse3_monitor insn.
776
777 2013-08-13 Julian Brown <julian@codesourcery.com>
778
779 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
780 perform invalid legitimization on greater-than-word-size modes for
781 TARGET_E500_DOUBLE.
782
783 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
784
785 * ira.c (setup_class_translate_array): Use aclass instead of cl
786 for classes not fully covered by allocno classes.
787
788 2013-08-13 Jakub Jelinek <jakub@redhat.com>
789
790 PR tree-optimization/57661
791 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
792 * tree-inline.c (tree_function_versioning): Initialize it.
793 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
794 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
795 that is not being copied.
796
797 PR sanitizer/56417
798 * asan.c (instrument_strlen_call): Fix typo in comment.
799 Use char * type even for the lhs of POINTER_PLUS_EXPR.
800
801 2013-08-13 Steve Ellcey <sellcey@mips.com>
802
803 * config/mips/mips.md (prefetch): Use lw instead of ld on
804 loongson in 32bit mode.
805
806 2013-08-13 Nick Clifton <nickc@redhat.com>
807
808 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
809
810 2013-08-13 Jan Hubicka <jh@suse.cz>
811
812 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
813 introduced edge; fix typo in sanity check.
814 (cgraph_resolve_speculation): Export; improve diagnostic.
815 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
816 speculation at type mismatch.
817 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
818 (cgraph_resolve_speculation): Declare.
819 (symtab_can_be_discarded): New function.
820 * value-prof.c (gimple_ic_transform): Remove actual transform code.
821 * ipa-inline-transform.c (speculation_removed): New global var.
822 (clone_inlined_nodes): See if speculation can be removed.
823 (inline_call): If speculations was removed, we growths may not match.
824 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
825 (speculation_useful_p): New function.
826 (resolve_noninline_speculation): New function.
827 (inline_small_functions): Resolve useless speculations.
828 * ipa-inline.h (speculation_useful_p): Declare
829 * ipa.c (can_replace_by_local_alias): Simplify.
830 (ipa_profile): Produce speculative calls in non-lto, too;
831 add simple cost model; produce local aliases.
832
833 2013-08-13 David Malcolm <dmalcolm@redhat.com>
834
835 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
836 PASS_MANAGER_H.
837
838 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
839
840 * config/i386/i386.c (ix86_function_versions): Use error + inform.
841
842 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
843
844 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
845 iterator instead of X87MODEF.
846
847 2013-08-12 Perez Read <netfirewall@gmail.com>
848
849 PR target/58132
850 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
851 operand 0 for intel asm alternative.
852 (*movabs<mode>_2): Ditto for operand 1.
853
854 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
855
856 * config/aarch64/arm_none.h
857 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
858
859 2013-08-12 Nick Clifton <nickc@redhat.com>
860
861 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
862
863 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
864
865 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
866 expand for QI/HImode operand to produce more effictive code for
867 unsigned char(short) --> float(double) conversion.
868
869 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
870
871 * doc/invoke.texi: Mention that -ftls-model does not force the final
872 model.
873
874 2013-08-12 Marek Polacek <polacek@redhat.com>
875 Marc Glisse <marc.glisse@inria.fr>
876
877 PR tree-optimization/57980
878 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
879 when creating -1 constant.
880
881 2013-08-10 Jan Hubicka <jh@suse.cz>
882
883 Workaround binutils PR14342.
884 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
885 (gimple_init_edge_profiler): Likewise.
886 (gimple_gen_ic_func_profiler): Likewise.
887
888 2013-08-09 Jan Hubicka <jh@suse.cz>
889
890 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
891
892 2013-08-09 Xinliang David Li <davidxl@google.com>
893
894 * config/i386/stringop.def: New file.
895 * config/i386/stringop.opt: New file.
896 * config/i386/i386-opts.h: Include stringopt.def.
897 * config/i386/i386.opt: Include stringopt.opt.
898 * config/i386/i386.c (ix86_option_override_internal):
899 Override default size based stringop inline strategies with options.
900 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
901 New function.
902
903 2013-08-09 Jan Hubicka <jh@suse.cz>
904
905 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
906
907 2013-08-09 Jan Hubicka <jh@suse.cz>
908
909 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
910 CGRAPH_FREQ_MAX.
911 (dump_cgraph_node): Dump profile-id.
912 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
913 and common_target_probability.
914 * lto-cgraph.c (lto_output_edge): Stream common targets.
915 (lto_output_node): Stream profile ids.
916 (input_node): Stream profile ids.
917 (input_edge): Stream common targets.
918 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
919 * ipa.c: Include value-prof.h
920 (ipa_profile_generate_summary): Turn indirect call statement histograms
921 into common targets.
922 (ipa_profile): Turn common targets into speculative edges.
923
924 2013-08-09 Jan Hubicka <jh@suse.cz>
925
926 * cgraph.h (cgraph_node): Add profile_id.
927 * value-prof.c (cgraph_node_map): Turn into pointer_map.
928 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
929 (del_node_map): Update.
930 (find_func_by_funcdef_no): Replace by ...
931 (find_func_by_profile_id): ... this one.
932 (gimple_ic_transform): Do not remove useful histograms when
933 speculation is not done; dump info when indirect call removal
934 can happen at LTO.
935 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
936 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
937 (__gcov_indirect_call_profiler_v2): .. this one.
938 * profile.h (init_node_map): Update.
939 * coverage.c (coverage_compute_profile_id): New function.
940 * coverage.h (coverage_compute_profile_id): Declare.
941 * tree-profile.c (init_ic_make_global_vars): Make
942 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
943 (gimple_init_edge_profiler): Update prototype of
944 __gcov_indirect_call_profiler.
945 (gimple_gen_ic_func_profiler): Simplify.
946 (tree_profiling): Use init_node_map
947
948 2013-08-09 Jan Hubicka <jh@suse.cz>
949
950 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
951 non-speculative refs.
952 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
953 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
954 (cgraph_set_call_stmt): Likewise.
955 (cgraph_create_edge_1): Fix release checking compilatoin;
956 clear lto_stmt_uid.
957 (cgraph_free_edge): Free indirect info.
958 (cgraph_turn_edge_to_speculative): New function.
959 (cgraph_speculative_call_info): New function.
960 (cgraph_make_edge_direct): Return direct edge; handle speculation.
961 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
962 (dump_cgraph_node): Dump speculation.
963 (verify_edge_count_and_frequency): Accept speculative edges.
964 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
965 (verify_cgraph_node): Handle speculation.
966 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
967 (cgraph_set_call_stmt): Update prototype.
968 (cgraph_make_edge_direct): Update prototype.
969 (cgraph_speculative_call_info): Declare.
970 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
971 to change; update call of ipa_find_references.
972 * ipa-ref.c (ipa_record_reference): Fix return value; clear
973 lto_stmt_uid and speculative flags.
974 (ipa_dump_references): Dump speculation.
975 (ipa_clone_references): Clone speculative flag.
976 (ipa_clone_referring): Likewise.
977 (ipa_clone_ref): New function.
978 (ipa_find_reference): Look into lto_stmt_uids
979 (ipa_clear_stmts_in_references): Do not clear speculative calls.
980 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
981 (ipa_find_reference): Update declaration.
982 (ipa_clone_ref): Declare.
983 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
984 stream speculative flag.
985 (lto_output_ref): Stream statements uids and speculation.
986 (input_ref): Likewise.
987 (input_edge): Stream speuclation.
988 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
989 (cgraph_set_call_stmt_including_clones): Handle speculation.
990 * ipa-inline.c (heap_edge_removal_hook): New function.
991 (inline_small_functions): Register it.
992 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
993 also initialize refs.
994 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
995 edge to change.
996 (try_make_edge_direct_simple_call): Likewise.
997 (try_make_edge_direct_simple_call): Likewise.
998 (update_indirect_edges_after_inlining): Likewise.
999 (remove_described_reference): Look proper lto_stmt_uid.
1000 (propagate_controlled_uses): Likewise.
1001 (propagate_controlled_uses): Liekwise.
1002 * tree-inline.c (copy_bb): Copy speculative edges.
1003 (redirect_all_calls): New function.
1004 (copy_cfg_body): Do redirection after loop info is updated.
1005 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1006
1007 2013-08-09 Jan Hubicka <jh@suse.cz>
1008
1009 * lto-streamer-out.c (output_function): Renumber PHIs.
1010 * lto-streamer-in.c (input_function): Likewise.
1011
1012 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1013
1014 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1015 (get_lane_unsigned): Likewise.
1016 (dup_lane_scalar): Likewise.
1017 (get_lane): enable for VALL.
1018 * config/aarch64/aarch64-simd.md
1019 (aarch64_dup_lane_scalar<mode>): Remove.
1020 (aarch64_get_lane_signed<mode>): Likewise.
1021 (aarch64_get_lane_unsigned<mode>): Likewise.
1022 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1023 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1024 (aarch64_get_lane<mode>): Enable for all vector modes.
1025 (aarch64_get_lanedi): Remove misleading constraints.
1026 * config/aarch64/arm_neon.h
1027 (__aarch64_vget_lane_any): Define.
1028 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1029 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1030 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1031 * config/aarch64/iterators.md (VDQQH): New.
1032 (VDQQHS): Likewise.
1033 (vwcore): Likewise.
1034
1035 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1036
1037 * configure.ac: Add GAS check for LEON instructions on SPARC.
1038 * configure: Regenerate.
1039 * config.in: Likewise.
1040 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1041 sparc*-*-* block.
1042 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1043 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1044 for LEON or LEON3.
1045 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1046 (AS_LEON_FLAG): New macro.
1047 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1048 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1049 Deal with LEON and LEON3 for the memory model.
1050 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1051 (atomic_compare_and_swap<mode>_1): Likewise.
1052 (*atomic_compare_and_swap<mode>_1): Likewise.
1053
1054 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1055
1056 * config/arm/neon.md (vcond): Fix floating-point vector
1057 comparisons against 0.
1058
1059 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1060
1061 * lra-constraints.c (emit_spill_move): Remove assert.
1062 (process_alt_operands): Add more debugging
1063 output. Increase reject for spilling into memory. Decrease
1064 reject for reloading scratch.
1065 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1066
1067 2013-08-08 Steve Ellcey <sellcey@mips.com>
1068
1069 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1070 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1071 micromips incompatible. Add nan2008.
1072 (MULTILIB_DIRNAMES): Add nan2008.
1073 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1074 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1075 and micromips incompatible. Add nan2008.
1076 (MULTILIB_DIRNAMES): Add nan2008.
1077 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1078
1079 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1080
1081 PR rtl-optimization/58079
1082 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1083 simplify_comparison has widened a comparison with an integer.
1084
1085 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1086
1087 * config/arm/neon.md (movmisalign<mode>): Disable when we
1088 don't allow unaligned accesses.
1089 (*movmisalign<mode>_neon_store): Likewise.
1090 (*movmisalign<mode>_neon_load): Likewise.
1091 (*movmisalign<mode>_neon_store): Likewise.
1092 (*movmisalign<mode>_neon_load): Likewise.
1093
1094 2013-08-08 Jan Hubicka <jh@suse.cz>
1095
1096 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
1097 (make_pass_rebuild_cgraph_edges): Also clear references.
1098 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
1099 * ipa-inline-transform.c (inline_transform): Remove all references
1100 after inlining.
1101 * cgraphunit.c (expand_function): Remove all references after
1102 expansion.
1103 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
1104 (ipa_find_reference): Rewrite to iterator.
1105 (remove_stmt_references): Likewise.
1106 (ipa_clear_stmts_in_references): New function.
1107 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
1108 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
1109 clear references.
1110 * ipa-split.c (split_function): Remove references in split function.
1111
1112 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1113
1114 PR target/57431
1115 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
1116 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
1117
1118 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1119
1120 PR target/56979
1121 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
1122 suggested mode for the assignment isn't compatible with the
1123 registers required.
1124
1125 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
1126
1127 PR target/58065
1128 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
1129
1130 2013-08-07 Xinliang David Li <davidxl@google.com>
1131
1132 * config/i386/i386.opt: New option -mtune-ctrl=.
1133 * config/i386/x86-tune.def: New file.
1134 * config/i386/i386.h: include x86-tune.def.
1135 * config/i386/i386.c (ix86_option_override_internal):
1136 Parsing -mtune-ctrl= option and set tune features.
1137
1138 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
1139
1140 PR other/12081
1141 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
1142 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
1143 to gen_2arg_fn_t.
1144
1145 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1146
1147 * rtl.h (update_alignments): Declare.
1148 * final.c (grow_label_align): New function extracted from...
1149 (shorten_branches): ...here. Call it.
1150 (update_alignments): New function.
1151 * reorg.c (sibling_labels): New variable.
1152 (get_label_before): Add SIBLING parameter. If it is non-zero, push
1153 the new label along with it onto the sibling_labels vector.
1154 (fill_simple_delay_slots): Adjust call to get_label_before.
1155 (fill_slots_from_thread): Likewise.
1156 (relax_delay_slots): Likewise.
1157 (make_return_insns): Likewise.
1158 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
1159
1160 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
1161
1162 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
1163 document its semantics.
1164 (diagnostic_report_diagnostic): Adjust accordingly.
1165
1166 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1167
1168 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
1169 (sparc_option_override): ...and port to new C++ pass API.
1170 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
1171
1172 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
1173
1174 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
1175
1176 2013-08-06 Caroline Tice <cmtice@google.com>
1177
1178 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
1179 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
1180 * tree-pass.h: Add pass_vtable_verify.
1181 * varasm.c (assemble_variable): Add code to properly set the comdat
1182 section and name for the .vtable_map_vars section.
1183 (assemble_vtyv_preinit_initializer): New function.
1184 (default_sectin_type_flags): Make sure .vtable_map_vars section has
1185 LINK_ONCE flag.
1186 * output.h: Add function decl for assemble_vtv_preinit_initializer.
1187 * vtable-verify.c: New file.
1188 * vtable-verify.h: New file.
1189 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
1190 initialiation levels.
1191 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
1192 * passes.def: Insert pass_vtable_verify.
1193 * aclocal.m4: Reorder includes.
1194 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
1195 -fvtv-counts options.
1196 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
1197 as appropriate, if -fvtable-verify=... is used.
1198 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
1199 -fvtable-verify=... is used.
1200 * Makefile.in (OBJS): Add vtable-verify.o to list.
1201 (vtable-verify.o): Add new build rule.
1202 (GTFILES): Add vtable-verify.c to list.
1203 * common.opt (fvtable-verify=): New flag.
1204 (vtv_priority): Values for fvtable-verify= flag.
1205 (fvtv-counts): New flag.
1206 (fvtv-debug): New flag.
1207 * tree.h (save_vtable_map_decl): New extern function decl.
1208
1209 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1210
1211 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
1212 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
1213 (pass_data_rl78_devirt): ...new pass_data instance and...
1214 (make_pass_rl78_devirt): ...new function.
1215 (rl78_asm_file_start): Port pass registration to new C++ API.
1216
1217 2013-08-07 David Malcolm <dmalcolm@redhat.com>
1218
1219 * coretypes.h (rtl_opt_pass): Add.
1220 (gcc::context): Add.
1221 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
1222 (epiphany_init): Port to new C++ pass API.
1223 (epiphany_optimize_mode_switching): Likewise.
1224 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
1225 (pass_manager::get_pass_mode_switching): New.
1226 (pass_manager::get_pass_peephole2): New.
1227 * mode-switching.c (pass_mode_switching): Add clone method.
1228 * recog.c (pass_peephole2): Add clone method.
1229 (pass_split_all_insns): Add clone method.
1230
1231 2013-08-06 David Malcolm <dmalcolm@redhat.com>
1232
1233 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
1234 (mips_option_override): ...here, porting to new C++ API for passes.
1235
1236 2013-08-06 Jan Hubicka <jh@suse.cz>
1237
1238 * cgraph.c (cgraph_get_body): New function based on lto.c
1239 implementation.
1240 * cgraph.h (cgraph_get_body): Declare.
1241 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
1242 LTO paths.
1243 * cgraphunit.c (expand_function): Get body prior expanding.
1244 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
1245 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
1246 really need.
1247 * passes.c (do_per_function_toporder): Get body.
1248 * tree-inline.c (expand_call_inline): Get body prior inlining it.
1249 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
1250
1251 2013-08-06 Martin Jambor <mjambor@suse.cz>
1252
1253 PR fortran/57987
1254 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
1255 is not re-finalized. Rename second parameter to no_collect.
1256
1257 2013-08-06 Martin Jambor <mjambor@suse.cz>
1258
1259 PR middle-end/58041
1260 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
1261 MEM_REF has proper alignment information.
1262
1263 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
1264
1265 PR other/12081
1266 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
1267 class insn_gen_fn.
1268 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
1269 rtx (*) (rtx, ...) with insn_gen_fn.
1270 * genoutput.c (output_insn_data): Cast gen_? function pointers to
1271 insn_gen_fn::stored_funcptr. Add initializer braces.
1272
1273 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1274
1275 Rewrite how instances of passes are cloned to remove assumptions
1276 about their sizes (thus allowing pass subclasses to have
1277 additional data fields, albeit non-GC-managed ones at this point).
1278
1279 * passes.c (make_pass_instance): Now that passes have clone
1280 methods, rewrite this function to eliminate XNEW and memcpy
1281 calls that used hardcoded sizes. Since this function no longer
1282 creates pass instances, rename it to...
1283 (add_pass_instance): ...this. Document the old way that passes were
1284 numbered and flagged, and rework this function to continue using it.
1285 (next_pass_1): Add an initial_pass argument for use by
1286 add_pass_instance.
1287 (position_pass): When adding multiple instances of a pass, use
1288 the pass's clone method, rather than relying on the XNEW/memcpy
1289 within the former make_pass_instance (now add_pass_instance).
1290 (pass_manager::pass_manager): When invoking next_pass_1, also supply
1291 the initial instance of the current pass within the pass manager.
1292
1293 2013-08-05 David Malcolm <dmalcolm@redhat.com>
1294
1295 This is the automated part of the conversion of passes from C
1296 structs to C++ classes.
1297
1298 Patch autogenerated by refactor_passes.py from
1299 https://github.com/davidmalcolm/gcc-refactoring-scripts
1300 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
1301
1302 * asan.c (pass_asan): Convert from a global struct to a subclass of
1303 gimple_opt_pass along with...
1304 (pass_data_asan): ...new pass_data instance and...
1305 (make_pass_asan): ...new function.
1306 (pass_asan_O0): Convert from a global struct to a subclass of
1307 gimple_opt_pass along with...
1308 (pass_data_asan_O0): ...new pass_data instance and...
1309 (make_pass_asan_O0): ...new function.
1310 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
1311 subclass of rtl_opt_pass along with...
1312 (pass_data_inc_dec): ...new pass_data instance and...
1313 (make_pass_inc_dec): ...new function.
1314 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
1315 a subclass of rtl_opt_pass along with...
1316 (pass_data_reorder_blocks): ...new pass_data instance and...
1317 (make_pass_reorder_blocks): ...new function.
1318 (pass_duplicate_computed_gotos): Convert from a global struct to a
1319 subclass of rtl_opt_pass along with...
1320 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
1321 (make_pass_duplicate_computed_gotos): ...new function.
1322 (pass_partition_blocks): Convert from a global struct to a subclass of
1323 rtl_opt_pass along with...
1324 (pass_data_partition_blocks): ...new pass_data instance and...
1325 (make_pass_partition_blocks): ...new function.
1326 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
1327 struct to a subclass of rtl_opt_pass along with...
1328 (pass_data_branch_target_load_optimize1): ...new pass_data instance
1329 and...
1330 (make_pass_branch_target_load_optimize1): ...new function.
1331 (pass_branch_target_load_optimize2): Convert from a global struct to a
1332 subclass of rtl_opt_pass along with...
1333 (pass_data_branch_target_load_optimize2): ...new pass_data instance
1334 and...
1335 (make_pass_branch_target_load_optimize2): ...new function.
1336 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
1337 of rtl_opt_pass along with...
1338 (pass_data_jump): ...new pass_data instance and...
1339 (make_pass_jump): ...new function.
1340 (pass_jump2): Convert from a global struct to a subclass of
1341 rtl_opt_pass along with...
1342 (pass_data_jump2): ...new pass_data instance and...
1343 (make_pass_jump2): ...new function.
1344 * cfgexpand.c (pass_expand): Convert from a global struct to a
1345 subclass of rtl_opt_pass along with...
1346 (pass_data_expand): ...new pass_data instance and...
1347 (make_pass_expand): ...new function.
1348 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
1349 of rtl_opt_pass along with...
1350 (pass_data_free_cfg): ...new pass_data instance and...
1351 (make_pass_free_cfg): ...new function.
1352 (pass_into_cfg_layout_mode): Convert from a global struct to a
1353 subclass of rtl_opt_pass along with...
1354 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
1355 (make_pass_into_cfg_layout_mode): ...new function.
1356 (pass_outof_cfg_layout_mode): Convert from a global struct to a
1357 subclass of rtl_opt_pass along with...
1358 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
1359 (make_pass_outof_cfg_layout_mode): ...new function.
1360 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
1361 struct to a subclass of gimple_opt_pass along with...
1362 (pass_data_build_cgraph_edges): ...new pass_data instance and...
1363 (make_pass_build_cgraph_edges): ...new function.
1364 (pass_rebuild_cgraph_edges): Convert from a global struct to a
1365 subclass of gimple_opt_pass along with...
1366 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
1367 (make_pass_rebuild_cgraph_edges): ...new function.
1368 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
1369 subclass of gimple_opt_pass along with...
1370 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
1371 and...
1372 (make_pass_remove_cgraph_callee_edges): ...new function.
1373 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
1374 struct to a subclass of rtl_opt_pass along with...
1375 (pass_data_stack_adjustments): ...new pass_data instance and...
1376 (make_pass_stack_adjustments): ...new function.
1377 * combine.c (pass_combine): Convert from a global struct to a subclass
1378 of rtl_opt_pass along with...
1379 (pass_data_combine): ...new pass_data instance and...
1380 (make_pass_combine): ...new function.
1381 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
1382 global struct to a subclass of rtl_opt_pass along with...
1383 (pass_data_compare_elim_after_reload): ...new pass_data instance
1384 and...
1385 (make_pass_compare_elim_after_reload): ...new function.
1386 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
1387 of rtl_opt_pass along with...
1388 (pass_data_rtl_cprop): ...new pass_data instance and...
1389 (make_pass_rtl_cprop): ...new function.
1390 * cse.c (pass_cse): Convert from a global struct to a subclass of
1391 rtl_opt_pass along with...
1392 (pass_data_cse): ...new pass_data instance and...
1393 (make_pass_cse): ...new function.
1394 (pass_cse2): Convert from a global struct to a subclass of
1395 rtl_opt_pass along with...
1396 (pass_data_cse2): ...new pass_data instance and...
1397 (make_pass_cse2): ...new function.
1398 (pass_cse_after_global_opts): Convert from a global struct to a
1399 subclass of rtl_opt_pass along with...
1400 (pass_data_cse_after_global_opts): ...new pass_data instance and...
1401 (make_pass_cse_after_global_opts): ...new function.
1402 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
1403 of rtl_opt_pass along with...
1404 (pass_data_ud_rtl_dce): ...new pass_data instance and...
1405 (make_pass_ud_rtl_dce): ...new function.
1406 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
1407 rtl_opt_pass along with...
1408 (pass_data_fast_rtl_dce): ...new pass_data instance and...
1409 (make_pass_fast_rtl_dce): ...new function.
1410 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
1411 a subclass of rtl_opt_pass along with...
1412 (pass_data_df_initialize_opt): ...new pass_data instance and...
1413 (make_pass_df_initialize_opt): ...new function.
1414 (pass_df_initialize_no_opt): Convert from a global struct to a
1415 subclass of rtl_opt_pass along with...
1416 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
1417 (make_pass_df_initialize_no_opt): ...new function.
1418 (pass_df_finish): Convert from a global struct to a subclass of
1419 rtl_opt_pass along with...
1420 (pass_data_df_finish): ...new pass_data instance and...
1421 (make_pass_df_finish): ...new function.
1422 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
1423 rtl_opt_pass along with...
1424 (pass_data_rtl_dse1): ...new pass_data instance and...
1425 (make_pass_rtl_dse1): ...new function.
1426 (pass_rtl_dse2): Convert from a global struct to a subclass of
1427 rtl_opt_pass along with...
1428 (pass_data_rtl_dse2): ...new pass_data instance and...
1429 (make_pass_rtl_dse2): ...new function.
1430 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
1431 subclass of rtl_opt_pass along with...
1432 (pass_data_dwarf2_frame): ...new pass_data instance and...
1433 (make_pass_dwarf2_frame): ...new function.
1434 * except.c (pass_set_nothrow_function_flags): Convert from a global
1435 struct to a subclass of rtl_opt_pass along with...
1436 (pass_data_set_nothrow_function_flags): ...new pass_data instance
1437 and...
1438 (make_pass_set_nothrow_function_flags): ...new function.
1439 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
1440 subclass of rtl_opt_pass along with...
1441 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
1442 and...
1443 (make_pass_convert_to_eh_region_ranges): ...new function.
1444 * final.c (pass_compute_alignments): Convert from a global struct to a
1445 subclass of rtl_opt_pass along with...
1446 (pass_data_compute_alignments): ...new pass_data instance and...
1447 (make_pass_compute_alignments): ...new function.
1448 (pass_final): Convert from a global struct to a subclass of
1449 rtl_opt_pass along with...
1450 (pass_data_final): ...new pass_data instance and...
1451 (make_pass_final): ...new function.
1452 (pass_shorten_branches): Convert from a global struct to a subclass of
1453 rtl_opt_pass along with...
1454 (pass_data_shorten_branches): ...new pass_data instance and...
1455 (make_pass_shorten_branches): ...new function.
1456 (pass_clean_state): Convert from a global struct to a subclass of
1457 rtl_opt_pass along with...
1458 (pass_data_clean_state): ...new pass_data instance and...
1459 (make_pass_clean_state): ...new function.
1460 * function.c (pass_instantiate_virtual_regs): Convert from a global
1461 struct to a subclass of rtl_opt_pass along with...
1462 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
1463 (make_pass_instantiate_virtual_regs): ...new function.
1464 (pass_leaf_regs): Convert from a global struct to a subclass of
1465 rtl_opt_pass along with...
1466 (pass_data_leaf_regs): ...new pass_data instance and...
1467 (make_pass_leaf_regs): ...new function.
1468 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
1469 subclass of rtl_opt_pass along with...
1470 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
1471 and...
1472 (make_pass_thread_prologue_and_epilogue): ...new function.
1473 (pass_match_asm_constraints): Convert from a global struct to a
1474 subclass of rtl_opt_pass along with...
1475 (pass_data_match_asm_constraints): ...new pass_data instance and...
1476 (make_pass_match_asm_constraints): ...new function.
1477 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
1478 subclass of rtl_opt_pass along with...
1479 (pass_data_rtl_fwprop): ...new pass_data instance and...
1480 (make_pass_rtl_fwprop): ...new function.
1481 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
1482 rtl_opt_pass along with...
1483 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
1484 (make_pass_rtl_fwprop_addr): ...new function.
1485 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
1486 rtl_opt_pass along with...
1487 (pass_data_rtl_pre): ...new pass_data instance and...
1488 (make_pass_rtl_pre): ...new function.
1489 (pass_rtl_hoist): Convert from a global struct to a subclass of
1490 rtl_opt_pass along with...
1491 (pass_data_rtl_hoist): ...new pass_data instance and...
1492 (make_pass_rtl_hoist): ...new function.
1493 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
1494 subclass of gimple_opt_pass along with...
1495 (pass_data_lower_cf): ...new pass_data instance and...
1496 (make_pass_lower_cf): ...new function.
1497 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
1498 from a global struct to a subclass of gimple_opt_pass along with...
1499 (pass_data_strength_reduction): ...new pass_data instance and...
1500 (make_pass_strength_reduction): ...new function.
1501 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
1502 of rtl_opt_pass along with...
1503 (pass_data_rtl_ifcvt): ...new pass_data instance and...
1504 (make_pass_rtl_ifcvt): ...new function.
1505 (pass_if_after_combine): Convert from a global struct to a subclass of
1506 rtl_opt_pass along with...
1507 (pass_data_if_after_combine): ...new pass_data instance and...
1508 (make_pass_if_after_combine): ...new function.
1509 (pass_if_after_reload): Convert from a global struct to a subclass of
1510 rtl_opt_pass along with...
1511 (pass_data_if_after_reload): ...new pass_data instance and...
1512 (make_pass_if_after_reload): ...new function.
1513 * init-regs.c (pass_initialize_regs): Convert from a global struct to
1514 a subclass of rtl_opt_pass along with...
1515 (pass_data_initialize_regs): ...new pass_data instance and...
1516 (make_pass_initialize_regs): ...new function.
1517 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
1518 of ipa_opt_pass_d along with...
1519 (pass_data_ipa_cp): ...new pass_data instance and...
1520 (make_pass_ipa_cp): ...new function.
1521 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
1522 global struct to a subclass of gimple_opt_pass along with...
1523 (pass_data_inline_parameters): ...new pass_data instance and...
1524 (make_pass_inline_parameters): ...new function.
1525 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
1526 subclass of gimple_opt_pass along with...
1527 (pass_data_early_inline): ...new pass_data instance and...
1528 (make_pass_early_inline): ...new function.
1529 (pass_ipa_inline): Convert from a global struct to a subclass of
1530 ipa_opt_pass_d along with...
1531 (pass_data_ipa_inline): ...new pass_data instance and...
1532 (make_pass_ipa_inline): ...new function.
1533 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
1534 struct to a subclass of gimple_opt_pass along with...
1535 (pass_data_local_pure_const): ...new pass_data instance and...
1536 (make_pass_local_pure_const): ...new function.
1537 (pass_ipa_pure_const): Convert from a global struct to a subclass of
1538 ipa_opt_pass_d along with...
1539 (pass_data_ipa_pure_const): ...new pass_data instance and...
1540 (make_pass_ipa_pure_const): ...new function.
1541 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
1542 to a subclass of ipa_opt_pass_d along with...
1543 (pass_data_ipa_reference): ...new pass_data instance and...
1544 (make_pass_ipa_reference): ...new function.
1545 * ipa-split.c (pass_split_functions): Convert from a global struct to
1546 a subclass of gimple_opt_pass along with...
1547 (pass_data_split_functions): ...new pass_data instance and...
1548 (make_pass_split_functions): ...new function.
1549 (pass_feedback_split_functions): Convert from a global struct to a
1550 subclass of gimple_opt_pass along with...
1551 (pass_data_feedback_split_functions): ...new pass_data instance and...
1552 (make_pass_feedback_split_functions): ...new function.
1553 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
1554 global struct to a subclass of simple_ipa_opt_pass along with...
1555 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
1556 instance and...
1557 (make_pass_ipa_function_and_variable_visibility): ...new function.
1558 (pass_ipa_free_inline_summary): Convert from a global struct to a
1559 subclass of simple_ipa_opt_pass along with...
1560 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
1561 (make_pass_ipa_free_inline_summary): ...new function.
1562 (pass_ipa_whole_program_visibility): Convert from a global struct to a
1563 subclass of ipa_opt_pass_d along with...
1564 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
1565 and...
1566 (make_pass_ipa_whole_program_visibility): ...new function.
1567 (pass_ipa_profile): Convert from a global struct to a subclass of
1568 ipa_opt_pass_d along with...
1569 (pass_data_ipa_profile): ...new pass_data instance and...
1570 (make_pass_ipa_profile): ...new function.
1571 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
1572 ipa_opt_pass_d along with...
1573 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
1574 (make_pass_ipa_cdtor_merge): ...new function.
1575 * ira.c (pass_ira): Convert from a global struct to a subclass of
1576 rtl_opt_pass along with...
1577 (pass_data_ira): ...new pass_data instance and...
1578 (make_pass_ira): ...new function.
1579 (pass_reload): Convert from a global struct to a subclass of
1580 rtl_opt_pass along with...
1581 (pass_data_reload): ...new pass_data instance and...
1582 (make_pass_reload): ...new function.
1583 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
1584 subclass of rtl_opt_pass along with...
1585 (pass_data_cleanup_barriers): ...new pass_data instance and...
1586 (make_pass_cleanup_barriers): ...new function.
1587 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
1588 of rtl_opt_pass along with...
1589 (pass_data_loop2): ...new pass_data instance and...
1590 (make_pass_loop2): ...new function.
1591 (pass_rtl_loop_init): Convert from a global struct to a subclass of
1592 rtl_opt_pass along with...
1593 (pass_data_rtl_loop_init): ...new pass_data instance and...
1594 (make_pass_rtl_loop_init): ...new function.
1595 (pass_rtl_loop_done): Convert from a global struct to a subclass of
1596 rtl_opt_pass along with...
1597 (pass_data_rtl_loop_done): ...new pass_data instance and...
1598 (make_pass_rtl_loop_done): ...new function.
1599 (pass_rtl_move_loop_invariants): Convert from a global struct to a
1600 subclass of rtl_opt_pass along with...
1601 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
1602 (make_pass_rtl_move_loop_invariants): ...new function.
1603 (pass_rtl_unswitch): Convert from a global struct to a subclass of
1604 rtl_opt_pass along with...
1605 (pass_data_rtl_unswitch): ...new pass_data instance and...
1606 (make_pass_rtl_unswitch): ...new function.
1607 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
1608 subclass of rtl_opt_pass along with...
1609 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
1610 and...
1611 (make_pass_rtl_unroll_and_peel_loops): ...new function.
1612 (pass_rtl_doloop): Convert from a global struct to a subclass of
1613 rtl_opt_pass along with...
1614 (pass_data_rtl_doloop): ...new pass_data instance and...
1615 (make_pass_rtl_doloop): ...new function.
1616 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
1617 a subclass of rtl_opt_pass along with...
1618 (pass_data_lower_subreg): ...new pass_data instance and...
1619 (make_pass_lower_subreg): ...new function.
1620 (pass_lower_subreg2): Convert from a global struct to a subclass of
1621 rtl_opt_pass along with...
1622 (pass_data_lower_subreg2): ...new pass_data instance and...
1623 (make_pass_lower_subreg2): ...new function.
1624 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
1625 struct to a subclass of ipa_opt_pass_d along with...
1626 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
1627 (make_pass_ipa_lto_gimple_out): ...new function.
1628 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
1629 of ipa_opt_pass_d along with...
1630 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
1631 (make_pass_ipa_lto_finish_out): ...new function.
1632 * mode-switching.c (pass_mode_switching): Convert from a global struct
1633 to a subclass of rtl_opt_pass along with...
1634 (pass_data_mode_switching): ...new pass_data instance and...
1635 (make_pass_mode_switching): ...new function.
1636 * modulo-sched.c (pass_sms): Convert from a global struct to a
1637 subclass of rtl_opt_pass along with...
1638 (pass_data_sms): ...new pass_data instance and...
1639 (make_pass_sms): ...new function.
1640 * omp-low.c (pass_expand_omp): Convert from a global struct to a
1641 subclass of gimple_opt_pass along with...
1642 (pass_data_expand_omp): ...new pass_data instance and...
1643 (make_pass_expand_omp): ...new function.
1644 (pass_lower_omp): Convert from a global struct to a subclass of
1645 gimple_opt_pass along with...
1646 (pass_data_lower_omp): ...new pass_data instance and...
1647 (make_pass_lower_omp): ...new function.
1648 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
1649 of gimple_opt_pass along with...
1650 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
1651 (make_pass_diagnose_omp_blocks): ...new function.
1652 * passes.c (pass_early_local_passes): Convert from a global struct to
1653 a subclass of simple_ipa_opt_pass along with...
1654 (pass_data_early_local_passes): ...new pass_data instance and...
1655 (make_pass_early_local_passes): ...new function.
1656 (pass_all_early_optimizations): Convert from a global struct to a
1657 subclass of gimple_opt_pass along with...
1658 (pass_data_all_early_optimizations): ...new pass_data instance and...
1659 (make_pass_all_early_optimizations): ...new function.
1660 (pass_all_optimizations): Convert from a global struct to a subclass
1661 of gimple_opt_pass along with...
1662 (pass_data_all_optimizations): ...new pass_data instance and...
1663 (make_pass_all_optimizations): ...new function.
1664 (pass_all_optimizations_g): Convert from a global struct to a subclass
1665 of gimple_opt_pass along with...
1666 (pass_data_all_optimizations_g): ...new pass_data instance and...
1667 (make_pass_all_optimizations_g): ...new function.
1668 (pass_rest_of_compilation): Convert from a global struct to a subclass
1669 of rtl_opt_pass along with...
1670 (pass_data_rest_of_compilation): ...new pass_data instance and...
1671 (make_pass_rest_of_compilation): ...new function.
1672 (pass_postreload): Convert from a global struct to a subclass of
1673 rtl_opt_pass along with...
1674 (pass_data_postreload): ...new pass_data instance and...
1675 (make_pass_postreload): ...new function.
1676 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
1677 subclass of rtl_opt_pass along with...
1678 (pass_data_gcse2): ...new pass_data instance and...
1679 (make_pass_gcse2): ...new function.
1680 * postreload.c (pass_postreload_cse): Convert from a global struct to
1681 a subclass of rtl_opt_pass along with...
1682 (pass_data_postreload_cse): ...new pass_data instance and...
1683 (make_pass_postreload_cse): ...new function.
1684 * predict.c (pass_profile): Convert from a global struct to a subclass
1685 of gimple_opt_pass along with...
1686 (pass_data_profile): ...new pass_data instance and...
1687 (make_pass_profile): ...new function.
1688 (pass_strip_predict_hints): Convert from a global struct to a subclass
1689 of gimple_opt_pass along with...
1690 (pass_data_strip_predict_hints): ...new pass_data instance and...
1691 (make_pass_strip_predict_hints): ...new function.
1692 * recog.c (pass_peephole2): Convert from a global struct to a subclass
1693 of rtl_opt_pass along with...
1694 (pass_data_peephole2): ...new pass_data instance and...
1695 (make_pass_peephole2): ...new function.
1696 (pass_split_all_insns): Convert from a global struct to a subclass of
1697 rtl_opt_pass along with...
1698 (pass_data_split_all_insns): ...new pass_data instance and...
1699 (make_pass_split_all_insns): ...new function.
1700 (pass_split_after_reload): Convert from a global struct to a subclass
1701 of rtl_opt_pass along with...
1702 (pass_data_split_after_reload): ...new pass_data instance and...
1703 (make_pass_split_after_reload): ...new function.
1704 (pass_split_before_regstack): Convert from a global struct to a
1705 subclass of rtl_opt_pass along with...
1706 (pass_data_split_before_regstack): ...new pass_data instance and...
1707 (make_pass_split_before_regstack): ...new function.
1708 (pass_split_before_sched2): Convert from a global struct to a subclass
1709 of rtl_opt_pass along with...
1710 (pass_data_split_before_sched2): ...new pass_data instance and...
1711 (make_pass_split_before_sched2): ...new function.
1712 (pass_split_for_shorten_branches): Convert from a global struct to a
1713 subclass of rtl_opt_pass along with...
1714 (pass_data_split_for_shorten_branches): ...new pass_data instance
1715 and...
1716 (make_pass_split_for_shorten_branches): ...new function.
1717 * ree.c (pass_ree): Convert from a global struct to a subclass of
1718 rtl_opt_pass along with...
1719 (pass_data_ree): ...new pass_data instance and...
1720 (make_pass_ree): ...new function.
1721 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
1722 subclass of rtl_opt_pass along with...
1723 (pass_data_stack_regs): ...new pass_data instance and...
1724 (make_pass_stack_regs): ...new function.
1725 (pass_stack_regs_run): Convert from a global struct to a subclass of
1726 rtl_opt_pass along with...
1727 (pass_data_stack_regs_run): ...new pass_data instance and...
1728 (make_pass_stack_regs_run): ...new function.
1729 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
1730 subclass of rtl_opt_pass along with...
1731 (pass_data_cprop_hardreg): ...new pass_data instance and...
1732 (make_pass_cprop_hardreg): ...new function.
1733 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
1734 subclass of rtl_opt_pass along with...
1735 (pass_data_reginfo_init): ...new pass_data instance and...
1736 (make_pass_reginfo_init): ...new function.
1737 * regmove.c (pass_regmove): Convert from a global struct to a subclass
1738 of rtl_opt_pass along with...
1739 (pass_data_regmove): ...new pass_data instance and...
1740 (make_pass_regmove): ...new function.
1741 * regrename.c (pass_regrename): Convert from a global struct to a
1742 subclass of rtl_opt_pass along with...
1743 (pass_data_regrename): ...new pass_data instance and...
1744 (make_pass_regrename): ...new function.
1745 * reorg.c (pass_delay_slots): Convert from a global struct to a
1746 subclass of rtl_opt_pass along with...
1747 (pass_data_delay_slots): ...new pass_data instance and...
1748 (make_pass_delay_slots): ...new function.
1749 (pass_machine_reorg): Convert from a global struct to a subclass of
1750 rtl_opt_pass along with...
1751 (pass_data_machine_reorg): ...new pass_data instance and...
1752 (make_pass_machine_reorg): ...new function.
1753 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
1754 of rtl_opt_pass along with...
1755 (pass_data_sched): ...new pass_data instance and...
1756 (make_pass_sched): ...new function.
1757 (pass_sched2): Convert from a global struct to a subclass of
1758 rtl_opt_pass along with...
1759 (pass_data_sched2): ...new pass_data instance and...
1760 (make_pass_sched2): ...new function.
1761 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
1762 to a subclass of rtl_opt_pass along with...
1763 (pass_data_stack_ptr_mod): ...new pass_data instance and...
1764 (make_pass_stack_ptr_mod): ...new function.
1765 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
1766 to a subclass of rtl_opt_pass along with...
1767 (pass_data_rtl_store_motion): ...new pass_data instance and...
1768 (make_pass_rtl_store_motion): ...new function.
1769 * tracer.c (pass_tracer): Convert from a global struct to a subclass
1770 of gimple_opt_pass along with...
1771 (pass_data_tracer): ...new pass_data instance and...
1772 (make_pass_tracer): ...new function.
1773 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
1774 to a subclass of gimple_opt_pass along with...
1775 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
1776 (make_pass_diagnose_tm_blocks): ...new function.
1777 (pass_lower_tm): Convert from a global struct to a subclass of
1778 gimple_opt_pass along with...
1779 (pass_data_lower_tm): ...new pass_data instance and...
1780 (make_pass_lower_tm): ...new function.
1781 (pass_tm_init): Convert from a global struct to a subclass of
1782 gimple_opt_pass along with...
1783 (pass_data_tm_init): ...new pass_data instance and...
1784 (make_pass_tm_init): ...new function.
1785 (pass_tm_mark): Convert from a global struct to a subclass of
1786 gimple_opt_pass along with...
1787 (pass_data_tm_mark): ...new pass_data instance and...
1788 (make_pass_tm_mark): ...new function.
1789 (pass_tm_edges): Convert from a global struct to a subclass of
1790 gimple_opt_pass along with...
1791 (pass_data_tm_edges): ...new pass_data instance and...
1792 (make_pass_tm_edges): ...new function.
1793 (pass_tm_memopt): Convert from a global struct to a subclass of
1794 gimple_opt_pass along with...
1795 (pass_data_tm_memopt): ...new pass_data instance and...
1796 (make_pass_tm_memopt): ...new function.
1797 (pass_ipa_tm): Convert from a global struct to a subclass of
1798 simple_ipa_opt_pass along with...
1799 (pass_data_ipa_tm): ...new pass_data instance and...
1800 (make_pass_ipa_tm): ...new function.
1801 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
1802 subclass of gimple_opt_pass along with...
1803 (pass_data_call_cdce): ...new pass_data instance and...
1804 (make_pass_call_cdce): ...new function.
1805 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
1806 subclass of gimple_opt_pass along with...
1807 (pass_data_build_cfg): ...new pass_data instance and...
1808 (make_pass_build_cfg): ...new function.
1809 (pass_split_crit_edges): Convert from a global struct to a subclass of
1810 gimple_opt_pass along with...
1811 (pass_data_split_crit_edges): ...new pass_data instance and...
1812 (make_pass_split_crit_edges): ...new function.
1813 (pass_warn_function_return): Convert from a global struct to a
1814 subclass of gimple_opt_pass along with...
1815 (pass_data_warn_function_return): ...new pass_data instance and...
1816 (make_pass_warn_function_return): ...new function.
1817 (pass_warn_function_noreturn): Convert from a global struct to a
1818 subclass of gimple_opt_pass along with...
1819 (pass_data_warn_function_noreturn): ...new pass_data instance and...
1820 (make_pass_warn_function_noreturn): ...new function.
1821 (pass_warn_unused_result): Convert from a global struct to a subclass
1822 of gimple_opt_pass along with...
1823 (pass_data_warn_unused_result): ...new pass_data instance and...
1824 (make_pass_warn_unused_result): ...new function.
1825 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
1826 a subclass of gimple_opt_pass along with...
1827 (pass_data_merge_phi): ...new pass_data instance and...
1828 (make_pass_merge_phi): ...new function.
1829 * tree-complex.c (pass_lower_complex): Convert from a global struct to
1830 a subclass of gimple_opt_pass along with...
1831 (pass_data_lower_complex): ...new pass_data instance and...
1832 (make_pass_lower_complex): ...new function.
1833 (pass_lower_complex_O0): Convert from a global struct to a subclass of
1834 gimple_opt_pass along with...
1835 (pass_data_lower_complex_O0): ...new pass_data instance and...
1836 (make_pass_lower_complex_O0): ...new function.
1837 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
1838 subclass of gimple_opt_pass along with...
1839 (pass_data_lower_eh): ...new pass_data instance and...
1840 (make_pass_lower_eh): ...new function.
1841 (pass_refactor_eh): Convert from a global struct to a subclass of
1842 gimple_opt_pass along with...
1843 (pass_data_refactor_eh): ...new pass_data instance and...
1844 (make_pass_refactor_eh): ...new function.
1845 (pass_lower_resx): Convert from a global struct to a subclass of
1846 gimple_opt_pass along with...
1847 (pass_data_lower_resx): ...new pass_data instance and...
1848 (make_pass_lower_resx): ...new function.
1849 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
1850 of gimple_opt_pass along with...
1851 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
1852 (make_pass_lower_eh_dispatch): ...new function.
1853 (pass_cleanup_eh): Convert from a global struct to a subclass of
1854 gimple_opt_pass along with...
1855 (pass_data_cleanup_eh): ...new pass_data instance and...
1856 (make_pass_cleanup_eh): ...new function.
1857 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
1858 to a subclass of simple_ipa_opt_pass along with...
1859 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
1860 (make_pass_ipa_lower_emutls): ...new function.
1861 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
1862 a subclass of gimple_opt_pass along with...
1863 (pass_data_if_conversion): ...new pass_data instance and...
1864 (make_pass_if_conversion): ...new function.
1865 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
1866 subclass of gimple_opt_pass along with...
1867 (pass_data_build_ssa): ...new pass_data instance and...
1868 (make_pass_build_ssa): ...new function.
1869 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
1870 global struct to a subclass of gimple_opt_pass along with...
1871 (pass_data_loop_distribution): ...new pass_data instance and...
1872 (make_pass_loop_distribution): ...new function.
1873 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
1874 subclass of gimple_opt_pass along with...
1875 (pass_data_mudflap_1): ...new pass_data instance and...
1876 (make_pass_mudflap_1): ...new function.
1877 (pass_mudflap_2): Convert from a global struct to a subclass of
1878 gimple_opt_pass along with...
1879 (pass_data_mudflap_2): ...new pass_data instance and...
1880 (make_pass_mudflap_2): ...new function.
1881 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
1882 subclass of gimple_opt_pass along with...
1883 (pass_data_mudflap_1): ...new pass_data instance and...
1884 (make_pass_mudflap_1): ...new function.
1885 (pass_mudflap_2): Convert from a global struct to a subclass of
1886 gimple_opt_pass along with...
1887 (pass_data_mudflap_2): ...new pass_data instance and...
1888 (make_pass_mudflap_2): ...new function.
1889 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
1890 gimple_opt_pass along with...
1891 (pass_data_nrv): ...new pass_data instance and...
1892 (make_pass_nrv): ...new function.
1893 (pass_return_slot): Convert from a global struct to a subclass of
1894 gimple_opt_pass along with...
1895 (pass_data_return_slot): ...new pass_data instance and...
1896 (make_pass_return_slot): ...new function.
1897 * tree-object-size.c (pass_object_sizes): Convert from a global struct
1898 to a subclass of gimple_opt_pass along with...
1899 (pass_data_object_sizes): ...new pass_data instance and...
1900 (make_pass_object_sizes): ...new function.
1901 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
1902 global struct to a subclass of gimple_opt_pass along with...
1903 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
1904 and...
1905 (make_pass_cleanup_cfg_post_optimizing): ...new function.
1906 (pass_fixup_cfg): Convert from a global struct to a subclass of
1907 gimple_opt_pass along with...
1908 (pass_data_fixup_cfg): ...new pass_data instance and...
1909 (make_pass_fixup_cfg): ...new function.
1910 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
1911 (make_pass_mudflap_1): ...new function.
1912 (pass_mudflap_2): Replace declaration with that of...
1913 (make_pass_mudflap_2): ...new function.
1914 (pass_asan): Replace declaration with that of...
1915 (make_pass_asan): ...new function.
1916 (pass_asan_O0): Replace declaration with that of...
1917 (make_pass_asan_O0): ...new function.
1918 (pass_tsan): Replace declaration with that of...
1919 (make_pass_tsan): ...new function.
1920 (pass_tsan_O0): Replace declaration with that of...
1921 (make_pass_tsan_O0): ...new function.
1922 (pass_lower_cf): Replace declaration with that of...
1923 (make_pass_lower_cf): ...new function.
1924 (pass_refactor_eh): Replace declaration with that of...
1925 (make_pass_refactor_eh): ...new function.
1926 (pass_lower_eh): Replace declaration with that of...
1927 (make_pass_lower_eh): ...new function.
1928 (pass_lower_eh_dispatch): Replace declaration with that of...
1929 (make_pass_lower_eh_dispatch): ...new function.
1930 (pass_lower_resx): Replace declaration with that of...
1931 (make_pass_lower_resx): ...new function.
1932 (pass_build_cfg): Replace declaration with that of...
1933 (make_pass_build_cfg): ...new function.
1934 (pass_early_tree_profile): Replace declaration with that of...
1935 (make_pass_early_tree_profile): ...new function.
1936 (pass_cleanup_eh): Replace declaration with that of...
1937 (make_pass_cleanup_eh): ...new function.
1938 (pass_sra): Replace declaration with that of...
1939 (make_pass_sra): ...new function.
1940 (pass_sra_early): Replace declaration with that of...
1941 (make_pass_sra_early): ...new function.
1942 (pass_early_ipa_sra): Replace declaration with that of...
1943 (make_pass_early_ipa_sra): ...new function.
1944 (pass_tail_recursion): Replace declaration with that of...
1945 (make_pass_tail_recursion): ...new function.
1946 (pass_tail_calls): Replace declaration with that of...
1947 (make_pass_tail_calls): ...new function.
1948 (pass_tree_loop): Replace declaration with that of...
1949 (make_pass_tree_loop): ...new function.
1950 (pass_tree_loop_init): Replace declaration with that of...
1951 (make_pass_tree_loop_init): ...new function.
1952 (pass_lim): Replace declaration with that of...
1953 (make_pass_lim): ...new function.
1954 (pass_tree_unswitch): Replace declaration with that of...
1955 (make_pass_tree_unswitch): ...new function.
1956 (pass_predcom): Replace declaration with that of...
1957 (make_pass_predcom): ...new function.
1958 (pass_iv_canon): Replace declaration with that of...
1959 (make_pass_iv_canon): ...new function.
1960 (pass_scev_cprop): Replace declaration with that of...
1961 (make_pass_scev_cprop): ...new function.
1962 (pass_empty_loop): Replace declaration with that of...
1963 (make_pass_empty_loop): ...new function.
1964 (pass_record_bounds): Replace declaration with that of...
1965 (make_pass_record_bounds): ...new function.
1966 (pass_graphite): Replace declaration with that of...
1967 (make_pass_graphite): ...new function.
1968 (pass_graphite_transforms): Replace declaration with that of...
1969 (make_pass_graphite_transforms): ...new function.
1970 (pass_if_conversion): Replace declaration with that of...
1971 (make_pass_if_conversion): ...new function.
1972 (pass_loop_distribution): Replace declaration with that of...
1973 (make_pass_loop_distribution): ...new function.
1974 (pass_vectorize): Replace declaration with that of...
1975 (make_pass_vectorize): ...new function.
1976 (pass_slp_vectorize): Replace declaration with that of...
1977 (make_pass_slp_vectorize): ...new function.
1978 (pass_complete_unroll): Replace declaration with that of...
1979 (make_pass_complete_unroll): ...new function.
1980 (pass_complete_unrolli): Replace declaration with that of...
1981 (make_pass_complete_unrolli): ...new function.
1982 (pass_parallelize_loops): Replace declaration with that of...
1983 (make_pass_parallelize_loops): ...new function.
1984 (pass_loop_prefetch): Replace declaration with that of...
1985 (make_pass_loop_prefetch): ...new function.
1986 (pass_iv_optimize): Replace declaration with that of...
1987 (make_pass_iv_optimize): ...new function.
1988 (pass_tree_loop_done): Replace declaration with that of...
1989 (make_pass_tree_loop_done): ...new function.
1990 (pass_ch): Replace declaration with that of...
1991 (make_pass_ch): ...new function.
1992 (pass_ccp): Replace declaration with that of...
1993 (make_pass_ccp): ...new function.
1994 (pass_phi_only_cprop): Replace declaration with that of...
1995 (make_pass_phi_only_cprop): ...new function.
1996 (pass_build_ssa): Replace declaration with that of...
1997 (make_pass_build_ssa): ...new function.
1998 (pass_build_alias): Replace declaration with that of...
1999 (make_pass_build_alias): ...new function.
2000 (pass_build_ealias): Replace declaration with that of...
2001 (make_pass_build_ealias): ...new function.
2002 (pass_dominator): Replace declaration with that of...
2003 (make_pass_dominator): ...new function.
2004 (pass_dce): Replace declaration with that of...
2005 (make_pass_dce): ...new function.
2006 (pass_dce_loop): Replace declaration with that of...
2007 (make_pass_dce_loop): ...new function.
2008 (pass_cd_dce): Replace declaration with that of...
2009 (make_pass_cd_dce): ...new function.
2010 (pass_call_cdce): Replace declaration with that of...
2011 (make_pass_call_cdce): ...new function.
2012 (pass_merge_phi): Replace declaration with that of...
2013 (make_pass_merge_phi): ...new function.
2014 (pass_split_crit_edges): Replace declaration with that of...
2015 (make_pass_split_crit_edges): ...new function.
2016 (pass_pre): Replace declaration with that of...
2017 (make_pass_pre): ...new function.
2018 (pass_profile): Replace declaration with that of...
2019 (make_pass_profile): ...new function.
2020 (pass_strip_predict_hints): Replace declaration with that of...
2021 (make_pass_strip_predict_hints): ...new function.
2022 (pass_lower_complex_O0): Replace declaration with that of...
2023 (make_pass_lower_complex_O0): ...new function.
2024 (pass_lower_complex): Replace declaration with that of...
2025 (make_pass_lower_complex): ...new function.
2026 (pass_lower_vector): Replace declaration with that of...
2027 (make_pass_lower_vector): ...new function.
2028 (pass_lower_vector_ssa): Replace declaration with that of...
2029 (make_pass_lower_vector_ssa): ...new function.
2030 (pass_lower_omp): Replace declaration with that of...
2031 (make_pass_lower_omp): ...new function.
2032 (pass_diagnose_omp_blocks): Replace declaration with that of...
2033 (make_pass_diagnose_omp_blocks): ...new function.
2034 (pass_expand_omp): Replace declaration with that of...
2035 (make_pass_expand_omp): ...new function.
2036 (pass_expand_omp_ssa): Replace declaration with that of...
2037 (make_pass_expand_omp_ssa): ...new function.
2038 (pass_object_sizes): Replace declaration with that of...
2039 (make_pass_object_sizes): ...new function.
2040 (pass_strlen): Replace declaration with that of...
2041 (make_pass_strlen): ...new function.
2042 (pass_fold_builtins): Replace declaration with that of...
2043 (make_pass_fold_builtins): ...new function.
2044 (pass_stdarg): Replace declaration with that of...
2045 (make_pass_stdarg): ...new function.
2046 (pass_early_warn_uninitialized): Replace declaration with that of...
2047 (make_pass_early_warn_uninitialized): ...new function.
2048 (pass_late_warn_uninitialized): Replace declaration with that of...
2049 (make_pass_late_warn_uninitialized): ...new function.
2050 (pass_cse_reciprocals): Replace declaration with that of...
2051 (make_pass_cse_reciprocals): ...new function.
2052 (pass_cse_sincos): Replace declaration with that of...
2053 (make_pass_cse_sincos): ...new function.
2054 (pass_optimize_bswap): Replace declaration with that of...
2055 (make_pass_optimize_bswap): ...new function.
2056 (pass_optimize_widening_mul): Replace declaration with that of...
2057 (make_pass_optimize_widening_mul): ...new function.
2058 (pass_warn_function_return): Replace declaration with that of...
2059 (make_pass_warn_function_return): ...new function.
2060 (pass_warn_function_noreturn): Replace declaration with that of...
2061 (make_pass_warn_function_noreturn): ...new function.
2062 (pass_cselim): Replace declaration with that of...
2063 (make_pass_cselim): ...new function.
2064 (pass_phiopt): Replace declaration with that of...
2065 (make_pass_phiopt): ...new function.
2066 (pass_forwprop): Replace declaration with that of...
2067 (make_pass_forwprop): ...new function.
2068 (pass_phiprop): Replace declaration with that of...
2069 (make_pass_phiprop): ...new function.
2070 (pass_tree_ifcombine): Replace declaration with that of...
2071 (make_pass_tree_ifcombine): ...new function.
2072 (pass_dse): Replace declaration with that of...
2073 (make_pass_dse): ...new function.
2074 (pass_nrv): Replace declaration with that of...
2075 (make_pass_nrv): ...new function.
2076 (pass_rename_ssa_copies): Replace declaration with that of...
2077 (make_pass_rename_ssa_copies): ...new function.
2078 (pass_sink_code): Replace declaration with that of...
2079 (make_pass_sink_code): ...new function.
2080 (pass_fre): Replace declaration with that of...
2081 (make_pass_fre): ...new function.
2082 (pass_check_data_deps): Replace declaration with that of...
2083 (make_pass_check_data_deps): ...new function.
2084 (pass_copy_prop): Replace declaration with that of...
2085 (make_pass_copy_prop): ...new function.
2086 (pass_vrp): Replace declaration with that of...
2087 (make_pass_vrp): ...new function.
2088 (pass_uncprop): Replace declaration with that of...
2089 (make_pass_uncprop): ...new function.
2090 (pass_return_slot): Replace declaration with that of...
2091 (make_pass_return_slot): ...new function.
2092 (pass_reassoc): Replace declaration with that of...
2093 (make_pass_reassoc): ...new function.
2094 (pass_rebuild_cgraph_edges): Replace declaration with that of...
2095 (make_pass_rebuild_cgraph_edges): ...new function.
2096 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
2097 (make_pass_remove_cgraph_callee_edges): ...new function.
2098 (pass_build_cgraph_edges): Replace declaration with that of...
2099 (make_pass_build_cgraph_edges): ...new function.
2100 (pass_local_pure_const): Replace declaration with that of...
2101 (make_pass_local_pure_const): ...new function.
2102 (pass_tracer): Replace declaration with that of...
2103 (make_pass_tracer): ...new function.
2104 (pass_warn_unused_result): Replace declaration with that of...
2105 (make_pass_warn_unused_result): ...new function.
2106 (pass_diagnose_tm_blocks): Replace declaration with that of...
2107 (make_pass_diagnose_tm_blocks): ...new function.
2108 (pass_lower_tm): Replace declaration with that of...
2109 (make_pass_lower_tm): ...new function.
2110 (pass_tm_init): Replace declaration with that of...
2111 (make_pass_tm_init): ...new function.
2112 (pass_tm_mark): Replace declaration with that of...
2113 (make_pass_tm_mark): ...new function.
2114 (pass_tm_memopt): Replace declaration with that of...
2115 (make_pass_tm_memopt): ...new function.
2116 (pass_tm_edges): Replace declaration with that of...
2117 (make_pass_tm_edges): ...new function.
2118 (pass_split_functions): Replace declaration with that of...
2119 (make_pass_split_functions): ...new function.
2120 (pass_feedback_split_functions): Replace declaration with that of...
2121 (make_pass_feedback_split_functions): ...new function.
2122 (pass_strength_reduction): Replace declaration with that of...
2123 (make_pass_strength_reduction): ...new function.
2124 (pass_ipa_lower_emutls): Replace declaration with that of...
2125 (make_pass_ipa_lower_emutls): ...new function.
2126 (pass_ipa_function_and_variable_visibility): Replace declaration with
2127 that of...
2128 (make_pass_ipa_function_and_variable_visibility): ...new function.
2129 (pass_ipa_tree_profile): Replace declaration with that of...
2130 (make_pass_ipa_tree_profile): ...new function.
2131 (pass_early_local_passes): Replace declaration with that of...
2132 (make_pass_early_local_passes): ...new function.
2133 (pass_ipa_whole_program_visibility): Replace declaration with that
2134 of...
2135 (make_pass_ipa_whole_program_visibility): ...new function.
2136 (pass_ipa_lto_gimple_out): Replace declaration with that of...
2137 (make_pass_ipa_lto_gimple_out): ...new function.
2138 (pass_ipa_increase_alignment): Replace declaration with that of...
2139 (make_pass_ipa_increase_alignment): ...new function.
2140 (pass_ipa_inline): Replace declaration with that of...
2141 (make_pass_ipa_inline): ...new function.
2142 (pass_ipa_free_lang_data): Replace declaration with that of...
2143 (make_pass_ipa_free_lang_data): ...new function.
2144 (pass_ipa_free_inline_summary): Replace declaration with that of...
2145 (make_pass_ipa_free_inline_summary): ...new function.
2146 (pass_ipa_cp): Replace declaration with that of...
2147 (make_pass_ipa_cp): ...new function.
2148 (pass_ipa_reference): Replace declaration with that of...
2149 (make_pass_ipa_reference): ...new function.
2150 (pass_ipa_pure_const): Replace declaration with that of...
2151 (make_pass_ipa_pure_const): ...new function.
2152 (pass_ipa_pta): Replace declaration with that of...
2153 (make_pass_ipa_pta): ...new function.
2154 (pass_ipa_lto_finish_out): Replace declaration with that of...
2155 (make_pass_ipa_lto_finish_out): ...new function.
2156 (pass_ipa_tm): Replace declaration with that of...
2157 (make_pass_ipa_tm): ...new function.
2158 (pass_ipa_profile): Replace declaration with that of...
2159 (make_pass_ipa_profile): ...new function.
2160 (pass_ipa_cdtor_merge): Replace declaration with that of...
2161 (make_pass_ipa_cdtor_merge): ...new function.
2162 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
2163 of...
2164 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2165 (pass_init_datastructures): Replace declaration with that of...
2166 (make_pass_init_datastructures): ...new function.
2167 (pass_fixup_cfg): Replace declaration with that of...
2168 (make_pass_fixup_cfg): ...new function.
2169 (pass_expand): Replace declaration with that of...
2170 (make_pass_expand): ...new function.
2171 (pass_instantiate_virtual_regs): Replace declaration with that of...
2172 (make_pass_instantiate_virtual_regs): ...new function.
2173 (pass_rtl_fwprop): Replace declaration with that of...
2174 (make_pass_rtl_fwprop): ...new function.
2175 (pass_rtl_fwprop_addr): Replace declaration with that of...
2176 (make_pass_rtl_fwprop_addr): ...new function.
2177 (pass_jump): Replace declaration with that of...
2178 (make_pass_jump): ...new function.
2179 (pass_jump2): Replace declaration with that of...
2180 (make_pass_jump2): ...new function.
2181 (pass_lower_subreg): Replace declaration with that of...
2182 (make_pass_lower_subreg): ...new function.
2183 (pass_cse): Replace declaration with that of...
2184 (make_pass_cse): ...new function.
2185 (pass_fast_rtl_dce): Replace declaration with that of...
2186 (make_pass_fast_rtl_dce): ...new function.
2187 (pass_ud_rtl_dce): Replace declaration with that of...
2188 (make_pass_ud_rtl_dce): ...new function.
2189 (pass_rtl_dce): Replace declaration with that of...
2190 (make_pass_rtl_dce): ...new function.
2191 (pass_rtl_dse1): Replace declaration with that of...
2192 (make_pass_rtl_dse1): ...new function.
2193 (pass_rtl_dse2): Replace declaration with that of...
2194 (make_pass_rtl_dse2): ...new function.
2195 (pass_rtl_dse3): Replace declaration with that of...
2196 (make_pass_rtl_dse3): ...new function.
2197 (pass_rtl_cprop): Replace declaration with that of...
2198 (make_pass_rtl_cprop): ...new function.
2199 (pass_rtl_pre): Replace declaration with that of...
2200 (make_pass_rtl_pre): ...new function.
2201 (pass_rtl_hoist): Replace declaration with that of...
2202 (make_pass_rtl_hoist): ...new function.
2203 (pass_rtl_store_motion): Replace declaration with that of...
2204 (make_pass_rtl_store_motion): ...new function.
2205 (pass_cse_after_global_opts): Replace declaration with that of...
2206 (make_pass_cse_after_global_opts): ...new function.
2207 (pass_rtl_ifcvt): Replace declaration with that of...
2208 (make_pass_rtl_ifcvt): ...new function.
2209 (pass_into_cfg_layout_mode): Replace declaration with that of...
2210 (make_pass_into_cfg_layout_mode): ...new function.
2211 (pass_outof_cfg_layout_mode): Replace declaration with that of...
2212 (make_pass_outof_cfg_layout_mode): ...new function.
2213 (pass_loop2): Replace declaration with that of...
2214 (make_pass_loop2): ...new function.
2215 (pass_rtl_loop_init): Replace declaration with that of...
2216 (make_pass_rtl_loop_init): ...new function.
2217 (pass_rtl_move_loop_invariants): Replace declaration with that of...
2218 (make_pass_rtl_move_loop_invariants): ...new function.
2219 (pass_rtl_unswitch): Replace declaration with that of...
2220 (make_pass_rtl_unswitch): ...new function.
2221 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
2222 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2223 (pass_rtl_doloop): Replace declaration with that of...
2224 (make_pass_rtl_doloop): ...new function.
2225 (pass_rtl_loop_done): Replace declaration with that of...
2226 (make_pass_rtl_loop_done): ...new function.
2227 (pass_web): Replace declaration with that of...
2228 (make_pass_web): ...new function.
2229 (pass_cse2): Replace declaration with that of...
2230 (make_pass_cse2): ...new function.
2231 (pass_df_initialize_opt): Replace declaration with that of...
2232 (make_pass_df_initialize_opt): ...new function.
2233 (pass_df_initialize_no_opt): Replace declaration with that of...
2234 (make_pass_df_initialize_no_opt): ...new function.
2235 (pass_reginfo_init): Replace declaration with that of...
2236 (make_pass_reginfo_init): ...new function.
2237 (pass_inc_dec): Replace declaration with that of...
2238 (make_pass_inc_dec): ...new function.
2239 (pass_stack_ptr_mod): Replace declaration with that of...
2240 (make_pass_stack_ptr_mod): ...new function.
2241 (pass_initialize_regs): Replace declaration with that of...
2242 (make_pass_initialize_regs): ...new function.
2243 (pass_combine): Replace declaration with that of...
2244 (make_pass_combine): ...new function.
2245 (pass_if_after_combine): Replace declaration with that of...
2246 (make_pass_if_after_combine): ...new function.
2247 (pass_ree): Replace declaration with that of...
2248 (make_pass_ree): ...new function.
2249 (pass_partition_blocks): Replace declaration with that of...
2250 (make_pass_partition_blocks): ...new function.
2251 (pass_match_asm_constraints): Replace declaration with that of...
2252 (make_pass_match_asm_constraints): ...new function.
2253 (pass_regmove): Replace declaration with that of...
2254 (make_pass_regmove): ...new function.
2255 (pass_split_all_insns): Replace declaration with that of...
2256 (make_pass_split_all_insns): ...new function.
2257 (pass_fast_rtl_byte_dce): Replace declaration with that of...
2258 (make_pass_fast_rtl_byte_dce): ...new function.
2259 (pass_lower_subreg2): Replace declaration with that of...
2260 (make_pass_lower_subreg2): ...new function.
2261 (pass_mode_switching): Replace declaration with that of...
2262 (make_pass_mode_switching): ...new function.
2263 (pass_sms): Replace declaration with that of...
2264 (make_pass_sms): ...new function.
2265 (pass_sched): Replace declaration with that of...
2266 (make_pass_sched): ...new function.
2267 (pass_ira): Replace declaration with that of...
2268 (make_pass_ira): ...new function.
2269 (pass_reload): Replace declaration with that of...
2270 (make_pass_reload): ...new function.
2271 (pass_clean_state): Replace declaration with that of...
2272 (make_pass_clean_state): ...new function.
2273 (pass_branch_prob): Replace declaration with that of...
2274 (make_pass_branch_prob): ...new function.
2275 (pass_value_profile_transformations): Replace declaration with that
2276 of...
2277 (make_pass_value_profile_transformations): ...new function.
2278 (pass_postreload_cse): Replace declaration with that of...
2279 (make_pass_postreload_cse): ...new function.
2280 (pass_gcse2): Replace declaration with that of...
2281 (make_pass_gcse2): ...new function.
2282 (pass_split_after_reload): Replace declaration with that of...
2283 (make_pass_split_after_reload): ...new function.
2284 (pass_branch_target_load_optimize1): Replace declaration with that
2285 of...
2286 (make_pass_branch_target_load_optimize1): ...new function.
2287 (pass_thread_prologue_and_epilogue): Replace declaration with that
2288 of...
2289 (make_pass_thread_prologue_and_epilogue): ...new function.
2290 (pass_stack_adjustments): Replace declaration with that of...
2291 (make_pass_stack_adjustments): ...new function.
2292 (pass_peephole2): Replace declaration with that of...
2293 (make_pass_peephole2): ...new function.
2294 (pass_if_after_reload): Replace declaration with that of...
2295 (make_pass_if_after_reload): ...new function.
2296 (pass_regrename): Replace declaration with that of...
2297 (make_pass_regrename): ...new function.
2298 (pass_cprop_hardreg): Replace declaration with that of...
2299 (make_pass_cprop_hardreg): ...new function.
2300 (pass_reorder_blocks): Replace declaration with that of...
2301 (make_pass_reorder_blocks): ...new function.
2302 (pass_branch_target_load_optimize2): Replace declaration with that
2303 of...
2304 (make_pass_branch_target_load_optimize2): ...new function.
2305 (pass_leaf_regs): Replace declaration with that of...
2306 (make_pass_leaf_regs): ...new function.
2307 (pass_split_before_sched2): Replace declaration with that of...
2308 (make_pass_split_before_sched2): ...new function.
2309 (pass_compare_elim_after_reload): Replace declaration with that of...
2310 (make_pass_compare_elim_after_reload): ...new function.
2311 (pass_sched2): Replace declaration with that of...
2312 (make_pass_sched2): ...new function.
2313 (pass_stack_regs): Replace declaration with that of...
2314 (make_pass_stack_regs): ...new function.
2315 (pass_stack_regs_run): Replace declaration with that of...
2316 (make_pass_stack_regs_run): ...new function.
2317 (pass_df_finish): Replace declaration with that of...
2318 (make_pass_df_finish): ...new function.
2319 (pass_compute_alignments): Replace declaration with that of...
2320 (make_pass_compute_alignments): ...new function.
2321 (pass_duplicate_computed_gotos): Replace declaration with that of...
2322 (make_pass_duplicate_computed_gotos): ...new function.
2323 (pass_variable_tracking): Replace declaration with that of...
2324 (make_pass_variable_tracking): ...new function.
2325 (pass_free_cfg): Replace declaration with that of...
2326 (make_pass_free_cfg): ...new function.
2327 (pass_machine_reorg): Replace declaration with that of...
2328 (make_pass_machine_reorg): ...new function.
2329 (pass_cleanup_barriers): Replace declaration with that of...
2330 (make_pass_cleanup_barriers): ...new function.
2331 (pass_delay_slots): Replace declaration with that of...
2332 (make_pass_delay_slots): ...new function.
2333 (pass_split_for_shorten_branches): Replace declaration with that of...
2334 (make_pass_split_for_shorten_branches): ...new function.
2335 (pass_split_before_regstack): Replace declaration with that of...
2336 (make_pass_split_before_regstack): ...new function.
2337 (pass_convert_to_eh_region_ranges): Replace declaration with that
2338 of...
2339 (make_pass_convert_to_eh_region_ranges): ...new function.
2340 (pass_shorten_branches): Replace declaration with that of...
2341 (make_pass_shorten_branches): ...new function.
2342 (pass_set_nothrow_function_flags): Replace declaration with that of...
2343 (make_pass_set_nothrow_function_flags): ...new function.
2344 (pass_dwarf2_frame): Replace declaration with that of...
2345 (make_pass_dwarf2_frame): ...new function.
2346 (pass_final): Replace declaration with that of...
2347 (make_pass_final): ...new function.
2348 (pass_rtl_seqabstr): Replace declaration with that of...
2349 (make_pass_rtl_seqabstr): ...new function.
2350 (pass_release_ssa_names): Replace declaration with that of...
2351 (make_pass_release_ssa_names): ...new function.
2352 (pass_early_inline): Replace declaration with that of...
2353 (make_pass_early_inline): ...new function.
2354 (pass_inline_parameters): Replace declaration with that of...
2355 (make_pass_inline_parameters): ...new function.
2356 (pass_update_address_taken): Replace declaration with that of...
2357 (make_pass_update_address_taken): ...new function.
2358 (pass_convert_switch): Replace declaration with that of...
2359 (make_pass_convert_switch): ...new function.
2360 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
2361 to a subclass of simple_ipa_opt_pass along with...
2362 (pass_data_ipa_tree_profile): ...new pass_data instance and...
2363 (make_pass_ipa_tree_profile): ...new function.
2364 * tree-sra.c (pass_sra_early): Convert from a global struct to a
2365 subclass of gimple_opt_pass along with...
2366 (pass_data_sra_early): ...new pass_data instance and...
2367 (make_pass_sra_early): ...new function.
2368 (pass_sra): Convert from a global struct to a subclass of
2369 gimple_opt_pass along with...
2370 (pass_data_sra): ...new pass_data instance and...
2371 (make_pass_sra): ...new function.
2372 (pass_early_ipa_sra): Convert from a global struct to a subclass of
2373 gimple_opt_pass along with...
2374 (pass_data_early_ipa_sra): ...new pass_data instance and...
2375 (make_pass_early_ipa_sra): ...new function.
2376 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
2377 subclass of gimple_opt_pass along with...
2378 (pass_data_ccp): ...new pass_data instance and...
2379 (make_pass_ccp): ...new function.
2380 (pass_fold_builtins): Convert from a global struct to a subclass of
2381 gimple_opt_pass along with...
2382 (pass_data_fold_builtins): ...new pass_data instance and...
2383 (make_pass_fold_builtins): ...new function.
2384 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
2385 subclass of gimple_opt_pass along with...
2386 (pass_data_copy_prop): ...new pass_data instance and...
2387 (make_pass_copy_prop): ...new function.
2388 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
2389 global struct to a subclass of gimple_opt_pass along with...
2390 (pass_data_rename_ssa_copies): ...new pass_data instance and...
2391 (make_pass_rename_ssa_copies): ...new function.
2392 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
2393 subclass of gimple_opt_pass along with...
2394 (pass_data_dce): ...new pass_data instance and...
2395 (make_pass_dce): ...new function.
2396 (pass_dce_loop): Convert from a global struct to a subclass of
2397 gimple_opt_pass along with...
2398 (pass_data_dce_loop): ...new pass_data instance and...
2399 (make_pass_dce_loop): ...new function.
2400 (pass_cd_dce): Convert from a global struct to a subclass of
2401 gimple_opt_pass along with...
2402 (pass_data_cd_dce): ...new pass_data instance and...
2403 (make_pass_cd_dce): ...new function.
2404 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
2405 subclass of gimple_opt_pass along with...
2406 (pass_data_dominator): ...new pass_data instance and...
2407 (make_pass_dominator): ...new function.
2408 (pass_phi_only_cprop): Convert from a global struct to a subclass of
2409 gimple_opt_pass along with...
2410 (pass_data_phi_only_cprop): ...new pass_data instance and...
2411 (make_pass_phi_only_cprop): ...new function.
2412 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
2413 subclass of gimple_opt_pass along with...
2414 (pass_data_dse): ...new pass_data instance and...
2415 (make_pass_dse): ...new function.
2416 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
2417 a subclass of gimple_opt_pass along with...
2418 (pass_data_forwprop): ...new pass_data instance and...
2419 (make_pass_forwprop): ...new function.
2420 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
2421 struct to a subclass of gimple_opt_pass along with...
2422 (pass_data_tree_ifcombine): ...new pass_data instance and...
2423 (make_pass_tree_ifcombine): ...new function.
2424 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
2425 subclass of gimple_opt_pass along with...
2426 (pass_data_ch): ...new pass_data instance and...
2427 (make_pass_ch): ...new function.
2428 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
2429 subclass of gimple_opt_pass along with...
2430 (pass_data_tree_loop): ...new pass_data instance and...
2431 (make_pass_tree_loop): ...new function.
2432 (pass_tree_loop_init): Convert from a global struct to a subclass of
2433 gimple_opt_pass along with...
2434 (pass_data_tree_loop_init): ...new pass_data instance and...
2435 (make_pass_tree_loop_init): ...new function.
2436 (pass_lim): Convert from a global struct to a subclass of
2437 gimple_opt_pass along with...
2438 (pass_data_lim): ...new pass_data instance and...
2439 (make_pass_lim): ...new function.
2440 (pass_tree_unswitch): Convert from a global struct to a subclass of
2441 gimple_opt_pass along with...
2442 (pass_data_tree_unswitch): ...new pass_data instance and...
2443 (make_pass_tree_unswitch): ...new function.
2444 (pass_predcom): Convert from a global struct to a subclass of
2445 gimple_opt_pass along with...
2446 (pass_data_predcom): ...new pass_data instance and...
2447 (make_pass_predcom): ...new function.
2448 (pass_vectorize): Convert from a global struct to a subclass of
2449 gimple_opt_pass along with...
2450 (pass_data_vectorize): ...new pass_data instance and...
2451 (make_pass_vectorize): ...new function.
2452 (pass_graphite): Convert from a global struct to a subclass of
2453 gimple_opt_pass along with...
2454 (pass_data_graphite): ...new pass_data instance and...
2455 (make_pass_graphite): ...new function.
2456 (pass_graphite_transforms): Convert from a global struct to a subclass
2457 of gimple_opt_pass along with...
2458 (pass_data_graphite_transforms): ...new pass_data instance and...
2459 (make_pass_graphite_transforms): ...new function.
2460 (pass_check_data_deps): Convert from a global struct to a subclass of
2461 gimple_opt_pass along with...
2462 (pass_data_check_data_deps): ...new pass_data instance and...
2463 (make_pass_check_data_deps): ...new function.
2464 (pass_iv_canon): Convert from a global struct to a subclass of
2465 gimple_opt_pass along with...
2466 (pass_data_iv_canon): ...new pass_data instance and...
2467 (make_pass_iv_canon): ...new function.
2468 (pass_scev_cprop): Convert from a global struct to a subclass of
2469 gimple_opt_pass along with...
2470 (pass_data_scev_cprop): ...new pass_data instance and...
2471 (make_pass_scev_cprop): ...new function.
2472 (pass_record_bounds): Convert from a global struct to a subclass of
2473 gimple_opt_pass along with...
2474 (pass_data_record_bounds): ...new pass_data instance and...
2475 (make_pass_record_bounds): ...new function.
2476 (pass_complete_unroll): Convert from a global struct to a subclass of
2477 gimple_opt_pass along with...
2478 (pass_data_complete_unroll): ...new pass_data instance and...
2479 (make_pass_complete_unroll): ...new function.
2480 (pass_complete_unrolli): Convert from a global struct to a subclass of
2481 gimple_opt_pass along with...
2482 (pass_data_complete_unrolli): ...new pass_data instance and...
2483 (make_pass_complete_unrolli): ...new function.
2484 (pass_parallelize_loops): Convert from a global struct to a subclass
2485 of gimple_opt_pass along with...
2486 (pass_data_parallelize_loops): ...new pass_data instance and...
2487 (make_pass_parallelize_loops): ...new function.
2488 (pass_loop_prefetch): Convert from a global struct to a subclass of
2489 gimple_opt_pass along with...
2490 (pass_data_loop_prefetch): ...new pass_data instance and...
2491 (make_pass_loop_prefetch): ...new function.
2492 (pass_iv_optimize): Convert from a global struct to a subclass of
2493 gimple_opt_pass along with...
2494 (pass_data_iv_optimize): ...new pass_data instance and...
2495 (make_pass_iv_optimize): ...new function.
2496 (pass_tree_loop_done): Convert from a global struct to a subclass of
2497 gimple_opt_pass along with...
2498 (pass_data_tree_loop_done): ...new pass_data instance and...
2499 (make_pass_tree_loop_done): ...new function.
2500 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
2501 struct to a subclass of gimple_opt_pass along with...
2502 (pass_data_cse_reciprocals): ...new pass_data instance and...
2503 (make_pass_cse_reciprocals): ...new function.
2504 (pass_cse_sincos): Convert from a global struct to a subclass of
2505 gimple_opt_pass along with...
2506 (pass_data_cse_sincos): ...new pass_data instance and...
2507 (make_pass_cse_sincos): ...new function.
2508 (pass_optimize_bswap): Convert from a global struct to a subclass of
2509 gimple_opt_pass along with...
2510 (pass_data_optimize_bswap): ...new pass_data instance and...
2511 (make_pass_optimize_bswap): ...new function.
2512 (pass_optimize_widening_mul): Convert from a global struct to a
2513 subclass of gimple_opt_pass along with...
2514 (pass_data_optimize_widening_mul): ...new pass_data instance and...
2515 (make_pass_optimize_widening_mul): ...new function.
2516 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
2517 subclass of gimple_opt_pass along with...
2518 (pass_data_phiopt): ...new pass_data instance and...
2519 (make_pass_phiopt): ...new function.
2520 (pass_cselim): Convert from a global struct to a subclass of
2521 gimple_opt_pass along with...
2522 (pass_data_cselim): ...new pass_data instance and...
2523 (make_pass_cselim): ...new function.
2524 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
2525 subclass of gimple_opt_pass along with...
2526 (pass_data_phiprop): ...new pass_data instance and...
2527 (make_pass_phiprop): ...new function.
2528 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
2529 subclass of gimple_opt_pass along with...
2530 (pass_data_pre): ...new pass_data instance and...
2531 (make_pass_pre): ...new function.
2532 (pass_fre): Convert from a global struct to a subclass of
2533 gimple_opt_pass along with...
2534 (pass_data_fre): ...new pass_data instance and...
2535 (make_pass_fre): ...new function.
2536 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
2537 subclass of gimple_opt_pass along with...
2538 (pass_data_reassoc): ...new pass_data instance and...
2539 (make_pass_reassoc): ...new function.
2540 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
2541 subclass of gimple_opt_pass along with...
2542 (pass_data_sink_code): ...new pass_data instance and...
2543 (make_pass_sink_code): ...new function.
2544 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
2545 subclass of gimple_opt_pass along with...
2546 (pass_data_strlen): ...new pass_data instance and...
2547 (make_pass_strlen): ...new function.
2548 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
2549 struct to a subclass of gimple_opt_pass along with...
2550 (pass_data_build_alias): ...new pass_data instance and...
2551 (make_pass_build_alias): ...new function.
2552 (pass_build_ealias): Convert from a global struct to a subclass of
2553 gimple_opt_pass along with...
2554 (pass_data_build_ealias): ...new pass_data instance and...
2555 (make_pass_build_ealias): ...new function.
2556 (pass_ipa_pta): Convert from a global struct to a subclass of
2557 simple_ipa_opt_pass along with...
2558 (pass_data_ipa_pta): ...new pass_data instance and...
2559 (make_pass_ipa_pta): ...new function.
2560 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
2561 subclass of gimple_opt_pass along with...
2562 (pass_data_uncprop): ...new pass_data instance and...
2563 (make_pass_uncprop): ...new function.
2564 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
2565 global struct to a subclass of gimple_opt_pass along with...
2566 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
2567 (make_pass_late_warn_uninitialized): ...new function.
2568 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
2569 to a subclass of gimple_opt_pass along with...
2570 (pass_data_init_datastructures): ...new pass_data instance and...
2571 (make_pass_init_datastructures): ...new function.
2572 (pass_early_warn_uninitialized): Convert from a global struct to a
2573 subclass of gimple_opt_pass along with...
2574 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
2575 (make_pass_early_warn_uninitialized): ...new function.
2576 (pass_update_address_taken): Convert from a global struct to a
2577 subclass of gimple_opt_pass along with...
2578 (pass_data_update_address_taken): ...new pass_data instance and...
2579 (make_pass_update_address_taken): ...new function.
2580 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
2581 struct to a subclass of gimple_opt_pass along with...
2582 (pass_data_release_ssa_names): ...new pass_data instance and...
2583 (make_pass_release_ssa_names): ...new function.
2584 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
2585 subclass of gimple_opt_pass along with...
2586 (pass_data_stdarg): ...new pass_data instance and...
2587 (make_pass_stdarg): ...new function.
2588 * tree-switch-conversion.c (pass_convert_switch): Convert from a
2589 global struct to a subclass of gimple_opt_pass along with...
2590 (pass_data_convert_switch): ...new pass_data instance and...
2591 (make_pass_convert_switch): ...new function.
2592 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
2593 to a subclass of gimple_opt_pass along with...
2594 (pass_data_tail_recursion): ...new pass_data instance and...
2595 (make_pass_tail_recursion): ...new function.
2596 (pass_tail_calls): Convert from a global struct to a subclass of
2597 gimple_opt_pass along with...
2598 (pass_data_tail_calls): ...new pass_data instance and...
2599 (make_pass_tail_calls): ...new function.
2600 * tree-vect-generic.c (pass_lower_vector): Convert from a global
2601 struct to a subclass of gimple_opt_pass along with...
2602 (pass_data_lower_vector): ...new pass_data instance and...
2603 (make_pass_lower_vector): ...new function.
2604 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
2605 gimple_opt_pass along with...
2606 (pass_data_lower_vector_ssa): ...new pass_data instance and...
2607 (make_pass_lower_vector_ssa): ...new function.
2608 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
2609 to a subclass of gimple_opt_pass along with...
2610 (pass_data_slp_vectorize): ...new pass_data instance and...
2611 (make_pass_slp_vectorize): ...new function.
2612 (pass_ipa_increase_alignment): Convert from a global struct to a
2613 subclass of simple_ipa_opt_pass along with...
2614 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
2615 (make_pass_ipa_increase_alignment): ...new function.
2616 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
2617 gimple_opt_pass along with...
2618 (pass_data_vrp): ...new pass_data instance and...
2619 (make_pass_vrp): ...new function.
2620 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
2621 subclass of simple_ipa_opt_pass along with...
2622 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
2623 (make_pass_ipa_free_lang_data): ...new function.
2624 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
2625 gimple_opt_pass along with...
2626 (pass_data_tsan): ...new pass_data instance and...
2627 (make_pass_tsan): ...new function.
2628 (pass_tsan_O0): Convert from a global struct to a subclass of
2629 gimple_opt_pass along with...
2630 (pass_data_tsan_O0): ...new pass_data instance and...
2631 (make_pass_tsan_O0): ...new function.
2632 * var-tracking.c (pass_variable_tracking): Convert from a global
2633 struct to a subclass of rtl_opt_pass along with...
2634 (pass_data_variable_tracking): ...new pass_data instance and...
2635 (make_pass_variable_tracking): ...new function.
2636 * web.c (pass_web): Convert from a global struct to a subclass of
2637 rtl_opt_pass along with...
2638 (pass_data_web): ...new pass_data instance and...
2639 (make_pass_web): ...new function.
2640 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
2641 declaration with that of...
2642 (make_pass_mode_switch_use): ...new function.
2643 (pass_resolve_sw_modes): Replace declaration with that of...
2644 (make_pass_resolve_sw_modes): ...new function.
2645 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
2646 from a global struct to a subclass of rtl_opt_pass along with...
2647 (pass_data_mode_switch_use): ...new pass_data instance and...
2648 (make_pass_mode_switch_use): ...new function.
2649 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
2650 from a global struct to a subclass of rtl_opt_pass along with...
2651 (pass_data_resolve_sw_modes): ...new pass_data instance and...
2652 (make_pass_resolve_sw_modes): ...new function.
2653 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
2654 struct to a subclass of rtl_opt_pass along with...
2655 (pass_data_insert_vzeroupper): ...new pass_data instance and...
2656 (make_pass_insert_vzeroupper): ...new function.
2657 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
2658 global struct to a subclass of rtl_opt_pass along with...
2659 (pass_data_work_around_errata): ...new pass_data instance and...
2660 (make_pass_work_around_errata): ...new function.
2661 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
2662 struct to a subclass of rtl_opt_pass along with...
2663 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
2664 (make_pass_mips_machine_reorg2): ...new function.
2665
2666 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2667
2668 * passes.c (pass_manager::operator new): New.
2669
2670 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2671
2672 Handwritten part of conversion of passes to C++ classes.
2673
2674 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
2675 (toplev.o): Add dep on PASS_MANAGER_H.
2676 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
2677 of early local pases to reflect this moving from a global to a
2678 member of gcc::pass_manager.
2679 (cgraph_add_new_function): Likewise.
2680 * lto-cgraph.c (lto_output_node): Update for conversion of
2681 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
2682 * passes.c (opt_pass::clone): New.
2683 (opt_pass::gate): New.
2684 (opt_pass::execute): New.
2685 (opt_pass::opt_pass): New.
2686 (pass_manager::execute_early_local_passes): New.
2687 (pass_manager::execute_pass_mode_switching): new.
2688 (finish_optimization_passes): Convert to...
2689 (pass_manager::finish_optimization_passes): ...this.
2690 (finish_optimization_passes): Update for conversion of passes to
2691 C++ classes.
2692 (register_dump_files_1): Use has_gate since we cannot portably
2693 check a vtable entry against NULL.
2694 (dump_one_pass): Likewise.
2695 (ipa_write_summaries_2): Likewise.
2696 (ipa_write_optimization_summaries_1): Likewise.
2697 (ipa_read_summaries_1): Likewise.
2698 (ipa_read_optimization_summaries_1): Likewise.
2699 (execute_ipa_stmt_fixups): Likewise.
2700 (pass_manager::pass_manager): Rewrite pass-creation, invoking
2701 pass-creation functions rather than wiring up globals, and
2702 storing the results in fields of pass_manager generated using
2703 pass-instances.def.
2704 (pass_manager::dump_profile_report): Update for conversion of
2705 passes to C++ classes.
2706 (pass_manager::execute_ipa_summary_passes): Likewise.
2707 (execute_one_ipa_transform_pass): Likewise.
2708 (execute_one_pass): Use has_gate and has_execute since we cannot
2709 portably check a vtable entry against NULL.
2710 * pass_manager.h (pass_manager::finish_optimization_passes): New.
2711 (pass_manager): Use pass-instances.def to add fields for the
2712 various pass instances.
2713 * toplev.c (finalize): Update for move of
2714 finish_optimization_passes to a method of gcc::pass_manager.
2715 * toplev.h (finish_optimization_passes): Move to method of class
2716 pass_manager.
2717 * tree-pass.h (struct pass_data): New.
2718 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
2719 (opt_pass::gate): Convert to virtual function.
2720 (opt_pass::~opt_pass): New.
2721 (opt_pass::clone): New.
2722 (opt_pass::execute): Convert to virtual function.
2723 (opt_pass::opt_pass): New.
2724 (opt_pass::ctxt_): new.
2725 (gimple_opt_pass): Convert to subclass of opt_pass.
2726 (gimple_opt_pass::gimple_opt_pass): New.
2727 (rtl_opt_pass): Convert to subclass of opt_pass.
2728 (rtl_opt_pass::rtl_opt_pass): New.
2729 (ipa_opt_pass_d): Convert to subclass of opt_pass.
2730 (ipa_opt_pass_d::ipa_opt_pass_d): New.
2731 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
2732 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
2733 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
2734 invocation of pass_mode_switching to reflect this moving from a
2735 global to a member of gcc::pass_manager.
2736 (ix86_option_override): Rework how pass_insert_vzeroupper is
2737 added to the pass_manager to reflect autogenerated changes.
2738 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
2739
2740 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
2741
2742 PR rtl-optimization/57708
2743 * recog.c (peep2_find_free_register): Validate all regs in a
2744 multi-reg mode.
2745
2746 2013-08-05 Jan Hubicka <jh@suse.cz>
2747
2748 PR lto/57602
2749 * cgraph.c (verify_cgraph_node): Accept local flags from other
2750 partitions.
2751 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
2752 (function_and_variable_visibility): Likewise.
2753 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
2754
2755 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2756
2757 * graph.c (init_graph_slim_pretty_print): Remove.
2758 (print_graph_cfg): Do not call it. Use local pretty printer.
2759 (start_graph_dump): Likewise.
2760
2761 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2762
2763 * gimple-pretty-print.c (buffer): Remove.
2764 (initialized): Likewise.
2765 (maybe_init_pretty_print): Likewise.
2766 (print_gimple_stmt): Do not call it. Use non-static local
2767 pretty_printer variable.
2768 (print_gimple_expr): Likewise.
2769 (print_gimple_seq): Likewise.
2770 (gimple_dump_bb): Likewise.
2771
2772 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
2773
2774 * asan.c (asan_pp): Remove.
2775 (asan_pp_initialized): Likewise.
2776 (asan_pp_initialize): Likewise.
2777 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
2778 (asan_emit_stack_protection): Tidy. Use local pretty printer.
2779 (asan_add_global): Likewise.
2780
2781 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
2782
2783 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
2784 * diagnostic.h (diagnostic_flush_buffer): Adjust.
2785 * pretty-print.c (pp_formatted_text_data): Likewise.
2786 (pp_indent): Rename from pp_base_indent.
2787 (pp_format): Rename from pp_base_format.
2788 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
2789 (pp_format_verbatim): Rename from pp_base_format_verbatim.
2790 (pp_flush): Rename from pp_base_flush.
2791 (pp_set_line_maximum_length): Rename from
2792 pp_base_set_line_maximum_length.
2793 (pp_clear_output_area): Rename from pp_base_clear_output_area.
2794 (pp_set_prefix): Rename from pp_base_set_prefix.
2795 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
2796 (pp_emit_prefix): Rename from pp_base_emit_prefix.
2797 (pp_append_text): Rename from pp_base_append_text.
2798 (pp_formatted_text): Rename from pp_base_formatted_text.
2799 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
2800 (pp_remaining_character_count_for_line): Rename from
2801 pp_base_remaining_character_count_for_line.
2802 (pp_newline): Rename from pp_base_newline.
2803 (pp_character): Rename from pp_base_character.
2804 (pp_string): Rename from pp_base_string.
2805 (pp_maybe_space): Rename from pp_base_maybe_space.
2806 * asan.c (asan_pp_string): Adjust.
2807 (asan_emit_stack_protection): Likewise.
2808 (asan_add_global): Likewise.
2809 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
2810 * tree-mudflap.c (mf_varname_tree): Likewise.
2811 * tree-pretty-print.c (pp_tree_identifier): Rename from
2812 pp_base_tree_identifier.
2813 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
2814 Declare as function.
2815
2816 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2817
2818 * pretty-print.h (pp_bar_bar): New.
2819 (pp_ampersand_ampersand): Likewise.
2820 (pp_less_equal): Likewise.
2821 (pp_greater_equal): Likewise.
2822 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
2823 printer functions instead of pp_string or operators and punctuators.
2824 (dump_gimple_call): Likewise.
2825 (dump_gimple_omp_for): Likewise.
2826 (dump_gimple_transaction): Likewise.
2827 (dump_gimple_phi): Likewise.
2828 (pp_gimple_stmt_1): Likewise.
2829 * sched-vis.c (print_insn): Likewise.
2830 * tree-mudflap.c (mf_varname_tree): Likewise.
2831 * tree-pretty-print.c (dump_block_node): Likewise.
2832 (dump_generic_node): Likewise.
2833
2834 2013-08-02 Jan Hubicka <jh@suse.cz>
2835
2836 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
2837 boundaries.
2838 * lto-streamer-out.c (tree_is_indexable): Results decls and
2839 parm decls are not indexable.
2840 (DFS_write_tree_body): Do not follow args and results.
2841 (hash_tree): Likewise.
2842 (output_functions): Rearrange so struct function is needed
2843 only when real body is output; be able to also ouptut abstract
2844 functions; output DECL_ARGUMENTS and DECL_RESULT.
2845 (lto_output): When not in WPA, ale store abstract functions.
2846 (write_symbol): Do not care about RESULT_DECL.
2847 (output_symbol_p): Handle correctly sbtract decls.
2848 * lto-streamer-in.c (input_function): Rearrange so struct
2849 function can be NULL at entry; allow streaming of
2850 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
2851 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
2852 sanity check during LTO.
2853 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
2854 RESULT_DECl and DECL_ARGUMENTS.
2855 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
2856 Likewise.
2857
2858 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
2859
2860 * pretty-print.h (pp_underscore): New.
2861 (pp_comma): Tidy.
2862 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
2863 printer functions instead of pp_character.
2864 (dump_binary_rhs): Likewise.
2865 (dump_ternary_rhs): Likewise.
2866 (dump_gimple_call_args): Likewise.
2867 (pp_points_to_solution): Likewise.
2868 (dump_gimple_call): Likewise.
2869 (dump_gimple_switch): Likewise.
2870 (dump_gimple_cond): Likewise.
2871 (dump_gimple_bind): Likewise.
2872 (dump_gimple_try): Likewise.
2873 (dump_gimple_omp_for): Likewise.
2874 (dump_gimple_omp_continue): Likewise.
2875 (dump_gimple_omp_single): Likewise.
2876 (dump_gimple_omp_sections): Likewise.
2877 (dump_gimple_omp_block): Likewise.
2878 (dump_gimple_omp_critical): Likewise.
2879 (dump_gimple_transaction): Likewise.
2880 (dump_gimple_asm): Likewise.
2881 (dump_gimple_phi): Likewise.
2882 (dump_gimple_omp_parallel): Likewise.
2883 (dump_gimple_omp_task): Likewise.
2884 (dump_gimple_omp_atomic_load): Likewise.
2885 (dump_gimple_omp_atomic_store): Likewise.
2886 (dump_gimple_mem_ops): Likewise.
2887 (pp_gimple_stmt_1): Likewise.
2888 (pp_cfg_jump): Likewise.
2889 (dump_implicit_edges): Likewise.
2890 (gimple_dump_bb_for_graph): Likewise.
2891 * graph.c (draw_cfg_node): Likewise.
2892 * langhooks.c (lhd_print_error_function): Likewise.
2893 * sched-vis.c (print_exp): Likewise.
2894 (print_value): Likewise.
2895 (print_pattern): Likewise.
2896 (print_insn): Likewise.
2897 (rtl_dump_bb_for_graph): Likewise.
2898 * tree-pretty-print.c (dump_function_declaration): Likewise.
2899 (dump_array_domain): Likewise.
2900 (dump_omp_clause): Likewise.
2901 (dump_location): Likewise.
2902 (dump_generic_node): Likewise.
2903 (print_struct_decl): Likewise.
2904 * diagnostic.c (diagnostic_show_locus): Use pp_space.
2905
2906 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
2907
2908 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
2909 candidate table when replacing a candidate statement.
2910 (replace_rhs_if_not_dup): Likewise.
2911 (replace_one_candidate): Likewise.
2912
2913 2013-08-02 Jan Hubicka <jh@suse.cz>
2914 Martin Liska <marxin.liska@gmail.com>
2915
2916 * cgraphunit.c (add_new_function): Fix logic when adding from
2917 late IPA pass.
2918 (assemble_thunk): Rename to ...
2919 (expand_thunk); .. this one; export; get it working with
2920 general functions; make produced gimple valid.
2921 * cgraph.h (expand_thunk): Declare.
2922
2923 2013-08-02 Jan Hubicka <jh@suse.cz>
2924
2925 * ipa-cp.c (gather_context_independent_values): Use
2926 ipa_get_param_move_cost.
2927 (get_replacement_map): Remove PARAM; move parameter folding
2928 into tree-inline.c
2929 (create_specialized_node): Update.
2930 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
2931 assert that we have gimple body; update move_cost.
2932 (count_formal_params): Assert that we have gimple body.
2933 (ipa_dump_param): New function.
2934 (ipa_alloc_node_params): Break out from ...
2935 (ipa_initialize_node_params): ... here.
2936 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
2937 (ipa_write_node_info): Stream move costs.
2938 (ipa_read_node_info): Read move costs.
2939 (ipa_update_after_lto_read): Do not recompute node params.
2940 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
2941 (ipa_get_param): Check we are not in WPA.
2942 (ipa_get_param_move_cost): New.
2943 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
2944 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
2945 parm numbers to be present.
2946
2947 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2948
2949 PR rtl-optimization/58048
2950 * lra-constraints.c (process_alt_operands): Don't check asm
2951 operand on register.
2952
2953 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
2954
2955 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
2956 the implied StoreLoad barrier for atomic operations if before.
2957
2958 2013-08-02 Jan Hubicka <jh@suse.cz>
2959 Martin Liska <marxin.liska@gmail.com>
2960
2961 * cgraph.c (cgraph_function_body_availability): Do not check
2962 cgraph flags.
2963 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
2964 symtab_node_availability): Declare.
2965 * ipa.c (can_replace_by_local_alias): New.
2966 (function_and_variable_visibility): Use it.
2967 * symtab.c (symtab_for_node_and_aliases,
2968 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
2969
2970 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
2971
2972 PR rtl-optimization/57963
2973 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
2974 (lra_constraints): Use them.
2975
2976 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
2977
2978 * config/arm/types.md (define_attr "type"): Add "load_acq"
2979 and "store_rel".
2980 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
2981 changes.
2982 (cortex_a53_store1): Likewise.
2983
2984 2013-08-01 Jan Hubicka <jh@suse.cz>
2985
2986 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
2987 partitions are not needed.
2988
2989 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
2990
2991 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
2992 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
2993 MAYBE_NON_Q_CLASS_P where appropriate.
2994
2995 2013-08-01 Jan Hubicka <jh@suse.cz>
2996
2997 * cgraph.h (release_function_body): Declare.
2998 * tree.c (free_lang_data_in_decl): Free, parameters and return values
2999 of unused delcarations.
3000
3001 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3002
3003 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3004 RTL form when subtracting a constant.
3005
3006 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3007
3008 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3009 Generate canonical plus rtx with negated immediate instead of minus
3010 where appropriate.
3011 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3012
3013 2013-08-01 Jan Hubicka <jh@suse.cz>
3014
3015 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3016 (cgraph_release_function_body): Likewise.
3017 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3018 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3019 to used_as_abstract_origin.
3020 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3021 symbols used as abstract origins.
3022 * cgraphunit.c (analyze_functions): Update.
3023 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3024 used_as_abstract_origin.
3025 * tree-inline.c (tree_function_versioning): Update
3026 used_as_abstract_origin; be ready for DECL_RESULT and
3027 DECL_ARGUMENTS to be NULL.
3028
3029 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3030 for abstract functions.
3031 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3032 real symbols.
3033
3034 2013-08-01 Jan Hubicka <jh@suse.cz>
3035
3036 * profile.c (compute_value_histograms): Fix thinko.
3037
3038 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3039
3040 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3041 aarch-common-protos.h to extra_headers.
3042 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3043 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3044 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3045
3046 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3047
3048 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3049 Include "../arm/types.md". Define "type" attribute for all patterns.
3050 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3051 attribute changes.
3052
3053 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3054
3055 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3056 to support power8 load fusion.
3057 (fusion_gpr_mem_load): Likewise.
3058
3059 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3060
3061 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3062 declarations for power8 load fusion.
3063 (emit_fusion_gpr_load): Likewise.
3064
3065 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3066 tuning for power8, turn on fusion mode by default. Turn on sign
3067 extending fusion mode if normal fusion mode is on, and we are at
3068 -O2 or -O3.
3069 (fusion_gpr_load_p): New function, return true if we can fuse an
3070 addis instruction with a dependent load to a GPR.
3071 (emit_fusion_gpr_load): Emit the instructions for power8 load
3072 fusion to GPRs.
3073
3074 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3075 (VSX load fusion peepholes): New peepholes to fuse together an
3076 addi instruction with a VSX load instruction.
3077
3078 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3079 peepholes to fuse an addis instruction with a load to a GPR base
3080 register. If we are supporting sign extending fusions, convert
3081 sign extending loads to zero extending loads and add an explicit
3082 sign extension.
3083
3084 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3085
3086 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3087 aarch-common-protos.h to extra_headers.
3088 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3089 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3090 (arm_early_store_addr_dep): Likewise.
3091 (arm_no_early_alu_shift_dep): Likewise.
3092 (arm_no_early_alu_shift_value_dep): Likewise.
3093 (arm_no_early_mul_dep): Likewise.
3094 (arm_no_early_store_addr_dep): Likewise.
3095 (arm_mac_accumulator_is_mul_result): Likewise.
3096 (arm_mac_accumulator_is_result): Likewise.
3097 * config/arm/aarch-common.c: ... here. New file.
3098 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
3099 here to ...
3100 (arm_early_store_addr_dep): Likewise.
3101 (arm_no_early_alu_shift_dep): Likewise.
3102 (arm_no_early_alu_shift_value_dep): Likewise.
3103 (arm_no_early_mul_dep): Likewise.
3104 (arm_no_early_store_addr_dep): Likewise.
3105 (arm_mac_accumulator_is_mul_result): Likewise.
3106 (arm_mac_accumulator_is_result): Likewise.
3107 * config/arm/aarch-common-protos.h: ... here. New file.
3108 * config/arm/t-arm (aarch-common.o): Define.
3109
3110 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3111
3112 * config/arm/arm.md: Include new file "types.md".
3113 (define_attr "type"): Move from here to ...
3114 (define_attr "mul32"): Likewise.
3115 (define_attr "mul64"): Likewise.
3116 * config/arm/types.md: ... here. New file.
3117
3118 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
3119
3120 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
3121 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
3122
3123 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3124
3125 * gen-pass-instances.awk: Fix offset of substr().
3126
3127 2013-07-31 David Malcolm <dmalcolm@redhat.com>
3128
3129 * Makefile.in (pass-instances.def): New.
3130 (passes.o): Replace dependency on passes.def with one on
3131 pass-instances.def
3132
3133 * gen-pass-instances.awk: New.
3134
3135 * passes.c (pass_manager::pass_manager): Use pass-instances.def
3136 rather than passes.def, updating local definition of NEXT_PASS
3137 macro to add an extra NUM parameter (currently unused).
3138
3139 2013-07-30 David Malcolm <dmalcolm@redhat.com>
3140
3141 * Makefile.in (PASS_MANAGER_H): New.
3142 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
3143 (passes.o): Likewise.
3144 (statistics.o): Likewise.
3145 (cgraphunit.o): Likewise.
3146 (context.o): Depend on PASS_MANAGER_H.
3147
3148 * pass_manager.h: New.
3149
3150 * cgraphunit.c (cgraph_add_new_function): Update for moves
3151 of globals to fields of pass_manager.
3152 (analyze_function): Likewise.
3153 (expand_function): Likewise.
3154 (ipa_passes): Likewise.
3155 (compile): Likewise.
3156
3157 * context.c (context::context): New.
3158 * context.h (context::context): New.
3159 (context::get_passes): New.
3160 (context::passes_): New.
3161
3162 * lto-cgraph.c (input_node): Update for moves of globals to
3163 fields of pass_manager.
3164
3165 * passes.c (all_passes): Remove, in favor of a field of the
3166 same name within the new class pass_manager.
3167 (all_small_ipa_passes): Likewise.
3168 (all_lowering_passes): Likewise.
3169 (all_regular_ipa_passes): Likewise.
3170 (all_late_ipa_passes): Likewise.
3171 (all_lto_gen_passes): Likewise.
3172 (passes_by_id): Likewise.
3173 (passes_by_id_size): Likewise.
3174 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3175 the new class pass_manager.
3176 (set_pass_for_id): Convert to...
3177 (pass_manager::set_pass_for_id): ...method.
3178 (get_pass_for_id): Convert to...
3179 (pass_manager::get_pass_for_id): ...method.
3180 (register_one_dump_file): Move body of implementation into...
3181 (pass_manager::register_one_dump_file): ...here.
3182 (register_dump_files_1): Convert to...
3183 (pass_manager::register_dump_files_1): ...method.
3184 (register_dump_files): Convert to...
3185 (pass_manager::register_dump_files): ...method.
3186 (create_pass_tab): Update for moves of globals to fields of
3187 pass_manager.
3188 (dump_passes): Move body of implementation into...
3189 (pass_manager::dump_passes): ...here.
3190 (register_pass): Move body of implementation into...
3191 (pass_manager::register_pass): ...here.
3192 (init_optimization_passes): Convert into...
3193 (pass_manager::pass_manager): ...constructor for new
3194 pass_manager class, and initialize the pass_lists array.
3195 (check_profile_consistency): Update for moves of globals to
3196 fields of pass_manager.
3197 (dump_profile_report): Move body of implementation into...
3198 (pass_manager::dump_profile_report): ...here.
3199 (ipa_write_summaries_1): Update for moves of pass lists from
3200 being globals to fields of pass_manager.
3201 (ipa_write_optimization_summaries): Likewise.
3202 (ipa_read_summaries): Likewise.
3203 (ipa_read_optimization_summaries): Likewise.
3204 (execute_all_ipa_stmt_fixups): Likewise.
3205
3206 * statistics.c (statistics_fini): Update for moves of globals to
3207 fields of pass_manager.
3208
3209 * toplev.c (general_init): Replace call to
3210 init_optimization_passes with construction of the pass_manager
3211 instance.
3212
3213 * tree-pass.h (all_passes): Remove, in favor of a field of the
3214 same name within the new class pass_manager.
3215 (all_small_ipa_passes): Likewise.
3216 (all_lowering_passes): Likewise.
3217 (all_regular_ipa_passes): Likewise.
3218 (all_lto_gen_passes): Likewise.
3219 (all_late_ipa_passes): Likewise.
3220 (passes_by_id): Likewise.
3221 (passes_by_id_size): Likewise.
3222 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
3223 the new class pass_manager.
3224 (get_pass_for_id): Remove.
3225
3226 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3227
3228 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
3229 configs.
3230
3231 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
3232
3233 * arm.md (mulhi3): New expand pattern.
3234
3235 2013-07-30 Jan Hubicka <jh@suse.cz>
3236 Martin Liska <marxin.liska@gmail.com>
3237
3238 * profile.c (compute_value_histograms): Do not ICE when
3239 there is mismatch only on some counters.
3240
3241 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3242
3243 PR rtl-optimization/57637
3244 * function.c (move_insn_for_shrink_wrap): Also check the
3245 GEN set of the LIVE problem for the liveness analysis
3246 if it exists, otherwise give up.
3247
3248 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3249
3250 PR tree-optimization/57993
3251 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
3252 replaced statement in the candidate table.
3253 (phi_add_costs): Return infinite cost when the hidden basis does
3254 not dominate all phis on which the candidate is dependent.
3255 (replace_one_candidate): Record replaced statement in the
3256 candidate table.
3257
3258 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
3259
3260 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
3261 (ashlv2si3): New expander.
3262 (*ashlv2si3_i): New define_insn_and_split.
3263 * predicates.md (float_operation): Allow patterns with three
3264 basic sub-patterns.
3265
3266 PR rtl-optimization/58021
3267 * mode-switching.c (create_pre_exit): Always split off preceding
3268 insns if we are not at the basic block head.
3269
3270 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
3271
3272 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
3273 (UCLIBC_DYNAMIC_LINKER): New macro.
3274 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
3275 `-mnan=2008'.
3276 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
3277 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
3278 `-mnan=2008'.
3279 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
3280 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
3281 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
3282 for SF and DF modes. Use ieee_quad_format for TF mode.
3283 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
3284 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
3285 (mips_option_override): Handle `-mnan=legacy'.
3286 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
3287 `-mabs=2008' and `-mnan=2008'.
3288 (OPTION_DEFAULT_SPECS): Add "nan" default.
3289 (ASM_SPEC): Handle `-mnan='.
3290 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
3291 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
3292 comment accordingly.
3293 (neg<mode>2): Likewise.
3294 * config/mips/mips.opt (mabs, mnan): New options.
3295 * doc/install.texi (Configuration): Document `--with-nan=' option.
3296 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
3297 `-mnan=' options.
3298 (MIPS Options): Document them.
3299 * config.gcc <mips*-*-*>: Handle `--with-nan='.
3300 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
3301 * configure: Regenerate.
3302 * config.in: Regenerate.
3303
3304 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3305
3306 * config/i386/i386.md (float post-reload splitters): Do not check
3307 for subregs of SSE registers.
3308
3309 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
3310 H.J. Lu <hongjiu.lu@intel.com>
3311
3312 PR target/57954
3313 PR target/57988
3314 * config/i386/i386.md (post-reload splitter
3315 to avoid partial SSE reg dependency stalls): New pattern.
3316
3317 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
3318
3319 * config/s390/s390.md ("movcc"): Swap load and store instructions.
3320
3321 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
3322
3323 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
3324 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
3325
3326 2013-07-26 Cary Coutant <ccoutant@google.com>
3327
3328 * dwarf2out.c (die_checksum_ordered): Don't include template
3329 instantiations in signature.
3330 (is_template_parameter): New function.
3331 (is_template_instantiation): New function.
3332 (generate_skeleton_bottom_up): Don't include template instantiations
3333 in type unit DIE.
3334 (generate_skeleton): Likewise.
3335 (break_out_comdat_types): Move recursive call to break out nested
3336 types earlier.
3337 (prune_unused_types_mark_generic_parms_dies): Call
3338 is_template_parameter.
3339
3340 2013-07-26 Ian Bolton <ian.bolton@arm.com>
3341
3342 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
3343 uses vector registers.
3344 * config/aarch64/iterators.md: Add attributes rtn and vas.
3345
3346 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3347 Richard Earnshaw <richard.earnshaw@arm.com>
3348
3349 * combine.c (simplify_comparison): Re-canonicalize operands
3350 where appropriate.
3351 * config/arm/arm.md (movcond_addsi): New splitter.
3352
3353 2013-07-25 Sterling Augustine <saugustine@google.com>
3354
3355 * dwarf2out.c (size_of_pubnames): Move code to...
3356 (include_pubname_in_output): ...here. New.
3357 (want_pubnames): Rearrange.
3358 (output_pubnames): Call include_pubname_in_output. Move assertion.
3359
3360 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3361
3362 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
3363
3364 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
3365
3366 PR target/38836
3367 * doc/extend.texi: Remove obsolete builtins. Fix
3368 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
3369
3370 2013-07-25 Jan Hubicka <jh@suse.cz>
3371
3372 * cgraph.c (release_function_body): Break out from ...
3373 (cgraph_release_function_body): ... this one; also release DECL_RESULT
3374 and DECL_ARGUMENTS.
3375 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
3376 old_tree in the map.
3377 (create_specialized_node): Update.
3378 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
3379 into index.
3380 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
3381 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
3382 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
3383 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
3384 DECL_RESULT.
3385
3386 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3387
3388 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
3389 addsi3_carryin_alt2_<optab>): Correct output template.
3390
3391 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3392
3393 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
3394 Adjust for arm_restrict_it.
3395 Remove trailing whitespace.
3396
3397 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
3398
3399 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
3400 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
3401
3402 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
3403
3404 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
3405
3406 PR rtl-optimization/57960
3407 * lra-constraints.c (process_alt_operands): Use the right mode
3408 when checking strict_low.
3409
3410 2013-07-25 Jan Hubicka <jh@suse.cz>
3411
3412 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
3413 * cgraph.c (cgraph_remove_node): Do not release function body
3414 when in cgraph streaming.
3415 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
3416 in other partitions are not considered reachable; fix handling of
3417 clones.
3418
3419 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3420
3421 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
3422
3423 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3424
3425 PR target/19599
3426 PR target/57731
3427 PR target/57837
3428 * config/arm/arm.md ("*sibcall_insn): Replace use of
3429 Ss with US. Adjust output for v5 and v4t.
3430 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
3431
3432 * config/arm/constraints.md ("Ss"): Rename to US.
3433
3434 2013-07-25 Terry Guo <terry.guo@arm.com>
3435
3436 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
3437 shift_add/shift_sub0/shift_sub1 RTXs.
3438
3439 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
3440 Anton Blanchard <anton@au1.ibm.com>
3441
3442 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
3443 (altivec_vpks<VI_char>ss): Likewise.
3444 (altivec_vpks<VI_char>us): Likewise.
3445 (altivec_vpku<VI_char>us): Likewise.
3446 (altivec_vpku<VI_char>um): Likewise.
3447
3448 2013-07-24 David Malcolm <dmalcolm@redhat.com>
3449
3450 Introduce context class.
3451
3452 * Makefile.in (CONTEXT_H): New.
3453 (OBJS): Add context.o.
3454 (toplev.o): Add CONTEXT_H to dependencies.
3455 (context.o): New.
3456
3457 * toplev.c (general_init): Create the singleton gcc::context instance.
3458
3459 * context.c: New.
3460
3461 * context.h: New.
3462
3463 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
3464
3465 PR rtl-optimization/57968
3466 * mode-switching.c (create_pre_exit): Allow instructions that
3467 don't set a return register to need a non-exit mode.
3468
3469 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3470 Anton Blanchard <anton@au1.ibm.com>
3471
3472 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
3473 operands to vperm for little endian.
3474 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
3475 of lvsl to create the control mask for a vperm for little endian.
3476
3477 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3478 Anton Blanchard <anton@au1.ibm.com>
3479
3480 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3481 two operands for little-endian.
3482
3483 2013-07-23 Steve Ellcey <sellcey@mips.com>
3484
3485 * config/mips/mips.c (mips_case_values_threshold): New.
3486 (TARGET_CASE_VALUES_THRESHOLD): Define.
3487
3488 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3489 Anton Blanchard <anton@au1.ibm.com>
3490
3491 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
3492 selection of field for vector splat in little endian mode.
3493
3494 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3495
3496 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
3497 expanders to rs6000.md.
3498 (ior<mode>3): Likewise.
3499 (and<mode>3): Likewise.
3500 (one_cmpl<mode>2): Likewise.
3501 (nor<mode>3): Likewise.
3502 (andc<mode>3): Likewise.
3503 (eqv<mode>3): Likewise.
3504 (nand<mode>3): Likewise.
3505 (orc<mode>3): Likewise.
3506
3507 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
3508 declaration.
3509
3510 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
3511 to split multi-word logical operations.
3512 (rs6000_split_logical_di): Likewise.
3513 (rs6000_split_logical): Likewise.
3514
3515 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
3516 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
3517 and allow TImode operations in 32-bit.
3518 (vsx_and<mode>3_64bit): Likewise.
3519 (vsx_ior<mode>3_32bit): Likewise.
3520 (vsx_ior<mode>3_64bit): Likewise.
3521 (vsx_xor<mode>3_32bit): Likewise.
3522 (vsx_xor<mode>3_64bit): Likewise.
3523 (vsx_one_cmpl<mode>2_32bit): Likewise.
3524 (vsx_one_cmpl<mode>2_64bit): Likewise.
3525 (vsx_nor<mode>3_32bit): Likewise.
3526 (vsx_nor<mode>3_64bit): Likewise.
3527 (vsx_andc<mode>3_32bit): Likewise.
3528 (vsx_andc<mode>3_64bit): Likewise.
3529 (vsx_eqv<mode>3_32bit): Likewise.
3530 (vsx_eqv<mode>3_64bit): Likewise.
3531 (vsx_nand<mode>3_32bit): Likewise.
3532 (vsx_nand<mode>3_64bit): Likewise.
3533 (vsx_orc<mode>3_32bit): Likewise.
3534 (vsx_orc<mode>3_64bit): Likewise.
3535
3536 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
3537 logical types in GPRs.
3538
3539 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
3540 logical insns to rs6000.md, and allow TImode operations in
3541 32-bit.
3542 (altivec_ior<mode>3): Likewise.
3543 (altivec_xor<mode>3): Likewise.
3544 (altivec_one_cmpl<mode>2): Likewise.
3545 (altivec_nor<mode>3): Likewise.
3546 (altivec_andc<mode>3): Likewise.
3547
3548 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
3549 attributes for moving the 128-bit logical operations into
3550 rs6000.md.
3551 (BOOL_REGS_OUTPUT): Likewise.
3552 (BOOL_REGS_OP1): Likewise.
3553 (BOOL_REGS_OP2): Likewise.
3554 (BOOL_REGS_UNARY): Likewise.
3555 (BOOL_REGS_AND_CR0): Likewise.
3556 (one_cmpl<mode>2): Add support for DI logical operations on
3557 32-bit, splitting the operations to 32-bit.
3558 (anddi3): Likewise.
3559 (iordi3): Likewise.
3560 (xordi3): Likewise.
3561 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
3562 changes to combine the 32/64-bit code, allow logical operations on
3563 TI mode in 32-bit, and to use similar match_operator patterns like
3564 scalar mode uses. Combine the Altivec and VSX code for logical
3565 operations, and move it here.
3566 (ior<mode>3, 128-bit types): Likewise.
3567 (xor<mode>3, 128-bit types): Likewise.
3568 (one_cmpl<mode>3, 128-bit types): Likewise.
3569 (nor<mode>3, 128-bit types): Likewise.
3570 (andc<mode>3, 128-bit types): Likewise.
3571 (eqv<mode>3, 128-bit types): Likewise.
3572 (nand<mode>3, 128-bit types): Likewise.
3573 (orc<mode>3, 128-bit types): Likewise.
3574 (and<mode>3_internal): Likewise.
3575 (bool<mode>3_internal): Likewise.
3576 (boolc<mode>3_internal1): Likewise.
3577 (boolc<mode>3_internal2): Likewise.
3578 (boolcc<mode>3_internal1): Likewise.
3579 (boolcc<mode>3_internal2): Likewise.
3580 (eqv<mode>3_internal1): Likewise.
3581 (eqv<mode>3_internal2): Likewise.
3582 (one_cmpl1<mode>3_internal): Likewise.
3583
3584 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3585
3586 * config/microblaze/microblaze.c (microblaze_expand_prologue):
3587 Rename flag_stack_usage to flag_stack_usage_info.
3588
3589 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
3590
3591 * config/microblaze/sync.md: New file.
3592 * config/microblaze/microblaze.md: Include sync.md
3593 * config/microblaze/microblaze.c: Add print_operand 'y'.
3594 * config/microblaze/constraints.md: Add memory_contraint
3595 'Q' which is a single register.
3596
3597 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
3598
3599 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
3600
3601 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3602
3603 * reload.c (find_reloads): Exit loop once we find this operand
3604 cannot be reloaded somehow for this alternative.
3605
3606 * reload.c (find_reloads): Exit loop once we find a hard register.
3607
3608 * rtlanal.c (computed_jump_p): Exit loop once we find label
3609 reference is used.
3610
3611 * i386.c (ix86_pad_returns): Exit loop after setting replace.
3612
3613 * cfgloopmanip.c (remove_path): Exit loop after setting
3614 irred_invalidated.
3615
3616 * gensupport.c (subst_dup): Avoid loop if code is not
3617 MATCH_DUP nor MATCH_OP_DUP.
3618
3619 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
3620
3621 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
3622
3623 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3624
3625 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
3626 true for SP_REGNUM if mode == ptr_mode.
3627 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
3628 with value R0_REGNUM + 31.
3629
3630 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3631
3632 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
3633 pad pointer-typed argument downward.
3634
3635 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3636
3637 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
3638 and __ILP32__ when the ILP32 model is in use.
3639
3640 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3641
3642 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
3643 (aarch64_load_symref_appropriately): In the case of
3644 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
3645 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
3646 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
3647 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
3648 if 'mode' doesn't equal to 'ptr_mode'.
3649 (aarch64_output_mi_thunk): Add an assertion on the alignment of
3650 'vcall_offset'; change to call aarch64_emit_move differently depending
3651 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
3652 to calculate the upper bound of 'vcall_offset'.
3653 (aarch64_cannot_force_const_mem): Change to also return true if
3654 mode != ptr_mode.
3655 (aarch64_legitimize_reload_address): In the case of large
3656 displacements, add new local variable 'xmode' and an assertion
3657 based on it; change to use 'xmode' to generate the new rtx and
3658 reload.
3659 (aarch64_asm_trampoline_template): Change to generate the template
3660 differently depending on TARGET_ILP32 or not; change to use
3661 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
3662 (aarch64_trampoline_size): Removed.
3663 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
3664 and replace immediate literals with it. Change to use 'ptr_mode'
3665 instead of 'DImode' and call convert_memory_address if the mode
3666 of 'fnaddr' doesn't equal to 'ptr_mode'.
3667 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
3668 to output symbol.
3669 (aarch64_elf_asm_destructor): Likewise.
3670 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
3671 on TARGET_ILP32 instead of aarch64_trampoline_size.
3672 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
3673 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
3674 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
3675 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
3676 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
3677 (storewb_pair<GPI:mode>_<P:mode>): ... this.
3678 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
3679 depending on the value of 'mode'.
3680 (add_losym_<mode>): New.
3681 (ldr_got_small_<mode>): New, based on ldr_got_small.
3682 (ldr_got_small): Remove.
3683 (ldr_got_small_sidi): New.
3684 * config/aarch64/iterators.md (P): New.
3685 (PTR): Change to 'ptr_mode' in the condition.
3686
3687 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
3688
3689 * config.gcc (aarch64*-*-*): Support --with-abi.
3690 (aarch64*-*-elf): Support --with-multilib-list.
3691 (aarch64*-*-linux*): Likewise.
3692 (supported_defaults): Add abi to aarch64*-*-*.
3693 * configure.ac: Mention AArch64 for --with-multilib-list.
3694 * configure: Re-generated.
3695 * config/aarch64/biarchilp32.h: New file.
3696 * config/aarch64/biarchlp64.h: New file.
3697 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
3698 (ABI_SPEC): Ditto.
3699 (MULTILIB_DEFAULTS): Ditto.
3700 (DRIVER_SELF_SPECS): Ditto.
3701 (ASM_SPEC): Update to also substitute -mabi.
3702 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
3703 file whose name depends on -mabi= and -mbig-endian.
3704 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
3705 TARGET_ILP32.
3706 (POINTER_SIZE): New define.
3707 (POINTERS_EXTEND_UNSIGNED): Ditto.
3708 (enum aarch64_abi_type): New enumeration tag.
3709 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
3710 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
3711 (TARGET_ILP32): New define.
3712 * config/aarch64/aarch64.opt (mabi): New.
3713 (aarch64_abi): New.
3714 (ilp32, lp64): New values for -mabi.
3715 * config/aarch64/t-aarch64 (comma): New define.
3716 (MULTILIB_OPTIONS): Ditto.
3717 (MULTILIB_DIRNAMES): Ditto.
3718 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
3719 * doc/invoke.texi: Document -mabi for AArch64.
3720
3721 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
3722
3723 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
3724
3725 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3726 Anton Blanchard <anton@au1.ibm.com>
3727
3728 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
3729 endianness when selecting field to splat.
3730
3731 2013-07-22 Eric Christopher <echristo@gmail.com>
3732
3733 * dwarf2out.c (die_odr_checksum): New function to use
3734 CHECKSUM_ macros and ULEB128 for DIE tag.
3735 (generate_type_signature): Use.
3736
3737 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
3738
3739 * config.gcc (sparc*-*-*): Accept leon3 processor.
3740 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
3741 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
3742 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
3743 * config/sparc/sparc.opt (enum processor_type): Add leon3.
3744 (mfix-ut699): Adjust comment.
3745 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
3746 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
3747 (CPP_CPU_SPEC): Likewise.
3748 (ASM_CPU_SPEC): Likewise.
3749 * config/sparc/sparc.c (leon3_cost): New constant.
3750 (sparc_option_override): Add leon3 support.
3751 (mem_ref): New function.
3752 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
3753 (sparc_do_work_around_errata): Look into the instruction in the delay
3754 slot and adjust accordingly. Add fix for the data cache nullify issues
3755 of the UT699. Change insertion position for the NOP.
3756 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
3757 (leon3_load): New reservation.
3758 (leon_store): Bump latency to 2.
3759 (grfpu): New automaton.
3760 (grfpu_alu): New unit.
3761 (grfpu_ds): Likewise.
3762 (leon_fp_alu): Adjust.
3763 (leon_fp_mult): Delete.
3764 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
3765 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
3766 * config/sparc/sparc.md (cpu): Add leon3.
3767 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
3768 (swapsi): Likewise.
3769 (atomic_test_and_set): Likewise.
3770 (ldstub): Likewise.
3771
3772 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
3773
3774 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
3775 default for R5900 targets.
3776 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
3777 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
3778 * config/mips/mips.c (mips_option_override): Report an error for
3779 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
3780 for -march=r5900 -mhard-float.
3781
3782 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
3783
3784 * df-problems.c (can_move_insns_across): Exit loop once we
3785 find a non-fixed, non-global register.
3786
3787 * ipa-pure-const.c (propagate_nothrow): Exit loop after
3788 setting can_throw.
3789
3790 * omega.c (omega_eliminate_red): Break after setting red_found.
3791 (omega_problem_has_red_equations): Similarly after setting found.
3792 (omega_query_variable): Similarly after setting coupled.
3793
3794 2013-07-22 Marek Polacek <polacek@redhat.com>
3795
3796 * gimplify.c: Don't include gimple.h twice.
3797
3798 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3799
3800 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
3801 instead of TARGET_THUMB1.
3802 (Pz): New constraint.
3803 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
3804 encodings.
3805 (compare_negsi_si): Likewise.
3806 (compare_addsi2_op0): Likewise.
3807 (compare_addsi2_op1): Likewise.
3808 (addsi3_carryin_<optab>): Likewise.
3809 (addsi3_carryin_alt2_<optab>): Likewise.
3810 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
3811 for arm_restrict_it.
3812 (subsi3_carryin): Likewise.
3813 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
3814 (minmax_arithsi): Disable for arm_restrict_it.
3815 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
3816 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
3817 (satsi_<SAT:code>_shift): Likewise.
3818 (arm_shiftsi3): Add alternative for 16-bit encoding.
3819 (arm32_movhf): Disable for arm_restrict_it.
3820 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
3821 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
3822
3823 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
3824
3825 * config/arm/arm.md (attribute "insn"): Delete.
3826 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
3827 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
3828 (not_shiftsi): Update for attribute change.
3829 (not_shiftsi_compare0): Likewise.
3830 (not_shiftsi_compare0_scratch): Likewise.
3831 (arm_one_cmplsi2): Likewise.
3832 (thumb1_one_cmplsi2): Likewise.
3833 (notsi_compare0): Likewise.
3834 (notsi_compare0_scratch): Likewise.
3835 (thumb1_movdi_insn): Likewise.
3836 (arm_movsi_insn): Likewise.
3837 (movhi_insn_arch4): Likewise.
3838 (movhi_bytes): Likewise.
3839 (arm_movqi_insn): Likewise.
3840 (thumb1_movqi_insn): Likewise.
3841 (arm32_movhf): Likewise.
3842 (thumb1_movhf): Likewise.
3843 (arm_movsf_soft_insn): Likewise.
3844 (thumb1_movsf_insn): Likewise.
3845 (thumb_movdf_insn): Likewise.
3846 (movsicc_insn): Likewise.
3847 (movsfcc_soft_insn): Likewise.
3848 (and_scc): Likewise.
3849 (cond_move): Likewise.
3850 (if_move_not): Likewise.
3851 (if_not_move): Likewise.
3852 (if_shift_move): Likewise.
3853 (if_move_shift): Likewise.
3854 (if_shift_shift): Likewise.
3855 (if_not_arith): Likewise.
3856 (if_arith_not): Likewise.
3857 (cond_move_not): Likewise.
3858 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
3859 (neon_mov<mode>): Likewise.
3860 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
3861 (thumb2_movsi_vfp): Likewise.
3862 (movsf_vfp): Likewise.
3863 (thumb2_movsf_vfp): Likewise.
3864 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
3865 change.
3866 (cortexa7_older_only): Likewise.
3867 (cortexa7_younger): Likewise.
3868 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
3869 (1020alu_shift_op): Likewise.
3870 (1020alu_shift_reg_op): Likewise.
3871 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
3872 (alu_shift_op): Likewise.
3873 (alu_shift_reg_op): Likewise.
3874 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
3875 (11_alu_shift_op): Likewise.
3876 (11_alu_shift_reg_op): Likewise.
3877 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
3878 (9_alu_shift_reg_op): Likewise.
3879 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3880 change.
3881 (cortex_a15_alu_shift): Likewise.
3882 (cortex_a15_alu_shift_reg): Likewise.
3883 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
3884 (cortex_a5_alu_shift): Likewise.
3885 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
3886 change.
3887 (cortex_a53_alu_shift): Likewise.
3888 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
3889 change.
3890 (cortex_a7_alu_reg): Likewise.
3891 (cortex_a7_alu_shift): Likewise.
3892 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3893 (cortex_a8_alu_shift): Likewise.
3894 (cortex_a8_alu_shift_reg): Likewise.
3895 (cortex_a8_mov): Likewise.
3896 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3897 (cortex_a9_dp_shift): Likewise.
3898 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
3899 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
3900 (cortex_r4_mov): Likewise.
3901 (cortex_r4_alu_shift): Likewise.
3902 (cortex_r4_alu_shift_reg): Likewise.
3903 * config/arm/fa526.md (526_alu_op): Update for attribute change.
3904 (526_alu_shift_op): Likewise.
3905 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
3906 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
3907 (626te_alu_shift_op): Likewise.
3908 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
3909 (726te_alu_op): Likewise.
3910 (726te_alu_shift_op): Likewise.
3911 (726te_alu_shift_reg_op): Likewise.
3912 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
3913 (mp626_alu_shift_op): Likewise.
3914 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
3915 (pj4_alu_e1_conds): Likewise.
3916 (pj4_alu): Likewise.
3917 (pj4_alu_conds): Likewise.
3918 (pj4_shift): Likewise.
3919 (pj4_shift_conds): Likewise.
3920 (pj4_alu_shift): Likewise.
3921 (pj4_alu_shift_conds): Likewise.
3922
3923 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3924
3925 * config/arm/predicates.md (shiftable_operator_strict_it):
3926 New predicate.
3927 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
3928 Disable cond_exec version for arm_restrict_it.
3929 (thumb2_smaxsi3): Convert to generate cond_exec.
3930 (thumb2_sminsi3): Likewise.
3931 (thumb32_umaxsi3): Likewise.
3932 (thumb2_uminsi3): Likewise.
3933 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
3934 (thumb2_neg_abssi2): Likewise.
3935 (thumb2_mov_scc): Add alternative for 16-bit encoding.
3936 (thumb2_movsicc_insn): Adjust alternatives.
3937 (thumb2_mov_negscc): Disable for arm_restrict_it.
3938 (thumb2_mov_negscc_strict_it): New pattern.
3939 (thumb2_mov_notscc_strict_it): New pattern.
3940 (thumb2_mov_notscc): Disable for arm_restrict_it.
3941 (thumb2_ior_scc): Likewise.
3942 (thumb2_ior_scc_strict_it): New pattern.
3943 (thumb2_cond_move): Adjust for arm_restrict_it.
3944 (thumb2_cond_arith): Disable for arm_restrict_it.
3945 (thumb2_cond_arith_strict_it): New pattern.
3946 (thumb2_cond_sub): Adjust for arm_restrict_it.
3947 (thumb2_movcond): Likewise.
3948 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
3949 (thumb2_zero_extendhisi2_v6): Likewise.
3950 (thumb2_zero_extendqisi2_v6): Likewise.
3951 (orsi_notsi_si): Likewise.
3952 (orsi_not_shiftsi_si): Likewise.
3953
3954 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
3955
3956 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
3957 instruction sequence is 1 byte shorter.
3958
3959 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
3960
3961 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
3962 it is not needed after split.
3963
3964 2013-07-20 Iain Sandoe <iain@codesourcery.com>
3965
3966 PR target/51784
3967 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
3968 second label for nonlocal goto receivers. Don't output pic base labels
3969 unless we're producing PIC; mark that action unreachable().
3970 (ix86_save_reg): If the function contains a nonlocal label, save the
3971 PIC base reg.
3972 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
3973 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
3974 (update_pic_label_number_if_needed): New.
3975 (machopic_output_function_base_name): Adjust for nonlocal receiver
3976 case.
3977 (machopic_should_output_picbase_label): New.
3978 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
3979 (nonlocal_goto_receiver): New insn and split.
3980
3981 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
3982
3983 * config/aarch64/aarch64-builtins.c
3984 (aarch64_fold_builtin): Fold abs in all modes.
3985 * config/aarch64/aarch64-simd-builtins.def
3986 (abs): Enable for all modes.
3987 * config/aarch64/arm_neon.h
3988 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
3989 (vabs_f64): Add missing intrinsic.
3990
3991 2013-07-19 Ian Bolton <ian.bolton@arm.com>
3992
3993 * config/aarch64/arm_neon.h (vabs_s64): New function
3994
3995 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
3996
3997 PR target/57516
3998 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
3999 * config/avr/avr.md (adjust_len): Add `round'.
4000 * config/avr/avr-protos.h (avr_out_round): New prototype.
4001 (avr_out_plus): Add `out_label' argument.
4002 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4003 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4004 Handle the case where `insn' is just a pattern.
4005 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4006 (avr_out_round): New function.
4007 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4008
4009 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4010
4011 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4012 Add check for flag_stack_usage to handle -fstack-usage support
4013
4014 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4015
4016 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4017 interaction for new Power8 flags and VSX.
4018
4019 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4020
4021 PR middle-end/57698
4022 * tree-inline.c (expand_call_inline): Emit errors during
4023 early_inlining only if optimization is not turned on.
4024
4025 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4026
4027 * passes.def: New.
4028
4029 * passes.c (init_optimization_passes): Move the construction of
4030 the pass hierarchy into a new passes.def file.
4031
4032 * Makefile.in (passes.o): Add dependency on passes.def.
4033
4034 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4035
4036 * passes.c (init_optimization_passes): Introduce macros for
4037 constructing the tree of passes (INSERT_PASSES_AFTER,
4038 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4039 TERMINATE_PASS_LIST).
4040
4041 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4042 Wei Mi <wmi@google.com>
4043
4044 PR rtl-optimization/57878
4045 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4046 top.
4047 (reload_pseudo_compare_func): Check nregs first for reload
4048 pseudos.
4049
4050 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4051
4052 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4053
4054 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4055
4056 * read-rtl.c (validate_const_int): Once an invalid character is
4057 seen, quit the loop.
4058
4059 * gengtype.c (write_roots): Similarly once we find the "deletable"
4060 or "if_marked" option.
4061
4062 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4063
4064 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4065 "xtab" and "sat". Move value "clz" from here to ...
4066 (attriubte "type"): ... here.
4067 (satsi_<SAT:code>): Delete "insn" attribute.
4068 (satsi_<SAT:code>_shift): Likewise.
4069 (arm_zero_extendqisi2addsi): Likewise.
4070 (arm_extendqisi2addsi): Likewise.
4071 (clzsi2): Update for attribute changes.
4072 (rbitsi2): Likewise.
4073 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4074 attribute.
4075 (arm_usatsihi): Likewise.
4076 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4077
4078 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4079
4080 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4081 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4082 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4083 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4084 in alphabetical order.
4085 (attribute "core_cycles"): Update for attribute changes.
4086 (arm_addsi3): Likewise.
4087 (addsi3_compare0): Likewise.
4088 (addsi3_compare0_scratch): Likewise.
4089 (addsi3_compare_op1): Likewise.
4090 (addsi3_compare_op2): Likewise.
4091 (compare_addsi2_op0): Likewise.
4092 (compare_addsi2_op1): Likewise.
4093 (addsi3_carryin_shift_<optab>): Likewise.
4094 (subsi3_carryin_shift): Likewise.
4095 (rsbsi3_carryin_shift): Likewise.
4096 (arm_subsi3_insn): Likewise.
4097 (subsi3_compare0): Likewise.
4098 (subsi3_compare): Likewise.
4099 (arm_andsi3_insn): Likewise.
4100 (thumb1_andsi3_insn): Likewise.
4101 (andsi3_compare0): Likewise.
4102 (andsi3_compare0_scratch): Likewise.
4103 (zeroextractsi_compare0_scratch
4104 (andsi_not_shiftsi_si): Likewise.
4105 (iorsi3_insn): Likewise.
4106 (iorsi3_compare0): Likewise.
4107 (iorsi3_compare0_scratch): Likewise.
4108 (arm_xorsi3): Likewise.
4109 (thumb1_xorsi3_insn): Likewise.
4110 (xorsi3_compare0): Likewise.
4111 (xorsi3_compare0_scratch): Likewise.
4112 (satsi_<SAT:code>_shift): Likewise.
4113 (rrx): Likewise.
4114 (arm_shiftsi3): Likewise.
4115 (shiftsi3_compare0): Likewise.
4116 (not_shiftsi): Likewise.
4117 (not_shiftsi_compare0): Likewise.
4118 (not_shiftsi_compare0_scratch): Likewise.
4119 (arm_one_cmplsi2): Likewise.
4120 (thumb_one_complsi2): Likewise.
4121 (notsi_compare0): Likewise.
4122 (notsi_compare0_scratch): Likewise.
4123 (thumb1_zero_extendhisi2): Likewise.
4124 (arm_zero_extendhisi2): Likewise.
4125 (arm_zero_extendhisi2_v6): Likewise.
4126 (arm_zero_extendhisi2addsi): Likewise.
4127 (thumb1_zero_extendqisi2): Likewise.
4128 (thumb1_zero_extendqisi2_v6): Likewise.
4129 (arm_zero_extendqisi2): Likewise.
4130 (arm_zero_extendqisi2_v6): Likewise.
4131 (arm_zero_extendqisi2addsi): Likewise.
4132 (thumb1_extendhisi2): Likewise.
4133 (arm_extendhisi2): Likewise.
4134 (arm_extendhisi2_v6): Likewise.
4135 (arm_extendqisi): Likewise.
4136 (arm_extendqisi_v6): Likewise.
4137 (arm_extendqisi2addsi): Likewise.
4138 (thumb1_extendqisi2): Likewise.
4139 (thumb1_movdi_insn): Likewise.
4140 (arm_movsi_insn): Likewise.
4141 (movsi_compare0): Likewise.
4142 (movhi_insn_arch4): Likewise.
4143 (movhi_bytes): Likewise.
4144 (arm_movqi_insn): Likewise.
4145 (thumb1_movqi_insn): Likewise.
4146 (arm32_movhf): Likewise.
4147 (thumb1_movhf): Likewise.
4148 (arm_movsf_soft_insn): Likewise.
4149 (thumb1_movsf_insn): Likewise.
4150 (movdf_soft_insn): Likewise.
4151 (thumb_movdf_insn): Likewise.
4152 (arm_cmpsi_insn): Likewise.
4153 (cmpsi_shiftsi): Likewise.
4154 (cmpsi_shiftsi_swp): Likewise.
4155 (arm_cmpsi_negshiftsi_si): Likewise.
4156 (movsicc_insn): Likewise.
4157 (movsfcc_soft_insn): Likewise.
4158 (arith_shiftsi): Likewise.
4159 (arith_shiftsi_compare0
4160 (arith_shiftsi_compare0_scratch
4161 (sub_shiftsi): Likewise.
4162 (sub_shiftsi_compare0
4163 (sub_shiftsi_compare0_scratch
4164 (and_scc): Likewise.
4165 (cond_move): Likewise.
4166 (if_plus_move): Likewise.
4167 (if_move_plus): Likewise.
4168 (if_move_not): Likewise.
4169 (if_not_move): Likewise.
4170 (if_shift_move): Likewise.
4171 (if_move_shift): Likewise.
4172 (if_shift_shift): Likewise.
4173 (if_not_arith): Likewise.
4174 (if_arith_not): Likewise.
4175 (cond_move_not): Likewise.
4176 (thumb1_ashlsi3): Set type attribute.
4177 (thumb1_ashrsi3): Likewise.
4178 (thumb1_lshrsi3): Likewise.
4179 (thumb1_rotrsi3): Likewise.
4180 (shiftsi3_compare0_scratch): Likewise.
4181 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
4182 (neon_mov<mode>): Likewise.
4183 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
4184 attribute changes.
4185 (thumb2_movsi_insn): Likewise.
4186 (thumb2_cmpsi_neg_shiftsi): Likewise.
4187 (thumb2_extendqisi_v6): Likewise.
4188 (thumb2_zero_extendhisi2_v6): Likewise.
4189 (thumb2_zero_extendqisi2_v6): Likewise.
4190 (thumb2_shiftsi3_short): Likewise.
4191 (thumb2_addsi3_compare0_scratch): Likewise.
4192 (orsi_not_shiftsi_si): Likewise.
4193 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4194 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
4195 changes.
4196 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4197 (1020alu_shift_op): Likewise.
4198 (1020alu_shift_reg_op): Likewise.
4199 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4200 (alu_shift_op): Likewise.
4201 (alu_shift_reg_op): Likewise.
4202 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4203 (11_alu_shift_op): Likewise.
4204 (11_alu_shift_reg_op): Likewise.
4205 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4206 (9_alu_shift_reg_op): Likewise.
4207 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4208 changes.
4209 (cortex_a15_alu_shift): Likewise.
4210 (cortex_a15_alu_shift_reg): Likewise.
4211 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
4212 changes.
4213 (cortex_a5_alu_shift): Likewise.
4214 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
4215 changes.
4216 (cortex_a53_alu_shift): Likewise.
4217 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4218 changes.
4219 (cortex_a7_alu_reg): Likewise.
4220 (cortex_a7_alu_shift): Likewise.
4221 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
4222 changes.
4223 (cortex_a8_alu_shift): Likewise.
4224 (cortex_a8_alu_shift_reg): Likewise.
4225 (cortex_a8_mov): Likewise.
4226 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4227 (cortex_a9_dp_shift): Likewise.
4228 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
4229 changes.
4230 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
4231 changes.
4232 (cortex_r4_mov): Likewise.
4233 (cortex_r4_alu_shift): Likewise.
4234 (cortex_r4_alu_shift_reg): Likewise.
4235 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4236 (526_alu_shift_op): Likewise.
4237 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4238 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4239 (626te_alu_shift_op): Likewise.
4240 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
4241 (726te_alu_op): Likewise.
4242 (726te_alu_shift_op): Likewise.
4243 (726te_alu_shift_reg_op): Likewise.
4244 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4245 (mp626_alu_shift_op): Likewise.
4246 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
4247 (pj4_alu_e1_conds): Likewise.
4248 (pj4_alu): Likewise.
4249 (pj4_alu_conds): Likewise.
4250 (pj4_shift): Likewise.
4251 (pj4_shift_conds): Likewise.
4252 (pj4_alu_shift): Likewise.
4253 (pj4_alu_shift_conds): Likewise.
4254 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4255 changes.
4256 (cortexa7_older_only): Likewise.
4257 (cortexa7_younger): Likewise.
4258
4259 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4260
4261 * ipa-pure-const.c (generate_summary): Rename to...
4262 (pure_const_generate_summary): ... this.
4263
4264 2013-07-17 Iain Sandoe <iain@codesourcery.com>
4265
4266 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
4267
4268 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
4269
4270 PR target/57909
4271 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
4272 usage in HI mode.
4273
4274 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4275
4276 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
4277 enabled without -march=zEC12.
4278 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
4279 flags to be set.
4280
4281 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4282
4283 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
4284 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
4285 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
4286 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
4287 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
4288 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
4289 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
4290 ISA_HAS_FP4.
4291 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
4292 and ISA_HAS_NMADD3_NMSUB3.
4293 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
4294 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
4295 (nmsub4<mode>, nmsub3<mode>): Likewise.
4296 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
4297
4298 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4299
4300 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
4301 TARGET_MIPS5400 checking.
4302
4303 2013-07-16 Jakub Jelinek <jakub@redhat.com>
4304 Peter Bergner <bergner@vnet.ibm.com>
4305
4306 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
4307 registers in the comment.
4308 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
4309 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
4310 rather than FIRST_PSEUDO_REGISTERS.
4311
4312 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
4313
4314 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
4315 enable extra ISA flags with TARGET_HTM.
4316
4317 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
4318
4319 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
4320 Fix comment typos.
4321
4322 2013-07-15 Cong Hou <congh@google.com>
4323
4324 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
4325 in compare function for sorting.
4326
4327 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
4328
4329 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
4330 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
4331 * config/rs6000/rs6000.opt: Add -mhtm option.
4332 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
4333 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
4334 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
4335 __HTM__ if the HTM instructions are available.
4336 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
4337 htm_spr_reg_operand): New define_predicates.
4338 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
4339 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
4340 Include htm.md.
4341 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
4342 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
4343 HTM builtin functions.
4344 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
4345 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
4346 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
4347 (rs6000_builtin_mask_calculate): Likewise.
4348 (rs6000_option_override_internal): Likewise.
4349 (bdesc_htm): Add new HTM builtin support.
4350 (htm_spr_num): New function.
4351 (htm_spr_regno): Likewise.
4352 (rs6000_htm_spr_icode): Likewise.
4353 (htm_expand_builtin): Likewise.
4354 (htm_init_builtins): Likewise.
4355 (rs6000_expand_builtin): Add support for HTM builtin functions.
4356 (rs6000_init_builtins): Likewise.
4357 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
4358 option.
4359 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
4360 (TARGET_HTM, MASK_HTM): Define macros.
4361 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
4362 (FIXED_REGISTERS): Likewise.
4363 (CALL_USED_REGISTERS): Likewise.
4364 (CALL_REALLY_USED_REGISTERS): Likewise.
4365 (REG_ALLOC_ORDER): Likewise.
4366 (enum reg_class): Likewise.
4367 (REG_CLASS_NAMES): Likewise.
4368 (REG_CLASS_CONTENTS): Likewise.
4369 (REGISTER_NAMES): Likewise.
4370 (ADDITIONAL_REGISTER_NAMES): Likewise.
4371 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
4372 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
4373 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
4374 * config/rs6000/htm.md: New file.
4375 * config/rs6000/htmintrin.h: New file.
4376 * config/rs6000/htmxlintrin.h: New file.
4377
4378 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
4379
4380 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
4381 Define SYMBOL_TINY_GOT, update comment.
4382 * config/aarch64/aarch64.c
4383 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
4384 (aarch64_expand_mov_immediate): Likewise.
4385 (aarch64_print_operand): Likewise.
4386 (aarch64_classify_symbol): Likewise.
4387 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
4388 (ldr_got_tiny): Define.
4389
4390 2013-07-13 Tobias Grosser <tobias@grosser.es>
4391
4392 PR tree-optimization/54094
4393 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
4394 scheduling dimension for the parallelism check from the polyhedral
4395 information in the AST.
4396 * graphite-dependences.c (carries_deps): Do not assume the schedule is
4397 in 2D + 1 form.
4398
4399 2013-07-13 Jason Merrill <jason@redhat.com>
4400
4401 * print-tree.c (debug_vec_tree): Use debug_raw.
4402 (debug_raw (vec<tree, va_gc> &)): New.
4403 (debug_raw (vec<tree, va_gc> *)): New.
4404 * tree.h: Declare them.
4405
4406 2013-07-13 Bin Cheng <bin.cheng@arm.com>
4407
4408 * ifcvt.c (ifcvt_after_combine): New static variable.
4409 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
4410 for size.
4411 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
4412 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
4413 rest_of_handle_if_after_reload): Pass new argument for if_convert.
4414
4415 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
4416
4417 * config/mips/mips.c (mips_expand_call): Remove empty statement.
4418
4419 2013-07-12 Michael Matz <matz@suse.de>
4420
4421 PR middle-end/55771
4422 * convert.c (convert_to_real): Reject non-float inner types.
4423
4424 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
4425
4426 * config/aarch64/aarch64-protos.h
4427 (aarch64_simd_immediate_valid_for_move): Remove.
4428 * config/aarch64/aarch64.c (simd_immediate_info): New member.
4429 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
4430 cases.
4431 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
4432
4433 2013-07-11 Steve Ellcey <sellcey@mips.com>
4434
4435 * config/mips/mips.c (mips_conditional_register_usage): Do not
4436 use t[0-7] registers in MIPS16 mode when optimizing for size.
4437
4438 2013-07-11 Sriraman Tallam <tmsriram@google.com>
4439
4440 * config/i386/i386.c (dispatch_function_versions): Fix array
4441 indexing of function_version_info to match actual_versions.
4442
4443 2013-07-11 Teresa Johnson <tejohnson@google.com>
4444
4445 * vec.h (struct va_gc): Move release out-of-line.
4446 (va_gc::release): Call ggc_free on released vec.
4447
4448 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
4449
4450 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
4451 Require GOT register as additional operand in UNSPEC.
4452 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
4453 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4454 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
4455 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
4456 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
4457 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
4458 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
4459
4460 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
4461
4462 PR target/57631
4463 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
4464 name seen by assembler/linker rather if available.
4465
4466 2013-07-11 Andreas Schwab <schwab@suse.de>
4467
4468 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
4469
4470 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
4471
4472 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
4473
4474 2013-07-10 Joseph Myers <joseph@codesourcery.com>
4475
4476 * doc/tm.texi.in: Move hook documentation to ....
4477 * target.def: ... here.
4478
4479 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
4480 text on @hook line.
4481 * doc/tm.texi: Regenerate.
4482
4483 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
4484
4485 PR c++/57869
4486 * doc/invoke.texi: Document -Wconditionally-supported.
4487
4488 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4489
4490 PR target/57844
4491 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
4492 of my_fp.
4493
4494 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4495
4496 PR target/57506
4497 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
4498 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
4499 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
4500 Remove duplicate devices.
4501 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
4502 * config/avr/t-multilib: Regenerate.
4503 * config/avr/avr-tables.opt: Regenerate.
4504 * doc/avr-mmcu.texi: Regenerate.
4505
4506 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
4507
4508 PR target/56987
4509 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
4510
4511 2013-07-10 Graham Stott <graham.stott@btinternet.com>
4512
4513 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
4514 the cost of MULT when optimizing for size.
4515
4516 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4517
4518 * config/cr16/cr16-protos.h: Don't include target.h.
4519
4520 2013-07-09 Joseph Myers <joseph@codesourcery.com>
4521
4522 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
4523 adjust register size for TDmode and TFmode for VSX registers.
4524
4525 2013-07-08 Kai Tietz <ktietz@redhat.com>
4526
4527 PR target/56892
4528 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
4529 hook_bool_const_tree_true.
4530
4531 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4532
4533 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
4534 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
4535 * config/s390/s390.md: Define FPR*_REGNUM constants.
4536 Fix FPR2_REGNUM constant (18 -> 17).
4537 ("*trunc<BFP:mode><DFP_ALL:mode>2")
4538 ("*trunc<DFP_ALL:mode><BFP:mode>2")
4539 ("trunc<BFP:mode><DFP_ALL:mode>2")
4540 ("trunc<DFP_ALL:mode><BFP:mode>2")
4541 ("*extend<BFP:mode><DFP_ALL:mode>2")
4542 ("*extend<DFP_ALL:mode><BFP:mode>2")
4543 ("extend<BFP:mode><DFP_ALL:mode>2")
4544 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
4545 FPR4_REGNUM.
4546
4547 2013-07-08 Graham Stott <graham.stott@btinternet.com>
4548
4549 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
4550
4551 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4552
4553 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
4554 and cfun_fpr_bit_p to cfun_fpr_save_p.
4555 (s390_frame_area, s390_register_info, s390_frame_info)
4556 (s390_emit_prologue, s390_emit_epilogue)
4557 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
4558 register numbers.
4559 * config/s390/s390.h: Define *_REGNUM macros for floating point
4560 register numbers.
4561
4562 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
4563
4564 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
4565
4566 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
4567
4568 PR rtl-optimization/57786
4569 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
4570 and break out of the loop when it is set to false.
4571
4572 2013-07-08 Jakub Jelinek <jakub@redhat.com>
4573
4574 PR target/57819
4575 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
4576 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
4577 (const_int 63)) 0)).
4578 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
4579 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
4580 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
4581
4582 PR rtl-optimization/57829
4583 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
4584 mask bits outside of mode are just sign-extension from mode to HWI.
4585
4586 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
4587
4588 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
4589 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
4590 adjust_address instead of change_address to keep info about alignment.
4591 (emit_strmov): Remove.
4592 (emit_memmov): New function.
4593 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
4594 (expand_movmem_epilogue): Likewise and return updated rtx for
4595 destination.
4596 (expand_constant_movmem_prologue): Likewise and return updated rtx for
4597 destination and source.
4598 (decide_alignment): Refactor, handle vector_loop.
4599 (ix86_expand_movmem): Likewise.
4600 (ix86_expand_setmem): Likewise.
4601 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
4602
4603 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
4604
4605 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
4606 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
4607
4608 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4609
4610 * config/i386/sse.md (sse_movlhps): Change alternative 3
4611 of operand 2 to "m".
4612
4613 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
4614
4615 PR target/57807
4616 * config/i386/sse.md (iptr): New mode attribute.
4617 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
4618 (<sse>_vm<plusminus_insn><mode>3): Ditto.
4619 (<sse>_vmmul<mode>3): Ditto.
4620 (<sse>_vmdiv<mode>3): Ditto.
4621 (sse_vmrcpv4sf2): Ditto.
4622 (<sse>_vmsqrt<mode>2): Ditto.
4623 (sse_vmrsqrtv4sf2): Ditto.
4624 (<sse>_vm<code><mode>3): Ditto.
4625 (avx_vmcmp<mode>3): Ditto.
4626 (<sse>_vmmaskcmp<mode>3): Ditto.
4627 (<sse>_comi): Ditto.
4628 (<sse>_ucomi): Ditto.
4629 (*xop_vmfrcz_<mode>): Ditto.
4630 (*fmai_fmadd_<mode>): Ditto.
4631 (*fmai_fmsub_<mode>): Ditto.
4632 (*fmai_fnmadd_<mode>): Ditto.
4633 (*fmai_fnmsub_<mode>): Ditto.
4634 (*fma4i_vmfmadd_<mode>): Ditto.
4635 (*fma4i_vmfmsub_<mode>): Ditto.
4636 (*fma4i_vmfnmadd_<mode>): Ditto.
4637 (*fma4i_vmfnmsub_<mode>): Ditto.
4638 (*xop_vmfrcz_<mode>): Ditto.
4639 (sse_cvtps2pi): Ditto.
4640 (sse_cvttps2pi): Ditto.
4641 (sse_cvtss2si): Ditto.
4642 (sse_cvtss2si_2): Ditto.
4643 (sse_cvtss2siq_2): Ditto.
4644 (sse_cvttss2si): Ditto.
4645 (sse_cvttss2siq): Ditto.
4646 (sse_cvtsd2si): Ditto.
4647 (sse_cvtsd2si_2): Ditto.
4648 (sse_cvtsd2siq_2): Ditto.
4649 (sse_cvttsd2si): Ditto.
4650 (sse_cvttsd2siq): Ditto.
4651 (sse_cvtsd2ss): Ditto.
4652 (sse_cvtss2sd): Ditto.
4653 (avx2_pbroadcast<mode>): Ditto.
4654 (avx2_pbroadcast<mode>_1): Ditto.
4655 (*avx_vperm_broadcast_v4sf): Ditto.
4656
4657 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
4658 (sse_movlhps): Ditto.
4659 (sse_storehps): Ditto.
4660 (sse_loadhps): Ditto.
4661 (sse_storelps): Ditto.
4662 (sse_loadlps): Ditto.
4663 (*vec_concatv4sf): Ditto.
4664 (*vec_interleave_highv2df): Ditto.
4665 (*vec_interleave_lowv2df): Ditto.
4666 (*vec_extractv2df_1_sse): Ditto.
4667 (*vec_extractv2df_0_sse): Ditto.
4668 (sse2_storelpd): Ditto.
4669 (sse2_loadlpd): Ditto.
4670 (sse2_movsd): Ditto.
4671 (*vec_concatv4si): Ditto.
4672 (vec_concatv2di): Ditto.
4673
4674 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
4675 for Intel asm dialect.
4676 (mmx_punpcklwd): Ditto.
4677 (mmx_punpckldq): Ditto.
4678
4679 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
4680 for intel assembler dialect.
4681
4682 2013-07-06 Jakub Jelinek <jakub@redhat.com>
4683
4684 PR target/29776
4685 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
4686 for BUILT_IN_C{LZ,LRSB}*.
4687 * tree.h (CASE_INT_FN): Add FN##IMAX case.
4688 * tree-vrp.c (extract_range_basic): Handle
4689 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
4690 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
4691 fall thru to code calling set_value*.
4692 * builtins.c (expand_builtin): Remove *IMAX cases.
4693 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
4694 if width is bigger than 2*HWI.
4695
4696 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
4697
4698 PR rtl-optimization/55342
4699 * lra-int.h (lra_subreg_reload_pseudos): New.
4700 * lra.c: Add undoing optional reloads to the block diagram.
4701 (lra_subreg_reload_pseudos): New.
4702 (lra_optional_reload_pseudos): Change comments.
4703 (lra): Init and clear lra_subreg_reload_pseudos. Clear
4704 lra_optional_reload_pseudos after undo transformations.
4705 * lra-assigns.c (pseudo_prefix_title): New.
4706 (lra_setup_reg_renumber): Use it.
4707 (spill_for): Ditto. Check subreg reload pseudos too.
4708 (assign_by_spills): Consider subreg reload pseudos too.
4709 * lra-constraints.c (simplify_operand_subreg): Use
4710 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
4711 (curr_insn_transform): Recognize and do optional reloads.
4712 (undo_optional_reloads): New.
4713 (lra_undo_inheritance): Call undo_optional_reloads.
4714
4715 2013-07-05 Thomas Quinot <quinot@adacore.com>
4716
4717 * tree-complex.c (expand_complex_operations_1): Fix typo.
4718
4719 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
4720
4721 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
4722 (tune_params): New member 'const vec_costs'.
4723 * config/aarch64/aarch64.c (generic_vector_cost): New.
4724 (generic_tunings): New member 'generic_vector_cost'.
4725 (aarch64_builtin_vectorization_cost): New.
4726 (aarch64_add_stmt_cost): New.
4727 (TARGET_VECTORIZE_ADD_STMT_COST): New.
4728 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
4729
4730 2013-07-03 Jakub Jelinek <jakub@redhat.com>
4731
4732 PR target/57777
4733 * config/i386/predicates.md (vsib_address_operand): Disallow
4734 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
4735
4736 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
4737
4738 PR middle-end/55030
4739 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
4740 expand_builtin_setjmp_receiver.
4741 (expand_label): Adjust, call expand_builtin_setjmp_receiver
4742 with NULL for the label parameter.
4743 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
4744 the frame-pointer. Adjust comments.
4745 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
4746 only if LABEL is non-NULL.
4747
4748 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
4749
4750 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
4751 (ARM_ABI_AAPCS64): Ditto.
4752 (arm_abi): Ditto.
4753 (ARM_DEFAULT_ABI): Ditto.
4754
4755 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
4756
4757 * config/aarch64/aarch64-builtins.c
4758 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
4759 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
4760 (st1): Likewise.
4761 * config/aarch64/aarch64-simd.md
4762 (aarch64_ld1<VALL:mode>): New.
4763 (aarch64_st1<VALL:mode>): Likewise.
4764 * config/aarch64/arm_neon.h
4765 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
4766
4767 2013-07-02 Sriraman Tallam <tmsriram@google.com>
4768
4769 * config/i386/i386.c (gate_insert_vzeroupper): Check if
4770 target ISA is AVX.
4771 (ix86_option_override_internal):Turn on all -mavx target flags by
4772 default as they are dependent on AVX anyway.
4773
4774 2013-07-02 Cary Coutant <ccoutant@google.com>
4775
4776 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
4777 deterministic hash.
4778 (loc_checksum_ordered): Likewise.
4779 (hash_loc_operands): Remove inline keyword.
4780
4781 2013-07-02 Jakub Jelinek <jakub@redhat.com>
4782
4783 PR tree-optimization/57741
4784 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
4785 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
4786 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
4787 Allow REAL_CST step_exprs if flag_associative_math.
4788 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
4789
4790 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4791
4792 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
4793
4794 2013-07-02 Ian Bolton <ian.bolton@arm.com>
4795
4796 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
4797
4798 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4799
4800 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
4801 encoding.
4802 (iorsi3_insn): Likewise.
4803 (arm_xorsi3): Likewise.
4804
4805 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
4806
4807 * arm.md (attribute "wtype"): Delete. Move attribute values from here
4808 to ...
4809 (attribute "type"): ... here, and prefix with "wmmx_".
4810 (attribute "core_cycles"): Update for attribute changes.
4811 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
4812 (tbcstv4hi): Likewise.
4813 (tbcstv2si): Likewise.
4814 (iwmmxt_iordi3): Likewise.
4815 (iwmmxt_xordi3): Likewise.
4816 (iwmmxt_anddi3): Likewise.
4817 (iwmmxt_nanddi3): Likewise.
4818 (iwmmxt_arm_movdi): Likewise.
4819 (iwmmxt_movsi_insn): Likewise.
4820 (mov<mode>_internal): Likewise.
4821 (and<mode>3_iwmmxt): Likewise.
4822 (ior<mode>3_iwmmxt): Likewise.
4823 (xor<mode>3_iwmmxt): Likewise.
4824 (add<mode>3_iwmmxt): Likewise.
4825 (ssaddv8qi3): Likewise.
4826 (ssaddv4hi3): Likewise.
4827 (ssaddv2si3): Likewise.
4828 (usaddv8qi3): Likewise.
4829 (usaddv4hi3): Likewise.
4830 (usaddv2si3): Likewise.
4831 (sub<mode>3_iwmmxt): Likewise.
4832 (sssubv8qi3): Likewise.
4833 (sssubv4hi3): Likewise.
4834 (sssubv2si3): Likewise.
4835 (ussubv8qi3): Likewise.
4836 (ussubv4hi3): Likewise.
4837 (ussubv2si3): Likewise.
4838 (mulv4hi3_iwmmxt): Likewise.
4839 (smulv4hi3_highpart): Likewise.
4840 (umulv4hi3_highpart): Likewise.
4841 (iwmmxt_wmacs): Likewise.
4842 (iwmmxt_wmacsz): Likewise.
4843 (iwmmxt_wmacu): Likewise.
4844 (iwmmxt_wmacuz): Likewise.
4845 (iwmmxt_clrdi): Likewise.
4846 (iwmmxt_clrv8qi): Likewise.
4847 (iwmmxt_clr4hi): Likewise.
4848 (iwmmxt_clr2si): Likewise.
4849 (iwmmxt_uavgrndv8qi3): Likewise.
4850 (iwmmxt_uavgrndv4hi3): Likewise.
4851 (iwmmxt_uavgv8qi3): Likewise.
4852 (iwmmxt_uavgv4hi3): Likewise.
4853 (iwmmxt_tinsrb): Likewise.
4854 (iwmmxt_tinsrh): Likewise.
4855 (iwmmxt_tinsrw): Likewise.
4856 (iwmmxt_textrmub): Likewise.
4857 (iwmmxt_textrmsb): Likewise.
4858 (iwmmxt_textrmuh): Likewise.
4859 (iwmmxt_textrmsh): Likewise.
4860 (iwmmxt_textrmw): Likewise.
4861 (iwmxxt_wshufh): Likewise.
4862 (eqv8qi3): Likewise.
4863 (eqv4hi3): Likewise.
4864 (eqv2si3): Likewise.
4865 (gtuv8qi3): Likewise.
4866 (gtuv4hi3): Likewise.
4867 (gtuv2si3): Likewise.
4868 (gtv8qi3): Likewise.
4869 (gtv4hi3): Likewise.
4870 (gtv2si3): Likewise.
4871 (smax<mode>3_iwmmxt): Likewise.
4872 (umax<mode>3_iwmmxt): Likewise.
4873 (smin<mode>3_iwmmxt): Likewise.
4874 (umin<mode>3_iwmmxt): Likewise.
4875 (iwmmxt_wpackhss): Likewise.
4876 (iwmmxt_wpackwss): Likewise.
4877 (iwmmxt_wpackdss): Likewise.
4878 (iwmmxt_wpackhus): Likewise.
4879 (iwmmxt_wpackwus): Likewise.
4880 (iwmmxt_wpackdus): Likewise.
4881 (iwmmxt_wunpckihb): Likewise.
4882 (iwmmxt_wunpckihh): Likewise.
4883 (iwmmxt_wunpckihw): Likewise.
4884 (iwmmxt_wunpckilb): Likewise.
4885 (iwmmxt_wunpckilh): Likewise.
4886 (iwmmxt_wunpckilw): Likewise.
4887 (iwmmxt_wunpckehub): Likewise.
4888 (iwmmxt_wunpckehuh): Likewise.
4889 (iwmmxt_wunpckehuw): Likewise.
4890 (iwmmxt_wunpckehsb): Likewise.
4891 (iwmmxt_wunpckehsh): Likewise.
4892 (iwmmxt_wunpckehsw): Likewise.
4893 (iwmmxt_wunpckelub): Likewise.
4894 (iwmmxt_wunpckeluh): Likewise.
4895 (iwmmxt_wunpckeluw): Likewise.
4896 (iwmmxt_wunpckelsb): Likewise.
4897 (iwmmxt_wunpckelsh): Likewise.
4898 (iwmmxt_wunpckelsw): Likewise.
4899 (ror<mode>3): Likewise.
4900 (ashr<mode>3_iwmmxt): Likewise.
4901 (lshr<mode>3_iwmmxt): Likewise.
4902 (ashl<mode>3_iwmmxt): Likewise.
4903 (ror<mode>3_di): Likewise.
4904 (ashr<mode>3_di): Likewise.
4905 (lshr<mode>3_di): Likewise.
4906 (ashl<mode>3_di): Likewise.
4907 (iwmmxt_wmadds): Likewise.
4908 (iwmmxt_wmaddu): Likewise.
4909 (iwmmxt_tmia): Likewise.
4910 (iwmmxt_tmiaph): Likewise.
4911 (iwmmxt_tmiabb): Likewise.
4912 (iwmmxt_tmiatb): Likewise.
4913 (iwmmxt_tmiabt): Likewise.
4914 (iwmmxt_tmiatt): Likewise.
4915 (iwmmxt_tmovmskb): Likewise.
4916 (iwmmxt_tmovmskh): Likewise.
4917 (iwmmxt_tmovmskw): Likewise.
4918 (iwmmxt_waccb): Likewise.
4919 (iwmmxt_wacch): Likewise.
4920 (iwmmxt_waccw): Likewise.
4921 (iwmmxt_waligni): Likewise.
4922 (iwmmxt_walignr): Likewise.
4923 (iwmmxt_walignr0): Likewise.
4924 (iwmmxt_walignr1): Likewise.
4925 (iwmmxt_walignr2): Likewise.
4926 (iwmmxt_walignr3): Likewise.
4927 (iwmmxt_wsadb): Likewise.
4928 (iwmmxt_wsadh): Likewise.
4929 (iwmmxt_wsadbz): Likewise.
4930 (iwmmxt_wsadhz): Likewise.
4931 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
4932 (iwmmxt_wabsdiffb): Likewise.
4933 (iwmmxt_wabsdiffh): Likewise.
4934 (iwmmxt_wabsdiffw): Likewise.
4935 (iwmmxt_waddsubhx): Likewise
4936 (iwmmxt_wsubaddhx): Likewise.
4937 (addc<mode>3): Likewise.
4938 (iwmmxt_avg4): Likewise.
4939 (iwmmxt_avg4r): Likewise.
4940 (iwmmxt_wmaddsx): Likewise.
4941 (iwmmxt_wmaddux): Likewise.
4942 (iwmmxt_wmaddsn): Likewise.
4943 (iwmmxt_wmaddun): Likewise.
4944 (iwmmxt_wmulwsm): Likewise.
4945 (iwmmxt_wmulwum): Likewise.
4946 (iwmmxt_wmulsmr): Likewise.
4947 (iwmmxt_wmulumr): Likewise.
4948 (iwmmxt_wmulwsmr): Likewise.
4949 (iwmmxt_wmulwumr): Likewise.
4950 (iwmmxt_wmulwl): Likewise.
4951 (iwmmxt_wqmulm): Likewise.
4952 (iwmmxt_wqmulwm): Likewise.
4953 (iwmmxt_wqmulmr): Likewise.
4954 (iwmmxt_wqmulwmr): Likewise.
4955 (iwmmxt_waddbhusm): Likewise.
4956 (iwmmxt_waddbhusl): Likewise.
4957 (iwmmxt_wqmiabb): Likewise.
4958 (iwmmxt_wqmiabt): Likewise.
4959 (iwmmxt_wqmiatb): Likewise.
4960 (iwmmxt_wqmiatt): Likewise.
4961 (iwmmxt_wqmiabbn): Likewise.
4962 (iwmmxt_wqmiabtn): Likewise.
4963 (iwmmxt_wqmiatbn): Likewise.
4964 (iwmmxt_wqmiattn): Likewise.
4965 (iwmmxt_wmiabb): Likewise.
4966 (iwmmxt_wmiabt): Likewise.
4967 (iwmmxt_wmiatb): Likewise.
4968 (iwmmxt_wmiatt): Likewise.
4969 (iwmmxt_wmiabbn): Likewise.
4970 (iwmmxt_wmiabtn): Likewise.
4971 (iwmmxt_wmiatbn): Likewise.
4972 (iwmmxt_wmiattn): Likewise.
4973 (iwmmxt_wmiawbb): Likewise.
4974 (iwmmxt_wmiawbt): Likewise.
4975 (iwmmxt_wmiawtb): Likewise.
4976 (iwmmxt_wmiawtt): Likewise.
4977 (iwmmxt_wmiawbbn): Likewise.
4978 (iwmmxt_wmiawbtn): Likewise.
4979 (iwmmxt_wmiawtbn): Likewise.
4980 (iwmmxt_wmiawttn): Likewise.
4981 (iwmmxt_wmerge): Likewise.
4982 (iwmmxt_tandc<mode>3): Likewise.
4983 (iwmmxt_torc<mode>3): Likewise.
4984 (iwmmxt_torvsc<mode>3): Likewise.
4985 (iwmmxt_textrc<mode>3): Likewise.
4986 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
4987 (wmmxt_pack): Likewise.
4988 (wmmxt_mult_c1): Likewise.
4989 (wmmxt_mult_c2): Likewise.
4990 (wmmxt_alu_c1): Likewise.
4991 (wmmxt_alu_c2): Likewise.
4992 (wmmxt_alu_c3): Likewise.
4993 (wmmxt_transfer_c1): Likewise.
4994 (wmmxt_transfer_c2): Likewise.
4995 (wmmxt_transfer_c3): Likewise.
4996 (marvell_f_iwmmxt_wstr): Likewise.
4997 (marvell_f_iwmmxt_wldr): Likewise.
4998
4999 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5000
5001 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5002
5003 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5004
5005 Revert:
5006 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5007 * lra-constraints.c (need_for_split_p): Check call used hard regs
5008 living through calls.
5009
5010 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5011 call used regs for call insn.
5012
5013 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5014
5015 PR target/57736
5016 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5017 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5018
5019 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5020
5021 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5022
5023 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5024
5025 * lra-constraints.c (need_for_split_p): Check call used hard regs
5026 living through calls.
5027
5028 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5029
5030 PR target/57744
5031 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5032 to tie with any other modes. Eliminate Altivec vector mode tests,
5033 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5034 code, to return 0 if testing MODE2 for a condition, if we've
5035 already tested MODE1 for the same condition.
5036
5037 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5038
5039 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5040 layout.
5041
5042 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5043
5044 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5045 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5046
5047 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5048
5049 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5050 Define.
5051 (aarch64_symbolic_constant_p): Remove.
5052 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5053 static. Fix line length and white space.
5054 (aarch64_symbolic_constant_p): Remove.
5055 * config/aarch64/predicates.md (aarch64_valid_symref):
5056 Use aarch64_classify_symbol_expression.
5057
5058 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5059
5060 * config/arm/constraints.md (Ts): New constraint.
5061 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5062 16-bit encodings.
5063 (compare_scc): Use "Ts" constraint for operand 0.
5064 (ior_scc_scc): Likewise.
5065 (and_scc_scc): Likewise.
5066 (and_scc_scc_nodom): Likewise.
5067 (ior_scc_scc_cmp): Likewise for operand 7.
5068 (and_scc_scc_cmp): Likewise.
5069 * config/arm/thumb2.md (thumb2_movsi_insn):
5070 Add alternatives for 16-bit encodings.
5071 (thumb2_movhi_insn): Likewise.
5072 (thumb2_movsicc_insn): Likewise.
5073 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5074 (thumb2_negscc): Use "Ts" constraint.
5075 Move mvn instruction outside cond_exec block.
5076 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5077 for 16-bit encodings.
5078
5079 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5080
5081 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5082 encoding.
5083 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5084 (mulsi3subsi): Likewise.
5085 (mulsidi3adddi): Likewise.
5086 (mulsidi3_v6): Likewise.
5087 (umulsidi3_v6): Likewise.
5088 (umulsidi3adddi_v6): Likewise.
5089 (smulsi3_highpart_v6): Likewise.
5090 (umulsi3_highpart_v6): Likewise.
5091 (mulhisi3tb): Likewise.
5092 (mulhisi3bt): Likewise.
5093 (mulhisi3tt): Likewise.
5094 (maddhisi4): Likewise.
5095 (maddhisi4tb): Likewise.
5096 (maddhisi4tt): Likewise.
5097 (maddhidi4): Likewise.
5098 (maddhidi4tb): Likewise.
5099 (maddhidi4tt): Likewise.
5100 (zeroextractsi_compare0_scratch): Likewise.
5101 (insv_zero): Likewise.
5102 (insv_t2): Likewise.
5103 (anddi_notzesidi_di): Likewise.
5104 (anddi_notsesidi_di): Likewise.
5105 (andsi_notsi_si): Likewise.
5106 (iordi_zesidi_di): Likewise.
5107 (xordi_zesidi_di): Likewise.
5108 (andsi_iorsi3_notsi): Likewise.
5109 (smax_0): Likewise.
5110 (smax_m1): Likewise.
5111 (smin_0): Likewise.
5112 (not_shiftsi): Likewise.
5113 (unaligned_loadsi): Likewise.
5114 (unaligned_loadhis): Likewise.
5115 (unaligned_loadhiu): Likewise.
5116 (unaligned_storesi): Likewise.
5117 (unaligned_storehi): Likewise.
5118 (extv_reg): Likewise.
5119 (extzv_t2): Likewise.
5120 (divsi3): Likewise.
5121 (udivsi3): Likewise.
5122 (arm_zero_extendhisi2addsi): Likewise.
5123 (arm_zero_extendqisi2addsi): Likewise.
5124 (compareqi_eq0): Likewise.
5125 (arm_extendhisi2_v6): Likewise.
5126 (arm_extendqisi2addsi): Likewise.
5127 (arm_movt): Likewise.
5128 (thumb2_ldrd): Likewise.
5129 (thumb2_ldrd_base): Likewise.
5130 (thumb2_ldrd_base_neg): Likewise.
5131 (thumb2_strd): Likewise.
5132 (thumb2_strd_base): Likewise.
5133 (thumb2_strd_base_neg): Likewise.
5134 (arm_negsi2): Add alternative for 16-bit encoding.
5135 (arm_one_cmplsi2): Likewise.
5136
5137 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5138
5139 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
5140 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
5141 (movdfcc): Likewise.
5142 * config/arm/vfp.md (*thumb2_movsf_vfp):
5143 Disable predication for arm_restrict_it.
5144 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
5145 (*thumb2_movdfcc_vfp): Likewise.
5146 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
5147 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
5148 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
5149 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
5150 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
5151 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
5152 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
5153 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
5154 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
5155 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
5156 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
5157 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
5158 Disable predication for arm_restrict_it.
5159
5160 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
5161
5162 * config/i386/bmiintrin.h (_bextr_u32): New.
5163 (_bextr_u64): Ditto.
5164
5165 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
5166
5167 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
5168 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
5169 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
5170 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
5171 * config/mips/n32-elf.h: ...this new file.
5172
5173 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
5174
5175 PR target/57224
5176 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
5177 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
5178
5179 2013-06-27 Catherine Moore <clm@codesourcery.com>
5180
5181 * config/mips/mips-tables.opt: Regenerate.
5182 * config/mips/mips-cpus.def: Add m14ke and m14kec.
5183 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
5184 * doc/invoke.texi: Add -m14kc.
5185
5186 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5187
5188 PR target/57623
5189 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
5190 constraints of operand 1 and 2.
5191
5192 PR target/57623
5193 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
5194 to match RTL canonicalization. Swap predicates and
5195 constraints of operand 1 and 2.
5196
5197 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
5198
5199 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
5200 Process OP_INOUT regs for splitting too.
5201
5202 2013-06-27 Jakub Jelinek <jakub@redhat.com>
5203
5204 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
5205 decl before the loop, initialize to NULL.
5206 (vectorizable_load): Initialize ptr_incr to NULL.
5207
5208 2013-06-27 Martin Jambor <mjambor@suse.cz>
5209
5210 PR lto/57208
5211 * ipa-ref.h (ipa_maybe_record_reference): Declare.
5212 * ipa-ref.c (ipa_maybe_record_reference): New function.
5213 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
5214 * ipa-cp.c (create_specialized_node): Record potential references from
5215 aggvals.
5216 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
5217
5218 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5219
5220 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
5221 parameter 'mode' of type 'enum machine_mode mode'; change to pass
5222 'mode' to force_reg.
5223 (aarch64_add_offset): Update calls to aarch64_force_temporary.
5224 (aarch64_expand_mov_immediate): Likewise.
5225
5226 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
5227
5228 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
5229 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
5230
5231 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5232
5233 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
5234 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5235 (struct machine_function): Add tbegin_p.
5236 (s390_canonicalize_comparison): Fold CC mode compares to
5237 conditional jump if possible.
5238 (s390_emit_jump): Return the emitted jump.
5239 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
5240 Handle CCRAWmode compares.
5241 (s390_option_override): Default to -mhtm if available.
5242 (s390_reg_clobbered_rtx): Handle floating point regs as well.
5243 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
5244 FPRs instead of df_regs_ever_live_p.
5245 (s390_optimize_nonescaping_tx): New function.
5246 (s390_init_frame_layout): Extend clobbered_regs array to cover
5247 FPRs as well.
5248 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
5249 (s390_expand_tbegin): New function.
5250 (enum s390_builtin): New enum definition.
5251 (code_for_builtin): New array definition.
5252 (s390_init_builtins): New function.
5253 (s390_expand_builtin): New function.
5254 (TARGET_INIT_BUILTINS): Define.
5255 (TARGET_EXPAND_BUILTIN): Define.
5256 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5257 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
5258 (s390_alc_comparison): Likewise.
5259 * config/s390/s390-modes.def: Add CCRAWmode.
5260 * config/s390/s390.h (processor_flags): Add PF_TX.
5261 (TARGET_CPU_HTM): Define macro.
5262 (TARGET_HTM): Define macro.
5263 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
5264 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
5265 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
5266 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
5267 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5268 (TBEGIN_MASK, TBEGINC_MASK): New constants.
5269 ("*cc_to_int"): Move up.
5270 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
5271 constants other than 0.
5272 ("*ccraw_to_int"): New insn and splitter definition.
5273 ("tbegin", "tbegin_nofloat", "tbegin_retry")
5274 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
5275 ("tx_assist"): New expander.
5276 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
5277 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
5278 * config/s390/s390.opt: Add -mhtm option.
5279 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
5280 * config/s390/htmxlintrin.h: New file.
5281 * config/s390/htmintrin.h: New file.
5282 * config/s390/s390intrin.h: New file.
5283 * doc/extend.texi: Document htm builtins.
5284 * config.gcc: Add the new header files to extra_headers.
5285
5286 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
5287
5288 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
5289 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
5290
5291 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
5292 Pat Haugen <pthaugen@us.ibm.com>
5293 Peter Bergner <bergner@vnet.ibm.com>
5294
5295 * config/rs6000/power8.md: New.
5296 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
5297 setting for power8 entry.
5298 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
5299 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
5300 test for Power4/Power5 only.
5301 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
5302 support.
5303 (force_new_group): Adjust comment.
5304 * config/rs6000/rs6000.md: Include power8.md.
5305
5306 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
5307
5308 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
5309 * config/arm/arm-protos.h (arm_max_conditional_execute): New
5310 declaration.
5311 (tune_params): Update comment.
5312 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
5313 (arm_max_conditional_execute): New function.
5314 (thumb2_final_prescan_insn): Use max_insn_skipped and
5315 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
5316
5317 2013-06-25 Jakub Jelinek <jakub@redhat.com>
5318
5319 PR tree-optimization/57705
5320 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
5321 SSA_NAME step, provided that it is not defined inside the loop.
5322 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
5323 (get_initial_def_for_induction): Handle SSA_NAME IV step.
5324
5325 2013-06-25 Martin Jambor <mjambor@suse.cz>
5326
5327 PR middle-end/57670
5328 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
5329 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
5330 calls in the dump.
5331 (ipa_note_param_call): Initialize member_ptr flag.
5332 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
5333 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
5334 (ipa_write_indirect_edge_info): Stream member_ptr flag.
5335 (ipa_read_indirect_edge_info): Likewise.
5336
5337 2013-06-25 Richard Biener <rguenther@suse.de>
5338
5339 PR middle-end/56977
5340 * passes.c (init_optimization_passes): Move pass_fold_builtins
5341 and pass_dce earlier with -Og.
5342
5343 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
5344
5345 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
5346 <BIT_FIELD_REF>: Remove trailing TAB.
5347 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
5348 remove blank line.
5349
5350 2013-06-24 Martin Jambor <mjambor@suse.cz>
5351
5352 PR tree-optimization/57358
5353 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
5354 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
5355 (ipa_analyze_params_uses): Generate pessimistic info when true.
5356
5357 2013-06-24 Martin Jambor <mjambor@suse.cz>
5358
5359 PR tree-optimization/57539
5360 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
5361 global.inlined_to of the new node to it. All callers changed.
5362 * ipa-inline-transform.c (clone_inlined_nodes): New variable
5363 inlining_into, pass it to cgraph_clone_node.
5364 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
5365 ipa_free_edge_args_substructures.
5366 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
5367 rdesc linked list. Do not assert rdesc edges have inlined caller.
5368 Assert we have found an rdesc in the rdesc list.
5369
5370 2013-06-24 Richard Biener <rguenther@suse.de>
5371
5372 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
5373 (pointer_set_lookup): Declare.
5374 (class pointer_map): New template class implementing a
5375 generic pointer to T map.
5376 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
5377 pointer_map<T>::contains, pointer_map<T>::insert,
5378 pointer_map<T>::traverse): New functions.
5379 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
5380 (pointer_set_lookup): New function.
5381 (pointer_set_contains): Use pointer_set_lookup.
5382 (pointer_set_insert): Likewise.
5383 (insert_aux): Remove.
5384 (struct pointer_map_t): Embed a pointer_set_t.
5385 (pointer_map_create): Adjust.
5386 (pointer_map_destroy): Likewise.
5387 (pointer_map_contains): Likewise.
5388 (pointer_map_insert): Likewise.
5389 (pointer_map_traverse): Likewise.
5390 * tree-streamer.h (struct streamer_tree_cache_d): Use a
5391 pointer_map<unsigned> instead of a pointer_map_t.
5392 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
5393 (streamer_tree_cache_lookup): Likewise.
5394 (streamer_tree_cache_create): Likewise.
5395 (streamer_tree_cache_delete): Likewise.
5396 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
5397 pointer_map<unsigned> instead of a pointer_map_t.
5398 (lto_init_tree_ref_encoder): Adjust.
5399 (lto_destroy_tree_ref_encoder): Likewise.
5400 * lto-section-out.c (lto_output_decl_index): Likewise.
5401 (lto_record_function_out_decl_state): Likewise.
5402 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
5403
5404 2013-06-24 Richard Biener <rguenther@suse.de>
5405
5406 PR tree-optimization/57488
5407 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
5408
5409 2013-06-24 Alan Modra <amodra@gmail.com>
5410
5411 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
5412 (gen_easy_altivec_constant): Likewise.
5413 * config/rs6000/predicates.md (easy_vector_constant_add_self,
5414 easy_vector_constant_msb): Likewise.
5415
5416 2013-06-23 Jakub Jelinek <jakub@redhat.com>
5417
5418 PR target/57688
5419 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
5420 add missing return true.
5421
5422 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
5423
5424 PR target/52483
5425 * config/sh/predicates.md (general_extend_operand): Invoke
5426 general_movsrc_operand for memory operands.
5427 (general_movsrc_operand): Allow reg+reg addressing, do not use
5428 general_operand for memory operands.
5429
5430 2013-06-23 Sriraman Tallam <tmsriram@google.com>
5431
5432 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
5433 when current target options does not apply.
5434 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
5435 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
5436 * config/i386/bmiintrin.h: Pass appropriate target
5437 attributes to header.
5438 * config/i386/mmintrin.h: Ditto.
5439 * config/i386/nmmintrin.h: Ditto.
5440 * config/i386/avx2intrin.h: Ditto.
5441 * config/i386/fxsrintrin.h: Ditto.
5442 * config/i386/tbmintrin.h: Ditto.
5443 * config/i386/xsaveintrin.h: Ditto.
5444 * config/i386/f16cintrin.h: Ditto.
5445 * config/i386/xtestintrin.h: Ditto.
5446 * config/i386/xsaveoptintrin.h: Ditto.
5447 * config/i386/bmi2intrin.h: Ditto.
5448 * config/i386/lzcntintrin.h: Ditto.
5449 * config/i386/smmintrin.h: Ditto.
5450 * config/i386/wmmintrin.h: Ditto.
5451 * config/i386/x86intrin.h: Remove all header include guards.
5452 * config/i386/prfchwintrin.h: Ditto.
5453 * config/i386/pmmintrin.h: Ditto.
5454 * config/i386/tmmintrin.h: Ditto.
5455 * config/i386/xmmintrin.h: Ditto.
5456 * config/i386/popcntintrin.h: Ditto.
5457 * config/i386/rdseedintrin.h: Ditto.
5458 * config/i386/ammintrin.h: Ditto.
5459 * config/i386/emmintrin.h: Ditto.
5460 * config/i386/immintrin.h: Remove all header include guards.
5461 * config/i386/fma4intrin.h: Ditto.
5462 * config/i386/lwpintrin.h: Ditto.
5463 * config/i386/xopintrin.h: Ditto.
5464 * config/i386/ia32intrin.h: Ditto.
5465 * config/i386/avxintrin.h: Ditto.
5466 * config/i386/rtmintrin.h: Ditto.
5467 * config/i386/fmaintrin.h: Ditto.
5468 * config/i386/mm3dnow.h: Ditto.
5469
5470 2013-06-22 Sriraman Tallam <tmsriram@google.com>
5471
5472 * common/config/i386/i386-common.c: Handle LZCNT.
5473
5474 2013-06-22 Andi Kleen <ak@linux.intel.com>
5475
5476 * doc/extend.texi: Use __atomic_store_n instead of
5477 __atomic_store in HLE example.
5478
5479 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
5480
5481 * config/sh/sh.c: Remove <cstdlib> workaround.
5482
5483 2013-06-21 Andi Kleen <ak@linux.intel.com>
5484
5485 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
5486
5487 2013-06-21 Andi Kleen <ak@linux.intel.com>
5488
5489 * doc/extend.texi: Document that __atomic_clear and
5490 __atomic_test_and_set should only be used with bool.
5491
5492 2013-06-20 Jan Hubicka <jh@suse.cz>
5493
5494 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
5495 types_same_for_odr.
5496 * tree.c (decls_same_for_odr): New function.
5497 (same_for_edr): New function.
5498 (types_same_for_odr): New function.
5499 (get_binfo_at_offset): Use it.
5500 * tree.h (types_same_for_odr): Declare.
5501
5502 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
5503 Jason Merrill <jason@redhat.com>
5504
5505 * system.h: Include <cstdlib> as well as <stdlib.h>.
5506
5507 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
5508
5509 PR target/57655
5510 * config/i386/i386.c (construct_container): Report error if
5511 long double is used with disabled x87 float returns.
5512
5513 2013-06-20 Jan Hubicka <jh@suse.cz>
5514
5515 * lto-cgraph.c (input_symtab): Do not set cgraph state.
5516
5517 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
5518
5519 PR rtl-optimization/57425
5520 PR rtl-optimization/57569
5521 * alias.c (write_dependence_p): Remove parameters mem_mode and
5522 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
5523 Changed all callers.
5524 (canon_anti_dependence): Get comments and semantics in sync.
5525 Add parameter mem_canonicalized. Changed all callers.
5526 * rtl.h (canon_anti_dependence): Update prototype.
5527
5528 2013-06-20 Richard Biener <rguenther@suse.de>
5529
5530 * data-streamer-in.c (streamer_read_uhwi): Optimize single
5531 byte case, inline streamer_read_uchar and defer section
5532 overrun check.
5533
5534 2013-06-20 Richard Biener <rguenther@suse.de>
5535
5536 PR tree-optimization/57584
5537 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
5538 SSA names into the expanded expression that take part in
5539 abnormal coalescing.
5540
5541 2013-06-19 Sharad Singhai <singhai@google.com>
5542
5543 * gcov.c (print_usage): Handle new option.
5544 (process_args): Ditto.
5545 (get_gcov_intermediate_filename): New function.
5546 (output_intermediate_file): New function.
5547 (output_gcov_file): New function
5548 (generate_results): Handle new option.
5549 (release_function): Relase demangled name.
5550 (read_graph_file): Handle demangled name.
5551 (output_lines): Ditto.
5552 * doc/gcov.texi: Document gcov intermediate format.
5553
5554 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
5555
5556 PR bootstrap/57604
5557 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
5558 (lra_emit_add): Use the functions. Add comment about Y as an
5559 address segment.
5560
5561 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
5562
5563 PR driver/57652
5564 * collect2.c (collect_atexit): New.
5565 (collect_exit): Delete.
5566 (main): Register collect_atexit with atexit.
5567 (collect_wait): Change collect_exit to exit.
5568 (do_wait): Same.
5569 * collect2.h (collect_exit): Delete.
5570 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
5571
5572 2013-06-19 Wei Mi <wmi@google.com>
5573
5574 PR rtl-optimization/57518
5575 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
5576 if regno is used in paradoxical subreg.
5577 (update_equiv_regs): Check pdx_subregs[regno] before
5578 set a reg to be equivalent with a mem.
5579
5580 2013-06-19 Matthias Klose <doko@ubuntu.com>
5581
5582 PR driver/57651
5583 * file-find.h (find_a_file): Add a mode parameter.
5584 * file-find.c (find_a_file): Likewise.
5585 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
5586 with X_OK for the executables.
5587 * collect2.c (main): Call find_a_file with X_OK.
5588
5589 2013-06-19 Steve Ellcey <sellcey@mips.com>
5590
5591 PR target/56942
5592 * config/mips/mips.md (casesi_internal_mips16_<mode>):
5593 Use NEXT_INSN instead of next_real_insn.
5594
5595 2013-06-19 Jan Hubicka <jh@suse.cz>
5596
5597 * cgraph.h (const_value_known_p): Replace by ...
5598 (ctor_for_folding): .. this one.
5599 * cgraphunit.c (process_function_and_variable_attributes): Use it.
5600 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
5601 * expr.c (expand_expr_real_1): Likewise.
5602 (string_constant): Likewise.
5603 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
5604 * ipa.c (process_references): Likewise.
5605 (symtab_remove_unreachable_nodes): Likewise.
5606 * ipa-inline-analysis.c (param_change_prob): Likewise.
5607 * gimple-fold.c (canonicalize_constructor_val): Likewise.
5608 (get_base_constructor): Likwise.
5609 * varpool.c (varpool_remove_node): Likewise.
5610 (varpool_remove_initializer): LIkewise.
5611 (dump_varpool_node): LIkwise.
5612 (const_value_known_p): Rewrite to ...
5613 (ctor_for_folding): ... this one.
5614
5615 2013-06-19 Jakub Jelinek <jakub@redhat.com>
5616
5617 PR driver/57651
5618 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
5619 PERSONALITY in $PATH derived prefixes.
5620
5621 2013-06-19 Jeff Law <law@redhat.com>
5622
5623 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
5624 in comment.
5625
5626 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
5627 (simplify_bitwise_binary): Use it to simpify certain binary ops on
5628 booleans.
5629
5630 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
5631
5632 * config/arm/vfp.md: Move VFP instruction classification documentation
5633 to ...
5634 * config/arm/arm.md: ... here. Update instruction classification
5635 documentation.
5636
5637 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
5638
5639 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
5640 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
5641 pattern. Use more efficient sequences on ARMv5 and Thumb2.
5642
5643 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
5644
5645 PR target/57609
5646 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5647 with NEXT_INSN. Use tablejump_p to check for jump table data
5648 insns.
5649
5650 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
5651
5652 PR c++/56544
5653 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
5654 that now in C++ the value is correct per the C++ standards.
5655
5656 2013-06-19 Richard Biener <rguenther@suse.de>
5657
5658 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
5659 for global context.
5660
5661 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5662
5663 Revert:
5664 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5665
5666 PR target/57609
5667 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5668 with next_active_insn.
5669
5670 2013-06-18 Sriraman Tallam <tmsriram@google.com>
5671
5672 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
5673 functions are inlined during failures to flag an error.
5674 * tree-inline.c (expand_call_inline): Allow the error to be flagged
5675 in early inline pass.
5676
5677 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
5678
5679 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
5680 in comments.
5681
5682 2013-06-18 Julian Brown <julian@codesourcery.com>
5683
5684 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
5685 Permit virtual register pre-reload if !strict.
5686 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
5687 change.
5688 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
5689 prototype.
5690 * config/arm/neon.md (movmisalign<mode>): Use
5691 neon_perm_struct_or_reg_operand instead of
5692 neon_struct_or_register_operand.
5693 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
5694 neon_permissive_struct_operand instead of neon_struct_operand.
5695 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
5696 neon_vector_mem_operand.
5697 * config/arm/predicates.md (neon_struct_operand): Adjust call to
5698 neon_vector_mem_operand.
5699 (neon_permissive_struct_operand): New.
5700 (neon_struct_or_register_operand): Rename to...
5701 (neon_perm_struct_or_reg_operand): This. Adjust call to
5702 neon_vector_mem_operand.
5703
5704 2013-06-18 Richard Biener <rguenther@suse.de>
5705
5706 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
5707 * lto-streamer.h: Include pointer-set.h.
5708 (struct lto_decl_slot): Remove.
5709 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
5710 Remove next_index entry.
5711 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5712 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5713 (lto_init_tree_ref_encoder): Adjust.
5714 (lto_destroy_tree_ref_encoder): Likewise.
5715 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
5716 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
5717 (lto_output_decl_index): Adjust.
5718 (lto_new_out_decl_state): Likewise.
5719 (lto_record_function_out_decl_state): Likewise.
5720 * lto-streamer-out.c (copy_function): Likewise.
5721
5722 2013-06-18 Richard Biener <rguenther@suse.de>
5723
5724 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
5725 * cgraphunit.c: Include cfgloop.h.
5726 (init_lowered_empty_function): Initialize the loop tree.
5727 (assemble_thunk): Insert new BBs into loops.
5728
5729 2013-06-18 Richard Biener <rguenther@suse.de>
5730
5731 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
5732 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
5733 the map from cache entry to cache index optional.
5734 (streamer_tree_cache_replace_tree): Adjust accordingly.
5735 (streamer_tree_cache_append): Likewise.
5736 (streamer_tree_cache_delete): Likewise.
5737 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
5738 streamer cache map from cache entry to cache index.
5739 * lto-streamer-out.c (create_output_block): Adjust.
5740
5741 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
5742
5743 * config/arm/arm.md (attribute "insn"): Move multiplication and
5744 division attributes to...
5745 (attribute "type"): ... here. Remove mult.
5746 (attribute "mul32"): New attribute.
5747 (attribute "mul64"): Add umaal.
5748 (*arm_mulsi3): Update attributes.
5749 (*arm_mulsi3_v6): Likewise.
5750 (*thumb_mulsi3): Likewise.
5751 (*thumb_mulsi3_v6): Likewise.
5752 (*mulsi3_compare0): Likewise.
5753 (*mulsi3_compare0_v6): Likewise.
5754 (*mulsi_compare0_scratch): Likewise.
5755 (*mulsi_compare0_scratch_v6): Likewise.
5756 (*mulsi3addsi): Likewise.
5757 (*mulsi3addsi_v6): Likewise.
5758 (*mulsi3addsi_compare0): Likewise.
5759 (*mulsi3addsi_compare0_v6): Likewise.
5760 (*mulsi3addsi_compare0_scratch): Likewise.
5761 (*mulsi3addsi_compare0_scratch_v6): Likewise.
5762 (*mulsi3subsi): Likewise.
5763 (*mulsidi3adddi): Likewise.
5764 (*mulsi3addsi_v6): Likewise.
5765 (*mulsidi3adddi_v6): Likewise.
5766 (*mulsidi3_nov6): Likewise.
5767 (*mulsidi3_v6): Likewise.
5768 (*umulsidi3_nov6): Likewise.
5769 (*umulsidi3_v6): Likewise.
5770 (*umulsidi3adddi): Likewise.
5771 (*umulsidi3adddi_v6): Likewise.
5772 (*smulsi3_highpart_nov6): Likewise.
5773 (*smulsi3_highpart_v6): Likewise.
5774 (*umulsi3_highpart_nov6): Likewise.
5775 (*umulsi3_highpart_v6): Likewise.
5776 (mulhisi3): Likewise.
5777 (*mulhisi3tb): Likewise.
5778 (*mulhisi3bt): Likewise.
5779 (*mulhisi3tt): Likewise.
5780 (maddhisi4): Likewise.
5781 (*maddhisi4tb): Likewise.
5782 (*maddhisi4tt): Likewise.
5783 (maddhidi4): Likewise.
5784 (*maddhidi4tb): Likewise.
5785 (*maddhidi4tt): Likewise.
5786 (divsi3): Likewise.
5787 (udivsi3): Likewise.
5788 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
5789 (thumb2_mulsi_short_compare0): Likewise.
5790 (thumb2_mulsi_short_compare0_scratch): Likewise.
5791 * config/arm/arm1020e.md (1020mult1): Update attribute change.
5792 (1020mult2): Likewise.
5793 (1020mult3): Likewise.
5794 (1020mult4): Likewise.
5795 (1020mult5): Likewise.
5796 (1020mult6): Likewise.
5797 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
5798 change.
5799 (cortex_a15_mult64): Likewise.
5800 (cortex_a15_sdiv): Likewise.
5801 (cortex_a15_udiv): Likewise.
5802 * config/arm/arm1026ejs.md (mult1): Update attribute change.
5803 (mult2): Likewise.
5804 (mult3): Likewise.
5805 (mult4): Likewise.
5806 (mult5): Likewise.
5807 (mult6): Likewise.
5808 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
5809 (pj4_ir_div): Likewise.
5810 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
5811 (11_mult2): Likewise.
5812 (11_mult3): Likewise.
5813 (11_mult4): Likewise.
5814 (11_mult5): Likewise.
5815 (11_mult6): Likewise.
5816 (11_mult7): Likewise.
5817 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
5818 (cortex_a8_mla): Likewise.
5819 (cortex_a8_mull): Likewise.
5820 (cortex_a8_smulwy): Likewise.
5821 (cortex_a8_smlald): Likewise.
5822 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
5823 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
5824 (cortex_r4_mul_3): Likewise.
5825 (cortex_r4_mla_4): Likewise.
5826 (cortex_r4_mla_3): Likewise.
5827 (cortex_r4_smlald): Likewise.
5828 (cortex_r4_mull): Likewise.
5829 (cortex_r4_sdiv): Likewise.
5830 (cortex_r4_udiv): Likewise.
5831 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
5832 (cortex_a7_idiv): Likewise.
5833 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
5834 (9_mult2): Likewise.
5835 (9_mult3): Likewise.
5836 (9_mult4): Likewise.
5837 (9_mult5): Likewise.
5838 (9_mult6): Likewise.
5839 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
5840 (cortex_a53_sdiv): Likewise.
5841 (cortex_a53_udiv): Likewise.
5842 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
5843 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
5844 (mp626_mult2): Likewise.
5845 (mp626_mult3): Likewise.
5846 (mp626_mult4): Likewise.
5847 * config/arm/fa526.md (526_mult1): Update attribute change.
5848 (526_mult2): Likewise.
5849 * config/arm/arm-generic.md (mult): Update attribute change.
5850 (mult_ldsched_strongarm): Likewise.
5851 (mult_ldsched): Likewise.
5852 (multi_cycle): Likewise.
5853 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
5854 * config/arm/fa606te.md (606te_mult1): Update attribute change.
5855 (606te_mult2): Likewise.
5856 (606te_mult3): Likewise.
5857 (606te_mult4): Likewise.
5858 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
5859 (cortex_a9_mac16): Likewise.
5860 (cortex_a9_multiply): Likewise.
5861 (cortex_a9_mac): Likewise.
5862 (cortex_a9_multiply_long): Likewise.
5863 * config/arm/fa626te.md (626te_mult1): Update attribute change.
5864 (626te_mult2): Likewise.
5865 (626te_mult3): Likewise.
5866 (626te_mult4): Likewise.
5867
5868 2013-06-18 Richard Biener <rguenther@suse.de>
5869
5870 PR lto/57334
5871 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
5872
5873 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5874
5875 PR target/57609
5876 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
5877 with next_active_insn.
5878
5879 2013-06-18 Alan Modra <amodra@gmail.com>
5880
5881 * config/rs6000/rs6000.h (enum data_align): New.
5882 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
5883 (DATA_ABI_ALIGNMENT): Define.
5884 (CONSTANT_ALIGNMENT): Correct comment.
5885 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
5886 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
5887
5888 2013-06-17 David Malcolm <dmalcolm@redhat.com>
5889
5890 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
5891 ATTRIBUTE_UNUSED marking.
5892
5893 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
5894
5895 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
5896 alternative and update.
5897 (aarch64_dup_lanedi): Delete.
5898 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
5899 * config/aarch64/aarch64-simd-builtins.def: Update.
5900
5901 2013-06-17 Richard Biener <rguenther@suse.de>
5902
5903 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
5904 (lto_input_scc): Declare.
5905 (lto_input_tree_1): Likewise.
5906 (struct lto_stats_d): Add num_tree_bodies_output and
5907 num_pickle_refs_output.
5908 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
5909 (lto_read_tree_1): Split out from ...
5910 (lto_read_tree): ... this.
5911 (lto_input_scc): New function.
5912 (lto_input_tree_1): Split out from ...
5913 (lto_input_tree): ... this. Handle LTO_tree_scc.
5914 (lto_data_in_create): Create the streamer cache without hashes.
5915 * lto-streamer-out.c (create_output_block): Create the streamer
5916 cache with hashes when not doing WPA.
5917 (lto_write_tree_1): Split out from ...
5918 (lto_write_tree): ... this.
5919 (get_symbol_initial_value): New function.
5920 (lto_output_tree_1): Split out from ...
5921 (lto_output_tree): ... this. Write trees as series of SCCs
5922 using a DFS walk via DFS_write_tree.
5923 (struct sccs, struct scc_entry): New types.
5924 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
5925 (DFS_write_tree_body): New function.
5926 (DFS_write_tree): Likewise.
5927 (hash_tree): Likewise.
5928 (scc_entry_compare): Likewise.
5929 (hash_scc): Likewise.
5930 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
5931 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
5932 TREE_CHAIN as regular reference.
5933 (streamer_read_integer_cst): Remove.
5934 (streamer_get_pickled_tree): Adjust.
5935 * tree-streamer-out.c (streamer_write_chain): Disable streaming
5936 of DECL_EXTERNALs in BLOCK_VARS for now.
5937 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
5938 reference.
5939 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
5940 Add hash value argument and record that if hashes are recorded
5941 in the cache.
5942 (streamer_tree_cache_insert_1): Adjust.
5943 (streamer_tree_cache_insert): Likewise.
5944 (streamer_tree_cache_insert_at): Rename to ...
5945 (streamer_tree_cache_replace_tree): ... this and adjust.
5946 (streamer_tree_cache_append): Adjust.
5947 (record_common_node): Likewise.
5948 (streamer_tree_cache_create): Add argument whether to
5949 record hash values together with trees.
5950 (streamer_tree_cache_delete): Adjust.
5951 * tree-streamer.h (struct streamer_tree_cache_d): Add
5952 vector of hashes.
5953 (streamer_read_integer_cst): Remove.
5954 (streamer_tree_cache_insert): Adjust.
5955 (streamer_tree_cache_append): Likewise.
5956 (streamer_tree_cache_insert_at): Rename to ...
5957 (streamer_tree_cache_replace_tree): ... this and adjust.
5958 (streamer_tree_cache_create): Add argument whether to record hashes.
5959 (streamer_tree_cache_get): Rename to ...
5960 (streamer_tree_cache_get_tree): ... this.
5961 (streamer_tree_cache_get_hash): New function.
5962 * tree.c (cache_integer_cst): New function.
5963 * tree.h (cache_integer_cst): Declare.
5964 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
5965 * lto-symtab.c (lto_varpool_replace_node): Only release
5966 DECL_INITIAL of non-prevailing decls.
5967 * varpool.c (varpool_remove_initializer): Do not release
5968 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
5969
5970 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
5971
5972 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
5973 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
5974 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
5975 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
5976 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
5977 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
5978 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
5979 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
5980 instead of TARGET_64BIT.
5981 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
5982 Require ISA_HAS_<D>DIV.
5983
5984 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
5985
5986 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
5987 (mips*-*-linux*): Move default with_llsc setting to where other
5988 defaults are set.
5989 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
5990 with_arch block.
5991 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
5992 Likewise. Remove default with_tune setting. Move default float
5993 setting to its own block. Handle with_llsc in the same block as above.
5994
5995 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
5996
5997 PR rtl-optimization/57425
5998 PR rtl-optimization/57569
5999 * alias.c (write_dependence_p): Add new parameters mem_mode,
6000 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6001 Changed all callers.
6002 (canon_anti_dependence): New function.
6003 * cse.c (check_dependence): Use canon_anti_dependence.
6004 * cselib.c (cselib_invalidate_mem): Likewise.
6005 * rtl.h (canon_anti_dependence): Declare.
6006
6007 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6008
6009 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6010 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6011 ".set mips3" for 64-bit targets.
6012
6013 2013-06-15 Dehao Chen <dehao@google.com>
6014
6015 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6016 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6017 (gimple_check_call_args): Likewise.
6018 * value-prof.c (check_ic_target): Likewise.
6019 * ipa-inline.c (early_inliner): Likewise.
6020 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6021 * cgraph.c (cgraph_create_edge_1): Likewise.
6022 (cgraph_make_edge_direct): Likewise.
6023
6024 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6025
6026 PR target/57615
6027 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6028 rs6000_output_move_128bit to handle emitting quad memory
6029 operations. Set attribute length to 8 bytes.
6030
6031 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6032
6033 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6034 New pattern.
6035 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6036 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6037 (aarch64_<su>mlsl<mode>): Likewise.
6038
6039 2013-06-14 Mike Stump <mikestump@comcast.net>
6040
6041 * Makefile.in (TARGET_H): Add insn-codes.h.
6042
6043 2013-06-14 Alan Modra <amodra@gmail.com>
6044
6045 PR middle-end/57134
6046 PR middle-end/57586
6047 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6048 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6049 bitfield expansion when EXPAND_MEMORY.
6050 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6051
6052 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6053
6054 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6055 test for clearing quad memory on 32-bit later.
6056
6057 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6058
6059 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6060 (fold_negate_expr): Likewise.
6061 (fold_real_zero_addition_p): Handle vectors.
6062 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6063
6064 2013-06-14 Alan Modra <amodra@gmail.com>
6065
6066 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6067
6068 2013-06-13 Jan Hubicka <jh@suse.cz>
6069
6070 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6071 Local comdats are not externally visible.
6072 * symtab.c (dump_symtab_base): Dump externally visible.
6073 (verify_symtab_base): Verify back links in the symtab hash.
6074
6075 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6076
6077 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6078 CONVERT_EXPR as equal nodes.
6079
6080 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6081
6082 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6083
6084 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6085
6086 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6087 Generalize to complex and vector.
6088 * tree.c (build_all_ones_cst): New function.
6089 * tree.h (build_all_ones_cst): Declare it.
6090
6091 2013-06-13 Alan Modra <amodra@gmail.com>
6092
6093 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
6094 * config/rs6000/rs6000.md (signbittf2): New insn.
6095 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
6096 (abstf2_internal, cmptf_internal2): Likewise.
6097 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
6098
6099 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6100 Pat Haugen <pthaugen@us.ibm.com>
6101 Peter Bergner <bergner@vnet.ibm.com>
6102
6103 * config/rs6000/rs6000.c (emit_load_locked): Add support for
6104 power8 byte, half-word, and quad-word atomic instructions.
6105 (emit_store_conditional): Likewise.
6106 (rs6000_expand_atomic_compare_and_swap): Likewise.
6107 (rs6000_expand_atomic_op): Likewise.
6108
6109 * config/rs6000/sync.md (larx): Add new modes for power8.
6110 (stcx): Likewise.
6111 (AINT): New mode iterator to include TImode as well as normal
6112 integer modes on power8.
6113 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
6114 that VSX registers are not considered. Use AINT mode iterator
6115 instead of INT1 to allow inclusion of quad word atomic operations
6116 on power8.
6117 (load_locked<mode>): Likewise.
6118 (store_conditional<mode>): Likewise.
6119 (atomic_compare_and_swap<mode>): Likewise.
6120 (atomic_exchange<mode>): Likewise.
6121 (atomic_nand<mode>): Likewise.
6122 (atomic_fetch_<fetchop_name><mode>): Likewise.
6123 (atomic_nand_fetch<mode>): Likewise.
6124 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
6125 each type.
6126 (ATOMIC): On power8, add QImode, HImode modes.
6127 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
6128 modes that promote to SImode.
6129 (load_lockedti): Convert TImode arguments to PTImode, so that we
6130 get a guaranteed even/odd register pair.
6131 (load_lockedpti): Likewise.
6132 (store_conditionalti): Likewise.
6133 (store_conditionalpti): Likewise.
6134
6135 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
6136 atomic load/store instructions.
6137 (HSI): Likewise.
6138
6139 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
6140
6141 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
6142 loads.
6143 (insn_count): New attribute, with most cases extracted from...
6144 (length): ...here. Redefine most cases in terms of insn_count.
6145 (single_insn): Delete.
6146 (can_delay): Use insn_count to check for single instructions.
6147 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
6148 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
6149 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
6150 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
6151 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
6152 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
6153 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
6154 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
6155 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
6156 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
6157 rather than "length".
6158 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
6159 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
6160 Use "insn_count" rather than "length".
6161 * config/mips/mips-dsp.md
6162 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
6163 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
6164 length attributes.
6165
6166 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
6167
6168 PR tree-optimization/57361
6169 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
6170
6171 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
6172
6173 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
6174 to split.
6175 (aarch64_simd_combine<mode>): New instruction expansion.
6176 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
6177 function prototype.
6178 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
6179 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
6180
6181 2013-06-12 Jan Hubicka <jh@suse.cz>
6182
6183 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
6184 decl has when in streaming stage.
6185 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
6186 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
6187
6188 2013-06-12 Roland Stigge <stigge@antcom.de>
6189
6190 PR target/57578
6191 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
6192
6193 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6194
6195 PR tree-optimization/57537
6196 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
6197 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
6198
6199 2013-06-12 Richard Biener <rguenther@suse.de>
6200
6201 * data-streamer.h (streamer_write_char_stream): CSE
6202 obs->current_pointer.
6203 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
6204 streamer_write_char_stream manually and optimize the resulting loop.
6205 (streamer_write_hwi_stream): Likewise.
6206
6207 2013-06-12 Jan Hubicka <jh@suse.cz>
6208
6209 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
6210 * cgraph.h (varpool_create_empty_node): Declare.
6211 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
6212 duplicated nodes.
6213 * symtab.c (symtab_unregister_node): Be lax about missin entries
6214 in node hash.
6215 (symtab_get_node): Update comment.
6216 * varpool.c (varpool_create_empty_node): Break out from ...
6217 (varpool_node_for_decl): ... here.
6218 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
6219
6220 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
6221
6222 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
6223 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
6224 part. Use straight-line flow at the end.
6225 <COMPONENT_REF>: Remove superfluous else.
6226 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
6227
6228 2013-06-12 Jakub Jelinek <jakub@redhat.com>
6229
6230 PR target/56564
6231 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
6232 target hook even for !TREE_PUBLIC decls. If no resolution info
6233 is available, return false for common and external decls.
6234
6235 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
6236
6237 * config/rl78/constraints.md (U): New constraint.
6238 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
6239 valloc attribute.
6240
6241 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
6242
6243 PR target/57589
6244 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
6245 to allow returning address to AT_PLATFORM name.
6246
6247 2013-06-11 Jan Hubicka <jh@suse.cz>
6248
6249 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
6250 * cgraph.h (symtab_node_base): Add weakref flag.
6251 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
6252 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
6253 (output_weakrefs): Use weakref flag.
6254 * fold-const.c (simple_operand_p): Handle WEAK.
6255 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
6256 * ipa.c (varpool_externally_visible_p): Drop weakref.
6257 (function_and_variable_visibility): Update comment; fix weakref
6258 sanity checks; do not clear DECL_WEAK on them.
6259 * lto-cgraph.c (lto_output_node): update.
6260 (lto_output_varpool_node): Update.
6261 (input_overwrite_node): Update.
6262 (input_node): Update.
6263 (input_varpool_node): Update.
6264 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
6265 (lto_symtab_merge_symbols): Add sanity check.
6266 (lto_symtab_prevailing_decl): Do not special case weakrefs.
6267 * passes.c (rest_of_decl_compilation): Set static flag, too.
6268 * symtab.c (dump_symtab_base): Dump weakref.
6269 (verify_symtab_base): Sanity check weakrefs.
6270 (symtab_make_decl_local): Remove duplicated code.
6271 (symtab_alias_ultimate_target): Simplify.
6272 * varpool.c (varpool_create_variable_alias): Set weakref flag.
6273
6274 2013-06-11 DJ Delorie <dj@redhat.com>
6275
6276 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
6277 (rl78_unwind_word_mode): New.
6278
6279 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6280
6281 * final.c (debug_prefix_maps): Make static.
6282
6283 2013-06-11 David Malcolm <dmalcolm@redhat.com>
6284
6285 * function.c (initial_trampoline): Remove stray copy.
6286
6287 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
6288
6289 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
6290
6291 2013-06-11 Martin Jambor <mjambor@suse.cz>
6292
6293 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
6294 within bounds at the beginning of the function.
6295
6296 2013-06-11 Alan Modra <amodra@gmail.com>
6297
6298 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
6299 reporting.
6300 (get_named_section): Don't NULL !DECL_P decl.
6301
6302 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
6303
6304 * doc/invoke.texi (core-avx2): Document.
6305 (slm): Likewise.
6306 (atom): Updated with MOVBE.
6307
6308 2013-06-11 Richard Biener <rguenther@suse.de>
6309
6310 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
6311
6312 2013-06-11 Anton Blanchard <anton@samba.org>
6313
6314 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
6315 correct shift value in little-endian mode.
6316
6317 2013-06-11 Jakub Jelinek <jakub@redhat.com>
6318
6319 PR target/56564
6320 * varasm.c (get_variable_align): Move #endif to the right place.
6321
6322 2013-06-10 Cary Coutant <ccoutant@google.com>
6323
6324 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
6325 for hash so that hash table traversal order is deterministic.
6326
6327 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
6328 Pat Haugen <pthaugen@us.ibm.com>
6329 Peter Bergner <bergner@vnet.ibm.com>
6330
6331 * config/rs6000/vector.md (GPR move splitter): Do not split moves
6332 of vectors in GPRS if they are direct moves or quad word load or
6333 store moves.
6334
6335 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
6336 declaration.
6337 (direct_move_p): Likewise.
6338 (quad_load_store_p): Likewise.
6339
6340 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
6341 classes into bins based on the physical register type.
6342 (reg_class_to_reg_type): Likewise.
6343 (IS_STD_REG_TYPE): Likewise.
6344 (IS_FP_VECT_REG_TYPE): Likewise.
6345 (reload_fpr_gpr): Arrays to determine what insn to use if we can
6346 use direct move instructions.
6347 (reload_gpr_vsx): Likewise.
6348 (reload_vsx_gpr): Likewise.
6349 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
6350 information that is a simplification of register classes. Also
6351 precalculate direct move reload helpers.
6352 (direct_move_p): New function to return true if the operation can
6353 be done as a direct move instruciton.
6354 (quad_load_store_p): New function to return true if the operation
6355 is a quad memory operation.
6356 (rs6000_legitimize_address): If quad memory, only allow register
6357 indirect for TImode addresses.
6358 (rs6000_legitimate_address_p): Likewise.
6359 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
6360 (rs6000_reload_register_type): Likewise.
6361 (register_to_reg_type): Return register type.
6362 (rs6000_secondary_reload_simple_move): New helper function for
6363 secondary reload and secondary memory needed to identify anything
6364 that is a simple move, and does not need reloading.
6365 (rs6000_secondary_reload_direct_move): New helper function for
6366 secondary reload to identify cases that can be done with several
6367 instructions via the direct move instructions.
6368 (rs6000_secondary_reload_move): New helper function for secondary
6369 reload to identify moves between register types that can be done.
6370 (rs6000_secondary_reload): Add support for quad memory operations
6371 and for direct move.
6372 (rs6000_secondary_memory_needed): Likewise.
6373 (rs6000_debug_secondary_memory_needed): Change argument names.
6374 (rs6000_output_move_128bit): New function to return the move to
6375 use for 128-bit moves, including knowing about the various
6376 limitations of quad memory operations.
6377
6378 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
6379 memory operations. call rs6000_output_move_128bit for the actual
6380 instruciton(s) to generate.
6381 (vsx_movti_64bit): Likewise.
6382
6383 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
6384 (UNSPEC_P8V_MTVSRWZ): Likewise.
6385 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
6386 (UNSPEC_P8V_MTVSRD): Likewise.
6387 (UNSPEC_P8V_XXPERMDI): Likewise.
6388 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
6389 (UNSPEC_FUSION_GPR): Likewise.
6390 (FMOVE128_GPR): New iterator for direct move.
6391 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
6392 (f32_sv): Likewise.
6393 (f32_dm): Likewise.
6394 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
6395 loads and direct move instructions.
6396 (zero_extendsidi2_lfiwzx): Likewise.
6397 (extendsidi2_lfiwax): Likewise.
6398 (extendsidi2_nocell): Likewise.
6399 (floatsi<mode>2_lfiwax): Likewise.
6400 (lfiwax): Likewise.
6401 (floatunssi<mode>2_lfiwzx): Likewise.
6402 (lfiwzx): Likewise.
6403 (fix_trunc<mode>_stfiwx): Likewise.
6404 (fixuns_trunc<mode>_stfiwx): Likewise.
6405 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
6406 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
6407 (parity<mode>2_cmpb): Set length/type attr.
6408 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
6409 for 'mr.' to fast_compare.
6410 (bpermd_<mode>): Change type attr to popcnt.
6411 (p8_fmrgow_<mode>): New insns for power8 direct move support.
6412 (p8_mtvsrwz_1): Likewise.
6413 (p8_mtvsrwz_2): Likewise.
6414 (reload_fpr_from_gpr<mode>): Likewise.
6415 (p8_mtvsrd_1): Likewise.
6416 (p8_mtvsrd_2): Likewise.
6417 (p8_xxpermdi_<mode>): Likewise.
6418 (reload_vsx_from_gpr<mode>): Likewise.
6419 (reload_vsx_from_gprsf): Likewise.
6420 (p8_mfvsrd_3_<mode>): LIkewise.
6421 (reload_gpr_from_vsx<mode>): Likewise.
6422 (reload_gpr_from_vsxsf): Likewise.
6423 (p8_mfvsrd_4_disf): Likewise.
6424 (multi-word GPR splits): Do not split direct moves or quad memory
6425 operations.
6426
6427 2013-06-10 David Malcolm <dmalcolm@redhat.com>
6428
6429 * tree-into-ssa.c (interesting_blocks): Make static.
6430
6431 2013-06-10 Jakub Jelinek <jakub@redhat.com>
6432
6433 PR target/56564
6434 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
6435 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
6436 Use DATA_ABI_ALIGNMENT for that case instead if defined.
6437 (get_variable_align): New function.
6438 (get_variable_section, emit_bss, emit_common,
6439 assemble_variable_contents, place_block_symbol): Use
6440 get_variable_align instead of DECL_ALIGN.
6441 (assemble_noswitch_variable): Add align argument, use it
6442 instead of DECL_ALIGN.
6443 (assemble_variable): Adjust caller. Use get_variable_align
6444 instead of DECL_ALIGN.
6445 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
6446 caller.
6447 (DATA_ABI_ALIGNMENT): Define.
6448 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
6449 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
6450 opt is false, only return the psABI mandated alignment increase.
6451 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
6452 (DATA_ABI_ALIGNMENT): ... this.
6453 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
6454 (DATA_ABI_ALIGNMENT): ... this.
6455 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
6456 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
6457 (DATA_ABI_ALIGNMENT): ... this.
6458 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
6459 * doc/tm.texi: Regenerated.
6460
6461 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
6462
6463 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
6464 cmp_code to construct REG_EQUAL note.
6465
6466 2013-06-09 Jakub Jelinek <jakub@redhat.com>
6467
6468 PR target/57568
6469 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
6470 that operands[2] doesn't overlap with operands[0].
6471
6472 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
6473 Jan Hubicka <jh@suse.cz>
6474
6475 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
6476 hack to mark symbols as used.
6477
6478 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
6479
6480 PR rtl-optimization/57559
6481 * lra-constraints.c (process_alt_operands): Don't discourage
6482 memory with known offset for offsetable memory constraint.
6483 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
6484
6485 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
6486
6487 * varasm.c (struct oc_local_state): Reorder fields.
6488 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
6489 and adjust accordingly.
6490 (output_constructor): Reorder initialization code and adjust call to
6491 output_constructor_bitfield.
6492
6493 2013-06-07 Jan Hubicka <jh@suse.cz>
6494
6495 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
6496
6497 2013-06-07 David Malcolm <dmalcolm@redhat.com>
6498
6499 * tree-object-size.c (unknown): Make const.
6500
6501 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6502
6503 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
6504 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
6505 for last alternative in the cpu_facility attribute.
6506
6507 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6508
6509 PR target/56315
6510 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
6511 (xordi3): Change operand 2 constraint to arm_xordi_operand.
6512 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
6513 * config/arm/constraints.md (Dg): New constraint.
6514 * config/arm/neon.md (xordi3_neon): Remove.
6515 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
6516 * config/arm/predicates.md (arm_xordi_operand): New predicate.
6517
6518 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6519
6520 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
6521 Clean up alternatives.
6522
6523 2013-06-07 Alan Modra <amodra@gmail.com>
6524
6525 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
6526 va_list_gpr_size.
6527
6528 2013-06-07 Alan Modra <amodra@gmail.com>
6529
6530 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
6531
6532 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6533
6534 * config/arm/constraints.md (Df): New constraint.
6535 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
6536 Correct length attribute for last two alternatives.
6537
6538 2013-06-07 Alan Modra <amodra@gmail.com>
6539
6540 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6541 override user -mfp-in-toc.
6542 (offsettable_ok_by_alignment): Consider just the current access
6543 rather than the whole object, unless BLKmode. Handle
6544 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
6545 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
6546 for -mcmodel=medium.
6547 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
6548 override user -mfp-in-toc or -msum-in-toc. Default to
6549 -mno-fp-in-toc for -mcmodel=medium.
6550
6551 2013-06-06 DJ Delorie <dj@redhat.com>
6552
6553 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
6554 TARGET_VALID_POINTER_MODE.
6555
6556 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
6557 Pat Haugen <pthaugen@us.ibm.com>
6558 Peter Bergner <bergner@vnet.ibm.com>
6559
6560 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6561 Document new power8 builtins.
6562
6563 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
6564 condition code register, to allow 128-bit logical operations to be
6565 done in the VSX or GPR registers.
6566 (nor<mode>3): Use the canonical form for nor.
6567 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
6568 vclz*, and vpopcnt* vector instructions.
6569 (nand<mode>3): Likewise.
6570 (orc<mode>3): Likewise.
6571 (clz<mode>2): LIkewise.
6572 (popcount<mode>2): Likewise.
6573
6574 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
6575 that only the GPRs are recognized.
6576
6577 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6578 support for new power8 builtins.
6579
6580 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
6581 builtin functions.
6582 (xscvdpspn): Likewise.
6583 (vclz): Likewise.
6584 (vclzb): Likewise.
6585 (vclzh): Likewise.
6586 (vclzw): Likewise.
6587 (vclzd): Likewise.
6588 (vpopcnt): Likewise.
6589 (vpopcntb): Likewise.
6590 (vpopcnth): Likewise.
6591 (vpopcntw): Likewise.
6592 (vpopcntd): Likewise.
6593 (vgbbd): Likewise.
6594 (vmrgew): Likewise.
6595 (vmrgow): Likewise.
6596 (eqv): Likewise.
6597 (eqv_v16qi3): Likewise.
6598 (eqv_v8hi3): Likewise.
6599 (eqv_v4si3): Likewise.
6600 (eqv_v2di3): Likewise.
6601 (eqv_v4sf3): Likewise.
6602 (eqv_v2df3): Likewise.
6603 (nand): Likewise.
6604 (nand_v16qi3): Likewise.
6605 (nand_v8hi3): Likewise.
6606 (nand_v4si3): Likewise.
6607 (nand_v2di3): Likewise.
6608 (nand_v4sf3): Likewise.
6609 (nand_v2df3): Likewise.
6610 (orc): Likewise.
6611 (orc_v16qi3): Likewise.
6612 (orc_v8hi3): Likewise.
6613 (orc_v4si3): Likewise.
6614 (orc_v2di3): Likewise.
6615 (orc_v4sf3): Likewise.
6616 (orc_v2df3): Likewise.
6617
6618 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
6619 allow power8 quad mode in 64-bit.
6620 (rs6000_builtin_vectorized_function): Add support to vectorize
6621 ISA 2.07 count leading zeros, population count builtins.
6622 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
6623 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
6624 (builtin_function_type): Add vgbbd builtin function which takes an
6625 unsigned argument.
6626 (altivec_expand_vec_perm_const): Add support for new power8 merge
6627 instructions.
6628
6629 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
6630 that does not include TImdoe for use with 32-bit.
6631 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
6632 instructions.
6633 (UNSPEC_VSX_CVDPSPN): Likewise.
6634 (vsx_xscvdpspn): Likewise.
6635 (vsx_xscvspdpn): Likewise.
6636 (vsx_xscvdpspn_scalar): Likewise.
6637 (vsx_xscvspdpn_directmove): Likewise.
6638 (vsx_and<mode>3): Split logical operations into 32-bit and
6639 64-bit. Add support to do logical operations on TImode as well as
6640 VSX vector types. Allow logical operations to be done in either
6641 VSX registers or in general purpose registers in 64-bit mode. Add
6642 splitters if GPRs were used. For AND, add clobber of CCmode to
6643 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
6644 encoding.
6645 (vsx_and<mode>3_32bit): Likewise.
6646 (vsx_and<mode>3_64bit): Likewise.
6647 (vsx_ior<mode>3): Likewise.
6648 (vsx_ior<mode>3_32bit): Likewise.
6649 (vsx_ior<mode>3_64bit): Likewise.
6650 (vsx_xor<mode>3): Likewise.
6651 (vsx_xor<mode>3_32bit): Likewise.
6652 (vsx_xor<mode>3_64bit): Likewise.
6653 (vsx_one_cmpl<mode>2): Likewise.
6654 (vsx_one_cmpl<mode>2_32bit): Likewise.
6655 (vsx_one_cmpl<mode>2_64bit): Likewise.
6656 (vsx_nor<mode>3): Likewise.
6657 (vsx_nor<mode>3_32bit): Likewise.
6658 (vsx_nor<mode>3_64bit): Likewise.
6659 (vsx_andc<mode>3): Likewise.
6660 (vsx_andc<mode>3_32bit): Likewise.
6661 (vsx_andc<mode>3_64bit): Likewise.
6662 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
6663 and xxlorc instructions.
6664 (vsx_eqv<mode>3_64bit): Likewise.
6665 (vsx_nand<mode>3_32bit): Likewise.
6666 (vsx_nand<mode>3_64bit): Likewise.
6667 (vsx_orc<mode>3_32bit): Likewise.
6668 (vsx_orc<mode>3_64bit): Likewise.
6669
6670 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
6671
6672 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
6673 instruction.
6674 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
6675 (p8_vmrgow): Likewise.
6676 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
6677 GPRs to be split under VSX.
6678 (p8v_clz<mode>2): Add power8 count leading zero support.
6679 (p8v_popcount<mode>2): Add power8 population count support.
6680 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
6681 support.
6682
6683 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
6684 instruction.
6685
6686 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
6687 builtin functions.
6688 (vec_nand): Likewise.
6689 (vec_vclz): Likewise.
6690 (vec_vclzb): Likewise.
6691 (vec_vclzd): Likewise.
6692 (vec_vclzh): Likewise.
6693 (vec_vclzw): Likewise.
6694 (vec_vgbbd): Likewise.
6695 (vec_vmrgew): Likewise.
6696 (vec_vmrgow): Likewise.
6697 (vec_vpopcnt): Likewise.
6698 (vec_vpopcntb): Likewise.
6699 (vec_vpopcntd): Likewise.
6700 (vec_vpopcnth): Likewise.
6701 (vec_vpopcntw): Likewise.
6702
6703 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6704
6705 PR rtl-optimization/57468
6706 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
6707 spilled pseudos.
6708
6709 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6710
6711 PR rtl-optimization/57459
6712 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
6713 type when setting live regs.
6714
6715 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
6716
6717 * config/s390/s390.opt (mlra): New option.
6718 * config/s390/s390.c (s390_decompose_address): Check displacement
6719 for all registers for LRA.
6720 (s390_secondary_reload): Don't used secondary reloads for LRA.
6721 (s390_lra_p): New function.
6722 (TARGET_LRA_P): Define.
6723 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
6724 of attribute cpu_facility to zarch for the last alternative.
6725 (*cmpmem_short): Ditto.
6726
6727 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
6728
6729 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
6730 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
6731 (arm_expand_prologue): Likewise.
6732
6733 2013-06-06 Teresa Johnson <tejohnson@google.com>
6734
6735 PR c++/53743
6736 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
6737 as this is now done by redirect_edge_and_branch_force.
6738 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
6739 barriers, and fix interaction with splitting.
6740 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
6741 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
6742 reflect changes made in the routine.
6743 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
6744 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
6745 since this is called in cfglayout mode, and replace partition fixup
6746 with assert as that is now done by force_nonfallthru_and_redirect.
6747 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
6748 already be marked with region crossing note.
6749 (insert_section_boundary_note): Make non-static, gate on flag
6750 has_bb_partition, rewrite to also check for multiple partitions.
6751 (rest_of_handle_reorder_blocks): Remove call to
6752 insert_section_boundary_note, now done later during free_cfg.
6753 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
6754 * bb-reorder.h (insert_section_boundary_note): Declare.
6755 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
6756 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
6757 invoke insert_section_boundary_note.
6758 (try_redirect_by_replacing_jump): Remove unnecessary
6759 check for region crossing note.
6760 (fixup_partition_crossing): New function.
6761 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
6762 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
6763 in non-cfglayout mode.
6764 (force_nonfallthru_and_redirect): Fixup partition boundaries,
6765 remove old code that tried to do this. Emit barrier correctly
6766 when we are in cfglayout mode.
6767 (last_bb_in_partition): New function.
6768 (rtl_split_edge): Correctly fixup partition boundaries.
6769 (commit_one_edge_insertion): Remove old code that tried to
6770 fixup region crossing edge since this is now handled in
6771 split_block, and set up insertion point correctly since
6772 block may now end in a jump.
6773 (verify_hot_cold_block_grouping): Guard against checking when not in
6774 linearized RTL mode.
6775 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
6776 notes.
6777 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
6778 rtl_verify_flow_info, so not called in cfglayout mode.
6779 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
6780 (fixup_reorder_chain): Remove old code that attempted to fixup region
6781 crossing note as this is now handled in force_nonfallthru_and_redirect.
6782 (duplicate_insn_chain): Don't duplicate switch section notes.
6783 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
6784 note.
6785 * basic-block.h (emit_barrier_after_bb): Declare.
6786
6787 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6788
6789 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
6790 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
6791 arm_usatsihi): Adjust alternatives for arm_restrict_it.
6792
6793 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6794
6795 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
6796 where appropriate.
6797 * config/arm/ldmstm.md: Regenerate.
6798
6799 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6800
6801 * config/arm/sync.md (atomic_loaddi_1):
6802 Disable predication for arm_restrict_it.
6803 (arm_load_exclusive<mode>): Likewise.
6804 (arm_load_exclusivesi): Likewise.
6805 (arm_load_exclusivedi): Likewise.
6806 (arm_load_acquire_exclusive<mode>): Likewise.
6807 (arm_load_acquire_exclusivesi): Likewise.
6808 (arm_load_acquire_exclusivedi): Likewise.
6809 (arm_store_exclusive<mode>): Likewise.
6810 (arm_store_exclusive<mode>): Likewise.
6811 (arm_store_release_exclusivedi): Likewise.
6812 (arm_store_release_exclusive<mode>): Likewise.
6813
6814 2013-06-06 Richard Biener <rguenther@suse.de>
6815
6816 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
6817 after LTO_null.
6818 (lto_tag_is_tree_code_p): Adjust.
6819 (lto_tag_is_gimple_code_p): Likewise.
6820 (lto_gimple_code_to_tag): Likewise.
6821 (lto_tag_to_gimple_code): Likewise.
6822 (lto_tree_code_to_tag): Likewise.
6823 (lto_tag_to_tree_code): Likewise.
6824 * data-streamer.h (streamer_write_hwi_in_range): Use
6825 uhwi streaming to stream the normalized range.
6826 (streamer_read_hwi_in_range): Likewise.
6827
6828 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6829
6830 * config/arm/arm.md (enabled_for_depr_it): New attribute.
6831 (predicable_short_it): Likewise.
6832 (predicated): Likewise.
6833 (enabled): Handle above.
6834 (define_cond_exec): Set predicated attribute to yes.
6835
6836 2013-06-05 Mike Stump <mikestump@comcast.net>
6837
6838 * gdbinit.in (__FUNCTION__): Add.
6839
6840 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
6841
6842 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
6843 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
6844
6845 2013-06-05 Jan Hubicka <jh@suse.cz>
6846
6847 * varasm.c (mark_decl_referenced): Revert the removal until targets
6848 are fixed.
6849
6850 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
6851
6852 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
6853 instead of mark_decl_referenced.
6854
6855 2013-06-05 Jan Hubicka <jh@suse.cz>
6856
6857 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
6858 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
6859 and symtab_used_from_object_file_p.
6860 (cgraph_make_node_local_1): Clear forced_by_abi.
6861 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
6862 * cgraph.h (symtab_node_base): Add forced_by_abi.
6863 (decide_is_variable_needed): Remove.
6864 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
6865 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
6866 (decide_is_symbol_needed): ... this one; handle symbols in general;
6867 always analyze virtuals; honnor forced_by_abi.
6868 (cgraph_finalize_function): Update.
6869 (varpool_finalize_decl): Update.
6870 (symbol_defined_and_needed): Remove.
6871 (analyze_functions): Update.
6872 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
6873 output_refs, input_overwrite_node): Handle forced_by_abi.
6874 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
6875 (address_taken_from_non_vtable_p): ... this one.
6876 (comdat_can_be_unshared_p_1): New function.
6877 (cgraph_comdat_can_be_unshared_p): Rename to ...
6878 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
6879 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
6880 (function_and_variable_visibility): Clear forced_by_abi as needed.
6881 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
6882 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
6883 * symtab.c (dump_symtab_base): Dump forced_by_abi.
6884 * varpool.c (decide_is_variable_needed): Remove.
6885
6886 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6887
6888 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
6889 (arm_option_override): Override arm_restrict_it where appropriate.
6890 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
6891 * config/arm/arm.opt (mrestrict-it): New command-line option.
6892 * doc/invoke.texi: Document -mrestrict-it.
6893
6894 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6895
6896 * tsan.c (tsan_atomic_table): Make const.
6897
6898 2013-06-05 Richard Biener <rguenther@suse.de>
6899
6900 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
6901 index associated with the tree we are supposed to replace.
6902 * tree-streamer-out.c (pack_ts_base_value_fields): Output
6903 TREE_ASM_WRITTEN as zero for everything but SSA names.
6904
6905 2013-06-05 David Malcolm <dmalcolm@redhat.com>
6906
6907 * tree-ssa-structalias.c (call_stmt_vars): Make static.
6908
6909 2013-06-04 Jan Hubicka <jh@suse.cz>
6910
6911 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
6912 (input_node, input_varpool_node): Handle correctly external same
6913 body aliases.
6914 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
6915 nodes at ltrans stage.
6916
6917 2013-06-04 Jan Hubicka <jh@suse.cz>
6918
6919 * ipa-inline.c (update_caller_keys): Fix availability test.
6920 (update_callee_keys): Likewise.
6921 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
6922 to follow ELF standard.
6923
6924 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
6925
6926 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
6927 (mips64r5900el-*-elf*): New configurations.
6928 * config/mips/mips-cpus.def (r5900): New processor.
6929 * config/mips/mips-tables.opt: Regenerate.
6930 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
6931 (mips_issue_rate): Handle PROCESSOR_R5900.
6932 (mips_reorg_process_insns): Force reorder mode for the R5900.
6933 * config/mips/mips.h (TARGET_MIPS5900): Define.
6934 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
6935 TARGET_MIPS5900.
6936 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
6937 TARGET_MIPS5900.
6938 * config/mips/mips.md (processor): Add r5900.
6939 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
6940
6941 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6942
6943 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
6944 into function to generate MOVI instruction.
6945 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
6946 (aarch64_preferred_simd_mode): Turn into wrapper.
6947 (aarch64_output_scalar_simd_mov_immediate): New function.
6948 * config/aarch64/aarch64-protos.h: Add prototype for above.
6949
6950 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6951
6952 * config/aarch64/aarch64.c (simd_immediate_info): Remove
6953 element_char member.
6954 (sizetochar): Return signed char.
6955 (aarch64_simd_valid_immediate): Remove elchar and other
6956 unnecessary variables.
6957 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
6958 Calculate element_char as required.
6959 * config/aarch64/aarch64-protos.h: Update and move prototype
6960 for aarch64_output_simd_mov_immediate.
6961 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
6962 Update arguments.
6963
6964 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6965
6966 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
6967 information completed by aarch64_simd_valid_immediate.
6968 (aarch64_legitimate_constant_p): Update arguments.
6969 (aarch64_simd_valid_immediate): Work with struct rather than many
6970 pointers.
6971 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
6972 (aarch64_simd_make_constant): Update arguments.
6973 (aarch64_output_simd_mov_immediate): Work with struct rather than
6974 many pointers. Output immediate directly rather than as operand.
6975 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6976 Update prototype.
6977 * config/aarch64/constraints.md (Dn): Update arguments.
6978
6979 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6980
6981 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
6982 longer static.
6983 (aarch64_simd_immediate_valid_for_move): Remove.
6984 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
6985 (aarch64_simd_make_constant): Update call.
6986 (aarch64_output_simd_mov_immediate): Update call.
6987 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
6988 Add prototype.
6989 * config/aarch64/constraints.md (Dn): Update call.
6990
6991 2013-06-04 Ian Bolton <ian.bolton@arm.com>
6992
6993 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
6994 return type to bool for prototype.
6995 (aarch64_legitimate_constant_p): Check for true instead of not -1.
6996 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
6997 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
6998
6999 2013-06-04 Catherine Moore <clm@codesourcery.com>
7000
7001 * config/mips/mips.opt (meva): New.
7002 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7003 (ASM_SPEC): Handle -meva.
7004 * doc/invoke.texi (meva): Document.
7005
7006 2013-06-04 Alan Modra <amodra@gmail.com>
7007
7008 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7009 constant output.
7010
7011 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7012
7013 * rtl.def: Add extra fourth optional field to define_cond_exec.
7014 * gensupport.c (process_one_cond_exec): Process attributes from
7015 define_cond_exec.
7016 * doc/md.texi: Document fourth field in define_cond_exec.
7017
7018 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7019
7020 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7021 out the processing order as in store_bit_field_1.
7022
7023 2013-06-04 Jan Hubicka <jh@suse.cz>
7024
7025 PR middle-end/57500
7026 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7027 non-VAR_DECL node if it does not exist yet.
7028
7029 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7030
7031 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7032 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7033 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7034 target_cpu_default setting.
7035
7036 2013-06-03 Teresa Johnson <tejohnson@google.com>
7037
7038 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7039 default to -fopt-info=optimized instead of all.
7040 * doc/invoke.texi: Ditto.
7041 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7042 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7043 (execute_vect_slp): Emit BB vectorization success under
7044 MSG_OPTIMIZED_LOCATIONS.
7045 * tree-vect-slp.c (vect_slp_transform_bb): Change
7046 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7047 * tree-vect-loop.c (vect_transform_loop): Ditto.
7048
7049 2013-06-03 Jason Merrill <jason@redhat.com>
7050
7051 PR c++/57415
7052 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7053 Use TARGET_EXPR for C++.
7054
7055 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7056
7057 PR rtl-optimization/57268
7058 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7059 if DEBUG_INSN_P (insn).
7060
7061 Reapply
7062 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7063
7064 PR rtl-optimization/57268
7065 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7066 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7067
7068 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7069
7070 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7071 (ix86_avoid_lea_for_addr): Likewise.
7072 (exact_dependency_1): Likewise.
7073 (ix86_adjust_cost): Likewise.
7074 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7075 removed.
7076 (do_reorder_for_imul): Fix typo, formatting and
7077 !reload_completed check removed.
7078 (ix86_sched_reorder): Fix typo and formatting.
7079 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7080 list.
7081
7082 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7083
7084 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7085
7086 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7087
7088 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7089 <STRING_CST>: Likewise.
7090 <VECTOR_CST>: Likewise.
7091
7092 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7093 Mikael Morin <mikael@gcc.gnu.org>
7094
7095 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
7096 * config.in: Regenerated.
7097 * configure: Regenerated.
7098
7099 2013-06-01 Jan Hubicka <jh@suse.cz>
7100
7101 PR middle-end/57366
7102 * cgraphunit.c (compile): When weakref is not supported,
7103 set up transparent aliases before final output pass.
7104 * varasm.c (assemble_alias): Do not try to do it here.
7105
7106 2013-06-01 Jan Hubicka <jh@suse.cz>
7107
7108 PR middle-end/57467
7109 * passes.c (for_per_function): Skip unanalyzed functions.
7110
7111 2013-06-01 Jan Hubicka <jh@suse.cz>
7112
7113 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
7114 (lto_symtab_merge_symbols_1): ... this one.
7115 (lto_symtab_merge_cgraph_nodes): Rename to ...
7116 (lto_symtab_merge_symbols): ... this one; simplify.
7117 * cgraph.c (same_body_aliases_done): Rename to ...
7118 (cpp_implicit_aliases_done): ... this one.
7119 (cgraph_create_function_alias): Update.
7120 (cgraph_same_body_alias): Update.
7121 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
7122 (verify_edge_corresponds_to_fndecl): Simplify.
7123 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
7124 (cgraph_node): Remove same_body_alias.
7125 (varpool_node): Remove alias_of and extra_name_alias.
7126 (same_body_aliases_done): Rename to ..
7127 (cpp_implicit_aliases_done): ... this one.
7128 (symtab_alias_ultimate_target): Add default parameter.
7129 (symtab_resolve_alias): New function.
7130 (fixup_same_cpp_alias_visibility): Declare.
7131 (cgraph_function_node): Add default parameter.
7132 (cgraph_node_asm_name): Likewise.
7133 (cgraph_function_or_thunk_node): Add default parameter; do
7134 not ICE when it is NULL.
7135 (varpool_variable_node): Likewise.
7136 * tree-emutls.c (create_emultls_var): Update.
7137 (ipa_lower_emutls): Update.
7138 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
7139 (cgraph_reset_node): Reset alias info.
7140 (cgraph_finalize_function): Update.
7141 (fixup_same_cpp_alias_visibility): Move to symtab.c.
7142 (analyze_function): Simplify.
7143 (cgraph_process_same_body_aliases): Simplify.
7144 (analyze_functions): Fixup same body aliases.
7145 (handle_alias_pairs): Simplify.
7146 (assemble_thunk): Update.
7147 (assemble_thunks_and_aliases): Update.
7148 (output_weakrefs): Rewrite.
7149 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
7150 (lto_output_varpool_node): Likewise.
7151 (compute_ltrans_boundary): Remve assert.
7152 (get_alias_symbol): New functoin.
7153 (input_node): Rewrite alias handling.
7154 (input_varpool_node): Likewise.
7155 * ipa-pure-const.c (propagate_pure_const): Fix formating.
7156 * ipa.c (process_references): Handle weakrefs correctly.
7157 (symtab_remove_unreachable_nodes): Likewise.
7158 * trans-mem.c (get_cg_data): Update.
7159 (ipa_tm_create_version_alias): Update.
7160 (ipa_tm_execute): Update.
7161 * symtab.c (dump_symtab_base): Dump aliases.
7162 (verify_symtab_base): Verify aliases.
7163 (symtab_node_availability): New function.
7164 (symtab_alias_ultimate_target): Simplify.
7165 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
7166 handle all the fixup cases.
7167 (symtab_resolve_alias): New function.
7168 * passes.c (ipa_write_summaries): Handle weakrefs.
7169 * varpool.c (varpool_analyze_node): Simplify.
7170 (assemble_aliases): Update.
7171 (varpool_create_variable_alias): Simplify.
7172 (varpool_extra_name_alias): Simplify.
7173 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
7174 (lto_symtab_merge_symbols): ... this one.
7175
7176 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
7177
7178 Revert
7179 PR rtl-optimization/57268
7180 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7181 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7182
7183 2013-06-01 Tobias Burnus <burnus@net-b.de>
7184
7185 Partially reverted:
7186 2013-05-31 Tobias Burnus <burnus@net-b.de>
7187
7188 PR middle-end/57073
7189 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
7190 further up.
7191
7192 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7193
7194 PR rtl-optimization/57268
7195 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7196 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7197
7198 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7199
7200 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
7201 unordered comparison operators when -fno-trapping-math is in effect
7202 on the e500.
7203 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
7204 and implement unordered comparison operators properly on the e500.
7205
7206 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
7207
7208 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
7209 for constant scalar integers.
7210 (simplify_relational_operation_1): Likewise.
7211
7212 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
7213
7214 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
7215 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
7216 Fix comment.
7217
7218 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7219 Igor Zamyatin <igor.zamyatin@intel.com>
7220
7221 Silvermont (SLM) architecture performance tuning.
7222 * config/i386/i386.h (enum ix86_tune_indices): Add
7223 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
7224 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
7225
7226 * config/i386/i386.c (initial_ix86_tune_features)
7227 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
7228 (ix86_lea_outperforms): Handle Silvermont tuning.
7229 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
7230 call.
7231 (ix86_use_lea_for_mov): Likewise.
7232 (ix86_avoid_lea_for_addr): Likewise.
7233 (ix86_lea_for_add_ok): Likewise.
7234 (exact_dependency_1): New function.
7235 (exact_store_load_dependency): Likewise.
7236 (ix86_adjust_cost): Handle Silvermont tuning.
7237 (do_reoder_for_imul): Likewise.
7238 (swap_top_of_ready_list): New function.
7239 (ix86_sched_reorder): Changed to handle Silvermont tuning.
7240
7241 * config/i386/i386.md (peepholes that split memory operand in fp
7242 converts): New.
7243
7244 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7245
7246 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
7247 Remove un-necessary braces.
7248
7249 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
7250
7251 * config/aarch64/aarch64.c (aarch64_classify_symbol):
7252 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
7253
7254 2013-05-31 Tobias Burnus <burnus@net-b.de>
7255
7256 PR middle-end/57073
7257 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
7258
7259 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7260
7261 PR target/56315
7262 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
7263 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
7264 * config/arm/neon.md (iordi3_neon): Remove.
7265 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
7266 * config/arm/predicates.md (imm_for_neon_logic_operand):
7267 Move to earlier in the file.
7268 (neon_logic_op2): Likewise.
7269 (arm_iordi_operand_neon): New predicate.
7270
7271 2013-05-31 Richard Biener <rguenther@suse.de>
7272
7273 PR tree-optimization/57478
7274 PR tree-optimization/57453
7275 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
7276 are life as well.
7277
7278 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7279
7280 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
7281 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
7282
7283 2013-05-30 Tobias Burnus <burnus@net-b.de>
7284 Thomas Koenig <tkoenig@gcc.gnu.org>
7285
7286 PR middle-end/57073
7287 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
7288 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
7289
7290 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
7291
7292 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
7293
7294 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
7295
7296 * target.def (register_usage_leveling_p): New hook.
7297 * targhooks.c (default_register_usage_leveling_p): New.
7298 * targhooks.h (default_register_usage_leveling_p): New prototype.
7299 * lra-assigns.c (register_usage_leveling_p): Use the hook.
7300 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
7301 * doc/tm.texi: Update.
7302 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
7303
7304 2013-05-30 Ian Bolton <ian.bolton@arm.com>
7305
7306 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
7307 (*insv_reg<mode>): New define_insn.
7308
7309 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
7310
7311 PR rtl-optimization/57439
7312 * postreload.c (move2add_valid_value_p): Check that we have
7313 a zero subreg_regno_offset when accessing the register in
7314 the requested mode.
7315
7316 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7317 Igor Zamyatin <igor.zamyatin@intel.com>
7318
7319 Silvermont (SLM) architecture pipeline model, tuning and
7320 insn selection.
7321 * config.gcc: Add slm config options and target.
7322
7323 * config/i386/slm.md: New.
7324
7325 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
7326
7327 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
7328 PROCESSOR_SLM.
7329 (ix86_target_macros_internal): Likewise.
7330
7331 * gcc/config/i386/i386.c (slm_cost): New cost.
7332 (m_SLM): New macro flag.
7333 (initial_ix86_tune_features): Set m_SLM.
7334 (x86_accumulate_outgoing_args): Likewise.
7335 (x86_arch_always_fancy_math_387): Likewise.
7336 (processor_target_table): Add slm cost.
7337 (cpu_names): Add slm cpu name.
7338 (x86_option_override_internal): Set SLM ISA.
7339 (ix86_issue_rate): New case PROCESSOR_SLM.
7340 (ia32_multipass_dfa_lookahead): Likewise.
7341 (fold_builtin_cpu): Add slm.
7342
7343 * config/i386/i386.h (TARGET_SLM): New target macro.
7344 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
7345 (processor_type): Add PROCESSOR_SLM.
7346
7347 * config/i386/i386.md (cpu): Add new value "slm".
7348 (slm.md): Include slm.md.
7349
7350 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
7351 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7352
7353 * config/arm/arm-protos.h: Add and update function protos.
7354 * config/arm/arm.c (use_simple_return_p): New added.
7355 (thumb2_expand_return): Check simple_return flag.
7356 * config/arm/arm.md: Add simple_return and conditional simple_return.
7357 * config/arm/iterators.md: Add iterator for return and simple_return.
7358
7359 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7360
7361 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
7362 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
7363 (arm_emit_vfp_multi_reg_pop): Likewise.
7364 (thumb2_emit_ldrd_pop): Likewise.
7365 (arm_expand_epilogue): Add misc REG_CFA notes.
7366 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
7367
7368 2013-05-29 Lawrence Crowl <crowl@google.com>
7369
7370 * config/arm/t-arm: Update for below.
7371
7372 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
7373 Change type to hash_table. Update dependent calls and types.
7374
7375 * config/i386/t-cygming: Update for below.
7376
7377 * config/i386/t-interix: Update for below.
7378
7379 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
7380 Change type to hash_table. Update dependent calls and types.
7381 (i386_find_on_wrapper_list::wrappers): Likewise.
7382
7383 * config/ia64/t-ia64: Update for below.
7384
7385 * config/ia64/ia64.c (bundle_state_table):
7386 Change type to hash_table. Update dependent calls and types.
7387
7388 * config/mips/mips.c (mips_reorg_process_insns::htab):
7389 Change type to hash_table. Update dependent calls and types.
7390
7391 * config/sol2.c (solaris_comdat_htab):
7392 Change type to hash_table. Update dependent calls and types.
7393
7394 * config/t-sol2: Update for above.
7395
7396 2013-05-29 Teresa Johnson <tejohnson@google.com>
7397
7398 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
7399 functions are not yet marked as defined.
7400
7401 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
7402 Pat Haugen <pthaugen@us.ibm.com>
7403 Peter Bergner <bergner@vnet.ibm.com>
7404
7405 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
7406 instructions.
7407 (VEC_A): Likewise.
7408 (VEC_C): Likewise.
7409 (vrotl<mode>3): Likewise.
7410 (vashl<mode>3): Likewise.
7411 (vlshr<mode>3): Likewise.
7412 (vashr<mode>3): Likewise.
7413
7414 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7415 support for power8 V2DI builtins.
7416
7417 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
7418 power8 V2DI builtins.
7419 (vupkhsw): Likewise.
7420 (vupklsw): Likewise.
7421 (vaddudm): Likewise.
7422 (vminsd): Likewise.
7423 (vmaxsd): Likewise.
7424 (vminud): Likewise.
7425 (vmaxud): Likewise.
7426 (vpkudum): Likewise.
7427 (vpksdss): Likewise.
7428 (vpkudus): Likewise.
7429 (vpksdus): Likewise.
7430 (vrld): Likewise.
7431 (vsld): Likewise.
7432 (vsrd): Likewise.
7433 (vsrad): Likewise.
7434 (vsubudm): Likewise.
7435 (vcmpequd): Likewise.
7436 (vcmpgtsd): Likewise.
7437 (vcmpgtud): Likewise.
7438 (vcmpequd_p): Likewise.
7439 (vcmpgtsd_p): Likewise.
7440 (vcmpgtud_p): Likewise.
7441 (vupkhsw): Likewise.
7442 (vupklsw): Likewise.
7443 (vaddudm): Likewise.
7444 (vmaxsd): Likewise.
7445 (vmaxud): Likewise.
7446 (vminsd): Likewise.
7447 (vminud): Likewise.
7448 (vpksdss): Likewise.
7449 (vpksdus): Likewise.
7450 (vpkudum): Likewise.
7451 (vpkudus): Likewise.
7452 (vrld): Likewise.
7453 (vsld): Likewise.
7454 (vsrad): Likewise.
7455 (vsrd): Likewise.
7456 (vsubudm): Likewise.
7457
7458 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
7459 support for power8 V2DI instructions.
7460
7461 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
7462 power8 V2DI instructions. Combine pack and unpack insns to use an
7463 iterator for each mode. Check whether a particular mode supports
7464 Altivec instructions instead of just checking TARGET_ALTIVEC.
7465 (UNSPEC_VPKUWUM): Likewise.
7466 (UNSPEC_VPKSHSS): Likewise.
7467 (UNSPEC_VPKSWSS): Likewise.
7468 (UNSPEC_VPKUHUS): Likewise.
7469 (UNSPEC_VPKSHUS): Likewise.
7470 (UNSPEC_VPKUWUS): Likewise.
7471 (UNSPEC_VPKSWUS): Likewise.
7472 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
7473 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
7474 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
7475 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
7476 (UNSPEC_VUPKHSB): Likewise.
7477 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
7478 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
7479 (UNSPEC_VUPKHSH): Likewise.
7480 (UNSPEC_VUPKLSB): Likewise.
7481 (UNSPEC_VUPKLSH): Likewise.
7482 (VI2): Likewise.
7483 (VI_char): Likewise.
7484 (VI_scalar): Likewise.
7485 (VI_unit): Likewise.
7486 (VP): Likewise.
7487 (VP_small): Likewise.
7488 (VP_small_lc): Likewise.
7489 (VU_char): Likewise.
7490 (add<mode>3): Likewise.
7491 (altivec_vaddcuw): Likewise.
7492 (altivec_vaddu<VI_char>s): Likewise.
7493 (altivec_vadds<VI_char>s): Likewise.
7494 (sub<mode>3): Likewise.
7495 (altivec_vsubcuw): Likewise.
7496 (altivec_vsubu<VI_char>s): Likewise.
7497 (altivec_vsubs<VI_char>s): Likewise.
7498 (altivec_vavgs<VI_char>): Likewise.
7499 (altivec_vcmpbfp): Likewise.
7500 (altivec_eq<mode>): Likewise.
7501 (altivec_gt<mode>): Likewise.
7502 (altivec_gtu<mode>): Likewise.
7503 (umax<mode>3): Likewise.
7504 (smax<mode>3): Likewise.
7505 (umin<mode>3): Likewise.
7506 (smin<mode>3): Likewise.
7507 (altivec_vpkuhum): Likewise.
7508 (altivec_vpkuwum): Likewise.
7509 (altivec_vpkshss): Likewise.
7510 (altivec_vpkswss): Likewise.
7511 (altivec_vpkuhus): Likewise.
7512 (altivec_vpkshus): Likewise.
7513 (altivec_vpkuwus): Likewise.
7514 (altivec_vpkswus): Likewise.
7515 (altivec_vpks<VI_char>ss): Likewise.
7516 (altivec_vpks<VI_char>us): Likewise.
7517 (altivec_vpku<VI_char>us): Likewise.
7518 (altivec_vpku<VI_char>um): Likewise.
7519 (altivec_vrl<VI_char>): Likewise.
7520 (altivec_vsl<VI_char>): Likewise.
7521 (altivec_vsr<VI_char>): Likewise.
7522 (altivec_vsra<VI_char>): Likewise.
7523 (altivec_vsldoi_<mode>): Likewise.
7524 (altivec_vupkhsb): Likewise.
7525 (altivec_vupkhs<VU_char>): Likewise.
7526 (altivec_vupkls<VU_char>): Likewise.
7527 (altivec_vupkhsh): Likewise.
7528 (altivec_vupklsb): Likewise.
7529 (altivec_vupklsh): Likewise.
7530 (altivec_vcmpequ<VI_char>_p): Likewise.
7531 (altivec_vcmpgts<VI_char>_p): Likewise.
7532 (altivec_vcmpgtu<VI_char>_p): Likewise.
7533 (abs<mode>2): Likewise.
7534 (vec_unpacks_hi_v16qi): Likewise.
7535 (vec_unpacks_hi_v8hi): Likewise.
7536 (vec_unpacks_lo_v16qi): Likewise.
7537 (vec_unpacks_hi_<VP_small_lc>): Likewise.
7538 (vec_unpacks_lo_v8hi): Likewise.
7539 (vec_unpacks_lo_<VP_small_lc>): Likewise.
7540 (vec_pack_trunc_v8h): Likewise.
7541 (vec_pack_trunc_v4si): Likewise.
7542 (vec_pack_trunc_<mode>): Likewise.
7543
7544 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
7545 V2DI builtins.
7546 (vec_vmaxsd): Likewise.
7547 (vec_vmaxud): Likewise.
7548 (vec_vminsd): Likewise.
7549 (vec_vminud): Likewise.
7550 (vec_vpksdss): Likewise.
7551 (vec_vpksdus): Likewise.
7552 (vec_vpkudum): Likewise.
7553 (vec_vpkudus): Likewise.
7554 (vec_vrld): Likewise.
7555 (vec_vsld): Likewise.
7556 (vec_vsrad): Likewise.
7557 (vec_vsrd): Likewise.
7558 (vec_vsubudm): Likewise.
7559 (vec_vupkhsw): Likewise.
7560 (vec_vupklsw): Likewise.
7561
7562 2013-05-29 Jan Hubicka <jh@suse.cz>
7563
7564 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
7565 flags; reorder rest of fields in more consistent way.
7566 (varpool_node): Remove analyzed, finalized and alias.
7567 (cgraph_ndoe): Likewise.
7568 (symtab_alias_ultimate_target): New function.
7569 (cgraph_function_node): Move offline.
7570 (cgraph_reset_node): Declare.
7571 (cgraph_comdat_can_be_unshared_p): Remove.
7572 (varpool_remove_initializer): Declare.
7573 (varpool_first_defined_variable, varpool_next_defined_variable
7574 cgraph_first_defined_function, cgraph_next_defined_function): Update.
7575 (cgraph_function_with_gimple_body_p): Update.
7576 (varpool_all_refs_explicit_p): Update.
7577 (symtab_alias_target): New function.
7578 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
7579 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
7580 (cgraph_function_or_thunk_node): Simplify using
7581 symtab_alias_ultimate_target.
7582 (varpool_variable_node): Likewise.
7583 * cgraph.c (cgraph_create_function_alias): Update.
7584 (cgraph_add_thunk): Update.
7585 (cgraph_remove_node): Update.
7586 (dump_cgraph_node): Do not dump removed flags.
7587 (cgraph_function_body_availability): Update.
7588 (cgraph_propagate_frequency): Update.
7589 (verify_cgraph_node): Check sanity of local flag.
7590 (cgraph_function_node): Move here from cgraph.h; revamp for
7591 cgraph_function_or_thunk_node.
7592 * lto-symtab.c (lto_varpool_replace_node): Update.
7593 (lto_symtab_resolve_can_prevail_p): Update.
7594 (lto_symtab_merge_cgraph_nodes): Update.
7595 * ipa-cp.c (determine_versionability, initialize_node_lattices,
7596 propagate_constants_accross_call, devirtualization_time_bonus,
7597 ipcp_propagate_stage): Update.
7598 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
7599 * ipa-inline-transform.c (clone_inlined_nodes,
7600 preserve_function_body_p): Update.
7601 * ipa-reference.c (propagate): Update.
7602 (write_node_summary_p): Update.
7603 * toplev.c (wrapup_global_declaration_2): Update.
7604 * cgraphunit.c (cgraph_analyze_function): Rename to ...
7605 (analyze_function) ... this one.
7606 (cgraph_process_new_functions): Update.
7607 (cgraph_reset_node): Export.
7608 (cgraph_finalize_function): Update.
7609 (cgraph_add_new_function): Update.
7610 (process_function_and_variable_attributes): Update.
7611 (varpool_finalize_decl): Update.
7612 (symbol_finalized): Remove.
7613 (symbol_finalized_and_needed): Rename to ...
7614 (symbol_defined_and_needed): ... update.
7615 (cgraph_analyze_functions): Update.
7616 (handle_alias_pairs): Update.
7617 (mark_functions_to_output): Update.
7618 (assemble_thunk): Update.
7619 (output_in_order): Update.
7620 (output_weakrefs): Update.
7621 (finalize_compilation_unit): Update.
7622 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
7623 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
7624 input_node, input_varpool_node): Update.
7625 * dbxout.c (dbxout_expand_expr): Update.
7626 * cgraphclones.c (cgraph_clone_node): Update.
7627 (cgraph_copy_node_for_versioning): Update.
7628 (cgraph_materialize_clone): Update.
7629 (cgraph_materialize_all_clones): Update.
7630 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
7631 propagate_pure_const, propagate_nothrow): Update.
7632 * lto-streamer-out.c (lto_output, write_symbol): Update.
7633 * ipa-utils.c (ipa_reverse_postorder): Update.
7634 * ipa-inline.c (can_inline_edge_p): Update.
7635 (update_caller_keys, ipa_inline): Update.
7636 * dwarf2out.c (reference_to_unused,
7637 premark_types_used_by_global_vars_helper): Update.
7638 * tree-eh.c (tree_could_trap_p): Update.
7639 * ipa-split.c (consider_split, execute_split_functions): Update.
7640 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
7641 has_addr_references_p): Update; move ahead in file for better
7642 readability.
7643 (process_references): Simplify.
7644 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
7645 bodies are removed.
7646 (cgraph_comdat_can_be_unshared_p): Make static.
7647 (cgraph_externally_visible_p): Update.
7648 (varpool_externally_visible_p): Update.
7649 (function_and_variable_visibility): Update.
7650 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
7651 ipa_tm_mark_force_output_node): Update.
7652 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
7653 estimate_edge_devirt_benefit, inline_generate_summary,
7654 inline_write_summary): Update.
7655 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
7656 * ipa-prop.c (ipa_compute_jump_functions): Update.
7657 (ipa_print_node_params, ipa_prop_read_section,
7658 ipa_update_after_lto_read, read_replacements_section): Update.
7659 * varasm.c (mark_decl_referenced): Update.
7660 (assemble_alias, dump_tm_clone_pairs): Update.
7661 * tree-inline.c (copy_bb): Update.
7662 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
7663 Update.
7664 * symtab.c (dump_symtab_base): Print new flags.
7665 (verify_symtab_base): Verify new flags.
7666 (symtab_alias_ultimate_target): New function.
7667 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
7668 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
7669 Update.
7670 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
7671 Update.
7672 * i386.c (ix86_get_function_versions_dispatcher,
7673 ix86_generate_version_dispatcher_body): Update.
7674 (fold_builtin_cpu): Use varpool_add_new_variable.
7675 * varpool.c (varpool_remove_initializer): Break out from ...
7676 (varpool_remove_node): ... this one.
7677 (dump_varpool_node, varpool_node_for_asm,
7678 cgraph_variable_initializer_availability, varpool_analyze_node,
7679 varpool_assemble_decl, varpool_remove_unreferenced_decls,
7680 varpool_finalize_named_section_flags, varpool_create_variable_alias):
7681 Update.
7682
7683 2013-05-29 Jan Hubicka <jh@suse.cz>
7684
7685 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
7686
7687 2013-05-29 Easwaran Raman <eraman@google.com>
7688
7689 PR tree-optimization/57442
7690 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
7691 when control exits the main loop.
7692
7693 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
7694
7695 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
7696 and RX600.
7697 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
7698 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
7699 * rx/t-rx: Add rx100 under multi library matches option for nofpu
7700 option.
7701
7702 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7703
7704 PR tree-optimization/57441
7705 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
7706 Don't limit size of incr_vec to number of candidates.
7707
7708 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
7709
7710 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
7711 and mips16 directories.
7712 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
7713 (MULTILIB_DIRNAMES): Ditto.
7714 (MULTILIB_EXCEPTIONS): Add new exceptions.
7715 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
7716 (MULTILIB_DIRNAMES): Ditto.
7717 (MULTILIB_EXCEPTIONS): Add new exceptions.
7718
7719 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7720 Marcus Shawcroft <marcus.shawcroft@arm.com>
7721
7722 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
7723 SYMBOL_TINY_ABSOLUTE.
7724 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
7725 SYMBOL_TINY_ABSOLUTE.
7726 (aarch64_expand_mov_immediate): Likewise.
7727 (aarch64_classify_symbol): Likewise.
7728 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
7729 Permit SYMBOL_TINY_ABSOLUTE.
7730 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
7731
7732 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
7733 Marcus Shawcroft <marcus.shawcroft@arm.com>
7734
7735 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
7736 Refactor if/switch. Replace gcc_assert with if.
7737
7738 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
7739
7740 * config/i386/i386.c (initial_ix86_tune_features): Enable
7741 FP Reassociation for AMD bdver1 and bdver2.
7742
7743 2013-05-29 Martin Jambor <mjambor@suse.cz>
7744
7745 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
7746 and IMAGPART_EXPR do not occur within other handled_components.
7747
7748 2013-05-29 Richard Biener <rguenther@suse.de>
7749
7750 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
7751 access on whether the use is in the BB we currently try to
7752 vectorize.
7753 (vect_bb_vectorization_profitable_p): Pass the BB we currently
7754 vectorize to vect_bb_slp_scalar_cost.
7755
7756 2013-05-29 Richard Biener <rguenther@suse.de>
7757
7758 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
7759 computing scalar cost offsetted by stmts that are kept live
7760 by scalar uses.
7761 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
7762 for computation of scalar cost.
7763
7764 2013-05-28 Steve Ellcey <sellcey@mips.com>
7765
7766 * config/mips/mips-cpus.def (mips32r2): Change processor type.
7767
7768 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
7769
7770 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
7771 array notation built-in reduction functions.
7772 * doc/passes.texi (Passes): Added documentation about changes done
7773 for Cilk Plus.
7774 * doc/invoke.texi (C Dialect Options): Added documentation about
7775 the -fcilkplus flag.
7776 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
7777 (BUILTINS_DEF): Depend on cilkplus.def.
7778 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
7779 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
7780 * cilkplus.def: New file.
7781
7782 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
7783
7784 PR rtl-optimization/57439
7785 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
7786
7787 2013-05-28 Easwaran Raman <eraman@google.com>
7788
7789 PR tree-optimization/57337
7790 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
7791 (find_insert_point): Correctly identify the insertion point
7792 when two statements with the same UID is compared.
7793
7794 2013-05-28 Richard Biener <rguenther@suse.de>
7795
7796 PR tree-optimization/56787
7797 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
7798 from the list of data references.
7799 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
7800 clobbers.
7801 (vect_analyze_loop_operations): Likewise.
7802 (vect_transform_loop): Remove clobbers.
7803
7804 2013-05-28 Martin Jambor <mjambor@suse.cz>
7805
7806 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
7807 and REALPART_EXPRs have scalar type.
7808
7809 2013-05-28 Richard Biener <rguenther@suse.de>
7810
7811 PR tree-optimization/57411
7812 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
7813 virtual operands.
7814 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
7815 virtual operand propagation.
7816
7817 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7818
7819 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
7820 destination register for bmasksi_vis.
7821 (vector_init_bshuffle): Likewise.
7822 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
7823
7824 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
7825
7826 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
7827 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
7828 mode if the instruction isn't available in the original mode.
7829 * config/sparc/sparc.opt (mfix-ut699): New option.
7830 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
7831 (divdf3): Turn into expander.
7832 (divdf3_nofix): New insn.
7833 (divdf3_fix): Likewise.
7834 (divsf3): Disable if -mfix-ut699.
7835 (sqrtdf2): Turn into expander.
7836 (sqrtdf2_nofix): New insn.
7837 (sqrtdf2_fix): Likewise.
7838 (sqrtsf2): Disable if -mfix-ut699.
7839
7840 2013-05-27 Richard Biener <rguenther@suse.de>
7841
7842 PR middle-end/57412
7843 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
7844 block for the new loop.
7845
7846 2013-05-27 Richard Biener <rguenther@suse.de>
7847
7848 PR tree-optimization/57343
7849 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
7850 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
7851 (number_of_iterations_cond): Do not build the folded tree.
7852
7853 2013-05-27 Richard Biener <rguenther@suse.de>
7854
7855 Revert
7856 PR middle-end/57381
7857 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
7858 OEP_CONSTANT_ADDRESS_OF retained.
7859
7860 PR tree-optimization/57417
7861 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
7862 for unchanged base.
7863 (set_ssa_val_to): Compare addresses using
7864 get_addr_base_and_unit_offset.
7865
7866 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7867
7868 PR rtl-optimization/56833
7869 * postreload.c (move2add_record_mode): New function.
7870 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
7871 (move2add_use_add2_insn): Use move2add_record_sym_value.
7872 (move2add_use_add3_insn): Likewise.
7873 (reload_cse_move2add): Use move2add_valid_value_p and
7874 move2add_record_mode. Invalidate call-clobbered and REG_INC
7875 affected regs by setting reg_mode to VOIDmode.
7876 (move2add_note_store): Don't pretend the inside of a SUBREG is
7877 the actual destination. Invalidate single/leading registers by
7878 setting reg_mode to VOIDmode.
7879 Use move2add_record_sym_value, move2add_valid_value_p and
7880 move2add_record_mode.
7881
7882 2013-05-27 Richard Biener <rguenther@suse.de>
7883
7884 PR tree-optimization/57396
7885 * tree-affine.c (double_int_constant_multiple_p): Properly
7886 return false for val == 0 and div != 0.
7887
7888 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7889
7890 * config/mips/mips.h: Use #elif in preprocessor conditions.
7891
7892 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7893
7894 PR target/53916
7895 * config/mips/constraints.md (kl): New constraint.
7896 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
7897 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
7898 constraint for operand 0. Split after CSE for MIPS16. Emit a move
7899 from LO for MIPS16.
7900 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
7901
7902 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
7903
7904 PR target/55777
7905 * config/mips/mips.c (mips_can_inline_p): New function.
7906 (TARGET_CAN_INLINE_P): Define.
7907
7908 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
7909
7910 * sched-int.h (ds_t, dw_t): Make unsigned int.
7911 Fix documentation that describes how all the ds_t bits are used.
7912 Reserve the last bit for delayed-branch scheduling.
7913 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
7914 (BITS_PER_DEP_WEAK): Fix definition and documentation.
7915 (gen_dep_weak_1): Remove prototype.
7916 * sched-deps.c (get_dep_weak_1): Make static.
7917 * target.def (speculate_insn, needs_block_p, gen_spec_check,
7918 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
7919 * doc/tm.texi: Regenerate.
7920 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
7921
7922 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
7923
7924 PR debug/56950
7925 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
7926
7927 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
7928 Sandra Loosemore <sandra@codesourcery.com>
7929
7930 * config.gcc (powerpc-*): Allow native for with-cpu.
7931
7932 2013-05-24 Jeff Law <law@redhat.com>
7933
7934 PR tree-optimization/57124
7935 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
7936 conversion feeding a condition if the range has an overflow
7937 if -fstrict-overflow. Add warnings for when we do make the
7938 transformation.
7939
7940 2013-05-24 Dehao Chen <dehao@google.com>
7941
7942 * tree-cfg.c (locus_discrim_map): Fix the typo.
7943 (locus_discrim_hasher): Likewise.
7944 (locus_discrim_hasher::hash): Likewise.
7945 (locus_discrim_hasher::equal): Likewise.
7946
7947 2013-05-24 Martin Jambor <mjambor@suse.cz>
7948
7949 PR tree-optimization/57294
7950 * cgraph.h (ipa_record_stmt_references): Declare.
7951 * cgraphbuild.c (ipa_record_stmt_references): New function.
7952 (build_cgraph_edges): Use ipa_record_stmt_references.
7953 (rebuild_cgraph_edges): Likewise.
7954 (cgraph_rebuild_references): Likewise.
7955 * ipa-prop.c (ipa_modify_call_arguments): Discard references
7956 associated with the old statement and build references from the
7957 newly built statements.
7958 * ipa-ref.c (ipa_remove_stmt_references): New function.
7959 * ipa-ref.h (ipa_remove_stmt_references): Declare.
7960
7961 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
7962
7963 * lra-constraints.c (emit_spill_move): Use smaller mode for
7964 mem-mem moves.
7965 (check_and_process_move): Consider mem-reg moves for secondary
7966 too.
7967 (curr_insn_transform): Don't lose insns emitted before for
7968 secondary memory moves.
7969 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
7970 reg set up in the current insn.
7971
7972 2013-05-24 Dehao Chen <dehao@google.com>
7973
7974 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
7975 hash function.
7976 (locus_descrim_hasher::equal): Likewise.
7977 (build_gimple_cfg): New discriminator assignment algorithm.
7978 (make_edges): Likewise.
7979 (next_discriminator_for_locus): Likewise.
7980 (same_line_p): Likewise.
7981 (assign_discriminators): Likewise.
7982 (make_cond_expr_edges): Likewise.
7983 (make_gimple_switch_edges): Likewise.
7984 (make_goto_expr_edges): Likewise.
7985 (make_gimple_asm_edges): Likewise.
7986
7987 2013-05-24 Ian Bolton <ian.bolton@arm.com>
7988
7989 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
7990 X format specifier to only display bottom 16 bits.
7991 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
7992 immediate to match for operand 2, since it will be masked.
7993
7994 2013-05-24 Richard Biener <rguenther@suse.de>
7995
7996 PR tree-optimization/57287
7997 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
7998 all SSA names that occur in abnormal PHIs.
7999
8000 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8001
8002 PR tree-ssa/57385
8003 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8004 that index is not negative.
8005
8006 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8007
8008 PR rtl-optimization/55177
8009 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8010 (simplify_byte_swapping_operation): New.
8011 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8012 (simplify_relational_operation_1): Deal with BSWAP.
8013
8014 2013-05-23 Richard Henderson <rth@redhat.com>
8015
8016 PR target/56742
8017 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8018 (ix86_reorg): Call it.
8019
8020 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8021
8022 PR target/57379
8023 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8024 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8025 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8026
8027 2013-05-23 Christian Bruel <christian.bruel@st.com>
8028
8029 PR debug/57351
8030 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8031
8032 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8033 Marcus Shawcroft <marcus.shawcroft@arm.com>
8034
8035 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8036 * config/aarch64/constraints.md (Usa): Remove.
8037 * doc/md.texi (AArch64 Usa): Remove.
8038
8039 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8040 Marcus Shawcroft <marcus.shawcroft@arm.com>
8041
8042 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8043 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8044 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8045 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8046
8047 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8048
8049 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8050 instruction (AdvSIMD).
8051 * config/aarch64/aarch64-builtins.c
8052 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8053 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8054
8055 2013-05-23 Martin Jambor <mjambor@suse.cz>
8056
8057 PR middle-end/57347
8058 * tree.h (contains_bitfld_component_ref_p): Declare.
8059 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8060 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8061 caller.
8062 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8063 not access a bit-field. Assert all final offsets are byte-aligned.
8064
8065 2013-05-23 Richard Biener <rguenther@suse.de>
8066
8067 PR tree-optimization/57380
8068 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8069 least one invariant or re-used load.
8070 * passes.c (init_optimization_passes): Move pass_phiprop before
8071 pass_forwprop.
8072
8073 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8074
8075 * config/aarch64/aarch64-simd.md
8076 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8077
8078 2013-05-23 Richard Biener <rguenther@suse.de>
8079
8080 PR middle-end/57381
8081 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8082 OEP_CONSTANT_ADDRESS_OF retained.
8083
8084 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8085
8086 PR middle-end/57344
8087 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8088 don't lower unit. Handle unit not being always BITS_PER_WORD.
8089
8090 2013-05-23 Richard Biener <rguenther@suse.de>
8091
8092 PR rtl-optimization/57341
8093 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
8094 instead of true_dependence.
8095
8096 2013-05-22 David Malcolm <dmalcolm@redhat.com>
8097
8098 * bb-reorder.c (branch_threshold): Make const.
8099 (exec_threshold): Ditto.
8100
8101 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8102 Pat Haugen <pthaugen@us.ibm.com>
8103 Peter Bergner <bergner@vnet.ibm.com>
8104
8105 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
8106 documentation for the power8 crypto builtins.
8107
8108 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
8109
8110 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
8111 macros for defining power8 builtin functions.
8112 (BU_P8V_AV_2): Likewise.
8113 (BU_P8V_AV_P): Likewise.
8114 (BU_P8V_VSX_1): Likewise.
8115 (BU_P8V_OVERLOAD_1): Likewise.
8116 (BU_P8V_OVERLOAD_2): Likewise.
8117 (BU_CRYPTO_1): Likewise.
8118 (BU_CRYPTO_2): Likewise.
8119 (BU_CRYPTO_3): Likewise.
8120 (BU_CRYPTO_OVERLOAD_1): Likewise.
8121 (BU_CRYPTO_OVERLOAD_2): Likewise.
8122 (XSCVSPDP): Fix typo, point to the correct instruction.
8123 (VCIPHER): Add power8 crypto builtins.
8124 (VCIPHERLAST): Likewise.
8125 (VNCIPHER): Likewise.
8126 (VNCIPHERLAST): Likewise.
8127 (VPMSUMB): Likewise.
8128 (VPMSUMH): Likewise.
8129 (VPMSUMW): Likewise.
8130 (VPERMXOR_V2DI): Likewise.
8131 (VPERMXOR_V4SI: Likewise.
8132 (VPERMXOR_V8HI: Likewise.
8133 (VPERMXOR_V16QI: Likewise.
8134 (VSHASIGMAW): Likewise.
8135 (VSHASIGMAD): Likewise.
8136 (VPMSUM): Likewise.
8137 (VPERMXOR): Likewise.
8138 (VSHASIGMA): Likewise.
8139
8140 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8141 __CRYPTO__ if the crypto instructions are available.
8142 (altivec_overloaded_builtins): Add support for overloaded power8
8143 builtins.
8144
8145 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
8146 support for power8 crypto builtins.
8147 (builtin_function_type): Likewise.
8148 (altivec_init_builtins): Add support for builtins that take vector
8149 long long (V2DI) arguments.
8150
8151 * config/rs6000/crypto.md: New file, define power8 crypto
8152 instructions.
8153
8154 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8155 Pat Haugen <pthaugen@us.ibm.com>
8156 Peter Bergner <bergner@vnet.ibm.com>
8157
8158 * doc/invoke.texi (Option Summary): Add power8 options.
8159 (RS/6000 and PowerPC Options): Likewise.
8160
8161 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
8162 constraints.md instead of rs6000.h. Reorder w* constraints. Add
8163 wm, wn, wr documentation.
8164
8165 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
8166 registers if direct move instructions are enabled.
8167 (wn): New constraint for no registers.
8168 (wq): New constraint for quad word even GPR registers.
8169 (wr): New constraint if 64-bit instructions are enabled.
8170 (wv): New constraint if power8 vector instructions are enabled.
8171 (wQ): New constraint for quad word memory locations.
8172
8173 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
8174 constraint for 0..15 for crypto instructions.
8175 (gpc_reg_operand): If VSX allow registers in VSX registers as well
8176 as GPR and floating point registers.
8177 (int_reg_operand): New predicate to match only GPR registers.
8178 (base_reg_operand): New predicate to match base registers.
8179 (quad_int_reg_operand): New predicate to match even GPR registers
8180 for quad memory operations.
8181 (vsx_reg_or_cint_operand): New predicate to allow vector logical
8182 operations in both GPR and VSX registers.
8183 (quad_memory_operand): New predicate for quad memory operations.
8184 (reg_or_indexed_operand): New predicate for direct move support.
8185
8186 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
8187 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
8188 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
8189 (POWERPC_MASKS): Add power8 options.
8190 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
8191 various options.
8192
8193 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
8194 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
8195
8196 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
8197 (-mpower8-fusion): New power8 options.
8198 (-mpower8-fusion-sign): Likewise.
8199 (-mpower8-vector): Likewise.
8200 (-mcrypto): Likewise.
8201 (-mdirect-move): Likewise.
8202 (-mquad-memory): Likewise.
8203
8204 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
8205 power8.
8206 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
8207 registers.
8208 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
8209 (rs6000_debug_vector_unit): Add p8_vector.
8210 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
8211 definitions. Also print fusion state.
8212 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
8213 (rs6000_builtin_mask_calculate): Add power8 builtin support.
8214 (rs6000_option_override_internal): Add support for power8.
8215 (rs6000_common_init_builtins): Add debugging for skipped builtins
8216 if -mdebug=builtin.
8217 (rs6000_adjust_cost): Add power8 support.
8218 (rs6000_issue_rate): Likewise.
8219 (insn_must_be_first_in_group): Likewise.
8220 (insn_must_be_last_in_group): Likewise.
8221 (force_new_group): Likewise.
8222 (rs6000_register_move_cost): Likewise.
8223 (rs6000_opt_masks): Likewise.
8224
8225 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
8226 power8 capable assembler, default to power7 options.
8227 (TARGET_DIRECT_MOVE): Likewise.
8228 (TARGET_CRYPTO): Likewise.
8229 (TARGET_P8_VECTOR): Likewise.
8230 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
8231 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
8232 (VECTOR_MEM_P8_VECTOR_P): Likewise.
8233 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
8234 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
8235 (TARGET_XSCVDPSPN): Likewise.
8236 (TARGET_XSCVSPDPN): Likewsie.
8237 (TARGET_SYNC_HI_QI): Likewise.
8238 (TARGET_SYNC_TI): Likewise.
8239 (MASK_CRYPTO): Likewise.
8240 (MASK_DIRECT_MOVE): Likewise.
8241 (MASK_P8_FUSION): Likewise.
8242 (MASK_P8_VECTOR): Likewise.
8243 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
8244 temporary used by some of the direct move instructions to get two FP
8245 temporary registers does not force creation of a stack frame.
8246 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
8247 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
8248 that any VSX registers are tieable, even if they are also an
8249 Altivec vector mode.
8250 (r6000_reg_class_enum): Add wm, wr, wv constraints.
8251 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
8252 (RS6000_BTM_CRYPTO): Likewise.
8253 (RS6000_BTM_COMMON): Likewise.
8254
8255 * config/rs6000/rs6000.md (cpu attribute): Add power8.
8256 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
8257 (enum rs6000_vector): Add power8 vector support.
8258
8259 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8260
8261 PR target/19599
8262 PR target/57340
8263 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
8264 (any_sibcall_could_use_r3): this and handle indirect calls.
8265 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
8266
8267 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8268
8269 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
8270
8271 2013-05-22 Richard Biener <rguenther@suse.de>
8272
8273 PR middle-end/57349
8274 * profile.c (branch_prob): Do not split blocks that are
8275 abnormally receiving from ECF_RETURNS_TWICE functions.
8276
8277 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8278
8279 * recog.c (offsettable_address_addr_space_p): Fix calculation of
8280 address mode. Move pointer mode initialization to the same place.
8281
8282 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8283
8284 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
8285 while it has any effect.
8286
8287 2013-05-21 Easwaran Raman <eraman@google.com>
8288
8289 PR tree-optimization/57322
8290 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
8291 UID of the statement added to the BB to be 1.
8292
8293 2013-05-21 Jakub Jelinek <jakub@redhat.com>
8294
8295 PR tree-optimization/57331
8296 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
8297 of conversion from pointer type to integral type with integer.
8298
8299 2013-05-21 Martin Jambor <mjambor@suse.cz>
8300
8301 PR lto/57289
8302 * ipa-prop.c (ipa_read_node_info): Process param_used and
8303 controlled_uses in the same order as when writing.
8304
8305 2013-05-21 Magnus Granberg <baldrick@free.fr>
8306
8307 PR plugins/56754
8308 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
8309
8310 2013-05-21 Richard Biener <rguenther@suse.de>
8311
8312 PR tree-optimization/57318
8313 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
8314 estimate stmts with side-effects as likely eliminated.
8315
8316 2013-05-21 Richard Biener <rguenther@suse.de>
8317
8318 PR tree-optimization/57330
8319 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
8320 preserve the call stmts fntype.
8321
8322 2013-05-21 Richard Biener <rguenther@suse.de>
8323
8324 PR tree-optimization/57303
8325 * tree-ssa-sink.c (statement_sink_location): Improve killing
8326 stmt detection and properly handle self-assignments.
8327
8328 2013-05-21 Christian Bruel <christian.bruel@st.com>
8329
8330 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
8331 spanning registers. LEAF_REG_REMAP is supported only for contiguous
8332 registers. Set register size out of the PARALLEL loop.
8333
8334 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
8335
8336 PR target/56547
8337 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
8338 (*fmasf4, *fmasf4_media): New insns.
8339
8340 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8341
8342 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
8343 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
8344 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
8345 (mips_idiv_insns): Update the comments to say that the returned
8346 instruction counts are in units of BASE_INSN_LENGTH.
8347 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
8348 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
8349 using 2 rather than 4 as the length of indirect MIPS16 and
8350 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
8351 length of a NOP. Don't divide MIPS16 lengths by 2.
8352 (mips16_split_long_branches): Assume a branch is long if the
8353 length is greater than 4 rather than 8.
8354 * config/mips/mips.md (length): Give MIPS16 lengths directly,
8355 rather than multiplying them by 2. Multiply instruction counts
8356 by BASE_INSN_LENGTH rather than 4.
8357 (*jump_mips16, tls_get_tp_mips16_<mode>)
8358 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
8359
8360 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8361
8362 * config/mips/mips.md (extended_mips16): Remove branch case.
8363 (length): Remove duplicated extended_mips16 test.
8364
8365 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
8366
8367 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
8368
8369 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
8370
8371 * recog.h (Recog_data): Rename to...
8372 (recog_data_d): ...this.
8373 (recog_data): Update accordingly.
8374 * recog.c (recog_data): Likewise.
8375 * reload.c (save_recog_data): Likewise.
8376 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
8377 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
8378
8379 2013-05-17 Julian Brown <julian@codesourcery.com>
8380
8381 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
8382 found in a REG_EQUAL note, invalidate it.
8383
8384 2013-05-17 Easwaran Raman <eraman@google.com>
8385
8386 * tree-ssa-reassoc.c (find_insert_point): New function.
8387 (insert_stmt_after): Likewise.
8388 (get_def_stmt): Likewise.
8389 (ensure_ops_are_available): Likewise.
8390 (not_dominated_by): Likewise.
8391 (rewrite_expr_tree): Do not move statements beyond what is
8392 necessary. Remove call to swap_ops_for_binary_stmt...
8393 (reassociate_bb): ... and move it here.
8394 (build_and_add_sum): Assign UIDs for new statements.
8395 (linearize_expr): Likewise.
8396 (do_reassoc): Renumber gimple statement UIDs.
8397
8398 2013-05-17 Jan Hubicka <jh@suse.cz>
8399
8400 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
8401 weakrefs.
8402 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
8403 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
8404 weakrefs.
8405 (output_weakrefs): Update.
8406
8407 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
8408 Martin Jambor <mjambor@suse.cz>
8409
8410 PR middle-end/57276
8411 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
8412 value that corresponds to the given aggval is found in values vector.
8413
8414 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
8415
8416 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
8417 sse, sse2, sse3, ssse3 and sse4a flags to options.
8418
8419 2013-05-17 David Malcolm <dmalcolm@redhat.com>
8420
8421 * gengtype-state.c: (s_expr_writer): New class, to handle
8422 prettifying of output layout of s-expressions.
8423 (state_writer): New class, to write out gtype.state.
8424 (state_written_type_count): Move this variable into member data of
8425 state_writer.
8426 (s_expr_writer::s_expr_writer): New code: constructor for new class
8427 (state_writer::state_writer(): ditto
8428 (s_expr_writer::write_new_line): New function
8429 (s_expr_writer::write_any_indent): ditto
8430 (s_expr_writer::begin_s_expr): ditto
8431 (s_expr_writer::end_s_expr): ditto
8432 (write_state_fileloc): convert to method of state_writer...
8433 (state_writer:: write_state_fileloc): ...and use methods of
8434 s_expr_writer to write indentation into the gtype.state output file
8435 to visually represent the hierarchical structure of the list
8436 structures
8437 (write_state_fields): ditto, renaming to...
8438 (state_writer::write_state_fields)
8439 (write_state_a_string): ditto, renaming to...
8440 (state_writer::write_state_a_string)
8441 (write_state_string_option): ditto, renaming to...
8442 (state_writer::write_state_string_option)
8443 (write_state_type_option): ditto, renaming to...
8444 (state_writer::write_state_type_option)
8445 (write_state_nested_option): ditto, renaming to...
8446 (state_writer::write_state_nested_option)
8447 (write_state_option): ditto, renaming to...
8448 (state_writer::write_state_option)
8449 (write_state_options): ditto, renaming to...
8450 (state_writer::write_state_options)
8451 (write_state_lang_bitmap): ditto, renaming to...
8452 (state_writer::write_state_lang_bitmap)
8453 (write_state_version): ditto, renaming to...
8454 (state_writer::write_state_version)
8455 (write_state_scalar_type): ditto, renaming to...
8456 (state_writer::write_state_scalar_type)
8457 (write_state_string_type): ditto, renaming to...
8458 (state_writer::write_state_string_type)
8459 (write_state_undefined_type): ditto, renaming to...
8460 (state_writer::write_state_undefined_type)
8461 (write_state_struct_union_type): ditto, renaming to...
8462 (state_writer::write_state_struct_union_type)
8463 (write_state_struct_type): ditto, renaming to...
8464 (state_writer::write_state_struct_type)
8465 (write_state_user_struct_type): ditto, renaming to...
8466 (state_writer::write_state_user_struct_type)
8467 (write_state_lang_struct_type): ditto, renaming to...
8468 (state_writer::write_state_lang_struct_type)
8469 (write_state_param_struct_type): ditto, renaming to...
8470 (state_writer::write_state_param_struct_type)
8471 (write_state_pointer_type): ditto, renaming to...
8472 (state_writer::write_state_pointer_type)
8473 (write_state_array_type): ditto, renaming to...
8474 (state_writer::write_state_array_type)
8475 (write_state_gc_used): ditto, renaming to...
8476 (state_writer::write_state_gc_used)
8477 (write_state_common_type_content): ditto, renaming to...
8478 (state_writer::write_state_common_type_content)
8479 (write_state_type): ditto, renaming to...
8480 (state_writer::write_state_type)
8481 (write_state_pair_list): ditto, renaming to...
8482 (state_writer::write_state_pair_list)
8483 (write_state_pair): ditto, renaming to...
8484 (state_writer::write_state_pair)
8485 (write_state_typedefs): ditto, renaming to...
8486 (state_writer::write_state_typedefs)
8487 (write_state_structures): ditto, renaming to...
8488 (state_writer::write_state_structures)
8489 (write_state_param_structs): ditto, renaming to...
8490 (state_writer::write_state_param_structs)
8491 (write_state_variables): ditto, renaming to...
8492 (state_writer::write_state_variables)
8493 (write_state_srcdir): ditto, renaming to...
8494 (state_writer::write_state_srcdir)
8495 (write_state_files_list): ditto, renaming to...
8496 (state_writer::write_state_files_list)
8497 (write_state_languages): ditto, renaming to...
8498 (state_writer::write_state_languages)
8499 (write_state): create a state_writer instance and use it when
8500 writing out the state file
8501
8502 2013-05-17 Mike Stump <mikestump@comcast.net>
8503
8504 PR rtl-optimization/57304
8505 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
8506 accessing DF_REF_REAL_LOC.
8507
8508 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8509
8510 PR rtl-optimization/57281
8511 PR rtl-optimization/57300
8512 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
8513 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
8514 what the other splitter did if the registers are dead.
8515
8516 2013-05-17 Richard Biener <rguenther@suse.de>
8517
8518 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
8519 MEM_REF offsets.
8520
8521 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8522
8523 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
8524 linking.
8525
8526 2013-05-17 Marek Polacek <polacek@redhat.com>
8527
8528 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
8529 length when doing non-zero store of storing '\0' to '\0'.
8530
8531 2013-05-17 Jakub Jelinek <jakub@redhat.com>
8532
8533 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
8534 vect_external_def oprnd1 with loop_vinfo, try to emit
8535 optional cast, negation and and stmts on the loop preheader
8536 edge instead of into the pattern def seq.
8537
8538 PR tree-optimization/57051
8539 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
8540 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
8541
8542 2013-05-16 Nick Clifton <nickc@redhat.com>
8543
8544 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
8545 (rl78_is_naked_func): New function.
8546 (rl78_expand_prologue): Skip prologue generation for naked functions.
8547 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
8548 * doc/extend.texi (naked): Add RL78 to the list of processors
8549 that supports this attribute.
8550
8551 2013-05-16 Jeff Law <law@redhat.com>
8552
8553 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
8554
8555 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8556
8557 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
8558 cache parameters using detect_caches_amd also for CYRIX,
8559 NSC and TM2 signatures.
8560
8561 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
8562 Dzianis Kahanovich <mahatma@eu.by>
8563
8564 PR target/45359
8565 PR target/46396
8566 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8567 VIA/Centaur processors and determine their cache parameters
8568 using detect_caches_amd.
8569
8570 2013-05-16 Teresa Johnson <tejohnson@google.com>
8571
8572 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
8573 (rtl_verify_edges): New function.
8574 (rtl_verify_bb_insns): Ditto.
8575 (rtl_verify_bb_pointers): Ditto.
8576 (rtl_verify_bb_insn_chain): Ditto.
8577 (rtl_verify_fallthru): Ditto.
8578 (rtl_verify_bb_layout): Ditto.
8579 (rtl_verify_flow_info_1): Outline checks into new functions.
8580 (rtl_verify_flow_info): Ditto.
8581
8582 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
8583
8584 * cfghooks.c (copy_bbs): Add update_dominance argument.
8585 * cfghooks.h (copy_bbs): Update prototype.
8586 * tree-cfg.c (gimple_duplicate_sese_region):
8587 Add update_dominance argument.
8588 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
8589 * tree-ssa-loop-ch.c (copy_loop_headers): Update
8590 gimple_duplicate_sese_region call.
8591 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
8592 Update copy_bbs call.
8593 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
8594 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
8595
8596 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8597
8598 * tree-vectorizer.h (NUM_PATTERNS): Increment.
8599 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
8600 vect_recog_rotate_pattern.
8601 (vect_recog_rotate_pattern): New function.
8602
8603 2013-05-16 Jason Merrill <jason@redhat.com>
8604
8605 * Makefile.in (LLINKER): New variable.
8606 (mostlyclean): Remove link mutex.
8607 * configure.ac: Handle --enable-link-mutex.
8608 * lock-and-run.sh: New script.
8609
8610 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8611
8612 PR target/19599
8613 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
8614 for NULL decl.
8615
8616 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8617
8618 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
8619
8620 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
8621
8622 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
8623 * config/arm/arm.c (next_consecutive_mem): New function.
8624 (gen_movmem_ldrd_strd): Likewise.
8625 * config/arm/arm.md (movmemqi): Update condition and code.
8626 (unaligned_loaddi, unaligned_storedi): New patterns.
8627
8628 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8629
8630 * config.gcc: Obsolete *-*-solaris2.9*.
8631 * doc/install.texi (Specific, *-*-solaris2*): Document it.
8632
8633 2013-05-16 Richard Biener <rguenther@suse.de>
8634
8635 * passes.c (init_optimization_passes): Move pass_parallelize_loops
8636 earlier, after GRAPHITE transforms and IV canonicalization.
8637
8638 2013-05-16 Jakub Jelinek <jakub@redhat.com>
8639
8640 * omp-low.c (extract_omp_for_data): For collapsed loops,
8641 if at least one of the loops is known at compile time to
8642 iterate zero times, set count to 0.
8643 (expand_omp_regimplify_p): New function.
8644 (expand_omp_for_generic): For collapsed loops, if at least
8645 one of the loops isn't known to iterate at least once,
8646 add runtime check with setting count to 0.
8647 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
8648 For unsigned types if it isn't known at compile time that
8649 the loop will iterate at least once, add runtime check to bypass
8650 the whole loop if initial condition isn't true.
8651
8652 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
8653
8654 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
8655
8656 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
8657
8658 PR middle-end/57286
8659 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
8660 transformations to avoid an infinite loop.
8661
8662 2013-05-16 Marek Polacek <polacek@redhat.com>
8663
8664 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
8665
8666 2013-05-15 Leif Ekblad <leif@rdos.net>
8667
8668 * config/i386/i386.c (ix86_decompose_address): Use
8669 DEFAULT_TLS_SEG_REG to access TLS segment register.
8670 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
8671 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
8672 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
8673
8674 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
8675
8676 PR target/57260
8677 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
8678 sibling calls to functions that would normally be lazily bound,
8679 unless $gp is call-clobbered.
8680
8681 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
8682
8683 * config/i386/i386.c (ix86_option_override_internal): Update
8684 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
8685 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
8686 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
8687 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
8688 of TARGET_3DNOW.
8689 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
8690
8691 2013-05-15 Andreas Schwab <schwab@suse.de>
8692
8693 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
8694 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
8695 third operand.
8696
8697 2013-05-15 Teresa Johnson <tejohnson@google.com>
8698
8699 * loop-unroll.c (report_unroll_peel): Check decision before
8700 emitting unroll/peel message.
8701
8702 2013-05-15 Teresa Johnson <tejohnson@google.com>
8703
8704 * function.h (has_bb_partition): New rtl_data flag.
8705 (bb_reorder_complete): Ditto.
8706 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
8707 instead of flag_reorder_blocks_and_partition.
8708 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
8709 with some enhancements.
8710 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
8711 * bb-reorder.c (connect_traces): Check for has_bb_partition
8712 instead of flag_reorder_blocks_and_partition.
8713 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
8714 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
8715 verify_hot_cold_block_grouping.
8716 (partition_hot_cold_basic_blocks): Set has_bb_partition.
8717
8718 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8719
8720 PR target/19599
8721 * config/arm/predicates.md (call_insn_operand): New predicate.
8722 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
8723 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
8724 if insn is not a tail call.
8725 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
8726 registers.
8727 * config/arm/arm.h (enum reg_class): New caller save register class.
8728 (REG_CLASS_NAMES): Likewise.
8729 (REG_CLASS_CONTENTS): Likewise.
8730 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
8731 without decls.
8732
8733 2013-05-15 Richard Biener <rguenther@suse.de>
8734
8735 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
8736 of MSG_OPTIMIZED_LOCATIONS.
8737 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
8738 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
8739 message.
8740 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
8741 of MSG_OPTIMIZED_LOCATIONS.
8742 (execute_vect_slp): Likewise.
8743 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
8744 (vect_create_cond_for_alias_checks): Likewise.
8745 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
8746 (vect_recog_widen_mult_pattern): Likewise.
8747 (vect_recog_widen_sum_pattern): Likewise.
8748 (vect_recog_over_widening_pattern): Likewise.
8749 (vect_recog_widen_shift_pattern): Likewise.
8750 (vect_recog_vector_vector_shift_pattern): Likewise.
8751 (vect_recog_divmod_pattern): Likewise.
8752 (vect_recog_mixed_size_cond_pattern): Likewise.
8753 (vect_recog_bool_pattern): Likewise.
8754 (vect_pattern_recog_1): Likewise.
8755
8756 2013-05-15 Martin Jambor <mjambor@suse.cz>
8757
8758 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
8759 non-functions to builtin_unreachable.
8760 * ipa-inline-transform.c (inline_call): Do not assert estimates were
8761 correct when new direct edges were discovered.
8762
8763 2013-05-15 Martin Jambor <mjambor@suse.cz>
8764
8765 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
8766 header, print symbol order instead of node uid, print more information
8767 about indirect edge targets.
8768 (ipa_make_edge_direct_to_target): Print symbol order instead of node
8769 uids.
8770 (ipa_make_edge_direct_to_target): Likewise.
8771 (remove_described_reference): Likewise.
8772 (propagate_controlled_uses): Likewise.
8773 (ipa_print_node_params): Also print symbol order.
8774 (ipcp_transform_function): Print symbol order instead of node uids.
8775 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
8776 (cgraph_get_create_real_symbol_node): Likewise.
8777 * ipa-cp.c (print_lattice): Likewise.
8778 (print_all_lattices): Likewise.
8779 (determine_versionability): Likewise.
8780 (initialize_node_lattices): Likewise.
8781 (estimate_local_effects): Likewise.
8782 (update_profiling_info): Likewise.
8783 (create_specialized_node): Likewise.
8784 (perhaps_add_new_callers): Likewise.
8785 (decide_about_value): Likewise.
8786 (decide_whether_version_node): Likewise.
8787 (identify_dead_nodes): Likewise.
8788 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
8789 (dump_inline_summary): Likewise.
8790 (estimate_node_size_and_time): Likewise.
8791 (inline_analyze_function): Likewise.
8792 * ipa-inline.c (report_inline_failed_reason): Likewise.
8793 (want_early_inline_function_p): Likewise.
8794 (edge_badness): Likewise.
8795 (update_edge_key): Likewise.
8796 (inline_small_functions): Likewise. Add dumping of order to two other
8797 dumps.
8798 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
8799 instead of node uids.
8800 (propagate_pure_const): Likewise.
8801 (propagate_pure_const): Likewise.
8802 * ipa-utils.c (dump_cgraph_node_set): Likewise.
8803 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
8804 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
8805 of node uids.
8806 * tree-pretty-print.c (dump_function_header): Likewise.
8807 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
8808 Print symbol order instead of node uids.
8809
8810 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8811
8812 * config/s390/s390.c (s390_register_move_cost): Don't impose the
8813 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
8814
8815 2013-05-15 Richard Biener <rguenther@suse.de>
8816
8817 PR tree-optimization/57275
8818 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
8819 return value for fail to do runtime alias checks for gather loads.
8820
8821 2013-05-15 Jan Hubicka <jh@suse.cz>
8822
8823 PR lto/57038
8824 PR lto/47375
8825 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
8826 weakrefs are not external.
8827 (lto_symtab_merge_decls): Fix thinko when dealing with
8828 non-lto_symtab decls.
8829 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
8830 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
8831 * varpool.c (dump_varpool_node): Dump more flags.
8832
8833 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8834
8835 * config/i386/i386.c (processor_alias_table): Add instruction
8836 FSGSBASE for AMD bdver3 architecture.
8837
8838 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8839
8840 * tree.c (warn_deprecated_use): Print file:line using locus color.
8841 * diagnostic.c (diagnostic_report_current_module): Print file:line
8842 and file:line:column using locus color.
8843
8844 2013-05-14 Mike Stump <mikestump@comcast.net>
8845
8846 * gdbinit.in: Add __null.
8847
8848 2013-05-14 Mike Stump <mikestump@comcast.net>
8849
8850 * recog.h: Rename struct recog_data to Recog_data.
8851 * recog.c: Likewise.
8852 * reload.c (can_reload_into): Likewise.
8853 * config/picochip/picochip.c: Likewise.
8854
8855 2013-05-14 Mike Stump <mikestump@comcast.net>
8856
8857 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
8858
8859 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
8860
8861 * resource.h (struct resources): Remove unch_memory member.
8862 (CLEAR_RESOURCE): Don't clear unch_memory.
8863 * resource.c (mark_referenced_resources): Don't set it.
8864 (mark_set_resources): Likewise.
8865 (mark_target_live_regs): Don't clear it.
8866 (init_resource_info): Likewise.
8867 * reorg.c (resource_conflicts_p): Don't compare it.
8868 (redundant_insn): Don't set it.
8869
8870 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
8871 Remove prototypes.
8872 * emit-rtl.c (next_label): Remove unused function.
8873 (skip_consecutive_labels, link_cc0_insns): Move to ...
8874 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
8875 only place where these functions are used, and make them static.
8876
8877 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8878
8879 * fold-const.c (fold_negate_expr): Handle vectors.
8880 (fold_truth_not_expr): Make it static.
8881 (fold_invert_truthvalue): New static function.
8882 (invert_truthvalue_loc): Handle vectors. Do not call
8883 fold_truth_not_expr directly.
8884 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
8885 <TRUTH_NOT_EXPR>: Do not cast to boolean.
8886 (fold_comparison): Handle vector constants.
8887 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
8888 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
8889 * tree.h (fold_truth_not_expr): Remove declaration.
8890
8891 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
8892
8893 * config/aarch64/aarch64-simd.md
8894 (aarch64_vcond_internal<mode>): Rename to...
8895 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
8896 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
8897 float modes. Clarify all iterator modes.
8898 (vcond<mode><mode>): Use new name for vcond expanders.
8899 (vcond<v_cmp_result><mode>): Likewise.
8900 (vcondu<mode><mode>: Likewise.
8901 * config/aarch64/iterators.md (VDQF_COND): New.
8902
8903 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
8904
8905 PR bootstrap/57266
8906 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
8907 variable for the shift amount. Check that we shift by non-negative
8908 amounts.
8909
8910 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
8911
8912 PR target/42017
8913 * config/arm/arm.h (EPILOGUE_USES): Only return true
8914 for LR_REGNUM after epilogue_completed.
8915
8916 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
8917
8918 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
8919 is error_mark_node.
8920
8921 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8922
8923 PR target/57261
8924 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
8925 and Solaris 11+/x86 with gld.
8926 * configure: Regenerate.
8927
8928 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8929
8930 * expmed.c (expand_shift_1): Canonicalize rotates by
8931 constant bitsize / 2 to bitsize - 1.
8932 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
8933 case ROTATERT>: Likewise.
8934
8935 Revert:
8936 2013-05-10 Jakub Jelinek <jakub@redhat.com>
8937
8938 * config/i386/i386.md (rotateinv): New code attr.
8939 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
8940 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
8941 roll $31, %eax, etc.
8942
8943 2013-05-14 Richard Biener <rguenther@suse.de>
8944
8945 PR middle-end/57235
8946 * tree-eh.c (sink_clobbers): Give up for successors with
8947 multiple predecessors and no virtual uses.
8948
8949 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
8950
8951 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
8952 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
8953
8954 2013-05-14 Jakub Jelinek <jakub@redhat.com>
8955
8956 PR middle-end/57251
8957 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
8958 the case when both op0 and op1 have VOIDmode.
8959
8960 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8961
8962 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
8963 in multiply-accumulate mode.
8964
8965 2013-05-13 Guozhi Wei <carrot@google.com>
8966
8967 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
8968
8969 2013-05-13 Kai Tietz <ktietz@redhat.com>
8970
8971 PR target/56975
8972 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
8973 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
8974 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
8975 * config/i386/i386.c (ix86_option_override_internal): Likewise.
8976 (ix86_expand_prologue): Likewise.
8977 (ix86_expand_split_stack_prologue): Likewise.
8978 (legitimate_pic_address_disp_p): Likewise.
8979 (legitimize_pic_address): Likewise.
8980 (legitimize_tls_address): Likewise.
8981 (legitimize_pe_coff_symbol): Likewise.
8982 (output_pic_addr_const): Likewise.
8983 (construct_plt_address): Likewise.
8984 (ix86_expand_call): Likewise.
8985 (x86_output_mi_thunk): Likewise.
8986 (x86_function_profiler): Likewise.
8987
8988 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
8989
8990 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
8991 similar switch cases.
8992 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
8993 (aarch64_simd_mov_to_<mode>low): Delete.
8994 (aarch64_simd_mov_to_<mode>high): Delete.
8995 (move_lo_quad_<mode>): Add w<-r alternative.
8996 (aarch64_simd_move_hi_quad_<mode>): Likewise.
8997 (aarch64_simd_mov_from_*): Update type attribute.
8998 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
8999 statement.
9000
9001 2013-05-13 Jan Hubicka <jh@suse.cz>
9002
9003 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9004 * config/i386/i386.c (ix86_compute_frame_layout,
9005 ix86_expand_epilogue, emit_i387_cw_initialization,
9006 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9007 ix86_local_alignment): Fix use of size/speed predicates.
9008
9009 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9010
9011 PR tree-optimization/45216
9012 PR tree-optimization/57157
9013 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9014 the (-Y) & (B - 1) variant if OP is |.
9015 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9016 return shifted. Use (-op1) & (prec - 1) as other_amount
9017 instead of prec - op1.
9018
9019 2013-05-13 Martin Jambor <mjambor@suse.cz>
9020
9021 PR middle-end/42371
9022 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9023 (ipa_constant_data): New type.
9024 (ipa_jump_func): Use ipa_constant_data to hold information about
9025 constant jump functions.
9026 (ipa_get_jf_constant): Adjust to jump function type changes.
9027 (ipa_get_jf_constant_rdesc): New function.
9028 (ipa_param_descriptor): New field controlled_uses.
9029 (ipa_get_controlled_uses): New function.
9030 (ipa_set_controlled_uses): Likewise.
9031 * ipa-ref.h (ipa_find_reference): Declare.
9032 * ipa-prop.c (ipa_cst_ref_desc): New type.
9033 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9034 changes.
9035 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9036 New parameter cs. Adjust all callers.
9037 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9038 (remove_described_reference): New function.
9039 (jfunc_rdesc_usable): Likewise.
9040 (try_make_edge_direct_simple_call): Decrement controlled use count,
9041 attempt to remove reference if it hits zero.
9042 (combine_controlled_uses_counters): New function.
9043 (propagate_controlled_uses): Likewise.
9044 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9045 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9046 (ipa_print_node_params): Print described use counter.
9047 (ipa_write_jump_function): Adjust to jump function type changes.
9048 (ipa_read_jump_function): New parameter CS, pass it to
9049 ipa_set_jf_constant. Adjust caller.
9050 (ipa_write_node_info): Stream controlled use count
9051 (ipa_read_node_info): Likewise.
9052 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9053 asserting.
9054 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9055 count. Remove cloning-added reference if it reaches zero.
9056 * ipa-ref.c (ipa_find_reference): New function.
9057
9058 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9059
9060 * config/i386/i386.c (processor_target_table): Modified default
9061 alignment values for AMD BD and BT architectures.
9062
9063 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9064
9065 * tree-vect-generic.c (uniform_vector_p): Move ...
9066 * tree.c (uniform_vector_p): ... here.
9067 * tree.h (uniform_vector_p): Declare it.
9068 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9069 into a scalar.
9070
9071 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9072
9073 PR tree-optimization/57230
9074 * tree-ssa-strlen.c (handle_char_store): Record length for
9075 array store from STRING_CST.
9076
9077 PR tree-optimization/57230
9078 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9079 check.
9080
9081 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9082
9083 * config/epiphany/epiphany.c (epiphany_init): Check size of
9084 NUM_MODES_FOR_MODE_SWITCHING.
9085 (epiphany_expand_prologue):
9086 Remove CONFIG_REGNUM initial value handling code.
9087 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9088 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9089 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9090 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9091 Don't return 1 for FP_MODE_NONE.
9092 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9093 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
9094 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
9095 * config/epiphany/epiphany.md (save_config): New pattern.
9096
9097 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
9098
9099 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
9100
9101 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
9102
9103 * config/i386/i386.md (memory): Handle sseishft1.
9104 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
9105 (*vec_extractv2di_1): Ditto.
9106
9107 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
9108
9109 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
9110 saved registers.
9111
9112 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
9113
9114 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
9115 Add mthumb/march=armv7-a multilib.
9116 Add mthumb/march=armv7-r multilib.
9117 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
9118
9119 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
9120
9121 * config/v850/t-rtems: Add more multilibs.
9122
9123 2013-05-10 Richard Biener <rguenther@suse.de>
9124
9125 PR tree-optimization/57214
9126 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
9127 not propagate from SSA names that occur in abnormal PHI nodes.
9128
9129 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
9130
9131 * stor-layout.c (element_precision): New function.
9132 * machmode.h (element_precision): Declare it.
9133 * tree.c (build_minus_one_cst): New function.
9134 (element_precision): Likewise.
9135 * tree.h (build_minus_one_cst): Declare new function.
9136 (element_precision): Likewise.
9137 * fold-const.c (operand_equal_p): Use element_precision.
9138 (fold_binary_loc): Handle vector types.
9139 * convert.c (convert_to_integer): Use element_precision.
9140 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
9141 separately.
9142
9143 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
9144
9145 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
9146 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
9147 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
9148 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9149 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
9150 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
9151 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
9152 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
9153 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
9154 (Uuw8): New constraints.
9155 (Usb4): Move into alphabetical order.
9156 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
9157 (sd8_operand, ub8_operand, uw8_operand): New predicates.
9158 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
9159 previously unnamed patterns.
9160 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
9161 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
9162 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
9163 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
9164 of set_attr_alternative/if_then_else. Use extended_mips16 instead
9165 of specific lengths.
9166
9167 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9168
9169 * config/i386/i386.md (rotateinv): New code attr.
9170 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9171 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9172 roll $31, %eax, etc.
9173
9174 PR tree-optimization/45216
9175 PR tree-optimization/57157
9176 * tree-ssa-forwprop.c (simplify_rotate): New function.
9177 (ssa_forward_propagate_and_combine): Call it.
9178
9179 2013-05-10 Richard Biener <rguenther@suse.de>
9180
9181 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
9182 disable peeling when we version for aliasing.
9183 (vector_alignment_reachable_p): Honor explicit user alignment.
9184 (vect_supportable_dr_alignment): Likewise.
9185 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
9186 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
9187 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
9188 then peeling to arrange for the cost-model check to come first.
9189
9190 2013-05-10 Alan Modra <amodra@gmail.com>
9191
9192 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
9193 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
9194 * configure: Regenerate.
9195
9196 2013-05-10 Alan Modra <amodra@gmail.com>
9197
9198 PR target/55033
9199 * varasm.c (default_elf_select_section): Move !DECL_P check..
9200 (get_named_section): ..to here before calling get_section_name.
9201 Adjust assertion.
9202 (default_section_type_flags): Add DECL_P check.
9203 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
9204 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
9205
9206 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
9207
9208 * config/epiphany/epiphany.c (epiphany_expand_prologue):
9209 When using gen_stack_adjust_str with a register offset, add a
9210 REG_FRAME_RELATED_EXPR note.
9211
9212 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
9213
9214 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
9215 (*vec_extractv4si_zext_mem): Ditto.
9216 (*vec_extractv2di): Add 0->x and x->x alternatives.
9217 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
9218 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
9219
9220 2013-05-09 Jason Merrill <jason@redhat.com>
9221
9222 N3639 C++1y VLA support
9223 * gimplify.c (gimplify_vla_decl): Don't touch an existing
9224 DECL_VALUE_EXPR.
9225
9226 * tree.c (build_constructor_va): New.
9227 * tree.h: Declare it.
9228
9229 2013-05-09 Martin Jambor <mjambor@suse.cz>
9230
9231 PR lto/57084
9232 * gimple-fold.c (canonicalize_constructor_val): Call
9233 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
9234
9235 2013-05-09 Jan Hubicka <jh@suse.cz>
9236 Richard Biener <rguenther@suse.de>
9237
9238 PR lto/54095
9239 * symtab.c (symtab_make_decl_local): Do not add private names.
9240
9241 2013-05-09 Jan Hubicka <jh@suse.cz>
9242
9243 PR lto/54095
9244 * symtab.c (insert_to_assembler_name_hash): Handle clones.
9245 (unlink_from_assembler_name_hash): Likewise.
9246 (symtab_prevail_in_asm_name_hash, symtab_register_node,
9247 symtab_unregister_node, symtab_initialize_asm_name_hash,
9248 change_decl_assembler_name): Update.
9249
9250 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
9251
9252 * config/aarch64/aarch64.md: New movtf split.
9253 (*movtf_aarch64): Update.
9254 (aarch64_movdi_tilow): Handle TF modes and rename to
9255 aarch64_movdi_<mode>low.
9256 (aarch64_movdi_tihigh): Handle TF modes and rename to
9257 aarch64_movdi_<mode>high
9258 (aarch64_movtihigh_di): Handle TF modes and rename to
9259 aarch64_mov<mode>high_di
9260 (aarch64_movtilow_di): Handle TF modes and rename to
9261 aarch64_mov<mode>low_di
9262 (aarch64_movtilow_tilow): Remove spurious whitespace.
9263 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
9264 splits.
9265 (aarch64_print_operand): Update.
9266
9267 2013-05-09 Alan Modra <amodra@gmail.com>
9268
9269 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
9270 powerpc64le.
9271 * configure: Regenerate.
9272
9273 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9274
9275 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
9276 splitter preparation statements.
9277 * config/i386/sse.md (*vec_extract* splitters): Ditto.
9278 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
9279 adjust_address_nv.
9280
9281 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9282
9283 * gimple-ssa-strength-reduction.c (count_candidates): Change
9284 return value to int.
9285 (analyze_candidates_and_replace): Change type of length to int.
9286
9287 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
9288
9289 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
9290 (*vec_extract<mode>): Use VI12_128 mode iterator.
9291 (*vec_extract<mode>_mem): Ditto.
9292 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
9293 attribute.
9294
9295 2013-05-08 Diego Novillo <dnovillo@google.com>
9296
9297 PR bootstrap/54659
9298
9299 Revert:
9300 2012-08-17 Diego Novillo <dnovillo@google.com>
9301
9302 PR bootstrap/54281
9303 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
9304 * config.in: Regenerate.
9305 * configure: Regenerate.
9306 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
9307
9308 2013-05-08 Jan Hubicka <jh@suse.cz>
9309
9310 PR lto/54095
9311 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
9312 * cgraph.h (symtab_node_base): Add unique_name.
9313 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9314 input_overwrite_node, input_varpool_node): Stream unique_name.
9315 * cgraphclones.c (cgraph_create_virtual_clone,
9316 cgraph_function_versioning): Set unique_name.
9317 * ipa.c (function_and_variable_visibility): Set unique_name.
9318
9319 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9320
9321 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
9322 (alloc_cand_and_find_basis): Restrict conditional candidate
9323 processing to CAND_MULTs.
9324
9325 2013-05-08 Jan Hubicka <jh@suse.cz>
9326
9327 PR lto/54095
9328 lto-symtab.c (lto_symtab_symbol_p): New function.
9329 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
9330 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
9331 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
9332 Skip static symbols.
9333
9334 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
9335
9336 PR tree-optimization/57200
9337 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
9338 Only call inform if the preceding warning_at returns true.
9339
9340 2013-05-07 Han Shen <shenhan@google.com>
9341
9342 * cfgexpand.c (record_or_union_type_has_array_p): New function.
9343 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
9344 * common.opt (fstack-protector-strong): New option.
9345 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
9346 * doc/invoke.texi (Optimization Options): Document
9347 "-fstack-protector-strong".
9348 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
9349
9350 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9351
9352 * config/mips/mips.c (mips_machine_reorg2): Return 0.
9353
9354 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
9355
9356 * ira.c (update_equiv_regs): Add insn having equiv memory even if
9357 it is not lhs of the insn.
9358 (setup_reg_equiv): Remove insn having equiv memory which it is not
9359 lhs of the insn.
9360 * lra-constraints.c (process_address): Try to improve generation
9361 code for address base + disp.
9362 (lra_constraints): Make correct the code for checking insn setting
9363 up backward equivalence. Remove insn only if it is in the init
9364 insn list.
9365 * lra-eliminations.c (update_reg_eliminate): Change return value.
9366 (lra_eliminate): Use the result.
9367
9368 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
9369
9370 * config/i386/sse.md (ssescalarnummask): New mode attribute.
9371 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
9372 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
9373 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
9374 register target operands.
9375 (*vec_extractv8hi_sse2): New pattern.
9376 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
9377 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
9378 (*vec_extract<mode>_mem): New insn and split pattern.
9379
9380 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
9381
9382 * config/arm/arm.c (arm_asan_shadow_offset): New function.
9383 (TARGET_ASAN_SHADOW_OFFSET): Define.
9384 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
9385 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
9386
9387 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9388
9389 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
9390 (incr_vec_index): Return -1 if increment not found.
9391 (create_add_on_incoming_edge): Assert if increment not found.
9392 (record_increment): Limit number of increments recorded.
9393 (all_phi_incrs_profitable): Return false if an increment not found.
9394 (replace_profitable_candidates): Don't process increments that were
9395 not recorded.
9396 (analyze_candidates_and_replace): Limit size of incr_vec.
9397
9398 2013-05-07 Richard Biener <rguenther@suse.de>
9399
9400 * calls.c (special_function_p): setjmp-like functions are leaf.
9401 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
9402 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
9403
9404 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
9405
9406 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
9407 (aarch64_simd_mov<mode>): New expander.
9408 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
9409 (aarch64_simd_mov_to_<mode>high): Likewise.
9410 (aarch64_simd_mov_from_<mode>low): Likewise.
9411 (aarch64_simd_mov_from_<mode>high): Likewise.
9412 (aarch64_dup_lane<mode>): Update.
9413 (aarch64_dup_lanedi): New instruction pattern.
9414 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
9415 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
9416
9417 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9418
9419 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
9420 (replace_mult_candidate): Remove unnecessary argument; remove
9421 unnecessary parameter from call to introduce_cast_before_cand.
9422 (replace_unconditional_candidate): Remove unnecessary parameter
9423 from call to replace_mult_candidate.
9424 (replace_conditional_candidate): Likewise.
9425 (insert_initializers): Use make_temp_ssa_name.
9426 (introduce_cast_before_cand): Remove unnecessary argument; use
9427 make_temp_ssa_name.
9428 (replace_one_candidate): Remove unnecessary argument; remove
9429 unnecessary parameter from calls to introduce_cast_before_cand.
9430 (replace_profitable_candidates): Remove unnecessary parameters
9431 from calls to replace_one_candidate.
9432
9433 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9434
9435 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
9436 phi def as possibly hiding a basis for a CAND_ADD whose operands
9437 have been commuted in the analysis.
9438 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
9439
9440 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
9441
9442 * config/aarch64/aarch64.md
9443 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
9444 shift value between 0-4.
9445
9446 2013-05-07 Richard Biener <rguenther@suse.de>
9447
9448 * double-int.h (rshift): New overload.
9449 * double-int.c (rshift): New function.
9450 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
9451 (create_reference_ops_from_ref): Remove.
9452 (vn_reference_insert): Use shared ops for constructing the
9453 reference and copy it.
9454
9455 2013-05-07 Richard Biener <rguenther@suse.de>
9456
9457 PR middle-end/57190
9458 * tree-eh.c (sink_clobbers): Properly propagate
9459 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
9460
9461 2013-05-07 Jakub Jelinek <jakub@redhat.com>
9462
9463 PR tree-optimization/57149
9464 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
9465 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
9466 collect_phi_def_edges, execute_late_warn_uninitialized): Use
9467 uninit_undefined_value_p instead of ssa_undefined_value_p.
9468
9469 PR debug/57184
9470 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
9471 for modifier == EXPAND_INITIALIZER.
9472
9473 2013-05-07 Anton Blanchard <anton@samba.org>
9474
9475 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
9476 for powerpc64 little endian.
9477 * configure: Regenerate.
9478
9479 2013-05-06 Graham Stott <grahams@btinternet.com>
9480
9481 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
9482 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
9483 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
9484 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
9485
9486 2013-05-06 Graham Stott <grahams@btinternet.com>
9487
9488 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
9489 codes which allow non-lvalues.
9490
9491 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
9492
9493 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
9494 components are all 1s.
9495 (integer_minus_onep): New function.
9496 * tree.h (integer_minus_onep): Declare it.
9497 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
9498 integer_minus_onep instead of integer_all_onesp.
9499
9500 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9501
9502 PR target/52933
9503 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
9504 variations of these patterns.
9505
9506 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9507
9508 * config/i386/i386.md (isa): Add x64_sse4 member.
9509 (enabled): Handle x64_sse4.
9510 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
9511 instruction for 64bit SSE4_1 targets. Update insn attributes.
9512 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
9513 instruction for SSE4_1 targets. Update insn attributes.
9514 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9515 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
9516 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
9517 const_1 selector.
9518 (*vec_extractv4si): Rename from *sse4_1_pextrd.
9519 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
9520 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
9521
9522 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9523
9524 PR target/57108
9525 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
9526
9527 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
9528
9529 * final.c (do_assembler_dialects): Don't handle curly braces and
9530 vertical bar escaped by % as dialect delimiters.
9531 (output_asm_insn): Print curly braces and vertical bar if escaped
9532 by % and ASSEMBLER_DIALECT defined.
9533 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
9534 * doc/tm.texi: Regenerated.
9535
9536 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
9537
9538 * config/mips/mips.c: Include tree-pass.h.
9539 (mips_reorg): Split in pre- and post-dbr_schedule parts.
9540 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
9541 (pass_mips_machine_reorg2): New machine specific pass.
9542 (insert_pass_mips_machine_reorg2): New pass plugin definition.
9543 (mips_option_override): Register the new pass.
9544 * rtl.h (cleanup_barriers): Remove prototype.
9545 (dbr_schedule): Likewise.
9546 * jump.c (cleanup_barriers): Make static.
9547 * reorg.c (dbr_schedule): Likewise.
9548
9549 2013-05-06 Richard Biener <rguenther@suse.de>
9550
9551 PR tree-optimization/57185
9552 * tree-parloops.c (add_field_for_reduction): Handle anonymous
9553 SSA names properly.
9554
9555 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
9556
9557 PR target/57106
9558 * config/i386/i386.c (add_parameter_dependencies): Add dependence
9559 between "first_arg" and "insn", not "last" and "insn".
9560
9561 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9562
9563 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
9564 (find_candidates_in_block): Re-enable slsr_process_phi.
9565 (create_phi_basis): Fix double counting of candidate adjustment.
9566
9567 2013-05-06 Richard Biener <rguenther@suse.de>
9568
9569 PR middle-end/57147
9570 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
9571 the edge is also fallthru, preserve it and just clear the
9572 abnormal flag.
9573 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
9574 also complex, preserve that and just clear the fallthru flag.
9575 * tree-inline.c (update_ssa_across_abnormal_edges): Also
9576 update virtual operands.
9577
9578 2013-05-06 Alan Modra <amodra@gmail.com>
9579
9580 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
9581 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
9582 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
9583 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
9584 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9585 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
9586
9587 2013-05-06 Alan Modra <amodra@gmail.com>
9588
9589 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
9590 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
9591 (DEFAULT_ASM_ENDIAN): Define.
9592 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
9593 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
9594 Update -K PIC clause from sysv4.h.
9595 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
9596 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
9597
9598 2013-05-06 Alan Modra <amodra@gmail.com>
9599
9600 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
9601 twice for little-endian.
9602 (ashrdi3_no_power, ashrdi3): Support little-endian.
9603
9604 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
9605
9606 PR target/55303
9607 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
9608 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
9609 related expanders.
9610 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
9611 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
9612 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
9613 New predicates.
9614
9615 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
9616 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
9617
9618 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
9619 * config/pa/pa.opt: Make mbig-switch a no-op.
9620 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
9621 (CASE_VECTOR_MODE): Always return SImode.
9622 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
9623 for the !TARGET_BIG_SWITCH case.
9624 * config/pa/pa-linux.h: Likewise.
9625 * config/pa/pa-openbsd.h: Likewise.
9626 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
9627 * config/pa/pa.md (short_jump): Remove define_insn.
9628 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
9629 (casesi0): Remove define_insn.
9630 (type): Remove btable_branch.
9631 (pa_combine_type): Likewise.
9632 (in_nullified_branch_delay): Likewise.
9633 (in_call_delay): Likewise.
9634 (define_delay): Likewise.
9635 (define_insn_reservation "Z3"): Likewise.
9636 (define_insn_reservation "Z4"): Likewise.
9637 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
9638 (pa_adjust_insn_length): Remove adjustment for btable branches.
9639 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
9640 and mno-big-switch
9641
9642 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
9643
9644 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
9645 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
9646 Add m->r,x alternatives.
9647 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
9648 splitters using SWI48x mode iterator.
9649 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
9650 TARGET_64BIT. Add m->x alternative.
9651 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
9652 Add o->x alternative. Enable for TARGET_SSE.
9653 (sse_storeq): Remove expander.
9654 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
9655 with memory input operand.
9656 (*vec_extractv2di_1 splitter): New.
9657 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
9658 * config/i386/i386.md (ssevecmodelower): New mode attribute.
9659
9660 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9661
9662 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
9663 (INT_LOWPART): Delete.
9664 (extract_MB): Adjust.
9665 (extract_ME): Adjust.
9666 (print_operand): Adjust.
9667
9668 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
9669
9670 * config/rs6000/predicates.md (reg_or_add_cint_operand,
9671 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9672 (reg_or_logical_cint_operand, easy_fp_constant,
9673 logical_const_operand): Delete "CONST_DOUBLE" case.
9674 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
9675 "HOST_BITS_PER_WIDE_INT == 64" test.
9676 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
9677 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
9678 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
9679 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
9680 test.
9681 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
9682 CONST_DOUBLE DImode/VOIDmode case.
9683 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
9684 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9685 CONST_DOUBLE VOIDmode case.
9686 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
9687 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
9688 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
9689 Delete CONST_DOUBLE case.
9690 (splitters for mov FMOVE64 const_double): Delete
9691 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
9692 "HOST_BITS_PER_WIDE_INT >= 64" test.
9693 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
9694 case.
9695 (mov DI const_double): Delete.
9696
9697 2013-05-04 Jakub Jelinek <jakub@redhat.com>
9698
9699 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
9700 on op shows all bits zero in mode of a lowpart subreg, return zero.
9701
9702 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
9703
9704 PR target/57150
9705 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
9706 to save TFmode registers and DImode to save TImode registers for
9707 caller save operations.
9708 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
9709 mark being partially clobbered since they only use the first
9710 double word.
9711
9712 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
9713 and TDmode only use the upper 64-bits of each VSX register.
9714
9715 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9716
9717 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
9718 (find_candidates_in_block): Disable slsr_process_phi.
9719
9720 2013-05-03 Guozhi Wei <carrot@google.com>
9721
9722 * coverage.c (coverage_obj_init): Move the construction of gcov
9723 constructor to ...
9724 (build_init_ctor): ... here.
9725
9726 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9727
9728 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
9729 (slsr_cand_d): Redefine def_phi.
9730 (stride_status, phi_adjust_status, count_phis_status): New enums.
9731 (find_phi_def): New.
9732 (find_basis_for_base_expr): New.
9733 (find_basis_for_candidate): Handle hidden bases.
9734 (alloc_cand_and_find_basis): Handle phi candidates.
9735 (slsr_process_phi): New.
9736 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
9737 (create_mul_imm_cand): Likewise.
9738 (create_add_ssa_cand): Exclude phi base candidates.
9739 (create_add_imm_cand): Likewise.
9740 (slsr_process_cast): Likewise.
9741 (slsr_process_copy): Likewise.
9742 (find_candidates_in_block): Handle phi candidates.
9743 (dump_candidate): Likewise.
9744 (unconditional_cands): Delete.
9745 (unconditional_cands_with_known_stride_p): Delete.
9746 (phi_dependent_cand_p): New.
9747 (cand_increment): Handle phi-dependent candidates.
9748 (replace_dependent): Delete.
9749 (replace_mult_candidate): New.
9750 (replace_unconditional_candidate): New.
9751 (incr_vec_index): Move to avoid forward reference.
9752 (create_add_on_incoming_edge): New.
9753 (create_phi_basis): New.
9754 (replace_dependents): Delete.
9755 (replace_conditional_candidate): New.
9756 (phi_add_costs): New.
9757 (replace_uncond_cands_and_profitable_phis): New.
9758 (record_increment): Handle phi adjustments.
9759 (record_phi_increments): New.
9760 (record_increments): Handle phi adjustments.
9761 (phi_incr_cost): New.
9762 (lowest_cost_path): Handle phis.
9763 (total_savings): Likewise.
9764 (analyze_increments): Likewise.
9765 (ncd_with_phi): New.
9766 (ncd_of_cand_and_phis): New.
9767 (nearest_common_dominator_for_cands): Handle phi increments.
9768 (all_phi_incrs_profitable): New.
9769 (replace_profitable_candidates): Handle phi-dependent candidates.
9770 (analyze_candidates_and_replace): Likewise.
9771
9772 2013-05-03 Teresa Johnson <tejohnson@google.com>
9773
9774 PR bootstrap/57154
9775 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
9776 do not exceed REG_BR_PROB_BASE.
9777
9778 2013-05-03 Jeff Law <law@redhat.com>
9779
9780 PR tree-optimization/57144
9781 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
9782 operand of the condition will bit into the new type when eliminating
9783 a cast feeding a condition.
9784
9785 2013-05-03 Jakub Jelinek <jakub@redhat.com>
9786
9787 PR rtl-optimization/57130
9788 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
9789 of COMPARE as in_code to the recursive call if needed.
9790
9791 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
9792
9793 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
9794 (enabled): Handle new members.
9795 * config/i386/sse.md (*vec_concatv2si): Merge from
9796 *vec_concatv2si_sse2 and vec_concatv2si_sse.
9797 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
9798
9799 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
9800
9801 PR tree-optimization/57027
9802 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
9803 for fnms opportunity, check we got the prerequisite kind
9804 of tree / gimple before using accessor functions.
9805
9806 2013-05-03 Richard Biener <rguenther@suse.de>
9807
9808 * double-int.h (lshift): New overload without precision
9809 and arith argument.
9810 (operator *=, operator +=, operator -=): Move ...
9811 * double-int.c (operator *=, operator +=, operator -=): ... here
9812 and implement more efficiently.
9813 (mul_double_with_sign): Remove.
9814 (lshift_double): Adjust to take unsinged shift argument, push
9815 dispatching code to callers.
9816 (mul_double_wide_with_sign): Add early out for callers that
9817 are not interested in high parts or overflow.
9818 (lshift): New function.
9819 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
9820 dispatch code here.
9821 (lrotate, rrotate): Use logical shifts.
9822 * expr.c (get_inner_reference): Use lshift.
9823 * fixed-value.c (do_fixed_divide): Likewise.
9824 * tree-dfa.c (get_ref_base_and_extent): Likewise.
9825 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
9826 (indirect_refs_may_alias_p): Likewise.
9827 (stmt_kills_ref_p_1): Likewise.
9828
9829 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9830
9831 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
9832
9833 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
9834
9835 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
9836 scalar form of FABD instruction.
9837
9838 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
9839
9840 * lra-constraints.c (process_alt_operands): Add checking alt
9841 number to choose the best alternative.
9842
9843 2013-05-02 Richard Biener <rguenther@suse.de>
9844
9845 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
9846 bitmap and its handling.
9847 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
9848
9849 2013-05-02 Richard Biener <rguenther@suse.de>
9850
9851 PR middle-end/57140
9852 * tree-inline.c (copy_loops): Properly handle removed loops.
9853 (copy_cfg_body): Mark destination loops for fixup if source
9854 loops needed fixup.
9855
9856 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
9857
9858 PR target/56732
9859 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
9860 generating simple_return for naked functions.
9861
9862 2013-05-02 Martin Jambor <mjambor@suse.cz>
9863
9864 PR middle-end/56988
9865 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
9866 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
9867 flags match.
9868 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
9869 ipa_agg_replacement_value structures.
9870 (known_aggs_to_agg_replacement_list): Likewise.
9871 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
9872 (read_agg_replacement_chain): Likewise.
9873 (ipcp_transform_function): Also check that by_ref flags match.
9874
9875 2013-05-02 Richard Biener <rguenther@suse.de>
9876
9877 * graphds.h (struct graph): Add obstack member.
9878 * graphds.c (new_graph): Initialize obstack and allocate
9879 vertices from it.
9880 (add_edge): Allocate edge from the obstack.
9881 (free_graph): Free the obstack instead of all edges and vertices.
9882
9883 2013-05-02 Teresa Johnson <tejohnson@google.com>
9884
9885 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
9886 divides.
9887 * cfg.c (update_bb_profile_for_threading): Ditto.
9888 * tree-inline.c (copy_bb): Ditto.
9889 (copy_edges_for_bb): Ditto.
9890 (initialize_cfun): Ditto.
9891 (copy_cfg_body): Ditto.
9892 (expand_call_inline): Ditto.
9893 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
9894 (estimate_node_size_and_time): Ditto.
9895 (inline_merge_summary): Ditto.
9896 * cgraphclones.c (cgraph_clone_edge): Ditto.
9897 (cgraph_clone_node): Ditto.
9898 * sched-rgn.c (compute_dom_prob_ps): Ditto.
9899 (compute_trg_info): Ditto.
9900
9901 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9902
9903 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
9904 S reg when fp attribute set.
9905 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
9906
9907 2013-05-02 Ian Bolton <ian.bolton@arm.com>
9908
9909 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
9910 New pattern.
9911 (*and_one_cmplsi3_compare0_uxtw): Likewise.
9912 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
9913 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
9914
9915 2013-05-02 Richard Biener <rguenther@suse.de>
9916
9917 * tree-scalar-evolution.c (scev_info_hasher): Remove.
9918 (struct instantiate_cache_entry): New type.
9919 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
9920 (struct instantiate_cache_type): New type.
9921 (set_instantiated_value, get_instantiated_value): Remove.
9922 (get_instantiated_value_entry): New function.
9923 (instantiate_scev_name): Use the new cache and adjust.
9924 (instantiate_scev_poly): Adjust.
9925 (instantiate_scev_binary): Likewise.
9926 (instantiate_array_ref): Likewise.
9927 (instantiate_scev_convert): Likewise.
9928 (instantiate_scev_not): Likewise.
9929 (instantiate_scev_3): Likewise.
9930 (instantiate_scev_2): Likewise.
9931 (instantiate_scev_r): Likewise.
9932 (instantiate_scev): Likewise.
9933 (resolve_mixers): Likewise.
9934
9935 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
9936
9937 PR target/57091
9938 * lra-constraints.c (best_small_class_operands_num): Remove.
9939 (process_alt_operands): Remove small_class_operands_num. Take
9940 small classes operands into losers and only if the operand is not
9941 matched. Modify debugging output.
9942 (curr_insn_transform): Remove best_small_class_operands_num.
9943 Print insn name.
9944
9945 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9946
9947 * config/aarch64/aarch64-builtins.c
9948 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
9949 * config/aarch64/aarch64-simd-builtins.def
9950 (reduc_splus_): Add new modes.
9951 (reduc_uplus_): New.
9952 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
9953 (reduc_uplus_v4sf): Likewise.
9954 (reduc_splus_v4sf): Likewise.
9955 (aarch64_addv<mode>): Likewise.
9956 (reduc_uplus_<mode>): Likewise.
9957 (reduc_splus_<mode>): Likewise.
9958 (aarch64_addvv2di): Likewise.
9959 (reduc_uplus_v2di): Likewise.
9960 (reduc_splus_v2di): Likewise.
9961 (aarch64_addvv2si): Likewise.
9962 (reduc_uplus_v2si): Likewise.
9963 (reduc_splus_v2si): Likewise.
9964 (reduc_<sur>plus_<mode>): New.
9965 (reduc_<sur>plus_v2di): Likewise.
9966 (reduc_<sur>plus_v2si): Likewise.
9967 (reduc_<sur>plus_v4sf): Likewise.
9968 (aarch64_addpv4sf): Likewise.
9969 * config/aarch64/arm_neon.h
9970 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
9971 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
9972 add UNSPEC_SADDV, UNSPEC_UADDV.
9973 (SUADDV): New.
9974 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
9975
9976 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9977
9978 * config/aarch64/arm_neon.h
9979 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
9980
9981 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9982
9983 * config/aarch64/aarch64-builtins
9984 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
9985
9986 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
9987
9988 * config/aarch64/aarch64-simd-builtins.def
9989 (reduc_smax_): New.
9990 (reduc_smin_): Likewise.
9991 (reduc_umax_): Likewise.
9992 (reduc_umin_): Likewise.
9993 (reduc_smax_nan_): Likewise.
9994 (reduc_smin_nan_): Likewise.
9995 (fmax): Remove.
9996 (fmin): Likewise.
9997 (smax): Update for V2SF, V4SF and V2DF modes.
9998 (smin): Likewise.
9999 (smax_nan): New.
10000 (smin_nan): Likewise.
10001 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10002 (<su><maxmin><mode>3): ...This, refactor.
10003 (s<maxmin><mode>3): New.
10004 (<maxmin_uns><mode>3): Likewise.
10005 (reduc_<maxmin_uns>_<mode>): Refactor.
10006 (reduc_<maxmin_uns>_v4sf): Likewise.
10007 (reduc_<maxmin_uns>_v2si): Likewise.
10008 (aarch64_<fmaxmin><mode>: Remove.
10009 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10010 new builtin names.
10011 (vmin<q>_f<32,64>): Likewise.
10012 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10013 (FMAXMIN): New.
10014 (su): Add mappings for smax, smin, umax, umin.
10015 (maxmin): New.
10016 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10017 (FMAXMIN): Rename as...
10018 (FMAXMIN_UNS): ...This.
10019 (maxminv): Remove.
10020 (fmaxminv): Likewise.
10021 (fmaxmin): Likewise.
10022 (maxmin_uns): New.
10023 (maxmin_uns_op): Likewise.
10024
10025 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10026
10027 * config/aarch64/arm_neon.h
10028 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10029 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10030 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10031
10032 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10033
10034 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10035 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10036
10037 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10038
10039 * config/aarch64/aarch64-simd.md
10040 (vcond<mode>_internal): Handle special cases for constant masks.
10041 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10042 (vcondu<mode><mode>): Likewise.
10043 (vcond<v_cmp_result><mode>): New.
10044
10045 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10046
10047 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10048 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10049 * config/aarch64/aarch64-simd-builtins.def
10050 (cmeq): Update to BUILTIN_VALLDI.
10051 (cmgt): Likewise.
10052 (cmge): Likewise.
10053 (cmle): Likewise.
10054 (cmlt): Likewise.
10055 * config/aarch64/arm_neon.h
10056 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10057 to builtins or C as appropriate.
10058
10059 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10060
10061 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10062 (cmgeu): ...This.
10063 (cmhi): Rename to...
10064 (cmgtu): ...This.
10065 * config/aarch64/aarch64-simd.md
10066 (simd_mode): Add SF.
10067 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10068 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10069 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10070 (cstore<mode>_neg): ...This.
10071 * config/aarch64/iterators.md
10072 (VALLF): new.
10073 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10074 (COMPARISONS): New.
10075 (UCOMPARISONS): Likewise.
10076 (optab): Add missing comparisons.
10077 (n_optab): New.
10078 (cmp_1): Likewise.
10079 (cmp_2): Likewise.
10080 (CMP): Likewise.
10081 (cmp): Remove.
10082 (VCMP_S): Likewise.
10083 (VCMP_U): Likewise.
10084 (V_cmp_result): Add DF, SF modes.
10085 (v_cmp_result): Likewise.
10086 (v): Likewise.
10087 (vmtype): Likewise.
10088 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10089
10090 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10091
10092 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10093 define_insn to define_insn_and_split.
10094 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
10095 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
10096 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
10097 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
10098 (thumb2_negscc): Likewise.
10099
10100 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10101
10102 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
10103
10104 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10105
10106 * config/arm/thumb2.md: Remove trailing whitespaces.
10107
10108 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10109
10110 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
10111 Use gen_int_mode rather than GEN_INT.
10112
10113 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
10114
10115 * value-prof.c (stream_in_histogram_value): Remove the strayed
10116 debug_gimple_stmt.
10117
10118 2013-04-30 Richard Biener <rguenther@suse.de>
10119
10120 PR middle-end/57122
10121 * cfghooks.c (split_edge): Properly check for the loop latch edge.
10122
10123 2013-04-30 Richard Biener <rguenther@suse.de>
10124
10125 PR middle-end/57107
10126 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
10127
10128 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
10129
10130 PR rtl-optimization/56957
10131 PR rtl-optimization/57105
10132 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
10133 variable. Use just INSN_UID for determining whether an insn
10134 should be only disconnected from the insn stream.
10135 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
10136
10137 2013-04-30 Jakub Jelinek <jakub@redhat.com>
10138
10139 PR tree-optimization/57104
10140 * tsan.c (instrument_expr): Don't instrument accesses to
10141 DECL_HARD_REGISTER VAR_DECLs.
10142
10143 2013-04-30 Richard Biener <rguenther@suse.de>
10144
10145 * function.h (loops_for_fn): New inline function.
10146 (set_loops_for_fn): Likewise.
10147 * cfgloop.h (place_new_loop): Add struct function parameter.
10148 (get_loop): Likewise.
10149 (get_loops): Likewise.
10150 (number_of_loops): Likewise.
10151 (fel_next): Adjust.
10152 (fel_init): Likewise.
10153 * cfg.c (get_loop_copy): Adjust.
10154 * cfgloop.c (flow_loops_dump): Likewise.
10155 (record_loop_exits): Likewise.
10156 (verify_loop_structure): Likewise.
10157 * cfgloopanal.c (mark_irreducible_loops): Likewise.
10158 (estimate_reg_pressure_cost): Likewise.
10159 (mark_loop_exit_edges): Likewise.
10160 * cfgloopmanip.c (place_new_loop): Likewise.
10161 (add_loop): Likewise.
10162 (duplicate_loop): Likewise.
10163 * graph.c (draw_cfg_nodes): Likewise.
10164 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
10165 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
10166 (extract_affine_chrec): Likewise.
10167 (build_scop_iteration_domain): Likewise.
10168 * graphite.c (graphite_initialize): Likewise.
10169 * ira-build.c (create_loop_tree_nodes): Likewise.
10170 (more_one_region_p): Likewise.
10171 (rebuild_regno_allocno_maps): Likewise.
10172 (mark_loops_for_removal): Likewise.
10173 (mark_all_loops_for_removal): Likewise.
10174 (remove_unnecessary_regions): Likewise.
10175 (ira_build): Likewise.
10176 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
10177 * loop-init.c (fix_loop_structure): Likewise.
10178 (gate_rtl_move_loop_invariants): Likewise.
10179 (gate_rtl_unswitch): Likewise.
10180 (gate_rtl_unroll_and_peel_loops): Likewise.
10181 (rtl_doloop): Likewise.
10182 * lto-streamer-in.c (input_cfg): Likewise.
10183 * lto-streamer-out.c (output_cfg): Likewise.
10184 * modulo-sched.c (sms_schedule): Likewise.
10185 * predict.c (tree_estimate_probability): Likewise.
10186 (tree_estimate_probability_driver): Likewise.
10187 (estimate_loops): Likewise.
10188 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
10189 (move_sese_region_to_fn): Likewise.
10190 (debug_loop_num): Likewise.
10191 * tree-chrec.c (chrec_evaluate): Likewise.
10192 (hide_evolution_in_other_loops_than_loop): Likewise.
10193 (chrec_component_in_loop_num): Likewise.
10194 (reset_evolution_in_loop): Likewise.
10195 (evolution_function_is_invariant_rec_p): Likewise.
10196 * tree-if-conv.c (main_tree_if_conversion): Likewise.
10197 * tree-inline.c (copy_loops): Likewise.
10198 (copy_cfg_body): Likewise.
10199 (tree_function_versioning): Likewise.
10200 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
10201 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
10202 Likewise.
10203 (add_to_evolution_1): Likewise.
10204 (scev_const_prop): Likewise.
10205 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
10206 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
10207 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
10208 (tree_ssa_lim_initialize): Likewise.
10209 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
10210 (verify_loop_closed_ssa): Likewise.
10211 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
10212 (tree_ssa_loop_im): Likewise.
10213 (tree_ssa_loop_unswitch): Likewise.
10214 (tree_vectorize): Likewise.
10215 (check_data_deps): Likewise.
10216 (tree_ssa_loop_ivcanon): Likewise.
10217 (tree_ssa_loop_bounds): Likewise.
10218 (tree_complete_unroll): Likewise.
10219 (tree_complete_unroll_inner): Likewise.
10220 (tree_parallelize_loops): Likewise.
10221 (tree_ssa_loop_prefetch): Likewise.
10222 (tree_ssa_loop_ivopts): Likewise.
10223 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
10224 * tree-vectorizer.c (vectorize_loops): Likewise.
10225
10226 2013-04-29 Mike Frysinger <vapier@gentoo.org>
10227
10228 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
10229 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
10230 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
10231 with EABI_LINK_SPEC.
10232
10233 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10234
10235 PR target/44578
10236 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
10237 alternative.
10238
10239 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
10240
10241 PR target/57097
10242 * lra-constraints.c (process_alt_operands): Discourage a bit more
10243 using memory for pseudos. Print cost dump for alternatives.
10244 Modify cost values for conflicts with early clobbers.
10245 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
10246
10247 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10248
10249 PR target/57098
10250 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
10251
10252 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10253
10254 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
10255 from/to S register.
10256 (movdi_aarch64): Support LDR/STR from/to D register.
10257
10258 2013-04-29 Ian Bolton <ian.bolton@arm.com>
10259
10260 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
10261 or higher by default.
10262
10263 2013-04-29 Richard Biener <rguenther@suse.de>
10264
10265 PR middle-end/57075
10266 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
10267 even if not adding abnormal edges for calls that can make
10268 abnormal gotos.
10269
10270 2013-04-29 Richard Biener <rguenther@suse.de>
10271
10272 PR middle-end/57103
10273 * tree-cfg.c (move_stmt_op): Fix condition under which to update
10274 TREE_BLOCK.
10275 (move_stmt_r): Remove redundant checking.
10276
10277 2013-04-29 Teresa Johnson <tejohnson@google.com>
10278
10279 PR bootstrap/57077
10280 * basic-block.h (apply_scale): New function.
10281 (apply_probability): Use apply_scale.
10282 * gimple-streamer-in.c (input_bb): Ditto.
10283 * lto-streamer-in.c (input_cfg): Ditto.
10284 * lto-cgraph.c (merge_profile_summaries): Ditto.
10285 * tree-optimize.c (execute_fixup_cfg): Ditto.
10286 * tree-inline.c (copy_bb): Update comment to use apply_scale.
10287 (copy_edges_for_bb): Ditto.
10288 (copy_cfg_body): Ditto.
10289
10290 2013-04-29 Tom de Vries <tom@codesourcery.com>
10291
10292 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
10293 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
10294 (tail_merge_optimize): Handle current_loops == NULL.
10295
10296 2013-04-26 Jeff Law <law@redhat.com>
10297
10298 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
10299 (simplify_cond_using_ranges): Generalize code to simplify
10300 COND_EXPRs where one argument is a constant and the other
10301 is an SSA_NAME created by an integral type conversion.
10302
10303 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10304
10305 * config/arm/arm.md (store_minmaxsi): Use only when
10306 optimize_insn_for_size_p.
10307
10308 2013-04-29 Christian Bruel <christian.bruel@st.com>
10309
10310 PR target/57108
10311 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
10312
10313 2013-04-29 Richard Biener <rguenther@suse.de>
10314
10315 PR middle-end/57089
10316 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
10317 loop tree make sure to schedule a fixup for the child as well.
10318 (expand_omp_for_generic): Properly add loops.
10319 (expand_omp_for_static_nochunk): Likewise.
10320 (expand_omp_for_static_chunk): Likewise.
10321 (expand_omp_for): For the degenerate case fixup loops.
10322 (expand_omp_sections): Fix default bb placement in loops.
10323 (expand_omp_atomic_pipeline): Properly add loops.
10324
10325 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10326
10327 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
10328
10329 2013-04-29 Tom de Vries <tom@codesourcery.com>
10330
10331 * tree-ssa-tail-merge.c: Update header comment.
10332
10333 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10334
10335 * config/aarch64/arm_neon.h
10336 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
10337 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
10338 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
10339 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
10340 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
10341 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
10342 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
10343 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
10344
10345 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10346
10347 * config/aarch64/aarch64-simd.md
10348 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
10349 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
10350 fix_trunc, fixuns_trunc.
10351 (ftrunc<VDQF:mode>2): New.
10352 * config/aarch64/iterators.md (optab): Add fix, fixuns.
10353 (fix_trunc_optab): New.
10354
10355 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10356
10357 * config/aarch64/aarch64-builtins.c
10358 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
10359 iceilf, lround, iroundf.
10360
10361 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
10362
10363 PR target/54349
10364 * config/i386/i386.h (enum ix86_tune_indices)
10365 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
10366 New, split from X86_TUNE_INTER_UNIT_MOVES.
10367 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
10368 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
10369 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
10370 (TARGET_INTER_UNIT_MOVES): Remove.
10371 * config/i386/i386.c (initial_ix86_tune_features): Update.
10372 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
10373 (ix86_expand_convert_uns_didf_sse): Use
10374 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10375 (ix86_expand_vector_init_one_nonzero): Ditto.
10376 (ix86_expand_vector_init_interleave): Ditto.
10377 (inline_secondary_memory_needed): Return true for moves from SSE class
10378 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
10379 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
10380 * config/i386/constraints.md (Yi, Ym): Depend on
10381 TARGET_INTER_UNIT_MOVES_TO_VEC.
10382 (Yj, Yn): New constraints.
10383 * config/i386/i386.md (*movdi_internal): Change constraints of
10384 operand 1 from Yi to Yj and from Ym to Yn.
10385 (*movsi_internal): Ditto.
10386 (*movdf_internal): Ditto.
10387 (*movsf_internal): Ditto.
10388 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
10389 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
10390 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
10391 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
10392 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
10393 * config/i386/sse.md (movdi_to_sse): Ditto.
10394 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
10395 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
10396 TARGET_INTER_UNIT_MOVES.
10397 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
10398 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
10399 instead of TARGET_INTER_UNIT_MOVES.
10400 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
10401 operand 1 from Yi to Yj and from Ym to Yn.
10402
10403 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10404
10405 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
10406 (float_truncate_hi_): Likewise.
10407 (float_extend_lo_): Likewise.
10408 (float_truncate_lo_): Likewise.
10409 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
10410 (aarch64_float_extend_lo_v2df): Likewise.
10411 (vec_unpacks_hi_v4sf): Likewise.
10412 (aarch64_float_truncate_lo_v2sf): Likewise.
10413 (aarch64_float_truncate_hi_v4sf): Likewise.
10414 (vec_pack_trunc_v2df): Likewise.
10415 (vec_pack_trunc_df): Likewise.
10416
10417 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10418
10419 * config/aarch64/aarch64-builtins.c
10420 (aarch64_fold_builtin): Fold float conversions.
10421 * config/aarch64/aarch64-simd-builtins.def
10422 (floatv2si, floatv4si, floatv2di): New.
10423 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
10424 * config/aarch64/aarch64-simd.md
10425 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
10426 * config/aarch64/iterators.md (FLOATUORS): New.
10427 (optab): Add float, floatuns.
10428 (su_optab): Likewise.
10429
10430 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10431
10432 * config/aarch64/aarch64-builtins.c
10433 (aarch64_builtin_vectorized_function): Use new names for
10434 fcvt builtins.
10435 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
10436 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
10437 (fcvtzu): Split as...
10438 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
10439 (fcvtas): Split as...
10440 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
10441 (fcvtau): Split as...
10442 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
10443 (fcvtps): Split as...
10444 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
10445 (fcvtpu): Split as...
10446 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
10447 (fcvtms): Split as...
10448 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
10449 (fcvtmu): Split as...
10450 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
10451 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
10452 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
10453 (lfrintnusf, lfrintnudf): Likewise.
10454 * config/aarch64/aarch64-simd.md
10455 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
10456 define_insn.
10457 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
10458 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
10459 (fcvt_pattern): Likewise.
10460
10461 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10462
10463 * config/aarch64/aarch64-simd.md
10464 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
10465 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
10466
10467 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10468
10469 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
10470 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
10471 (vrnd<a,m,n,p>_f32): Implement using builtins.
10472 (vrnd<i,x><q>_f<32, 64>): New.
10473
10474 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
10475
10476 * config/aarch64/aarch64-builtins.c
10477 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
10478 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
10479 (frintz): Rename to...
10480 (btrunc): ...this.
10481 (frintp): Rename to...
10482 (ceil): ...this.
10483 (frintm): Rename to...
10484 (floor): ...this.
10485 (frinti): Rename to...
10486 (nearbyint): ...this.
10487 (frintx): Rename to...
10488 (rint): ...this.
10489 (frinta): Rename to...
10490 (round): ...this.
10491 * config/aarch64/aarch64-simd.md
10492 (aarch64_frint<frint_suffix><mode>): Delete.
10493 (<frint_pattern><mode>2): Convert to insn.
10494 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
10495 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
10496 (frint_pattern): Likewise.
10497 (frint_suffix): Likewise.
10498
10499 2013-04-29 Richard Biener <rguenther@suse.de>
10500
10501 PR tree-optimization/57081
10502 * loop-init.c: Include tree-flow.h.
10503 (loop_optimizer_finalize): Free number of iteration estimates.
10504 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
10505
10506 2013-04-29 Jakub Jelinek <jakub@redhat.com>
10507
10508 PR tree-optimization/57083
10509 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
10510 non-singleton shift count range, zero extend low_bound for uns case.
10511
10512 * config/i386/predicates.md (general_vector_operand): New predicate.
10513 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
10514 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
10515 if they aren't nonimmediate operands. If their original values
10516 satisfy const_vector_equal_evenodd_p, don't shift them.
10517 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
10518 predicates. For the SSE4.1 case force operands[{1,2}] into registers
10519 if not nonimmediate_operand.
10520 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
10521 instead of register_operand.
10522 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
10523
10524 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
10525
10526 * stor-layout.c (finalize_size_functions): Allocate a structure and
10527 reset cfun before dumping the functions.
10528
10529 2013-04-27 Jakub Jelinek <jakub@redhat.com>
10530
10531 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
10532
10533 PR target/56866
10534 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
10535 use xop_pmacsdqh if uns_p.
10536 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
10537 the immediate rotate count.
10538
10539 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
10540
10541 * rtl.h (struct rtx_def): Add comment for field jump.
10542 (LRA_SUBREG_P): New macro.
10543 * recog.c (register_operand): Check LRA_SUBREG_P.
10544 * lra.c (lra): Add note at the end of RTL code. Align non-empty
10545 stack frame.
10546 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
10547 (lra_final_code_change): Skip subreg change for operators.
10548 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
10549 if there are no operand changes.
10550 * lra-constraints.c (curr_insn_set): New.
10551 (match_reload): Set LRA_SUBREG_P.
10552 (emit_spill_move): Ditto.
10553 (check_and_process_move): Use curr_insn_set. Process only single
10554 set insns. Don't initialize sec_mem_p and change_p.
10555 (simplify_operand_subreg): Use LRA_SUBREG_P.
10556 (reg_in_class_p): New function.
10557 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
10558 of #ifdef. Add code to remove cycling.
10559 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
10560 non-null disp. Reload inner instead of disp when base and index
10561 are null. Try to put lo_sum into register.
10562 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
10563 (check_and_process_move): Move code for move cost check to
10564 simple_move_p. Remove equiv_substitution.
10565 (simple_move_p): New function.
10566 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
10567 curr_insn_set. Call check_and_process_move only for single set
10568 insns. Use the new function. Move call of check_and_process_move
10569 after operand equiv substitution and address process.
10570
10571 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10572
10573 PR go/57045
10574 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
10575 with nonlocal goto receivers or returns twice calls, ignore
10576 unininitialized values from abnormal edges to nl goto receiver
10577 or returns twice call.
10578
10579 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10580
10581 PR tree-optimization/57051
10582 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
10583 and VEC_RSHIFT_EXPR if shift count is a multiple of element
10584 bitsize.
10585
10586 2013-04-26 Richard Biener <rguenther@suse.de>
10587
10588 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
10589 (expand_omp_taskreg): Likewise. Mark loops for fixup.
10590 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
10591 (fixup_loop_arrays_after_move): New function.
10592 (move_sese_region_to_fn): Properly outline the loop tree parts
10593 of the SESE region.
10594
10595 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
10596
10597 * config/i386/i386.md (type, unit): Fix long lines.
10598
10599 2013-04-26 Richard Biener <rguenther@suse.de>
10600
10601 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
10602 (lto-streamer-out.o): Likewise.
10603 * cfgloop.c (init_loops_structure): Export, add struct function
10604 argument and adjust.
10605 (flow_loops_find): Adjust.
10606 * cfgloop.h (enum loop_estimation): Add EST_LAST.
10607 (init_loops_structure): Declare.
10608 * lto-streamer-in.c: Include cfgloop.h.
10609 (input_cfg): Input the loop tree.
10610 * lto-streamer-out.c: Include cfgloop.h.
10611 (output_cfg): Output the loop tree.
10612 (output_struct_function_base): Do not drop PROP_loops.
10613
10614 2013-03-26 Richard Biener <rguenther@suse.de>
10615
10616 * tree-cfg.c (execute_build_cfg): Build the loop tree.
10617 (pass_build_cfg): Provide PROP_loops.
10618 (move_sese_region_to_fn): Remove loops that are outlined into fn
10619 for now.
10620 * tree-inline.c: Include cfgloop.h.
10621 (initialize_cfun): Do not drop PROP_loops.
10622 (copy_loops): New function.
10623 (copy_cfg_body): Copy loop structure.
10624 (tree_function_versioning): Initialize destination loop tree.
10625 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
10626 (pass_parallelize_loops): Do IL verification.
10627 * loop-init.c (loop_optimizer_init): Fixup loops if required.
10628 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
10629 the CFG make sure we fixup loops as well.
10630 * tree-ssa-tail-merge.c: Include cfgloop.h.
10631 (replace_block_by): When merging loop latches mark loops for fixup.
10632 * lto-streamer-out.c (output_struct_function_base): Drop
10633 PROP_loops for now.
10634 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
10635 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
10636 * ipa-split.c: Include cfgloop.h.
10637 (split_function): Add the new return block to the loop tree root.
10638 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
10639 whether we have removed the forwarder block.
10640 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
10641 * cfgloop.h (place_new_loop): Declare.
10642 * cfgloopmanip.c (place_new_loop): Export.
10643 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
10644 (tree-switch-conversion.o): Likewise.
10645 (tree-complex.o): Likewise.
10646 (tree-inline.o): Likewise.
10647 (tree-ssa-tailmerge.o): Likewise.
10648 (ipa-split.o): Likewise.
10649 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
10650 (tree-ssa-copy.o): Likewise.
10651 * tree-switch-conversion.c: Include cfgloop.h
10652 (process_switch): If we emit a bit-test cascade, schedule loops
10653 for fixup.
10654 * tree-complex.c: Include cfgloop.h.
10655 (expand_complex_div_wide): Properly add new basic-blocks to loops.
10656 * asan.c: Include cfgloop.h.
10657 (create_cond_insert_point): Properly add new basic-blocks to
10658 loops, schedule loop fixup.
10659 * cfgloop.c (verify_loop_structure): Check that looks are not
10660 marked for fixup.
10661 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
10662 to loops.
10663 (expand_omp_for_generic): Likewise.
10664 (expand_omp_sections): Likewise.
10665 (expand_omp_atomic_pipeline): Schedule loops for fixup.
10666 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
10667 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
10668 is initialized, not when loops are present.
10669 * tree-parloops.c (parallelize_loops): Remove checking here.
10670 * passes.c (init_optimization_passes): Schedule a copy-propagation
10671 pass before complete unrolling of inner loops.
10672
10673 2013-04-26 Jakub Jelinek <jakub@redhat.com>
10674
10675 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
10676 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
10677 (colorize_init): Add argument to _WIN32 version.
10678 * toplev.c: Include diagnostic-color.h.
10679 (process_options): Default to -fdiagnostics-color=auto if
10680 GCC_COLORS env var is in the environment.
10681 * common.opt (fdiagnostics-color=): Add Var and Init.
10682 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
10683 env var is in the environment, the default is auto rather than never.
10684
10685 * diagnostic.h (file_name_as_prefix): Add context argument.
10686 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
10687 the string as locus.
10688 * langhooks.c (lhd_print_error_function): Adjust caller.
10689
10690 2013-04-25 Lawrence Crowl <crowl@google.com>
10691
10692 * var-tracking.c (shared_hash_def::htab):
10693 Change type to hash_table. Update dependent calls and types.
10694
10695 2013-04-25 Lawrence Crowl <crowl@google.com>
10696
10697 * Makefile.in: Update as needed below.
10698
10699 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
10700 Move declaration to after the type's method definitons.
10701
10702 * attribs.c (htab_t scoped_attributes::attribute_hash):
10703 Change type to hash_table. Update dependent calls and types.
10704
10705 * bitmap.c (htab_t bitmap_desc_hash):
10706 Change type to hash_table. Update dependent calls and types.
10707
10708 * cselib.c (htab_t cselib_hash_table):
10709 Change type to hash_table. Update dependent calls and types.
10710
10711 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
10712 (hash_string_slot_node): Move implementation into lto-streamer.h
10713 struct string_slot_hasher.
10714 (eq_string_slot_node): Likewise.
10715
10716 * data-streamer-out.c: Update output_block::string_hash_table
10717 dependent calls and types.
10718
10719 * dwarf2cfi.c (htab_t trace_index):
10720 Change type to hash_table. Update dependent calls and types.
10721
10722 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
10723 Change type to hash_table. Update dependent calls and types.
10724 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
10725 (htab_t optimize_external_refs::map): Likewise.
10726 (htab_t output_comp_unit::extern_map): Likewise.
10727 (htab_t output_comdat_type_unit::extern_map): Likewise.
10728 (htab_t output_macinfo::macinfo_htab): Likewise.
10729 (htab_t optimize_location_lists::htab): Likewise.
10730 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
10731
10732 * except.c (htab_t ehspec_hash_type):
10733 Change type to hash_table. Update dependent calls and types.
10734 (assign_filter_values::ttypes): Likewise.
10735 (assign_filter_values::ehspec): Likewise.
10736 (sjlj_assign_call_site_values::ar_hash): Likewise.
10737 (convert_to_eh_region_ranges::ar_hash): Likewise.
10738
10739 * gcse.c (htab_t pre_ldst_table):
10740 Change type to hash_table. Update dependent calls and types.
10741
10742 * ggc-common.c (htab_t saving_htab):
10743 Change type to hash_table. Update dependent calls and types.
10744 (htab_t loc_hash): Likewise.
10745 (htab_t ptr_hash): Likewise.
10746 (call_count): Rename ggc_call_count.
10747 (call_alloc): Rename ggc_call_alloc.
10748 (loc_descriptor): Rename make_loc_descriptor.
10749 (add_statistics): Rename ggc_add_statistics.
10750
10751 * ggc-common.c (saving_htab):
10752 Change type to hash_table. Update dependent calls and types.
10753
10754 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
10755 (push_gimplify_context): Likewise.
10756 (pop_gimplify_context): Likewise.
10757 (struct gimple_temp_hash_elt): Added.
10758 (struct gimplify_hasher): Likewise.
10759 (struct gimplify_ctx.temp_htab):
10760 Change type to hash_table. Update dependent calls and types.
10761
10762 * gimple-fold.c: Include gimplify-ctx.h.
10763
10764 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
10765 Change type to hash_table. Update dependent calls and types.
10766 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
10767 avoid potential global name collision.
10768
10769 * gimplify.c: Include gimplify-ctx.h.
10770 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
10771 (htab_t gimplify_ctx::temp_htab):
10772 Update dependent calls and types for new type hash_table.
10773 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
10774 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
10775
10776 * gimplify-ctx.h: New.
10777 (struct gimple_temp_hash_elt): Move from gimplify.c.
10778 (class gimplify_hasher): New.
10779 (struct gimplify_ctx): Move from gimple.h.
10780 (htab_t gimplify_ctx::temp_htab):
10781 Change type to hash_table. Update dependent calls and types.
10782
10783 * graphite-clast-to-gimple.c: Include graphite-htab.h.
10784 (htab_t ivs_params::newivs_index):
10785 Change type to hash_table. Update dependent calls and types.
10786 (htab_t ivs_params::params_index): Likewise.
10787 (htab_t print_generated_program::params_index): Likewise.
10788 (htab_t gloog::newivs_index): Likewise.
10789 (htab_t gloog::params_index): Likewise.
10790
10791 * graphite.c: Include graphite-htab.h.
10792 4htab_t graphite_transform_loops::bb_pbb_mapping):
10793 Change type to hash_table. Update dependent calls and types.
10794
10795 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
10796 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
10797 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
10798
10799 * graphite-dependences.c: Include graphite-htab.h.
10800 (loop_is_parallel_p): Change hash table type of parameter.
10801
10802 * graphite-htab.h: New.
10803 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
10804 (extern find_pbb_via_hash): Move from graphite-poly.h.
10805 (extern loop_is_parallel_p): Move from graphite-poly.h.
10806 (extern get_loop_body_pbbs): Move from graphite-poly.h.
10807
10808 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
10809 (extern loop_is_parallel_p): Move to graphite-htab.h.
10810 (extern get_loop_body_pbbs): Move to graphite-htab.h.
10811
10812 * haifa-sched.c (htab_t delay_htab):
10813 Change type to hash_table. Update dependent calls and types.
10814 (htab_t delay_htab_i2): Likewise.
10815
10816 * ira-color.c (htab_t allocno_hard_regs_htab):
10817 Change type to hash_table. Update dependent calls and types.
10818
10819 * ira-costs.c (htab_t cost_classes_htab):
10820 Change type to hash_table. Update dependent calls and types.
10821
10822 * loop-invariant.c (htab_t merge_identical_invariants::eq):
10823 Change type to hash_table. Update dependent calls and types.
10824
10825 * loop-iv.c (htab_t bivs):
10826 Change type to hash_table. Update dependent calls and types.
10827
10828 * loop-unroll.c (htab_t opt_info::insns_to_split):
10829 Change type to hash_table. Update dependent calls and types.
10830 (htab_t opt_info::insns_with_var_to_expand): Likewise.
10831
10832 * lto-streamer.h (struct string_slot): Move from data-streamer.h
10833 (struct string_slot_hasher): New.
10834 (htab_t output_block::string_hash_table):
10835 Change type to hash_table. Update dependent calls and types.
10836
10837 * lto-streamer-in.c (freeing_string_slot_hasher): New.
10838 (htab_t file_name_hash_table):
10839 Change type to hash_table. Update dependent calls and types.
10840
10841 * lto-streamer-out.c: Update output_block::string_hash_table dependent
10842 calls and types.
10843
10844 * lto-streamer.c (htab_t tree_htab):
10845 Change type to hash_table. Update dependent calls and types.
10846
10847 * omp-low.c: Include gimplify-ctx.h.
10848
10849 * passes.c (htab_t name_to_pass_map):
10850 Change type to hash_table. Update dependent calls and types.
10851 (pass_traverse): Rename to passes_pass_traverse.
10852
10853 * plugin.c (htab_t event_tab):
10854 Change type to hash_table. Update dependent calls and types.
10855
10856 * postreload-gcse.c (htab_t expr_table):
10857 Change type to hash_table. Update dependent calls and types.
10858 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
10859
10860 * sese.c (debug_rename_map_1): Make extern.
10861 (htab_t copy_bb_and_scalar_dependences::rename_map):
10862 Change type to hash_table. Update dependent calls and types.
10863
10864 * sese.h (extern debug_rename_map): Move to .c file.
10865
10866 * store-motion.c (htab_t store_motion_mems_table):
10867 Change type to hash_table. Update dependent calls and types.
10868
10869 * trans-mem.c (htab_t tm_new_mem_hash):
10870 Change type to hash_table. Update dependent calls and types.
10871
10872 * tree-browser.c (htab_t TB_up_ht):
10873 Change type to hash_table. Update dependent calls and types.
10874
10875 * tree-cfg.c (htab_t discriminator_per_locus):
10876 Change type to hash_table. Update dependent calls and types.
10877
10878 * tree-complex.c: Include tree-hasher.h
10879 (htab_t complex_variable_components):
10880 Change type to hash_table. Update dependent calls and types.
10881
10882 * tree-eh.c (htab_t finally_tree):
10883 Change type to hash_table. Update dependent calls and types.
10884
10885 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
10886 struct int_tree_hasher.
10887 (extern int_tree_map_eq): Likewise.
10888 (uid_decl_map_hash): Removed.
10889 (extern decl_tree_map_eq): Likewise.
10890
10891 * tree-hasher.h: New.
10892 (struct int_tree_hasher): New.
10893 (typedef int_tree_htab_type): New.
10894
10895 * tree-inline.c: Include gimplify-ctx.h.
10896
10897 * tree-mudflap.c: Include gimplify-ctx.h.
10898
10899 * tree-parloops.c: Include tree-hasher.h.
10900 (htab_t eliminate_local_variables_stmt::decl_address):
10901 Change type to hash_table. Update dependent calls and types.
10902 (htab_t separate_decls_in_region::decl_copies): Likewise.
10903
10904 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
10905 Change type to hash_table. Update dependent calls and types.
10906
10907 * tree-sra.c (candidates):
10908 Change type to hash_table. Update dependent calls and types.
10909
10910 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
10911 in tree-flow.h.
10912 (int_tree_map_hash): Likewise.
10913
10914 * tree-ssa-dom.c (htab_t avail_exprs):
10915 Change type to hash_table. Update dependent calls and types.
10916
10917 * tree-ssa-live.c (var_map_base_init::tree_to_index):
10918 Change type to hash_table. Update dependent calls and types.
10919
10920 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
10921 Change type to hash_table. Update dependent calls and types.
10922
10923 * tree-ssa-phiopt.c (seen_ssa_names):
10924 Change type to hash_table. Update dependent calls and types.
10925
10926 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
10927 Change type to hash_table. Update dependent calls and types.
10928
10929 * tree-ssa-uncprop.c (equiv):
10930 Change type to hash_table. Update dependent calls and types.
10931
10932 2013-04-25 Jakub Jelinek <jakub@redhat.com>
10933
10934 PR rtl-optimization/57003
10935 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
10936 call note_stores with kill_clobbered_value callback again after
10937 killing regs_invalidated_by_call.
10938
10939 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10940
10941 * config/aarch64/aarch64-simd.md
10942 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
10943 (aarch64_simd_bsl<mode>): Likewise.
10944 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
10945
10946 2013-04-25 Marek Polacek <polacek@redhat.com>
10947
10948 PR tree-optimization/57066
10949 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
10950
10951 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
10952
10953 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
10954
10955 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10956
10957 * config/aarch64/aarch64-builtins.c
10958 (aarch64_fold_builtin): New.
10959 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
10960 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
10961 * config/aarch64/aarch64-simd-builtins.def (abs): New.
10962 * config/aarch64/arm_neon.h
10963 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
10964
10965 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
10966 Tejas Belagod <tejas.belagod@arm.com>
10967
10968 * config/aarch64/aarch64-builtins.c
10969 (aarch64_gimple_fold_builtin): New.
10970 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
10971 * config/aarch64/aarch64-simd-builtins.def (addv): New.
10972 * config/aarch64/aarch64-simd.md (addpv4sf): New.
10973 (addvv4sf): Update.
10974 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
10975
10976 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10977
10978 * config/aarch64/aarch64.md
10979 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
10980
10981 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10982
10983 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
10984 (*ngcsi_uxtw): New pattern.
10985
10986 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10987 Julian Brown <julian@codesourcery.com>
10988
10989 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
10990 (TB_DREG): Add T_V4HF.
10991 (v4hf_UP): New macro.
10992 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10993 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10994 Handle initialisation of V4HF. Adjust initialisation of reinterpret
10995 built-ins.
10996 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
10997 (arm_vector_mode_supported_p): Handle V4HF.
10998 (arm_mangle_map): Handle V4HFmode.
10999 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11000 * config/arm/arm_neon_builtins.def: Add entries for
11001 vcvtv4hfv4sf, vcvtv4sfv4hf.
11002 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11003 (neon_vcvtv4hfv4sf): Likewise.
11004 * config/arm/neon-gen.ml: Handle half-precision floating point
11005 features.
11006 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11007 * config/arm/arm_neon.h: Regenerate.
11008 * config/arm/neon.ml (type elts): Add F16.
11009 (type vectype): Add T_float16x4, T_floatHF.
11010 (type vecmode): Add V4HF.
11011 (type features): Add Requires_FP_bit feature.
11012 (elt_width): Handle F16.
11013 (elt_class): Likewise.
11014 (elt_of_class_width): Likewise.
11015 (mode_of_elt): Refactor.
11016 (type_for_elt): Handle F16, fix error messages.
11017 (vectype_size): Handle T_float16x4.
11018 (vcvt_sh): New function.
11019 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11020 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11021 (string_of_mode): Handle V4HF.
11022 * doc/arm-neon-intrinsics.texi: Regenerate.
11023
11024 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11025
11026 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11027 format specifier in 'X' case.
11028
11029 2013-04-25 Alan Modra <amodra@gmail.com>
11030
11031 PR target/57052
11032 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11033 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11034 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11035 Repeat for many other rotate/shift and mask patterns using subregs.
11036 Name lshiftrt insns.
11037 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11038 on WORDS_BIG_ENDIAN.
11039
11040 2013-04-25 Alan Modra <amodra@gmail.com>
11041
11042 * config.gcc: Support little-endian powerpc-linux targets.
11043 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11044 (LINK_OS_LINUX_SPEC): Define.
11045 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11046 Preserve MASK_LITTLE_ENDIAN.
11047 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11048 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11049 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11050 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11051 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11052 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11053 Correct fp word order for little-endian. Don't shift toc entries
11054 smaller than a word for little-endian.
11055 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11056 (bswapdi2 splits): Correct low-part subreg for little-endian.
11057 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11058 low/high where such is correct only for be.
11059 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11060 little-endian for -mcall-aixdesc.
11061
11062 2013-04-25 Alan Modra <amodra@gmail.com>
11063
11064 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11065 replace_equiv_address_nv.
11066
11067 2013-04-25 Alan Modra <amodra@gmail.com>
11068
11069 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11070
11071 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11072
11073 Revert:
11074 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11075 * rtl.h (struct rtx_def): ...
11076
11077 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11078
11079 PR rtl-optimizations/57046
11080 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11081 for multi-reg splits.
11082
11083 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11084
11085 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11086
11087 2013-04-24 Sterling Augustine <saugustine@google.com>
11088
11089 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11090 (comp_dir_string, debug_str_dwo_section): New.
11091 (DEBUG_STR_DWO_SECTION): Rename to ...
11092 (DEBUG_DWO_STR_SECTION): ... this.
11093 (DEBUG_NORM_STR_SECTION): Delete.
11094 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
11095 (DEBUG_STR_DWO_SECTION_FLAGS): New.
11096 (find_AT_string): Move most logic to ...
11097 (find_AT_string_in_table): ... here. New.
11098 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
11099 add_skeleton_AT_string. Delete logic.
11100 (output_skeleton_debug_sections): Remove call to
11101 add_top_level_skeleton_die_attrs.
11102 (add_comp_dir_attribute): Move logic to comp_dir_string.
11103 (dwarf2out_init): Initialize debug_str_dwo_section.
11104 (output_indirect_string): Call find_string_form.
11105 (output_indirect_strings): Rewrite.
11106 (prune_unused_types): Empty skeleton_debug_str_hash.
11107 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
11108 (dwarf2out_finish): Call output_indirect_strings.
11109
11110 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11111
11112 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
11113
11114 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11115
11116 * rtl.h (struct rtx_def): Add comment for field jump.
11117 (LRA_SUBREG_P): New macro.
11118 * recog.c (register_operand): Check LRA_SUBREG_P.
11119 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11120 stack frame.
11121 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11122 (lra_final_code_change): Skip subreg change for operators.
11123 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11124 if there are no operand changes.
11125 * lra-constraints.c (curr_insn_set): New.
11126 (match_reload): Set LRA_SUBREG_P.
11127 (emit_spill_move): Ditto.
11128 (check_and_process_move): Use curr_insn_set. Process only single
11129 set insns. Don't initialize sec_mem_p and change_p.
11130 (simplify_operand_subreg): Use LRA_SUBREG_P.
11131 (reg_in_class_p): New function.
11132 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11133 of #ifdef. Add code to remove cycling.
11134 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11135 non-null disp. Reload inner instead of disp when base and index
11136 are null. Try to put lo_sum into register.
11137 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11138 (check_and_process_move): Move code for move cost check to
11139 simple_move_p. Remove equiv_substitution.
11140 (simple_move_p): New function.
11141 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11142 curr_insn_set. Call check_and_process_move only for single set
11143 insns. Use the new function. Move call of check_and_process_move
11144 after operand equiv substitution and address process.
11145
11146 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
11147
11148 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
11149 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
11150 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
11151
11152 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11153
11154 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
11155
11156 2013-04-24 Marek Polacek <polacek@redhat.com>
11157
11158 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
11159 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
11160 (select_loops_exit_conditions): Likewise.
11161 (number_of_iterations_for_all_loops): Likewise.
11162 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
11163 (scev_analysis): Likewise.
11164
11165 2013-04-02 Catherine Moore <clm@codesourcery.com>
11166 Chao-ying Fu <fu@mips.com>
11167
11168 * config/mips/micromips.md (jraddiusp): New pattern.
11169 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
11170 instruction if possible.
11171
11172 2013-04-24 Alan Modra <amodra@gmail.com>
11173
11174 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
11175
11176 2013-04-24 Julian Brown <julian@codesourcery.com>
11177 Chung-Lin Tang <cltang@codesourcery.com>
11178
11179 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
11180 dependency behavior in enumeration type DIE generation. Add TODO note
11181 to comments about future DW_FORM_sdata/udata re-work of related code.
11182
11183 2013-04-23 Lawrence Crowl <crowl@google.com>
11184
11185 * Makefile.in: Update as needed below.
11186
11187 * hash-table.h (class hash_table):
11188 Correct many methods with parameter types compare_type to the correct
11189 value_type. (Correct code was unlikely to notice the change.)
11190 (hash_table::elements_with_deleted) New.
11191 (class hashtable::iterator): New.
11192 (hashtable::begin()): New.
11193 (hashtable::end()): New.
11194 (FOR_EACH_HASH_TABLE_ELEMENT): New.
11195
11196 * statistics.c (statistics_hashes):
11197 Change type to hash_table. Update dependent calls and types.
11198
11199 * tree-into-ssa.c (var_infos):
11200 Change type to hash_table. Update dependent calls and types.
11201
11202 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
11203 Change type to hash_table. Update dependent calls and types.
11204
11205 * tree-ssa-loop-im.c (struct mem_ref.refs):
11206 Change type to hash_table. Update dependent calls and types.
11207
11208 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
11209 Change type to hash_table. Update dependent calls and types.
11210
11211 * tree-ssa-sccvn.c (vn_tables_s::nary):
11212 Change type to hash_table. Update dependent calls and types.
11213 (vn_tables_s::phis): Likewise.
11214 (vn_tables_s::references): Likewise.
11215
11216 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
11217 (vn_reference_eq): Update parameter and return types.
11218
11219 * tree-ssa-structalias.c (pointer_equiv_class_table):
11220 Change type to hash_table. Update dependent calls and types.
11221 (location_equiv_class_table): Likewise.
11222
11223 * tree-vect-data-refs.c: Consequential changes for making
11224 peeling a hash_table.
11225
11226 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
11227 (destroy_loop_vec_info): Dependent hash_table update.
11228
11229 * tree-vectorizer.h (peeling_htab):
11230 Change type to hash_table. Update dependent calls and types.
11231
11232 2013-04-23 Shiva Chen <shiva0217@gmail.com>
11233
11234 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
11235 to check the register content is equal or not.
11236 * lra-constraints.c (match_reload): Use lra_assign_reg_val
11237 to assign register content record.
11238 * lra-eliminations.c (update_reg_eliminate): Use
11239 lra_update_reg_val_offset to update register content offset.
11240 * lra-int.h (struct lra_reg): Add offset member.
11241 (lra_reg_val_equal_p): New static inline function.
11242 (lra_update_reg_val_offset): New static inline function.
11243 (lra_assign_reg_val): New static inline function.
11244 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
11245 to assign register content record.
11246 (initialize_lra_reg_info_element): Initial offset to zero.
11247
11248 2013-04-23 Catherine Moore <clm@codesourcery.com>
11249
11250 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
11251 operands. Record compression.
11252
11253 2013-04-23 Xinliang David Li <davidxl@google.com>
11254
11255 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
11256
11257 2013-04-23 Richard Biener <rguenther@suse.de>
11258
11259 PR middle-end/57036
11260 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
11261 parameter, only add abnormal goto edges from the copied body
11262 if the call could perform abnormal gotos.
11263 (copy_cfg_body): Adjust.
11264
11265 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
11266
11267 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
11268
11269 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
11270
11271 * coretypes.h (gimple_stmt_iterator): Add struct to make
11272 compatible with C.
11273
11274 2013-04-23 Richard Biener <rguenther@suse.de>
11275
11276 PR tree-optimization/57026
11277 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
11278 from SSA names occuring in abnormal PHI nodes.
11279
11280 2013-04-22 Andi Kleen <ak@linux.intel.com>
11281
11282 * lto/lto.c (print_lto_report_1): Fix LTO report names.
11283
11284 2013-04-22 Andi Kleen <ak@linux.intel.com>
11285
11286 * lto/lto.c (print_lto_report_1): Declare early.
11287 (read_cgraph_and_symbols): Call print_lto_report_1 early.
11288
11289 2013-04-22 Andi Kleen <ak@linux.intel.com>
11290
11291 * common.opt (-flto-report-wpa): Add.
11292 * doc/invoke.texi (-flto-report-wpa): Add.
11293 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
11294 (lto_main): dito.
11295
11296 2013-04-22 Xinliang David Li <davidxl@google.com>
11297
11298 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
11299 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
11300 * Makefile.in: New dependency
11301
11302 David Daney <ddaney.cavm@gmail.com>
11303
11304 * configure.ac (gcc_cv_as_micromips_support): Use the
11305 --fatal-warnings option.
11306 * configure: Regenerate.
11307
11308 2013-04-22 Marek Polacek <polacek@redhat.com>
11309
11310 PR sanitizer/56990
11311 * tsan.c (instrument_expr): Don't instrument expression
11312 in case its size is zero.
11313
11314 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
11315
11316 PR target/57032
11317 Revert:
11318 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
11319
11320 * config/alpha/alpha.c (TARGET_LRA_P): New define.
11321
11322 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11323
11324 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
11325 (gimple_stmt_iterator): New typedef.
11326 * gimple.h (gimple_stmt_iterator): Rename to...
11327 (gimple_stmt_iterator_d): ... This.
11328 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
11329 trees be valid for GIMPLE and GENERIC.
11330 (TARGET_GIMPLE_FOLD_BUILTIN): New.
11331 * gimple-fold.c (gimple_fold_call): Call target hook
11332 gimple_fold_builtin.
11333 * hooks.c (hook_bool_gsiptr_false): New.
11334 * hooks.h (hook_bool_gsiptr_false): New.
11335 * target.def (fold_stmt): New.
11336 * doc/tm.texi: Regenerate.
11337
11338 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
11339
11340 PR target/57018
11341 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
11342 a set sp if no stack realignment.
11343
11344 2013-04-22 Nick Clifton <nickc@redhat.com>
11345
11346 * config.gcc (tilegx-linux): Extend extra_objs rather than
11347 overwriting it.
11348 (tilepro-linux): Likewise.
11349
11350 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11351
11352 * config/aarch64/aarch64-builtins.c
11353 (CF): Remove.
11354 (CF0, CF1, CF2, CF3, CF4, CF10): New.
11355 (VAR<1-12>): Add MAP parameter.
11356 (BUILTIN_*): Likewise.
11357 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
11358 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
11359 (aarch64_ushl_n<mode>): Likewise.
11360 (aarch64_sshr_n<mode>): Likewise.
11361 (aarch64_ushr_n<mode>): Likewise.
11362 (aarch64_<maxmin><mode>): Likewise.
11363 (aarch64_sqrt<mode>): Likewise.
11364 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
11365 (vshr<q>_n_*): Likewise.
11366
11367 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
11368
11369 * config/aarch64/aarch64-builtins.c
11370 (aarch64_simd_builtin_type_mode): Handle SF types.
11371 (sf_UP): Define.
11372 (BUILTIN_GPF): Define.
11373 (aarch64_init_simd_builtins): Handle SF types.
11374 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
11375 (frecps): Likewise.
11376 (frecpx): Likewise.
11377 * config/aarch64/aarch64-simd.md
11378 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
11379 (aarch64_frecpe<mode>): New.
11380 (aarch64_frecps<mode>): Likewise.
11381 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
11382 (v8type): Add frecp<esx>.
11383 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
11384 (aarch64_frecps<mode>): Likewise.
11385 * config/aarch64/iterators.md (FRECP): New.
11386 (frecp_suffix): Likewise.
11387 * config/aarch64/arm_neon.h
11388 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
11389
11390 2013-04-22 Christian Bruel <christian.bruel@st.com>
11391
11392 PR target/56995
11393 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
11394 (REG_CLASS_NAMES): Idem.
11395 (REG_CLASS_CONTENTS): Idem.
11396 (REGCLASS_HAS_FP_REG): Idem.
11397 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
11398 (sh_conditional_register_usage): Idem.
11399
11400 2013-04-21 Jeff Law <law@redhat.com>
11401
11402 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
11403 (ssa_forward_propagate_and_combine): Use it.
11404
11405 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11406
11407 * lra.c: Update the flow chart diagram.
11408
11409 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
11410
11411 PR rtl-optimization/56847
11412 * lra-constraints.c (process_alt_operands): Discourage alternative
11413 with non-matche doffsettable memory constraint fro memory with
11414 known offset.
11415
11416 2013-04-19 Richard Biener <rguenther@suse.de>
11417
11418 PR tree-optimization/56982
11419 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
11420 function.
11421 * gimplify.c (gimplify_call_expr): Notice special calls.
11422 (gimplify_modify_expr): Likewise.
11423 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
11424 abnormal control flow receivers.
11425 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
11426 in the same way as cfun->has_nonlocal_labels.
11427 (gimple_purge_dead_abnormal_call_edges): Likewise.
11428 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
11429 receivers start a basic-block.
11430
11431 2013-04-19 Richard Biener <rguenther@suse.de>
11432
11433 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
11434 member ...
11435 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
11436 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
11437 (SLP_TREE_LOAD_PERMUTATION): Add.
11438 (vect_transform_slp_perm_load): Adjust prototype.
11439 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
11440 (vect_free_slp_instance): Likewise.
11441 (vect_create_new_slp_node): Likewise.
11442 (vect_supported_slp_permutation_p): Remove.
11443 (vect_slp_rearrange_stmts): Adjust.
11444 (vect_supported_load_permutation_p): Likewise. Inline
11445 vect_supported_slp_permutation_p here.
11446 (vect_analyze_slp_instance): Compute load permutations per
11447 slp node instead of per instance.
11448 (vect_get_slp_defs): Adjust.
11449 (vect_transform_slp_perm_load): Likewise.
11450 (vect_schedule_slp_instance): Remove redundant code.
11451 (vect_schedule_slp): Remove hack for PR56270, add it ...
11452 * tree-vect-stmts.c (vectorizable_load): ... here, do not
11453 CSE loads for SLP. Adjust.
11454
11455 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11456
11457 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
11458 spelling in two comments.
11459
11460 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
11461
11462 PR target/56797
11463 * config/arm/arm.c (load_multiple_sequence): Require SP
11464 as base register for loads if SP is in the register list.
11465
11466 2013-04-19 Martin Jambor <mjambor@suse.cz>
11467
11468 PR tree-optimization/56718
11469 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
11470 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
11471 and made public. Adjusted all callers.
11472 (ipa_intraprocedural_devirtualization): New function.
11473 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
11474 (ipa_intraprocedural_devirtualization): Likewise.
11475 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
11476
11477 2013-04-19 Richard Biener <rguenther@suse.de>
11478
11479 PR tree-optimization/57000
11480 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
11481
11482 2013-04-19 Terry Guo <terry.guo@arm.com>
11483
11484 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
11485 Replace with ...
11486 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
11487 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
11488 (cortex_m4_fmacs): Use new reservations.
11489 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
11490
11491 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
11492
11493 PR rtl-optimization/56999
11494 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
11495 related code.
11496 (lra_coalesce): Remove split_origin_bitmap and related code.
11497 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
11498 ranges if necessary.
11499
11500 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
11501
11502 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
11503 New array.
11504 (ix86_expand_call): Remove clobbered_registers array and use
11505 x86_64_ms_sysv_extra_clobbered_registers instead.
11506 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
11507 Declare here.
11508 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
11509 predicate.
11510 * config/i386/i386.md (*call_rex64_ms_sysv): Use
11511 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
11512 (*call_value_rex64_ms_sysv): Ditto.
11513
11514 2013-04-18 Cary Coutant <ccoutant@google.com>
11515
11516 * dwarf2out.c (output_pubnames): Check die_perennial_p of
11517 parent instead of die_mark.
11518
11519 2013-04-18 Diego Novillo <dnovillo@google.com>
11520
11521 * gimple.c (create_gimple_tmp): New.
11522 (get_expr_type): New.
11523 (build_assign): New.
11524 (build_type_cast): New.
11525 * gimple.h (enum ssa_mode): Define.
11526 (gimple_seq_set_location): New.
11527 * asan.c (build_check_stmt): Change some gimple_build_* calls
11528 to use build_assign and build_type_cast.
11529
11530 2013-04-18 Richard Biener <rguenther@suse.de>
11531
11532 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
11533 handle negative step. Remove redundant checks.
11534 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
11535 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
11536 for negative step and grouped loads fail to vectorize.
11537
11538 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11539
11540 * emit-rtl.c (reset_insn_used_flags): New function.
11541 (reset_all_used_flags): Use it.
11542 (verify_insn_sharing): New function.
11543 (verify_rtl_sharing): Fix verification for SEQUENCEs.
11544
11545 2013-04-18 Jakub Jelinek <jakub@redhat.com>
11546
11547 PR tree-optimization/56984
11548 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
11549 and (x >> M) >= N don't register any assertion if N << M is the
11550 minimum value.
11551
11552 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
11553
11554 * lower-subreg.c (resolve_simple_move): If called self-recursive,
11555 do not delete_insn insns that have not yet been emitted, only
11556 unlink them with remove_insn.
11557 * df-scan.c (df_insn_delete): Revert r197492.
11558
11559 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
11560
11561 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
11562 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
11563
11564 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11565
11566 * config/arm/arm.md (movsicc_insn): Convert define_insn into
11567 define_insn_and_split.
11568 (and_scc,ior_scc,negscc): Likewise.
11569 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
11570
11571 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
11572
11573 * config/arm/arm.c (use_return_insn): Return 0 for targets that
11574 can benefit from using a sequence of LDRD instructions in epilogue
11575 instead of a single LDM instruction.
11576
11577 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
11578
11579 PR 45688
11580 * doc/extend.texi: Fix typo.
11581
11582 2013-04-17 Richard Biener <rguenther@suse.de>
11583
11584 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
11585 (vect_build_slp_tree): ... here.
11586 (vect_build_slp_tree_1): Compute which stmts of the SLP group
11587 match. Remove special-casing of mismatched complex loads.
11588 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
11589 re-try the match with swapped commutative operands.
11590 (vect_supported_load_permutation_p): Remove special-casing of
11591 mismatched complex loads.
11592 (vect_analyze_slp_instance): Adjust.
11593
11594 2013-04-17 Richard Biener <rguenther@suse.de>
11595
11596 PR rtl-optimization/56921
11597 * cfgloop.h (struct loop): Add simple_loop_desc member.
11598 (struct niter_desc): Mark with GTY(()).
11599 (simple_loop_desc): Do not use aux field but simple_loop_desc.
11600 * loop-iv.c (get_simple_loop_desc): Likewise.
11601 (free_simple_loop_desc): Likewise.
11602
11603 Revert
11604 2013-04-16 Richard Biener <rguenther@suse.de>
11605
11606 PR rtl-optimization/56921
11607 * loop-init.c (pass_rtl_move_loop_invariants): Add
11608 TODO_do_not_ggc_collect to todo_flags_finish.
11609 (pass_rtl_unswitch): Same.
11610 (pass_rtl_unroll_and_peel_loops): Same.
11611 (pass_rtl_doloop): Same.
11612
11613 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
11614
11615 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
11616 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
11617 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
11618 references.
11619 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
11620 * tree-streamer.c (record_common_node): Adjust reference in comment.
11621
11622 2013-04-17 Terry Guo <terry.guo@arm.com>
11623
11624 * config/arm/cortex-m4.md: Add a new bypass.
11625
11626 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11627
11628 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
11629 New pattern.
11630 (*subs_<optab><mode>_multp2): New pattern.
11631 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11632 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
11633
11634 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11635
11636 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
11637 (*subs_mul_imm_<mode>): New pattern.
11638
11639 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
11640
11641 PR target/56948
11642 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
11643 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
11644 (vsx_movti_32bit): Change j->wa to O->wa.
11645
11646 2013-04-16 Richard Biener <rguenther@suse.de>
11647
11648 PR rtl-optimization/56921
11649 * loop-init.c (pass_rtl_move_loop_invariants): Add
11650 TODO_do_not_ggc_collect to todo_flags_finish.
11651 (pass_rtl_unswitch): Same.
11652 (pass_rtl_unroll_and_peel_loops): Same.
11653 (pass_rtl_doloop): Same.
11654
11655 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
11656
11657 * config/arm/arm.c (emit_multi_reg_push): New declaration
11658 for an existing function.
11659 (arm_emit_strd_push): New function.
11660 (arm_expand_prologue): Used here.
11661 (arm_emit_ldrd_pop): New function.
11662 (arm_expand_epilogue): Used here.
11663 (arm_get_frame_offsets): Update condition.
11664 (arm_emit_multi_reg_pop): Add a special case for load of a single
11665 register with writeback.
11666
11667 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
11668
11669 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
11670 description.
11671
11672 2013-04-16 Richard Biener <rguenther@suse.de>
11673
11674 PR tree-optimization/56756
11675 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
11676 (first_mem_ref_loc): New.
11677 (execute_sm): Place the load temporarily before a previous
11678 access instead of in the latch edge to ensure its SSA dependencies
11679 are defined at points dominating the load.
11680
11681 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
11682
11683 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
11684 correct fix by moving header and footer insn to the footer of
11685 the merged basic block. Clear BB_END of the merged-away block.
11686
11687 PR middle-end/43631
11688 * emit-rtl.c (make_note_raw): New function.
11689 (link_insn_into_chain): New static inline function.
11690 (add_insn): Use it.
11691 (add_insn_before, add_insn_after): Factor insn chain linking code...
11692 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
11693 using link_insn_into_chain.
11694 (note_outside_basic_block_p): New helper function for emit_note_after
11695 and emit_note_before.
11696 (emit_note_after): Use nobb variant of add_insn_after if the note
11697 should not be contained in a basic block.
11698 (emit_note_before): Use nobb variant of add_insn_before if the note
11699 should not be contained in a basic block.
11700 (emit_note_copy): Use make_note_raw.
11701 (emit_note): Likewise.
11702 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
11703 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
11704 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
11705 the moved barrier the tail of the basic block it follows.
11706 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
11707
11708 2013-04-15 Jakub Jelinek <jakub@redhat.com>
11709
11710 PR tree-optimization/56962
11711 * gimple-ssa-strength-reduction.c (record_increment): Only set
11712 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
11713 either rhs1 or rhs2 is equal to c->base_expr.
11714
11715 2013-04-15 Richard Biener <rguenther@suse.de>
11716
11717 PR tree-optimization/56933
11718 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
11719 member.
11720 (GROUP_READ_WRITE_DEPENDENCE): Remove.
11721 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
11722 * tree-vect-data-refs.c (vect_analyze_group_access): Move
11723 dependence check ...
11724 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
11725 ... here.
11726 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
11727 GROUP_READ_WRITE_DEPENDENCE.
11728
11729 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11730
11731 * emit-rtl.c (reset_all_used_flags): New function.
11732 (verify_rtl_sharing): Call reset_all_used_flags before and after
11733 performing the checks.
11734
11735 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11736
11737 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
11738 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
11739 * config/arm/constraints.md (De): New constraint.
11740 * config/arm/neon.md (anddi3_neon): Delete.
11741 (neon_vand<mode>): Expand to standard anddi3 pattern.
11742 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
11743 Move earlier in the file.
11744 (neon_inv_logic_op2): Likewise.
11745 (arm_anddi_operand_neon): New predicate.
11746
11747 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11748
11749 * configure.ac (gcc_cv_ld_as_needed): Set
11750 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
11751 Use -z ignore, -z record on *-*-solaris2*.
11752 (HAVE_LD_AS_NEEDED): Update comment.
11753 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
11754 * configure: Regenerate.
11755 * config.in: Regenerate.
11756 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
11757 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
11758 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
11759 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
11760 equivalents. Fix markup.
11761 * doc/tm.texi: Regenerate.
11762
11763 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
11764
11765 * config/i386/i386.opt: New option mstack-protector-guard=.
11766 * config/i386/i386-opts.h: Add enum stack_protector_guard.
11767 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
11768 TARGET_SSP_TLS_GUARD.
11769 * config/i386/i386.c (ix86_option_override_internal): Set
11770 ix86_stack_protector_guard.
11771 * config/i386/i386.md (stack_protect_set): Enable for
11772 TARGET_SSP_TLS_GUARD only.
11773 (stack_protect_set_<mode>): Ditto.
11774 (stack_protect_test): Ditto.
11775 (stack_protect_test_<mode>): Ditto.
11776 * doc/invoke.texi (i386 Option): Document.
11777
11778 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
11779
11780 PR target/56890
11781 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
11782 (S_MODES): Set H_MODE bit.
11783 (SF_MODES): Set only S_MODE and SF_MODE bits.
11784 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
11785 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
11786 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
11787 <MODE_FLOAT>: Likewise.
11788
11789 2013-04-15 Joey Ye <joey.ye@arm.com>
11790
11791 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
11792
11793 2013-04-15 Joey Ye <joey.ye@arm.com>
11794
11795 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
11796 for real far jump.
11797 (thumb_far_jump_used_p): Count instruction size and set
11798 far_jump_used.
11799
11800 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
11801
11802 * reorg.c (fill_simple_delay_slots): Reindent block of code.
11803 * resource.c (mark_target_live_regs): Reformat conditional block.
11804
11805 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11806
11807 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
11808 notes, they are emitted only just before final.
11809 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
11810
11811 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
11812
11813 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
11814 * cfgrtl.c (delete_insn): Call it here instead.
11815 * lra-spills.c (lra_final_code_change): Use delete_insn.
11816 * haifa-sched.c (sched_remove_insn): Likewise.
11817 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
11818 returning to the nop pool.
11819 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
11820 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
11821
11822 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11823
11824 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
11825 * doc/tm.texi: Regenerated.
11826
11827 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
11828
11829 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
11830 QImode checks.
11831
11832 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
11833
11834 * df-core.c (df_find_def): Compare register numbers.
11835 (df_find_use): Likewise.
11836
11837 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
11838
11839 PR target/56903
11840 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
11841 lra_in_progress for return.
11842
11843 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11844
11845 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
11846 define_insn into define_insn_and_split and emit movsicc patterns.
11847
11848 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
11849
11850 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
11851
11852 2013-04-12 Richard Biener <rguenther@suse.de>
11853
11854 * tree-pass.h (TODO_do_not_ggc_collect): New.
11855 * passes.c (execute_one_ipa_transform_pass): Honor
11856 TODO_do_not_ggc_collect.
11857 (execute_one_pass): Likewise.
11858
11859 Revert
11860 2013-04-10 Richard Biener <rguenther@suse.de>
11861
11862 * passes.c (init_optimization_passes): Remove reload pass.
11863 * ira.c (do_reload): Merge into ...
11864 (ira): ... this.
11865 (rest_of_handle_reload): Remove.
11866 (pass_reload): Likewise.
11867 * config/i386/i386.c (ix86_option_override): Refer to ira instead
11868 of reload for vzeroupper pass placement.
11869
11870 2013-04-12 Jakub Jelinek <jakub@redhat.com>
11871
11872 PR tree-optimization/56918
11873 PR tree-optimization/56920
11874 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
11875 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
11876 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
11877 use wide_mul_with_sign method.
11878
11879 2013-04-12 Richard Biener <rguenther@suse.de>
11880
11881 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
11882 not be considered a gimple constant.
11883
11884 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
11885
11886 * fold-const.c (const_binop): Handle vector shifts by a scalar.
11887 (fold_binary_loc): Call const_binop also for mixed vector-scalar
11888 operations.
11889
11890 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
11891 Jakub Jelinek <jakub@redhat.com>
11892
11893 * opts.c: Include diagnostic-color.h.
11894 (common_handle_option): Handle OPT_fdiagnostics_color_.
11895 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
11896 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
11897 (diagnostic-color.o): New.
11898 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
11899 (diagnostic_color_rule): New enum.
11900 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
11901 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
11902 the location string.
11903 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
11904 either NULL, or color kind.
11905 * diagnostic-color.c: New file.
11906 * diagnostic-color.h: New file.
11907 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
11908 arguments.
11909 * doc/invoke.texi (-fdiagnostics-color): Document.
11910 * pretty-print.h (pp_show_color): Define.
11911 (struct pretty_print_info): Add show_color field.
11912 * diagnostic.c: Include diagnostic-color.h.
11913 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
11914 macros. Colorize error:, warning: etc. strings and also the location
11915 string.
11916 (diagnostic_show_locus): Colorize the caret line.
11917 * pretty-print.c: Include diagnostic-color.h.
11918 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
11919 inside of %< %> quotes or quoted through q format modifier.
11920
11921 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11922
11923 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
11924
11925 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11926
11927 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
11928 code in CC_NZ mode.
11929 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
11930 pattern.
11931
11932 2013-04-11 Marek Polacek <polacek@redhat.com>
11933
11934 PR tree-optimization/48184
11935 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
11936
11937 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
11938
11939 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
11940 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
11941 (skip_simple_arithmetic): Tidy up.
11942 * tree.h (skip_simple_constant_arithmetic): Declare.
11943
11944 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11945
11946 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
11947
11948 2013-04-11 Richard Biener <rguenther@suse.de>
11949
11950 * tree-vect-loop.c (get_initial_def_for_induction): Properly
11951 generate vector constants.
11952
11953 2013-04-11 Richard Biener <rguenther@suse.de>
11954
11955 PR tree-optimization/56878
11956 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
11957 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
11958 New function.
11959 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
11960 Prefer to align the DR with the most invariant base address.
11961
11962 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
11963
11964 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
11965 comment.
11966
11967 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
11968
11969 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
11970 floating-point vector comparisons against 0.
11971
11972 2013-04-11 Jakub Jelinek <jakub@redhat.com>
11973
11974 PR tree-optimization/56899
11975 * fold-const.c (extract_muldiv_1): Apply distributive law
11976 only if TYPE_OVERFLOW_WRAPS (ctype).
11977
11978 2013-04-11 Bin Cheng <bin.cheng@arm.com>
11979
11980 PR target/56124
11981 * ira-costs.c (scan_one_insn): Check whether the source rtx of
11982 loading has side effect.
11983
11984 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
11985
11986 * config/sparc/sparc.c: Include tree-pass.h.
11987 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
11988 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
11989 head of file. Change return type. Split off gate function.
11990 (sparc_gate_work_around_errata): New function.
11991 (pass_work_around_errata): New pass definition.
11992 (insert_pass_work_around_errata) New pass insert definition to
11993 insert pass_work_around_errata just after delayed-branch scheduling.
11994 (sparc_option_override): Insert the pass.
11995 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
11996
11997 2013-04-10 David S. Miller <davem@davemloft.net>
11998
11999 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12000 or -mcpu=hypersparc.
12001
12002 * target.def (cstore_mode): New hook.
12003 * target.h: Include insn-codes.h
12004 * targhooks.c: Likewise.
12005 (default_cstore_mode): New function.
12006 * targhooks.h: Declare it.
12007 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12008 * doc/tm.texi: Rebuild.
12009 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12010 target hook, rather than inspecting the insn_data.
12011 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12012 (TARGET_CSTORE_MODE): Redefine.
12013 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12014 result patterns.
12015 * config/sparc/predicates.md (cstore_result_operand): New special
12016 predicate.
12017 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12018 Use it for operand 0.
12019 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12020 (*snesi_special): Likewise.
12021 (*snesi_zero): Likewise.
12022 (*seqsi_zero): Likewise.
12023 (*sltu_insn): Likewise.
12024 (*sgeu_insn): Likewise.
12025 (*seqdi_special): Make operand 0 and comparison operation be of
12026 DImode.
12027 (*snedi_special): Likewise.
12028 (*snedi_special_vis3): Likewise.
12029 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12030 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12031 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12032 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12033 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12034 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12035 (*sltu_extend_sp64): Likewise.
12036 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12037 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12038 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12039 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12040 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12041
12042 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12043
12044 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12045 (aarch64_start_file): Use the new function.
12046
12047 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12048 Jason Merrill <jason@redhat.com>
12049
12050 * common.opt: Add -gdwarf.
12051 * opts.c (common_handle_option): Handle it.
12052 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12053
12054 2013-04-10 Richard Biener <rguenther@suse.de>
12055
12056 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12057 (execute_one_ipa_transform_pass): But unconditionally here.
12058 (execute_one_pass): And here.
12059 (init_optimization_passes): Remove reload pass.
12060 * tree-pass.h (TODO_ggc_collect): Remove.
12061 (pass_reload): Likewise.
12062 * ira.c (do_reload): Merge into ...
12063 (ira): ... this.
12064 (rest_of_handle_reload): Remove.
12065 (pass_reload): Likewise.
12066 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12067 of reload for vzeroupper pass placement.
12068 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12069 and todo_flags_finish of all passes.
12070
12071 2013-04-10 Richard Biener <rguenther@suse.de>
12072
12073 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12074 first_const_oprnd field, rename first_def_type to first_op_type.
12075 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12076 (vect_get_and_check_slp_defs): Always use the type of the
12077 operand. Allow mixed vect_external_def, vect_constant_def types.
12078 (vect_get_constant_vectors): Handle mixed vect_external_def,
12079 vect_constant_def types.
12080
12081 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12082
12083 PR tree-optimization/55524
12084 * tree-ssa-math-opts.c
12085 (convert_mult_to_fma): Don't use an fms construct
12086 when we don't have an fms operation, but fnma, and it looks
12087 likely that we'll be able to use the latter.
12088
12089 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12090
12091 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12092 function.
12093 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
12094 inline fail caused by overwritable functions.
12095
12096 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
12097
12098 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
12099 unnecessary bits in the constant power of two case.
12100
12101 2013-04-10 Richard Biener <rguenther@suse.de>
12102
12103 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
12104 broken code swapping operands.
12105 (vect_build_slp_tree): Do not compute load permutations here.
12106 (vect_analyze_slp_instance): Compute load permutations here,
12107 after building the SLP tree.
12108
12109 2013-04-09 Christian Bruel <christian.bruel@st.com>
12110
12111 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
12112 of next/prev_real_insn.
12113
12114 2013-04-09 Jan Hubicka <jh@suse.cz>
12115
12116 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12117 Drop aliased parameter.
12118 (function_and_variable_visibility): Do not handle alias pairs.
12119 * cgraph.c (varpool_externally_visible_p): Update prototype.
12120 * varpool.c (varpool_add_new_variable): Update.
12121
12122 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12123
12124 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
12125
12126 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
12127
12128 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
12129
12130 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
12131
12132 2013-04-09 Marek Polacek <polacek@redhat.com>
12133
12134 PR tree-optimization/48762
12135 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
12136
12137 2013-04-09 Richard Biener <rguenther@suse.de>
12138
12139 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
12140 dealing with cost.
12141 (vect_build_slp_tree): Likewise.
12142 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
12143 calculating the cost of a SLP instance.
12144 (vect_analyze_slp_instance): Use it from here, after building
12145 the SLP tree.
12146
12147 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12148
12149 PR middle-end/56883
12150 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
12151 expand_omp_for_static_chunk): Use simple_p = true in
12152 force_gimple_operand_gsi calls when assigning to addressable decls.
12153
12154 2013-04-09 Jeff Law <law@redhat.com>
12155
12156 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
12157 when the boolean was created by converting a wider object which
12158 had a boolean range.
12159
12160 2013-04-09 Richard Biener <rguenther@suse.de>
12161
12162 * tree-vectorizer.h (slp_void_p): Remove.
12163 (slp_tree): Typedef before _slp_tree declaration.
12164 (struct _slp_tree): Use a vector of slp_tree as children.
12165 (vect_get_place_in_interleaving_chain): Remove.
12166 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
12167 Move ...
12168 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
12169 and make static.
12170 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
12171 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
12172 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
12173 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
12174 Use slp_node instead of slp_void_p and adjust.
12175
12176 2013-04-09 Richard Biener <rguenther@suse.de>
12177
12178 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
12179 work that is not necessary.
12180
12181 2013-04-09 Jakub Jelinek <jakub@redhat.com>
12182
12183 PR tree-optimization/56854
12184 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
12185 forward into clobber stmts if it would change MEM_REF lhs into
12186 non-MEM_REF.
12187
12188 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
12189
12190 * tree.c (type_hash_lookup, type_hash_add): Make static.
12191 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
12192
12193 2013-04-09 Richard Biener <rguenther@suse.de>
12194
12195 * tree.h (unsave_expr_now): Remove.
12196 * tree-inline.c (mark_local_for_remap_r): Remove.
12197 (unsave_expr_1): Likewise.
12198 (unsave_r): Likewise.
12199 (unsave_expr_now): Likewise.
12200 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
12201 (propagate_tree_value): Likewise.
12202
12203 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
12204
12205 * doc/rtl.texi (sequence): Rewrite documentation to match the
12206 current use of SEQUENCE rtl objects.
12207 * rtl.def (SEQUENCE): Likewise.
12208
12209 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
12210 Update documentation.
12211 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
12212 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
12213
12214 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
12215
12216 2013-04-08 Teresa Johnson <tejohnson@google.com>
12217
12218 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
12219 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
12220 methods.
12221 (estimate_edge_size_and_time): Add comment to suggest using rounding
12222 methods.
12223 (estimate_node_size_and_time): Ditto.
12224 (remap_edge_change_prob): Use helper rounding divide methods.
12225 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
12226 (gimple_mod_pow2_value_transform): Ditto.
12227 (gimple_mod_subtract_transform): Ditto.
12228 (gimple_ic_transform): Ditto.
12229 (gimple_stringops_transform): Ditto.
12230 * stmt.c (conditional_probability): Ditto.
12231 (emit_case_dispatch_table): Ditto.
12232 * lto-cgraph.c (merge_profile_summaries): Ditto.
12233 * tree-optimize.c (execute_fixup_cfg): Ditto.
12234 * cfgcleanup.c (try_forward_edges): Ditto.
12235 * cfgloopmanip.c (scale_loop_profile): Ditto.
12236 (loopify): Ditto.
12237 (duplicate_loop_to_header_edge): Ditto.
12238 (lv_adjust_loop_entry_edge): Ditto.
12239 * tree-vect-loop.c (vect_transform_loop): Ditto.
12240 * profile.c (compute_branch_probabilities): Ditto.
12241 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
12242 * lto-streamer-in.c (input_cfg): Ditto.
12243 * gimple-streamer-in.c (input_bb): Ditto.
12244 * ipa-cp.c (update_profiling_info): Ditto.
12245 (update_specialized_profile): Ditto.
12246 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
12247 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
12248 rounding methods.
12249 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12250 (compute_trg_info): Ditto.
12251 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
12252 (purge_dead_edges): Ditto.
12253 * loop-unswitch.c (unswitch_loop): Ditto.
12254 * cgraphclones.c (cgraph_clone_edge): Ditto.
12255 (cgraph_clone_node): Ditto.
12256 * tree-inline.c (copy_bb): Ditto.
12257 (copy_edges_for_bb): Ditto.
12258 (initialize_cfun): Ditto.
12259 (copy_cfg_body): Ditto.
12260 (expand_call_inline): Ditto.
12261
12262 2013-04-08 Kai Tietz <ktietz@redhat.com>
12263
12264 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
12265 TARGET_CYGWIN64 by TARGET_64BIT.
12266
12267 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12268
12269 * config/epiphany/epiphany.md (GPR_1): New constant.
12270 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
12271 * config/epiphany/epiphany.c (gen_compare_reg):
12272 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
12273 is already in place.
12274 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
12275 Don't require being called during rtl expansion; If y operlaps r0,
12276 return 0.
12277 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
12278 (epiphany_expand_epilogue): Likewise.
12279
12280 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
12281 Don't use CC_FPmode for ORDERED / UNORDERED.
12282 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
12283
12284 * config/epiphany/constraints.md (CnL): New constraint.
12285 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
12286 * config/epiphany/predicates.md (add_operand): Allow 1024.
12287
12288 * config/epiphany/epiphany.md (logical_op): New code iterator.
12289 (op_mnc): New code attribute.
12290 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
12291 (mov_f+1, mov_f+2): New peephole2 patterns.
12292
12293 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
12294 (cstoresi4): Also allow re-use of zero result when doing a NE
12295 comparison to a non-zero operand.
12296 Use (clobber (scratch)) for first insn if the gpr output is not needed.
12297
12298 * config/epiphany/epiphany.md (<insn_opname>v2si3):
12299 Use gen_addsi3_i / gen_subsi3_i.
12300
12301 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12302
12303 PR c++/34949
12304 PR c++/50243
12305 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
12306 contain anything but clobbers, at most one __builtin_stack_restore,
12307 optionally debug stmts and final resx, and if it has at least one
12308 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
12309 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
12310 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
12311 which isn't defaut definition, remove them.
12312 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
12313 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
12314 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
12315 with MEM_REF LHS with SSA_NAME address.
12316
12317 2013-04-08 Jeff Law <law@redhat.com>
12318
12319 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
12320
12321 2013-04-08 Richard Biener <rguenther@suse.de>
12322
12323 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
12324 extra newline.
12325 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
12326 determined vector type.
12327 (vect_analyze_data_refs): Likewise.
12328 (vect_get_new_vect_var): Adjust.
12329 (vect_create_destination_var): Preserve SSA name versions.
12330 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
12331 not dump anything here.
12332
12333 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
12334
12335 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
12336 Add member lr_slot_known.
12337 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
12338 if necessary.
12339 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12340 Remove code that sets lr_slot_offset according to what a previous
12341 version of epiphany_emit_save_restore used to do.
12342 (epiphany_emit_save_restore): When doing an lr save or restore,
12343 set/verify lr_slot_known and lr_slot_offset.
12344
12345 2013-04-08 Xinyu Qi <xyqi@marvell.com>
12346
12347 PR target/54338
12348 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
12349 in ALL_REGS.
12350
12351 2013-04-08 Richard Biener <rguenther@suse.de>
12352
12353 * alias.c (find_base_term): Fix thinko in previous change.
12354
12355 2013-04-08 Jakub Jelinek <jakub@redhat.com>
12356
12357 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
12358 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
12359 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
12360 if possible to compute val.
12361 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
12362 For QImode integers don't require anything about precision. Use
12363 const_with_all_bytes_same to find out if the constant doesn't have
12364 repeated bytes in it.
12365
12366 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12367
12368 * config/s390/s390.c (s390_expand_insv): Only accept insertions
12369 within mode size.
12370
12371 2013-04-08 Marek Polacek <polacek@redhat.com>
12372
12373 PR rtl-optimization/48182
12374 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
12375 value to 1.
12376
12377 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12378
12379 PR target/55487
12380 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
12381 nuses, make sure we have a label.
12382
12383 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12384
12385 PR target/56843
12386 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
12387 (rs6000_emit_swdiv_low_precision): Remove.
12388 (rs6000_emit_swdiv): Rewrite to handle between one and four
12389 iterations of Newton-Raphson generally; modify required number of
12390 iterations for some cases.
12391 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
12392
12393 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
12394
12395 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
12396 set-but-unused variable.
12397
12398 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
12399 basic blocks of released function bodies garbage-collectable.
12400
12401 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
12402 (struct rtl_opt_pass): Add TODO_df_finish.
12403
12404 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
12405
12406 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12407
12408 * config/arm/constraints.md (q): New constraint.
12409 * config/arm/ldrdstrd.md: New file.
12410 * config/arm/arm.md (ldrdstrd.md) New include.
12411 (arm_movdi): Use "q" instead of "r" constraint
12412 for double-word memory access.
12413 (movdf_soft_insn): Likewise.
12414 * config/arm/vfp.md (movdi_vfp): Likewise.
12415 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
12416 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
12417 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
12418 (mem_ok_for_ldrd_strd): Likewise.
12419 (output_move_double): Update assertion.
12420
12421 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12422
12423 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
12424
12425 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12426
12427 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
12428 define_insn_and_split.
12429 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
12430
12431 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12432
12433 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
12434 define_insn_and_split.
12435 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
12436 (shiftsi3_compare): New pattern.
12437 (rrx): New pattern.
12438 * config/arm/unspecs.md (UNSPEC_RRX): New.
12439
12440 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12441
12442 * config/arm/arm.md (negdi_extendsidi): New pattern.
12443 (negdi_zero_extendsidi): Likewise.
12444
12445 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12446
12447 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
12448 define_insn_and_split.
12449 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
12450 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
12451
12452 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12453
12454 * config/arm/arm.md (arm_subdi3): Convert define_insn into
12455 define_insn_and_split.
12456 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
12457 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
12458
12459 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12460
12461 * config/arm/arm.md (subsi3_carryin): New pattern.
12462 (subsi3_carryin_const): Likewise.
12463 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
12464 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
12465
12466 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12467
12468 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
12469
12470 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
12471
12472 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
12473 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
12474
12475 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12476
12477 * config/arm/arm.c (arm_expand_builtin): Change fcode
12478 type to unsigned int.
12479
12480 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12481
12482 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
12483
12484 2013-04-04 Ian Lance Taylor <iant@google.com>
12485
12486 * doc/standards.texi (Standards): The Go frontend supports the Go 1
12487 language standard.
12488
12489 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
12490
12491 PR middle-end/56729
12492 * df-scan.c (df_insn_delete): Disable failing assert.
12493
12494 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12495
12496 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
12497 New function prototype.
12498 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
12499 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
12500 (arm_builtin_vectorized_function): New function.
12501
12502 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12503
12504 * config/arm/arm_neon_builtins.def: New file.
12505 * config/arm/arm.c (neon_builtin_data): Move contents to
12506 arm_neon_builtins.def.
12507 (enum arm_builtins): Include neon builtin definitions.
12508 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
12509 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
12510
12511 2013-04-04 Marek Polacek <polacek@redhat.com>
12512
12513 PR tree-optimization/48186
12514 * predict.c (maybe_hot_frequency_p): Return false if
12515 HOT_BB_FREQUENCY_FRACTION is 0.
12516 (cgraph_maybe_hot_edge_p): Likewise.
12517
12518 2013-04-04 Richard Biener <rguenther@suse.de>
12519
12520 PR tree-optimization/56826
12521 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
12522 more accurately.
12523
12524 2013-04-04 Richard Biener <rguenther@suse.de>
12525
12526 PR tree-optimization/56213
12527 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
12528 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
12529
12530 2013-04-04 Richard Biener <rguenther@suse.de>
12531
12532 PR tree-optimization/56837
12533 * tree-loop-distribution.c (classify_partition): For non-zero
12534 values require that the value has the same precision as its
12535 mode to be useful as memset value.
12536
12537 2013-04-03 Nick Clifton <nickc@redhat.com>
12538
12539 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
12540 (fmssf4): Use fmsf.s on E3V5 architectures.
12541 (fnmasf4): Use fnmaf.s on E3V5 architectures.
12542 (fnmssf4): Use fnmsf.s on E3V5 architectures.
12543
12544 2013-04-03 Jeff Law <law@redhat.com>
12545
12546 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
12547 (lra-eliminations.o): Likewise.
12548
12549 2013-04-03 Teresa Johnson <tejohnson@google.com>
12550
12551 * gcov-io.c (compute_working_sets): Moved most of body of old
12552 compute_working_sets here from profile.c.
12553 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
12554 (gcov_working_set_t): Moved typedef here from basic-block.h
12555 (compute_working_set): Declare.
12556 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
12557 (get_working_sets): Renamed from compute_working_set,
12558 replace most of body with call to new compute_working_sets.
12559 (get_exec_counts): Replace call to compute_working_sets
12560 to get_working_sets.
12561 * profile.h (get_working_sets): Renamed from compute_working_set.
12562 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
12563 to get_working_sets.
12564 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
12565 * gcov-dump.c (dump_working_sets): New function.
12566
12567 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
12568
12569 * hwint.c (sext_hwi, zext_hwi): New functions.
12570 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
12571 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
12572 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
12573 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
12574 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
12575 (sext_hwi, zext_hwi): New functions.
12576
12577 2013-04-03 Jeff Law <law@redhat.com>
12578
12579 PR tree-optimization/56799
12580 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
12581 back test for widening conversion erroneously dropped in prior change.
12582
12583 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12584
12585 PR target/56809
12586 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
12587 instead of next_real_insn.
12588
12589 2013-04-03 Marek Polacek <polacek@redhat.com>
12590
12591 PR sanitizer/55702
12592 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
12593
12594 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12595
12596 PR target/56809
12597 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
12598 next_real_insn.
12599 (thumb1_output_casesi): Likewise.
12600 (thumb2_output_casesi): Likewise.
12601
12602 2013-04-03 Richard Biener <rguenther@suse.de>
12603
12604 PR tree-optimization/56817
12605 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
12606 Split out ...
12607 (tree_unroll_loops_completely_1): ... new function to manually
12608 walk the loop tree, properly defering outer loops of unrolled
12609 loops to later iterations.
12610
12611 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12612
12613 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
12614 (vectorizable_load): Likewise.
12615 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
12616 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
12617
12618 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12619
12620 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
12621 BIT_FIELD_REF.
12622
12623 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12624
12625 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
12626
12627 2013-04-03 Bin Cheng <bin.cheng@arm.com>
12628
12629 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
12630
12631 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12632
12633 PR tree-optimization/56790
12634 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
12635 folding.
12636
12637 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
12638
12639 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
12640 Handle VEC_MERGE.
12641 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
12642 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
12643 equal arguments.
12644
12645 2013-04-03 Jakub Jelinek <jakub@redhat.com>
12646
12647 PR c/19449
12648 * tree.h (force_folding_builtin_constant_p): New decl.
12649 * builtins.c (force_folding_builtin_constant_p): New variable.
12650 (fold_builtin_constant_p): Fold immediately also if
12651 force_folding_builtin_constant_p.
12652
12653 2013-04-03 Richard Biener <rguenther@suse.de>
12654
12655 PR tree-optimization/56812
12656 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
12657 DRs of the same interleaving chain are independent.
12658
12659 2013-04-02 Jason Merrill <jason@redhat.com>
12660
12661 * gdbinit.in (pbb): Use debug fn.
12662
12663 2013-04-02 Lawrence Crowl <crowl@google.com>
12664
12665 * sese.h (struct ivtype_map_elt_s): Remove unused.
12666 (extern debug_ivtype_map): Remove unused.
12667 (extern eq_ivtype_map_elts): Remove unused.
12668 * sese.c (debug_ivtype_map): Removed unused.
12669 (debug_ivtype_map_1): Removed unused.
12670 (debug_ivtype_elt): Remove unused.
12671 (eq_ivtype_map_elts): Remove unused.
12672
12673
12674 2013-04-02 Kai Tietz <ktietz@redhat.com>
12675
12676 PR target/52790
12677 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
12678 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
12679 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
12680 function.
12681 (legitimize_pe_coff_symbol): Likewise.
12682 (is_imported_p): New helper-function.
12683 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
12684 for Windows x64 targets.
12685 (ix86_expand_prologue): Optimize for pe-coff targets.
12686 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
12687 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
12688 medium/large code-model.
12689 (legitimize_pic_address): Likewise.
12690 (legitimize_tls_address): Likewise.
12691 (ix86_expand_call): Likewise.
12692 (x86_output_mi_thunk): Likewise.
12693 (get_dllimport_decl): Add new beimport argument.
12694 (construct_plt_address): Don't assert for x64 pe-coff targets.
12695 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
12696 targets.
12697 (SYMBOL_FLAG_STUBVAR): New macro.
12698 (SYMBOL_REF_STUBVAR_P): Likewise.
12699 * config/i386/winnt.c (stub_list): New structure.
12700 (stub_head): New local variable.
12701 (i386_pe_record_stub): New function.
12702 (i386_pe_file_end): Emit refptr-stubs.
12703
12704 2013-04-02 Jakub Jelinek <jakub@redhat.com>
12705
12706 PR rtl-optimization/56745
12707 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
12708 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
12709
12710 PR c++/34949
12711 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
12712 and both of them are MEM_REFs, just compare first argument for
12713 equality and attempt to deal even with differing offsets.
12714
12715 PR c++/34949
12716 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
12717 of gimple_clobber_p to be MEM_REF.
12718 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
12719 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
12720 after gimplification.
12721 * asan.c (get_mem_ref_of_assignment): Don't instrument
12722 gimple_clobber_p stmts.
12723 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
12724 gimple_clobber_p stmt if they have MEM_REF lhs and
12725 are dead because of another gimple_clobber_p stmt.
12726 * tree-ssa-live.c (clear_unused_block_pointer): Treat
12727 gimple_clobber_p stmts like debug stmts.
12728 (remove_unused_locals): Remove clobbers with MEM_REF lhs
12729 that refer to unused VAR_DECLs or uninitialized values.
12730 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
12731 gimple_clobber_p stmts if they refer to removed parameters.
12732 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
12733 formatting.
12734
12735 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
12736
12737 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
12738 using SWI48 mode attribute.
12739
12740 2013-04-02 Wei Mi <wmi@google.com>
12741
12742 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
12743 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
12744 *<rotate_insn><mode>3_mask in i386.md.
12745
12746 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
12747
12748 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
12749
12750 2013-04-02 Richard Biener <rguenther@suse.de>
12751
12752 PR tree-optimization/56778
12753 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
12754 Runtime alias tests are not supported for gather loads.
12755 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
12756 stmts referenced from SSA operands before updating SSA form.
12757
12758 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
12759 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12760
12761 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
12762 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
12763 * config/arm/cortex-a53.md: New file.
12764 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
12765 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
12766 * config/arm/arm.c (arm_issue_rate): Likewise.
12767 * config/arm/arm-tune.md: Regenerate
12768 * config/arm/arm-tables.opt: Regenerate.
12769 * config/arm/arm-cores.def: Add cortex-a53.
12770
12771 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
12772
12773 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
12774 non-static link.
12775
12776 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12777
12778 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
12779 scalar load/store operations using B/H registers.
12780 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
12781
12782 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
12783
12784 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
12785 scalar move.
12786 * config/aarch64/aarch64.c
12787 (aarch64_simd_scalar_immediate_valid_for_move): New.
12788 * config/aarch64/aarch64-protos.h
12789 (aarch64_simd_scalar_immediate_valid_for_move): New.
12790 * config/aarch64/constraints.md (Dh, Dq): New.
12791 * config/aarch64/iterators.md (hq): New.
12792
12793 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
12794
12795 * reorg.c (get_branch_condition): Deal with conditional returns.
12796 (fill_simple_delay_slots): Remove dead code dealing with jumps.
12797
12798 2013-04-01 Wei Mi <wmi@google.com>
12799
12800 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
12801 Truncate operand 2 using %b asm operand modifier.
12802 (*<shift_insn><mode>3_mask): Ditto.
12803 (*<rotate_insn><mode>3_mask): Ditto.
12804
12805 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
12806
12807 PR middle-end/56798
12808 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
12809
12810 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
12811
12812 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
12813 of next_real_insn.
12814 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
12815
12816 2013-03-30 Lawrence Crowl <crowl@google.com>
12817
12818 * dse.c (clear_alias_sets): Remove never set.
12819 (disqualified_clear_alias_sets): Remove never set.
12820 (clear_alias_mode_pool): Remove never set.
12821 (dse_step0): Remove condition that is never true.
12822 (canon_address): Remove condition that is never true.
12823 (dse_step7): Remove condition that is never true.
12824 (rest_of_handle_dse): Remove condition that is never true.
12825 (rest_of_handle_dse::did_global): Remove never read from above.
12826 (dse_step2_spill): Remove never called from above.
12827 (dse_step5_spill): Remove never called from above.
12828
12829 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
12830
12831 * doc/md.texi (Standard Names) <casesi>: Update documentation for
12832 JUMP_TABLE_DATA changes.
12833 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
12834 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
12835 (Insns) <jump_table_data>: New entry.
12836 * doc/tm.texi: Regenerate.
12837
12838 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
12839
12840 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
12841 for table jump at the end of a basic block using tablejump_p.
12842 * targhooks.c (default_invalid_within_doloop): Likewise.
12843 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
12844 target hook implementation that is identical to the default hook.
12845 (rs6000_invalid_within_doloop): Remove.
12846
12847 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
12848 unused variable from tablejump_p call.
12849
12850 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
12851 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
12852 (INSN_DELETED_P): Likewise.
12853 (emit_jump_table_data): New prototype.
12854 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
12855 after 4th as unused.
12856 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
12857 * sched-vis.c (print_insn): Likewise.
12858 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
12859 insn for compatibility with back ends that use next_active_insn to
12860 identify jump table data.
12861 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
12862 (remove_insn): Likewise.
12863 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
12864 to be emitted.
12865 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
12866 (emit_jump_table_data): New function.
12867
12868 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
12869 basic block, a JUMP_TABLE_DATA never is.
12870 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
12871 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
12872 off from code handling real insns.
12873 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
12874 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
12875 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
12876 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
12877 is not a NONDEBUG_INSN_P.
12878 * ira-costs.c (scan_one_insn): Likewise.
12879 * jump.c (mark_all_labels): Likewise.
12880 (mark_jump_label_1): Likewise.
12881 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12882 * lra.c (get_insn_freq): Expect all insns reaching here to be in
12883 a basic block.
12884 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
12885 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
12886 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
12887 JUMP_TABLE_DATA_P insns.
12888 (calculate_elim_costs_all_insns): Likewise.
12889 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
12890 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
12891 (delete_output_reload): Code style fixups.
12892 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
12893 insn flags on this non-insn.
12894 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
12895 as scheduling barriers, for pre-change compatibility.
12896 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
12897 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
12898
12899 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
12900 redundant JUMP_TABLE_DATA_P test.
12901 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
12902 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12903 (frv_for_each_packet): Likewise.
12904 * config/i386/i386.c (min_insn_size): Likewise.
12905 (ix86_avoid_jump_mispredicts): Likewise.
12906 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12907 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
12908 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12909 (mips16_insn_length): Robustify.
12910 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
12911 (mips16_split_long_branches): Likewise.
12912 * config/pa/pa.c (pa_combine_instructions): Likewise.
12913 * config/rs6000/rs6000.c (get_next_active_insn): Treat
12914 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
12915 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
12916 as contributing to pool range lengths.
12917 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
12918 Remove redundant JUMP_TABLE_DATA_P test.
12919 (sh_loop_align): Likewise.
12920 (split_branches): Likewise.
12921 (sh_insn_length_adjustment): Likewise.
12922 * config/spu/spu.c (get_branch_target): Likewise.
12923
12924 2013-03-29 Jan Hubicka <jh@suse.cz>
12925
12926 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
12927 gcov streaming; stream hot bb threshold to ltrans.
12928 * predict.c (get_hot_bb_threshold): Break out from ....
12929 (maybe_hot_count_p): ... here.
12930 (set_hot_bb_threshold): New function.
12931 * lto-section-in.c (lto_section_name): Add profile.
12932 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
12933 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
12934 and data-streamer.h
12935 (histogram_entry): New structure.
12936 (histogram, histogram_pool): New global vars.
12937 (histogram_hash): New structure.
12938 (histogram_hash::hash): New method.
12939 (histogram_hash::equal): Likewise.
12940 (account_time_size): New function.
12941 (cmp_counts): New function.
12942 (dump_histogram): New function.
12943 (ipa_profile_generate_summary): New function.
12944 (ipa_profile_write_summary): New function.
12945 (ipa_profile_read_summary): New function.
12946 (ipa_profile): Decide on threshold.
12947 (pass_ipa_profile): Add ipa_profile_write_summary and
12948 ipa_profile_read_summary.
12949 * Makefile.in (ipa.o): Update dependencies.
12950 * lto-streamer.h (LTO_section_ipa_profile): New section.
12951
12952 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
12953
12954 * tree.h (VAR_P): New.
12955
12956 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
12957
12958 PR lto/56777
12959 * doc/invoke.texi ([-fwhole-program]): Fix typo.
12960
12961 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
12962
12963 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
12964 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12965 (control_flow_insn_p): Likewise.
12966 * cfgrtl.c (duplicate_insn_chain): Likewise.
12967 * final.c (get_attr_length_1): Likewise.
12968 (shorten_branches): Likewise.
12969 (final_scan_insn): Likewise.
12970 * function.c (instantiate_virtual_regs): Likewise.
12971 * gcse.c (insert_insn_end_basic_block): Likewise.
12972 * ira-costs.c (scan_one_insn): Likewise.
12973 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
12974 * lra.c (check_rtl): Likewise.
12975 * reload1.c (elimination_costs_in_insn): Likewise.
12976 * reorg.c (follow_jumps): Likewise.
12977
12978 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
12979 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
12980 (thumb_far_jump_used_p): Likewise.
12981 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
12982 (workaround_speculation): Likewise.
12983 (add_sched_insns_for_speculation): Likewise.
12984 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
12985 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
12986 (frv_for_each_packet): Likewise.
12987 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
12988 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
12989 (final_emit_insn_group_barriers): Likewise.
12990 * config/m32r/m32r.c (m32r_is_insn): Likewise.
12991 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
12992 (mips16_insn_length): Likewise.
12993 * config/pa/pa.c (pa_reorg): Likewise.
12994 (pa_combine_instructions): Likewise.
12995 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
12996 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
12997 (sh_reorg): Likewise.
12998 (split_branches): Likewise.
12999 * config/spu/spu.c (get_branch_target): Likewise.
13000
13001 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13002 JUMP_TABLE_DATA_P.
13003
13004 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13005
13006 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13007 Fix declaration name.
13008
13009 2013-03-28 Lawrence Crowl <crowl@google.com>
13010
13011 * graphds.h (struct graph.indicies): Remove unused.
13012 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13013 (SCOP_ORIGINAL_PDDRS): Remove unused.
13014 * sese.h (extern insert_loop_close_phis): Removed unused.
13015 (extern insert_guard_phis): Removed unused.
13016 (extern ivtype_map_elt_info): Removed unused.
13017 (new_ivtype_map_elt): Removed unused.
13018 * sese.c (ivtype_map_elt_info): Removed unused.
13019
13020 2013-03-28 Lawrence Crowl <crowl@google.com>
13021
13022 * Makefile.in: Add several missing include dependences.
13023 (DUMPFILE_H): New.
13024 (test-dump.o): New. This object is not added to any executable,
13025 but is present for ad-hoc testing.
13026 * bitmap.c
13027 (debug (const bitmap_head_def &)): New.
13028 (debug (const bitmap_head_def *)): New.
13029 * bitmap.h
13030 (extern debug (const bitmap_head_def &)): New.
13031 (extern debug (const bitmap_head_def *)): New.
13032 * cfg.c
13033 (debug (edge_def &)): New.
13034 (debug (edge_def *)): New.
13035 * cfghooks.c
13036 (debug (basic_block_def &)): New.
13037 (debug (basic_block_def *)): New.
13038 * dumpfile.h
13039 (dump_node (const_tree, int, FILE *)): Correct source file.
13040 * dwarf2out.c
13041 (debug (die_struct &)): New.
13042 (debug (die_struct *)): New.
13043 * dwarf2out.h
13044 (extern debug (die_struct &)): New.
13045 (extern debug (die_struct *)): New.
13046 * gimple-pretty-print.c
13047 (debug (gimple_statement_d &)): New.
13048 (debug (gimple_statement_d *)): New.
13049 * gimple-pretty-print.h
13050 (extern debug (gimple_statement_d &)): New.
13051 (extern debug (gimple_statement_d *)): New.
13052 * ira-build.c
13053 (debug (ira_allocno_copy &)): New.
13054 (debug (ira_allocno_copy *)): New.
13055 (debug (ira_allocno &)): New.
13056 (debug (ira_allocno *)): New.
13057 * ira-int.h
13058 (extern debug (ira_allocno_copy &)): New.
13059 (extern debug (ira_allocno_copy *)): New.
13060 (extern debug (ira_allocno &)): New.
13061 (extern debug (ira_allocno *)): New.
13062 * ira-lives.c
13063 (debug (live_range &)): New.
13064 (debug (live_range *)): New.
13065 * lra-int.h
13066 (debug (lra_live_range &)): New.
13067 (debug (lra_live_range *)): New.
13068 * lra-lives.c
13069 (debug (lra_live_range &)): New.
13070 (debug (lra_live_range *)): New.
13071 * omega.c
13072 (debug (omega_pb_d &)): New.
13073 (debug (omega_pb_d *)): New.
13074 * omega.h
13075 (extern debug (omega_pb_d &)): New.
13076 (extern debug (omega_pb_d *)): New.
13077 * print-rtl.c
13078 (debug (const rtx_def &)): New.
13079 (debug (const rtx_def *)): New.
13080 * print-tree.c
13081 (debug_tree (tree): Move within file.
13082 (debug_raw (const tree_node &)): New.
13083 (debug_raw (const tree_node *)): New.
13084 (dump_tree_via_hooks (const tree_node *, int)): New.
13085 (debug (const tree_node &)): New.
13086 (debug (const tree_node *)): New.
13087 (debug_verbose (const tree_node &)): New.
13088 (debug_verbose (const tree_node *)): New.
13089 (debug_head (const tree_node &)): New.
13090 (debug_head (const tree_node *)): New.
13091 (debug_body (const tree_node &)): New.
13092 (debug_body (const tree_node *)): New.
13093 (debug_vec_tree (tree): Move and reimplement in terms of dump.
13094 (debug (vec<tree, va_gc> &)): New.
13095 (debug (vec<tree, va_gc> *)): New.
13096 * rtl.h
13097 (extern debug (const rtx_def &)): New.
13098 (extern debug (const rtx_def *)): New.
13099 * sbitmap.c
13100 (debug_raw (simple_bitmap_def &)): New.
13101 (debug_raw (simple_bitmap_def *)): New.
13102 (debug (simple_bitmap_def &)): New.
13103 (debug (simple_bitmap_def *)): New.
13104 * sbitmap.h
13105 (extern debug (simple_bitmap_def &)): New.
13106 (extern debug (simple_bitmap_def *)): New.
13107 (extern debug_raw (simple_bitmap_def &)): New.
13108 (extern debug_raw (simple_bitmap_def *)): New.
13109 * sel-sched-dump.c
13110 (debug (vinsn_def &)): New.
13111 (debug (vinsn_def *)): New.
13112 (debug_verbose (vinsn_def &)): New.
13113 (debug_verbose (vinsn_def *)): New.
13114 (debug (expr_def &)): New.
13115 (debug (expr_def *)): New.
13116 (debug_verbose (expr_def &)): New.
13117 (debug_verbose (expr_def *)): New.
13118 (debug (vec<rtx> &)): New.
13119 (debug (vec<rtx> *)): New.
13120 * sel-sched-dump.h
13121 (extern debug (vinsn_def &)): New.
13122 (extern debug (vinsn_def *)): New.
13123 (extern debug_verbose (vinsn_def &)): New.
13124 (extern debug_verbose (vinsn_def *)): New.
13125 (extern debug (expr_def &)): New.
13126 (extern debug (expr_def *)): New.
13127 (extern debug_verbose (expr_def &)): New.
13128 (extern debug_verbose (expr_def *)): New.
13129 (extern debug (vec<rtx> &)): New.
13130 (extern debug (vec<rtx> *)): New.
13131 * sel-sched-ir.h
13132 (_list_iter_cond_expr): Make inline instead of static.
13133 * sreal.c
13134 (debug (sreal &)): New.
13135 (debug (sreal *)): New.
13136 * sreal.h
13137 (extern debug (sreal &)): New.
13138 (extern debug (sreal *)): New.
13139 * tree.h
13140 (extern debug_raw (const tree_node &)): New.
13141 (extern debug_raw (const tree_node *)): New.
13142 (extern debug (const tree_node &)): New.
13143 (extern debug (const tree_node *)): New.
13144 (extern debug_verbose (const tree_node &)): New.
13145 (extern debug_verbose (const tree_node *)): New.
13146 (extern debug_head (const tree_node &)): New.
13147 (extern debug_head (const tree_node *)): New.
13148 (extern debug_body (const tree_node &)): New.
13149 (extern debug_body (const tree_node *)): New.
13150 (extern debug (vec<tree, va_gc> &)): New.
13151 (extern debug (vec<tree, va_gc> *)): New.
13152 * tree-cfg.c
13153 (debug (struct loop &)): New.
13154 (debug (struct loop *)): New.
13155 (debug_verbose (struct loop &)): New.
13156 (debug_verbose (struct loop *)): New.
13157 * tree-dump.c: Add header dependence.
13158 * tree-flow.h
13159 (extern debug (struct loop &)): New.
13160 (extern debug (struct loop *)): New.
13161 (extern debug_verbose (struct loop &)): New.
13162 (extern debug_verbose (struct loop *)): New.
13163 * tree-data-ref.c
13164 (debug (data_reference &)): New.
13165 (debug (data_reference *)): New.
13166 (debug (vec<data_reference_p> &)): New.
13167 (debug (vec<data_reference_p> *)): New.
13168 (debug (vec<ddr_p> &)): New.
13169 (debug (vec<ddr_p> *)): New.
13170 * tree-data-ref.h
13171 (extern debug (data_reference &)): New.
13172 (extern debug (data_reference *)): New.
13173 (extern debug (vec<data_reference_p> &)): New.
13174 (extern debug (vec<data_reference_p> *)): New.
13175 (extern debug (vec<ddr_p> &)): New.
13176 (extern debug (vec<ddr_p> *)): New.
13177 * tree-ssa-alias.c
13178 (debug (pt_solution &)): New.
13179 (debug (pt_solution *)): New.
13180 * tree-ssa-alias.h
13181 (extern debug (pt_solution &)): New.
13182 (extern debug (pt_solution *)): New.
13183 * tree-ssa-alias.c
13184 (debug (_var_map &)): New.
13185 (debug (_var_map *)): New.
13186 (debug (tree_live_info_d &)): New.
13187 (debug (tree_live_info_d *)): New.
13188 * tree-ssa-alias.h
13189 (extern debug (_var_map &)): New.
13190 (extern debug (_var_map *)): New.
13191 (extern debug (tree_live_info_d &)): New.
13192 (extern debug (tree_live_info_d *)): New.
13193
13194 2013-03-28 Jan Hubicka <jh@suse.cz>
13195
13196 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
13197
13198 2013-03-28 Ian Bolton <ian.bolton@arm.com>
13199
13200 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
13201 record only when desired or required.
13202
13203 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
13204
13205 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
13206 *vec_extractv2di_1_rex64. Use x64 isa attribute.
13207
13208 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13209
13210 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
13211 (*andsi3_compare0_uxtw): New pattern.
13212 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
13213 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
13214
13215 2013-03-28 Jan Hubicka <jh@suse.cz>
13216
13217 * data-streamer-in.c (streamer_read_gcov_count): New function.
13218 * gimple-streamer-out.c: Include value-prof.h.
13219 (output_gimple_stmt): Output histogram.
13220 (output_bb): Use streamer_write_gcov_count.
13221 * value-prof.c: Include data-streamer.h
13222 (dump_histogram_value): Add HIST_TYPE_MAX.
13223 (stream_out_histogram_value): New function.
13224 (stream_in_histogram_value): New function.
13225 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
13226 (stream_out_histogram_value, stream_in_histogram_value): Declare.
13227 * data-streamer-out.c (streamer_write_gcov_count): New function.
13228 (streamer_write_gcov_count_stream): New function.
13229 * lto-cgraph.c (lto_output_edge): Update counter streaming.
13230 (lto_output_node): Likewise.
13231 (input_node, input_edge): Likewise.
13232 * lto-streamer-out.c (output_cfg): Update streaming.
13233 * lto-streamer-in.c (input_cfg): Likewise.
13234 * data-streamer.h (streamer_write_gcov_count,
13235 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
13236 * gimple-streamer-in.c: Include value-prof.h
13237 (input_gimple_stmt): Input histograms.
13238 (input_bb): Update profile streaming.
13239
13240 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13241
13242 * genmodes.c (emit_max_int): New function.
13243 (emit_insn_modes_h): Added call to emit_max_function.
13244 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
13245 Added doc.
13246 * machmode.def: Fixed comment.
13247
13248 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
13249
13250 * combine.c (try_combine): Removed useless assert.
13251 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
13252
13253 2013-03-28 Marek Polacek <polacek@redhat.com>
13254 Richard Biener <rguenther@suse.de>
13255
13256 PR tree-optimization/56695
13257 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
13258 build signed result of a vector comparison.
13259 * tree-cfg.c (verify_gimple_comparison): Check that a result
13260 of a vector comparison has signed type.
13261
13262 2013-03-28 Richard Biener <rguenther@suse.de>
13263
13264 PR tree-optimization/37021
13265 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
13266 do not restrict gaps between groups.
13267 * tree-vect-stmts.c (vectorizable_load): Properly account for
13268 a gap between groups.
13269
13270 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
13271
13272 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
13273 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
13274 is not enabled.
13275
13276 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
13277
13278 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
13279 * doc/extend.texi (Named Address Spaces): Ditto.
13280 (Variable Attributes): Ditto.
13281
13282 2013-03-27 Kai Tietz <ktietz@redhat.com>
13283
13284 * config.build: Add support for cygwin x64 target.
13285 * config.gcc: Likewise.
13286 * config.host: Likewise.
13287 * configure.ac: Likewise
13288 * configure: Regenerated.
13289
13290 2013-03-27 Kai Tietz <ktietz@redhat.com>
13291
13292 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
13293 * config/i386/t-cygwin-w64: New file.
13294 * config/i386/cygwin-w64.h: New file.
13295 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
13296 and add support for x64-cygwin target.
13297 (CPP_SPEC): Likewise.
13298 (CXX_WRAP_SPEC_LIST): Undefine before define.
13299 (LIBGCJ_SONAME): Use 15 as version.
13300
13301 2013-03-27 Richard Biener <rguenther@suse.de>
13302
13303 PR tree-optimization/56716
13304 * tree-ssa-structalias.c (perform_var_substitution): Adjust
13305 dumping for ref nodes.
13306
13307 2013-03-27 Martin Jambor <mjambor@suse.cz>
13308
13309 PR tree-optimization/55334
13310 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
13311 restricted pointers to arrays.
13312
13313 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
13314
13315 * Makefile.in (.SUFFIXES): Add .cc.
13316 (.c.o): Apply same recipe for implicit rule .cc.o.
13317
13318 2013-03-27 Richard Biener <rguenther@suse.de>
13319
13320 PR tree-optimization/37021
13321 * tree-vect-data-refs.c (vect_check_strided_load): Allow
13322 REALPART/IMAGPART_EXPRs around the supported refs.
13323 * tree-ssa-structalias.c (find_func_aliases): Assume that
13324 floating-point values are not used to transfer pointers.
13325
13326 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
13327
13328 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
13329 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
13330 * gcc/doc/tm.texi: Regenerate.
13331 * gcc/targhooks.h (default_has_ifunc_p): New.
13332 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
13333 * gcc/config/linux-protos.h: New file.
13334 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
13335 this hook for linux which disables support of indirect functions in
13336 android.
13337 * gcc/config/linux-android.c: New file.
13338 * gcc/config/t-linux-android.c: Ditto.
13339 * gcc/config.gcc: Added new object file linux-android.o.
13340 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
13341 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
13342 * gcc/varasm.c (do_assemble_alias): Likewise.
13343 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
13344 doesn't support indirect functions.
13345 * configure: Regenerate.
13346
13347 2013-03-27 Bin Cheng <bin.cheng@arm.com>
13348
13349 PR target/56102
13350 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
13351 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
13352 mult-word mode.
13353
13354 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13355
13356 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
13357
13358 2013-03-27 Terry Guo <terry.guo@arm.com>
13359
13360 * config/arm/arm-cores.def: Added core cortex-r7.
13361 * config/arm/arm-tune.md: Regenerated.
13362 * config/arm/arm-tables.opt: Regenerated.
13363 * doc/invoke.texi: Added entry for core cortex-r7.
13364
13365 2013-03-27 Walter Lee <walt@tilera.com>
13366
13367 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
13368 double-decrement of next_scratch_regno.
13369
13370 2013-03-27 Walter Lee <walt@tilera.com>
13371
13372 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
13373 input operands.
13374 (insn_v1mulus): Ditto.
13375 (insn_v2muls): Ditto.
13376
13377 2013-03-27 Walter Lee <walt@tilera.com>
13378
13379 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
13380 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
13381
13382 2013-03-27 Walter Lee <walt@tilera.com>
13383
13384 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
13385 (*sibcall_value): Ditto.
13386
13387 2013-03-27 Walter Lee <walt@tilera.com>
13388
13389 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
13390 (insn_mnz_v8qi): ... this ...
13391 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
13392 vector equivalent.
13393 (insn_v<n>mnz): Replaced by ...
13394 (insn_v1mnz): ... this ...
13395 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
13396 equivalent.
13397 (insn_mz_<mode>): Replaced by ...
13398 (insn_mz_v8qi): ... this ...
13399 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
13400 vector equivalent.
13401 (insn_v<n>mz): Replaced by ...
13402 (insn_v1mz): ... this ...
13403 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
13404 equivalent.
13405
13406 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
13407
13408 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
13409
13410 2013-03-26 Roland McGrath <mcgrathr@google.com>
13411
13412 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
13413 than fprintf with a non-constant, non-format string.
13414
13415 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13416
13417 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
13418 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
13419 operand 0 predicate.
13420 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
13421 attribute. Use general_x64nomem_operand as operand 1 predicate.
13422 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
13423 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13424 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
13425 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
13426 (mov<mode>_insv_1): Remove expander. Merge insn with
13427 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
13428 Use general_x64nomem_operand as operand 1 predicate.
13429 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
13430 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
13431 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13432 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
13433 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
13434 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
13435 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
13436 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
13437 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
13438 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
13439 (general_x64nomem_operand): Ditto.
13440
13441 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
13442
13443 * config/rtems.opt: Add -pthread option.
13444
13445 2013-03-26 Richard Biener <rguenther@suse.de>
13446
13447 * alias.c (find_base_term): Avoid redundant and not used recursion.
13448 (base_alias_check): Get the initial base term from the caller.
13449 (true_dependence_1): Compute and pass base terms to base_alias_check.
13450 (write_dependence_p): Likewise.
13451 (may_alias_p): Likewise.
13452
13453 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
13454
13455 * config/aarch64/aarch64.c (aarch64_classify_address): Support
13456 PC-relative load in SI modes and above only.
13457
13458 2013-03-26 Xinyu Qi <xyqi@marvell.com>
13459
13460 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
13461 * config/arm/iwmmxt.md (WCGR0): Update.
13462 (WCGR1, WCGR2, WCGR3): Likewise.
13463
13464 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
13465
13466 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
13467 Use x64 and nox64 isa attributes.
13468
13469 2013-03-26 Richard Biener <rguenther@suse.de>
13470
13471 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
13472 alignment computations and rely on get_object_alignment_1
13473 for the !TYPE_P case.
13474 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
13475
13476 2013-03-26 Walter Lee <walt@tilera.com>
13477
13478 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
13479 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
13480
13481 2013-03-25 Jeff Law <law@redhat.com>
13482
13483 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
13484 check for INTEGRAL_TYPE_P that was missing due to checking in
13485 wrong version of prior patch.
13486
13487 2013-03-25 Walter Lee <walt@tilera.com>
13488
13489 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
13490 TILEGX_INSN_SHUFFLEBYTES1.
13491 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
13492 shufflebytes1.
13493 (tilegx_builtins): Ditto.
13494 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
13495
13496 2013-03-25 Walter Lee <walt@tilera.com>
13497
13498 * config/tilegx/tilegx.md (floatsisf2): New pattern.
13499 (floatunssisf2): New pattern.
13500 (floatsidf2): New pattern.
13501 (floatunssidf2): New pattern.
13502
13503 2013-03-25 Walter Lee <walt@tilera.com>
13504
13505 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
13506 tests for constraint J, K, N, P.
13507
13508 2013-03-25 Walter Lee <walt@tilera.com>
13509
13510 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
13511 Use indirect/pcrel encoding.
13512 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
13513 Ditto.
13514
13515 2013-03-25 Steve Ellcey <sellcey@mips.com>
13516
13517 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
13518 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
13519 * config/mips/mips.c (mips_option_override): Set IMADD default.
13520 * config/mips/mips.h (PTF_AVOID_IMADD): New.
13521 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
13522 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
13523 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
13524 * doc/invoke.texi (-mimadd/-mno-imadd): New.
13525
13526 2013-03-25 Jeff Law <law@redhat.com>
13527
13528 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
13529 slightly to avoid creating and folding useless trees. Simplify
13530 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
13531
13532 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
13533
13534 * config/i386/i386.md (*zero_extendsidi2): Merge with
13535 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
13536 * config/i386/predicates.md (x86_64_zext_operand): Rename from
13537 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
13538 targets. Clarify comment.
13539
13540 2013-03-25 Martin Jambor <mjambor@suse.cz>
13541
13542 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
13543 pass-through jump functions differently.
13544 (ipa_read_jump_function): Likewise. Also use setter functions to set
13545 up jump functions.
13546
13547 2013-03-25 Martin Jambor <mjambor@suse.cz>
13548
13549 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
13550 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
13551 process it.
13552 (ipa_get_indirect_edge_target): New function.
13553 (devirtualization_time_bonus): New parameter known_aggs, pass it to
13554 ipa_get_indirect_edge_target. Update all callers.
13555 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
13556 ipa_get_indirect_edge_target_1 instead of calling
13557 ipa_get_indirect_edge_target.
13558 (create_specialized_node): Pass aggvlas to
13559 ipcp_discover_new_direct_edges.
13560
13561 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13562
13563 * config/arm/arm.md (f_sels, f_seld): New types.
13564 (*cmov<mode>): New pattern.
13565 * config/arm/predicates.md (arm_vsel_comparison_operator): New
13566 predicate.
13567
13568 2013-03-25 Kai Tietz <ktietz@redhat.com>
13569
13570 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
13571 POSIX-printf for mingw-hosted builds.
13572
13573 2013-03-25 Richard Biener <rguenther@suse.de>
13574
13575 PR middle-end/56694
13576 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
13577 must-not-throw stmt location.
13578
13579 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13580
13581 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
13582 Emit load-acquire versions when acq is true.
13583 (arm_emit_store_exclusive): Add rel parameter.
13584 Emit store-release versions when rel is true.
13585 (arm_split_compare_and_swap): Use acquire-release instructions
13586 instead.
13587 of barriers when appropriate.
13588 (arm_split_atomic_op): Likewise.
13589 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
13590 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
13591 (VUNSPEC_SLX): Likewise.
13592 (VUNSPEC_LDA): Likewise.
13593 (VUNSPEC_STL): Likewise.
13594 * config/arm/sync.md (atomic_load<mode>): New pattern.
13595 (atomic_store<mode>): Likewise.
13596 (arm_load_acquire_exclusive<mode>): Likewise.
13597 (arm_load_acquire_exclusivesi): Likewise.
13598 (arm_load_acquire_exclusivedi): Likewise.
13599 (arm_store_release_exclusive<mode>): Likewise.
13600
13601 2013-03-25 Catherine Moore <clm@codesourcery.com>
13602
13603 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
13604 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
13605 * config/mip/predicates.md (lwsp_swsp_operand,
13606 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
13607 sb16_operand, db4_operand, db7_operand, ib3_operand,
13608 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
13609 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
13610 andi16_operand): New predicates.
13611 * config/mips/mips.md (compression): New attribute.
13612 (enabled): New attribute.
13613 (length): Consider compression in computing length.
13614 (shift_compression): New code attribute.
13615 (*add<mode>3): New operands. Record compression.
13616 (sub<mode>3): Likewise.
13617 (one_cmpl<mode>2): Likewise.
13618 (*and<mode>3): Likewise.
13619 (*ior<mode>3): Likewise.
13620 (unnamed pattern for xor): Likewise.
13621 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
13622 (*<optab><mode>3): Likewise.
13623 (*mov<mode>_internal: Likewise.
13624 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
13625 (mips_unsigned_immediate_p): New.
13626 (umips_lwsp_swsp_address_p): New.
13627 (m16_based_address_p): New.
13628 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
13629 (mips_unsigned_immediate_p): New prototype.
13630 (lwsp_swsp_address_p): New prototype.
13631 (m16_based_address_p): New prototype.
13632 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
13633 (mips_signed_immediate_p): New function.
13634 (m16_based_address_p): New function.
13635 (lwsp_swsp_address_p): New function.
13636 (mips_print_operand_punctuation): Recognize short delay slot insns
13637 for microMIPS.add<mode>3"
13638
13639 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13640
13641 PR target/56720
13642 * config/arm/iterators.md (v_cmp_result): New mode attribute.
13643 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
13644
13645 2013-03-25 Richard Biener <rguenther@suse.de>
13646
13647 PR tree-optimization/56689
13648 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
13649 any edge.
13650
13651 2013-03-25 Richard Biener <rguenther@suse.de>
13652
13653 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
13654 of bitmap.
13655 (memory_references): Likewise.
13656 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
13657 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
13658 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
13659 (gather_mem_refs_in_loops): Fold into ...
13660 (analyze_memory_references): ... this. Move initialization
13661 to tree_ssa_lim_initialize.
13662 (fill_always_executed_in): Rename to ...
13663 (fill_always_executed_in_1): ... this.
13664 (fill_always_executed_in): Move contains_call computation to
13665 this new function from ...
13666 (tree_ssa_lim_initialize): ... here.
13667 (tree_ssa_lim): Call fill_always_executed_in.
13668
13669 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13670
13671 * postreload.c (reload_combine): Fix code detecting returns.
13672
13673 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
13674
13675 * function.c (emit_use_return_register_into_block): On cc0 targets,
13676 do not emit the sequence between cc0 setter and user.
13677
13678 2013-03-25 Kai Tietz <ktietz@redhat.com>
13679
13680 * config/i386/predicates.md (local_symbolic_operand): Interpret
13681 dll-imported symbols as none-local.
13682
13683 2013-03-25 Richard Biener <rguenther@suse.de>
13684
13685 * tree-ssa-loop-im.c (struct depend): Remove.
13686 (struct lim_aux_data): Make depends a vec of gimples.
13687 (free_lim_aux_data): Adjust.
13688 (add_dependency): Likewise.
13689 (set_level): Likewise.
13690
13691 2013-03-25 Richard Biener <rguenther@suse.de>
13692
13693 PR middle-end/56434
13694 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
13695 the pointer returned by calls with ECF_MALLOC set.
13696
13697 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13698
13699 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
13700
13701 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
13702
13703 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
13704 using MMXMODE mode iterator.
13705 (*move<mode>_internal): Merge with *movv2sf_internal and
13706 *movv2sf_internal_rex64 using MMXMODE mode iterator.
13707
13708 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13709
13710 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
13711 (record_last_mem_set_info): Likewise.
13712
13713 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
13714 of XNEWVEC followed by memset.
13715 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
13716
13717 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
13718
13719 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
13720 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
13721 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
13722 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
13723 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
13724 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
13725 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
13726 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
13727 BARRIER_P instead of GET_CODE.
13728
13729 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13730
13731 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
13732 inaccuracy in the probing code.
13733
13734 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
13735 (ctrapdi4): Likewise.
13736
13737 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13738
13739 * calls.c (expand_call): Add missing guard to code handling return
13740 of non-BLKmode structures in MSB.
13741 * function.c (expand_function_end): Likewise.
13742
13743 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
13744
13745 * combine.c (try_combine): Adjust comment. Do not add the set of
13746 insn #0 if the destination indirectly is set or dies in insn #2.
13747 Tidy up code to distribute a new note.
13748
13749 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13750
13751 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
13752 also for alternatives 16 and 17.
13753
13754 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
13755
13756 * config/i386/sse.md (*mov<mode>_internal): Merge with
13757 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
13758 Emit insn template depending on type attribute. Use
13759 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
13760 movd instead of movq mnemonic for interunit moves. Rewrite mode
13761 attribute calculation. Remove unit attribute calculation.
13762 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
13763 Set prefix_data16 attribute for DImode ssemov types.
13764 Use Ym instead of y for SSE-MMX conversion alternatives.
13765 Reorder operand constraints.
13766
13767 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
13768
13769 * df.h (df_insn_delete): Adjust prototype.
13770 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
13771 and let it decide whether mark the basic block dirty.
13772 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
13773 * df-scan.c (df_insn_info_delete): New helper function, split
13774 off from df_insn_delete.
13775 (df_scan_free_bb_info): Use it.
13776 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
13777 Likewise.
13778 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
13779 that the insn is actually an insn and it has a non-NULL basic block.
13780 Do not mark basic block dirty if only deleting a DEBUG_INSN.
13781
13782 2013-03-22 Richard Biener <rguenther@suse.de>
13783
13784 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
13785 dep_ref members.
13786 (mem_ref_alloc): Do not allocate them.
13787 (refs_independent_p): Do not query or maintain a cache.
13788
13789 2013-03-22 Richard Biener <rguenther@suse.de>
13790
13791 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
13792 (gather_mem_refs_in_loops): Do not compute it.
13793 (analyze_memory_references): Do not allocate it.
13794 (tree_ssa_lim_finalize): Do not free it.
13795 (for_all_locs_in_loop): Do not query all_refs_in_loop.
13796
13797 2013-03-22 Richard Biener <rguenther@suse.de>
13798
13799 * is-a.h (as_a): Use gcc_checking_assert.
13800
13801 2013-03-22 Ian Bolton <ian.bolton@arm.com>
13802
13803 * config/aarch64/aarch64.c (aarch64_print_operand): New
13804 format specifier for printing a constant in hex.
13805 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
13806 format specifier for printing second operand.
13807
13808 2013-03-22 Richard Biener <rguenther@suse.de>
13809
13810 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
13811 bitmaps.
13812 (gather_mem_refs_in_loops): Perform store accumulation here.
13813 (create_vop_ref_mapping_loop): Remove.
13814 (create_vop_ref_mapping): Likewise.
13815 (analyze_memory_references): Initialize refs_stored_in_loop.
13816 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
13817 (record_indep_loop): Remove.
13818 (record_dep_loop): New function.
13819 (ref_indep_loop_p_1): Adjust to only walk over references
13820 in the loop, not its subloops.
13821 (ref_indep_loop_p): Rename to ...
13822 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
13823 maintaining a more fine-grained cache.
13824 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
13825 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
13826
13827 2013-03-22 Richard Biener <rguenther@suse.de>
13828
13829 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
13830 (struct mem_ref): Make accesses_in_loop a vec of a vec of
13831 aggregate mem_ref_loc.
13832 (free_mem_ref_locs): Inline into ...
13833 (memref_free): ... this and adjust.
13834 (mem_ref_alloc): Adjust.
13835 (mem_ref_locs_alloc): Remove.
13836 (record_mem_ref_loc): Adjust.
13837 (get_all_locs_in_loop): Rewrite into ...
13838 (for_all_locs_in_loop): ... this iterator.
13839 (rewrite_mem_ref_loc): New functor.
13840 (rewrite_mem_refs): Use for_all_locs_in_loop.
13841 (sm_set_flag_if_changed): New functor.
13842 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
13843 (ref_always_accessed): New functor.
13844 (ref_always_accessed_p): Use for_all_locs_in_loop.
13845
13846 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13847
13848 * tree-pass.h (PROP_gimple_lvec): New.
13849 * passes.c (dump_properties): Handle PROP_gimple_lvec.
13850 (init_optimization_passes): Move pass_lower_vector.
13851 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
13852 PROP_gimple_lvec.
13853 (pass_lower_vector): Provide PROP_gimple_lvec.
13854 (pass_lower_vector_ssa): Likewise.
13855 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
13856
13857 2013-03-21 Mark Wielaard <mjw@redhat.com>
13858
13859 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
13860
13861 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
13862
13863 * config/i386/i386.md (*movdi_internal): Disparage slightly
13864 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
13865 conversion alternatives.
13866
13867 2013-03-21 Jakub Jelinek <jakub@redhat.com>
13868
13869 PR middle-end/48087
13870 * diagnostic.def (DK_WERROR): New kind.
13871 * diagnostic.h (werrorcount): Define.
13872 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
13873 promoted to DK_ERROR, increment DK_WERROR counter instead of
13874 DK_ERROR counter.
13875 * toplev.c (toplev_main): Call print_ignored_options even if
13876 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
13877 even if just werrorcount is non-zero.
13878
13879 PR debug/55608
13880 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
13881 on failure.
13882 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
13883 (string_cst_pool_decl): New function.
13884 (optimize_one_addr_into_implicit_ptr): New function.
13885 (resolve_addr_in_expr): Optimize DWARF location expression
13886 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
13887 which doesn't live in memory, but has DW_AT_location or
13888 DW_AT_const_value, or refers to a string literal, into
13889 DW_OP_GNU_implicit_pointer.
13890 (optimize_location_into_implicit_ptr): New function.
13891 (resolve_addr): If removing DW_AT_location of a variable because
13892 it was DW_OP_addr of address of the variable, but the variable doesn't
13893 live in memory, try to emit const value attribute for the initializer.
13894
13895 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13896
13897 * tree.h (VECTOR_TYPE_P): New macro.
13898 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
13899 TYPE_MODE): Use it.
13900 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
13901 VEC_COND_EXPR cannot be lvalues.
13902 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
13903
13904 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
13905
13906 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
13907 Restrict the transformation to equal modes.
13908
13909 2013-03-21 Richard Biener <rguenther@suse.de>
13910
13911 PR tree-optimization/39326
13912 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
13913 (MEM_ANALYZABLE): Adjust.
13914 (record_mem_ref_loc): Move bitmap ops ...
13915 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
13916 unanalyzable refs, do not record locations for it.
13917 (analyze_memory_references): Allocate ref zero as shared
13918 unanalyzable ref.
13919 (refs_independent_p): Do not test for unanalyzed mems here.
13920 (ref_indep_loop_p_1): Special-case disambiguation against
13921 the unanalyzed ref.
13922 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
13923
13924 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
13925
13926 * config/arm/arm-protos.h (tune_params): Add
13927 prefer_neon_for_64bits field.
13928 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
13929 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
13930 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
13931 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
13932 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
13933 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
13934 (arm_option_override): Handle -mneon-for-64bits new option.
13935 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
13936 (prefer_neon_for_64bits): Declare new variable.
13937 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
13938 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
13939 (arch_enabled): Handle new arch types. Remove support for onlya8
13940 and nota8.
13941 (one_cmpldi2): Use new arch names.
13942 (zero_extend<mode>di2, extend<mode>di2): Ditto.
13943 * config/arm/arm.opt (mneon-for-64bits): Add option.
13944 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
13945 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
13946 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
13947 of onlya8.
13948 * doc/invoke.texi (-mneon-for-64bits): Document.
13949
13950 2013-03-21 Richard Biener <rguenther@suse.de>
13951
13952 PR tree-optimization/39326
13953 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
13954 (sort_bbs_in_loop_postorder_cmp): New function.
13955 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
13956
13957 2013-03-21 Richard Biener <rguenther@suse.de>
13958
13959 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
13960 (vect_insert_into_interleaving_chain): Likewise.
13961 (vect_drs_dependent_in_basic_block): Inline ...
13962 (vect_slp_analyze_data_ref_dependence): ... here. New function,
13963 split out from ...
13964 (vect_analyze_data_ref_dependence): ... here. Simplify.
13965 (vect_check_interleaving): Simplify.
13966 (vect_analyze_data_ref_dependences): Likewise. Split out ...
13967 (vect_slp_analyze_data_ref_dependences): ... this new function.
13968 (dr_group_sort_cmp): New function.
13969 (vect_analyze_data_ref_accesses): Compute data-reference groups
13970 here instead of in vect_analyze_data_ref_dependence. Use
13971 a more efficient algorithm.
13972 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
13973 vect_slp_analyze_data_ref_dependences. Call
13974 vect_analyze_data_ref_accesses earlier.
13975 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
13976 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
13977 (vect_slp_analyze_data_ref_dependences): New prototype.
13978
13979 2013-03-21 Richard Biener <rguenther@suse.de>
13980
13981 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
13982 ref is stored in the loop.
13983 (find_refs_for_sm): Walk only over all stores.
13984 (store_motion_loop): Allocate from lim_bitmap_obstack.
13985 (store_motion): Likewise.
13986
13987 2013-03-21 Richard Biener <rguenther@suse.de>
13988
13989 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
13990 Update virtual SSA form.
13991
13992 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13993
13994 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
13995 * configure: Regenerate.
13996 * config.in: Regenerate.
13997 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
13998 if !HAVE_LD_EH_FRAME_CIEV3.
13999
14000 2013-03-21 Richard Biener <rguenther@suse.de>
14001
14002 * tree-cfg.c (verify_expr_no_block): New function.
14003 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14004 nor DECL_VALUE_EXPR have locations with associated blocks.
14005 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14006 (clear_unused_block_pointer): Remove code dealing with
14007 blocks in DECL_DEBUG_EXPR locations.
14008
14009 2013-03-21 Richard Biener <rguenther@suse.de>
14010
14011 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14012 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14013 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14014 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14015 instead of DECL_DEBUG_EXPR_IS_FROM.
14016 * gimplify.c (gimplify_modify_expr): Likewise.
14017 * tree-cfg.c (verify_expr_location_1): Likewise.
14018 * tree-complex.c (create_one_component_var): Likewise.
14019 * tree-sra.c (create_access_replacement): Likewise.
14020 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14021 (clear_unused_block_pointer): Likewise.
14022 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14023 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14024 * var-tracking.c (var_debug_decl): Likewise.
14025 (track_expr_p): Likewise.
14026 * tree-inline.c (add_local_variables): Likewise. Set
14027 DECL_HAS_DEBUG_EXPR_P after copying it.
14028 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14029 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14030
14031 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14032
14033 PR bootstrap/56656
14034 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14035 * configure: Regenerate.
14036 * config.in: Regenerate.
14037 * config/i386/i386.md (*movdf_internal): Use
14038 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14039 movd instead of movq mnemonic for interunit moves.
14040 (*movdi_internal): Ditto.
14041
14042 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14043
14044 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14045 (abd<mode>_3): New pattern.
14046 (aba<mode>_3): New pattern.
14047 (fabd<mode>_3): New pattern.
14048
14049 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14050
14051 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14052 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14053 occurrence of REGISTER_PREFIX as its empty string.
14054
14055 2013-03-20 Jeff Law <law@redhat.com>
14056
14057 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14058 addititional equivalences for equality comparisons between an SSA_NAME
14059 and a constant where the SSA_NAME was set from a widening conversion.
14060
14061 2013-03-20 Walter Lee <walt@tilera.com>
14062
14063 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14064
14065 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14066
14067 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14068 depending on type attribute.
14069 (*movti_internal): Ditto.
14070 (*movtf_internal): Ditto.
14071 (*movxf_internal): Ditto.
14072 (*movdf_internal): Ditto.
14073 (*movsf_internal): Ditto.
14074
14075 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14076
14077 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14078 maybe_vex for sselog1 and ssemov types.
14079 (*movdi_internal): Reorder operand constraints.
14080 (*movsi_internal): Ditto. Set prefix attribute to
14081 maybe_vex for sselog1 and ssemov types.
14082 (*movtf_internal): Set prefix attribute to maybe_vex
14083 for sselog1 and ssemov types.
14084 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14085 DImode ssemov types. Reorder operand constraints.
14086 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14087 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14088 attribute for SImode ssemov types. Reorder operand constraints.
14089
14090 2013-03-20 Martin Jambor <mjambor@suse.cz>
14091
14092 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14093 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
14094
14095 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
14096
14097 * config/rs6000/predicates.md (indexed_address, update_address_mem
14098 update_indexed_address_mem): New predicates.
14099 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
14100 attribute for load/store instructions.
14101 * config/rs6000/dfp.md (movsd_store): Likewise.
14102 (movsd_load): Likewise.
14103 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
14104 (unnamed HI->DI extend define_insn): Likewise.
14105 (unnamed SI->DI extend define_insn): Likewise.
14106 (unnamed QI->SI extend define_insn): Likewise.
14107 (unnamed QI->HI extend define_insn): Likewise.
14108 (unnamed HI->SI extend define_insn): Likewise.
14109 (unnamed HI->SI extend define_insn): Likewise.
14110 (extendsfdf2_fpr): Likewise.
14111 (movsi_internal1): Likewise.
14112 (movsi_internal1_single): Likewise.
14113 (movhi_internal): Likewise.
14114 (movqi_internal): Likewise.
14115 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
14116 attribute for load/store instructions.
14117 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
14118 instructions.
14119 (mov<mode>_softfloat): Likewise.
14120 (mov<mode>_hardfloat32): Likewise.
14121 (mov<mode>_hardfloat64): Likewise.
14122 (mov<mode>_softfloat64): Likewise.
14123 (movdi_internal32): Likewise.
14124 (movdi_internal64): Likewise.
14125 (probe_stack_<mode>): Likewise.
14126
14127 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
14128
14129 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
14130 floating point, and decimal floating point to reload iterator.
14131
14132 * config/rs6000/constraints.md (wl constraint): New constraints to
14133 return FLOAT_REGS if certain options are used to reduce the number
14134 of separate patterns that exist in the file.
14135 (wx constraint): Likewise.
14136 (wz constraint): Likewise.
14137
14138 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14139 -mdebug=reg, print wg, wl, wx, and wz constraints.
14140 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14141 Initialize the reload functions for 64-bit binary/decimal floating
14142 point types.
14143 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14144 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14145 create the buffer on the stack to overcome not having a 32-bit
14146 load and store.
14147 (rs6000_emit_move): Likewise.
14148 (rs6000_secondary_memory_needed_rtx): Likewise.
14149 (rs6000_alloc_sdmode_stack_slot): Likewise.
14150 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14151 via xxlxor, just like DFmode 0.0.
14152
14153 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
14154 define as 1 if we are running on a power7 or newer.
14155 (enum r6000_reg_class_enum): Add new constraints.
14156
14157 * config/rs6000/dfp.md (movsd): Delete, combine with binary
14158 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14159 with other moves by using conditional constraits (wg). Use LFIWZX
14160 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
14161 (movsd splitter): Likewise.
14162 (movsd_hardfloat): Likewise.
14163 (movsd_softfloat): Likewise.
14164
14165 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
14166 binary and decimal floating point moves.
14167 (fmove_ok): New attributes to combine binary and decimal floating
14168 point moves, and to combine power6x (mfpgpr) moves along normal
14169 floating moves.
14170 (real_value_to_target): Likewise.
14171 (f32_lr): Likewise.
14172 (f32_lm): Likewise.
14173 (f32_li): Likewise.
14174 (f32_sr): Likewise.
14175 (f32_sm): Likewise.
14176 (f32_si): Likewise.
14177 (movsf): Combine binary and decimal floating point moves. Combine
14178 power6x (mfpgpr) moves with other moves by using conditional
14179 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
14180 (mov<mode> for SFmode/SDmode); Likewise.
14181 (SFmode/SDmode splitters): Likewise.
14182 (movsf_hardfloat): Likewise.
14183 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
14184 (movsf_softfloat): Likewise.
14185 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
14186
14187 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
14188 wx and wz constraints.
14189
14190 * config/rs6000/constraints.md (wg constraint): New constraint to
14191 return FLOAT_REGS if -mmfpgpr (power6x) was used.
14192
14193 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
14194 constraint.
14195
14196 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
14197 -mdebug=reg, print wg, wl, wx, and wz constraints.
14198 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
14199 Initialize the reload functions for 64-bit binary/decimal floating
14200 point types.
14201 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
14202 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
14203 create the buffer on the stack to overcome not having a 32-bit
14204 load and store.
14205 (rs6000_emit_move): Likewise.
14206 (rs6000_secondary_memory_needed_rtx): Likewise.
14207 (rs6000_alloc_sdmode_stack_slot): Likewise.
14208 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
14209 via xxlxor, just like DFmode 0.0.
14210
14211 * config/rs6000/dfp.md (movdd): Delete, combine with binary
14212 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
14213 with other moves by using conditional constraits (wg). Use LFIWZX
14214 and STFIWX for loading SDmode on power7.
14215 (movdd splitters): Likewise.
14216 (movdd_hardfloat32): Likewise.
14217 (movdd_softfloat32): Likewise.
14218 (movdd_hardfloat64_mfpgpr): Likewise.
14219 (movdd_hardfloat64): Likewise.
14220 (movdd_softfloat64): Likewise.
14221
14222 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
14223 64-bit binary and decimal floating point moves.
14224 (FMOVE64X): Likewise.
14225 (movdf): Combine 64-bit binary and decimal floating point moves.
14226 Combine power6x (mfpgpr) moves with other moves by using
14227 conditional constraits (wg).
14228 (mov<mode> for DFmode/DDmode): Likewise.
14229 (DFmode/DDmode splitters): Likewise.
14230 (movdf_hardfloat32): Likewise.
14231 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
14232 (movdf_softfloat32): Likewise.
14233 (movdf_hardfloat64_mfpgpr): Likewise.
14234 (movdf_hardfloat64): Likewise.
14235 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
14236 (movdf_softfloat64): Likewise.
14237 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
14238 (reload_<mode>_load): Move to later in the file so they aren't in
14239 the middle of the floating point move insns.
14240 (reload_<mode>_store): Likewise.
14241
14242 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
14243 constraint.
14244
14245 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
14246 constraint if -mdebug=reg.
14247 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
14248 Enable using dd reload support if needed.
14249
14250 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
14251 binary and decimal floating point moves in rs6000.md.
14252 (movtd_internal): Likewise.
14253
14254 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
14255 decimal floating point moves.
14256 (movtf): Likewise.
14257 (movtf_internal): Likewise.
14258 (mov<mode>_internal, TDmode/TFmode): Likewise.
14259 (movtf_softfloat): Likewise.
14260 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
14261
14262 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
14263 movdi_internal64, using wg constraint for move direct operations.
14264 (movdi_internal64): Likewise.
14265
14266 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
14267 MODES_TIEABLE_P for selected modes. Print the numerical value of
14268 the various virtual registers. Use GPR/FPR first/last values,
14269 instead of hard coding the register numbers. Print which modes
14270 have reload functions registered.
14271 (rs6000_option_override_internal): If -mdebug=reg, trace the options
14272 settings before/after setting cpu, target and subtarget settings.
14273 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
14274 and for secondary reload failures in rs6000_secondary_reload_inner.
14275 (rs6000_secondary_reload_fail): Likewise.
14276 (rs6000_secondary_reload_inner): Likewise.
14277
14278 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
14279 macros for first/last GPR and FPR registers.
14280 (LAST_GPR_REGNO): Likewise.
14281 (FIRST_FPR_REGNO): Likewise.
14282 (LAST_FPR_REGNO): Likewise.
14283
14284 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
14285 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
14286 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
14287 (vcond<mode><mode>): Likewise.
14288 (vcondu<mode><mode>): Likewise.
14289 (vector_gtu<mode>): Likewise.
14290 (vector_gte<mode>): Likewise.
14291 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
14292 to prevent the compiler from converting DImode operations to TImode.
14293 (ior<mode>3): Likewise.
14294 (and<mode>3): Likewise.
14295 (one_cmpl<mode>2): Likewise.
14296 (nor<mode>3): Likewise.
14297 (andc<mode>3): Likewise.
14298
14299 * config/rs6000/constraints.md (wt constraint): New constraint
14300 that returns VSX_REGS if TImode is allowed in VSX registers.
14301
14302 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
14303 constant under VSX.
14304
14305 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
14306 similar to TImode, but it is restricted to being in the GPRs.
14307
14308 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
14309 TImode to occupy a single VSX register.
14310
14311 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
14312 -mvsx-timode for power7/power8.
14313 (power7 cpu): Likewise.
14314 (power8 cpu): Likewise.
14315
14316 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
14317 sure that TFmode/TDmode take up two registers if they are ever
14318 allowed in the upper VSX registers.
14319 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
14320 registers.
14321 (rs6000_init_hard_regno_mode_ok): Likewise.
14322 (rs6000_debug_reg_global): Add debugging for PTImode and wt
14323 constraint. Print if LRA is turned on.
14324 (rs6000_option_override_internal): Give an error if -mvsx-timode
14325 and VSX is not enabled.
14326 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
14327 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
14328 to reg+offset addressing. Use PTImode when checking offset
14329 addresses for validity.
14330 (reg_offset_addressing_ok_p): Likewise.
14331 (rs6000_legitimate_offset_address_p): Likewise.
14332 (rs6000_legitimize_address): Likewise.
14333 (rs6000_legitimize_reload_address): Likewise.
14334 (rs6000_legitimate_address_p): Likewise.
14335 (rs6000_eliminate_indexed_memrefs): Likewise.
14336 (rs6000_emit_move): Likewise.
14337 (rs6000_secondary_reload): Likewise.
14338 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
14339 reloads to fpr registers to continue to use reg+offset addressing,
14340 but 64-bit reloads to altivec registers need reg+reg addressing.
14341 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
14342 it. Treat LO_SUM like a PLUS operation.
14343 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
14344 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
14345 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
14346 registers to share a register with a smaller sized type, since VSX
14347 puts scalars in the upper 64-bits.
14348 (print_operand): Add support for PTImode.
14349 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
14350 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
14351 registers, but don't have arithmetic support.
14352 (rs6000_memory_move_cost): Add test for VSX.
14353 (rs6000_opt_masks): Add -mvsx-timode.
14354
14355 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
14356 for TImode.
14357 (VSs): Likewise.
14358 (VSr): Use wt constraint for TImode.
14359 (VSv): Drop TImode support.
14360 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
14361 (vsx_movti_64bit): Likewise.
14362 (vsx_movti_32bit): Likewise.
14363 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
14364 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
14365 one '?' on the appropriate output constraint. Do not allow TImode
14366 logical operations on 32-bit systems.
14367 (vsx_ior<mode>3): Likewise.
14368 (vsx_xor<mode>3): Likewise.
14369 (vsx_one_cmpl<mode>2): Likewise.
14370 (vsx_nor<mode>3): Likewise.
14371 (vsx_andc<mode>3): Likewise.
14372 (vsx_concat_<mode>): Likewise.
14373 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
14374
14375 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
14376 OPTION_MASK_VSX_TIMODE.
14377 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
14378 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
14379
14380 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
14381 (TI2 iterator): New iterator for TImode, PTImode.
14382 (wd mode attribute): Add values for vector types.
14383 (movti_string): Replace TI move operations with operations for TImode
14384 and PTImode. Add support for TImode being allowed in VSX registers.
14385 (mov<mode>_string, TImode/PTImode): Likewise.
14386 (movti_ppc64): Likewise.
14387 (mov<mode>_ppc64, TImode/PTImode): Likewise.
14388 (TI mode splitters): Likewise.
14389
14390 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
14391 constraint.
14392
14393 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
14394
14395 PR tree-optimization/56355
14396 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
14397 Also handle integers with undefined overflow.
14398
14399 2013-03-20 Catherine Moore <clm@codesourcery.com>
14400 Maciej W. Rozycki <macro@codesourcery.com>
14401 Tom de Vries <tom@codesourcery.com>
14402 Nathan Sidwell <nathan@codesourcery.com>
14403 Iain Sandoe <iain@codesourcery.com>
14404 Nathan Froyd <froydnj@codesourcery.com>
14405 Chao-ying Fu <fu@mips.com>
14406
14407 * doc/extend.texi: (micromips, nomicromips, nocompression):
14408 Document new function attributes.
14409 * doc/invoke.texi (minterlink-compressed, mmicromips,
14410 m14k, m14ke, m14kec): Document new options.
14411 (minterlink-mips16): Update documentation.
14412 * doc/md.texi (ZC, ZD): Document new constraints.
14413 * configure.ac (gcc_cv_as_micromips): Check if linker
14414 supports the .set micromips directive.
14415 * configure: Regenerate.
14416 * config.in: Regenerate.
14417 * config/mips/mips-tables.opt: Regenerate.
14418 * config/mips/micromips.md: New file.
14419 * constraints.md (ZC, ZD): New constraints.
14420 * config/mips/predicates.md (movep_src_register): New predicate.
14421 (movep_src_operand): New predicate.
14422 (non_volatile_mem_operand): New predicate.
14423 * config/mips/mips.md (multimem): New type.
14424 (length): Differentiate between 17-bit and 18-bit branch offsets.
14425 (MOVEP1, MOVEP2): New mode iterator.
14426 (mov_<load>l): Use ZC constraint.
14427 (mov_<load>r): Likewise.
14428 (mov_<store>l): Likewise.
14429 (mov_<store>r): Likewise.
14430 (*branch_equality<mode>_inverted): Add microMIPS support.
14431 (*branch_equality<mode>): Likewise.
14432 (*jump_absolute): Likewise.
14433 (indirect_jump_<mode>): Likewise.
14434 (tablejump_<mode>): Likewise.
14435 (<optab>_internal): Likewise.
14436 (sibcall_internal): Likewise.
14437 (sibcall_value_internal): Likewise.
14438 (prefetch): Use constraint ZD.
14439 * config/mips/mips.opt (minterlink-compressed): New option.
14440 (minterlink-mips16): Now an alias for minterlink-compressed.
14441 (mmicromips): New option.
14442 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
14443 (compare_and_swap_12): Likewise.
14444 (sync_add<mode>): Likewise.
14445 (sync_<optab>_12): Likewise.
14446 (sync_old_<optab>_12): Likewise.
14447 (sync_new_<optab>_12): Likewise.
14448 (sync_nand_12): Likewise.
14449 (sync_old_nand_12): Likewise.
14450 (sync_new_nand_12): Likewise.
14451 (sync_sub<mode>): Likewise.
14452 (sync_old_add<mode>): Likewise.
14453 (sync_old_sub<mode>): Likewise.
14454 (sync_new_add<mode>): Likewise.
14455 (sync_new_sub<mode>): Likewise.
14456 (sync_<optab><mode>): Likewise.
14457 (sync_old_<optab><mode>): Likewise.
14458 (sync_new_<optab><mode>): Likewise.
14459 (sync_nand<mode>): Likewise.
14460 (sync_old_nand<mode>): Likewise.
14461 (sync_new_nand<mode>): Likewise.
14462 (sync_lock_test_and_set<mode>): Likewise.
14463 (test_and_set_12): Likewise.
14464 (atomic_compare_and_swap<mode>): Likewise.
14465 (atomic_exchange<mode>_llsc): Likewise.
14466 (atomic_fetch_add<mode>_llsc): Likewise.
14467 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
14468 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
14469 (umips_save_restore_pattern_p): Likewise.
14470 (umips_load_store_pair_p): Likewise.
14471 (umips_output_load_store_pair): Likewise.
14472 (umips_movep_target_p): Likewise.
14473 (umips_12bit_offset_address_p): Likewise.
14474 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
14475 (mips_base_mips16): Rename this...
14476 (mips_base_compression_flags): ...to this. Update all uses.
14477 (mips_attribute_table): Add micromips, nomicromips and nocompression.
14478 (mips_mips16_decl_p): Delete.
14479 (mips_nomips16_decl_p): Delete.
14480 (mips_get_compress_on_flags): New function.
14481 (mips_get_compress_off_flags): New function.
14482 (mips_get_compress_mode): New function.
14483 (mips_get_compress_on_name): New function.
14484 (mips_get_compress_off_name): New function.
14485 (mips_insert_attributes): Support multiple compression types.
14486 (mips_merge_decl_attributes): Likewise.
14487 (umips_12bit_offset_address_p): New function.
14488 (mips_start_function_definition): Emit .set micromips directive.
14489 (mips_call_may_need_jalx_p): New function.
14490 (mips_function_ok_for_sibcall): Add microMIPS support.
14491 (mips_print_operand_punctuation): Support short delay slots and
14492 compact jumps.
14493 (umips_swm_mask, umips_swm_encoding): New.
14494 (umips_build_save_restore): New function.
14495 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
14496 (was_mips16_p): Remove.
14497 (old_compression_mode): New.
14498 (mips_set_compression_mode): New function.
14499 (mips_set_current_function): Add microMIPS support.
14500 (mips_option_override): Likewise.
14501 (umips_save_restore_pattern_p): New function.
14502 (umips_output_save_restore): New function.
14503 (umips_load_store_pair_p_1): New function.
14504 (umips_load_store_pair_p): New function.
14505 (umips_output_load_store_pair_1): New function.
14506 (umips_output_load_store_pair): New function.
14507 (umips_movep_target_p) New function.
14508 (mips_prepare_pch_save): Add microMIPS support.
14509 * config/mips/mips.h (TARGET_COMPRESSION): New.
14510 (TARGET_CPU_CPP_BUILTINS): Update macro
14511 to use new compression flags and to support microMIPS.
14512 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
14513 (MIPS_ARCH_FLOAT_SPEC): Likewise.
14514 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
14515 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
14516 (ASM_SPEC): Support mmicromips and mno-micromips.
14517 (M16STORE_REG_P): New macro.
14518 (MIPS_CALL): Support TARGET_MICROMIPS.
14519 (MICROMIPS_J): New macro.
14520 (mips_base_mips16): Rename this...
14521 (mips_base_compression_flags): ...to this.
14522 (UMIPS_12BIT_OFFSET_P): New macro.
14523 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
14524 (MULTILIB_DIRNAMES): Likewise.
14525 2013-03-20 Richard Biener <rguenther@suse.de>
14526
14527 PR tree-optimization/56661
14528 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
14529 the result does not have to be distinct.
14530
14531 2013-03-20 Richard Biener <rguenther@suse.de>
14532
14533 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
14534 remap_gimple_op_r.
14535
14536 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14537 Steven Bosscher <steven@gcc.gnu.org>
14538
14539 PR rtl-optimization/56605
14540 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
14541
14542 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14543
14544 PR bootstrap/56656
14545 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
14546 that require movd instead of movq.
14547
14548 2013-03-20 Richard Biener <rguenther@suse.de>
14549
14550 * tree-ssa-structalias.c (struct variable_info): Add pointer
14551 to the first field of an aggregate with sub-vars. Make
14552 this and the pointer to the next subfield its ID.
14553 (vi_next): New function.
14554 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
14555 storedanything_id, integer_id): Increment by one.
14556 (new_var_info, get_call_vi, lookup_call_clobber_vi,
14557 get_call_clobber_vi): Adjust.
14558 (solution_set_expand): Simplify and speedup.
14559 (solution_set_add): Inline into ...
14560 (set_union_with_increment): ... this. Adjust accordingly.
14561 (do_sd_constraint): Likewise.
14562 (do_ds_constraint): Likewise.
14563 (do_complex_constraint): Simplify.
14564 (build_pred_graph): Adjust.
14565 (solve_graph): Likewise. Simplify and speedup.
14566 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
14567 get_constraint_for_component_ref, get_constraint_for_1,
14568 first_vi_for_offset, first_or_preceding_vi_for_offset,
14569 create_function_info_for, create_variable_info_for_1,
14570 create_variable_info_for, intra_create_variable_infos): Adjust.
14571 (init_base_vars): Push NULL for ID zero.
14572 (compute_points_to_sets): Adjust.
14573
14574 2013-03-20 Richard Biener <rguenther@suse.de>
14575
14576 * cfgloop.c (verify_loop_structure): Streamline and avoid
14577 ICEing on corrupt loop tree.
14578 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
14579 loop tree.
14580
14581 2013-03-20 Richard Biener <rguenther@suse.de>
14582
14583 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
14584 check whether an SSA update is needed.
14585
14586 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
14587
14588 * config/mips/constraints.md (T): Rename to...
14589 (Yf): ...this.
14590 (U): Rename to...
14591 (Yd): ...this.
14592 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
14593 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
14594
14595 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14596
14597 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
14598 (*subsi3_carryin_uxtw): Likewise.
14599
14600 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14601
14602 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
14603 (*rorsi3_insn_uxtw): Likewise.
14604
14605 2013-03-19 Ian Bolton <ian.bolton@arm.com>
14606
14607 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
14608 (*extrsi5_insn_uxtw): Likewise.
14609
14610 2013-03-19 Richard Biener <rguenther@suse.de>
14611
14612 PR tree-optimization/56273
14613 * passes.c (init_optimization_passes): Move second VRP after DOM.
14614
14615 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
14616
14617 * config/i386/i386.md (*movti_internal): Merge from
14618 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
14619 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
14620 nox64 isa attributes.
14621
14622 2013-03-18 Richard Biener <rguenther@suse.de>
14623
14624 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
14625 (unite): Likewise.
14626 (merge_node_constraints): Likewise.
14627 (build_succ_graph): Likewise.
14628 (valid_graph_edge): Inline into single caller.
14629 (unify_nodes): Likewise. Use bitmap_set_bit return value
14630 and cache varinfo.
14631 (scc_visit): Fix formatting and variable use.
14632 (do_sd_constraint): Use gcc_checking_assert.
14633 (do_ds_constraint): Likewise.
14634 (do_complex_constraint): Likewise.
14635 (condense_visit): Likewise. Cleanup.
14636 (dump_pred_graph): New function.
14637 (perform_var_substitution): Dump the pred-graph before
14638 variable substitution.
14639 (find_equivalent_node): Use gcc_checking_assert.
14640 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
14641
14642 2013-03-18 Richard Biener <rguenther@suse.de>
14643
14644 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
14645 Remove cond_expr_stmt_list argument and do not gimplify the
14646 built expression.
14647 (vect_loop_versioning): Adjust.
14648 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
14649 Cleanup to use less temporaries.
14650 (vect_create_data_ref_ptr): Cleanup.
14651
14652 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14653
14654 PR tree-optimization/56635
14655 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
14656 require types_compatible_p types.
14657
14658 2013-03-18 Nick Clifton <nickc@redhat.com>
14659
14660 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
14661 spurious backslash.
14662
14663 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
14664 Add missing line to comment describing function.
14665
14666 2013-03-18 Richard Biener <rguenther@suse.de>
14667
14668 PR tree-optimization/56210
14669 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
14670 Handle string / character search functions.
14671 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
14672
14673 2013-03-18 Richard Biener <rguenther@suse.de>
14674
14675 PR middle-end/56483
14676 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
14677 and implement properly.
14678 * gimple.h (gimple_cond_single_var_p): Remove.
14679
14680 2013-03-18 Richard Biener <rguenther@suse.de>
14681
14682 * tree-data-ref.h (find_data_references_in_loop): Declare.
14683 * tree-data-ref.c (get_references_in_stmt): Use a stack
14684 vector pre-allocated in the callers.
14685 (find_data_references_in_stmt): Adjust.
14686 (graphite_find_data_references_in_stmt): Likewise.
14687 (create_rdg_vertices): Likewise.
14688 (find_data_references_in_loop): Export.
14689 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
14690 Compute dependences here...
14691 (vect_analyze_data_refs): ...not here. When we encounter
14692 a non-vectorizable data reference in basic-block vectorization
14693 truncate the data reference vector. Do not bother to
14694 fixup data-dependence information for gather loads.
14695 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
14696 of data references, as reported.
14697
14698 2013-03-18 Richard Biener <rguenther@suse.de>
14699
14700 PR tree-optimization/3713
14701 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
14702 has_constants and expr.
14703 (stmt_has_constants): Properly valueize SSA names when deciding
14704 whether the stmt has constants.
14705
14706 2013-03-18 Richard Biener <rguenther@suse.de>
14707
14708 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
14709 whole function when there is nothing to do.
14710 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
14711 * tree-vectorizer.c (vectorize_loops): Update virtual and
14712 loop-closed SSA once.
14713 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
14714
14715 2013-03-18 Richard Biener <rguenther@suse.de>
14716
14717 PR middle-end/56113
14718 * domwalk.c (bb_postorder): New global static.
14719 (cmp_bb_postorder): New function.
14720 (walk_dominator_tree): Replace scheme imposing an order for
14721 visiting dominator sons by one sorting them at the time they
14722 are pushed on the stack.
14723
14724 2013-03-18 Richard Biener <rguenther@suse.de>
14725
14726 PR tree-optimization/39326
14727 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
14728 (struct mem_ref): Replace mem member with ao_ref typed member.
14729 (MEM_ANALYZABLE): Adjust.
14730 (memref_eq): Likewise.
14731 (mem_ref_alloc): Likewise.
14732 (gather_mem_refs_stmt): Likewise.
14733 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
14734 (execute_sm_if_changed_flag_set): Adjust.
14735 (execute_sm): Likewise.
14736 (ref_always_accessed_p): Likewise.
14737 (refs_independent_p): Likewise.
14738 (can_sm_ref_p): Likewise.
14739
14740 2013-03-18 Jakub Jelinek <jakub@redhat.com>
14741
14742 PR c/56566
14743 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
14744 return 1 even for !unsignedp.
14745
14746 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14747
14748 * config/i386/i386.md (isa): Add x64 and nox64.
14749 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
14750 (*pushtf): Enable *roF alternative for x64 isa only.
14751 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
14752 mode attribute of integer alternatives to DImode for TARGET_64BIT.
14753 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
14754 (*movtf_internal): Merge from *movtf_internal_rex64 and
14755 *movtf_internal_sse. Use x64 and nox64 isa attributes.
14756 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
14757 nox64 isa attributes.
14758 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
14759 nox64 isa attributes.
14760 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
14761
14762 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14763
14764 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14765
14766 2013-03-17 Jakub Jelinek <jakub@redhat.com>
14767
14768 PR target/56640
14769 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
14770 class names. Remove trailing comma after "ALL_REGS".
14771
14772 2013-03-16 Jan Hubicka <jh@suse.cz>
14773
14774 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
14775 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
14776 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
14777 of cgraph_get_create_node.
14778 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
14779
14780 2013-03-16 Jason Merrill <jason@redhat.com>
14781
14782 PR debug/49090
14783 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
14784 with DW_AT_default_value.
14785
14786 2013-03-16 Jakub Jelinek <jakub@redhat.com>
14787
14788 * BASE-VER: Set to 4.9.0.
14789
14790 2013-03-14 Andi Kleen <ak@linux.intel.com>
14791
14792 PR target/56619
14793 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
14794 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
14795 Document _x* TSX intrinsics.
14796
14797 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
14798 David Holsgrove <david.holsgrove@xilinx.com>
14799
14800 * configure.ac: Add MicroBlaze TLS support detection.
14801 * configure: Regenerate.
14802 * config/microblaze/microblaze-protos.h
14803 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
14804 symbol_mentioned_p, label_mentioned_p): Add prototypes.
14805 * config/microblaze/microblaze.c (microblaze_address_type): Add
14806 ADDRESS_TLS and tls_reloc address types.
14807 (microblaze_address_info): Add tls_reloc.
14808 (TARGET_HAVE_TLS): Define.
14809 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
14810 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
14811 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
14812 load_tls_operand, microblaze_call_tls_get_addr,
14813 microblaze_legitimize_tls_address): New functions.
14814 (microblaze_classify_unspec): Handle UNSPEC_TLS.
14815 (get_base_reg): Use microblaze_tls_symbol_p.
14816 (microblaze_classify_address): Handle TLS.
14817 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
14818 label_mentioned_p and microblaze_tls_referenced_p.
14819 (microblaze_legitimize_address): Handle TLS.
14820 (microblaze_address_insns): Handle ADDRESS_TLS.
14821 (pic_address_needs_scratch): Handle TLS.
14822 (print_operand_address): Handle TLS.
14823 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
14824 (microblaze_expand_move): Handle TLS.
14825 (microblaze_legitimate_constant_p): Check
14826 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
14827 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
14828 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
14829 (PIC_OFFSET_TABLE_REGNUM): Set.
14830 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
14831 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
14832 (addsi3, movsi_internal2, movdf_internal): Update constraints
14833 * config/microblaze/predicates.md (arith_plus_operand): Define
14834 (move_operand): Redefine as move_src_operand,
14835 check microblaze_tls_referenced_p.
14836
14837 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14838
14839 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
14840 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
14841
14842 2013-03-14 Ian Bolton <ian.bolton@arm.com>
14843
14844 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
14845 CC mode for AND.
14846
14847 2013-03-14 Jakub Jelinek <jakub@redhat.com>
14848
14849 PR tree-optimization/53265
14850 * common.opt (Waggressive-loop-optimizations): New option.
14851 * tree-ssa-loop-niter.c: Include tree-pass.h.
14852 (do_warn_aggressive_loop_optimizations): New function.
14853 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
14854 if number_of_latch_executions returned constant.
14855 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
14856 early. If number_of_latch_executions returned constant, set
14857 nb_iterations_upper_bound back to it.
14858 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
14859 field.
14860 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
14861 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
14862
14863 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
14864 (MULTILIB_OSDIRNAMES): Set.
14865 * genmultilib: If defaultosdirname doesn't start with :: , set
14866 defaultosdirname2 instead, clear it and emit two . multilib_raw
14867 entries instead of just one.
14868
14869 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
14870
14871 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14872 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
14873 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
14874 (SUBTARGET_OVERRIDE_OPTIONS): New.
14875
14876 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
14877
14878 PR target/49880
14879 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
14880 (musermode): Convert to Var(TARGET_USERMODE).
14881 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
14882 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
14883 * config/sh/sh.c (sh_option_override): Use
14884 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
14885 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
14886 condition.
14887 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
14888 TARGET_SH4.
14889 (udivsi3_i4_single, divsi3_i4_single): Use
14890 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
14891
14892 2013-03-13 Dave Korn <dave.korn.cygwin@....>
14893
14894 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
14895 default setting.
14896
14897 2013-03-13 Richard Biener <rguenther@suse.de>
14898
14899 PR tree-optimization/56608
14900 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
14901 calls when vectorizing basic-blocks.
14902
14903 2013-03-13 Jakub Jelinek <jakub@redhat.com>
14904
14905 PR plugins/45078
14906 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
14907 tm_file.
14908
14909 2013-03-12 Jakub Jelinek <jakub@redhat.com>
14910
14911 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
14912
14913 2013-03-11 Jan Hubicka <jh@suse.cz>
14914
14915 PR lto/56557
14916 * lto-streamer-out.c (output_symbol_p): Skip references from
14917 constructors of external variables.
14918
14919 2013-03-11 Jan Hubicka <jh@suse.cz>
14920
14921 PR middle-end/56571
14922 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
14923 from pseudos.
14924 * emit-rtl.c (verify_rtx_sharing): Likewise.
14925 (copy_insn_1): Likewise.
14926 * rtl.c (copy_rtx): Likewise.
14927
14928 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
14929
14930 PR target/56591
14931 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
14932 output_operand_lossage message.
14933
14934 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
14935
14936 PR target/56470
14937 * arm.c (shift_op): Validate RTL pattern on the fly.
14938 (arm_print_operand, case 'S'): Don't use shift_operator to validate
14939 the RTL.
14940
14941 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14942
14943 PR target/56347
14944 * config/pa/pa.md (call_value): Check for calls to powf and direct to
14945 new call patterns that clobber %fr12.
14946 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
14947 split and postreload patterns.
14948 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
14949 registers %fr12 and %fr12R as call used.
14950
14951 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
14952
14953 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
14954 (canon_address, record_store, replace_read, check_mem_read_rtx,
14955 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
14956 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
14957 rest_of_handle_dse): Likewise.
14958
14959 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
14960
14961 PR middle-end/56524
14962 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
14963 Add base_optabs.
14964 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
14965 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
14966 (save_optabs_if_changed): Replace with...
14967 (init_tree_optimization_optabs): ...this.
14968 * optabs.c (save_optabs_if_changed): Rename to...
14969 (init_tree_optimization_optabs): ...this. Take the optimization node
14970 as argument. Do nothing if the base optabs are already correct.
14971 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
14972 to recompute optabs.
14973 * function.h (function): Remove optabs field.
14974 * function.c (invoke_set_current_function_hook): Call
14975 init_tree_optimization_optabs. Use the result to initialize
14976 this_fn_optabs.
14977
14978 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
14979
14980 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
14981 if GTMA_HAS_NO_INSTRUMENTATION.
14982 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
14983 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
14984 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
14985 * gimple-pretty-print.c (dump_gimple_transaction): Handle
14986 GTMA_HAS_NO_INSTRUMENTATION.
14987
14988 2013-03-08 Jakub Jelinek <jakub@redhat.com>
14989
14990 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
14991 libasan_preinit.o.
14992
14993 2013-03-08 Marek Polacek <polacek@redhat.com>
14994 Jakub Jelinek <jakub@redhat.com>
14995
14996 PR tree-optimization/56478
14997 * predict.c (is_comparison_with_loop_invariant_p): Change the
14998 type of loop_step to tree.
14999 (predict_loops): Adjust.
15000 (predict_iv_comparison): Perform the computations on double_ints.
15001
15002 2013-03-08 Richard Biener <rguenther@suse.de>
15003
15004 PR tree-optimization/56570
15005 * tree-cfg.c (verify_expr_location_1): Verify locations for
15006 DECL_DEBUG_EXPR.
15007 * tree-sra.c (create_access_replacement): Strip locations
15008 from DECL_DEBUG_EXPRs.
15009
15010 2013-03-08 Richard Biener <rguenther@suse.de>
15011
15012 * tree-inline.c (expand_call_inline): Do not associate
15013 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15014 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15015
15016 2013-03-08 Richard Biener <rguenther@suse.de>
15017
15018 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15019 or block changes with -Og. Fix for location / block encoding
15020 changes and PHI arguments with locations.
15021
15022 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15023
15024 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15025 for all counters.
15026 (struct output_info): Likewise.
15027 (register_overhead): Remove bad gcc_assert.
15028 (bitmap_find_bit): If there is only a single bitmap element, do not
15029 count a miss as a search.
15030 (print_statistics): Update for counter type changes.
15031 (dump_bitmap_statistics): Likewise. Print headers such that they
15032 are properly lined up with the printed counters.
15033
15034 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15035
15036 PR tree-optimization/56559
15037 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15038 check that it has only a single use.
15039
15040 2013-03-07 Richard Biener <rguenther@suse.de>
15041
15042 * doc/invoke.texi (fwhole-program): Discourage use in combination
15043 with -flto.
15044
15045 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15046
15047 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15048
15049 PR tree-optimization/56539
15050 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15051 instead of GSI_CONTINUE_LINKING as last argument to
15052 force_gimple_operand_gsi. Adjust function comment.
15053
15054 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15055 aarch64-cores.def.
15056
15057 PR middle-end/56548
15058 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15059 promoted mode, convert the result back to the original mode.
15060
15061 2013-03-06 Richard Biener <rguenther@suse.de>
15062
15063 PR middle-end/56294
15064 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15065 (insert_updated_phi_nodes_compare_uids): New function.
15066 (update_ssa): Sort symbols_to_rename after UID before
15067 traversing it to insert PHI nodes.
15068
15069 2013-03-06 Richard Biener <rguenther@suse.de>
15070
15071 PR middle-end/50494
15072 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15073 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15074
15075 Revert
15076 2013-02-13 Richard Biener <rguenther@suse.de>
15077
15078 PR lto/50494
15079 * varasm.c (output_constant_def_1): Get the decl representing
15080 the constant as argument.
15081 (output_constant_def): Wrap output_constant_def_1.
15082 (make_decl_rtl): Use output_constant_def_1 with the decl
15083 representing the constant.
15084 (build_constant_desc): Optionally re-use a decl already
15085 representing the constant.
15086 (tree_output_constant_def): Adjust.
15087
15088 2013-03-06 Joey Ye <joey.ye@arm.com>
15089
15090 PR lto/50293
15091 * gcc.c (convert_white_space): New function.
15092 (main): Handles white space in function name.
15093
15094 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
15095
15096 PR target/56529
15097 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
15098 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
15099 to SH_DIV_CALL_TABLE for TARGET_SH2.
15100 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
15101 list.
15102 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
15103 call-table options.
15104
15105 2013-03-05 Sterling Augustine <saugustine@google.com>
15106 Cary Coutant <ccoutant@google.com>
15107
15108 PR debug/55364
15109 * dwarf2out.c (resolve_addr): Don't call
15110 remove_loc_list_addr_table_entries a second time for the same
15111 expression.
15112
15113 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15114
15115 PR debug/56510
15116 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
15117 (avoid_complex_debug_insns): New function.
15118 (expand_debug_locations): Call it.
15119
15120 PR rtl-optimization/56484
15121 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
15122 lifetimes of hard registers on small register class machines.
15123
15124 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15125
15126 * config/microblaze/microblaze-protos.h: Rename
15127 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
15128 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
15129 fast_interrupt.
15130 (microblaze_fast_interrupt_function_p): New function.
15131 (microblaze_is_interrupt_handler): Rename to
15132 microblaze_is_interrupt_variant and add fast_interrupt check.
15133 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
15134 (save_restore_insns): Likewise.
15135 (compute_frame_size): Likewise.
15136 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
15137 (microblaze_globalize_label): Likewise.
15138 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
15139 * config/microblaze/microblaze.md: Use wrapper
15140 microblaze_is_interrupt_variant.
15141
15142 2013-03-05 Kai Tietz <ktietz@redhat.com>
15143
15144 * sdbout.c (sdbout_one_type): Switch to current function's section
15145 supporting cold/hot.
15146
15147 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
15148
15149 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
15150 -mxl-reorder.
15151
15152 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15153
15154 PR middle-end/56461
15155 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
15156 if VALGRIND_GET_VBITS is defined, temporarily make object
15157 memory all defined, and restore previous valgrind addressability
15158 and definability afterwards. Free this_object at the end.
15159
15160 PR middle-end/56461
15161 * lra.c (lra): Call lra_clear_live_ranges if live_p,
15162 right before calling lra_create_live_ranges, also call it
15163 when clearing live_p. Only call lra_clear_live_ranges
15164 at the end if live_p.
15165
15166 PR middle-end/56461
15167 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
15168
15169 2013-03-05 Richard Biener <rguenther@suse.de>
15170
15171 PR tree-optimization/56521
15172 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
15173 value-id.
15174
15175 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
15176
15177 PR c++/55135
15178 * except.h (remove_unreachable_eh_regions): New prototype.
15179 * except.c (remove_eh_handler_splicer): New function, split out
15180 of remove_eh_handler.
15181 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
15182 warning about running it on many EH regions one at a time.
15183 (remove_unreachable_eh_regions_worker): New function, walk the
15184 EH tree in depth-first order and remove non-marked regions.
15185 (remove_unreachable_eh_regions): New function.
15186 * tree-eh.c (mark_reachable_handlers): New function, split out
15187 from remove_unreachable_handlers.
15188 (remove_unreachable_handlers): Use mark_reachable_handlers and
15189 remove_unreachable_eh_regions.
15190 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
15191 and remove_unreachable_eh_regions.
15192
15193 2013-03-05 Richard Biener <rguenther@suse.de>
15194
15195 PR middle-end/56525
15196 * loop-init.c (fix_loop_structure): Remove loops in two stages,
15197 not freeing them until the end.
15198
15199 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15200
15201 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
15202
15203 2013-03-05 Richard Biener <rguenther@suse.de>
15204
15205 PR tree-optimization/56270
15206 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
15207 of loads after scheduling an SLP instance.
15208
15209 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15210
15211 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
15212 tic6x.exp.
15213 (check_gcc_parallelize): Run guality.exp as a separate job from
15214 vect.exp with unsorted.exp and $(dg_target_exps) separately from
15215 struct-layout-1.exp with stackalign.exp.
15216
15217 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
15218
15219 PR middle-end/56461
15220 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
15221 load_index sbitmap even if some bit in it isn't set.
15222
15223 PR middle-end/56461
15224 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
15225 (discover_iteration_bound_by_body_walk): Change queues to
15226 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
15227 spelling in comment. Call safe_push on queues[bound_index] directly.
15228 Release queues[queue_index] in every iteration unconditionally.
15229 Release bounds vector.
15230
15231 PR middle-end/56461
15232 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
15233 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
15234 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
15235 inner_phis vector.
15236
15237 2013-03-05 Richard Biener <rguenther@suse.de>
15238
15239 PR lto/56515
15240 * tree-inline.c (remap_blocks_to_null): New function.
15241 (expand_call_inline): When expanding a call stmt without
15242 an associated block inline remap all callee blocks to NULL.
15243
15244 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15245
15246 PR rtl-optimization/56494
15247 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
15248 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
15249 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
15250
15251 PR middle-end/56461
15252 * sel-sched-ir.c (free_sched_pools): Release
15253 succs_info_pool.stack[succs_info_pool.max_top] vectors too
15254 if succs_info_pool.max_top isn't -1.
15255
15256 PR bootstrap/56509
15257 * opts.c (opts_obstack, opts_concat): Moved to...
15258 * opts-common.c (opts_obstack, opts_concat): ... here.
15259
15260 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15261
15262 PR middle-end/56461
15263 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
15264
15265 2013-03-04 Martin Jambor <mjambor@suse.cz>
15266
15267 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
15268 all appropriate places.
15269
15270 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
15271
15272 PR tree-optimization/56424
15273 * ipa-split.c (split_function): Do not set the RSO flag if result is
15274 not by reference and its type is a register type.
15275
15276 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
15277
15278 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
15279 (microblaze_legitimate_pic_operand): Likewise
15280 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
15281 new function microblaze_legitimate_pic_operand
15282 * config/microblaze/microblaze-protos.h
15283 (microblaze_legitimate_pic_operand): Declare.
15284
15285 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15286
15287 * config/microblaze/predicates.md (call_insn_simple_operand):
15288 New predicate for supported rtx code types.
15289 * config/microblaze/microblaze.md (call_internal1): Use
15290 call_insn_simple_operand predicate.
15291
15292 2013-03-04 Jakub Jelinek <jakub@redhat.com>
15293
15294 PR middle-end/56461
15295 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
15296 partitions.ordered_remove.
15297
15298 PR middle-end/56461
15299 * tree-vect-stmts.c (vectorizable_conversion): Don't call
15300 vec_oprnds0.create (1) for modifier == NONE.
15301
15302 PR middle-end/56461
15303 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
15304 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
15305 vec_oprnds1 right before pushing anything to it for
15306 scalar_shift_arg.
15307
15308 PR middle-end/56461
15309 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
15310 set nbbs to 0 instead of having separate code path.
15311 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
15312 instead of false as last argument if returning NULL.
15313
15314 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
15315
15316 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
15317 the attribute is now called "target" instead of "option".
15318 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
15319 * doc/tm.texi.in (Target Attributes): Likewise document the correct
15320 attribute/pragma name for TARGET_OPTION_VALID_P and
15321 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
15322 * doc/tm.texi: Regenerated.
15323
15324 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
15325
15326 * config/microblaze/microblaze.c:
15327 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
15328 * config/microblaze/microblaze.h: Add -mxl-reorder to
15329 DRIVER_SELF_SPECS.
15330 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
15331 instructions emitted if TARGET_REORDER.
15332 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
15333 or 0 for -m/-mno case, but initialises as 2 to detect default use case
15334 separately.
15335
15336 2013-03-01 Xinliang David Li <davidxl@google.com>
15337
15338 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
15339 walk length.
15340
15341 2013-03-01 Jakub Jelinek <jakub@redhat.com>
15342
15343 PR middle-end/56461
15344 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
15345 vector even when returning true. Fix up function comment formatting.
15346
15347 PR middle-end/56461
15348 * ira-build.c (ira_loop_nodes_count): New variable.
15349 (create_loop_tree_nodes): Initialize it.
15350 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
15351
15352 PR middle-end/56461
15353 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
15354 method on dr_chain and result_chain.
15355 * tree-vect-stmts.c (vectorizable_store): Only call
15356 result_chain.create if j == 0.
15357
15358 PR middle-end/56461
15359 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
15360 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
15361 before overwriting it.
15362
15363 2013-03-01 Tobias Burnus <burnus@net-b.de>
15364
15365 * doc/extended.texi (C Extensions): Change order in @menu
15366 to match @node.
15367 (Other MIPS Built-in Functions): Move last MIPS entry before
15368 "picoChip Built-in Functions".
15369 (SH Built-in Functions): Move after RX Built-in Functions.
15370 * doc/gcc.texi (Introduction): Change order in @menu
15371 to match @node.
15372 * doc/md.texi (Constraints): Ditto.
15373 * gty.texi (Type Information): Ditto.
15374 (User-provided marking routines for template types): Make
15375 subsection.
15376 * doc/invoke.texi (AArch64 Options): Move before
15377 "Adapteva Epiphany Options".
15378
15379 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
15380 Jakub Jelinek <jakub@redhat.com>
15381
15382 PR sanitizer/56454
15383 * asan.c (gate_asan): Lookup no_sanitize_address instead of
15384 no_address_safety_analysis attribute.
15385 * doc/extend.texi (no_address_safety_attribute): Rename to
15386 no_sanitize_address attribute, mention no_address_safety_analysis
15387 attribute as deprecated alias.
15388
15389 2013-02-28 Jakub Jelinek <jakub@redhat.com>
15390
15391 PR middle-end/56461
15392 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
15393 type to vec<vec<tree> > *.
15394 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
15395 to be vec<tree> instead of vec<tree> *, set vec_defs
15396 to vNULL and call vec_defs.create (number_of_vects), adjust other
15397 uses of vec_defs.
15398 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
15399 vectorizable_condition): Adjust vect_get_slp_defs callers.
15400
15401 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15402
15403 * config/aarch64/aarch64.c
15404 (aarch64_float_const_representable): Remove unused variable.
15405
15406 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15407
15408 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
15409
15410 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15411
15412 * config/aarch64/aarch64-builtins.c
15413 (aarch64_init_simd_builtins): Make static.
15414
15415 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
15416
15417 * config/aarch64/aarch64.c
15418 (aarch64_simd_make_constant): Make static.
15419
15420 2013-02-28 Martin Jambor <mjambor@suse.cz>
15421
15422 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
15423 with no initialization to the RHS of debug statements.
15424
15425 2013-02-28 Martin Jambor <mjambor@suse.cz>
15426
15427 PR tree-optimization/56294
15428 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
15429 Adjust dumping.
15430 (get_access_replacement): Do not call create_access_replacement.
15431 Assert a replacement exists.
15432 (get_repl_default_def_ssa_name): Create the replacement declaration
15433 itself.
15434
15435 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15436
15437 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
15438 final_end_function.
15439
15440 2013-02-28 Marek Polacek <polacek@redhat.com>
15441
15442 PR rtl-optimization/56466
15443 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
15444 if we're changing a loop.
15445 (peel_loops_completely): Likewise.
15446
15447 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
15448
15449 PR c++/55813
15450 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
15451
15452 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15453
15454 PR target/56445
15455 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
15456 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
15457 INTX_FTYPE_FX, FX_FTYPE_INTX.
15458 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
15459
15460 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15461
15462 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
15463 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
15464 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
15465 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
15466 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
15467 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
15468 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
15469 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
15470 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
15471 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
15472 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
15473 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
15474 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
15475 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
15476 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
15477 (avrxmega6): Increase max flash segments from 5 to 6.
15478 * config/avr/t-multilib: Regenerate.
15479 * config/avr/avr-tables.opt: Regenerate.
15480 * doc/avr-mmcu.texi: Regenerate.
15481
15482 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
15483
15484 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
15485 (avr_device_to_arch): Rename to avr_device_to_ld.
15486 (avr_device_to_as): New prototype.
15487 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
15488 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
15489 * config/avr/driver-avr.c (avr_device_to_as): New.
15490 (avr_device_to_arch): Rename to avr_device_to_ld.
15491
15492 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15493
15494 PR middle-end/56461
15495 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
15496 method on dr_chain and result_chain.
15497
15498 PR middle-end/56461
15499 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
15500 pointer_set_destroy on not_executed_last_iteration.
15501
15502 PR middle-end/56461
15503 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
15504
15505 PR middle-end/56461
15506 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
15507 FOR_EACH_DEFINED_FUNCTION when freeing state.
15508
15509 PR middle-end/56461
15510 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
15511 pool_free.
15512 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
15513 overwriting it.
15514
15515 PR middle-end/56461
15516 * ipa-cp.c (decide_whether_version_node): Call vec_free on
15517 known_aggs[i].items and release known_aggs vector.
15518
15519 PR middle-end/56461
15520 * ipa-reference.c (propagate): Free node_info even for alias nodes.
15521
15522 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
15523
15524 * config/microblaze/microblaze.c (microblaze_emit_compare):
15525 Use xor for EQ/NE comparisions.
15526 * config/microblaze/microblaze.md (cstoresf4): Add constraints
15527 (cbranchsf4): Adjust operator to comparison_operator.
15528
15529 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15530
15531 PR middle-end/56461
15532 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
15533 vector.
15534 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
15535 vec_safe_push, always update *slot.
15536 (redirect_edge_var_map_clear): Use vec_free.
15537 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
15538 (free_var_map_entry): Use vec_free.
15539 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
15540 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
15541
15542 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
15543
15544 PR middle-end/45472
15545 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
15546 when the may_trap_p bit of the exprs being merged differs.
15547 Reorder tests for speculativeness in the logical and operator.
15548
15549 2013-02-27 Jakub Jelinek <jakub@redhat.com>
15550
15551 * incpath.c (add_standard_paths): Use reconcat instead of concat
15552 where appropriate and avoid leaking memory.
15553
15554 * opts.h: Include obstack.h.
15555 (opts_concat): New prototype.
15556 (opts_obstack): New declaration.
15557 * opts.c (opts_concat): New function.
15558 (opts_obstack): New variable.
15559 (init_options_struct): Call gcc_init_obstack on opts_obstack.
15560 (finish_options): Use opts_concat instead of concat
15561 and XOBNEWVEC instead of XNEWVEC.
15562 * opts-common.c (generate_canonical_option, decode_cmdline_option,
15563 generate_option): Likewise.
15564 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
15565 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
15566
15567 PR target/56455
15568 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
15569 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
15570
15571 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15572
15573 PR middle-end/56461
15574 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
15575
15576 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15577
15578 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
15579 (arm_block_move_unaligned_straight): Likewise.
15580 (arm_adjust_block_mem): Likewise.
15581
15582 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
15583
15584 PR target/48901
15585 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
15586 temp, cond and label.
15587 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
15588
15589 PR target/52500
15590 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
15591 * config/c6x/c6x.h (dbx_register_map): Update declaration.
15592
15593 PR target/52501
15594 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
15595 of prologue/epilogue functions.
15596
15597 PR target/52550
15598 * config/tilegx/tilegx.c (tilegx_expand_prologue):
15599 Remove unused variable cfa_offset.
15600 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
15601
15602 PR target/54639
15603 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
15604 type promotion to unsigned.
15605
15606 PR target/54640
15607 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
15608 for HOST_WIDE_INT of 32 bit / same size as int.
15609 (arm_block_move_unaligned_straight): Likewise.
15610 (arm_adjust_block_mem): Likewise.
15611
15612 PR target/54662
15613 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
15614 ALL_CFLAGS.
15615
15616 2013-02-26 Marek Polacek <polacek@redhat.com>
15617
15618 PR tree-optimization/56426
15619 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
15620
15621 2013-02-26 Richard Biener <rguenther@suse.de>
15622
15623 PR target/56444
15624 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
15625 unused variable loops.
15626
15627 2013-02-26 Jakub Jelinek <jakub@redhat.com>
15628
15629 PR tree-optimization/56448
15630 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
15631 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
15632 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
15633 later operands of the references, or even first operand for
15634 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
15635
15636 PR tree-optimization/56443
15637 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
15638 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
15639 to type_for_mode langhook.
15640
15641 2013-02-25 Matt Turner <mattst88@gmail.com>
15642
15643 * doc/invoke.texi: Document r4700.
15644
15645 2013-02-25 Richard Biener <rguenther@suse.de>
15646
15647 PR tree-optimization/56175
15648 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
15649 split out from ...
15650 (simplify_bitwise_binary): ... here. Also guard the conversion
15651 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
15652
15653 2013-02-25 Catherine Moore <clm@codesourcery.com>
15654
15655 Revert:
15656 2013-02-24 Catherine Moore <clm@codesourcery.com>
15657 Maciej W. Rozycki <macro@codesourcery.com>
15658 Tom de Vries <tom@codesourcery.com>
15659 Nathan Sidwell <nathan@codesourcery.com>
15660 Iain Sandoe <iain@codesourcery.com>
15661 Nathan Froyd <froydnj@codesourcery.com>
15662 Chao-ying Fu <fu@mips.com>
15663
15664 * doc/extend.texi: (micromips, nomicromips, nocompression):
15665 Document new function attributes.
15666 * doc/invoke.texi (minterlink-compressed, mmicromips,
15667 m14k, m14ke, m14kec): Document new options.
15668 (minterlink-mips16): Update documentation.
15669 * doc/md.texi (ZC, ZD): Document new constraints.
15670 * configure.ac (gcc_cv_as_micromips): Check if linker
15671 supports the .set micromips directive.
15672 * configure: Regenerate.
15673 * config.in: Regenerate.
15674 * config/mips/mips-tables.opt: Regenerate.
15675 * config/mips/micromips.md: New file.
15676 * constraints.md (ZC, AD): New constraints.
15677 * config/mips/predicates.md (movep_src_register): New predicate.
15678 (movep_src_operand): New predicate.
15679 (non_volatile_mem_operand): New predicate.
15680 * config/mips/mips.md (multimem): New type.
15681 (length): Differentiate between 17-bit and 18-bit branch offsets.
15682 (MOVEP1, MOVEP2): New mode iterator.
15683 (mov_<load>l): Use ZC constraint.
15684 (mov_<load>r): Likewise.
15685 (mov_<store>l): Likewise.
15686 (mov_<store>r): Likewise.
15687 (*branch_equality<mode>_inverted): Add microMIPS support.
15688 (*branch_equality<mode>): Likewise.
15689 (*jump_absolute): Likewise.
15690 (indirect_jump_<mode>): Likewise.
15691 (tablejump_<mode>): Likewise.
15692 (<optab>_internal): Likewise.
15693 (sibcall_internal): Likewise.
15694 (sibcall_value_internal): Likewise.
15695 (prefetch): Use constraint ZD.
15696 * config/mips/mips.opt (minterlink-compressed): New option.
15697 (minterlink-mips16): Now an alias for minterlink-compressed.
15698 (mmicromips): New option.
15699 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15700 (compare_and_swap_12): Likewise.
15701 (sync_add<mode>): Likewise.
15702 (sync_<optab>_12): Likewise.
15703 (sync_old_<optab>_12): Likewise.
15704 (sync_new_<optab>_12): Likewise.
15705 (sync_nand_12): Likewise.
15706 (sync_old_nand_12): Likewise.
15707 (sync_new_nand_12): Likewise.
15708 (sync_sub<mode>): Likewise.
15709 (sync_old_add<mode>): Likewise.
15710 (sync_old_sub<mode>): Likewise.
15711 (sync_new_add<mode>): Likewise.
15712 (sync_new_sub<mode>): Likewise.
15713 (sync_<optab><mode>): Likewise.
15714 (sync_old_<optab><mode>): Likewise.
15715 (sync_new_<optab><mode>): Likewise.
15716 (sync_nand<mode>): Likewise.
15717 (sync_old_nand<mode>): Likewise.
15718 (sync_new_nand<mode>): Likewise.
15719 (sync_lock_test_and_set<mode>): Likewise.
15720 (test_and_set_12): Likewise.
15721 (atomic_compare_and_swap<mode>): Likewise.
15722 (atomic_exchange<mode>_llsc): Likewise.
15723 (atomic_fetch_add<mode>_llsc): Likewise.
15724 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15725 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15726 (umips_save_restore_pattern_p): Likewise.
15727 (umips_load_store_pair_p): Likewise.
15728 (umips_output_load_store_pair): Likewise.
15729 (umips_movep_target_p): Likewise.
15730 (umips_12bit_offset_address_p): Likewise.
15731 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15732 (mips_base_mips16): Rename this...
15733 (mips_base_compression_flags): ...to this. Update all uses.
15734 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15735 (mips_mips16_decl_p): Delete.
15736 (mips_nomips16_decl_p): Delete.
15737 (mips_get_compress_on_flags): New function.
15738 (mips_get_compress_off_flags): New function.
15739 (mips_get_compress_mode): New function.
15740 (mips_get_compress_on_name): New function.
15741 (mips_get_compress_off_name): New function.
15742 (mips_insert_attributes): Support multiple compression types.
15743 (mips_merge_decl_attributes): Likewise.
15744 (umips_12bit_offset_address_p): New function.
15745 (mips_start_function_definition): Emit .set micromips directive.
15746 (mips_call_may_need_jalx_p): New function.
15747 (mips_function_ok_for_sibcall): Add microMIPS support.
15748 (mips_print_operand_punctuation): Support short delay slots and
15749 compact jumps.
15750 (umips_swm_mask, umips_swm_encoding): New.
15751 (umips_build_save_restore): New function.
15752 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15753 (was_mips16_p): Remove.
15754 (old_compression_mode): New.
15755 (mips_set_compression_mode): New function.
15756 (mips_set_current_function): Add microMIPS support.
15757 (mips_option_override): Likewise.
15758 (umips_save_restore_pattern_p): New function.
15759 (umips_output_save_restore): New function.
15760 (umips_load_store_pair_p_1): New function.
15761 (umips_load_store_pair_p): New function.
15762 (umips_output_load_store_pair_1): New function.
15763 (umips_output_load_store_pair): New function.
15764 (umips_movep_target_p) New function.
15765 (mips_prepare_pch_save): Add microMIPS support.
15766 * config/mips/mips.h (TARGET_COMPRESSION): New.
15767 (TARGET_CPU_CPP_BUILTINS): Update macro
15768 to use new compression flags and to support microMIPS.
15769 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15770 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15771 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15772 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15773 (ASM_SPEC): Support mmicromips and mno-micromips.
15774 (M16STORE_REG_P): New macro.
15775 (MIPS_CALL): Support TARGET_MICROMIPS.
15776 (MICROMIPS_J): New macro.
15777 (mips_base_mips16): Rename this...
15778 (mips_base_compression_flags): ...to this.
15779 (UMIPS_12BIT_OFFSET_P): New macro.
15780 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15781 (MULTILIB_DIRNAMES): Likewise.
15782
15783 2013-02-25 Tom de Vries <tom@codesourcery.com>
15784
15785 PR rtl-optimization/56131
15786 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
15787 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
15788 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
15789
15790 2013-02-25 Tobias Burnus <burnus@net-b.de>
15791
15792 * doc/invoke.texi (-fsanitize=): Move from optimization
15793 to debugging options.
15794
15795 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15796
15797 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
15798
15799 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
15800 Alexander Monakov <amonakov@ispras.ru>
15801
15802 PR middle-end/56077
15803 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
15804 flush pending lists also on non-jumps. Adjust comment.
15805
15806 2013-02-24 Catherine Moore <clm@codesourcery.com>
15807 Maciej W. Rozycki <macro@codesourcery.com>
15808 Tom de Vries <tom@codesourcery.com>
15809 Nathan Sidwell <nathan@codesourcery.com>
15810 Iain Sandoe <iain@codesourcery.com>
15811 Nathan Froyd <froydnj@codesourcery.com>
15812 Chao-ying Fu <fu@mips.com>
15813
15814 * doc/extend.texi: (micromips, nomicromips, nocompression):
15815 Document new function attributes.
15816 * doc/invoke.texi (minterlink-compressed, mmicromips,
15817 m14k, m14ke, m14kec): Document new options.
15818 (minterlink-mips16): Update documentation.
15819 * doc/md.texi (ZC, ZD): Document new constraints.
15820 * configure.ac (gcc_cv_as_micromips): Check if linker
15821 supports the .set micromips directive.
15822 * configure: Regenerate.
15823 * config.in: Regenerate.
15824 * config/mips/mips-tables.opt: Regenerate.
15825 * config/mips/micromips.md: New file.
15826 * constraints.md (ZC, AD): New constraints.
15827 * config/mips/predicates.md (movep_src_register): New predicate.
15828 (movep_src_operand): New predicate.
15829 (non_volatile_mem_operand): New predicate.
15830 * config/mips/mips.md (multimem): New type.
15831 (length): Differentiate between 17-bit and 18-bit branch offsets.
15832 (MOVEP1, MOVEP2): New mode iterator.
15833 (mov_<load>l): Use ZC constraint.
15834 (mov_<load>r): Likewise.
15835 (mov_<store>l): Likewise.
15836 (mov_<store>r): Likewise.
15837 (*branch_equality<mode>_inverted): Add microMIPS support.
15838 (*branch_equality<mode>): Likewise.
15839 (*jump_absolute): Likewise.
15840 (indirect_jump_<mode>): Likewise.
15841 (tablejump_<mode>): Likewise.
15842 (<optab>_internal): Likewise.
15843 (sibcall_internal): Likewise.
15844 (sibcall_value_internal): Likewise.
15845 (prefetch): Use constraint ZD.
15846 * config/mips/mips.opt (minterlink-compressed): New option.
15847 (minterlink-mips16): Now an alias for minterlink-compressed.
15848 (mmicromips): New option.
15849 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15850 (compare_and_swap_12): Likewise.
15851 (sync_add<mode>): Likewise.
15852 (sync_<optab>_12): Likewise.
15853 (sync_old_<optab>_12): Likewise.
15854 (sync_new_<optab>_12): Likewise.
15855 (sync_nand_12): Likewise.
15856 (sync_old_nand_12): Likewise.
15857 (sync_new_nand_12): Likewise.
15858 (sync_sub<mode>): Likewise.
15859 (sync_old_add<mode>): Likewise.
15860 (sync_old_sub<mode>): Likewise.
15861 (sync_new_add<mode>): Likewise.
15862 (sync_new_sub<mode>): Likewise.
15863 (sync_<optab><mode>): Likewise.
15864 (sync_old_<optab><mode>): Likewise.
15865 (sync_new_<optab><mode>): Likewise.
15866 (sync_nand<mode>): Likewise.
15867 (sync_old_nand<mode>): Likewise.
15868 (sync_new_nand<mode>): Likewise.
15869 (sync_lock_test_and_set<mode>): Likewise.
15870 (test_and_set_12): Likewise.
15871 (atomic_compare_and_swap<mode>): Likewise.
15872 (atomic_exchange<mode>_llsc): Likewise.
15873 (atomic_fetch_add<mode>_llsc): Likewise.
15874 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15875 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15876 (umips_save_restore_pattern_p): Likewise.
15877 (umips_load_store_pair_p): Likewise.
15878 (umips_output_load_store_pair): Likewise.
15879 (umips_movep_target_p): Likewise.
15880 (umips_12bit_offset_address_p): Likewise.
15881 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15882 (mips_base_mips16): Rename this...
15883 (mips_base_compression_flags): ...to this. Update all uses.
15884 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15885 (mips_mips16_decl_p): Delete.
15886 (mips_nomips16_decl_p): Delete.
15887 (mips_get_compress_on_flags): New function.
15888 (mips_get_compress_off_flags): New function.
15889 (mips_get_compress_mode): New function.
15890 (mips_get_compress_on_name): New function.
15891 (mips_get_compress_off_name): New function.
15892 (mips_insert_attributes): Support multiple compression types.
15893 (mips_merge_decl_attributes): Likewise.
15894 (umips_12bit_offset_address_p): New function.
15895 (mips_start_function_definition): Emit .set micromips directive.
15896 (mips_call_may_need_jalx_p): New function.
15897 (mips_function_ok_for_sibcall): Add microMIPS support.
15898 (mips_print_operand_punctuation): Support short delay slots and
15899 compact jumps.
15900 (umips_swm_mask, umips_swm_encoding): New.
15901 (umips_build_save_restore): New function.
15902 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15903 (was_mips16_p): Remove.
15904 (old_compression_mode): New.
15905 (mips_set_compression_mode): New function.
15906 (mips_set_current_function): Add microMIPS support.
15907 (mips_option_override): Likewise.
15908 (umips_save_restore_pattern_p): New function.
15909 (umips_output_save_restore): New function.
15910 (umips_load_store_pair_p_1): New function.
15911 (umips_load_store_pair_p): New function.
15912 (umips_output_load_store_pair_1): New function.
15913 (umips_output_load_store_pair): New function.
15914 (umips_movep_target_p) New function.
15915 (mips_prepare_pch_save): Add microMIPS support.
15916 * config/mips/mips.h (TARGET_COMPRESSION): New.
15917 (TARGET_CPU_CPP_BUILTINS): Update macro
15918 to use new compression flags and to support microMIPS.
15919 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15920 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15921 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15922 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15923 (ASM_SPEC): Support mmicromips and mno-micromips.
15924 (M16STORE_REG_P): New macro.
15925 (MIPS_CALL): Support TARGET_MICROMIPS.
15926 (MICROMIPS_J): New macro.
15927 (mips_base_mips16): Rename this...
15928 (mips_base_compression_flags): ...to this.
15929 (UMIPS_12BIT_OFFSET_P): New macro.
15930 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15931 (MULTILIB_DIRNAMES): Likewise.
15932
15933 2013-02-24 Jakub Jelinek <jakub@redhat.com>
15934
15935 PR target/52555
15936 * target-globals.c (save_target_globals): For init_reg_sets and
15937 target_reinit remporarily set this_fn_optabs to this_target_optabs.
15938
15939 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
15940
15941 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
15942 * config/aarch64/t-aarch64
15943 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
15944
15945 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
15946
15947 PR inline-asm/56148
15948 * lra-constraints.c (process_alt_operands): Reload operand
15949 conflicting with earlier clobber only if no more other conflicting
15950 operands.
15951
15952 2013-02-22 Jakub Jelinek <jakub@redhat.com>
15953
15954 PR sanitizer/56393
15955 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
15956 if not linking a shared library.
15957
15958 2013-02-22 Seth LaForge <sethml@google.com>
15959
15960 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
15961
15962 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
15963
15964 * config/arm/arm.md (split for extendsidi): Update condition.
15965 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
15966 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
15967 (qhs_zextenddi_cstr): Likewise.
15968
15969 2013-02-21 Jakub Jelinek <jakub@redhat.com>
15970
15971 PR middle-end/56420
15972 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
15973 avoid signed wrapping.
15974 (expand_mult): Handle properly multiplication by
15975 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
15976 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
15977 in the compiler if coeff is HOST_WIDE_INT_MIN.
15978 (expand_divmod): Don't make ext_op1 static, change it's type to
15979 uhwi. Avoid undefined behavior in -INTVAL (op1).
15980
15981 PR rtl-optimization/50339
15982 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
15983 field.
15984 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
15985 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
15986 into splitting_ashiftrt field.
15987 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
15988 ASHIFTRT.
15989 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
15990 choices.
15991
15992 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
15993
15994 PR middle-end/56108
15995 * trans-mem.c (execute_tm_mark): Do not expand transactions that
15996 are sure to go irrevocable.
15997
15998 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
15999
16000 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16001 scalars are valid operands.
16002
16003 2013-02-21 Martin Jambor <mjambor@suse.cz>
16004
16005 PR tree-optimization/56310
16006 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16007 only matching indices and non-negative final offsets.
16008 (intersect_aggregates_with_edge): Pass src_idx to
16009 agg_replacements_to_vector. Pass src_idx insstead of index to
16010 intersect_with_agg_replacements.
16011
16012 2013-02-21 Martin Jambor <mjambor@suse.cz>
16013
16014 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16015 instead of hard-wired defaults.
16016
16017 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16018
16019 * doc/invoke.texi (MIPS Options): Update documentation of the
16020 floating-point multiply-accumulate instruction restrictions.
16021
16022 2013-02-21 Kostya Serebryany <kcc@google.com>
16023
16024 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16025 asan_shadow_offset on x86_64 linux.
16026
16027 2013-02-21 Richard Biener <rguenther@suse.de>
16028
16029 PR tree-optimization/56415
16030 Revert
16031 2013-02-11 Richard Biener <rguenther@suse.de>
16032
16033 PR tree-optimization/56273
16034 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16035 first VRP run.
16036
16037 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16038
16039 PR bootstrap/56258
16040 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16041 instead of @itemx.
16042
16043 PR inline-asm/56405
16044 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16045 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16046
16047 2013-02-20 Jan Hubicka <jh@suse.cz>
16048
16049 PR tree-optimization/56265
16050 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16051 when target is referenced for first time.
16052
16053 2013-02-20 Richard Biener <rguenther@suse.de>
16054
16055 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16056 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16057 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16058 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16059 not return anything.
16060 (rename_ssa_copies): Do not remove unused locals.
16061 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16062 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16063 * passes.c (execute_function_todo): Do not schedule unused locals
16064 removal if cleanup_tree_cfg did something.
16065 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16066 about the number of removed locals.
16067
16068 2013-02-20 Richard Biener <rguenther@suse.de>
16069
16070 PR tree-optimization/56398
16071 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16072
16073 2013-02-20 Martin Jambor <mjambor@suse.cz>
16074
16075 PR tree-optimization/55334
16076 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16077 restricted pointers to arrays.
16078
16079 2013-02-20 Richard Biener <rguenther@suse.de>
16080 Jakub Jelinek <jakub@redhat.com>
16081
16082 PR tree-optimization/56396
16083 * tree-ssa-ccp.c (n_const_val): New static variable.
16084 (get_value): Return NULL for SSA names we don't have a lattice
16085 entry for.
16086 (ccp_initialize): Initialize n_const_val.
16087 * tree-ssa-copy.c (n_copy_of): New static variable.
16088 (init_copy_prop): Initialize n_copy_of.
16089 (get_value): Return NULL_TREE for SSA names we don't have a
16090 lattice entry for.
16091
16092 2013-02-20 Martin Jambor <mjambor@suse.cz>
16093
16094 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
16095
16096 2013-02-20 Richard Biener <rguenther@suse.de>
16097
16098 * genpreds.c (write_lookup_constraint): Do not compare first
16099 letter of the constraint again.
16100
16101 2013-02-20 Richard Biener <rguenther@suse.de>
16102
16103 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
16104 and ceil_log2.
16105 (get_use_iv_cost): Terminate hashtable walk when coming across
16106 an empty entry.
16107
16108 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
16109
16110 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
16111 reassociation for avx2 targets.
16112
16113 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16114
16115 * config/microblaze/microblaze.c: microblaze_has_clz = 0
16116 Add version check for v8.10.a to enable microblaze_has_clz
16117 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
16118 version and TARGET_PATTERN_COMPARE check
16119 * config/microblaze/microblaze.md: New clzsi2 instruction
16120
16121 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16122
16123 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
16124 function before branching.
16125
16126 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16127
16128 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
16129 DUMP_INSN_RTX_UID.
16130 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
16131
16132 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
16133
16134 PR middle-end/55889
16135 * sel-sched.c: Include ira.h.
16136 (implicit_clobber_conflict_p): New function.
16137 (moveup_expr): Use it.
16138 * Makefile.in (sel-sched.o): Depend on ira.h.
16139
16140 2013-02-19 Richard Biener <rguenther@suse.de>
16141
16142 PR tree-optimization/56384
16143 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
16144 (vn_hash_type): Split out from ...
16145 (vn_hash_constant_with_type): ... here.
16146 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
16147 (vn_phi_eq): Compare types from vn_phi_s structure.
16148 (vn_phi_lookup): Populate vn_phi_s type.
16149 (vn_phi_insert): Likewise.
16150
16151 2013-02-19 Jakub Jelinek <jakub@redhat.com>
16152
16153 PR tree-optimization/56350
16154 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
16155 if haven't found reduction or nested cycle operand, rather than
16156 asserting we must find it.
16157
16158 PR tree-optimization/56381
16159 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
16160 to fold_build3.
16161
16162 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
16163 Jakub Jelinek <jakub@redhat.com>
16164
16165 PR target/52555
16166 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
16167 (swap_optab_enable): Same.
16168 (init_all_optabs): Use argument instead of global.
16169 * tree.h (struct tree_optimization_option): New field target_optabs.
16170 * expr.h (init_all_optabs): Add argument to prototype.
16171 (TREE_OPTIMIZATION_OPTABS): New.
16172 (save_optabs_if_changed): Protoize.
16173 * optabs.h: Declare this_fn_optabs.
16174 * optabs.c (save_optabs_if_changed): New.
16175 Declare this_fn_optabs.
16176 (init_optabs): Add argument to init_all_optabs() call.
16177 * function.c (invoke_set_current_function_hook): Handle per
16178 function optabs.
16179 * function.h (struct function): New field optabs.
16180 * config/mips/mips.c (mips_set_mips16_mode): Handle when
16181 optimization_current_node has changed.
16182 * target-globals.h (save_target_globals_default_opts): Protoize.
16183 * target-globals.c (save_target_globals_default_opts): New.
16184
16185 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16186
16187 PR target/56347
16188 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
16189 registers %fr12 and %fr12R as call used.
16190
16191 PR target/56214
16192 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
16193 and HImode, require all displacements to be an integer multiple of
16194 their mode size.
16195 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
16196 only allow QImode and HImode when reload is in progress and strict is
16197 true. Likewise for symbolic addresses. Use base14_operand to check
16198 displacements in REG+BASE addresses.
16199
16200 2013-02-18 Richard Biener <rguenther@suse.de>
16201
16202 PR tree-optimization/56366
16203 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16204 handle sign-conversion of outer-loop initial induction value.
16205
16206 2013-02-18 Richard Biener <rguenther@suse.de>
16207
16208 PR middle-end/56349
16209 * cfghooks.c (merge_blocks): If we merge a latch into another
16210 block adjust references to it.
16211 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
16212 (verify_loop_structure): Verify that a recorded latch is in fact
16213 a latch.
16214
16215 2013-02-18 Richard Biener <rguenther@suse.de>
16216
16217 PR tree-optimization/56321
16218 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
16219 order SSA name release and virtual operand unlinking.
16220
16221 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16222
16223 * config/microblaze/microblaze.md (save_stack_block): Define.
16224 (restore_stack_block): Likewise.
16225
16226 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16227
16228 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
16229 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
16230 * config/microblaze/microblaze.c (microblaze_option_override):
16231 Bail out early for PIC modes when target does not support PIC.
16232
16233 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16234
16235 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
16236 Replace with a microblaze version.
16237 (microblaze_trampoline_init): Adapt for microblaze.
16238 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
16239 microblaze.
16240
16241 2013-02-16 Jakub Jelinek <jakub@redhat.com>
16242 Dodji Seketeli <dodji@redhat.com>
16243
16244 PR asan/56330
16245 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
16246 (instrument_mem_region_access): Do not forget to always put
16247 instrumentation of the of 'base' and 'base + len' in a "if (len !=
16248 0) statement, even for cases where either 'base' or 'base + len'
16249 are not instrumented -- because they have been previously
16250 instrumented. Simplify the logic by putting all the statements
16251 instrument 'base + len' inside a sequence, and then insert that
16252 sequence right before the current insertion point. Then, to
16253 instrument 'base + len', just get an iterator on that statement.
16254 And do not forget to update the pointer to iterator the function
16255 received as argument.
16256
16257 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
16258
16259 PR rtl-optimization/56348
16260 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
16261
16262 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
16263
16264 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
16265 (clean_graph_dump_file): Pass base to start_graph_dump.
16266
16267 2013-02-14 Richard Henderson <rth@redhat.com>
16268
16269 PR target/55941
16270 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
16271
16272 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
16273
16274 * collect2-aix.h: Define F_LOADONLY.
16275
16276 2013-02-14 Richard Biener <rguenther@suse.de>
16277
16278 PR lto/50494
16279 * varasm.c (output_constant_def_1): Get the decl representing
16280 the constant as argument.
16281 (output_constant_def): Wrap output_constant_def_1.
16282 (make_decl_rtl): Use output_constant_def_1 with the decl
16283 representing the constant.
16284 (build_constant_desc): Optionally re-use a decl already
16285 representing the constant.
16286 (tree_output_constant_def): Adjust.
16287
16288 2013-02-14 Dodji Seketeli <dodji@redhat.com>
16289
16290 Fix an asan crash
16291 * asan.c (instrument_builtin_call): Really put the length of the
16292 second source argument into src1_len.
16293
16294 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16295
16296 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
16297 argument. If it is false, don't create edge from then_bb to
16298 fallthru_bb.
16299 (insert_if_then_before_iter): Pass true to it.
16300 (build_check_stmt): Pass false to it.
16301 (transform_statements): Flush hash table only on extended basic
16302 block boundaries, rather than at the beginning of every bb.
16303 Don't flush hash table on nonfreeing_call_p calls.
16304 * tree-flow.h (nonfreeing_call_p): New prototype.
16305 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
16306
16307 2013-02-13 David S. Miller <davem@davemloft.net>
16308
16309 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
16310
16311 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
16312
16313 PR target/56184
16314 * ira.c (max_regno_before_ira): Move from ...
16315 (ira): ... here.
16316 (fix_reg_equiv_init): Use max_regno_before_ira instead of
16317 vec_safe_length.
16318
16319 2013-02-13 Jakub Jelinek <jakub@redhat.com>
16320
16321 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
16322
16323 2013-02-13 Richard Biener <rguenther@suse.de>
16324
16325 PR lto/56295
16326 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
16327 globals in MEM_REFs.
16328
16329 2013-02-13 Richard Biener <rguenther@suse.de>
16330
16331 * loop-init.c (loop_optimizer_init): Clear loop state when
16332 re-initializing preserved loops.
16333 * loop-unswitch.c (unswitch_single_loop): Return whether
16334 we unswitched the loop. Do not verify loop state here.
16335 (unswitch_loops): When we unswitched a loop discover new loops.
16336
16337 2013-02-13 Kostya Serebryany <kcc@google.com>
16338
16339 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
16340 on x86_64 linux.
16341 * sanitizer.def: Rename __asan_init to __asan_init_v1.
16342
16343 2013-02-12 Dodji Seketeli <dodji@redhat.com>
16344
16345 Avoid instrumenting duplicated memory access in the same basic block
16346 * Makefile.in (asan.o): Add new dependency on hash-table.h
16347 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
16348 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
16349 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
16350 (free_mem_ref_resources, has_mem_ref_been_instrumented)
16351 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
16352 (get_mem_ref_of_assignment): New functions.
16353 (get_mem_refs_of_builtin_call): Extract from
16354 instrument_builtin_call and tweak a little bit to make it fit with
16355 the new signature.
16356 (instrument_builtin_call): Use the new
16357 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
16358 of is_gimple_builtin_call.
16359 (instrument_derefs, instrument_mem_region_access): Insert the
16360 instrumented memory reference into the hash table.
16361 (maybe_instrument_assignment): Renamed instrument_assignment into
16362 this, and change it to advance the iterator when instrumentation
16363 actually happened and return true in that case. This makes it
16364 homogeneous with maybe_instrument_assignment, and thus give a
16365 chance to callers to be more 'regular'.
16366 (transform_statements): Clear the memory reference hash table
16367 whenever we enter a new BB, when we cross a function call, or when
16368 we are done transforming statements. Use
16369 maybe_instrument_assignment instead of instrumentation. No more
16370 need to special case maybe_instrument_assignment and advance the
16371 iterator after calling it; it's now handled just like
16372 maybe_instrument_call. Update comment.
16373
16374 2013-02-13 Richard Biener <rguenther@suse.de>
16375
16376 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
16377 Fix loop discovery code.
16378
16379 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
16380
16381 PR inline-asm/56148
16382 * lra-constraints.c (process_alt_operands): Match early clobber
16383 operand with itself. Check conflicts with earlyclobber only if
16384 the operand is not reloaded. Prefer to reload conflicting operand
16385 if earlyclobber and matching operands are the same.
16386
16387 2013-02-12 Richard Biener <rguenther@suse.de>
16388
16389 PR lto/56297
16390 * lto-streamer-out.c (write_symbol): Do not output symbols
16391 for hard register variables.
16392
16393 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
16394
16395 PR target/54222
16396 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
16397 (umulsidi3_insn, mulsidi3_insn): New insns.
16398
16399 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
16400
16401 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
16402 (struct tune_params): Add vec_costs field.
16403 * config/arm/arm.c (arm_builtin_vectorization_cost)
16404 (arm_add_stmt_cost): New functions.
16405 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
16406 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
16407 (arm_default_vec_cost): New struct of type cpu_vec_costs.
16408 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
16409 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
16410 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
16411 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
16412
16413 2013-02-12 Richard Biener <rguenther@suse.de>
16414
16415 PR lto/56295
16416 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
16417 decls again if possible.
16418
16419 2013-02-12 Richard Biener <rguenther@suse.de>
16420
16421 PR middle-end/56288
16422 * tree-ssa.c (verify_ssa_name): Fix check, move
16423 SSA_NAME_IN_FREE_LIST check up.
16424
16425 2013-02-12 Jakub Jelinek <jakub@redhat.com>
16426 Steven Bosscher <steven@gcc.gnu.org>
16427
16428 PR rtl-optimization/56151
16429 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
16430 equal to op0 or op1, and last_insn pattern is CODE operation
16431 with MEM dest and one of the operands matches that MEM.
16432
16433 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
16434
16435 * doc/extend.texi: Document Function Multiversioning and "default"
16436 parameter string to target attribute.
16437 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
16438 target attribute parameter is "default".
16439 (ix86_compare_version_priority): Remove checks for target attribute.
16440 (ix86_mangle_function_version_assembler_name): Change error to sorry.
16441 Remove check for target attribute equal to NULL. Add assert.
16442 (ix86_generate_version_dispatcher_body): Change error to sorry.
16443
16444 2013-02-11 Iain Sandoe <iain@codesourcery.com>
16445 Jack Howarth <howarth@bromo.med.uc.edu>
16446 Patrick Marlier <patrick.marlier@gmail.com>
16447
16448 PR libitm/55693
16449 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
16450 define ENDFILE_SPEC as TM_DESTRUCTOR.
16451 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
16452
16453 2013-02-11 Alexander Potapenko <glider@google.com>
16454 Jack Howarth <howarth@bromo.med.uc.edu>
16455 Jakub Jelinek <jakub@redhat.com>
16456
16457 PR sanitizer/55617
16458 * config/darwin.c (cdtor_record): Rename ctor_record.
16459 (sort_cdtor_records): Rename sort_ctor_records.
16460 (finalize_dtors): New routine to sort destructors by
16461 priority before use in assemble_integer.
16462 (machopic_asm_out_destructor): Use finalize_dtors if needed.
16463
16464 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
16465
16466 PR rtl-optimization/56275
16467 * simplify-rtx.c (avoid_constant_pool_reference): Check that
16468 offset is non-negative and less than cmode size before
16469 calling simplify_subreg.
16470
16471 2013-02-11 Richard Biener <rguenther@suse.de>
16472
16473 PR tree-optimization/56264
16474 * cfgloop.h (fix_loop_structure): Adjust prototype.
16475 * loop-init.c (fix_loop_structure): Return the number of
16476 newly discovered loops.
16477 * tree-cfgcleanup.c (repair_loop_structures): When new loops
16478 are discovered, do a full loop-closed SSA rewrite.
16479
16480 2013-02-11 Richard Biener <rguenther@suse.de>
16481
16482 PR tree-optimization/56273
16483 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16484 first VRP run.
16485 (check_array_ref): Fix missing newline in dumps.
16486 (search_for_addr_array): Likewise.
16487
16488 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
16489
16490 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
16491
16492 2013-02-09 Jakub Jelinek <jakub@redhat.com>
16493
16494 PR target/56256
16495 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
16496
16497 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
16498
16499 PR rtl-optimization/56246
16500 * lra-constraints.c (simplify_operand_subreg): Try to reuse
16501 reload pseudo.
16502 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
16503 constraints are satisfied.
16504
16505 2013-02-08 Jeff Law <law@redhat.com>
16506
16507 PR debug/53948
16508 * emit-rtl.c (reg_is_parm_p): New function.
16509 * regs.h (reg_is_parm_p): New prototype.
16510 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
16511 callee-clobbered registers.
16512
16513 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
16514
16515 PR target/56043
16516 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
16517 If there is no implicit builtin declaration, just return NULL.
16518
16519 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16520
16521 * config/i386/sse.md (FMAMODEM): New mode iterator.
16522 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
16523 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
16524
16525 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16526
16527 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
16528 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
16529 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
16530
16531 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16532
16533 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
16534 (microblaze*-*-elf): Likewise.
16535 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
16536 LINK_SPEC.
16537 * config/microblaze/microblaze-c.c: Add builtin defines for
16538 _LITTLE_ENDIAN and _BIG_ENDIAN.
16539 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
16540 add to TARGET_DEFAULT flags.
16541 Expand ASM_SPEC and LINK_SPEC.
16542 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
16543 * config/microblaze/microblaze.md: Update extendsidi2 and
16544 movdi_internal instructions to use low-order / high-order reg
16545 print_operands.
16546 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
16547 options and inversemask / mask of LITTLE_ENDIAN.
16548 * config/microblaze/t-microblaze: Expand multilib options to
16549 include mlittle-endian (le) and update exceptions patterns.
16550
16551 2013-02-08 Jakub Jelinek <jakub@redhat.com>
16552
16553 PR rtl-optimization/56195
16554 * lra-constraints.c (get_reload_reg): Don't reuse regs
16555 if they have smaller mode than requested, if they have
16556 wider mode than requested, try to return a SUBREG.
16557
16558 PR tree-optimization/56250
16559 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
16560 if type is unsigned and code isn't MULT_EXPR.
16561
16562 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16563
16564 PR tree-optimization/56064
16565 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
16566 bits according to mode.
16567 * fixed-value.h (fixed_from_double_int)
16568 (const_fixed_from_double_int): Adjust comments.
16569
16570 2013-02-08 Richard Biener <rguenther@suse.de>
16571
16572 PR lto/56231
16573 * lto-streamer.h (struct data_in): Remove current_file, current_line
16574 and current_col members.
16575 * lto-streamer-out.c (lto_output_location): Stream changed bits
16576 en-block for efficiency.
16577 * lto-streamer-in.c (clear_line_info): Remove.
16578 (lto_input_location): Cache current file, line and column
16579 globally via local statics. Read changed bits en-block.
16580 (input_function): Do not call clear_line_info.
16581 (lto_read_body): Likewise.
16582 (lto_input_toplevel_asms): Likewise.
16583
16584 2013-02-08 Michael Matz <matz@suse.de>
16585
16586 PR tree-optimization/52448
16587 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
16588 (nt_call_phase): New static.
16589 (add_or_mark_expr): Only mark accesses with newer phase than any
16590 call seen.
16591 (nonfreeing_call_p): New.
16592 (nt_init_block): Update nt_call_phase, mark blocks as visited.
16593 (nt_fini_block): Keep blocks marked as visited.
16594 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
16595
16596 2013-02-08 Richard Biener <rguenther@suse.de>
16597
16598 * ira.c (ira): Free broken dominator information.
16599
16600 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
16601
16602 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
16603
16604 2013-02-08 Marek Polacek <polacek@redhat.com>
16605
16606 * cfgloop.c (verify_loop_structure): Add more checking of headers.
16607
16608 2013-02-08 Richard Biener <rguenther@suse.de>
16609
16610 PR middle-end/56181
16611 * cfgloop.h (flow_loops_find): Adjust.
16612 (bb_loop_header_p): Declare.
16613 * cfgloop.c (bb_loop_header_p): New function split out from ...
16614 (flow_loops_find): ... here. Adjust function signature,
16615 support incremental loop structure update.
16616 (verify_loop_structure): Cleanup. Verify a loop is a loop.
16617 * cfgloopmanip.c (fix_loop_structure): Move ...
16618 * loop-init.c (fix_loop_structure): ... here.
16619 (apply_loop_flags): Split out from ...
16620 (loop_optimizer_init): ... here.
16621 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
16622 in incremental mode, only remove dead loops here.
16623
16624 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
16625
16626 PR target/54222
16627 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
16628 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
16629 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
16630 (*round<mode>3.libgcc): New insns for fixed-modes.
16631 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
16632 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
16633 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
16634 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
16635 implementations. Define to __builtin_avr_absFX,
16636 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
16637 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
16638 __builtin_avr_countlsFX, respectively.
16639 * config/avr/avr-c.c (target.h): Include it.
16640 (enum avr_builtin_id): New enum.
16641 (avr_resolve_overloaded_builtin): New static function.
16642 (avr_register_target_pragmas): Use it to set
16643 targetm.resolve_overloaded_builtin.
16644 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
16645 tree nodes used by DEF_BUILTIN.
16646 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
16647 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
16648 <AVR_BUILTIN_xxBITS>: Same.
16649
16650 2013-02-08 Richard Biener <rguenther@suse.de>
16651
16652 * cfgloop.c (verify_loop_structure): Properly handle
16653 a loop exiting to another loop header.
16654 * ira-int.h (ira_loops): Remove.
16655 * ira.c (ira_loops): Remove.
16656 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
16657 (do_reload): Use loop_optimizer_finalize.
16658 * ira-build.c (create_loop_tree_nodes): Use get_loops and
16659 number_of_loops to access the loop tree.
16660 (more_one_region_p): Likewise.
16661 (finish_loop_tree_nodes): Likewise.
16662 (rebuild_regno_allocno_maps): Likewise.
16663 (mark_loops_for_removal): Likewise.
16664 (mark_all_loops_for_removal): Likewise.
16665 (remove_unnecessary_regions): Likewise.
16666 (ira_build): Likewise.
16667 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16668
16669 2013-02-08 Richard Biener <rguenther@suse.de>
16670
16671 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
16672 * ipa-pure-const.c (analyze_function): Avoid calling
16673 mark_irreducible_loops twice.
16674 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
16675
16676 2013-02-07 David S. Miller <davem@davemloft.net>
16677
16678 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
16679 on 'reg'.
16680 * var-tracking.c (vt_add_function_parameter): Test the presence of
16681 HAVE_window_save properly and do not remap argument registers when
16682 we have a leaf function.
16683
16684 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
16685
16686 PR bootstrap/56227
16687 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
16688 instead of "ll".
16689 * config/i386/i386.c (ix86_print_operand): Ditto.
16690
16691 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16692
16693 * lra-constraints.c (process_alt_operands): Fix recently added comment.
16694
16695 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
16696
16697 PR rtl-optimization/56225
16698 * lra-constraints.c (process_alt_operands): Check that reload hard
16699 reg can hold value for strict_low_part.
16700
16701 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16702
16703 PR debug/56154
16704 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
16705 dwarf2out_end_function.
16706 (in_first_function_p, maybe_at_text_label_p,
16707 first_loclabel_num_not_at_text_label): New variables.
16708 (dwarf2out_var_location): In the first function find out
16709 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
16710 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
16711 functions.
16712
16713 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
16714
16715 PR rtl-optimization/56178
16716 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
16717 SUBREG of a register. Tidy up related block of code.
16718 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
16719 note if the source is a register or a SUBREG of a register.
16720
16721 2013-02-07 Jakub Jelinek <jakub@redhat.com>
16722
16723 PR target/56228
16724 * config/rs6000/rs6000.md (ptrm): New mode attr.
16725 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
16726 call_value_indirect_aix<pttrsize>,
16727 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
16728 m in constraints.
16729
16730 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
16731
16732 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
16733 if -bnortl. Convert to strcmp and strncmp.
16734
16735 2013-02-07 Alan Modra <amodra@gmail.com>
16736
16737 PR target/54009
16738 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
16739 addresses won't wrap when offsetting.
16740 (rs6000_secondary_reload): Provide secondary reloads needed for
16741 wrapping LO_SUM addresses.
16742
16743 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
16744
16745 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
16746 MACH, just __MACH__.
16747
16748 2013-02-06 Richard Biener <rguenther@suse.de>
16749
16750 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
16751 instead of calling fix_loop_structure.
16752
16753 2013-02-06 Jakub Jelinek <jakub@redhat.com>
16754
16755 PR middle-end/56217
16756 * omp-low.c (use_pointer_for_field): Return false if
16757 lower_send_shared_vars doesn't generate any copy-out code.
16758
16759 2013-02-06 Tom de Vries <tom@codesourcery.com>
16760
16761 PR rtl-optimization/56131
16762 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
16763 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
16764 of the label is NULL. Add comment.
16765
16766 2013-02-05 Jakub Jelinek <jakub@redhat.com>
16767
16768 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
16769
16770 PR sanitizer/55374
16771 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
16772 (STATIC_LIBTSAN_LIBS): Likewise.
16773 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
16774 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
16775 is defined, don't add anything else beyond that.
16776 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
16777 (LINK_COMMAND_SPEC): Use them.
16778
16779 PR tree-optimization/56205
16780 * tree-stdarg.c (check_all_va_list_escapes): Return true if
16781 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
16782 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
16783
16784 2013-02-05 Richard Biener <rguenther@suse.de>
16785
16786 PR tree-optimization/53342
16787 PR tree-optimization/53185
16788 * tree-vectorizer.h (vect_check_strided_load): Remove.
16789 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
16790 not disallow peeling for vectorized strided loads.
16791 (vect_check_strided_load): Make static and simplify.
16792 (vect_analyze_data_refs): Adjust.
16793 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
16794 correctly when vectorizing strided loads.
16795
16796 2013-02-05 Richard Biener <rguenther@suse.de>
16797
16798 * doc/install.texi: Refer to ISL, not PPL.
16799
16800 2013-02-05 Jan Hubicka <jh@suse.cz>
16801
16802 PR tree-optimization/55789
16803 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
16804
16805 2013-02-05 Jan Hubicka <jh@suse.cz>
16806
16807 PR tree-optimization/55789
16808 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
16809 the dead call anyway.
16810
16811 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
16812
16813 PR sanitizer/55374
16814 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
16815
16816 2013-02-04 Alexander Potapenko <glider@google.com>
16817 Jack Howarth <howarth@bromo.med.uc.edu>
16818 Jakub Jelinek <jakub@redhat.com>
16819
16820 PR sanitizer/55617
16821 * config/darwin.c (sort_ctor_records): Stabilized qsort
16822 on constructor priority by using original position.
16823 (finalize_ctors): New routine to sort constructors by
16824 priority before use in assemble_integer.
16825 (machopic_asm_out_constructor): Use finalize_ctors if needed.
16826
16827 2013-02-04 Jakub Jelinek <jakub@redhat.com>
16828
16829 PR libstdc++/54314
16830 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
16831 about visibility on artificial decls.
16832 * config/sol2.c (solaris_assemble_visibility): Likewise.
16833
16834 2013-02-04 Kai Tietz <ktietz@redhat.com>
16835
16836 PR target/56186
16837 * config/i386/i386.c (function_value_ms_64): Add additional valtype
16838 argument and improve checking of return-argument types for 16-byte
16839 modes.
16840 (ix86_function_value_1): Add additional valtype argument on call
16841 of function_value_64.
16842 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
16843 handling infunction_value_64 function.
16844
16845 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
16846
16847 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
16848
16849 2013-02-04 Richard Biener <rguenther@suse.de>
16850
16851 PR tree-optimization/56188
16852 * tree-ssa-structalias.c (label_visit): Consider case with
16853 initially non-empty points-to set.
16854 (perform_var_substitution): Dump node mapping and clean up.
16855
16856 2013-02-04 Richard Guenther <rguenther@suse.de>
16857
16858 PR lto/56168
16859 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
16860 node prevail as last resort.
16861 (lto_symtab_merge_decls): Remove guard on LTRANS here.
16862 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
16863
16864 2013-02-04 Richard Biener <rguenther@suse.de>
16865
16866 PR tree-optimization/56113
16867 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
16868 Merge into ...
16869 (equiv_class_lookup_or_add): ... this.
16870 (label_visit): Adjust and fix error in previous patch.
16871 (perform_var_substitution): Adjust.
16872
16873 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
16874
16875 * config/sh/divtab.c: Fix formatting and comments throughout the file.
16876 * config/sh/sh4-300.md: Likewise.
16877 * config/sh/sh4a.md: Likewise.
16878 * config/sh/constraints.md: Likewise.
16879 * config/sh/sh.md: Likewise.
16880 * config/sh/netbsd-elf.h: Likewise.
16881 * config/sh/predicates.md: Likewise.
16882 * config/sh/sh-protos.h: Likewise.
16883 * config/sh/ushmedia.h: Likewise.
16884 * config/sh/linux.h: Likewise.
16885 * config/sh/sh.c: Likewise.
16886 * config/sh/superh.h: Likewise.
16887 * config/sh/elf.h: Likewise.
16888 * config/sh/sh4.md: Likewise.
16889 * config/sh/sh.h: Likewise.
16890
16891 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16892
16893 * config/pa/constraints.md: Adjust unused letters. Change "T"
16894 constraint to match_test floating_point_store_memory_operand().
16895 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
16896 (base14_operand): New.
16897 (floating_point_store_memory_operand): New.
16898 (integer_store_memory_operand): Revise to use base14_operand and
16899 reg_plus_base_memory_operand.
16900 (move_dest_operand): Allow symbolic_memory_operands.
16901 (symbolic_memory_operand): Check for LO_SOM.
16902 (symbolic_operand): Change default case to break.
16903 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
16904 CONST_DOUBLE values to be reloaded by putting them into memory when
16905 the destination is a floating point register.
16906 (movdf): Remove code to handle CONST_DOUBLE.
16907 (movsf): Likewise.
16908 (reload_indf_r1): New.
16909 (reload_insf_r1): New.
16910 Consistently use "Q" and "T" constraints with integer and floating
16911 point move instructions, respectively.
16912 (movdi): Remove FAIL.
16913 Change predicate for source operand unamed DImode move from
16914 general_operand to move_src_operand.
16915 (umulsidi3): Change predicate for destination operand to
16916 register_operand.
16917 Likewise for similar unamed patterns.
16918 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
16919 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
16920 (hppa_legitimize_address): Simplify mask calculation.
16921 (pa_emit_move_sequence): Revised handling of secondary reloads from
16922 REG+D addresses for floating point loads and stores. Directly handle
16923 loading CONST0_RTX (mode) to a floating point register.
16924 (pa_secondary_reload): Handle reloading DF and SFmode constant values
16925 to floating point registers. Don't restrict secondary reloads to
16926 floating point registers to integer modes. Revise some comments and
16927 cleanup some code.
16928 (TARGET_LEGITIMATE_ADDRESS_P): Define.
16929 (pa_legitimate_address_p): New.
16930 (pa_legitimize_reload_address): New.
16931 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
16932 (STRICT_REG_OK_FOR_BASE_P): New.
16933 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
16934 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
16935
16936 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
16937 Andrew Dixie <andrewd@gentrack.com>
16938
16939 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
16940 flag set.
16941
16942 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
16943
16944 * expmed.c (extract_bit_field_1): Pass the full width of the
16945 structure to get_best_reg_extraction_insn.
16946
16947 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16948
16949 PR target/54601
16950 * configure.ac (use_cxa_atexit): Add AIX.
16951 * configure: Regenerate.
16952
16953 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
16954
16955 2013-02-01 Jakub Jelinek <jakub@redhat.com>
16956
16957 PR debug/54793
16958 * final.c (need_profile_function): New variable.
16959 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
16960 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
16961 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
16962 notes, targetm.asm_out.function_prologue doesn't emit anything,
16963 HAVE_prologue and profiler should be emitted before prologue,
16964 set need_profile_function instead of emitting it.
16965 (final_scan_insn): If need_profile_function, emit
16966 profile_function on the first NOTE_INSN_BASIC_BLOCK or
16967 NOTE_INSN_FUNCTION_BEG note.
16968
16969 2013-02-01 Richard Henderson <rth@redhat.com>
16970
16971 * config/rs6000/rs6000.md (smulditi3): New.
16972 (umulditi3): New.
16973
16974 * config/alpha/alpha.md (umulditi3): New.
16975
16976 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
16977
16978 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
16979 (ASM_OUTPUT_ALIGNED_LOCAL): New.
16980
16981 2013-02-01 Richard Biener <rguenther@suse.de>
16982
16983 PR tree-optimization/56113
16984 * tree-ssa-structalias.c (label_visit): Reduce work for
16985 single-predecessor nodes.
16986
16987 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
16988
16989 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
16990 range isn't testing for zero.
16991
16992 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
16993
16994 PR middle-end/56113
16995 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
16996
16997 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
16998 Nick Clifton <nickc@redhat.com>
16999
17000 * config/v850/constraints.md (Q): Define as a memory constraint.
17001 * config/v850/predicates.md (label_ref_operand): New predicate.
17002 (e3v5_shift_operand): New predicate.
17003 (ior_operator): New predicate.
17004 * config/v850/t-v850: Add e3v5 multilib.
17005 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17006 (v850_gen_movdi): Prototype.
17007 * config/v850/v850.c: Add support for e3v5 architecture.
17008 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17009 TARGET_V850E_UP.
17010 (construct_save_jarl): Add e3v5 long JARL support.
17011 (v850_adjust_insn_length): New function. Adjust length of call
17012 insns when using e3v5 instructions.
17013 (v850_gen_movdi): New function: Generate instructions to move a
17014 DImode value.
17015 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17016 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17017 (TARGET_USE_FPU): Enable for e3v5.
17018 (CONST_OK_FOR_W): New macro.
17019 (ADJUST_INSN_LENGTH): Define.
17020 * config/v850/v850.md (UNSPEC_LOOP): Define.
17021 (attr cpu): Add v850e3v5.
17022 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17023 (movdi): New pattern.
17024 (movdi_internal): New pattern.
17025 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17026 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17027 (cstoresf4): Likewise.
17028 (cstoredf4): Likewise.
17029 (insv): New pattern.
17030 (rotlso3_a): New pattern.
17031 (rotlsi3_b): New pattern
17032 (rotlsi3_v850e3v5): New pattern.
17033 (doloop_begin): New pattern.
17034 (fix_loop_counter): New pattern.
17035 (doloop_end): New pattern.
17036 (branch_normal): Add e3v5 long branch support.
17037 (branch_invert): Likewise.
17038 (branch_z_normal): Likewise.
17039 (branch_z_invert): Likewise.
17040 (branch_nz_normal): Likewise.
17041 (branch_nz_invert): Likewise.
17042 (call_internal_short): Add e3v5 register-indirect JARL support.
17043 (call_internal_long): Likewise.
17044 (call_value_internal_short): Likewise.
17045 (call_value_internal_long): Likewise.
17046 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17047 (mloop): New option.
17048 * config.gcc: Add support for configuring v840e3v5 target.
17049 * doc/invoke.texi: Document new v850 specific command line options.
17050
17051 2013-01-31 Paul Koning <ni1d@arrl.net>
17052
17053 PR debug/55059
17054 PR debug/54508
17055 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17056 children if parent is a class.
17057 (prune_unused_types_prune): Don't add DW_AT_declaration.
17058
17059 2013-01-31 Richard Biener <rguenther@suse.de>
17060
17061 PR tree-optimization/56157
17062 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17063 match up operand with SLP child.
17064
17065 2013-01-31 Jason Merrill <jason@redhat.com>
17066
17067 PR debug/54410
17068 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17069 parameters the first time.
17070 (gen_scheduled_generic_parms_dies): Check completeness here.
17071
17072 2013-01-31 Richard Biener <rguenther@suse.de>
17073
17074 PR middle-end/53073
17075 * common.opt (faggressive-loop-optimizations): New flag,
17076 enabled by default.
17077 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17078 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17079 infer_loop_bounds_from_undefined by it.
17080
17081 2013-01-31 Richard Biener <rguenther@suse.de>
17082
17083 PR tree-optimization/56150
17084 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17085 visit virtual operands.
17086 (find_uses_to_rename_bb): Likewise.
17087
17088 2013-01-31 Richard Biener <rguenther@suse.de>
17089
17090 PR tree-optimization/56150
17091 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17092 mixed store non-store stmts.
17093
17094 2013-01-30 Jakub Jelinek <jakub@redhat.com>
17095
17096 PR sanitizer/55374
17097 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
17098 LIBASAN_EARLY_SPEC is defined.
17099 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
17100 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
17101 before %o.
17102 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
17103
17104 PR c++/55742
17105 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
17106 invalid args instead of ICEing on it.
17107 (ix86_valid_target_attribute_tree): Return error_mark_node if
17108 ix86_valid_target_attribute_inner_p failed.
17109 (ix86_valid_target_attribute_p): Return false only if
17110 ix86_valid_target_attribute_tree returned error_mark_node. Allow
17111 target("default") attribute.
17112 (sorted_attr_string): Change argument from const char * to tree,
17113 merge in all target attribute arguments rather than just one.
17114 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
17115 instead of free. Avoid using strcat.
17116 (ix86_mangle_function_version_assembler_name): Mangle
17117 target("default") as if no target attribute is present. Adjust
17118 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
17119 instead of xmalloc and XDELETEVEC instead of free.
17120 (ix86_function_versions): Don't return true if one of the decls
17121 doesn't have target attribute. If they don't and one of the decls
17122 is DECL_FUNCTION_VERSIONED, report an error. Adjust
17123 sorted_attr_string caller. Use XDELETEVEC instead of free.
17124 (ix86_supports_function_versions): Remove.
17125 (make_name): Fix up formatting.
17126 (make_dispatcher_decl): Remove resolver_name and its initialization.
17127 Avoid leaking memory.
17128 (is_function_default_version): Return true if there is
17129 target("default") attribute rather than no target attribute at all.
17130 (make_resolver_func): Avoid leaking memory.
17131 (ix86_generate_version_dispatcher_body): Likewise.
17132 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
17133 * target.def (supports_function_versions): Remove.
17134 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
17135 * doc/tm.texi: Regenerated.
17136
17137 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
17138
17139 PR rtl-optimization/56144
17140 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
17141 for values with side effects.
17142
17143 2013-01-30 Richard Biener <rguenther@suse.de>
17144
17145 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
17146 (sparseset_pop): Likewise.
17147 * cfganal.c (compute_idf): Likewise. Increase work-stack size
17148 to be able to use quick_push in the worker loop.
17149
17150 2013-01-30 Marek Polacek <polacek@redhat.com>
17151
17152 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
17153
17154 2013-01-30 Richard Biener <rguenther@suse.de>
17155
17156 PR lto/56147
17157 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
17158
17159 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
17160
17161 PR tree-optimization/56064
17162 * fixed-value.c (fixed_from_double_int): New function.
17163 * fixed-value.h (fixed_from_double_int): New prototype.
17164 (const_fixed_from_double_int): New static inline function.
17165 * fold-const.c (native_interpret_fixed): New static function.
17166 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
17167 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
17168 (native_encode_fixed): New static function.
17169 (native_encode_expr) <FIXED_CST>: Use it.
17170 (native_interpret_int): Move double_int worker code to...
17171 * double-int.c (double_int::from_buffer): ...this new static method.
17172 * double-int.h (double_int::from_buffer): Prototype it.
17173
17174 2013-01-30 Richard Biener <rguenther@suse.de>
17175
17176 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
17177 New pointer-map and obstack.
17178 (init_alias_vars): Allocate pointer-map and obstack.
17179 (delete_points_to_sets): Free them.
17180 (find_what_var_points_to): Cache result.
17181 (find_what_p_points_to): Adjust for changed interface of
17182 find_what_var_points_to.
17183 (compute_points_to_sets): Likewise.
17184 (ipa_pta_execute): Likewise.
17185
17186 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17187
17188 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
17189 * configure: Regenerate.
17190 * config.in: Regenerate.
17191 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
17192 #nobits/#progbits if supported.
17193
17194 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
17195
17196 PR target/56121
17197 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
17198 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
17199 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
17200
17201 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17202
17203 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
17204 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
17205
17206 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17207
17208 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
17209 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
17210
17211 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17212
17213 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
17214 declaration.
17215 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
17216 * config/arm/cortex-a7.md: New bypasses using
17217 arm_mac_accumulator_is_result.
17218
17219 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17220
17221 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
17222 (cortex_a7_neon_mla): Likewise.
17223 (cortex_a7_fpfmad): New reservation.
17224 (cortex_a7_fpmacs): Use ffmas and update required units.
17225 (cortex_a7_fpmuld): Update required units and latency.
17226 (cortex_a7_fpmacd): Likewise.
17227 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
17228 (cortex_a7_neon). Likewise.
17229 (bypass) Update participating units.
17230
17231 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
17232
17233 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
17234 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
17235 from fmac to ffma.
17236 * config/arm/vfp11.md (vfp_farith): Use ffmas.
17237 (vfp_fmul): Use ffmad.
17238 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
17239 (cortex_r4_fmacd): Use ffmad.
17240 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
17241 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
17242 (cortex_a9_fmacd): Use ffmad.
17243 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
17244 (cortex_a8_vfp_macd): Use ffmad.
17245 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
17246 (cortex_a5_fpmacd): Use ffmad.
17247 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
17248 (cortex_a15_vfp_macd): Use ffmad.
17249 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
17250
17251 2013-01-29 Jason Merrill <jason@redhat.com>
17252
17253 PR libstdc++/54314
17254 * varasm.c (default_assemble_visibility): Don't warn about
17255 visibility on artificial decls.
17256
17257 2013-01-29 Richard Biener <rguenther@suse.de>
17258
17259 PR tree-optimization/56113
17260 * tree-ssa-structalias.c (equiv_class_lookup): Also return
17261 the bitmap leader.
17262 (label_visit): Free duplicate bitmaps and record the leader instead.
17263 (perform_var_substitution): Adjust.
17264
17265 2013-01-29 Richard Biener <rguenther@suse.de>
17266
17267 PR tree-optimization/55270
17268 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
17269 the CFG, schedule loops for fixup.
17270
17271 2013-01-29 Nick Clifton <nickc@redhat.com>
17272
17273 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
17274 SP_REG.
17275
17276 2013-01-28 Leif Ekblad <leif@rdos.net>
17277
17278 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
17279 * config/i386/i386.h (TARGET_RDOS): New macro.
17280 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
17281 * config/i386/i386.c (ix86_option_override_internal): For 64bit
17282 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
17283 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
17284 DEFAULT_LARGE_SECTION_THRESHOLD.
17285 * config/i386/i386.md (R14_REG, R15_REG): New constants.
17286 * config/i386/rdos.h: New file.
17287 * config/i386/rdos64.h: New file.
17288
17289 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
17290
17291 PR other/54814
17292 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
17293 TEST_HARD_REG_BIT.
17294
17295 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17296
17297 PR rtl-optimization/56117
17298 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
17299 call cselib_lookup_from_insn on the MEM before calling
17300 add_insn_mem_dependence.
17301
17302 2013-01-28 Richard Biener <rguenther@suse.de>
17303
17304 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
17305 to a stmt that didn't have one.
17306 (copy_phis_for_bb): Likewise for PHI arguments.
17307 (copy_debug_stmt): Likewise for debug stmts.
17308
17309 2013-01-28 Richard Biener <rguenther@suse.de>
17310
17311 PR tree-optimization/56034
17312 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
17313 (partition_builtin_p): Adjust.
17314 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
17315 it is the last partition.
17316 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
17317 up the vertex for the definition.
17318 (classify_partition): Classify whether a partition is a
17319 PKIND_REDUCTION, thus has uses outside of the loop.
17320 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
17321 Merge all PKIND_REDUCTION partitions into the last partition.
17322 (tree_loop_distribution): Seed partitions from reductions as well.
17323
17324 2013-01-28 Jakub Jelinek <jakub@redhat.com>
17325
17326 PR tree-optimization/56125
17327 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
17328 pow(x,c) into sqrt(x) * powi(x, n/2) or
17329 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
17330 optimizing for size.
17331 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
17332 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
17333 integer.
17334
17335 PR tree-optimization/56094
17336 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
17337 to UNKNOWN_LOCATION while gimplifying expr.
17338
17339 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
17340
17341 PR target/56114
17342 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
17343 operand 0 in movabs insn template for -masm=intel asm alternative.
17344 (*movabs<mode>_2): Ditto for operand 1.
17345
17346 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
17347
17348 PR target/54663
17349 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
17350 of microblaze-c.o
17351
17352 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17353
17354 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
17355 tm_file.
17356
17357 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17358
17359 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
17360 Undef to avoid warning.
17361
17362 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17363
17364 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
17365 * configure: Regenerate.
17366
17367 2013-01-25 Jakub Jelinek <jakub@redhat.com>
17368
17369 PR tree-optimization/56098
17370 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
17371 for stmts with volatile ops.
17372 (cond_store_replacement): Don't optimize if assign has volatile ops.
17373 (cond_if_else_store_replacement_1): Don't optimize if either
17374 then_assign or else_assign have volatile ops.
17375 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
17376 volatile ops.
17377
17378 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17379
17380 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
17381
17382 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17383
17384 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
17385 missing ':' in asm example.
17386
17387 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
17388
17389 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
17390 entries into lane and laneq entries.
17391 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
17392 Remove AdvSIMD scalar modes.
17393 (aarch64_sq<r>dmulh_laneq<mode>): New.
17394 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
17395 modes.
17396 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
17397 builtin implementations to relfect changes in RTL in aarch64-simd.md.
17398 * config/aarch64/iterators.md (VCOND): New.
17399 (VCONQ): New.
17400
17401 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
17402
17403 PR target/54222
17404 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
17405 Add NULL LIBNAME argument to existing definitions.
17406 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
17407 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
17408 * config/avr/avr.c (DEF_BUILTIN): Same.
17409 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
17410 (avr_expand_builtin): Expand to a vanilla call if a libgcc
17411 implementation is available (DECL_ASSEMBLER_NAME is set).
17412 (avr_fold_absfx): New static function.
17413 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
17414 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
17415 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
17416 AVR_BUILTIN_ABSLLK.
17417 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
17418 (abshk, absk, abslk, absllk): Provide as static inline functions.
17419
17420 2013-01-25 Marek Polacek <polacek@redhat.com>
17421
17422 PR tree-optimization/56035
17423 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
17424
17425 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
17426
17427 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
17428 (*movtf_internal_rex64): Add (!o,C) alternative
17429 (*movxf_internal_rex64): Ditto.
17430 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
17431
17432 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
17433
17434 * doc/invoke.texi: fix typo.
17435 * doc/objc.texi: fix typo.
17436
17437 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
17438
17439 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
17440 for the first two alternatives.
17441
17442 2013-01-24 Diego Novillo <dnovillo@google.com>
17443
17444 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
17445 (ggc-zone.o): Remove.
17446 * configure.ac: Remove option --with-gc.
17447 * configure: Re-generate.
17448 * doc/install.texi: Remove documentation for --with-gc.
17449 * gengtype.c (write_enum_defn): Remove. Update all users.
17450 (write_Types_process_field): Remove generation of gt_e_* argument.
17451 (output_type_enum): Remove. Update all users.
17452 (write_enum_defn): Remove. Update all users.
17453 (enum alloc_zone): Remove. Update all users.
17454 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
17455 * ggc-common.c (ggc_splay_alloc): Remove first argument.
17456 Update all callers.
17457 (struct ptr_data): Remove field TYPE. Update all users.
17458 (gt_pch_note_object): Remove argument TYPE. Update all users.
17459 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
17460 Update all users.
17461 * ggc-none.c (ggc_alloc_typed_stat): Remove.
17462 (struct alloc_zone): Remove.
17463 (ggc_internal_alloc_zone_stat): Remove.
17464 (ggc_internal_cleared_alloc_zone_stat): Remove.
17465 * ggc-page.c (ggc_alloc_typed_stat): Remove.
17466 (ggc_pch_count_object): Remove last argument. Update all users.
17467 (ggc_pch_alloc_object): Remove last argument. Update all users.
17468 (struct alloc_zone): Remove.
17469 * ggc-zone.c: Remove.
17470 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
17471 (struct alloc_zone): Remove.
17472 (ggc_alloc_typed_stat): Remove.
17473 (ggc_alloc_typed): Remove.
17474 (ggc_splay_alloc): Remove first argument.
17475 (rtl_zone): Remove. Update all users.
17476 (tree_zone): Remove. Update all users.
17477 (tree_id_zone): Remove. Update all users.
17478 (ggc_internal_zone_alloc_stat): Remove. Update all users.
17479 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
17480 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
17481 * tree-ssanames.c: Remove references to zone allocator in comments.
17482
17483 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
17484
17485 * config/avr/avr.c (avr_out_fract): Make register numbers that
17486 might be outside of source operand signed.
17487
17488 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
17489
17490 * config/i386/constraints.md (Yf): New constraint.
17491 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
17492 of f constraint to conditionaly disable x87 register preferences.
17493 (*movdf_internal): Ditto.
17494 (*movsf_internal): Ditto.
17495
17496 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
17497
17498 PR inline-asm/55934
17499 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
17500 that have operands with impossible constraints.
17501 Add a FIXME for a speed-up opportunity.
17502 * lra-constraints.c (process_alt_operands): Verify that a class
17503 selected from constraints on asms is valid for the operand mode.
17504 (curr_insn_transform): Remove incorrect comment.
17505
17506 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
17507
17508 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
17509 TOC operand is a valid symbol ref in the constant pool.
17510
17511 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17512
17513 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
17514
17515 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
17516
17517 PR target/54222
17518 * config/avr/stdfix.h: New file.
17519 * t-avr (stdfix-gcc.h): New rule to build it.
17520 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
17521
17522 2013-01-23 Kostya Serebryany <kcc@google.com>
17523
17524 * config/darwin.h: remove dependency on
17525 CoreFoundation (asan on Mac OS).
17526
17527 2013-01-23 Jakub Jelinek <jakub@redhat.com>
17528
17529 PR target/49069
17530 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
17531 instead of cmpdi_operand for first comparison operand.
17532 Don't assert that comparison operands aren't both constants.
17533
17534 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17535
17536 * doc/install.texi (Downloading the Source): Update references to
17537 downloading separate components.
17538
17539 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
17540
17541 * doc/extend.texi (__int128): Improve grammar.
17542
17543 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
17544
17545 PR target/56028
17546 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
17547 alternative to (o,r).
17548 (*movdi_internal_rex64): Remove (!o,n) alternative.
17549 (DImode immediate->memory splitter): Remove.
17550 (DImode immediate->memory peephole2): Remove.
17551 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
17552 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
17553 alternative to (!o,*r).
17554 (*movtf_internal_sse): New pattern.
17555 (*movxf_internal_rex64): New pattern.
17556 (*movxf_internal): Disable for TARGET_64BIT.
17557 (*movdf_internal_rex64): Remove (!o,F) alternative.
17558
17559 2013-01-22 Jakub Jelinek <jakub@redhat.com>
17560
17561 PR middle-end/56074
17562 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
17563 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
17564 * tree-vect-loop-manip.c (find_loop_location): Also ignore
17565 stmt locations where LOCATION_LOCUS of the stmt location is
17566 UNKNOWN_LOCATION or BUILTINS_LOCATION.
17567
17568 PR target/55686
17569 * config/i386/i386.md (UNSPEC_STOS): New.
17570 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
17571 *strsetqi_1): Add UNSPEC_STOS.
17572
17573 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
17574
17575 PR c++/56067
17576 * doc/invoke.texi: Remove left over -Wsynth example.
17577
17578 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17579
17580 PR tree-optimization/56051
17581 * fold-const.c (fold_binary_loc): Don't fold
17582 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
17583 a narrowing conversion, or widening conversion from signed
17584 to unsigned.
17585
17586 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17587
17588 PR rtl-optimization/56023
17589 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
17590 dependent on debug instruction.
17591
17592 2013-01-21 Martin Jambor <mjambor@suse.cz>
17593
17594 PR middle-end/56022
17595 * function.c (allocate_struct_function): Call
17596 invoke_set_current_function_hook earlier.
17597
17598 2013-01-21 Jakub Jelinek <jakub@redhat.com>
17599
17600 * reload1.c (init_reload): Only initialize reload_obstack
17601 during the first call.
17602
17603 2013-01-21 Marek Polacek <polacek@redhat.com>
17604
17605 * cfgloop.c (verify_loop_structure): Fix up grammar.
17606
17607 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
17608
17609 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
17610 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
17611
17612 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17613
17614 PR target/56058
17615 * config/arm/marvell-pj4.md: Update copyright year.
17616 Fix up use of alu to alu_reg and simple_alu_imm.
17617
17618 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
17619
17620 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
17621
17622 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
17623
17624 PR target/55433
17625 * lra-constraints.c (curr_insn_transform): Don't reuse original
17626 insn for secondary memory move when memory mode should be different.
17627
17628 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17629
17630 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
17631 atomic_storedi_1): New patterns.
17632
17633 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
17634
17635 btver2 pipeline descriptions.
17636 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
17637 descriptions.
17638 * config/i386/i386.md (btver2_decode): New type attributes.
17639 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
17640 type attributes.
17641 * config/i386/btver2.md: New file describing btver2 pipelines.
17642
17643 2013-01-19 Andrew Pinski <apinski@cavium.com>
17644
17645 PR tree-optimization/52631
17646 * tree-ssa-sccvn (visit_use): Before looking up the original
17647 statement, try looking up the simplified expression.
17648
17649 2013-01-19 Anthony Green <green@moxielogic.com>
17650
17651 * config/moxie/moxie.c (moxie_expand_prologue): Set
17652 current_function_static_stack_size.
17653
17654 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17655
17656 PR tree-optimization/56029
17657 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
17658 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
17659
17660 2013-01-18 Sharad Singhai <singhai@google.com>
17661
17662 PR tree-optimization/55995
17663 * dumpfile.c (dump_loc): Print location only if available.
17664 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
17665
17666 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
17667
17668 PR target/55433
17669 * lra-constraints.c (curr_insn_transform): Reuse original insn for
17670 secondary memory move.
17671 (inherit_reload_reg): Use rclass instead of cl for
17672 check_secondary_memory_needed_p.
17673
17674 2013-01-18 Jakub Jelinek <jakub@redhat.com>
17675
17676 PR middle-end/56015
17677 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
17678 the case where writing real complex part of target modifies op1.
17679
17680 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
17681
17682 * config/aarch64/aarch64-simd.md
17683 (aarch64_vcond_internal<mode>): Handle unordered cases.
17684 * config/aarch64/iterators.md (v_cmp_result): New.
17685
17686 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
17687 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17688
17689 * config/arm/marvell-pj4.md: New file.
17690 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
17691 * config/arm/arm.md (generic_sched): Add marvell_pj4.
17692 (generic_vfp): Likewise.
17693 * config/arm/arm-cores.def: Add marvell-pj4.
17694 * config/arm/arm-tune.md: Regenerate.
17695 * config/arm/arm-tables.opt: Regenerate.
17696 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
17697 * doc/invoke.texi: Document marvell-pj4.
17698
17699 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
17700
17701 * config/aarch64/arm_neon.h: Map scalar types to standard types.
17702
17703 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17704
17705 PR debug/54114
17706 PR debug/54402
17707 PR debug/49888
17708 * var-tracking.c (negative_power_of_two_p): New.
17709 (global_get_addr_cache, local_get_addr_cache): New.
17710 (get_addr_from_global_cache, get_addr_from_local_cache): New.
17711 (vt_canonicalize_addr): Rewrite using the above. Adjust the
17712 heading comment.
17713 (vt_stack_offset_p): Remove.
17714 (vt_canon_true_dep): Always canonicalize loc's address.
17715 (clobber_overlapping_mems): Make sure we have a MEM.
17716 (local_get_addr_clear_given_value): New.
17717 (val_reset): Clear local cached entries.
17718 (compute_bb_dataflow): Create and release the local cache.
17719 Disable duplicate MEMs clobbering.
17720 (emit_notes_in_bb): Clobber MEMs likewise.
17721 (vt_emit_notes): Create and release the local cache.
17722 (vt_initialize, vt_finalize): Create and release the global
17723 cache, respectively.
17724 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
17725
17726 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17727
17728 PR libmudflap/53359
17729 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
17730 not found in the symtab.
17731
17732 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
17733
17734 PR debug/56006
17735 PR rtl-optimization/55547
17736 PR rtl-optimization/53827
17737 PR debug/53671
17738 PR debug/49888
17739 * alias.c (offset_overlap_p): New, factored out of...
17740 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
17741 the conservative special case for symbolic constants. Don't
17742 adjust zero sizes on alignment.
17743
17744 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
17745
17746 PR rtl-optimization/52573
17747 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
17748 REG_UNUSED for the same register.
17749
17750 2013-01-17 Richard Biener <rguenther@suse.de>
17751 Marek Polacek <polacek@redhat.com>
17752
17753 PR rtl-optimization/55833
17754 * loop-unswitch.c (unswitch_loops): Move loop verification...
17755 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
17756 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
17757 Set it to true when we're removing a loop from hierarchy tree in
17758 an irreducible region.
17759 (fix_bb_placements): Adjust caller.
17760 (fix_loop_placements): Likewise.
17761
17762 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
17763
17764 * config/avr/builtins.def (DEF_BUILTIN): Factor out
17765 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
17766 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
17767 Remove ID. Adjust comments.
17768 * config/avr/avr-c.c (avr_builtin_name): Remove.
17769 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
17770 * config/avr/avr.c (avr_tolower): New static function.
17771 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
17772 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
17773 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
17774 default expansion.
17775
17776 2013-01-17 Jan Hubicka <jh@suse.cz>
17777
17778 PR tree-optimization/55273
17779 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
17780
17781 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
17782
17783 PR target/55981
17784 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
17785 store through atomic_store<mode>_1.
17786 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
17787
17788 2013-01-17 Martin Jambor <mjambor@suse.cz>
17789
17790 PR tree-optimizations/55264
17791 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
17792 for virtual methods.
17793 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
17794 virtual methods before inlining is over.
17795 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
17796 virtual functions.
17797 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
17798 non-virtual.
17799
17800 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
17801
17802 PR rtl-optimization/56005
17803 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
17804 pending reads for prefetch.
17805
17806 2013-01-16 Ian Bolton <ian.bolton@arm.com>
17807
17808 * config/aarch64/aarch64.md
17809 (*cstoresi_neg_uxtw): New pattern.
17810 (*cmovsi_insn_uxtw): New pattern.
17811 (*<optab>si3_uxtw): New pattern.
17812 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
17813 (*<optab>si3_insn_uxtw): New pattern.
17814 (*bswapsi2_uxtw): New pattern.
17815
17816 2013-01-16 Richard Biener <rguenther@suse.de>
17817
17818 * tree-inline.c (tree_function_versioning): Remove set but
17819 never used variable.
17820
17821 2013-01-16 Richard Biener <rguenther@suse.de>
17822
17823 PR tree-optimization/55964
17824 * tree-flow.h (rename_variables_in_loop): Remove.
17825 (rename_variables_in_bb): Likewise.
17826 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
17827 (copy_loop_before): Adjust and delete update-ssa status.
17828 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
17829 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
17830 (rename_variables_in_loop): Remove.
17831 (slpeel_update_phis_for_duplicate_loop): Likewise.
17832 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
17833 use available cfg machinery instead of duplicating it.
17834 Update PHI nodes and perform poor-mans SSA update here.
17835 (slpeel_tree_peel_loop_to_edge): Adjust.
17836
17837 2013-01-16 Richard Biener <rguenther@suse.de>
17838
17839 PR tree-optimization/54767
17840 PR tree-optimization/53465
17841 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
17842 (vrp_visit_phi_node): For PHI arguments coming via backedges
17843 drop all symbolical range information.
17844 (execute_vrp): Compute backedges.
17845
17846 2013-01-16 Richard Biener <rguenther@suse.de>
17847
17848 * doc/install.texi: Update CLooG and ISL requirements to
17849 0.18.0 and 0.11.1.
17850
17851 2013-01-16 Christian Bruel <christian.bruel@st.com>
17852
17853 PR target/55301
17854 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
17855 (broken_move): Handle UNSPECV_SP_SWITCH_B.
17856 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
17857
17858 2013-01-16 DJ Delorie <dj@redhat.com>
17859
17860 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
17861 (UNSPECV_SP_SWITCH_E): New.
17862 (sp_switch_1): Change to an unspec.
17863 (sp_switch_2): Change to an unspec. Don't use post-inc when we
17864 replace $r15.
17865
17866 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
17867
17868 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
17869 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
17870 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
17871 (expand_mem_thread_fence): Ditto.
17872 (expand_mem_signal_fence): Ditto.
17873 (expand_atomic_load): Ditto.
17874 (expand_atomic_store): Ditto.
17875
17876 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
17877
17878 PR rtl-optimization/55547
17879 PR rtl-optimization/53827
17880 PR debug/53671
17881 PR debug/49888
17882 * alias.c (memrefs_conflict_p): Set sizes to negative after
17883 AND adjustments.
17884
17885 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17886
17887 PR target/55940
17888 * function.c (thread_prologue_and_epilogue_insns): Always
17889 add crtl->drap_reg to set_up_by_prologue.set, even if
17890 stack_realign_drap is false.
17891
17892 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17893
17894 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
17895 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
17896 *call): Fix indention.
17897
17898 2013-01-15 Tom de Vries <tom@codesourcery.com>
17899
17900 PR target/55876
17901 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
17902 Update comment.
17903
17904 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
17905
17906 PR rtl-optimization/55153
17907 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
17908
17909 2013-01-15 Martin Jambor <mjambor@suse.cz>
17910
17911 PR tree-optimization/55920
17912 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
17913 accesses as grp_to_be_debug_replaced.
17914
17915 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17916
17917 PR tree-optimization/55920
17918 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
17919 there is non-useless type conversion needed from debug rhs to lhs,
17920 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
17921
17922 2013-01-15 Joseph Myers <joseph@codesourcery.com>
17923 Mikael Pettersson <mikpe@it.uu.se>
17924
17925 PR target/43961
17926 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
17927 Thumb.
17928 (ASM_OUTPUT_CASE_LABEL): Remove.
17929 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
17930 * final.c (shorten_branches): Update alignment of labels before
17931 jump tables if CASE_VECTOR_SHORTEN_MODE.
17932
17933 2013-01-15 Richard Biener <rguenther@suse.de>
17934
17935 PR bootstrap/55961
17936 * system.h: Do not include gmp.h for building host tools.
17937
17938 2013-01-15 Richard Biener <rguenther@suse.de>
17939
17940 PR middle-end/55882
17941 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
17942 account for bitpos when computing alignment.
17943
17944 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
17945
17946 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
17947 (ix86_target_macros_internal): Likewise.
17948
17949 * config/i386/i386.c (m_CORE2I7): Removed.
17950 (m_CORE_HASWELL): New macro.
17951 (m_CORE_ALL): Likewise.
17952 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
17953 (initial_ix86_arch_features): Likewise.
17954 (processor_target_table): Initializations for Core avx2.
17955 (cpu_names): New names "core-avx2".
17956 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
17957 PROCESSOR_CORE_HASWELL.
17958 (ix86_issue_rate): New case.
17959 (ia32_multipass_dfa_lookahead): Likewise.
17960 (ix86_sched_init_global): Likewise.
17961
17962 * config/i386/i386.h (TARGET_HASWELL): New macro.
17963 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
17964 (processor_type): New PROCESSOR_HASWELL.
17965
17966 2013-01-15 Jakub Jelinek <jakub@redhat.com>
17967
17968 PR tree-optimization/55955
17969 * tree-vect-loop.c (vectorizable_reduction): Give up early on
17970 *SHIFT_EXPR and *ROTATE_EXPR codes.
17971
17972 PR tree-optimization/48766
17973 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
17974 -ftrapv disable -fwrapv.
17975
17976 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17977
17978 PR target/55974
17979 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
17980 etc. to 1 and not to __flash.
17981 Use LL suffix for __INT24_MAX__ with -mint8.
17982 Use ULL suffix for __UINT24_MAX__ with -mint8.
17983
17984 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
17985
17986 * config/avr/avr-arch.h
17987 (struct base_arch_s): Use typedef avr_arch_t instead.
17988 (struct arch_info_s): Use typedef avr_arch_info_t instead.
17989 (struct mcu_type_s): Use typedef avr_mcu_t instead.
17990 * config/avr/avr.c: Same.
17991 * config/avr/avr-devices.c: Same.
17992 * config/avr/driver-avr.c: Same.
17993 * config/avr/gen-avr-mmcu-texi.c: Same.
17994 * config/avr/avr-mcus.def: Adjust comment.
17995
17996 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
17997
17998 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
17999 * config/aarch64/iterators.md (VALLDI): New.
18000
18001 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18002 Andi Kleen <ak@linux.intel.com>
18003
18004 PR target/55948
18005 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18006 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18007 memmodel flag.
18008
18009 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18010
18011 * config/avr/avr-stdint.h: Remove trailing blanks.
18012 * config/avr/avr-log.h: Same.
18013 * config/avr/avr-arch.h: Same.
18014 * config/avr/avr-devices.c: Same.
18015 * config/avr/avr-dimode.md: Same.
18016 * config/avr/predicates.md: Same.
18017 * config/avr/avr-c.c: Same. And fix typo.
18018
18019 * config/avr/avr-protos.h: Same. And:
18020 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18021 (init_cumulative_args): Rename to avr_init_cumulative_args.
18022 (expand_prologue): Rename to avr_expand_prologue.
18023 (expand_epilogue): Rename to avr_expand_epilogue.
18024 (adjust_insn_length): Rename to avr_adjust_insn_length.
18025 (notice_update_cc): Rename to avr_notice_update_cc.
18026 (final_prescan_insn): Rename to avr_final_prescan_insn.
18027 * config/avr/avr.c: Same.
18028 * config/avr/avr.h: Same.
18029 * config/avr/avr.md: Remove trailing blanks.
18030 (prologue): Use avr_expand_prologue.
18031 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18032
18033 2013-01-14 Richard Biener <rguenther@suse.de>
18034
18035 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18036 verify_location, collect_subblocks): New functions.
18037 (verify_gimple_in_cfg): Verify that locations only reference
18038 BLOCKs in the functions BLOCK tree.
18039
18040 2013-01-14 Richard Biener <rguenther@suse.de>
18041
18042 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18043 PHI argument.
18044 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18045 unshare reference.
18046 (insert_out_of_ssa_copy_on_edge): Likewise.
18047 (rewrite_close_phi_out_of_ssa): Likewise.
18048 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18049 debug expressions.
18050 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18051 propagated constants.
18052 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18053 can not be shared.
18054
18055 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18056
18057 * config/avr/avr-modes.def: Add GPL copyright notice.
18058
18059 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18060
18061 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18062 MEMMODEL_MASK to determine memory model.
18063 (atomic_store<mode>): Ditto from operands[2].
18064 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18065
18066 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18067
18068 PR fortran/55935
18069 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18070 (fold_gimple_assign): Don't call unshare_expr here.
18071 (fold_ctor_reference): Call unshare_expr.
18072
18073 2013-01-13 Terry Guo <terry.guo@arm.com>
18074
18075 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18076 * doc/fragments.texi: Document MULTILIB_REUSE.
18077 * gcc.c (multilib_reuse): New internal spec.
18078 (set_multilib_dir): Also search multilib from multilib_reuse.
18079 * genmultilib (tmpmultilib3): Refactor code.
18080 (tmpmultilib4): Ditto.
18081 (multilib_reuse): New multilib argument.
18082
18083 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18084
18085 * Makefile.in: Update copyright.
18086
18087 2013-01-12 Tom de Vries <tom@codesourcery.com>
18088
18089 PR middle-end/55890
18090 * calls.c (expand_call): Check if arg_nr is valid.
18091
18092 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18093
18094 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
18095 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
18096 documentation. Add missing '__' in front of
18097 __builtin_ia32_packssdw256.
18098
18099 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18100
18101 PR target/55719
18102 * config/s390/s390.c (s390_preferred_reload_class): Do not return
18103 NO_REGS for larl operands.
18104 (s390_reload_larl_operand): Use s390_load_address instead of
18105 emit_move_insn.
18106
18107 2013-01-11 Richard Biener <rguenther@suse.de>
18108
18109 * tree-cfg.c (verify_node_sharing_1): Split out from ...
18110 (verify_node_sharing): ... here.
18111 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
18112
18113 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
18114
18115 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
18116 Substitute TREECHECKING.
18117 * configure: Regenerate.
18118 * Makefile.in (TREECHECKING): New.
18119
18120 2013-01-11 Richard Guenther <rguenther@suse.de>
18121
18122 PR tree-optimization/44061
18123 * tree-vrp.c (extract_range_basic): Compute zero as
18124 value-range for __builtin_constant_p of function parameters.
18125
18126 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
18127
18128 Update copyright years.
18129
18130 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
18131
18132 PR rtl-optimization/55672
18133 * lra-eliminations.c (mark_not_eliminable): Permit addition with
18134 const to be eliminable.
18135
18136 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
18137
18138 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
18139 * configure: Regenerate.
18140
18141 2013-01-10 Richard Biener <rguenther@suse.de>
18142
18143 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
18144
18145 2013-01-10 Richard Biener <rguenther@suse.de>
18146
18147 PR bootstrap/55792
18148 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
18149 locations for virtual PHI arguments.
18150 (rewrite_update_phi_arguments): Likewise.
18151
18152 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
18153
18154 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
18155 on to assembler.
18156
18157 2013-01-10 Jakub Jelinek <jakub@redhat.com>
18158
18159 PR tree-optimization/55921
18160 * tree-complex.c (expand_complex_asm): New function.
18161 (expand_complex_operations_1): Call it for GIMPLE_ASM.
18162
18163 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18164
18165 PR target/55718
18166 * config/s390/s390.c (s390_symref_operand_p)
18167 (s390_loadrelative_operand_p): Merge the two functions.
18168 (s390_check_qrst_address, print_operand_address): Add parameters
18169 to s390_loadrelative_operand_p invokation.
18170 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
18171 (s390_reload_larl_operand, s390_secondary_reload): Use
18172 s390_loadrelative_operand_p instead of s390_symref_operand_p.
18173 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
18174
18175 2013-01-09 Mike Stump <mikestump@comcast.net>
18176
18177 * dse.c (record_store): Remove unnecessary assert.
18178
18179 2013-01-09 Jan Hubicka <jh@suse.cz>
18180
18181 PR tree-optimization/55569
18182 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
18183 * cfgloop.h (scale_loop_profile): Likewise.
18184
18185 2013-01-09 Jan Hubicka <jh@suse.cz>
18186
18187 PR lto/45375
18188 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
18189 functions.
18190 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
18191
18192 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
18193
18194 PR middle-end/55114
18195 * expr.h (maybe_emit_group_store): Declare.
18196 * expr.c (maybe_emit_group_store): New function.
18197 * builtins.c (expand_builtin_int_roundingfn): Call it.
18198 (expand_builtin_int_roundingfn_2): Likewise.
18199
18200 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
18201
18202 PR rtl-optimization/55829
18203 * lra-constraints.c (match_reload): Add code for absent output.
18204 (curr_insn_transform): Add code for reloads of matched inputs
18205 without output.
18206
18207 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
18208
18209 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
18210 attribute of movddup insn to DF.
18211 (*vec_interleave_lowv2df): Ditto.
18212 (vec_dupv2df): Ditto.
18213
18214 2013-01-09 Jan Hubicka <jh@suse.cz>
18215
18216 PR tree-optimiation/55875
18217 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
18218 EVERY_ITERATION parameter.
18219 (number_of_iterations_exit): Check if exit is executed every iteration.
18220 (idx_infer_loop_bounds): Similarly here.
18221 (n_of_executions_at_most): Simplify
18222 to only test for cases where statement is dominated by the
18223 particular bound; handle correctly the "postdominance" test.
18224 (scev_probably_wraps_p): Use max loop iterations info
18225 as a global bound first.
18226
18227 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
18228 Nick Clifton <nickc@redhat.com>
18229
18230 * config/v850/v850.md (cbranchsf4): New pattern.
18231 (cstoresf4): New pattern.
18232 (cbranchdf4): New pattern.
18233 (cstoredf4): New pattern.
18234 (movsicc): Disallow floating point comparisons.
18235 (cmpsf_le_insn): Fix order of operators.
18236 (cmpsf_lt_insn): Likewise.
18237 (cmpsf_eq_insn): Likewise.
18238 (cmpdf_le_insn): Likewise.
18239 (cmpdf_lt_insn): Likewise.
18240 (cmpdf_eq_insn): Likewise.
18241 (cmpsf_ge_insn): Use LE comparison.
18242 (cmpdf_ge_insn): Likewise.
18243 (cmpsf_gt_insn): Use LT comparison.
18244 (cmpdf_gt_insn): Likewise.
18245 (cmpsf_ne_insn): Delete pattern.
18246 (cmpdf_ne_insn): Delete pattern.
18247 * config/v850/v850.c (v850_gen_float_compare): Use
18248 gen_cmpdf_eq_insn for NE comparison.
18249 (v850_float_z_comparison_operator)
18250 (v850_float_nz_comparison_operator): Move from here ...
18251 * config/v850/predicates.md: ... to here. Move GT and GE
18252 comparisons into v850_float_z_comparison_operator.
18253 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
18254 Delete prototype.
18255 (v850_float_nz_comparison_operator): Likewise.
18256
18257 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18258
18259 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
18260 with calls to gen_insvsi/gen_insvdi.
18261
18262 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18263
18264 * config/i386/i386.c (initial_ix86_tune_features): Set up
18265 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
18266
18267 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
18268 Jakub Jelinek <jakub@redhat.com>
18269
18270 PR tree-optimization/48189
18271 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
18272 If nitercst is 0, don't predict the exit edge.
18273
18274 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18275
18276 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
18277 in asm_fprintf with reg_names.
18278 (aarch64_print_operand_address): Likewise.
18279 (aarch64_return_addr): Likewise.
18280 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
18281
18282 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18283
18284 * config/pa/pa.h (VAL_U6_BITS_P): Define.
18285 (INT_U6_BITS): Likewise.
18286 * config/pa/predicates.md (uint6_operand): New predicate.
18287 (shift5_operand, shift6_operand): Likewise.
18288 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
18289 arith32_operand.
18290 (lshrdi3): Use shift6_operand.
18291 (shrpsi4, shrpdi4): New insn patterns.
18292 (extzv): Delete expander.
18293 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
18294 predicates in unamed zero extract patterns. Tighten common constraint.
18295 (extv): Delete expander.
18296 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
18297 predicates in unamed sign extract patterns. Tighten common constraint.
18298 (insv): Delete expander.
18299 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
18300 predicates in unamed insert patterns. Tighten common constraint.
18301 Change uint32_operand predicate to uint6_operand predicate in unamed
18302 DImode pattern to insert constant values of type 1...1xxxx.
18303
18304 2013-01-04 Jan Hubicka <jh@suse.cz>
18305
18306 PR tree-optimization/55823
18307 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
18308 issue.
18309
18310 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18311 Uros Bizjak <ubizjak@gmail.com>
18312
18313 PR rtl-optimization/55845
18314 * df-problems.c (can_move_insns_across): Stop scanning at
18315 volatile_insn_p source instruction or give up if
18316 across_from .. across_to range contains any volatile_insn_p
18317 instructions.
18318
18319 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18320
18321 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
18322 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
18323 Declare.
18324 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
18325 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
18326
18327 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18328
18329 PR fortran/55341
18330 * asan.c (asan_clear_shadow): New function.
18331 (asan_emit_stack_protection): Use it.
18332
18333 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
18334
18335 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
18336 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
18337 with tab instead of space.
18338
18339 2013-01-08 Nick Clifton <nickc@redhat.com>
18340
18341 * config/rl78/rl78.c (rl78_expand_prologue): Always select
18342 register bank 0 at the start of an interrupt handler.
18343 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
18344 MDBH registers.
18345
18346 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18347
18348 * config/aarch64/aarch64-simd.md
18349 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
18350 (aarch64_simd_bsl): Likewise.
18351 (aarch64_vcond_internal<mode>): Likewise.
18352 (vcond<mode><mode>): Likewise.
18353 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
18354 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
18355
18356 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
18357
18358 * config/aarch64/aarch64-builtins.c
18359 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
18360
18361 2013-01-08 Martin Jambor <mjambor@suse.cz>
18362
18363 PR debug/55579
18364 * tree-sra.c (analyze_access_subtree): Return true also after
18365 potentially creating a debug-only replacement.
18366
18367 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18368
18369 PR middle-end/55890
18370 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
18371
18372 PR tree-optimization/54120
18373 * tree-vrp.c (range_fits_type_p): Don't allow
18374 src_precision < precision from signed vr to unsigned_p
18375 if vr->min or vr->max is negative.
18376 (simplify_float_conversion_using_ranges): Test can_float_p
18377 against CODE_FOR_nothing.
18378
18379 2013-01-08 Jakub Jelinek <jakub@redhat.com>
18380 Richard Biener <rguenther@suse.de>
18381
18382 PR middle-end/55851
18383 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
18384 types instead of just INTEGER_TYPE types.
18385
18386 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
18387
18388 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
18389 TF_SIZE): Define.
18390
18391 2013-01-07 Steve Ellcey <sellcey@mips.com>
18392
18393 PR target/42661
18394 * config/mips/mips.opt: Change mad to mmad to match documentation.
18395
18396 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18397
18398 PR target/55897
18399 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
18400 .progmemx.data now.
18401
18402 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18403
18404 PR target/55897
18405 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
18406 (avr_addrspace_t): Add .section_name field.
18407 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
18408 array size.
18409 (avr_addrspace): Same. Initialize .section_name. Remove last
18410 NULL entry. Put __memx into .progmemx.data.
18411 (progmem_section_prefix): Remove.
18412 (avr_asm_init_sections): No need to initialize progmem_section.
18413 (avr_asm_named_section): Use avr_addrspace[].section_name to get
18414 section name prefix.
18415 (avr_asm_select_section): Ditto. And use get_unnamed_section to
18416 retrieve the progmem section.
18417 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
18418 boundary to run over avr_addrspace[].
18419 (avr_register_target_pragmas): Ditto.
18420
18421 2013-01-06 Jakub Jelinek <jakub@redhat.com>
18422
18423 * varasm.c (output_constant_def_contents): For asan_protect_global
18424 protected strings, adjust DECL_ALIGN if needed, before testing for
18425 anchored symbols.
18426 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
18427 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
18428 normal decls.
18429 (output_object_block): For asan protected decls, emit asan padding
18430 after their contents.
18431 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
18432 (asan_finish_file): Test it here instead.
18433
18434 2013-01-07 Nick Clifton <nickc@redhat.com>
18435 Matthias Klose <doko@debian.org>
18436 Doug Kwan <dougkwan@google.com>
18437 H.J. Lu <hongjiu.lu@intel.com>
18438
18439 PR driver/55470
18440 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
18441
18442 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
18443
18444 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
18445
18446 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
18447
18448 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
18449
18450 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18451
18452 PR target/54461
18453 * doc/install.texi (Cross-Compiler-Specific Options): Document
18454 --with-avrlibc.
18455
18456 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
18457
18458 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
18459 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
18460 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
18461 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
18462 vqmovun_high_s64): Fix source operand number and update copyright.
18463
18464 2013-01-07 Richard Biener <rguenther@suse.de>
18465
18466 PR middle-end/55890
18467 * gimple.h (gimple_call_builtin_p): New overload.
18468 * gimple.c (validate_call): New function.
18469 (gimple_call_builtin_p): Likewise.
18470 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18471 Use gimple_call_builtin_p.
18472 (find_func_clobbers): Likewise.
18473 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
18474 (strlen_optimize_stmt): Likewise.
18475
18476 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18477
18478 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
18479 (vld1q_dup_*): Likewise.
18480 (vld1_*): Likewise.
18481 (vld1q_*): Likewise.
18482 (vld1_lane_*): Likewise.
18483 (vld1q_lane_*): Likewise.
18484
18485 2013-01-07 Richard Biener <rguenther@suse.de>
18486
18487 * lto-streamer.h (LTO_minor_version): Bump to 2.
18488
18489 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
18490
18491 * config/aarch64/aarch64-protos.h
18492 (aarch64_const_double_zero_rtx_p): Rename to...
18493 (aarch64_float_const_zero_rtx_p): ...this.
18494 (aarch64_float_const_representable_p): New.
18495 (aarch64_output_simd_mov_immediate): Likewise.
18496 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
18497 move immediate case.
18498 * config/aarch64/aarch64.c
18499 (aarch64_const_double_zero_rtx_p): Rename to...
18500 (aarch64_float_const_zero_rtx_p): ...this.
18501 (aarch64_print_operand): Allow printing of new constants.
18502 (aarch64_valid_floating_const): New.
18503 (aarch64_legitimate_constant_p): Check for valid floating-point
18504 constants.
18505 (aarch64_simd_valid_immediate): Likewise.
18506 (aarch64_vect_float_const_representable_p): New.
18507 (aarch64_float_const_representable_p): Likewise.
18508 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
18509 (aarch64_output_simd_mov_immediate): New.
18510 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
18511 (*movdf_aarch64): Likewise.
18512 * config/aarch64/constraints.md (Ufc): New.
18513 (Y): call aarch64_float_const_zero_rtx.
18514 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
18515
18516 2013-01-07 Richard Biener <rguenther@suse.de>
18517
18518 PR tree-optimization/55888
18519 PR tree-optimization/55862
18520 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
18521 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
18522 not if it is contained therein.
18523
18524 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18525
18526 * config/avr/t-avr: Typo.
18527
18528 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
18529
18530 PR55243
18531 * config/avr/t-avr: Don't automatically rebuild
18532 $(srcdir)/config/avr/t-multilib
18533 $(srcdir)/config/avr/avr-tables.opt
18534 $(srcdir)/doc/avr-mmcu.texi
18535 (avr-mcus): New phony target to build them on request.
18536 (s-avr-mlib, s-avr-mmcu-texi): Remove.
18537 * avr/avr-mcus.def: Adjust comments.
18538
18539 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
18540
18541 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
18542
18543 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18544
18545 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
18546
18547 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
18548
18549 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
18550
18551 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
18552
18553 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
18554 to generate profiling.
18555 * config/rs6000/aix64.h (LIB_SPEC): Same.
18556
18557 2013-01-04 Andrew Pinski <apinski@cavium.com>
18558
18559 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
18560 New function.
18561 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
18562
18563 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
18564
18565 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
18566 unconditionally.
18567 (ix86_expand_move): Ditto.
18568 (ix86_zero_extend_to_Pmode): Ditto.
18569 (ix86_expand_call): Ditto.
18570 (ix86_expand_special_args_builtin): Ditto.
18571 (ix86_expand_builtin): Ditto.
18572
18573 2013-01-04 Richard Biener <rguenther@suse.de>
18574
18575 PR tree-optimization/55862
18576 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
18577 translating them through PHI nodes.
18578
18579 2013-01-04 Martin Jambor <mjambor@suse.cz>
18580
18581 PR tree-optimization/55755
18582 * tree-sra.c (sra_modify_assign): Do not check that an access has no
18583 children when trying to avoid producing a VIEW_CONVERT_EXPR.
18584
18585 2013-01-04 Marek Polacek <polacek@redhat.com>
18586
18587 PR middle-end/55859
18588 * opts.c (default_options_optimization): Clarify error message.
18589
18590 2013-01-04 Richard Biener <rguenther@suse.de>
18591
18592 PR middle-end/55863
18593 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
18594 reassociation.
18595
18596 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18597
18598 PR target/53789
18599 * config/pa/pa.md (movsi): Revert previous change.
18600 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
18601 references.
18602
18603 2013-01-03 Richard Henderson <rth@redhat.com>
18604
18605 * config/i386/i386.c (ix86_expand_move): Always assign to op1
18606 after eliminating TLS symbols.
18607
18608 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
18609
18610 PR bootstrap/50167
18611 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
18612 * graphite-poly.c (debug_gmp_value): Likewise.
18613
18614 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
18615
18616 PR target/55712
18617 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
18618 selected code model, define __code_mode_small__, __code_model_medium__,
18619 __code_model_large__, __code_model_32__ or __code_model_kernel__.
18620 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
18621 xchg temporary register with %k. Declare temporary register as
18622 early clobbered.
18623 [__x86_64__]: For medium and large code models, preserve %rbx register.
18624
18625 2013-01-03 Richard Biener <rguenther@suse.de>
18626
18627 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
18628 (dump_subscript): Adjust.
18629 (finalize_ddr_dependent): Do not dump redundant info.
18630 (analyze_siv_subscript): Adjust.
18631 (subscript_dependence_tester): Likewise.
18632 (compute_affine_dependence): Likewise.
18633
18634 2013-01-03 Richard Biener <rguenther@suse.de>
18635
18636 Revert
18637 2013-01-03 Richard Biener <rguenther@suse.de>
18638
18639 PR tree-optimization/55857
18640 * tree-vect-stmts.c (vectorizable_load): Do not setup
18641 re-alignment for invariant loads.
18642
18643 2013-01-02 Richard Biener <rguenther@suse.de>
18644
18645 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18646 invariant load do not generate a vector load from the scalar location.
18647
18648 2013-01-03 Richard Biener <rguenther@suse.de>
18649
18650 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
18651 for not vectorizing.
18652 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
18653 not build INDIRECT_REFs, call get_name once only.
18654 (vect_create_data_ref_ptr): Likewise. Dump base object kind
18655 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
18656
18657 2013-01-03 Richard Biener <rguenther@suse.de>
18658
18659 PR tree-optimization/55857
18660 * tree-vect-stmts.c (vectorizable_load): Do not setup
18661 re-alignment for invariant loads.
18662
18663 2013-01-03 Richard Biener <rguenther@suse.de>
18664
18665 PR lto/55848
18666 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
18667 prefer a built-in decl.
18668
18669 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18670
18671 * gcc.c (process_command): Update copyright notice dates.
18672 * gcov.c (print_version): Likewise.
18673 * gcov-dump.c (print_version): Likewise.
18674
18675 PR rtl-optimization/55838
18676 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
18677 iv0.step, iv1.step and step.
18678
18679 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18680 Marc Glisse <marc.glisse@inria.fr>
18681
18682 PR tree-optimization/55832
18683 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
18684 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
18685 integer_{one,zero}_node.
18686
18687 2013-01-03 Jakub Jelinek <jakub@redhat.com>
18688
18689 PR debug/54402
18690 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
18691 * var-tracking.c (reverse_op): Don't add reverse ops to
18692 VALUEs that have already
18693 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
18694
18695 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
18696
18697 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
18698
18699 2013-01-02 Teresa Johnson <tejohnson@google.com>
18700
18701 * dumpfile.c (dump_loc): Print filename with location.
18702 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
18703 new location_t parameter to emit complete unroll message with
18704 new dump framework.
18705 (canonicalize_loop_induction_variables): Compute loops location
18706 and pass to try_unroll_loop_completely.
18707 * loop-unroll.c (report_unroll_peel): New function.
18708 (peel_loops_completely): Use new dump format with location
18709 for main dumpfile message, and invoke report_unroll_peel on success.
18710 (decide_unrolling_and_peeling): Ditto.
18711 (decide_peel_once_rolling): Remove old dumpfile message subsumed
18712 by report_unroll_peel.
18713 (decide_peel_completely): Ditto.
18714 (decide_unroll_constant_iterations): Ditto.
18715 (decide_unroll_runtime_iterations): Ditto.
18716 (decide_peel_simple): Ditto.
18717 (decide_unroll_stupid): Ditto.
18718 * cfgloop.c (get_loop_location): New function.
18719 * cfgloop.h (get_loop_location): Declare.
18720
18721 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18722
18723 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
18724 NULL.
18725
18726 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18727
18728 PR middle-end/55198
18729 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
18730 BLKmode objects when EXPAND_MEMORY is specified.
18731
18732 2013-01-02 Sriraman Tallam <tmsriram@google.com>
18733
18734 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
18735 in loop predicate.
18736 (fold_builtin_cpu): Do not share cpu model decls across statements.
18737
18738 2013-01-02 Jason Merrill <jason@redhat.com>
18739
18740 PR c++/55804
18741 * tree.c (build_array_type_1): Revert earlier change.
18742
18743 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
18744
18745 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
18746 "cortex-a57".
18747 * config/aarch64/aarch64-tune.md: Re-generate.
18748
18749 2013-01-02 Richard Biener <rguenther@suse.de>
18750
18751 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
18752 invariant load do not generate a vector load from the scalar location.
18753
18754 2013-01-02 Richard Biener <rguenther@suse.de>
18755
18756 PR bootstrap/55784
18757 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
18758 * configure: Regenerate.
18759
18760 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18761
18762 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
18763 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
18764 (expand_builtin_int_roundingfn_2): Keep the original target around
18765 for the fallback case.
18766
18767 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
18768
18769 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
18770 to be clear for sign changes.
18771
18772 2013-01-01 Jan Hubicka <jh@suse.cz>
18773
18774 * ipa-inline-analysis.c: Fix formatting.
18775
18776 2013-01-01 Jakub Jelinek <jakub@redhat.com>
18777
18778 PR tree-optimization/55831
18779 * tree-vect-loop.c (get_initial_def_for_induction): Use
18780 gsi_after_labels instead of gsi_start_bb.
18781 \f
18782 Copyright (C) 2013 Free Software Foundation, Inc.
18783
18784 Copying and distribution of this file, with or without modification,
18785 are permitted in any medium without royalty provided the copyright
18786 notice and this notice are preserved.